Eastern Mediterranean Health Journal [2005; Vol.11, Issue 3]

...., jJ.I J .,.:J �I 4,U?I ..::..,L...t..-]1 �ill � ,./' J .9WI t,...,.Ji �- .1.... _pl J _r-l �'ii �I if J.J..,:,; JI � )I �I �I ,./' .!1.J� _;);-J ..:.,�\JI �l:i J aJ4_,ll ..::..,\.,,k,J.IJ i--"'l..il.lJ .-IJ \II J>� J ,lA �J_;JIJ �I ..::..,L.,..ij.l J 0-4..l.:l,.I dJ>W.IJ �I ..::..,L.$JIJ ,�I �1 .. �i JS" J! <.f>."J-" ,./'J . .1...._pl J.r- �� lf.:.. � \..., a..,.1>-J ,..::..,L.,},J.I ,:r �I >l..,.i:/IJ t.._,lWI �I � C;' ..,; Jl.dl _.?l}IJ ,�I y ..f.J-1 _;);- ..::..,l.Ja;.il 1.15'" J ,�I .lAWI J\.... J -">--JI>-) �'ii J t.....J4 EASTER"I MEDITERRANEAN HEALTH JOURNAL IS the official health journal published by the Eastern Mediterranean Regional Office of the World Health Organization. It is a forum for the presentation and promotion of new policies and initiatives in health services; and for the exchange of ideas, concepts, epidemiological data, research findings and other information, with special reference to the Eastern Mediterranean Region. It addresses all members of the health profession, medical and other health educational institutes, interested NGOs, WHO Collaborating Centres and individuals within and outside the Region. LA REVUE DE SANTE DE LA MEDITERRANEE ORIENTA LE EST une revue de sante officielle publiee par le Bureau regional de !'Organisation mondiale de la Sante pour la Mediterranee orientale. Elle offre une tribune pour la presentation et la promotion de nouvelles politiques et initiatives dans le domaine des services de sante ainsi qu'a l'echange d'idees, de concepts, de donnees epidemiologiques, de resultats de recherches et d'autres informations, se rapportant plus particulierement a la Region de la Mediterranee orientale. Elle s'adresse a tous les professionnels de la sante, aux membres des instituts medicaux et autres instituts de formation medico-sanitaire, aux ONG, Centres collaborateurs de l'OMS et personnes concernes au sein et hors de la Region. The Eastern Mediterranean Health Journal is abstracted/indexed in the Index Medicus and MEDLIN£ (Medical literature Analysis and Retrieval Systems on line) and the ExtraMed-Fu/1 text on CD-ROM, the Cumulative Index to Nursing and ;'/lied Health literature (CJNAHL), CAB International, Lexis Nexis™ and the Index Medicus for the WHO Eastern Mediterranean Region (IMEMR). EMHJ is also available on the World Wide Web: http://www/emro.who.int/emhj.htm ALL ARTICLES ARE PEER REVIEWED ISSN 1020-3397 © WORLD HEALTH ORGA"IIZATION, 2005 Publications of the World Health Organization enjoy copyright protection in accordance with the provisions of Protocol 2 of the Universal Copyright Convention. All rights reserved. The designations employed and the presentation of the material in this publication do not imply the expression of any opinion whatsoever on the part of the Secretariat of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitations of its frontiers or boundaries. The mention of specific companies or of certain manufacturers' products does not imply that they are endorsed or recommended by the World Health Organization in preference to others of a similar nature that are not mentioned. The authors alone are responsible for the views expressed, which do not necessarily reflect the opinion of the World Health Organization or of its Member States. COVER: Remote sensing image of the area of the world that includes the Eastern Mediterranean Region. Cover image produced by the Canada Centre for Remote Sensing, Ottawa, Canada. (NOAA AVHRR Composite). Reproduced with permission. Cover designed by Ahmed Hassanein Printed on acid-free paper Printed by League of Arab States Printshop

258 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Coronary heart disease risk-factor profile in a lower middles class urban community in Pakistan K. Aziz1, S. Aziz2, Najma Patel,1 A.M.A. Faruqui1 and H. Chagani1 1Department of Cardiology, National Institute of Cardiovascular Diseases, Karachi, Pakistan (Correspondence to K. Aziz: nicvdedo@khi.comsats.net.pk). 2Pakistan Medical Research Council, Sindh Branch, Karachi, Pakistan. Received: 14/04/03; accepted: 29/03/04 ABSTRACT We determined the risk-factor profile and prevalence of coronary heart disease in Metroville, a lower middle class urban community in Karachi, and compared them to the Pakistan health survey PNHS 1990–94, and the US health and nutrition survey 1988–94 NHANES111. Subjects < 18 years and pregnant women were excluded as were people with extreme ranges SMI, heart rate, height and waist. The preva- lence of hypertension was 23% in men and women, hypercholesterolaemia was 17% in men and 22% in women (P < 0.001). Hyperglycaemia was present in 5% of men and women and obesity in 33% of men and 47% of women (P < 0.001). Compared to PNHS, the prevalences of obesity, hypertension, hypercholesterol- aemia and WHR were higher in our population. Mean values of RMI, cholesterol, WHR were higher in the US population while mean values were lower for diastolic blood pressure and blood glucose. Le profil des facteurs de risque de coronaropathies dans une communauté urbaine de classe moyenne inférieure au Pakistan RÉSUMÉ Nous avons déterminé le profil des facteurs de risque et la prévalence des coronaropathies à Metroville, communauté urbaine de classe moyenne inférieure à Karachi, et les avons comparést à l’enquête nationale de santé réalisée au Pakistan de 1990 à 1994, enquête et á l’ d’examen de la santé et de la nutrition effectuée de 1988-1994 aux États-Unis NHANES III. Les sujets ayant moins de 18 ans et les femmes enceintes ont été exclus ainsi que les fourchettes extrêmes pour l’IMC, la fréquence cardiaque, la taille et le tour de taille. La prévalence de l’hypertension s’élevait à 23 % chez les hommes et les femmes, l’hypercholestérolémie était de 17 % chez les hommes et de 22 % chez les femmes (p < 0,001). Une hyperglycémie était présente chez 5 % des hommes et des femmes et l’obésité chez 33 % des hommes et 47 % des femmes (p < 0,001). Par rapport à l’enquête nationale sur la santé au Pakistan, la prévalence de l’obésité, de l’hypertension, de l’hypercholestérolémie ainsi que le rapport tour de taille/tour de hanches étaient plus élevés dans notre population. Les valeurs moyennes pour l’IMC, le cholestérol et le rapport tour de taille/tour de hanches étaient supérieures dans la population nord-américaine tandis qu’elles étaient inférieures pour la pression artérielle diastolique et la glycémie. 02 Coronary heart disease.pmd 12/20/2005, 2:34 PM258 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 259 Introduction The population in Pakistan, according to the 1998 Federal census, was 132.4 million and growing at the rate of 2.5% while the rate of growth of the urban population was 3.5%. At this rate the total population was expected to be 143.1 million in 2001 with an urban share of 47.7 million, representing 33% of the total population in Pakistan [1]. Pakistan is a developing country in which urbanization of a socioeconomically de- pressed rural population has resulted in ad- mixture of ever increasing lower middle class urban communities with much small- er rich and high middle class communities. This epidemiological transition has resulted in a non-homogeneous population and has imposed on our developing country a dou- ble burden of high risk factors such as obe- sity and hypercholesterolemia and mal- nutrition and hypocholesterolaemia [2]. There is evidence to suggest that over the past 3 decades in Pakistan the incidence of coronary heart disease (CHD) has in- creased [2–15]. It has been observed that CHD occurs at an earlier age in the Paki- stani population and the gender gap is nar- rower compared to women who reside in the more developed countries [15]. The ef- fect of urbanization is apparent in the stud- ies of immigrant Asians from the Indian subcontinent in Western Europe and North America in whom the incidence and sever- ity of CHD is higher than the native popula- tion. Asian women in particular have severe forms of CHD [17–20]. In Pakistan chang- ing lifestyles of lower middle class urban communities may be the operative factor in the observed increase in the incidence of CHD in Pakistan. The etiology of the CHD is not known, however the risk factors have a strong statistical correlation with the development of CHD and it has been shown that the modification of the risk fac- tors leads to a reduction in this risk. Re- cently a long-term population-based study has shown a strong correlation of the inci- dence of CHD with prevalence of risk fac- tors indicating that majority of the incidence of CHD in a population could be explained by the prevalence of the risk fac- tors [21]. Thus by using data on the preva- lence of risk factors in a community, it is possible to project the future burden of CHD in a population. A community-based population study in Pakistan was undertaken by the National Institute of Cardiovascular Disease during 1967–68. Two communities, one semi- rural and the other urban, were studied but the targeted communities were not repre- sentative of the entire Pakistani population [22]. The overall prevalence of ischemic heart diseases in men was 2.3% in the rural and 4.7% in the urban populations and cor- responding values for women over 25 years were 1.1% and 2.0%. The preva- lence of hypertension was 16.3% in men and 20.4% in women in the rural communi- ty and 15.9% in men and 16.7% in women in the urban community [22]. A recent study of 3 urban communities around Karachi has shown that the prevalence of hypertension was 19.9% in men over 18 years and 20.9% in women so it appears that the prevalence has risen [14]. Compar- ison of hospital-based studies shows that the incidence of CHD among hospital ad- missions with heart ailments has increased dramatically from 1.7% in 1944–48 to 17.9% in 1958 to 41.8% in 1967 to 75.8% in men and 51.8% in women in 1981 [11]. The demographic and risk factor profile of a homogeneous population can be used to plan national preventive interventions but this cannot be done for the non-homoge- neous population such as Pakistan. As in- dustrialization proceeds in Pakistan, the rural communities will increasingly become urbanized and with employment their pur- 02 Coronary heart disease.pmd 12/20/2005, 2:34 PM259 260 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 chasing power will increase which may ad- versely change their lifestyles. This change, we hypothesize, can be detected by comparing the risk-factor profile of the entire nation against the community under change. The purpose our study was to present the demographics of Metroville, a low mid- dle class urban community in Paksitan, and profile the risk factors of CHD and their prevalence rates. We also aimed to deter- mine if the degree of increase in the risk- factor profile and prevalence rates in the urban community could be measured by comparing it with the Pakistan National Health Survey (PNHS) [2] and also a highly urbanized US population of the Third Na- tional Health and Nutrition Examination Survey 1988–1994 (NHANES111) [23]. Methods Metroville is an urban community located 20 miles from down-town Karachi. Its eth- nic groups comprise mainly Pathan and Punjabi people with some Sindis and oth- ers. It is in large measure composed of mi- grants from rural villages from all 4 provinces of Pakistan. The Metroville Health Study (MHS) was a US–Pakistan cooperative prospective study involving the National Institute of Cardiovascular Dis- eases in Karachi, Pakistan, the National Heart Lung and Blood Institute (NHLBI ), Bethesda, Maryland and the University of North Carolina at Chapel Hill. It was de- signed to evaluate the efficacy of dietary intervention in the community [24]. Data were obtained at baseline examination of 399 (of 4296) Metroville households who responded to a letter of invitation to partic- ipate in the study and who volunteered for registration and were randomly assigned as treatment and control groups. The demographic data included sex, age, ethnicity, profession, and household ownership and composition. A question- naire was designed to evaluate awareness of cardiovascular diseases (CVD), includ- ing history of heart attacks, stroke, diabe- tes, hypertension, and tobacco smoking, chewing or sniffing. Educational level was determined by recording no formal educa- tion, primary school (4 years), high school (10 years) and college education (12 years). Physical examination included height (cm), weight (kg) and arm, waist and hip circumferences (cm). Blood pres- sure (mmHg) and heart rate were deter- mined in the right arm in a sitting position; the diastolic pressure was taken at the dis- appearance of the 5th Korotkoff sound (K5), and K4 was also recorded. Blood glu- cose and serum cholesterol levels were de- termined at random times after meals using the Reflotron dry strip method. The cali- bration was done against the standard strips provided by Boeringer–Manheim Company. Specially trained doctors and laboratory technicians collected the demo- graphic data, carried out the physical ex- amination and made the chemical analyses. The health questionnaire was completed by the doctors and included history of heart attack, stroke, diabetes, hypertension and smoking chewing sniffing or smoking to- bacco. Awareness about CHD was evaluat- ed by questions regarding specifically the role of hypertension, smoking and obesity on the occurrence of heart attacks [23]. Subjects 18 years or older were eligible for inclusion in the MHS analysis. Only pre-intervention risk factors and relevant demographic data at baseline examination of MHS were included. Pregnant women and those under 18 years were excluded from the analysis. Also excluded from the analysis were those with measurements of 02 Coronary heart disease.pmd 12/20/2005, 2:34 PM260 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 261 body mass index (BMI) beyond the range of 14 and 50 kg/m2, heart rate beyond 30 and 140 beats/minute range, height beyond 102 and 200 cm range, and waist circum- ference beyond 27 and 145 cm range. Statistical analysis The MHS, PNHS and NHANES111 data were analysed in age groups of 18–19, 20– 29, 30–39, 40–49, 50–59, 60–69 and ≥ 70 years. The percentages, means and stan- dard deviations (SD) were calculated for various risk factors. Comparisons of risk factors were made by calculating age-ad- justed means and the 25th, 50th, 75th and 90th percentile values for MHS, PNHS and NHANES111. Group means were com- pared by using the Student t-test and P-val- ues < 0.05 were considered significant. Results Demographic data The mean number of families per house- hold was 1.22 (range 1–5), and the mean number of adults and children per house- hold was 3.93 (range 1–12) and 4.17 (range 0–19) respectively. There were 2171 subjects included in the study, 1093 were children under 18 years and 1078 were adults, with adult to children ratio of 0.99. There were 1119 males and 1052 were females with male to female ratio of 1.06. Mean monthly income of the house- holds was 8321.5 Pakistani rupees (range 1500–90 000) (US$ 1 = 46 rupees at the time of the study). The ethnic distribution comprised 33.4% Punjabis, 45.3% Pathan 8.2 % Muhajir and 13.1% others. For 18 participants, ethnicity was not recorded. Job and occupation distribution showed 10.8% self-employed, 5.6% government employees, 1.7% labourers, 36.4% stu- dents (school and college) and 21.5% were housewives. Pre-school children com- prised 6.3% of the sample, 1.6 % were re- tired, 0.7% drivers and 8.2% had other jobs; in 7.3% the job was unknown. Thus out of 2171 participants, 50.0% (1086) in- cluding housewives were employed so that if one were to exclude the 1093 children most employable persons did have some employment. The medical history questionnaire of 946 adults showed that CVD was reported in 27.1%, diabetes in 7.3% (5.0% on med- ication), stroke in 2.0%, heart attack in 4.7% and hypertension in 20.4% (10.3% on medication). Blood pressure ≥ 140/90 mmHg was noted in 15.7% and blood pres- sure >160/95 mmHg in 11.6%. There were 463 men and 483 women >18 years who responded to the questionnaire on smoking. Of these, 21.8% of men and 1.0% of wom- en smoked tobacco, 11.7% of men and 4.6% of women chewed tobacco and 3.2% of men and 0.8% of women sniffed tobac- co. Education level showed that 27.5% were illiterate, 27.9% had primary educa- tion (4 years of schooling), 12.4% had middle school education and school-leaving certificate (10 years of schooling) and 15.5% had attended college. From the gen- der perspective, 40% of the women were illiterate while only 15% of the men were. In addition, 22.8% of the men had a college education compared with only 8.5% of the women. Risk factors profile Body mass index The body mass index (BMI) profile of MHS and PNHS is presented in Tables 1 and 2. BMI increased with age in both sex- es until the age of 70 years after which it fell slightly. The overall mean BMI in MHS men was 23.45 (SD 4.74) kg/m2 and 25.22 (SD 5.84) kg/m2 in women, women being 02 Coronary heart disease.pmd 12/20/2005, 2:34 PM261 262 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Ta bl e 1 C o ro n ar y h ea rt d is ea se r is k- fa ct o r p ro fi le in M et ro vi lle H ea lt h S tu dy (M H S ) A g e g ro u p B M I ( kg /m 2 ) W H R S B P (m m H g ) D B P (m m H g ) T C ( m g /d L ) G lu co se (m g /d L ) (y ea rs ) M ea n S D M ea n S D M ea n S D M ea n S D M ea n S D M ea n S D M en 1 8– 19 19 .6 8 3. 91 0. 82 1. 07 11 4. 12 11 .8 2 72 .3 1 7. 70 12 7. 90 19 .9 9 91 .5 1 21 .6 3 (n = 3 9) (n = 4 2) (n = 4 2) (n = 4 2) (n = 4 1) (n = 4 1) 20 –2 9 21 .7 9 3. 87 0. 87 0. 06 11 5. 40 10 .9 8 77 .2 0 10 .0 7 14 9. 11 35 .8 6 97 .7 7 19 .8 9 (n = 1 14 ) (n = 1 14 ) (n = 1 13 ) (n = 1 13 ) (n = 1 08 ) (n = 1 08 ) 30 –3 9 23 .2 9 4. 19 0. 93 0. 10 11 5. 96 12 .3 7 79 .4 9 9. 58 17 2. 21 36 .8 2 10 5. 38 28 .6 1 (n = 1 12 ) (n = 1 13 ) (n = 1 13 ) (n = 1 13 ) (n = 1 07 ) (n = 1 08 ) 40 –4 9 25 .2 9 4. 23 0. 99 0. 19 12 3. 06 15 .7 3 85 .3 5 11 .0 1 18 1. 31 47 .5 2 11 1. 54 45 .7 (n = 1 26 ) (n = 1 28 ) (n = 1 27 ) (n = 1 27 ) (n = 1 23 ) (n = 1 22 ) 50 –5 9 25 .1 5 4. 06 0. 98 0. 09 13 2. 69 20 .9 2 84 .9 4 11 .6 6 17 3. 80 45 .3 3 12 2. 46 44 .1 6 (n = 5 ) (n = 5 9) (n = 5 9) (n = 5 9) (n = 5 4) (n = 5 4) 60 –6 9 24 .6 8 8. 56 0. 95 0. 17 13 8. 74 19 .7 8 86 .3 9 14 .8 6 17 7. 85 31 .3 6 10 1. 89 18 .1 6 (n = 2 3) (n = 2 3) (n = 2 3) (n = 2 3) (n = 2 0) (n = 1 9) ≥7 0 23 .2 0 5. 49 0. 96 0. 23 13 4. 31 22 .9 8 81 .7 7 11 .1 6 17 3. 92 38 .2 4 14 0. 67 5 7. 65 (n = 1 3) (n = 1 3) (n = 1 3) (n = 1 3) (n = 1 2) ( n = 1 2) A ll 23 .4 5 4. 74 0. 93 0. 14 12 1. 08 10 .5 8 80 .9 1 11 .3 2 16 5. 82 42 .5 6 10 6. 15 35 .8 8 (n = 4 85 ) (n = 4 92 ) (n = 4 90 ) (n = 4 90 ) (n = 4 65 ) (n = 4 64 ) W om en 18 –1 9 20 .5 3 3. 18 0. 77 0. 06 11 1. 92 10 .4 8 71 .5 0 10 .3 9 15 5. 73 31 .8 3 96 .8 4 16 .9 0 (n = 3 1) (n = 3 3) (n = 2 3) (n = 3 3) (n = 3 0) (n = 3 2) 20 –2 9 22 .9 8 5. 05 0. 08 0. 07 11 0. 87 12 .2 1 73 .3 4 10 .3 5 15 2. 73 38 .9 8 96 .7 8 19 .2 9 (n = 1 21 ) (n = 1 29 ) (n = 1 28 ) (n = 1 28 ) (n = 1 22 ) (n = 1 23 ) 30 –3 9 25 .8 5 5. 45 0. 85 0. 07 11 7. 10 16 .5 6 77 .9 1 11 .0 6 17 5. 98 46 .2 3 10 1. 47 28 .3 4 (n = 1 37 ) (n = 1 37 ) (n = 1 34 ) (n = 1 34 ) (n = 1 31 ) (n = 1 31 ) 40 –4 9 27 .6 4 6. 31 0. 88 0. 09 12 6. 55 18 .1 4 83 .8 0 1 2. 25 18 7. 80 41 .4 8 11 2. 77 43 .9 5 (n = 1 00 ) (n = 1 00 ) (n = 9 9) (n = 9 9) (n = 9 2) (n = 9 0) 50 –5 9 27 .1 0 5. 97 0. 89 0. 07 13 2. 69 28 .4 0 82 .5 2 12 .6 9 19 2. 21 54 .3 6 12 8. 67 57 .3 4 (n = 4 5) (n = 4 5) (n = 4 4) (n = 4 4) (n = 4 3) (n = 4 3) 60 –6 9 26 .4 3 5. 03 0. 90 0. 07 14 4. 47 22 .4 4 84 .9 0 11 .9 5 18 7. 21 37 .3 6 12 8. 36 41 .8 6 (n = 2 9) (n = 2 9) (n = 2 9) (n = 2 9) (n = 2 8) (n = 2 8) 02 Coronary heart disease.pmd 12/20/2005, 2:34 PM262 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 263 more overweight than men (P < 0.0001) The BMI in PNHS also increased with age in both sexes and the overall BMI in men was 20.81 (SD 3.81) kg/m2 compared to 21.47 (SD 4.45) kg/m2 in women who were again more overweight than men (P < 0.0001) (Table 2). The comparison of overall BMI in MHS and PNHS showed that both women and men in MHS were more overweight (P < 0.0001). In MHS overweight, defined as BMI > 25 kg/m2 [25], was present in 33% of men compared to 12% of men in PNHS. In MHS 47% women were obese as against 20% of women in PNHS (P < 0.0001) (Ta- ble 3). Women were more overweight than men in both PNHS and MHS (P < 0.001). In both MHS and PNHS the prevalence of overweight in men increased with age. In women in MHS the trend was similar al- though the prevalence was much higher. Waist-to-hip ratio The overall mean waist-to-hip ratio (WHR) in men in MHS was 0.93 (SD 0.14) which was significantly higher than in women 0.85 (SD 0.08) (P < 0.0001) while in PNHS it was similar in both men and wom- en. In MHS men the WHR increased with age and then fell slightly after 70 years (Ta- ble 1). The same trend was noted in wom- en (P < 0.0001). In PNHS the WHR increased with age in both men and wom- en. WHR ratio was significantly greater in MHS men than PNHS men (P < 0.001) and in PNHS women than MHS women (P < 0.001) (Table 3). In PNHS 15% of men exceeded WHR = 0.96 [25] while in MHS 34% exceeded this ratio (P < 0.0001). In MHS men the preva- lence of men with WHR > 0.96 was 2% in the 18–19-year age group and increased to 77% in those >70 years. This trend was also found in PNHS men although the rate was much lower.Ta bl e 1 C o ro n ar y h ea rt d is ea se r is k- fa ct o r p ro fi le in M et ro vi lle H ea lt h S tu dy (M H S ) ( co nc lu de d) A g e g ro u p B M I ( kg /m 2 ) W H R S B P (m m H g ) D B P (m m H g ) T C ( m g /d L ) G lu co se (m g /d L ) (y ea rs ) M ea n S D M ea n S D M ea n S D M ea n S D M ea n S D M ea n S D ≥ 70 23 .8 3 4. 87 0. 89 0. 07 12 3. 9 2 1. 21 71 .8 7 11 .9 9 17 7. 79 48 .9 3 12 5. 33 74 .4 9 (n = 1 5) (n = 1 5) (n = 1 5) (n = 1 5) (n = 1 4) (n = 1 5) A ll 25 .2 2 5. 84 0. 85 0. 08 12 0. 31 19 .8 3 78 .1 2 12 .1 8 17 3. 16 45 .3 2 10 7. 04 37 .6 9 (n = 4 89 ) (n = 4 88 ) (n = 4 82 ) (n = 4 82 ) (n = 4 69 ) (n = 4 62 ) B M I = b od y m as s in de x. W H R = w ai st h ip r at io . S B P = s ys to lic b lo od p re ss ur e. D B P = d ia st ol ic b lo od p re ss ur e. T C = t ot al c ho le st er ol . S D = s ta nd ar d de vi at io n. 02 Coronary heart disease.pmd 12/20/2005, 2:34 PM263 264 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Ta bl e 2 C o ro n ar y h ea rt d is ea se r is k- fa ct o r p ro fi le in th e P ak is ta n N at io n al H ea lt h S u rv ey (P N H S ) A g e g ro u p B M I ( kg /m 2 ) W H R S B P (m m H g ) D B P (m m H g ) T C ( m g /d L ) G lu co se (m g /d L ) (y ea rs ) M ea n S D M ea n S D M ea n S D M ea n S D M ea n S D M ea n S D M en 1 8– 19 19 .4 3 3. 46 0. 84 0. 04 11 6. 25 11 .1 0 77 .9 8 8. 67 13 2. 3 28 .8 7 92 .0 6 15 .4 5 (n = 2 64 ) (n = 2 70 ) (n = 2 71 ) (n = 2 71 ) (n = 2 60 ) (n = 2 62 ) 20 –2 9 20 .6 8 3. 59 0. 87 0. 11 11 7. 04 11 .1 4 79 .7 9. 12 14 3. 96 35 .5 5 92 .3 5 18 .6 9 (n = 8 18 ) (n = 8 45 ) (n = 8 45 ) (n = 8 48 ) (n = 8 20 ) (n = 8 27 ) 30 –3 9 21 .1 2 3. 47 0. 89 0. 06 11 5. 35 11 .4 8 81 .5 1 10 .5 1 15 7. 08 37 .4 8 95 .0 3 29 .5 8 (n = 3 2) (n = 6 53 ) (n = 6 57 ) (n = 6 57 ) (n = 6 34 ) (n = 6 39 ) 40 –4 9 21 .7 9 4. 25 0. 91 0. 07 11 8. 36 14 .3 5 82 .8 7 9. 92 16 2. 85 44 .2 2 10 0. 86 30 .6 3 (n = 4 57 ) (n = 4 76 ) (n = 4 76 ) (n = 4 76 ) (n = 4 57 ) (n = 4 62 ) 50 –5 9 20 .7 4 3. 58 0. 92 0. 07 12 1. 88 18 .3 8 83 .2 9 10 .8 2 15 9. 89 39 .7 1 11 0. 13 60 .8 0 (n = 3 68 ) (n = 3 78 ) (n = 3 80 ) ( n = 3 80 ) (n = 3 67 ) (n = 3 68 ) 60 –6 9 21 .0 7 5. 05 0. 91 0. 08 12 6. 54 21 .2 7 81 .5 1 12 .1 2 15 7. 16 37 .2 3 10 8. 93 45 .5 7 (n = 3 03 ) (n = 3 04 ) (n = 3 10 ) (n = 3 10 ) (n = 3 00 ) (n = 3 04 ) 70 + 19 .1 2 3. 51 0. 90 0. 08 12 7. 84 2 5. 29 78 .7 1 12 .3 1 15 4. 6 37 .4 2 10 9. 91 52 .1 2 (n = 2 10 ) (n = 2 21 ) (n = 2 31 ) ( n = 2 31 ) ( n = 2 14 ) (n = 2 17 ) A ll 20 .8 1 3. 81 0. 89 0. 08 11 9. 11 15 .7 9 80 .9 5 10 .3 3 15 2. 38 38 .7 3 99 .1 6 36 .3 9 (n = 3 05 2) (n = 3 19 1) (n = 3 17 2) (n = 3 11 3) (n = 3 05 2) (n = 3 07 9) W om en 18 –1 9 19 .4 2 2. 87 0. 85 0. 08 11 3. 28 11 .3 4 73 .3 8 7. 96 14 6. 12 32 .9 0 95 .5 3 13 .9 0 (n = 2 64 ) (n = 2 39 ) (n = 2 41 ) (n = 2 41 ) (n = 2 32 ) (n = 2 36 ) 20 –2 9 20 .3 1 3. 87 0. 87 0. 07 11 1. 94 11 .6 0 73 .7 9 8. 36 15 0. 56 33 .9 9 96 .4 8 23 .4 8 (n = 8 18 ) (n = 8 46 ) (n = 8 52 ) (n = 8 52 ) (n = 8 18 ) (n = 8 32 ) 30 –3 9 22 .3 1 5. 22 0. 9 0. 08 11 5. 67 15 .0 1 77 .1 6 10 .0 9 15 7. 78 38 .0 2 10 0. 73 29 .8 (n = 6 32 ) (n = 6 73 ) (n = 6 74 ) (n = 6 74 ) (n = 6 46 ) (n = 6 51 ) 40 –4 9 22 .8 4 5. 33 0. 91 0. 09 12 1. 10 20 .2 3 79 .3 2 11 .8 5 16 3. 46 38 .6 2 10 5. 32 41 .7 3 (n = 4 57 ) (n = 5 39 ) (n = 5 53 ) (n = 5 53 ) (n = 5 27 ) (n = 5 30 ) 50 –5 9 22 .5 6 5. 29 0. 92 0. 09 12 7. 74 24 .8 5 81 .6 12 .9 4 17 1. 05 42 .2 2 11 3. 53 45 .9 4 (n = 3 68 ) (n = 3 80 ) (n = 3 83 ) (n = 3 83 ) (n = 3 64 ) (n = 3 69 ) 60 –6 9 21 .3 7 4. 96 0. 92 0. 08 13 2. 78 24 .0 2 81 .3 4 13 .4 6 17 3 46 .8 2 11 5. 28 57 .8 (n = 3 03 ) (n = 2 65 ) (n = 2 69 ) (n = 2 69 ) (n = 2 47 ) (n = 2 53 ) 02 Coronary heart disease.pmd 12/20/2005, 2:34 PM264 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 265 Systolic blood pressure The mean systolic blood pressure (SBP) in MHS men was 121.08 (SD 10.58) mmHg and 120.31 (SD 19.83) mmHg in women while in PNHS men the SBP was 119.11 (SD 15.79) mmHg and 119.4 (SD 19.38) mm Hg in women. The SBP in men and women was not significantly different in both studies. SBP was higher in MHS men than PNHS men (P < 0.01) but was not dif- ferent in women. The SBP increased with age in both sexes in both studies until the age of 70 years when it fell. Diastolic blood pressure The mean diastolic blood pressure (DBP) increased with age in MHS as well as PNHS until 70 years when it tended to drop in both men and women (Tables 1 and 2). The mean DBP in MHS men was 80.91 (SD 11.32) mmHg compared to 78.12 (SD 12.18) mmHg in women (P < 0.0001). In PNHS the mean DBP was also higher in men [80.95 (SD 10.31) mmHg] than in women [77.42 (SD 11.10) mmHg] in women (P < 0.0001). Comparison of DBP in MHS and PNHS men showed no signifi- cant difference. Hypertension was defined as SBP or DBP or both ≥ 140/90 mmHg [26]. The overall prevalence rate of hypertension was 18% in PNHS men and 23% in MHS men (P = 0.008) while 23% of MHS women had hypertension compared to 15% of PNHS women (P < 0.0001) (Table 3). In MHS and PNHS men and women there was a steady increase in the prevalence of hyper- tension with age. Serum cholesterol The mean total serum cholesterol in MHS men was 165.82 (SD 42.56) mg/dL and was higher in women 173.16 (SD 45.32) mg/dL (P < 0.011). In PNHS men the total cholesterol was 152.38 (SD 38.7) mg/dLTa bl e 2 C o ro n ar y h ea rt d is ea se r is k- fa ct o r p ro fi le in th e P ak is ta n N at io n al H ea lt h S u rv ey (P N H S ) ( co nc lu de d) A g e g ro u p B M I ( kg /m 2 ) W H R S B P (m m H g ) D B P (m m H g ) T C ( m g /d L ) G lu co se (m g /d L ) (y ea rs ) M ea n S D M ea n S D M ea n S D M ea n S D M ea n S D M ea n S D 70 + 20 .1 9 4. 48 0. 91 0. 10 13 3. 12 22 .8 5 80 .3 8 11 .8 9 17 5. 34 42 .2 8 11 2. 39 45 .0 6 (n = 2 10 ) (n = 1 85 ) (n = 1 84 ) (n = 1 84 ) (n = 1 80 ) (n = 1 78 ) A ll 21 .4 7 4. 84 0. 89 0. 09 11 9. 4 19 .3 8 77 .4 2 11 .1 6 15 9. 89 39 .3 9 10 3. 40 36 .8 1 (n = 3 06 1) (n = 3 12 6) (n = 3 15 6) (n = 3 15 ) (n = 3 61 4) (n = 3 04 9) B M I= b od y m as s in de x. W H R = w ai st h ip r at io . S B P = s ys to lic b lo od p re ss ur e. D B P = d ia st ol ic b lo od p re ss ur e. T C = t ot al c ho le st er ol . S D = s ta nd ar d de vi at io n. 02 Coronary heart disease.pmd 12/20/2005, 2:34 PM265 266 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Table 3 Prevalence of risk factors for coronary heart disease in men and women in the Pakistan in the National Health Survey (PNHS) and Metroville Health Study (MHS) Risk factor Men Women PNHS MHS P-value PNHS MHS P-value No. % No. % No. % No. % Body mass index > 25 kg/m2 3052 12 487 33 < 0.001 3061 20 486 47 < 0.001 WHR > 0.96 (men); > 0.86 (women) 3147 15 492 34 < 0.001 3127 85 488 76 < 0.001 Hypertension BP > 140/90 mmHg 3173 18 490 23 0.0013 3156 15 483 23 NS Cholesterol > 200 mg/dL 3052 11 465 17 0.0004 3014 14 460 22 NS Blood glucose > 140 mg/dL 3079 5 464 8 NS 3049 5 462 8 NS Smoking 3083 35 488 35 < 0.001 3152 4 485 7 < 0.001 and again higher in women at 159.89 (SD 39.39) mg/dL (P < 0.0001). Serum choles- terol increased with age in both studies and comparison of MHS and PNHS showed significantly higher serum cholesterol in MHS men and women (P < 0.0001) (Ta- bles 1 and 2). Hypercholesterolaemia was defined as serum cholesterol exceeding 200 mg/dL [27]. The overall prevalence of hypercho- lesterolaemia was 17% in PNHS men ver- sus 11% in MHS men (P < 0.001) and 22% in MHS women versus 14% in PNHS women (P < 0.05). The prevalence in- creased with age in both studies (Table 3). Blood glucose Random blood glucose profiles in MHS and PNHS are shown in Tables 1 and 2. Blood glucose levels in MHS men and women in- creased with age. The overall mean of 106.15 (SD 35.88) in MHS men was not significantly different from women [107.04 (SD 37.69) mg/dL]. In PNHS men the mean BG was 99.16 (SD 36.39) mg/dL and 103.40 (SD 36.81) mg/dL in women, which was significantly higher (P < 0.0001). Comparison of mean blood glu- cose levels in MHS with PNHS showed significantly higher values in MHS men (P < 0.0001) and women (P = 0.048). The mean prevalence of hyperglycae- mia, defined as blood glucose > 140 mg/ dL, was 8% in MHS and 5% in PNHS men (P < 0.0001). The rates were the same in women (Table 3). The prevalence of hy- perglycaemia increased with age in both MHS and PNHS men and women. Smoking Smoking was uncommon in women in both MHS and PNHS. The prevalence of smoking in women aged 18–19 years was 0% in both studies but the rate increased steadily with age. The prevalence of smok- ing among men was much higher in both studies; 17% of 18–19-year-olds smoked in MHS which increased to 69% in >70- year-olds, while 12% of 18–19-year-olds smoked in PNHS and the rate increased 02 Coronary heart disease.pmd 12/20/2005, 2:34 PM266 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 267 with age to 33% among 60–69-year-olds. The overall prevalence was 35% in men in both studies and 7% and 4% in MHS and PNHS women respectively (Figure 1). Comparative analysis Comparisons of the MHS, PNHS and NHANES111 data were done by comparing the age-adjusted overall mean of each risk factor (Table 4) and age-adjusted percentile plots of each risk factors (data available on request). The BMI percentile plots showed that the BMI values were highest in NHANES111 followed by MHS and PNHS. The comparison was valid for women as well. WHR comparisons revealed that in men WHR was highest in NHANES111 followed by MHS the values of the two were almost equal in the 75th to 90th percentiles. For women MHS and NHANES111 showed similar values except in highest percentiles where NHANES 111 were slightly higher. Total serum cholesterol values were higher for both men and women in NHANES111 compared to MHS and PNHS were MHS had higher values in lower and middle percentiles and almost equal in high- er percentiles. SBP values were slightly higher in NHANES111 compared to MHS men and PNHS had values lower than the two. For the women MHS had the highest values fol- lowed by PNHS. The NHAHES111 females were close to PNHS in the lower percen- tiles and were lower than PNHS in the high- er percentiles. For the DBP in both men and women the MHS had highest values fol- lowed closely by PNHS men and signifi- cantly lower in women while NHANES111 had the lowest values. The blood glucose level was highest in MHS followed by Table 4 WHO age-adjusted mean for risk factor in MHS, PNHS and NHANES111 Variable Men Women MHS PNHS US MHS PNHS US Systolic blood pressure (mmHg) 117.2 (3.5) 115.1 (2.1) 118.0 (3.0) 117.0 (3.6) 115.0 (2.9) 112.9 (3.2) Diastolic blood pressure (mmHg) 79.0 (1.5) 77.5 (1.7) 70.4 (2.6) 76.7 (1.4) 73.5 (2.0) 66.8 (2.1) Total cholesterol (mg/dL) 157.0 (6.0) 151.9 (2.9) 189.5 (5.5) 162.8 (6.1) 159.5 (2.9) 192.7 (6.1) Glucose (mg/dL) 103.1 (4.2) 99.3 (2.1) 96.5 (1.8) 105.7 (3.5) 103.8 (2.1) 92.9 (2.1) Haemoglobin (g/dL) 14.0 (0.2) 14.0 (0.2) 14.7 (0.2) 12.6 (0.1) 12.3 (0.1) 13.2 (0.05) Body mass index (kg/m2) 21.3 (1.0) 19.4 (0.7) 24.4 (0.9) 22.7 (1.2) 20.0 (0.7) 24.4 (0.9) Waist to hip ratio 0.91 0.88 0.92 0.85 0.89 0.85 (0.015) (0.009) (0.13) (0.012) (0.007) (0.009) Heart rate (beats/ minute) 73.9 (4.1) 84.5 (1.1) 74.1 (1.0) 76.1 (4.3) 85.5 (1.0) 78.3 (1.0) Values are means (standard error of the mean). PNHS = Pakistan National Health Survey. MHS = Metroville Health Study. NHANES111 = Third National Health and Nutrition Examination Survey 1988-1994. 02 Coronary heart disease.pmd 12/20/2005, 2:34 PM267 268 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 PNHS and NHANES111 in both men and women. Discussion Our study showed that PNHS risk factor profiles, which represents the national av- erage, could be used to compare the CHD risk factors profile of Metroville, a typical lower middle class urban community of a large metropolis of 10 million people. The degree of increase in levels of risk factors and increase in prevalence rates of CHD risk factors in Metroville could be quanti- fied. The data showed that in MHS com- pared to PNH, the risk factors profile of BMI, WHR, blood pressure, cholesterol and blood glucose were higher in Metro- ville. The risk factors profile and preva- lence rates were worst for the women compared to men in both studies and in Metroville the women had higher levels of risk factors compared to PNHS women, which represents the national average. Re- cent reports from urban centres in Pakistan show increasing incidence of CHD and it seems likely that the high prevalence rates of hypertension, hypercholesterolaemia and obesity in Metroville and similar com- munities in Pakistan are being translated into increasing clinical CHD in urban Paki- stan because a high profile of risk factors has been shown to translate into high prev- alence rates [21]. Smoking was the only risk factor that was predominantly more prevalent in men; women did not smoke to any significant degree. The demographic profile of Metroville was that of a lower middle class urban community and showed all the characteris- tics of the Pakistani population at large, the differences were quantitative rather than qualitative; for instance unemployment, il- literacy, low wages and a predominance of young people are all evident in MHS but are quantitatively less. Thus the MHS study population was characteristic of other ur- Figure 1 Prevalence of smoking in men and women by age group in the Metroville Health Study (MHS) and the Pakistan National Health Survey (PNHS) 02 Coronary heart disease.pmd 12/20/2005, 2:34 PM268 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 269 ban communities in Pakistan [2,14]. The important difference between the urban and rural populations was in wages. Al- though low by modern standards, better wages in the urban population gave the population purchasing power which is a factor in changing the lifestyles. For exam- ple, smoking was less than 4% in those un- der 18 years but with the acquisition of employment this figure rose to 34% in 20– 39-year-olds. The availability of jobs in the towns is a major reason for rural to urban migration. Our data show that compared to the PNHS population, adoption of urban lif- estyles by the MHS population had ad- versely affected their risk-factor profile so that the overall mean levels of blood pres- sure, cholesterol, blood glucose and BMI were all higher in Metroville. This compar- ative risk-factor profile showed that urban- ization increased socioeconomic status but worsened the risk factors. The process of urbanization is an integral part of industrial- ization in developing countries; thus it seems likely that the risk-factor profile of these urbanized communities in Pakistan and other similar populations will continue to change for the worse. The high prevalence of risk factors such as hypercholesterolaemia, hyperten- sion and obesity in a population with low average values of BMI and serum choles- terol seems paradoxical. The explanation lies in the demographic profile of Pakistan as delineated in the PNHS which showed that the population is not homogeneous and is composed of a low number of communi- ties of high and middle socioeconomic sta- tus together with a much larger number of urban and rural communities of poor and lower socioeconomic status. The PNHS data are an average of this intermix and are skewed toward the majority poor and low socioeconomic population. The PNHS data have clearly defined this double burden of CHD and malnutrition in our population [2]. Increased levels of risk factors have been shown to translate into high preva- lence rates of risk factors to the levels that are significantly associated with clinical events of CHD and mortality [21]. Our study showed that the prevalence of hyper- tension, smoking, hypercholesterolaemia, obesity and diabetes were significantly higher in MHS than PNHS and suggests that the MHS population and similar com- munities in the developing countries are at a higher risk of future CHD [28–30]. This observation explains the observed increas- ing incidence of CHD in urban Pakistan [2– 14]. Both MHS and PNHS data showed a high prevalence of hypertension in our pop- ulation. The reasons for high levels of DBP in our population are not clear although technical factors cannot be entirely exclud- ed. Nonetheless the high DBP profile of all of the communities suggests that this ob- servation is real. One can conjecture as to the reasons, which include high salt diet, low birth weight, obesity due to lack of ex- ercise and genetic factors [18]. Compared to NHANES111 data from the USA, a country in which CHD is high, the risk-factor profile of the Pakistani pop- ulation in the PNHS population had lower BMI, serum cholesterol and WHR, and higher DBP, blood glucose levels and heart rate. The population risk-factor profile al- lows the projection of the future burden of CHD and related circulatory diseases. Thus the risk of future CHD in Pakistan is ex- pected to be less than the USA but is signif- icant for diabetes and hypertension. Comparison of the risk-factor profile of MHS and PNHS showed that risk factors were for the most part lower than NHANES111 values. This gap is due to the 02 Coronary heart disease.pmd 12/20/2005, 2:34 PM269 270 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 lower socioeconomic status in Pakistan; however, as the urbanization increase, this gap will decrease and, as MHS showed, the risk-factor profile can change for the worse. For certain risk factors, such as blood pressure, this gap is decreasing. It is possible that there are technical reasons for the high blood pressure profile, such as variations in the accuracy of various opera- tors, but this is unlikely because the preva- lence of hypertension in Pakistan is similar to the US population. Some of the reasons for narrowing of the gap may be genetic but the association of obesity with hyper- tension suggests that environmental factors are significant. It has been suggested that women of reproductive age are protected from the risk of CHD and in the industrialized world women tend to develop CHD 10 years later than men [15]. In Pakistan the male:female ratio of patients with myocardial infraction has been reported to be 3:1 and the gender age gap to be 5 years [15]. Data from India, where lifestyles of urban communities are similar to Pakistan, also support the obser- vation of early onset of CHD in Asian wom- en [16–18,29,30]. The high prevalence of risk factors in women in Pakistan thus ap- pears to have translated into early and se- vere CHD in women [15]. For developing countries, CHD risk factors have been shown to largely explain the rate of CHD in a population and, with advancing age, the male/female gap of onset of CHD is eventu- ally reversed. This effect has been reported to be associated with a proportionately higher age-related increase in the risk fac- tors in women in the industrialized world. The early and severe CHD in Asian women has been variously ascribed to insulin resis- tance and genetically determined increased lipoprotein Lp(a) [17,18,31]. Urbanization and the sedentary home living conditions of Pakistani women may be the most impor- tant factor initiating obesity and the cluster- ing of all other risk factors [31]. This hy- pothesis needs further elaboration and confirmation. In conclusion our study showed that the degree of adverse change in the CHD risk-factor profile that has occurred in our urban lower-middle class community in Pakistan could be detected by comparing its CHD risk factors with the risk factor profile of PNHS and strongly indicates the effect of urbanization in worsening the risk-factor profile of our urban communi- ties. Our study also showed that the com- parative gender risk-factor profile was worse for women and the prevalence of risk factors was higher in women except for smoking. To the extent that risk factors in a population contribute to an increased incidence of cardiovascular disease, our study has provided an explanation for the observed early onset of CHD in Pakistani women. Our study supports the reported conclusion that ischaemic heart disease has emerged as the leading cause of morbidity and mortality in the developing world and suggests that urgent preventive strategies are needed to check the worsening life- styles in the emerging urban communities in Pakistan and other developing countries. Urbanization needs to be planned so that towns can allow for healthy lifestyles for the entire population. Acknowledgements The authors acknowledge the contributions made to the MHS by the following: M. Faisal Memon, T. Ashraf; SA Hafeez, SA Ghuari, H. Sultana from the National Insti- tute of Cardiovascular Diseases, Karachi; T. Manolio from the Department of Epide- miology, National Institute of Health, USA; C.E. Davis, B.H. Dennis and L. She, from 02 Coronary heart disease.pmd 12/20/2005, 2:34 PM270 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 271 the Collaborative Studies Coordinating Center, Chapel Hill, North Carolina. A significant portion of the cost was supported by a grant from the US–PAKI- STAN P.L. 480 program, the US National Heart, Lung, and Blood Institute (NHLBI) and under NHLBI contract NO-1HV59224 with the University of North Carolina, Chapel Hill. References 1. Zaman A, Ara I. Rising urbanization in Pakistan. The Journal, 2002; 7 (3) : 31– 46 (A publication of the National Institute of Public Administration, Karachi, Paki- stan). 2. Pakistan Medical Research Council. Na- tional Health Survey of Pakistan 1990– 1994. Islamabad, Pakistan Medical Research Council, 1998 (collaborating organizations: Pakistan Medical Coun- cil, Federal Bureau of Statistics Pakistan, Department of Health and Human Ser- vices, USA). 3. Zafar HM et al. The incidence of ischaemic heart disease in adults in Multan. A comparative study of risk be- tween Pakistanis and Americans. Paki- stan heart journal, 1975, 8(1–4):4–11. 4. Raza M et al. Epidemiological study of coronary heart disease in Pakistani com- munity. Journal of the Pakistan Medical Association, 1970, 20:389–94. 5. Abbasi AS et al. Serum cholesterol in health and in coronary heart disease. Pakistan heart journal, 1968, 1(2):66– 72. 6. Multi-center study of risk factors for coro- nary heart disease. Islamabad, Pakistan Medical Research Council, 1980 (Mono- gram No. 3). 7. Samad A et al. Risk factors analysis in random population of 4 cities in Paki- stan. Pakistan journal of cardiology, 1992, 36:7–14. 8. Beg MA et al. Atherosclerosis in Karachi. Journal of the Pakistan Medical Associa- tion, 1967,17:236–44. 9. Pirzada MA, Khan AH. Coronary heart disease in West Pakistan. Pakistan jour- nal of medical research, 1962, 2:9–37. 10. Khan N, Nasiruddin A. Epidemiology of coronary heart disease in Peshawar. Pa- kistan heart journal, 1973, 6:64–71. 11. Faruqui AMA. Heart disease in south- east Asia: Experience in Pakistan. In: Hurst JW, ed. Clinical essays on the heart, Vol. 1. New York, McGraw–Hill, 1983. 12. Raza M, Syed SA. Ischaemic heart dis- ease in Pakistan. Pakistan heart journal, 1972, 5(3):68–76. 13. Akram Z et al. Risk factor analysis of ischaemic heart disease in patients pre- senting for coronary angiography at Punjab Institute of cardiology, Lahore. Pakistan journal of cardiology, 1999, 10:115–20. 14. Aziz KU et al. Prevalence of risk factors for coronary heart disease in lower middle class urban communities in Paki- stan. Pakistan heart journal, 2000, 33(1– 2):46–50. 15. Memon MA Samad A. Acute myocardial infarction in women. Pakistan journal of cardiology, 1999, 10:95–107. 16. Kannel WB. Metabolic risk factors for coronary heart disease in women: per- spective from the Framingham Study. American heart journal, 1987, 114:413– 9. 17. Mckeigue PM et al. Diabetes hyper- insulinaemia and coronary risk factors in 02 Coronary heart disease.pmd 12/20/2005, 2:34 PM271 272 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Bangladeshis in East London. British heart journal, 1988, 60:390–6. 18. Mckeigue PM, Miller GJ, Marmot MG. Coronary heart disease in south Asians overseas. A review. Journal of clinical epidemiology, 1989, 42(7):597–609. 19. Enas EA, Mehta J. Malignant coronary artery disease in young Asian Indians: thoughts on pathogenesis, prevention, and therapy. Coronary Artery Disease in Asian Indians (CADI) Study. Clinical car- diology, 1995, 18:131–3. 20. Lopez AD, Murray CJL. Mortality by cause for eight regions of the world. Glo- bal burden of disease study. Lancet, 1997, 340:1269–76. 21. Jousilahti P et al. Sex, age, cardiovascu- lar risk factors and coronary heart dis- ease, a prospective follow up study of 1478 middle aged men and women in Finland. Circulation, 1999, 99:1165–72. 22. Syed SA et al. Establishment of compre- hensive research and rehabilitation programme for persons of various forms of heart disease. Karachi, Pakistan, Na- tional Institute of Cardiovascular Dis- eases (Project VRA Pak 8–66). 23. Aziz KU et al. Planning and implementa- tion of a cardiovascular prevention strat- egy in a lower middle class community in Pakistan – a preliminary report. Journal of the College of Physicians and Sur- geons Pakistan, 2000, 10(11):440–3. 24. The third National Health and Nutrition Examination Survey, 1988–1994. Hy- attsville, Maryland, National Center for Health Statistics, 1996. 25. Physical status: the use of and interpre- tation of anthropometry. Report of a WHO expert committee. Geneva, World Health Organization, 1995 (WHO Technical Re- port Series, No. 854). 26. Hypertension prevalence and the status of awareness, treatment, and control in the United States. Final report of the Sub- committee on Definition and Prevalence of the 1984 Joint National Committee. Hypertension, 1985, 7(3 Pt 1):457–68. 27. National Heart, Lung and Blood Institute. Report of the expert panel on detection evaluation and treatment of high blood cholesterol adults. National cholesterol education programme. Washington DC, US Department of Health and Human Services, 1988. 28. Tofier GH et al. Effect of gender and race on prognosis after myocardial infarction in women and men. Journal of the Ameri- can College of Cardiology, 1995, 8:3– 11. 29. Reddy KS, Yusuf S. Emerging epidemic of cardiovascular diseases in develop- ing countries. Circulation, 1998, 97:596– 601. 30. Yusuf S et al. Global burden of cardio- vascular diseases. Part II: variations in cardiovascular disease by specific eth- nic groups and geographic regions and prevention strategies. Circulation, 2001, 104:2855–64. 31. Manson JE et al. A prospective study of obesity and risk of coronary heart dis- ease in women. New England journal of medicine, 1990, 322:882–9. 02 Coronary heart disease.pmd 12/20/2005, 2:34 PM272 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 273 Cancer of the jaw and oral cavity in the Syrian Arab Republic: an epidemiological study ABSTRACT: Oral cancer ranks sixth globally among cancers in terms of prevalence. It is a challenge to physicians, a serious problem to patients and a human and financial loss to the community. This study was conducted to identify the epidemiological factors relating to oral cancer in the Syrian Arab Republic in line with the International Dental Federation recommendations which emphasize the need to focus efforts on oral cancer. The study, conducted between 1996 and 2001, involved outpatients with oral and jaw cancer at several government hospitals. Data were collected from the hospital archives, recorded on special forms and analysed. Recommendations are presented. Cancer de la mâchoire et de la cavité buccale en République arabe syrienne : étude épidé- miologique RÉSUMÉ Le cancer de la cavité buccale occupe la sixième place au niveau mondial parmi les cancers en termes de prévalence. Cela représente un défi pour le médecin, un problème sérieux pour le patient et une perte humaine et financière pour la communauté. Cette étude a été réalisée pour identifier les facteurs épidémiologiques liés au cancer de la cavité buccale en République arabe syrienne, conformément aux recommandations de la Fédération dentaire internationale qui soulignent la nécessité de centrer les efforts sur le cancer de la cavité buccale. L’étude réalisée entre 1996 et 2001 concernait des patients des consul- tations externes atteints de cancer de la cavité buccale et de la mâchoire dans plusieurs hôpitaux publics. Les données ont été recueillies dans les archives hospitalières, enregistrées sur des formulaires spéciaux et analysées. Des recommandations sont présentées. M.S. Shahrour. Department of Oral and Maxillofacial Surgery, Ibn al-Nafis Hospital, Ministry of Health, Damascus, Syrian Arab Republic. Received: 26/05/04; accepted: 12/10/04 03 Cancer of the jaw (Arabic).pmd 12/20/2005, 2:34 PM273 274 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 03 Cancer of the jaw (Arabic).pmd 12/20/2005, 2:34 PM274 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 275 03 Cancer of the jaw (Arabic).pmd 12/20/2005, 2:34 PM275 276 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 03 Cancer of the jaw (Arabic).pmd 12/20/2005, 2:34 PM276 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 277 03 Cancer of the jaw (Arabic).pmd 12/20/2005, 2:34 PM277 278 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 03 Cancer of the jaw (Arabic).pmd 12/20/2005, 2:34 PM278 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 279 03 Cancer of the jaw (Arabic).pmd 12/20/2005, 2:34 PM279 280 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 03 Cancer of the jaw (Arabic).pmd 12/20/2005, 2:34 PM280 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 281 03 Cancer of the jaw (Arabic).pmd 12/20/2005, 2:34 PM281 282 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 03 Cancer of the jaw (Arabic).pmd 12/20/2005, 2:34 PM282 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 283 03 Cancer of the jaw (Arabic).pmd 12/20/2005, 2:34 PM283 284 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 03 Cancer of the jaw (Arabic).pmd 12/20/2005, 2:34 PM284 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 285 3. Swango PA. Cancers of the oral cavity and pharynx in the United States: an epi- demiologic overview. Journal of public health dentistry, 1996; 56:309–18. 4. Shaklar G. Oral cancer. Philadelphia, WB Saunders, 1984. 5. Best evidence 2: linking medical re- search to practice [database on CD Rom]. Philadelphia, American College of Physicians, 1998. 6. Hindle I, Downer MC, Speight PM. The epidemiology of oral cancer. British jour- nal of oral & maxillofacial surgery, 1996, 34:471–6. 7. Boyle P et al. Recent advances in epide- miology of head and neck cancer. Cur- rent opinion in oncology, 1992, 4:471–7. 8. Macfarlane GJ et al. Rising trends of oral cancer mortality among males world- wide: the return of an old public health problem. Cancer causes & control, 1994, 5(3):259–65. 9. Johnson NW, ed. Oral cancer: detection of patients and lesions at risk. Cam- bridge, Cambridge University Press, 1991. 10. McKaig RG, Baric RS, Olshan AF. Human papilloma virus and head and neck can- cer: epidemiology and molecular biol- ogy. Head & neck 1998, 20:250–65. 11. Goldstein AM et al. Familial risk in oral and pharyngeal cancer. European jour- nal of cancer. Part B, Oral oncology, 1994, 308:319–22. 12. King GN et al. Increased prevalence of dysplastic and malignant lip lesions in renal-transplant recipients. New En- gland journal of medicine, 1995, 332:1052–7. 13. Gupta PC et al. Epidemiologic character- istics of treated oral cancer patients de- tected in a house-to-house survey in Kerala, India. Indian journal of cancer, 1986, 23:206–11. 14. Gupta PC et al. Primary prevention trial of oral cancer in India: a 10-year follow- up study. Journal of oral pathology & medicine, 1992, 21:433–9. 15. National Screening Committee. First re- port of the NSC. London, Department of Health, 1998. 16. Wilson JMG, Jungner G. Principles and practice of screening for disease. Geneva, World Health Organization, 1968 (Public Health Papers, No. 34). 17. Dimitroulis G, Reade P, Wiesenfeld D. Referral patterns of patients with oral squamous cell carcinoma, Australia. Eu- ropean journal of cancer. Part B, Oral oncology, 1992, 28B:23–7. 18. Jovanovic A et al. Delay in diagnosis of oral squamous cell carcinoma: a report from the Netherlands. European journal of cancer. Part B, Oral oncology, 1992, 28B:37–8. 19. Horowitz AM et al. The need for health promotion in oral cancer prevention and early detection. Journal of public health dentistry, 1996, 56:319–30. 20. Edwards D et al. Management of can- cers of the head and neck in the United Kingdom: a questionnaire survey of con- sultants. British medical journal, 1997, 315:1589. 21. El-Sayed S, Nelson N. Adjuvant and ad- junctive chemotherapy in the manage- ment of squamous cell carcinoma of the head and neck region: a meta-analysis of prospective and randomised trials. References 03 Cancer of the jaw (Arabic).pmd 12/20/2005, 2:34 PM285 286 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Journal of clinical oncology, 1996, 14:838–47. 22. Dimitroulis G, Avery BS. Oral cancer: a synopsis of pathology and management. Oxford, Wright, 1998. 23. The world oral health report 2003. Geneva, World Health Organization, 2003 (WHO/NMH/NPH/ORH/03.2). 24. Steward BW, Kleihues P. World cancer report. Lyon, WHO International Agency for Research on Cancer, 2003. 25. Silverman S Jr. Demographics and oc- currence of oral and pharyngeal can- cers. The outcomes, the trends, the challenge. Journal of the American Den- tal Association, 2001, 132(suppl.):7S– 11S. Correction Apolipoprotein B gene polymorphisms in people in the East Mediterranean area of Turkey. L. Tamer, K. Tanriverdi, B. Ercan, A. Unlu, N. Sucu, H. Pekdemir and U. Atik. Eastern Mediterranean Health Journal, 2004, Vol. 10 Nos 1/2, pages 125–130. In the English and French abstracts, certain symbols and diacritics were lost or distorted on printing. The correct abstracts are printed below. ABSTRACT Point mutations in the receptor binding domain of low density lipoprotein may increase cholesterol levels in blood. Three mutations of Apo B-100 protein result in defective binding (Arg 3500→Gln, Arg 3500→Trp and Arg 3531→Cys). We estimated the frequency of Apo B point mutations (codon 3500) C9774T (Arg 3500→Trp) and G9775A (Arg 3500→Gln) in 179 atherosclerotic, 145 hyperlipidaemic individuals and 272 healthy individuals in the east Mediterranean region of Turkey. Lipid and lipoprotein levels were measured with routine biochemical analyser and Apo B mutation was detected using real-time PCR. Neither mutation was found. In this region, Apo B-100 protein mutations are rare and causes of hyperlipidaemia and atherosclerosis may therefore be unrelated to them. Les polymorphismes du gène de l’apolipoprotéine B dans la population de la région est- méditerranéenne en Turquie RESUME Les mutations ponctuelles sur le site de liaison du récepteur des lipoprotéines de basse densité peuvent faire augmenter le taux de cholestérol sanguin. Trois mutations de la protéine apo B-100 entraînent une liaison défectueuse (Arg 3500 → Gln, Arg 3500 → Trp and Arg 3531 → Cys). Nous avons estimé la fréquence des mutations ponctuelles de l’apo B (codon 3500) C9774T (Arg 3500 → Trp) et G9775A (Arg 3500 → Gln) chez 179 patients athérosclérotiques, 145 sujets hyperlipidémiques et 272 sujets en bonne santé de la région est- méditerranéenne en Turquie. Le taux de lipides et de lipoprotéines a été mesuré à l’aide d’un analyseur biochimique et la mutation de l’apo B a été recherchée en utilisant la PCR en temps réel. Aucune mutation n’a été trouvée. Dans cette région, les mutations de la protéine apo B-100 sont rares et les causes de l’hyperlipidémie et de l’athérosclérose peuvent donc ne pas être liées à ces mutations. 03 Cancer of the jaw (Arabic).pmd 12/20/2005, 2:34 PM286 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 287 Tobacco cessation in a population over age 15 years: a community- based survey in Alexandria R.M. Youssef,1 S.A. Abou Khatwa1 and H.M. Fouad1 1Department of Community Medicine, Faculty of Medicine, University of Alexandria, Alexandria, Egypt (Correspondence to R.M. Youssef: randayoussef@link.net). Received: 10/04/03; accepted: 20/12/04 ABSTRACT A survey of 2120 adults in Alexandria, Egypt, studied the determinants and patterns of tobacco cessation among ever smokers. Ever smokers were 30.7% of respondents; only 3.5% had given up smoking (quit ratio of 11.4%). The quit ratio was significantly lower for ex-daily smokers (7.5%) than for ex- occasional smokers (44.8%). Among current smokers, 56.3% were between the contemplation and prepa- ratory phases for quitting, whereas 25.1% were in the action phase of attempting to quit, for an average duration of 2.5 months, before relapsing. Health concerns were the motive for all ex-smokers and 95.0% of attempters, but craving for nicotine was the commonest reason for relapse. Tobacco cessation was predict- ed by older age of tobacco initiation, shorter duration of use, presence of health problems and a perception of the benefits of quitting. Le sevrage tabagique dans une population âgée de plus de 15 ans : enquête communautaire à Alexandrie RÉSUMÉ Une enquête auprès de 2120 adultes à Alexandrie (Égypte) a permis d’étudier les déterminants et les caractéristiques du sevrage tabagique chez des personnes ayant fumé à un moment de leur existence. Celles-ci représentaient 30,7 % des sujets enquêtés ; seulement 3,5 % avaient cessé de fumer (taux d’abandon de 11,4 %). Le taux d’abandon était significativement plus faible pour les ex-fumeurs quotidiens (7,5 %) que pour les ex-fumeurs occasionnels (44,8 %). Parmi les fumeurs au moment de l’étude, 56,3 % se trouvaient entre les phases où ils envisageaient d’arrêter de fumer et où ils s’y préparaient, tandis que 25,1 % en étaient à l’action d’essayer d’arrêter de fumer, sur une durée moyenne de 2,5 mois, avant de rechuter. Les préoccupations de santé constituaient les motifs communs à tous les ex-fumeurs et à 95,0 % des personnes qui essayaient d’arrêter de fumer, mais le besoin impérieux de nicotine était la raison la plus courante de rechute. Les facteurs prédictifs du sevrage tabagique étaient l’initiation au tabagisme à un âge plus tardif, la plus courte durée de la consommation tabagique, la présence de problèmes de santé et la perception des avantages de ne pas fumer. 04 Tobacco cessation.pmd 12/20/2005, 2:34 PM287 288 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Introduction Tobacco products have no safe level of consumption. Tobacco is the only legal consumer product that causes ill health and premature death when used exactly as the manufacturer intends [1,2]. Most people know that tobacco is hazardous but few appreciate how hazardous it really is [2,3]. Tens of thousands of studies have docu- mented the association between cigarette smoking and a large number of serious and fatal diseases affecting almost all body sys- tems [2,4]. The lifetime risk of a smoker being killed by the use of tobacco products is at least 50%. Therefore, a lifelong smok- er is more likely to die as a direct result of tobacco use than from all other potential causes of death combined [2,3]. It is now reasonable to claim that tobacco use repre- sents the most extensively documented cause of ill health ever investigated in the history of biomedical research [4]. Quitting smoking has major and imme- diate benefits for smokers of all ages, bene- fits that accrue to those who already have smoking-related diseases as well as to those who do not. At any age, smoking cessation reduces the overall risk of death; however, the reduction of risk occurs more quickly for some diseases than others [4,5]. One year after quitting, the risk of coronary heart diseases decreases by 50% and within 15 years the associated risk of dying for an ex-smoker approaches that of a never smoker. Moreover, the risk of de- veloping stroke, chronic obstructive lung diseases and lung cancer also decreases but more slowly [6]. Smoking cessation falls under the cate- gory of primary prevention, as does the prevention of initiation [4]. Indeed, smok- ing cessation meets the definition of prima- ry prevention by reducing the risk of morbidity and mortality in asymptomatic people. The considerable evidence of the benefits of tobacco cessation prompted the study of ever smokers in Alexandria, Egypt, the aim being to determine the rate of tobacco cessation in a representative sample of the population, the motives for cessation, and the determinants of tobacco cessation and attempts to quit. Methods A comprehensive community-based tobac- co survey was conducted in Alexandria city between May and August 2000. The World Health Organization (WHO) modified clus- ter sample survey was adopted to enrol 2120 subjects aged 15 years and over. Fur- ther details of the sampling methods and questionnaire have been reported in earlier papers [7,8]. Data were collected using a pre-tested, pre-coded interview questionnaire in 4 sec- tions: participants’ demographic character- istics and level of physical activity; participants’ pattern of tobacco use and at- tempts to quit; participants’ knowledge of tobacco-associated morbidity; and partici- pants’ perceptions of the dangers of smok- ing, smokers’ susceptibility to health problems and potential benefits of refrain- ing from tobacco use [8]. The WHO core questions for tobacco surveys were used to classify the smoking status of participants into current smoker or ex-smokers (minimum duration of to- bacco cessation 3 months), and daily or occasional smokers [9]. The prevalence and the corresponding 95% confidence interval (CI) as well as the mean, standard deviation (SD) and the 95% CI of the mean were computed. The chi- squared and Student t-test were used to test the significance of the obtained results. Tobacco cessation was modelled as a func- 04 Tobacco cessation.pmd 12/20/2005, 2:34 PM288 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 289 tion of participants’ characteristics, pattern of tobacco use as well as their knowledge and perceptions. Results Ever-use of tobacco products was report- ed by 651 respondents (30.7%). The pro- portion of ever daily smokers (27.5%; 95% CI: 25.60–29.40) was significantly higher than that of ever occasional smokers (3.2%; 95% CI: 2.45–3.95) At the time of the survey, 577 respondents (27.2%; 95% CI: 25.31–29.09) were current smokers and only 74 (3.5%; 95% CI: 2.72–4.28) were ex-smokers. The quit ratio, the ratio of ex-smokers to ever smokers, was 11.4% (95% CI: 8.92–13.79). This was significantly lower for ex-daily smokers (7.5%; 95% CI: 5.39– 9.67) than for ex-occasional smokers (44.8%; 95% CI: 32.86–65.67). Ex-smok- ers had abstained for a mean (SD) duration of 4.07 (3.34) years, range 0.25–18.00. The majority of ex-smokers (82.4%) had successfully given up for 1 year or more whereas 17.6% had refrained for smoking for less than 1 year. The majority of ex- smokers stated perceived poor health (63.5%) and the fear of tobacco-related diseases (35.1%) as the main reason for quitting smoking. In only 1 ex-smoker (1.4%) the motive for giving up smoking was suffering from heart disease. Table 1 compares the demographic and health characteristics of ex-smokers with current smokers; both were comparable in all respects except for age and existing health problems. Ex-smokers were signifi- cantly younger than current smokers: mean (SD) age 35.2 (13.1) years versus 39.3 (12.3) years (t = 2.51, P = 0.012). In the age group between 25 to < 45 years, the proportion of ex-smokers (52.8%) was higher than current smokers (47.7%), whereas among those aged 45+ years the proportion of current smokers (36.9%) was higher than that of ex-smokers (22.9%). These differences were signifi- cant statistically (P = 0.024) (Table 1). A significantly higher percentage of ex- smokers (17.6%) reported enduring a chronic health problem (chronic bronchi- tis, hypertension, heart disease or diabetes) compared with current smokers (9.9%) (P = 0.044) (Table 1). Table 2 indicates that ex-smokers start- ed using tobacco at a significantly older age than current smokers [mean (SD) age 18.2 (4.0) years versus 20.5 (6.9)] (P < 0.0001) and continued to be tobacco users for a significantly shorter time [mean (SD) 21.1 (12.9) years versus 10.6 (10.1)] (P < 0.0001). Regarding knowledge of tobacco-relat- ed morbidity, the scores of ex-smokers and current smokers were identical (Table 3). On the perceptions scale, ex-smokers scored higher on the perception of dangers associated with tobacco use, susceptibility to tobacco-related morbidity as well as the benefits of cessation; however, the differ- ences were not statistically significant. Tobacco cessation was modelled as a function of participants’ demographic characteristics, pattern of tobacco use and knowledge and perceptions about smoking (Table 4). Quitting smoking was predicted by participants’ marital status, age of initia- tion and duration of tobacco use, presence of chronic health problems and the percep- tions of benefits associated with never starting to smoke or giving up smoking. This model correctly classified 89.3% of ex-smokers. Among current smokers, 145 (25.1%) reported having attempted to give up smok- ing. The mean (SD) number of attempts was 2.67 (1.81) (range 1–12). Fear of to- bacco-related morbidity was the motive in 04 Tobacco cessation.pmd 12/20/2005, 2:34 PM289 290 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 48.6% of participants, followed by per- ceived poor health (40.0%) or suffering a chronic disease aggravated by tobacco use (e.g. chronic bronchitis, hypertension, heart disease and diabetes) (6.4%). For only a few respondents, the motive for quitting was to protect their family from passive exposure to smoke (2.9%) and the rising cost of tobacco (2.1%). At each at- tempt, they had given up tobacco for a mean (SD) duration of 2.6 (1.92) months, range 0.40–10, followed by a relapse. The main reason for the relapse, given by nearly two-thirds of the attempters (64.2%) was a craving for nicotine. Other reasons were experiencing a personal problem (19.3%), Table 1 Demographic and health characteristics of current and ex-smokers in Alexandria, Egypt Characteristic Current smokers Ex-smokers χ2-value P-value (n = 577) (n = 74) No. % No. % Sex Men 563 97.6 70 94.6 χ21 = 2.17 0.141 Women 14 2.4 4 5.4 Age (years) < 25 88 15.4 13 24.3 χ22 = 7.44 0.024 25– < 45 274 47.7 39 52.8 45+ 215 36.9 17 22.9 Education Illiterate/read and write 293 50.8 40 54.1 χ23 = 0.75 0.862 Primary/preparatory 80 13.8 10 13.5 Secondary 130 22.6 17 22.9 University/higher 74 12.8 7 9.5 Occupationa (n = 482) (n = 60) Professional/ semiprofessional 65 13.6 5 8.3 χ22 = 1.31 0.518 Skilled/semiskilled 56 11.6 8 13.3 Manual/otherb 361 74.8 47 78.4 Marital statusc (n = 543) (n = 69) Unmarriedd 113 20.8 20 29.0 χ21 = 2.41 0.120 Married 430 79.2 49 71.0 Chronic health problems Absent 520 90.1 61 82.4 χ21 = 4.04 0.044 Present 57 9.9 13 17.6 Physical exercise Sedentary/light 536 92.9 70 94.6 χ21 = 0.29 0.587 Moderate/vigorous 41 7.1 4 5.4 aApplicable to those who were employed at the time of the survey. bOthers includes drivers, traders and fishermen. cNot applicable to students and those below the age 18 years. dUnmarried includes never married, divorced, separated, widows and widowers. 04 Tobacco cessation.pmd 12/20/2005, 2:34 PM290 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 291 boredom and loneliness (12.4%) or frustra- tion (4.1%). Current smokers who had ever attempt- ed to give up were had similar characteris- tics to those who had not attempted to give up smoking in terms of age, sex, level of education, occupation and marital status (Table 5). In addition, no significant differ- Table 3 Knowledge and perceptions about the health risks of smoking of current and ex-smokers Knowledge and Current smokers Ex-smokers t-value P-value perceptions (n = 577) (n = 74) Knowledge about health effects Mean (SD) score 24.96 (6.97) 24.81 (6.80) 0.17 0.860 Min–Max 14–42 14–42 Perception of dangers of smoking Mean (SD) score 9.64 (3.45) 10.12 (3.12) 1.12 0.262 Min–Max 0–14 4–14 Perception of susceptibility to illness score Mean (SD) score 6.86 (2.39) 7.27 (2.15) 1.41 0.159 Min–Max 0–10 4–10 Perception of benefits of not smoking Mean (SD) score 9.54 (3.42) 10.31 (3.11) 1.83 0.068 Min–Max 0–14 5–14 n = total number of respondents. SD = standard deviation. Table 2 Age of initiation and duration of tobacco use of current and ex-smokers Initiation/duration Current smokers Ex-smokers t-value P-value of tobacco use (n = 577) (n = 74) Age of initiation (years) Mean (SD) 18.21 (4.04) 20.50 (6.94) t = 4.14 < 0.0001 Min–Max 10–36 10–50 95% CI 17.88–18.54 18.89–22.11 Duration of tobacco use (years) Mean (SD) 21.06 (12.86) 10.59 (10.11) t = 6.73 < 0.0001 Min–Max 1–50 0.17–46 95% CI 20.00–22.11 8.24–12.93 n = total number of respondents. SD = standard deviation. CI = confidence interval. 04 Tobacco cessation.pmd 12/20/2005, 2:34 PM291 292 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 ences were detected between the 2 groups in relation to health problems or the level of physical activity. Table 6 shows that current smokers who had attempted to quit smoking were comparable to those who had not tried to quit in respect of their smoking status. Both groups initiated the habit around the age of 18 years, continued to be tobacco users for a similar duration and smoked nearly the same number of cigarettes. The mean scores of current smokers who ever attempted tobacco cessation on the knowledge scale as well as on the per- ception of benefits scale was comparable with their counterparts who reported never trying to quit. However, those who had at- tempted tobacco cessation had significant- ly higher scores on the perception of danger scale: mean (SD) score 10.18 (3.37) versus 9.47 (3.46) (P = 0.032). They also had higher scores on the percep- tion of susceptibility scale: mean (SD) score 7.20 (2.31) versus 6.75 (2.40) (P = 0.047) (Table 7). The health warning on cigarette packets was seen as a motivation to stop smoking for significantly more current smokers who attempted to quit smoking than for those who had never attempted to quit (50.3% versus 38.9%; P = 0.016) (Table 8). Moreover, a significantly higher per- centage of current smokers who attempted to give up (34.5%) had received advice from a medical professional to quit the hab- it (P = 0.012). More current smokers who suffer chronic health problems had been advised by a doctor to quit (59.6%) than healthy current smokers (22.9%) (P < 0.0001). Regarding patterns of tobacco use, half of the current smokers (51.3%) smoked the same number of cigarettes as the previ- ous year, whereas 30.0% reported smoking less and 18.7% reported smoking more than the previous year. No significant dif- ference in the pattern of tobacco use was observed between those who attempted to quit smoking and those who did not (Table 8). More than half of current smokers Table 4 Independent predictors of tobacco cessation of ex-smokers Independent Coefficient Standard Adjusted 95% CI P-value predictor error OR Health problems Absenta Present 1.13 0.402 3.11 1.41–6.83 0.0048 Marital status Marrieda Unmarriedb 0.758 0.345 2.13 1.08–4.19 0.0280 Duration of tobacco use –0.110 0.018 0.89 0.86–0.92 < 0.0001 Perception of benefits of not smoking 0.088 0.041 1.09 1.01–1.18 0.0322 Age of initiation 0.060 0.028 1.06 1.01–1.21 0.0315 OR = odds ratio. CI = confidence interval. aReference category. bUnmarried include never married, divorced, separated, widows and widowers. 04 Tobacco cessation.pmd 12/20/2005, 2:34 PM292 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 293 (56.3%) expected they would give up in the next 5 years. Indeed, a significantly higher percentage of current smokers who had at- tempted to give up smoking (63.4%) ex- pected that they would quit in the next 5 years compared with those who had not attempted to give up (53.9%) (P = 0.046). Discussion The total prevalence of tobacco use in a population is the net product of the two opposing forces of initiation and cessation. In any population, health awareness gener- ally starts with the utilization of curative Table 5 Demographic and health characteristics of current smokers in relation to previous cessation attempts Characteristic Previous attempt χ2-value P-value No (n = 432) Yes (n = 145) No. % No. % Sex Men 423 97.9 140 96.6 χ21 = 0.85 0.355 Women 9 2.1 5 3.4 Age (years) < 35 157 36.3 58 40.0 χ22 = 1.08 0.584 35– < 55 215 49.8 71 49.0 55+ 60 13.9 16 11.0 Education Illiterate/read and write 210 48.6 83 57.3 χ23 = 3.45 0.326 Primary/preparatory 63 14.6 17 11.7 Secondary 100 23.1 30 20.7 University/higher 59 13.7 15 10.3 Occupationa (n = 355) (n = 127) Professional/ semiprofessional 50 14.1 15 11.8 χ22 = 1.40 0.497 Skilled/semiskilled 44 12.4 12 9.5 Manual/otherb 261 73.5 100 78.7 Marital statusc (n = 406) (n = 137) Unmarriedd 88 21.7 25 18.2 χ21 = 0.73 0.392 Married 318 78.3 112 81.8 Chronic health problems Absent 392 90.7 128 88.3 χ21 = 0.74 0.389 Present 40 9.3 17 11.7 Physical exercise Sedentary/light 403 93.3 133 91.7 χ21 = 0.40 0.526 Moderate/vigorous 29 6.7 12 8.3 aApplicable to those who were employed at the time of the survey. bOthers include drivers, traders and fishermen. cExcluding students and those below the age of 18 years. dUnmarried includes never married, divorced, separated, widows and widowers. n = total number of respondents. 04 Tobacco cessation.pmd 12/20/2005, 2:34 PM293 294 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Table 6 Pattern of tobacco use of current smokers in relation to previous cessation attempts Pattern of tobacco use Previous attempt χ2 and t- P-value No (n = 432) Yes (n = 145) value Smoking status (No. & %) Daily smoker 407 94.2 133 91.7 χ21 = 1.21 0.290 Occasional smoker 25 5.8 12 8.3 Age of initiation (years) Mean (SD) 18.14 (3.98) 18.43 (4.24) t = 0.74 Min–Max 10–35 10–36 0.462 Duration of tobacco use (years) Mean (SD) 21.47 (12.83) 19.83 (12.91) t = 1.34 0.182 Min–Max 1–50 1–50 Number of cigarettes smoked (per day) Mean (SD) 23.66 (12.51) 21.80 (12.55) t = 1.55 0.122 Min–Max 0.28–60 0.28–60 n = total number of respondents. SD = standard deviation. Table 7 Knowledge and perceptions about the health risks of smoking of current smokers in relation to previous cessation attempts Knowledge and Previous attempt t-value P-value perceptions No (n = 432) Yes (n = 145) Knowledge about health effects Mean (SD) score 25.15 (6.99) 24.37 (6.88) 1.16 0.245 Min–Max 14–42 14–42 Perception of dangers of smoking Mean (SD) score 9.47 (3.46) 10.18 (3.37) 2.15 0.032 Min–Max 0–14 2–14 Perception of susceptibility to illness Mean (SD) score 6.75 (2.40) 7.20 (2.31) 1.99 0.047 Min–Max 0–10 3–10 Perception of benefits of not smoking Mean (SD) score 9.41 (3.43) 9.97 (3.33) 1.73 0.083 Min–Max 0–14 3–14 n = total number of respondents. SD = standard deviation. 04 Tobacco cessation.pmd 12/20/2005, 2:34 PM294 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 295 services and then the preventive ones. Ap- plying this to tobacco use implies that ces- sation rates will increase before any decline in the rates of initiation. The presence of organized tobacco ces- sation programmes enhances quitting [4]. However, most people who quit smoking do so on their own without aids such as cessation clinics or nicotine gum [10]. In the current survey, 11.4% of ever smokers had successfully achieved tobacco cessa- tion on their own without the use of aids such as nicotine patches or smoking cessa- tion programmes. This rate is slightly high- er than the 9.5% reported from China [11] but lower than 15% from Denmark [12]. The differential in the quit ratio by previous smoking status—7.5% for ex-daily smok- ers and 44.8% for ex-occasional smok- ers—is in accordance with previous find- ings [13,14] and identical with figures re- ported by Sargent et al. [14]. Our study confirms previous studies that have documented the relation between a higher tendency to quit smoking and older age of starting smoking [15–17] as well as shorter duration of use [18]. This is likely to be due to a lower level of nicotine depen- dence, as the risk of dependency increases with earlier age of initiation [19]. Controversy exists about the most likely age for tobacco cessation. Several studies have reported a higher quit ratio in older age groups [12,20,21]. This disagrees with the present survey, where ex-smokers were significantly younger than those who continued to smoke. Our finding supports the 1990 Surgeon General’s report in the Table 8 Motivations for quitting and future intentions to quit of current smokers in relation to previous cessation attempts Motivations Previous attempt χ2-value P-value No (n = 432) Yes (n = 145) No. % No. % Ever received medical advice to quit smoking No 329 76.2 95 65.5 χ21 = 6.30 0.012 Yes 103 23.8 50 34.5 Felt motivated to quit by health warnings on packet No 264 61.1 72 49.7 χ1 = 5.86 0.016 Yes 168 38.9 73 50.3 Extent of tobacco use in previous year Not changed 215 49.8 81 296 χ22 = 2.65 0.266 Smoked more 87 20.1 21 14.5 Smoked less 130 30.1 43 29.7 Expected tobacco use in next 5 years Continue to smoke 199 46.1 53 36.6 χ21 = 3.99 0.046 Quit smoking 233 53.9 92 63.4 n = total number of respondents. 04 Tobacco cessation.pmd 12/20/2005, 2:34 PM295 296 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 United States of America (USA), which postulated that highly motivated smokers might have already quit smoking at a younger age, leaving a relatively “hard- core” group of older smokers [4]. This finding can also be attributed to a miscon- ception among older smokers that any health damage accrued from smoking is ir- reversible after decades of smoking. Re- search shows that it is never too late to quit and there are benefits to health no matter what age a person quits [4,6]. Tobacco cessation is a complex pro- cess that passes through different phases. The model of behaviour change described by Prochaska and Diclemente is a cycle, which starts with the pre-contemplation phase when an individual is not yet thinking about behaviour change [22]. In the con- templation phase a smoker begin to evalu- ate the pros and cons of quitting, and this is followed by the preparation phase when the smoker takes the decision to quit and sets a plan to achieve it. The action phase starts with attempts to quit, ultimately followed by a sustained state of abstinence. In our sample of current smokers, about 44% did not intend to quit in the next 5 years and presumably are in the pre-con- templation phase. This figure is almost identical to that reported in studies from Switzerland [23] and the USA [24]. On the other hand, nearly 56% expressed an inten- tion to quit in the next 5 years and they must be somewhere between the contem- plation and preparation phases. This figure is much higher than that reported from Chi- na [11], the USA [25] and Switzerland [23]. This finding may be a promising precursor to a reduction of smoking prevalence in Egypt in the near future. In our study, the action phase is repre- sented by the 17.6% of the sample who were recent ex-smokers (who had ab- stained for less than 1 year) and the 25.1% of current smokers who had attempted to give up. The attempters had made on aver- age 2.5 attempts to quit and had refrained for a variable time before relapsing, princi- pally due to a craving for nicotine. Howev- er, relapse should not be perceived as failure but as part of the cessation cycle. In most cases, early attempts at quitting lead to relapse due to withdrawal symptoms and the smoker typically progresses through stages of ambivalence and readiness to stop before finally succeeding with cessation and becoming a permanent ex-smoker [26]. As attempting to quit, rather than in- tending to quit, predicts ultimate success in tobacco cessation [12], it is essential to un- derstand this process and support those at- tempting to quit to achieve the desired behaviour. Considering the motivational factors, it is clear that health concerns, namely fear of tobacco-related morbidity and perceived poor health, are the greatest impetus for quitting and attempting to quit smoking in this survey as well as others [10–12]. This illustrates how strong is the addiction to to- bacco, so that attempting and maintaining a state of abstinence needs a sustained per- sonal effort. Further evidence for this is the significantly higher proportion of those with chronic health problems among ex- smokers compared with current smokers. In this study, medical advice was an effi- cient trigger to attempting to give up, but actual experience of ill health provided a greater motivation for smokers to transfer to the action phase of quitting. Most theories of behaviour assume that knowledge is a necessary, but not suffi- cient condition, for behavioural change. There is no better example of this than the individuals who continue to smoke despite being aware of the risk they face [27]. In our sample of ever smokers, knowledge about the health problems associated with 04 Tobacco cessation.pmd 12/20/2005, 2:34 PM296 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 297 tobacco use was the same for smokers who had attempted and those who had not attempted to quit. However, there were dif- ferences in perceptions of smoking risks between the groups. This is logical since perception is a more enduring state and a more powerful influence on behaviour than knowledge. Those who attempted to quit were more likely to believe in the dangers of tobacco use as well as their own suscep- tibility to adverse health effects compared with those who did not attempt to quit. At- tempters were more likely to admit that the health warning on cigarette packets moti- vated them to quit. In contrast, ex-smokers were more likely to believe in the benefits associated with giving up smoking. Our findings provide evidence that smokers can achieve tobacco cessation on their own if they are motivated to do so. Since health concerns were the main moti- vational factor, health education pro- grammes should stress the adverse consequences of tobacco on health by ex- emplifying the dangers a smoker faces, be- fore shifting to illustrate the benefits of cessation at any age. This follows smokers along the path of the cessation process. Relapses resulting from the craving for nic- otine should be expected, particularly among those who are heavily dependent on tobacco, who need encouragement to con- tinue trying until they succeed. A later age of starting smoking has a major effect on future cessation, as late initiators are less dependent on nicotine and more likely to renounce smoking after a short duration. Physicians are the most trusted source of antismoking messages [27], and the one- to-one encounter in the physician’s office was shown in this study to be an important motivation for those attempting cessation. However, physicians need to advise all their patients who smoke, not only those with health problems aggravated by smoking. Quitting smoking reduces mortality within just a few years, while preventing initiation does not influence mortality statistics until some 30 to 50 years later, when young peo- ple reach the age when smoking-related morbidity affects them [9]. There can be no debate about the need to emphasize smoking cessation activities as well as pre- vention of initiation of the habit. Further research on people who never smoked dai- ly is essential for determining what factors prevented them from becoming regular smokers. Acknowledgement The authors acknowledge World Health Organization/Eastern Mediterranean Re- gional Office for funding this research. References 1. Governments for a tobacco-free world. WHO fact sheet no. 159. Geneva, World Health Organization, 1998. 2. Sports and the arts without tobacco— play it tobacco free. World No-Tobacco Day advisory kit. Geneva, World Health Organization, 1996. 3. Tobacco or health: a global status report. Geneva, World Health Organization, 1997. 4. The Surgeon General’s 1990 report on the health benefits of smoking cessation: executive summary. Morbidity and mor- tality weekly report, Recommendations and reports, 1990, 39(RR-12):2–10. 5. Hermanson B et al. Beneficial six-year outcome of smoking cessation in older men and women with coronary artery disease. Results from the CASS registry. 04 Tobacco cessation.pmd 12/20/2005, 2:34 PM297 298 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 New England journal of medicine, 1988, 319:1365–9. 6. Tobacco epidemic: health dimensions: tobacco is a greater cause of death and disability than any single disease. WHO fact sheet, no. 154. Geneva, World Health Organization, 1998. 7. Youssef RM, Abou-Khatwa SA, Fouad HM. Prevalence of smoking and age of initiation in Alexandria, Egypt. Eastern Mediterranean health journal, 2002, 8 (4/5):626–37. 8. Youssef RM, Abou-Khatwa SA, Fouad HM. Current and never smokers: differ- entials in demographic characteristics, knowledge and perceptions. Eastern Mediterranean health journal, 2003, 9(5/ 6):923–34. 9. Guidelines for controlling and monitor- ing the tobacco epidemic. World Health Organization, Geneva 1998. 10. Pederson LL et al. Quitting smoking: why, how and what might help. Tobacco con- trol, 1996, 5:209–14. 11. Yang G et al. Smoking cessation in China: findings from the 1996 national prevalence survey. Tobacco control, 2001, 10:170–4. 12. Osler M, Prescott E. Psychosocial, behavioural, and health determinants of successful smoking cessation: a longitu- dinal study of Danish adults. Tobacco control, 1998, 7:262–7. 13. Wersall JP, Eklund G. The decline of smoking among Swedish men. Interna- tional journal of epidemiology, 1998, 27(1):20–6. 14. Sargent JD, Mott LA, Stevens M. Predic- tors of smoking cessation in adoles- cents. Archives of pediatrics & ado- lescent medicine, 1998, 152:388–93. 15. Breslau N, Peterson EL. Smoking cessa- tion in young adults: age at initiation of cigarette smoking and other suspected influences. American journal of public health, 1996, 86(2):214–20. 16. Eisner MD et al. Predictors of cigarette smoking and smoking cessation among adults with asthma. American journal of public health, 2000, 90(8):1307–11. 17. McGrady GA, Ahluwalia JS, Pederson LL. Smoking initiation and cessation in African Americans attending an inner- city walk-in clinic. American journal of preventive medicine, 1998, 14(2):130– 7. 18. Benuck I, Gidding SS, Binns HJ. Identifi- cation of adolescent tobacco users in a pediatric practice. Archives of pediatrics & adolescent medicine, 2001, 155(1): 32–5. 19. Breslau N, Fenn N, Peterson EL. Early smoking initiation and nicotine depen- dence in a cohort of young adults. Drug and alcohol dependence, 1993, 33(2): 129–37. 20. Fernandez E et al. Quitting smoking in northern Italy: a cross-sectional analysis of 2621 subjects. European journal of epidemiology, 1997, 13(3):267–73. 21. Husten CG et al. Cigarette smoking and smoking cessation among older adults: United States, 1965–94. Tobacco con- trol, 1997, 6(3):175–80. 22. Coleman-Wallace D et al. Evaluation of developmentally appropriate programs for adolescent tobacco cessation. Jour- nal of school health, 1999, 69(8):314–9. 23. Etter JF, Perneger TV. Snowball sampling by mail: application to a survey of smok- ers in the general population. Interna- tional journal of epidemiology, 2000, 29(1):43–8. 24. Pallonen UE et al. Stages of acquisition and cessation for adolescent smoking: an empirical integration. Addictive be- haviors, 1998, 23(3):303–24. 04 Tobacco cessation.pmd 12/20/2005, 2:34 PM298 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 299 25. Kiovula M, Paunonen M. Smoking habits among Finnish middle-aged men: expe- riences and attitudes. Journal of ad- vanced nursing, 1998, 27(2):327–34. 26. Prignot J. A tentative illustration of the smoking initiation and cessation cycles. Tobacco control, 2000, 9:113. 27. Strengthening tobacco control in Central & Eastern Europe: proceedings of a training seminar, Warsaw, Poland, 20-25 October 1995. Geneva, World Health Or- ganization, 1996 (WHO/PSA/96.9). HB 2006 (Healthy Buildings Conference) HB 2006 is the eighth in a series of Healthy Buildings Conferences that started in 1988 in Stockholm. The issues addressed relate to indoor air quality and its impact on health. The main focus is on buildings as confined spaces where we spend around 90% of our life. This latest conference will be held from 4 to 6 June 2006 in Lisbon, Portugal. Further information about the conference can be obtained from the Secretariat, Rua Dr Roberto Frias, 4200–465 Porto, Portugal. Tel.: +351 22 508 17 63, Fax: +351 22 508 21 53, e-mail: hb2006@fe. up.pt. Information is also available at the following website: http:// www.hb2006.org 04 Tobacco cessation.pmd 12/20/2005, 2:34 PM299 300 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Incidence of haemoglobinopathies detected through neonatal screening in the United Arab Emirates H. Al Hosani,1 M. Salah,1 H.M. Osman,1 H.M. Farag1 and S.M. Anvery2 1Central Department of Maternal and Child Health, Ministry of Health, Abu-Dhabi, United Arab Emirates. 2Department of Haematology, Central and Al Jazira Hospitals, Abu-Dhabi, United Arab Emirates. Received: 10/04/03; accepted: 15/03/04 ABSTRACT In January 2002, a pilot programme of neonatal screening for sickle cell disease was launched in the United Arab Emirates (UAE) in 3 districts of Abu Dhabi emirate. This paper reports the incidence of sickle cell diseases, other haemoglobinopathies and haemoglobinopathy carriers over a 12-month period using high performance liquid chromatography as a primary screening method. The overall incidence of sickle cell disease among 22 200 screened neonates was 0.04% (0.07% for UAE citizens and 0.02% for non-UAE citizens). The incidence of sickle cell trait was 1.1% overall (1.5% for UAE citizens and 0.8% for non-UAE citizens). Universal neonatal screening for sickle cell haemoglobin at the national level should be considered. Incidence des hémoglobinopathies détectées dans le cadre du dépistage néonatal aux Émirats arabes unis RÉSUMÉ En janvier 2002, un programme pilote de dépistage néonatal de la drépanocytose a été lancé aux Émirats arabes unis dans trois districts de l’Émirat d’Abou Dhabi. Cet article présente l’incidence des drépanocytoses, d’autres hémoglobinopathies et des porteurs d’hémoglobinopathies sur une période de 12 mois, la chromatographie à haute performance ayant été utilisée comme méthode de dépistage primaire. L’incidence globale de la drépanocytose chez les 22 200 nouveau-nés examinés était de 0,04 % (0,07 % pour les citoyens des Émirats arabes unis et 0,02 % pour les non-citoyens des Émirats). L’incidence du trait drépanocytaire était de 1,1 % en général (1,5 % pour les citoyens des Émirats arabes unis et 0,8 % pour les non-citoyens des Émirats). Le dépistage de l’hémoglobine S chez tous les nouveau-nés au niveau national devrait être envisagé. 05 Incidence of haemoglobinopathies.pmd 12/20/2005, 2:35 PM300 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 301 Introduction For at least 20 years it has been known that children with sickle cell anaemia have an increased susceptibility to severe bacterial infection, particularly from Streptococcus pneumoniae. The risk of major infection with this organism is greater in the first 3 years of life and can occur as early as 4 months of age. This infection may be the first clinical manifestation of disease and carries a case fatality rate as high as 30% [1]. The demonstration in 1986 that prophy- lactic penicillin markedly reduces the inci- dence of pneumococcal sepsis provided a powerful incentive for the widespread im- plementation of neonatal screening for sickle cell disease. Subsequent experience demonstrated that neonatal screening, when linked to timely diagnostic testing, parental education and comprehensive care, markedly reduces morbidity and mor- tality from sickle cell disease in infancy and early childhood [2]. Sickle cell disease is an autosomal re- cessive genetic disorder characterized by the presence of sickle haemoglobin (HbS) in red blood cells. The clinical manifesta- tions of sickle cell disease result from vari- able degrees of haemolysis and intermittent episodes of vascular occlusion that cause tissue ischaemia and acute and chronic or- gan dysfunction. Heterozygous individuals have sickle cell trait, a generally benign, as- ymptomatic genetic carrier state [3]. Genes for sickle cell disease are common in peo- ple of African, Mediterranean, Middle East- ern and Indian ancestry and people from the Caribbean and parts of Central and South America [4]. The primary purpose of screening is to identify infants with sickle cell disease. Screening also identifies infants with other haemoglobinopathies and haemoglobinopa- thy carriers. Detection of infants with hae- moglobin traits by screening of newborns and testing of other family members, af- fords an opportunity to identify and coun- sel couples at risk for having future children with clinically significant disease. Previously undiagnosed family members with the trait or disease can also be identi- fied [5]. In the United Arab Emirates (UAE), the national neonatal screening programme started by screening for phenylketonuria in January 1995. Screening for congenital hy- pothyroidism was introduced in January 1998. In January 2002, the Ministry of Health decided to launch a pilot study for neonatal screening of sickle cell disease be- fore expanding it at the national level. The aim of this pilot study was to deter- mine the incidence of sickle cell diseases, other haemoglobinopathies and haemoglo- binopathy carriers in 3 districts of Abu Dhabi emirate through a 12-month period of the sickle neonatal screening pro- gramme. Methods This pilot study was performed in Abu Dhabi, Al-Ain and Western Region medical districts from 1 January 2002 to 31 De- cember 2002. Primary screening was car- ried out using high performance liquid chromatography (HPLC). In UAE, a standard form for neonatal screening is issued to every baby born in hospital, where 99% of deliveries occur, and mothers are informed about the proce- dures and importance of neonatal screen- ing. Newborn infants are brought to the designated maternal and child health centre on the fifth day for collection of blood sam- ples by heel prick onto filter paper (S&S 05 Incidence of haemoglobinopathies.pmd 12/20/2005, 2:35 PM301 302 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 903, Schleicher & Schuell, New Hamp- shire, USA). In this pilot study, the same filter paper used for thyroid stimulating hormone and phenylalanine testing was used for testing for sickle cell disease. For each sample, an 1/8-inch diameter disc was punched out from the collection card and placed into a separate sample vial, 0.5 mL deionized wa- ter was added and after standing for 30 minutes at room temperature the disc was removed [6]. Sample vials were analysed in the Abu Dhabi laboratories using the Variant HPLC system (Bio-Rad Laboratories, Cali- fornia, USA) for detection of haemoglobin (Hb) types F, A, S, C, D and E. Samples positive for haemoglobinopa- thies were referred to the Abu Dhabi central hospital for confirmation by HPLC and iso- electric methods. After appropriate educa- tion and genetic counselling for the family, the maternal and child health genetic spe- cialist doctor in each district recommended confirmatory testing of the infant before the age of 2 months (for haemoglobinopa- thy diseases) and before the age of 4 months (for haemoglobinopathy carrier traits). Testing of parents and other family members for all haemoglobinopathies, in- cluding thalassaemia, was recommended in order to help establish the correct diagnosis in some infants [7]. All infants confirmed with sickle cell disease started prophylactic penicillin ther- apy by the age of 2 months and follow-up was arranged with the cooperation of a consultant medical haematologist. Results During the 12-month period of the pilot study, 22 200 newborn infants were screened out of 23 244 total live births, i.e. a total uptake of screening of 95.5% (Table 1). There were 9165 infants of UAE ances- try and 13 035 non-citizens. A total of 342 newborns screened posi- tive for haemoglobinopathies (Table 2). In Abu Dhabi medical district, out of 178 pos- itive screening results 6 were sickle cell disease (FS), 1 had non-sickle haemoglo- binopathy (FE) and 171 were haemoglobin- opathy carriers (FAS, FAC, FAD or FAE). In Al-Ain medical district, 144 positive screenings were detected: 3 were sickle cell disease (FS), 3 had non-sickle haemo- globinopathy (FC or FD) and 138 were haemoglobinopathy carriers. In Western Region, there were no cases of sickle cell disease and 20 haemoglobinopathy carri- ers. The data indicated that the incidence of sickle-cell disease was 0.04% for total Table 1 Percentage uptake of the neonatal screening programme for haemoglobinopathies by medical districts of Abu Dhabi emirate Variable Abu Dhabi Al-Ain Western Total Region No. of live births 13 613 8 494 1 137 23 244 No. of neonates tested 12 830 8 354 1 016 22 200 % uptake 94.3 98.4 89.4 95.5 05 Incidence of haemoglobinopathies.pmd 12/20/2005, 2:35 PM302 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 303 Table 2 Haemoglobin screening results of neonates by medical districts of Abu Dhabi emirate Condition Haemo- No. of neonates globin Abu Dhabi Al-Ain Western Total pattern Region Sickle-cell diseasea FS 6 3 0 9 Non-sickle haemoglobinopathy Homozygous HbC FC 0 1 0 1 Homozygous HbD FD 0 2 0 2 Homozygous HbE FE 1 0 0 1 Haemoglobinopathy carrier Sickle-cell trait FAS 132 94 14 240 HbC carrier FAC 6 5 0 11 HbD carrier FAD 32 39 5 76 HbE carrier FAE 1 0 1 2 Total 178 144 20 342 Hb = haemoglobin. aSickle-cell disease: included 5 cases confirmed as homozygous SS and 1 case confirmed as S/β-thalassaemia. Table 3 Incidence of confirmed sickle-cell disease, non-sickle haemoglobinopathy and haemoglobinopathy trait carriers in Abu Dhabi emirate by nationality Condition UAE citizen Non-UAE citizen Total (n = 9165) (n = 13 035) (n = 22 200) No. % No. % No. % Sickle-cell diseasea 6 0.07 3 0.02 9 0.04 Non-sickle haemoglobinopathy Homozygous HbC 0 0.00 1 < 0.01 1 < 0.01 Homozygous HbD 1 0.01 1 < 0.01 2 0.01 Homozygous HbE 0 0.00 1 < 0.01 1 < 0.01 Haemoglobinopathy carrier Sickle-cell trait 137 1.49 103 0.79 240 1.08 HbC carrier 3 0.03 8 0.06 11 0.05 HbD carrier 37 0.40 39 0.30 76 0.34 HbE carrier 0 0.00 2 0.02 2 0.01 Total 184 2.01 158 1.21 342 1.54 aSickle-cell disease: included 5 cases confirmed as homozygous SS and 1 case confirmed as S/β-thalassaemia. n = total number of neonates tested. UAE = United Arab Emirates. Hb = haemoglobin. 05 Incidence of haemoglobinopathies.pmd 12/20/2005, 2:35 PM303 304 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 screened neonates. The incidence by citi- zenship was 0.07% (1:1528 for UAE citi- zens and 0.02% for non-UAE citizens respectively (Table 3). It is noteworthy that 5 out of 6 sickle-cell disease cases for UAE citizens were confirmed as sickle-cell anaemia (homozygous HbS) with a inci- dence rate of 0.06% (1:1833) and only 1 case was confirmed as sickle/²-thalas- saemia, while all 3 sickle-cell disease cases in non-UAE citizens were confirmed as sickle-cell anaemia (homozygous HbS). The incidence of sickle-cell traits were 1.08% overall (1:92) (1.49% for UAE citi- zens and 0.79% for non-UAE citizens re- spectively). For the non-sickle haemoglo- binopathies, the incidence for UAE citizens were 0.40% for HbD trait and 0.03% for HbC trait (Table 3). Our data indicated that no parents with previously undiagnosed haemoglobin dis- ease were discovered and only 10 couples at risk for future children with a sickle-cell disease were identified by follow-up of newborns with haemoglobinopathy carrier traits: 4 were AS/AS, 4 were AS/²-thalas- saemia trait and 2 AS/±-thalassaemia trait (Table 4). No cases of sickle cell disease or sickle cell trait were recorded from hospital records among the neonates screening nor- mal and there was no disagreement be- tween confirmed and positive screened cases. Thus the confirmed neonatal screening results had an apparent sensitivi- ty of 100% and a specificity > 99%. Of the neonates screening positive for haemoglobin diseases, 57% (4 of 7 positive cases) had a positive family history and of those screening positive for haemoglobin- opathy trait, 48% (85 of 177 traits) had positive consanguinity in the UAE-citizen group. Discussion The primary objective of screening new- born infants for haemoglobinopathies is the identification of sickle-cell disease before the onset of symptoms. Mortality from sickle-cell disease is greatest during the second 6 months of life and is due to infec- tions. Thus, newborn screening allows prophylactic antibiotics to be started at 2 months to prevent such infections [78]. This pilot scheme for testing for sickle- cell disease in Abu Dhabi used the same heel-prick blood sample as for the existing national neonatal screening programme for phenylketonuria and congenital hypothy- roidism. HPLC was used as the primary screening method because it has quantita- tive capabilities as well as higher sensitivity and specificity than isoelectric focusing or two-tier electrophoresis (cellulose acetate and citrate agar electrophoresis). In addi- tion, variant HPLC testing is fully automat- ed, allowing for hundreds of samples to be run daily and the programme can detect Table 4 Results of follow-up of neonates with haemoglobinopathy carrier traits, showing parents at risk of future infants with haemoglobinopathies Screening result No. of couples AS / AS 4 AS / β-thalassaemia trait 4 AS / α-thalassaemia trait 2 AC / AC 2 AC / β-thalassaemia trait 2 AC / α-thalassaemia trait 2 AD / AD 4 AS = sickle-cell trait. AC = haemoglobin C trait. AD = haemoglobin D trait. 05 Incidence of haemoglobinopathies.pmd 12/20/2005, 2:35 PM304 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 305 HbF, HbA, HbS, HbC, HbD and HbE in a 2.5 minute run-time [9]. The haemoglobin pattern for normal newborn infants, FA, shows that the red cells of contain fetal (HbF) and adult (HbA) haemoglobins, with the haemoglobin with the highest concentration listed first. The haemoglobin pattern or phenotype is due to predominance of HbF at birth. Newborns with sickle-cell trait have an FAS pheno- type, with more HbA than HbS. Infants with sickle-cell disease, sickle cell/²-thalas- saemia and HbS/hereditary persistence of fetal haemoglobin (Hb-S/HPFH) each have an FS phenotype on screening at birth. Fol- low-up and confirmatory testing should be mandatory for all clinically significant re- sults [10]. Our study achieved a total coverage of screening for Abu Dhabi emirate of 95.5%, which is comparable to the international coverage standard (99%) [11]. The figures for sickle-cell disease (0.07%) and sickle-cell trait (1.49%) for UAE citizens in our study are slightly lower than earlier studies done in UAE, which found an incidence of 1.9% for sickle-cell trait [12]. Saudi Arabia has an incidence of 2%–27% for sickle-cell trait and about 1.4% for sickle-cell anaemia with the high- est rates in the Eastern Region and the low- est in the Central Region of the country [13,14]. Bahrain, on the other hand, has a rate of 11%–18% for sickle-cell trait, and an incidence of sickle-cell disease of 2.1% among screened newborns [15]. Oman is reported to have an incidence of 10% for sickle-cell trait [16] and about 0.4% for sickle-cell anaemia [17]. No false negative or false positive re- sults were detected in the confirmation of the neonatal screening results in our study. These results are in accordance with re- ported studies in many countries using HPLC as a primary screening method [9,18]. As regards the consanguinity rate for citizens, there was no difference between the consanguinity rate among the positive haemoglobinopathy carrier screening in our study (48%) and the normal UAE popula- tion (51%) [19]. It is noteworthy that all infants diag- nosed with sickle-cell disease started pro- phylactic penicillin therapy by the age of 2 months. This is an important, because the success of a newborn screening pro- gramme for sickle-cell disease is measured not only by the number of infants diag- nosed correctly, but more important, by the timely fashion in which these children re- ceive appropriate medical care. No parents with previously undiag- nosed haemoglobin disease were discov- ered and only 10 couples at risk for future children with a sickle-cell disease were identified by follow-up of newborns with carrier traits. So an infant with the trait pro- vides a “genetic window” into a family that may be at-risk for having a child with sick- le-cell disease. Another objective of the haemoglobin- opathy screening follow-up services is the education of families so that they will be more knowledgeable about haemoglobin- opathy carrier traits and not confuse benign carrier status with disease [20]. Conclusions and recommendations The results of this 1-year pilot study for newborn screening for sickle-cell haemo- globinopathies suggest that universal neo- natal screening for sickle-cell haemoglobin should be considered at the national level, in addition to developing comprehensive health strategies such as genetic counsel- ling services, comprehensive early care and premarital and prenatal diagnosis for hae- moglobin disorders. 05 Incidence of haemoglobinopathies.pmd 12/20/2005, 2:35 PM305 306 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Acknowledgements We would like to thank members of region- al screening programme Dr Reem Zayed and Dr Shadya Hassan, all the participating laboratory staff, our colleague Mrs Aysha A. from the department of Maternal and Child Health for her assistance, and M.R. Nasser Al-Oudaly, laboratory manager of Central and Al Jazira Hospitals, Abu Dhabi, United Arab Emirates. References 1. Consensus Development Panel, Na- tional Institutes of Health. Newborn screening for sickle-cell disease and other hemoglobinopathies. Journal of the American Medical Association, 1987, 258:1205–9. 2. Vichinsky E et al. Newborn screening for sickle-cell disease: effect on mortality. Pediatrics, 1988, 81:749–55. 3. Lane PA. Sickle-cell disease. Pediatric clinics of North America, 1996, 43:639– 64. 4. Weatherall DJ, Clegg JB. Inherited hae- moglobin disorders: an increasing glo- bal health problem. Bulletin of the World Health Organization, 2001, 79 (8):704– 12. 5. Grossman LK et al. Neonatal screening and genetic counseling for sickle-cell trait. American journal of diseases of children, 1985, 139:241–4. 6. Papadea C, Cate JC. Identification and quantification of hemoglobins A, F, S, and C by automated chromatography. Clinical chemistry, 1996, 42:57–63. 7. Strickland DK, Ware RE, Kinney TR. Pit- falls in newborn hemoglobinopathy screening: failure to detect beta(+)- thalassemia. Journal of pediatrics, 1995, 127:304–8. 8. Leikin SL et al. Mortality in children and adolescents with sickle-cell disease. Co- operative Study of Sickle-cell Disease. Pediatrics, 1989, 84:500–8. 9. Lorey F et al. Universal screening for he- moglobinopathies using high perfor- mance liquid chromatography: results of 2.2 million screens. European journal of human genetics, 1994, 2:262–71. 10. Loomis SJ et al. An automated system for sickle-cell screening. American clinical laboratory, 1990, October:33–9. 11. Weston M. Screenwatch: neonatal screening. National Audit of Neonatal (Guthrie) screening program. Bandolier journal, 1995, April:14–7 (http:// www.jr2.ox.ac.uk/bandolier/band14/ b14-7.html, accessed 23 March 2005). 12. White JM et al. Red cell genetic abnor- malities in Peninsular Arabs: sickle hae- moglobin, G6PD deficiency, and alpha and beta thalassaemia. Journal of medi- cal genetics, 1986, 23:245–51. 13. El-Hazmi M. Haemoglobinopathies, thalassaemias and enzymopathies in Saudi Arabia: the present status. Saudi medical journal, 1992, 13:488–99. 14. Nasserullah Z et al. Neonatal screening for sickle-cell disease, glucose-6-phos- phate dehydrogenase deficiency and al- pha thalassemia in Qatif and Al Hasa. Annals of Saudi medicine, 1998, 18(4):289–92. 15. Mohammed AM et al. Haemoglobinopathies and glucose-6- phosphate dehydrogenase deficiency in hospital births in Bahrain. Annals of Saudi medicine, 1992, 12:536–9. 05 Incidence of haemoglobinopathies.pmd 12/20/2005, 2:35 PM306 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 307 16. Rajab A. Prevention of genetic disorders in Oman. The Ambassadors (online magazine), 1998, 1(3):5 (http:// ambassadors.net/archives/issue3/se- lect-oman.htm, accessed 29 March 2005). 17. Teebi AS, Farag TI. Genetic disorders among Arab populations. Oxford, Oxford University Press, 1997. 18. Shafer FE et al. Newborn screening for sickle-cell disease: 4 years of experi- ence from California’s newborn screen- Newsletter of the Centre for Arab Genomic Studies (CAGS) In May 2005, CAGS launched its first newsletter, CAGS Newsletter Issue 1, May 2005. It marks the beginning of a new service that CAGS will offer on a bimonthly basis to the international scientific community. Researchers working on health-related aspects are in- vited to contribute to future issues of the CAGS Newsletter and com- ments on the content of the present and future issues of the CAGS Newsletter are welcome. The newsletter can be accessed on line at: http://www.cags.org.ae/newsletter.html. CAGS was established to characterize and prevent genetic disorders and transform the future practice of health care in the region. Further information about CAGS is available at: Centre for Arab Genomic Studies, PO Box 22252, Dubai, United Arab Emirates. Telephone: +971-4-398 6 777; Fax: +971-4-398 0 999; E-mail: cags@emirates.net.ae. Further information about CAGS is available at: www.cags.org.ae ing program. American journal of pediat- ric hematology/oncology, 1996, 18(1):36–41. 19. Fahmy NA, Benson PF, Garrah DB. Con- sanguinity in the UAE: prevalence and analysis of risk factors. Emirates medical journal, 1993, 1(suppl.):39–41. 20. Githens JH et al. Newborn screening for haemoglobinopathies in Colorado. American journal of diseases of chil- dren, 1990, 144:466–70. 05 Incidence of haemoglobinopathies.pmd 12/20/2005, 2:35 PM307 308 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Evaluation of the national health policy of thalassaemia screening in the Islamic Republic of Iran N. Ghotbi1 and T. Tsukatani2 1Graduate School of Economics; 2Division of Environment and Resources, Institute of Economic Research, Kyoto University, Kyoto, Japan (Correspondence to N. Ghotbi: naderq@hotmail.com). Received: 21/11/03; accepted: 08/12/03 ABSTRACT We reviewed the medical and economic burden of thalassaemia major with emphasis on prenatal diagnosis for disease prevention as the most economic health care policy approach. The current programme in the Islamic Republic of Iran screens couples just before marriage, identifies carriers and refers them for genetic counselling. We searched the current literature for a refined model and enquired into compliance issues in interviews with physicians, couples and families with affected children. The programme was unsatisfactory in comparison with comparable programmes in the Mediterranean region. We devised a simple decision tree that incorporates cost–effectiveness and technical, methodological and social issues that affect compliance. While revisions to the policy could improve efficiency, follow-up is needed, especially to provide prenatal diagnosis for carrier couples. Évaluation de la politique nationale de santé en matière de dépistage de la thalassémie en République islamique d’Iran RÉSUMÉ Nous avons examiné le fardeau médical et économique de la thalassémie majeure en s’intéressant plus particulièrement au diagnostic prénatal pour la prévention de la maladie en tant qu’approche la plus économique de la politique de soins de santé. Le programme actuel en République islamique d’Iran effectue le dépistage des couples juste avant le mariage, identifie les porteurs et les oriente vers un service de conseil génétique. Nous avons effectué une recherche dans la littérature actuelle pour trouver un modèle amélioré et avons enquêté sur les questions de compliance dans des entretiens avec des médecins, des couples et des familles ayant des enfants affectés par cette maladie. Le programme n’était pas satisfaisant par rapport à des programmes comparables dans la Région de la Méditerranée orientale. Nous avons mis au point un arbre décisionnel simple incorporant le rapport coût-efficacité et des points techniques, méthodologiques et sociaux qui affectent la compliance. Alors que les révisions de la politique pourraient améliorer l’efficacité, un suivi est nécessaire, notamment pour assurer le diagnostic anténatal pour les couples porteurs. 06 Evaluation of the national health.pmd 12/20/2005, 2:35 PM308 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 309 Introduction Every year approximately 60 000 thalas- saemic babies are born worldwide. Optimal management of thalassaemia major re- quires specialized medical services in equipped centres, and over the lifetime of a patient the cost of care is extremely expen- sive [1]. In countries where the incidence of thalassaemia is high with a big burden of the disease to the population and the econo- my, intensive preventive programmes have been established and have been very suc- cessful [2–4]. This disease is a prime ex- ample in which prevention has primary importance and priority over treatment, thus prenatal diagnosis has been the main- stay of most control programmes thus far. Chorionic villus sampling (CVS) and amniocentesis have been used to obtain samples for genetic analysis. Prenatal diag- nosis of thalassaemia major is made using molecular techniques to test these samples, but accurate characterization of the molec- ular abnormality depends upon knowledge of the abnormal genetic forms, or thalas- saemia variants, in each community [5–7]. Prenatal diagnosis as the preventive strate- gy theoretically requires the termination of 25% of such pregnancies by abortion; ther- apeutic abortion is the final tool of preven- tion in standard control programmes. Thalassaemia in the Islamic Republic of Iran Beta thalassaemia is the most common he- reditary disease in the Islamic Republic of Iran. All patients with thalassaemia major are registered by regional health care cen- tres. The country has about 26 000 thalas- saemia major patients and an estimated 3 million thalassaemia carriers. The latter fig- ure is easily calculated as the beta thalas- saemia carrier frequency in the country has been estimated to be 4%–5% in screening studies and the July 2000 census reported a population of 65 619 636. This gives us an estimated 2 620 000–3 280 000 carriers (mean 2 950 000). Furthermore, it has been estimated that about 8000 pregnancies are at risk each year. The prevalence of thalassaemia major varies throughout the country with the highest incidence in regions near the Caspi- an Sea and the Gulf. The provinces of Ma- zandaran, Gilan, Hormozgan, Khuzestan, Kohkiluyeh-Boyerahmad, Fars, Bushehr, Sistan-Baluchestan, Kerman and Isfahan are the 10 provinces most afflicted. In 1995–1996, the Blood Transfusion Organization of Iran and the Iran Thalas- saemic Patients Supporting Society mass screened 8300 students from 114 high schools in Tehran for the beta thalassaemia trait. The project lasted for 2 months. Based on the results of that study, the government included beta-thalassaemia screening laboratory tests in the compulso- ry package of medical tests for all couples prior to marriage [8]. This standard control programme for thalassaemia, however, has not been practical because of religious un- certainties about issuing permits for legal therapeutic abortions. Instead, identifying “both carrier” fiancés and providing coun- selling to dissuade them from marriage, or not allowing them to marry, has been em- phasized. Even those not ethically opposed to therapeutic abortion have recommended this alternative method, because access to fetal sampling techniques and laboratory di- agnosis are very expensive. The cost issue is especially important in the Islamic Republic of Iran because the dominant public insurance systems do not cover expenses not in line with well- defined government policies. A few couples have tried sending the samples to countries with high-tech facilities for laboratory di- 06 Evaluation of the national health.pmd 12/20/2005, 2:35 PM309 310 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 agnosis like the United Kingdom, but in- complete knowledge of the variety of the diseased genes in Iranian communities, the lengthy wait for the results and the high costs have interfered. Screening for beta thalassaemia trait Approximately 3% of the world population are carriers of a beta thalassaemia muta- tion. Mass screening and genetic counsel- ling programmes have been carried out in many countries with high carrier states, but some factors have undermined the effec- tiveness of antenatal screening for preven- tion of thalassaemia [9–11]. For example, many medical practitioners and the general public are still unaware of screening proce- dures or do not comply and so new patients are still diagnosed in areas where the dis- ease has been locally preventable by ante- natal screening [12,13]. Defining national health policies to mass screen for carriers could be a solution and forms the basis of thalassaemia control programmes. In an economic sense, there is a trade- off between the costs of diagnosis and pre- vention and the treatment of undiagnosed cases. The system must consider the cost of acceptable, although suboptimal, care for all thalassaemia cases including those births that could have been avoided. The cost of care for other health care problems more common in thalassaemics, the time lost by parents caring for their children and the emotional burden for families dealing with this health problem should be consid- ered too. In ideal circumstances the benefit to cost ratio is larger than 1. This supports the notion of community-based screening for prevention; therefore, choosing a suit- able economic tool for mass screening is an important step. The gold standard for diagnosis of the carrier state is haemoglobin electrophore- sis, but because of its high cost, it is un- suitable for mass screening in developing countries with large populations like the Is- lamic Republic of Iran [14]. Most screen- ing programmes use an initial simple but sensitive test, such as red cell mean cor- puscular volume (MCV), or an osmotic fragility test [15]. The Naked Eye Single Tube Red Cell Osmotic Fragility Test (Nestroft) is a sensitive, cost-effective, rapid and reliable screening test for detec- tion of beta thalassaemia trait, or the carrier state, in a population. A positive predictive value and specificity of 85%–100%, a neg- ative predictive value of 83%–99% and a sensitivity of about 95% have been report- ed for Nestroft [16,17]. Nestroft as a single screening parameter is superior to any oth- er simple tests like MCV and is more cost- effective [18]. As a result, Nestroft has emerged as the single most effective, inex- pensive and easily reproducible test of pop- ulation screening for beta thalassaemia trait. Nestroft in combination with MCV with the cut-off at < 80 fL has been proven to be 100% sensitive, but the combination is not cost-effective. The screening method in the Islamic Republic of Iran is unclear and is some- times confusing as usually only a complete blood count (CBC) is requested and many variables in the report are evaluated without specific measures or standard criteria. Nonetheless, as reported in our interviews, MCV seems to be the most important item to informed physicians who judge the re- sults of the tests. They usually provide counselling as a simple explanation of the inheritance mechanism and prognosis of thalassaemia major, thereby keeping coun- selling expenses to just the doctor’s visit 06 Evaluation of the national health.pmd 12/20/2005, 2:35 PM310 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 311 fees, especially in small cities where genet- ics counselling by specialists is unavailable. Our study was a search of the current literature for information pertaining to the cost–effectiveness of the Iranian national screening programme in comparison to programmes in other countries. We also interviewed physicians, laboratory doctors and individuals personally affected by thalassaemia major. Our aim was to devise a simple decision tree that could be used to improve cost efficiency. Methods We searched electronic databases (includ- ing Medline, PubMed, EMBASE and CINAHL), information gathered by interna- tional and domestic thalassaemia support organizations, Iranian national health data- bases, collections of university periodicals and theses, research projects, presenta- tions and reports of seminars and govern- ment reports with the following queries: “screening and thalassaemia”, “Iran and thalassaemia” and “economics and thalas- saemia”. We sorted the retrieved data by abstracts and then we studied the full texts. For more detailed information on the Iranian control programme and to examine its medical, laboratory and technical, so- cial, ethical and patient group perspectives, a series of interviews were performed. We visited 5 health centres regularly providing services for thalassaemic patients, includ- ing clinics in Bahrami Hospital, Imam Khomeini Hospital and Amir-Kabir Hospital, and we interviewed 18 medical doctors, in- cluding 2 haematologists, 2 medical labora- tory doctors (PhDs), 3 paediatricians, 4 obstetric-gynaecologists and 7 general physicians working in the screening sys- tem; 27 families (at least 2 members from any one family) of thalassaemic patients; and 9 couples referred for haemoglobin electrophoresis tests (because of suspi- cious CBC test results). These visits served as a problem finding stage in which we identified commonly reported issues. We then further searched the literature for op- tions with higher success rates and fewer problems, especially in other Mediterranean countries. Finally, by looking at the programme’s pitfalls with cost–effectiveness concerns and at the reviewed articles for potential methodological and technical improve- ments, a refined model was designed as a simple decision analysis tree. Results Thalassaemia control programme Fars Province, in the south-eastern part of the country, is home to 5% of the popula- tion and nearly 10% of the thalassaemic patients of the Islamic Republic of Iran. In 1997–1998, a study evaluated thalassaemia in Fars. Ten years prior to the study, it had been estimated that the beta thalassaemia carriage rate was 7% in the province and a prevention programme had been initiated [19]. The 1997–1998 programme evaluat- ed the prevalence of beta thalassaemia mi- nor by examining 24 485 boys in their final year of high school. The frequency of beta thalassaemia minor in 1997–1998 was then estimated at 6.88% [19]; there were 2193 thalassaemia major patients (1263 boys and 930 girls) and the frequency of thalas- saemia major was 7.2 in 10 000 population. The mean age of patients was 10.1 years with a SD of 6.6 years. A large part of the decrease in prevalence of thalassaemia ma- jor was among those under age 10 years and was less significant after adjustment for population age distribution. The slight decrease was determined to be the result of a birth control programme, rather than the programme for the prevention of beta 06 Evaluation of the national health.pmd 12/20/2005, 2:35 PM311 312 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 thalassaemia. Studying the reasons for the relative failure of the programme and set- ting up new strategies were recommended as the programme had been operating for 10 years and was only moderately success- ful. Other studies found that the number of children born annually with congenital dis- orders in the country had fallen by 38% and potential births of children with thalas- saemia major had fallen from over 1200 to about 860 per year while the overall popula- tion had increased [20,21]. More recent evidence from population surveys, howev- er, indicates that the Islamic Republic of Iran has actually reduced its population growth to 1.2%, a rate only slightly higher than that of the United States of America. The country’s population growth rate dropped from an all-time high of 3.2% in 1986 to 1.2% in 2001, one of the fastest drops ever recorded. The results of the former study [19] explained much of the success claimed in the latter studies [20,21]. The claimed decrease in thalas- saemic births is not as great as in some en- demic areas in the Mediterranean where long-established control programmes have achieved 80%–100% prevention of newly affected births [22]. Interviews We interviewed 18 medical doctors, 27 families of thalassaemic patients, and 9 couples referred for haemoglobin electro- phoresis tests because of suspicious CBC tests. This was our final problem finding stage and commonly reported issues in- cluded: the financial difficulties of the pa- tients’ families; lack of access to or lack of insurance funding for chelation therapy; unnecessary anxiety and confusion in the first round of routine screening; difficulties performing prenatal tests; and difficulties obtaining legal abortions of thalassaemic fetuses. With this information, we charted the current programme as a simple decision tree (Figure 1). Then, using information from our literature search, we devised a decision tree that is more cost-effective (Figure 2). Discussion Economic aspects of thalassaemia The economic burden of thalassaemia is determined by the birth prevalence of af- fected infants, their survival, the availabili- ty, cost and effectiveness of treatment, Figure 1 The flow of the screening procedure in the current thalassaemia programme presented as a simple decision tree 06 Evaluation of the national health.pmd 12/20/2005, 2:35 PM312 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 313 social support and the extent to which pa- tients can be integrated into society. The need for care and its costs at any given time depend on the number of living patients in the population. The availability of better health services leads to a cumulative in- crease in the number of patients needing care and also to a higher annual cost per patient because cost of care usually rises as treatment options improve. The lifetime cost of optimum treatment in developed countries has been estimated to be about US$ 1 350 000 (undiscounted) per patient, and with a discount rate of 6%, about US$ 370 000. In a sensitivity analysis, the dis- counted cost ranged from about US$ 315 000 to US$ 380 000 [23]. In most countries, children with thalas- saemia major are easily diagnosed, and reg- ular blood transfusion is available, is life saving and gives excellent short-term qual- ity of life. Without more expensive therapy to remove excess iron, death from iron overload results at an early age; iron chela- tion therapy is needed for long-term surviv- al. There are many costs associated with thalassaemia, including blood transfusions, medicines and other essential treatment, hospital care and home visits. Thalassaem- ic patients receive red blood cell transfu- sions every 2–3 weeks, amounting to as many as 24.6 litres of blood a year. In the Islamic Republic of Iran, 50%–60% of all donated blood is transfused to thalassaemic patients. Each bag of donated blood costs at least US$ 25 to the country’s Blood Transfusion Organization (personal com- munication). Desferrioxamine is imported and is currently (2001) about US$ 3 per vial of 500 mg. A 12-year-old, 30 kg patient on regular blood transfusion will require three vials daily throughout his lifetime. Includ- ing the cost of consumables like disposable syringes and scalp-vein needles, the expen- diture for Desferal therapy is US$ 280 per month. The infusion pump is also imported and costs US$ 400. Many costs to the pa- tients are financed by plans in the public budget, and yet most thalassaemic patients cannot afford the cost of optimum treat- ment and die before they reach 20 years of age. Major public health insurance policies discontinue coverage when the child reach- es the legal age. This is yet another finan- Figure 2 The flow of our proposed screening procedure, which introduces cost-effective refinements to the current programme, presented as a simple decision tree 06 Evaluation of the national health.pmd 12/20/2005, 2:35 PM313 314 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 cial problem that thalassaemic patients and their families face, as they often do not have the resources to cover the costs. This can lead to a further fall in the quality of the care they receive and can have grave con- sequences on the patient’s longevity. A great number of patients hence rely on charity and support organizations that pro- vide the needed assistance. The estimated annual and 10-year projected costs for treating thalassaemia and the estimated an- nual cost of prevention in Eastern Mediter- ranean countries have been compared by the World Health Organization [24,25]. In the Islamic Republic of Iran, the av- erage birth prevalence of thalassaemia ma- jor is 0.74 per 1000, and the proportion of patients expected to survive at least into their late teens has risen from practically none 30 years ago to over 90% today based on Iranian thalassemic registries. Over 25 000 patients are currently under care, representing a forward commitment of approximately US$ 220 million per year for at least the next 30 years. In the absence of prevention, annual treatment costs could rise to over US$ 700 million per year. The treatment of thalassaemic patients is a huge expense on the national health budget, and therefore we propose that government and public insurance organizations actively look for eligible resources to continue insurance coverage until patients are employed. Spending on preventive measures is more cost-effective than treatment in the Islamic Republic of Iran. As annual preven- tion costs are constant but annual treatment costs rise yearly, the cost–effectiveness of prevention increases with every year that it is in place. The 10-year projection figures suggest that it is impossible for the country to finance optimal treatment for all patients who may be born, and that effective pre- vention is a necessary condition for those already living to be adequately treated. Health policy aspects National programmes of community infor- mation, carrier screening and counselling, and availability of prenatal diagnosis have greatly reduced the birth prevalence of thalassaemia major in all countries where they have been established. In the Islamic Republic of Iran, a national prevention pro- gramme has been under development and the national patient register has shown a re- duction, although insufficient, in affected births [26]. The health policy aspect of the control programme has a few problems that need to be addressed. The routine prevention of the birth of thalassaemic children by prena- tal diagnosis has not been practically possi- ble. The religious regulation, hokm, prohibits any kind of abortion after the fourth month of pregnancy. It does allow abortions before that stage, but only when the mother’s life is endangered and provid- ed that a lengthy procedure involving a council committee documents the case. Af- ter specialists petitioned the current reli- gious leader, or faghih, about thalassaemia major, they received a permit for abortion only for cases of thalassaemia major before the fourth month of pregnancy. Initially the permit seemed enough, but in practice it has not been successful, as physicians have tended to be conservative and to hesi- tate to issue definitive opinions. Further- more, obtaining test results is a lengthy matter. Chances of success for the legal option have thereby become so remote that in one daily newspaper a few specialists warned of and criticized the “legal void” that makes families resort to illegal abor- tions for thalassaemic babies. This problem is not unique to the Islamic Republic of Iran. In the United Kingdom, only a minor- ity of couples of Hindu, Sikh and Pakistani origin accepted fetal diagnosis; termination 06 Evaluation of the national health.pmd 12/20/2005, 2:35 PM314 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 315 of pregnancy within the second trimester and alternative methods of control were recommended when this attitude prevailed [27]. The health policy in the Islamic Repub- lic of Iran is to use alternative means to pre- vent the birth of thalassaemic children by avoiding marriages of carriers to each other [20]. In a large, 3-year study of 100 000 cases in Isfahan, after identification and genetic counselling of carriers, an average of 90% of high-risk couples decided not to marry and no new cases of thalassaemia were detected in the children of the screened population. It was concluded that when both members of the couple were trait-positive their preferred choice was not to marry, rather than to marry and use oth- er or no methods to prevent the birth of a thalassaemic child. Cultural and religious ideas seemed to have a role in these deci- sions and the establishment and use of a genetic counselling centre seemed to pre- vent most new thalassaemia cases by dis- couraging at-risk couples from marriage [20]. Although more studies are needed to confirm this, some have already suggested a different explanation [19]. Moreover, 10% of at-risk couples in the Isfahan study that insisted on marrying, reported trying to register the marriage in another area outside Isfahan. The study did not follow them nor did it include their offspring in its results. Results of the 4-year premarital screen- ing programme to reduce the incidence of thalassaemia major in the city of Denizli, Turkey, are interesting [6]. At-risk couples were counselled and offered prenatal diag- nosis and termination of pregnancy in cas- es of an affected fetus. This study demonstrated that premarital screening in the context of offering therapeutic abortion was very useful for detecting carrier cou- ples and was an effective way of control- ling thalassaemia major. It also showed that only a small proportion of cases were likely to face the final decision to terminate preg- nancy or not. In another study, prevention programmes based on carrier screening and genetic counselling in the absence of prenatal diagnosis produced no consistent effect on the birth rate of thalassaemia ma- jor [28]. Conclusions The pitfalls in the current thalassaemia con- trol programme could be addressed with a simple decision tree. Figures 1 and 2 show the scheme of the current programme and our proposed revisions. The current programme screens both members of every couple applying for mar- riage with MCV indexes. In the standard package of premarital tests in the Islamic Republic of Iran, men provided a sample of blood to be tested by VDRL for syphilis screening and a sample of urine to be tested for addiction. Women only provided urine samples before the current programme be- gan. We propose limiting the use of MCV index for thalassaemia screening to only men as this has many advantages over screening both men and women. Relying on microcytosis as revealed on MCV for thalassaemia minor screening has led to a large number of false positives as the num- ber of iron deficient women is very high. The Ministry of Health and Medical Educa- tion reported a 17% rate of iron deficiency anaemia (34% iron deficiency) in women of childbearing age in the Islamic Republic of Iran (project report to the World Bank, 2000) [29]. Iron deficiency is the most common cause of acquired microcytosis worldwide and is prevalent in developing countries like the Islamic Republic of Iran. With fewer false positives, there will be fewer referrals to doctors and less psycho- 06 Evaluation of the national health.pmd 12/20/2005, 2:35 PM315 316 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 logical trauma of a hereditary disease stig- ma. Moreover, performing half the number of tests would reduce current costs of screening. The use of CBC tests in the current pro- gramme causes confusion and could be largely avoided if the first screen were for men only or if the cheaper and more accu- rate Nestroft were to be used for screen- ing. At very least, the current screening tool and its criteria should be standardized with specific cut-off points for MCV and with collaboration between the marriage registration offices requesting the screen, the laboratories performing the test, and the physicians judging borderline cases. In the current programme, carrier cou- ples are prohibited to marry: this endangers the compliance rate and adds to the stigma of being labelled a carrier. In our analysis, providing genetic counselling and asking the couple to consider the procedures nec- essary to prevent the birth of a thalassaem- ic child and the costs and risks, including death of healthy fetuses because of sam- pling procedure, is enough and is more effi- cient. In the long term, we could include a suitable screening test in the health check package of schools and could provide in- formation with results of the test as is done in the Gaza Strip [30]. The matter could then be considered before choosing a part- ner for marriage instead of waiting until just before the couple prepares for the cer- emony as in the current policy. The current policy does not provide the necessary com- pliance rate and is not ethical in this re- spect. Iranian scientists have been able to de- tect up to 85% of cases (70%–90% for dif- ferent geographic and ethnic origins) [31,32]. More successful prenatal diagno- sis in the Iranian population requires more studies to identify mutant genes. Finally, the current thalassaemia control programme has ignored a prevention policy for already married couples that may or may not already have affected children. The considerable number of families with more than one affected child seen in some studies is strong proof. Provision of a working option of prenatal diagnosis and therapeutic abortion for this group of cou- ples is necessary. References 1. Weatherall DJ, Clegg JB. Inherited hae- moglobin disorders: an increasing glo- bal health problem. Bulletin of the World Health Organization, 2001, 79(8):704– 12. 2. Indaratna K. Screening for thalassemia: an economics viewpoint. Southeast Asian journal of tropical medicine and public health, 1997, 28(suppl. 3):75–81. 3. Mitchell JJ et al. Twenty-year outcome analysis of genetic screening pro- grammes for Tay-Sachs and beta- thalassaemia disease carriers in high schools. American journal of human ge- netics, 1996, 59(4):793–8. 4. Altay C et al. Premarital screening of he- moglobinopathies: a pilot study in Turkey. Human heredity, 1996, 46(2):112–4. 5. Curuk MA et al. Genetic heterogeneity of beta-thalassaemia at Cukurova in south- ern Turkey. Hemoglobin, 2001, 25(2): 241–5. 6. Keskin A et al. Premarital screening of beta-thalassaemia trait in the province of 06 Evaluation of the national health.pmd 12/20/2005, 2:35 PM316 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 317 Denizli, Turkey. Acta haematologica, 2000, 104(1):31–3. 7. Khan SN, Riazuddin S. Molecular char- acterization of beta-thalassaemia in Pakistan. Hemoglobin, 1998, 22(4): 333–45. 8. Emadi A, Koorosdari H, Pakbaz Z. Thalassaemia. Mass screening of beta thalassaemia minor in high school stu- dents of Tehran (1995–1996). Biannual journal of Iran. Thalassemic Patients Supporting Society, 1997, vol. 12 (in Per- sian with English abstract). 9. Scriver CR et al. Beta thalassaemia dis- ease prevention: genetic medicine ap- plied. American journal of human genetics, 1984, 36:1024. 10. Hendy J. Prevention of thalassaemia in Australia. Southeast Asian journal of tropical medicine and public health, 1999, 30(suppl. 2):94–6. 11. Ko TM, Xu X. Molecular study and prena- tal diagnosis of alpha- and beta- thalassaemias in Chinese. Journal of the Formosan Medical Association, 1998, 97(1):5–15. 12. Lee AC et al. Prevention of beta- thalassaemia major by antenatal screening in Hong Kong. Pediatric haematology and oncology, 1998, 15(3): 249–54. 13. Yong KN et al. Thalassaemia carrier screening and prenatal diagnosis among the British Columbia (Canada) population of Chinese descent. Clinical genetics, 1999, 55(1):20–5. 14. Altay C et al. Premarital screening of he- moglobinopathies: a pilot study in Turkey. Human heredity, 1996, 46(2):112–4. 15. Jaovisidha A et al. Prevention and con- trol of thalassaemia in Ramathibodi Hos- pital, Thailand. Southeast Asian journal of tropical medicine and public health, 2000, 31(3):561–5. 16. Thool AA et al. A simple screening test for the detection of heterozygous beta thalassaemia. Indian journal of pathol- ogy and microbiology, 1998, 41(4):423– 6. 17. Gomber S, Sanjeev, Madan N. Validity of Nestroft in screening and diagnosis of beta-thalassaemia trait. Journal of tropi- cal pediatrics, 1997, 43(6):363–6. 18. Manglani M et al. NESTROFT: an effec- tive screening test for beta thalassaemia trait. Indian pediatrics, 1997, 34(8):702– 7. 19. Karimi M, Alavian Ghavanini A, Kadivar MR. Regional mapping of the gene fre- quency of beta thalassaemia in Fars province, Iran during 1997–1998. Ira- nian journal of medical sciences, 2000, 25(3–4):134–7. 20. Ghanei M et al. Pre-marriage prevention of thalassaemia: report of a 100 000 case experience in Isfahan. Public health, 1997, 111(3):153–6. 21. Christianson A, Streetly A, Darr A. Les- sons from thalassaemia screening in Iran. British medical journal, 2004, 329(7475):1115–7. 22. Angastiniotis MA, Hadjiminas MG. Pre- vention of thalassaemia in Cyprus. Lan- cet, 1981, 1:369–71. 23. Karnon J et al. Lifetime treatment costs of beta-thalassaemia major. Clinical and laboratory haematology, 1999, 21(6): 377–85. 24. Alwan A, Modell B. Community control of genetic and congenital disorders. Alex- andria, World Health Organization, Regional Office for the Eastern Mediter- ranean, 1997. 25. Alwan A, Modell B. Recommendations for introducing genetics services in de- veloping countries. Nature reviews. Ge- netics, 2003 4(1):61–8. 06 Evaluation of the national health.pmd 12/20/2005, 2:35 PM317 318 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 26. Samavat A, Modell B. Iranian national thalassaemia screening programme. British medical journal, 2004, 329: 1134–7. 27. Cao A et al. The prenatal diagnosis of thalassaemia. British journal of haema- tology, 1986, 63:215. 28. Vullo C, Barry F. Population screening for carriers of recessively inherited disor- ders. Lancet, 1980, 2(8196):679–80. 29. Iran – second primary health care and nutrition project. Washington, World Bank, 2001 (Report number PID8574) Available at: http://www-wds.worldbank. org/servlet/WDSContentServer/WDSP/ I B / 2 0 0 0 / 0 3 / 2 1 / 0 0 0 0 9 4 9 4 6 _ 00031505340433/Rendered /PDF/ multi0page.pdf (accessed: 11/4/05). 30. Sirdah M et al. Screening secondary school students in the Gaza strip for beta-thalassaemia trait. Clinical and laboratory hematology, 1998, 20(5): 279–83. 31. Najmabadi H et al. The beta thalassae- mia mutation spectrum in the Iranian population. Hemoglobin, 2001, 25(3): 285–96. 32. Yavarian M et al. Molecular spectrum of beta-thalassaemia in the Iranian Prov- ince of Hormozgan. Hemoglobin, 2001, 25(1):35–43. A study of national health research systems in selected countries of the WHO Eastern Mediterranean Region A study of national health research systems in selected countries of the WHO Eastern Mediterranean Region comes at a time of global interest in promoting the development of appropriate national health research systems in low and middle-income developing countries. It presents the findings of a detailed and systematic situ- ation analysis of the national health research systems in five coun- tries of WHO’s Eastern Mediterranean Region: Egypt, Islamic Republic of Iran, Morocco, Pakistan and Sudan. The aim of the study was to pave the way for developing innovative and effective strate- gies for strengthening national health research systems in the Re- gion. The experience gained will be shared with other countries in the Region and will provide a critical input to WHO’s collaboration in the Region in the area of health research in the coming years. The publication can be obtained from: Distribution and Sales, World Health Organization Regional Office for the Eastern Mediterranean, Abdul Razzak Al Sanhouri Street, PO Box 7608, Nasr City, Cairo 11371, Egypt. Telephone: (202) 670 25 35; Fax: (202) 670 24 92/4; Email: emr:dsa@ emro.who.int. 06 Evaluation of the national health.pmd 12/20/2005, 2:35 PM318 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 319 Mental health publications from the Arab world cited in PubMed, 1987– 2002 M.M. Afifi1 1Department of Research and Studies, Ministry of Health, Muscat, Oman (Correspondence to M.M. Afifi: afifidr@yahoo.co.uk). Received: 30/09/03; accepted: 17/05/04 ABSTRACT The study aimed to identify mental health publications in Arab countries cited in PubMed from 1987 to 2002 and to analyse the year of publication, journal and field of research, specifically for child and adolescent mental health studies. A PubMed search was performed for all mental health publications followed by an analysis of child and adolescent mental health citations. The total number of mental health research citations published in Arab countries over the last 15 years was 338, 1.2% of the total number of citations for biomedical research; 17% were on child and adolescent mental health. The most dynamic fields of mental health research are anxiety and mood disorders and substance abuse. Child psychiatry, especially attention deficit hyperactivity disorder and child autism, are not gaining much interest from Arab researchers. Publications sur la santé mentale provenant du monde arabe citées dans PubMed, 1987-2002 RÉSUMÉ Cette étude visait à identifier les publications sur la santé mentale dans les pays arabes citées dans PubMed de 1987 à 2002 et à analyser l’année de publication, la revue et le domaine de recherche, notamment pour les études sur la santé mentale de l’enfant et de l’adolescent. Une recherche dans PubMed a été effectuée pour toutes les publications sur la santé mentale suivie d’une analyse des citations concer- nant la santé mentale de l’enfant et de l’adolescent. Le nombre total de citations de recherches en santé mentale publiées dans des pays arabes au cours des 15 dernières années était de 338 – 1,2 % du nombre total de citations pour la recherche biomédicale ; 17 % concernaient la santé mentale de l’enfant et de l’adolescent. Les domaines de recherche en santé mentale les plus dynamiques étaient l’anxiété et les troubles de l’humeur ainsi que la toxicomanie. La psychiatrie infantile, notamment le trouble d’hyperactivité avec déficit de l’attention et l’autisme de l’enfant, ne suscite pas beaucoup d’intérêt chez les chercheurs arabes. 07 Mental health publications.pmd 12/20/2005, 2:35 PM319 320 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Introduction The geographical distribution of publica- tions as an indicator of the research pro- ductivity of individual regions, countries or institutions has recently become a field of interest [1]. It has been investigated for European Union countries [2], the United States of America (USA) [3], Gulf Cooper- ation Council countries [4], Colombia [5], South Korea [6] and a Turkish medical school [7]. The research output of individ- ual and selected countries [8–12] for single specialties has also been reviewed. To the best of the author’s knowledge, such an investigation has not been carried out in Arab countries in the field of psychiatry and mental health. Gaining access to suitable databases is crucial for successful scientific work in the biomedical fields. The PubMed® online da- tabase is developed by the United States National Center for Biotechnology Infor- mation (NCBI) at the National Library of Medicine (NLM) and currently indexes over 12 million biomedical citations ex- tracted from more than 4600 journals, pub- lished in more than 70 countries and dating back to the mid-1960s [13]. PubMed is not only a simple search engine for biomedical citations, but also a powerful tool for con- ducting certain statistical analyses [14]. The World Health Organization focused its 2001 World health report for the first time on a single public health topic: mental health [15]. This decision was taken based on the increasing recognition that the mag- nitude and burden of mental disorders are high, effective treatments exist for most mental disorders, the great majority of those in need of effective treatments do not receive them and there are high and unnec- essary costs around the world in terms of suffering, disability and economic loss. Although there is some research activity in the field of psychiatry and mental health in the Eastern Mediterranean Region, planned, purposeful research programmes linked to the development and improvement of services and training are rare. In addi- tion, research programmes and activities are not aimed at the production of informa- tion systems through a national health plan; they are scattered, uncoordinated and often seem to be unrelated to needs [16]. This study was a geographical analysis of the number of psychiatry and mental health publications produced in 22 Arab countries over a 15-year period from 1987 to 2002. The aim was to estimate what pro- portion of each country’s research was de- voted to psychiatry and mental health and which countries were carrying out the ma- jority of mental health research in the Arab world; also to study the settings and types of mental health research for each Arab country during this period. Methods An initial search was made for all biomedi- cal studies published in 22 Arab countries from 1 January 1987 to 31 December 2002. The search made using the names of Arab countries in English as well in French and directing them to PubMed within a single hour limit on 4 July 2003 at the Faculty of Health Sciences, American University of Beirut. The country affiliation search was per- formed in English as well as French be- cause Arab biomedical scientists do not publish their results exclusively in English language journals and an important per- centage of publications appear in French journals. Moreover, even in English lan- guage journals, authors from the Arab 07 Mental health publications.pmd 12/20/2005, 2:35 PM320 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 321 Maghreb countries and Lebanon prefer to report their research in French. The ratio of French to English citations is about 1:4 in Lebanon, 1:1 in Algeria, rising to 2:1 in Tu- nisia, Morocco and Mauritania. The aim was to minimize false negative results, i.e. excluding publications in which the coun- try affiliations are written in French. Anoth- er common error revealed in previous studies was false positives, especially for Lebanon and Jordan, since there are several cities in the USA called Lebanon and many USA universities have a Jordan Hall. Most entries where “USA” was in the full affilia- tion name were manually deleted and accu- rate publication counts were produced for each country. An example of a query is “Morocco[affiliation] OR Maroc[affili- ation]”, in which the Boolean operator “OR” had the function of collecting all pub- lished articles carrying any of the English or French names for Morocco in the affili- ation field [14,17,18]. The second stage was to search in the Medical subject heading (MeSH) database of PubMed for articles about psychiatry and psychology for the whole world (the categories comprising “Mental disorders”, “Behavior and behavior mechanisms”, “Psychological phenomena and processes” and “Behavioral disciplines and activities”). The two search strategies were then combined using the Boolean operator “AND” to obtain the psychiatry and mental health publications for each Arab country over the last 15 years. The citations of psy- chiatry and mental health research for each Arab country were extracted from PubMed and transferred to Microsoft Word document files to be thoroughly read and inspected for any inconsistencies and to validate their attribution to mental health re- search. False positive publications for each country were deleted manually and the pro- cess ended with the total number of mental health publications in the 22 Arab countries. The citations were analysed using SPSS, version 6, in which country name, year of publication, field of study, type of study, study setting and remarks variables were identified. The data were analysed using SPSS and Microsoft Excel to calculate the following: • the number of mental health publica- tions for each country; • the proportion of mental health citations out of all biomedical citations in each country; • the proportion of mental health publica- tions in each country out of the total mental health publications of Arab countries; and • the proportion of mental health citations in the Arab world in different types, set- tings and subtopics of mental health. Results Country distribution The number of PubMed-listed biomedical research citations affiliated to Arab coun- tries over the last 15 years totalled 27 395 articles. Saudi Arabia and Egypt together had the highest numbers of biomedical cita- tions, 7421 and 6536 respectively, together accounting for 51.0% of the Arab world’s biomedical publications (Table 1). The number of PubMed-listed mental health research papers published and affili- ated to Arab countries over the last 15 years (1987–2002) totalled 338 articles: 1.2% of the total number of biomedical citations (Table 1). Kuwait and Saudi Arabia together had the highest number percentage of men- tal health publications relative to the total mental health publications from Arab coun- 07 Mental health publications.pmd 12/20/2005, 2:35 PM321 322 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 tries, 72 and 54 respectively, accounting for 37.3% of mental health publications in the Arab world. The highest proportion of mental health studies in a country was in Palestine, where 22.4% (13/58) of biomed- ical citations were mental health studies, followed by 5.7% (13/227) in Bahrain. Of the 338 mental health publications, 58 (17.2%) were related to studies of men- tal disorders in children or adolescent age Table 1 Biomedical research publications affiliated to Arab countries: distribution by country for all mental health publications and for child and adolescent mental health publications Country All biomedical All mental Child and science health adolescent mental health (n = 27 395) (n = 338) (n = 58) No. % No. % No. % Saudi Arabia 7 421 27.1 54 16.0 4 7.4 Egypt 6 536 23.9 52 15.4 10 19.2 Kuwait 2 184 8.0 72 21.3 11 15.3 Morocco 2 020 7.4 9 2.7 1 11.1 Tunisia 1 966 7.2 17 5.0 2 11.8 Lebanon 1 580 5.8 17 5.0 1 5.9 Jordan 1 344 4.9 9 2.7 2 22.2 United Arab Emirates 1 089 4.0 43 12.7 11 25.6 Sudan 659 2.4 3 0.9 1 33.3 Oman 577 2.1 12 3.6 2 16.7 Iraq 408 1.5 2 0.6 0 0 Algeria 406 1.5 4 1.2 0 0 Libyan Arab Jamahiriya 322 1.2 12 3.6 0 0 Qatar 228 0.8 1 0.3 0 0 Bahrain 227 0.8 13 3.8 3 23.1 Syrian Arab Republic 145 0.5 2 0.6 0 0 Yemen 92 0.3 2 0.6 0 0 Palestine 58 0.2 13 3.8 10 76.9 Mauritania 49 0.2 0 0 0 0 Somalia 46 0.2 1 0.3 0 0 Djibouti 28 0.1 0 0 0 0 Comoros 10 < 0.1 0 0 0 0 n = total number of published citations. 07 Mental health publications.pmd 12/20/2005, 2:35 PM322 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 323 groups. While Kuwait, Saudi Arabia and Egypt were the leading countries in mental health publications generally, Palestine and the United Arab Emirates led the Arab coun- tries in child and adolescent mental health research. More than half of the studies were pub- lished in the last 5 years with a major in- crease in the number of publications after 1998 (Table 2). Types, settings and fields of research Analysis of the study settings showed that most mental health research in Arab coun- tries was based in hospital or primary care facilities (55.6%) or the community (37.3%); a small proportion were based in other settings (a workplace, a prison and a medicolegal facility) (0.9%) (Table 3). The community-based mental health studies outnumbered the heath-facility-based stud- ies in the leading countries except in Saudi Arabia. A higher proportion of child and adolescent mental health studies (77.6%) were community-based compared with all mental health publications. Regarding the type of study, the majori- ty were epidemiological in nature (61.2%), followed by psychometric (10.7%) and clinical (10.4%) (Table 3). The majority (78.0%) of the psychometric studies were designed to measure depression, anxiety or obsessive–compulsive disorders. Mental health systems research comprised less than 5% of mental health studies. Almost half of these studies were in Saudi Arabia and dealt with topics such as manpower training and capacity building, evaluation of health facilities and assessment of perfor- mance, referral systems, utilization of men- tal health facilities and prescriptions in mental health facilities. More clinical and health systems research studies were car- ried out in Saudi Arabia than any other Arab country, while more epidemiological stud- ies were carried out in Kuwait and more review publications were from Egypt. Ba- sic science research constituted only 5% of all mental health studies and was not seen in the countries ranking first in mental health research. The most dynamic fields of mental health research in the present analysis (ac- counting together for more than one-third of the total mental health publications) were substance abuse disorders (15.4% of the overall mental health publications), anxiety disorders (14.8%) and mood disorders (10.7%) (Table 3). Among child and ado- lescent mental health publications, disor- Table 2 Mental health publications affiliated to Arab countries: distribution by year of publication Year All mental Child and health adolescent mental (n = 338) health (n = 58) No. % No. % 1987 3 0.9 0 0 1988 15 4.4 1 1.7 1989 11 3.3 1 1.7 1990 15 4.4 2 3.4 1991 11 3.3 1 1.7 1992 17 5.0 2 3.4 1993 10 3.0 2 6.9 1994 15 4.4 0 0 1995 16 4.7 4 6.9 1996 21 6.2 3 5.2 1997 22 6.5 2 3.4 1998 20 5.9 7 12.1 1999 41 12.1 8 13.8 2000 32 9.5 6 10.3 2001 49 14.5 12 20.7 2002 40 11.8 7 12.1 n = total number of published citations. 07 Mental health publications.pmd 12/20/2005, 2:35 PM323 324 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 ders which appear for the first time among children and adolescents were the most common: learning disorders (8.6%), eating disorders (8.6%) and attention deficit hy- peractivity disorder and disruptive behav- iour disorders (6.9%). Most of the publications on anxiety disorders were af- filiated to institutions in Palestine, mostly studies on post-traumatic stress disorders, whereas the majority of the studies on learning disorders were from Egypt. Journals About 71% of the mental health and psy- chiatry publications were published in 39 PubMed indexed journals (Table 4). The top 7 journals for Arab mental health publi- cations overall were Psychological reports, Eastern Mediterranean health journal, Acta psychiatrica scandinavica, British journal of psychiatry, L’Encephale, Inter- national journal of social psychiatry and Saudi medical journal. All are in English Table 3 Mental health publications affiliated to Arab countries: distribution by type of study, study setting and type of disorder Variable All mental Child and health adolescent mental (n = 338) health (n = 58) No. % No. % Type of study Epidemiological 207 61.2 48 82.8 Psychometric 36 10.7 5 8.6 Clinical 35 10.4 3 5.2 Literature review 17 5.0 1 1.7 Basic science 16 4.7 0 0 Health systems research 16 4.6 0 0 Knowledge, attitudes & practice 11 3.3 1 1.7 Study setting Hospital 174 51.5 11 19.0 Community 126 37.3 45 77.6 Primary care facility 14 4.1 1 1.7 Other 3 0.9 1 1.7 Not applicable 21 6.2 0 0 Disorder studied Substance abuse disorders 52 15.4 4 6.9 Anxiety disorders 50 14.8 11 19.0 Mood disorders 36 10.7 6 10.3 Eating disorders 8 2.4 5 8.6 Learning disorders 5 1.5 5 8.6 Attention deficit hyperactivity and disruptive behaviour disorders 4 1.2 4 6.9 n = total number of published citations. 07 Mental health publications.pmd 12/20/2005, 2:35 PM324 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 325 except for L’Encephale which is French. The remaining 29% of the Arab publica- tions were published in 95 journals. For child and adolescent mental health publica- tions, roughly half were published in Psy- Journal All mental health (n = 338) No. % Psychological reportsa 30 8.9 Eastern Mediterranean health journala 28 8.3 Acta psychiatrica scandinavica 27 8.0 British journal of psychiatry 19 5.6 L’Encephalea 15 4.4 International journal of social psychiatry 15 4.4 Saudi medical journal 11 3.3 Social psychiatry and psychiatric epidemiology 10 3.0 Drug and alcohol dependence 7 2.1 Child abuse & neglecta 6 1.8 Comprehensive psychiatry 5 1.5 Journal of affective disorders 5 1.5 Death studies 4 1.2 Le Journal medical libanais 4 1.2 Journal of tropical pediatricsa 4 1.2 Biological psychiatry 3 0.9 General pharmacology 3 0.9 International journal of emergency mental health 3 0.9 International journal of geriatric psychiatry 3 0.9 International journal of psychiatry in medicine 3 0.9 Journal All mental health (n = 338) No. % Perceptual and motor skills 3 0.9 Tunisie medicale 3 0.9 American journal of medical genetics 2 0.6 American journal of psychiatry 2 0.6 Archives of women’s mental health 2 0.6 Behavioural brain research 2 0.6 Bulletin de l’Academie nationale de medecine 2 0.6 Clinical psychology review 2 0.6 General hospital psychiatry 2 0.6 International journal of eating disorders 2 0.6 International journal of nursing studies 2 0.6 Journal of nervous and mental disease 2 0.6 Journal of psychosomatic research 2 0.6 Journal of substance abuse 2 0.6 Journal of traumatic stress 2 0.6 Lanceta 2 0.6 Medical teacher 2 0.6 Neuroepidemiologya 2 0.6 Other journals (n = 95) 95 28.1 Table 4 Journals in which mental health articles from Arab countries were published aMost common journals in child and mental health. n = total number of published citations. chological reports, Child abuse & neglect, Eastern Mediterranean health journal, Journal of tropical pediatrics, Lancet, L’Encephale or Neuroepidemiology. 07 Mental health publications.pmd 12/20/2005, 2:35 PM325 326 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Discussion In recent years, economic priorities have increasingly forced governments, normally the primary supporter of basic research, to adopt policies that link science and technol- ogy programmes more closely to broad or- ganizational and societal goals. Consequently, the assessment of research output through citation analysis has pro- gressively developed and become a priority issue for the scientific research community [19,20]. Quantifying and weighting the re- sults of research are difficult tasks. The advantages of PubMed as a powerful data- base and search tool may be easily degrad- ed if the user does not utilize proper keywords or correct syntaxes, and the re- sults can be biased or even misleading [21]. Nevertheless, citation analysis remains a worthwhile way of evaluating the publica- tion performance of individual scientists, research units or countries. The present data show that the geo- graphical distribution of publications in psychiatry and mental health does not fol- low the distribution pattern of publications in general biomedical research for the Arab countries. From the top two ranking coun- tries of the Arab world in biomedical cita- tions only one country, Saudi Arabia, led the Arab countries in mental health research along with Kuwait. Taking into account that Gulf Cooperation Council countries have a relatively short history of research, these countries are especially prolific in terms of PubMed-indexed biomedical cita- tions (a finding which has been shown be- fore [4]), as well as mental health research publications. To create an accurate picture, however, the frequencies of citations should be normalized, e.g. by country pop- ulation, school enrolment, adult illiteracy or gross domestic product (GDP) [18]. In the current study Saudi Arabia was no longer one of the two leading countries for mental health publications when the average num- ber of publications was corrected for its population size in 2000. However, it was decided not to report normalized figures in this study, as it was difficult to obtain accu- rate estimates of population size for the year 2000 for some countries. There is a real need to address major areas of research activity in the Arab coun- tries as well as other countries of the East- ern Mediterranean Region in terms of mental health policy, economic evaluation of models of mental health delivery and evaluation of intersectoral linkages [22]. Child psychiatry, especially attention deficit hyperactivity disorder and child autism, are not gaining much interest from researchers in Arab countries, although other authors have reported that research into behav- ioural problems among children and adoles- cents is high in some Arab countries as well as other developing countries [23]. This calls for more attention to these issues by health policy-makers as well as researchers and clinicians, in order to improve diagno- sis and treatment in Arab and other devel- oping countries. The basic demographic structure is rap- idly changing in many Arab societies and a great shift in the age distribution of the pop- ulation has occurred over the last two de- cades. Child and adolescent psychiatric disorders are seen more frequently, and earlier in a child’s lifespan. Prolonged life expectancy and the resulting increase in the geriatric population who have good access to health services means that geriatric psy- chiatry needs to attract more attention from researchers in the Region [16]. More ef- forts have to be made to increase policy- makers’ and the public’s awareness of mental health problems and to promote ac- tivities and research into these issues. 07 Mental health publications.pmd 12/20/2005, 2:35 PM326 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 327 Certain limitations of the present study have to be acknowledged. Only about 25% of around 280 medical journals published in the Arab world are indexed in PubMed. Thus the study may have underestimated the amount of research from Arab coun- tries if researchers have published their studies in these local journals. Another lim- itation is not using a factor such as popula- tion size or GDP to normalize the number of publications for each country. However, the use of ratios could be misleading, espe- cially with small numbers or using normal- ization by GDP indicator. Nevertheless, few studies have been undertaken in the Arab world to make a geographical analysis of biomedical publications and none of these, to the best of this author’s knowl- edge, have reviewed the publications for a single specialty. Conclusion and recommendations Mental health research is still relatively ne- glected in the Arab world. Only 1.2% of the biomedical citations are related to the field of psychiatry and mental health. Issues re- lated to utilization of general and mental health services and its association with mental disorders need to be studied in- depth. Much research is missing concern- ing training in mental health or teaching in medical schools. Mental health systems re- search constituted less than 5% of the mental health research in the current study and this is also an important domain. Child and adolescent psychiatry is not gaining much attention: only 17% of the overall mental health publications. Some topics, such as child autism and attention deficit hyperactivity disorder, need to be thor- oughly addressed and followed-up by well- designed studies. Unfortunately, given the cross-sectional nature of most of the men- tal health publications in the Arab world, the temporal or causal relationship between a child or adolescent’s mental disorders and the long-term consequences in adulthood are not clear. Studies investigating gender roles in mental disorders and their preva- lence, course and response to treatment are also very scarce in Arab world. Mental health problems of youth, women and eld- erly people have to be investigated, and Arab countries need to revise their mental health priorities on a regular basis accord- ing to their demographic as well as their ep- idemiological transitions. References 1. Tutarel O. Geographical distribution of publications in the field of medical edu- cation. BMC medical education, 2002, 2(1):3–10. 2. Hefler L, Tempfer C, Kainz C. Geography of biomedical publications in the Euro- pean Union, 1990–1998. Lancet, 1999, 353:1856. 3. Thompson DF. Geography of U.S. bio- medical publications, 1990 to 1997. New England journal of medicine, 1999, 340:817–8. 4. Deleu D, Northway MG, Hanssens Y. Geographical distribution of biomedical publications from the Gulf Corporation Council countries. Saudi medical jour- nal, 2001, 22:10–2. 5. Rosselli D. Latin American biomedical publications: the case of Colombia in Medline. Medical education, 1998, 32:274–7. 6. Lee CS. Productivity of SCI Korean medi- cal papers, 1996–1997. Journal of Ko- rean medical science, 1999, 14:351–8. 07 Mental health publications.pmd 12/20/2005, 2:35 PM327 328 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 7. Güllüoglu BM, Aktan A. Scientific publi- cations at a Turkish medical school. Aca- demic medicine, 2000, 75:760. 8. Powner DJ, Kellum JA. Declining critical care research publications by authors from U.S. institutions, 1990–1999. Aca- demic medicine, 2001, 76:1261–3. 9. Weisinger JR, Bellorin-Font E. Latin American nephrology: scientific produc- tion and impact of the publications. Kid- ney international, 1999, 56:1584–90. 10. Sorrentino D et al. Scientific publications in gastroenterology and hepatology in Western Europe, USA and Japan in the years 1992–1996: a global survey. Di- gestion, 2000, 61:77–83. 11. Mela GS, Mancardi GL. Neurological re- search in Europe, as assessed with a four-year overview of neurological sci- ence international journals. Journal of neurology, 2002, 249:390–5. 12. Favaloro EJ. Medical research in New South Wales 1993–1996 assessed by Medline publication capture. Medical journal of Australia, 1998, 169:617–22. 13. The NCBI handbook. Bethesda, Mary- land, USA, National Library of Medicine, 2003. (http://www.ncbi.nlm.nih.gov/ books/bv.fcgi?rid=handbook, accessed 8 May 2005). 14. Tadmouri GO, Tadmouri NB. Biomedical research in the Kingdom of Saudi Arabia (1982–2000). Saudi medical journal, 2002, 23:20–4. 15. Saraceno B. Mental health in EMRO: the future is now. Eastern Mediterranean health journal, 2001, 7(3):332–5. 16. Mohit A. Mental health in the Eastern Mediterranean Region of the World Health Organization with a view of the future trends. Eastern Mediterranean health journal, 2001, 7(3):353–62. 17. Tadmouri GO, Bissar-Tadmouri N. A ma- jor statistical pitfall in the strategy of search on PubMed. Saudi medical jour- nal, 2004, 25(suppl. 1):S55–8. 18. Shaban SF, Abu-Zidan FM. A quantita- tive analysis of medical publications from Arab countries. Saudi medical jour- nal, 2003, 24(3):294–6. 19. Mela GS, Cimmino MA, Ugolini D. Impact assessment of oncology research in the European Union. European journal of cancer, 1999, 35(8):1182–6. 20. Luukkonen T. Bibliometrics and evalua- tion of research performance. Annals of medicine, 1990, 22:145–50. 21. Robinson KA, Dickersin K. Development of a highly sensitive search strategy for the retrieval of reports of controlled trials using PubMed. International journal of epidemiology, 2002, 31:150–3. 22. Mubbashar MH, Saeed K. Development of mental health services in Pakistan. Eastern Mediterranean health journal, 2001, 7(3):392–6. 23. Al-Sharbati MM, Al-Hussaini AA, Antony SX. Profile of child and adolescent psy- chiatry in Oman. Saudi medical journal, 2003, 24(4):391–5. 07 Mental health publications.pmd 12/20/2005, 2:35 PM328 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 329 Subclinical hypothyroidism in lithium- treated psychiatric patients in Tehran, Islamic Republic of Iran M. Aliasgharpour,1 M. Abbassi, 1 H. Shafaroodi 2 and F. Razi1 1National Reference Laboratory of Iran, World Health Organization Collaborating Centre, Biochemistry Trace Elements Analysis Laboratory, Bou-Ali Hospital, Tehran, Islamic Republic of Iran (Correspondence to M. Aliasgharpour: mehri9@yahoo.com). 2Department of Pharmacology, School of Medicine, Azad University of Medical Sciences, Tehran, Islamic Republic of Iran. Received: 09/02/04; accepted: 17/05/04 ABSTRACT We investigated thyroid function in 46 (20 female & 26 male) psychiatric outpatients on lithium treatment by assessing triiodothyronine, thyroxine and thyroid stimulating hormone (TSH) levels. The pres- ence of thyroid antibodies (anti-thyroid peroxidase and anti-thyroglobulin) was also assessed. Out of the 46 patients, 8 (17%) displayed overt hypothyroidism. Of the remaining patients, subclinical hypothyroidism was found in 16 patients (35%) and euthyroidism in 22 (48%). Thyroid antibodies were present in 6 patients in the euthyroid group and 5 patients in the hypothyroid group. The Pearson product-moment correlation results indicated positive association between TSH level and duration of lithium use and age of the patients with subclinical hypothyroidism. Duration of lithium use and age could be a reasonable indicator for screening asymptomatic patients for subclinical hypothyroidism after starting lithium treatment. Hypothyroïdie infraclinique chez des patients psychiatriques traités au lithium à Téhéran (Répu- blique islamique d’Iran) RÉSUMÉ Nous avons examiné la fonction thyroïdienne chez 46 patients (20 femmes et 26 hommes) des consultations psychiatriques externes traités au lithium en évaluant le taux de triiodothyronine, de thyroxine et de thyréostimuline hypophysaire (TSH). La présence d’anticorps anti-thyroïde (anti-thyroperoxydase et anti-thyroglobuline) a également été évaluée. Huit (17 %) des 46 patients présentaient une hypothyroïdie patente. Parmi les patients restants, on a trouvé une hypothyroïdie infraclinique chez 16 d’entre eux (35 %) et une euthyroïdie chez 22 (48 %). Des anticorps anti-thyroïde étaient présents chez 6 patients dans le groupe de l’euthyroïdie et chez 5 patients dans le groupe de l’hypothyroïdie. Les résultats de la corrélation des moments mixtes de Pearson ont indiqué une association positive entre le taux de TSH, la durée de l’utilisation du lithium et l’âge des patients présentant une hypothyroïdie infraclinique. La durée de l’utilisation du lithium et l’âge pourraient servir d’indicateurs plausibles pour le dépistage de l’hypothyroïdie infraclinique chez des patients asymptomatiques après instauration d’un traitement au lithium. 08 Subclinical hypothyroidism.pmd 12/20/2005, 2:35 PM329 330 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Introduction Lithium carbonate is widely used in the treatment of mood disorders such as bipo- lar disorder. It has been reported to induce gastrointestinal [1], neuromuscular [2] and endocrine adverse effects [3] in 35%–93% of patients taking it. To optimize its effect, therefore, it is important to be aware of these side-effects. Thyroid hypofunction is one of the most common endocrine side- effects associated with lithium treatment. In most cases lithium-induced hypothy- roidism is subclinical [4–6]. The mechanism of the effect of lithium on the thyroid gland is not fully understood. However, it has been shown that lithium reduces iodine uptake into the gland, inhib- its iodine addition to tyrosine, reduces tri- iodothyronine (T3) and thyroxine (T4) release and enhances thyroid stimulating hormone (TSH) release [5–8]. Further- more, some studies have reported a higher incidence of thyroid antibodies in patients treated with lithium (24%) compared with those not taking lithium (12%) [9,10]. The risk of lithium-induced antibodies increases with the duration of therapy and is more common in women than in men [9,10]. It has been reported that up to 10%– 20% of patients develop lithium-induced subclinical hypothyroidism [11–13]. The majority of these patients have few symp- toms or none at all. However, it has been shown that conversion of subclinical hy- pothyroidism to overt hypothyroidism in the presence of circulating antibodies is high [14]. Therefore, routine screening of patients receiving lithium has been advocat- ed. In this regard, TSH appears to be the most efficient parameter in revealing sub- clinical hypothyroidism in patients on lithi- um treatment [11–13]. In the present study we evaluated the possible effect of lithium on thyroid func- tion tests (TSH, T3 andT4 levels) in 46 psy- chiatric outpatients at the Reference Labo- ratory of the Islamic Republic of Iran in Tehran. Methods Over the 2-year period 2001–2002, a total of 46 (20 female and 26 male) psychiatric outpatients attending the Reference Labo- ratory to check their serum lithium level were randomly selected for the study. A questionnaire was completed for each pa- tient with the patient’s age, any previous thyroid gland problems, time of possible onset and duration of lithium use. All the patients reported that before starting lithi- um treatment they had had their thyroid tested and the results were normal. Any pa- tients with a previous thyroid disorder were excluded from the study. Thus 5 patients who had had their thyroid gland removed or had a thyroid disorder not associated with lithium use were excluded. To investigate whether lithium is an im- portant risk factor for the development of subclinical hypothyroidism, 95% confi- dence intervals (95% CI) were calculated. Then the Pearson product-moment corre- lation was performed to asses whether there was a correlation between pairs of variables such as TSH level, lithium con- centration, duration of lithium use and age of patients with subclinical hypothryroid- ism. Blood samples were drawn 8–10 hours after oral dose/s of lithium were given and sera were separated from the cells within 2 hours. Serum lithium concentration was determined using flame atomic absorption spectroscopy (Varian 20-plus) [15,16]. Sera were sent to the radioimmunoas- say (RIA) laboratory for the determination of TSH (K1100To28202, Kavoshyar Iran Co.), T3 (K1100To38203, Kavoshyar Iran Co.), T4 (K1100To28203 Kavoshyar Iran 08 Subclinical hypothyroidism.pmd 12/20/2005, 2:35 PM330 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 331 Co.) levels, as well as anti-thyroid peroxi- dase (anti-TPO) (Monobind, Inc. Product code: 1125-300, USA) and anti-thyroglobu- lin (anti-Tg) (Monobind, Inc. Product code:1025-300, USA). Results Table 1 gives the patient information. Dura- tion of lithium use varied with a mean of 6 years (standard deviation 4.3) and range of 2–15 years. None of the patients had visible thyroid gland enlargement. Out of the 46 patients, 8 (17%) (4 females and 4 males) displayed overt hypothyroidism. Based on the completed questionnaires they were on levothyroxin and had normal TSH levels. Therefore, no further study was performed on this group of patients. Subclinical hy- pothyroidism, for which no thyroid supple- mentation was used, was observed in 16 patients (35%, 95% CI: 23.0%–54.0%) (6 female and 10 males) and euthyroidism was observed in 22 patients (48%) (14 fe- males and 8 males). Of the16 subclinical cases, 5 had mildly elevated TSH levels (TSH 3.0–30 mU/L) and 11 had highly ele- vated TSH levels (TSH > 30 mU/L): normal TSH levels for adult are 0.5–3.0 mU/L. In both groups, free T4 levels were normal. We carried out the Pearson product- moment correlation between the TSH level of the 16 patients with subclinical hypothy- roidism and lithium concentration, duration of lithium use and age. Significant differ- ences were observed for patient’s age (r = 0.71, P < 0.01) and duration of lithium use (r = 0.76, P < 0.001); lithium concentration was not significantly correlated (P > 0.05) (Table 2). Furthermore, using the Fisher exact test there was no significant differ- ence for sex and occurrence of subclinical hypothyroidism (P > 0.05). Thyroid antibodies (anti-TPO and anti- Tg) were present in 6 patients (27%) (5 fe- males and 1 male) in the euthyroid group (n = 22). In addition, 5 patients (31%) (3 fe- males and 2 males) in the subclinical group tested positive for these antibodies (Table 3). Discussion Subclinical hypothyroidism is defined as el- evated concentration of TSH and normal level of serum T4. This disorder in lithium- treated patients may be present without symptoms [6,11,12,17,18] In the present study thyroid function was evaluated in 46 psychiatric (20 female and 26 male) outpatients with different du- ration of lithium treatment (2–15 years) and 16 (35%) patients displayed subclinical hy- pothyroidism after starting lithium treat- ment. They appeared to be symptom-free and did not complain of any side-effects. Table 1 Patient information Sex Total Mean (SD) No. of Mean No. of Previous number duration of patients with (SD) age patients with thyroid of patients lithium use overt (years) subclinical gland (years) hypothyroidism hypothyroidism disorder Female 20 5.4 (4.3) 4 53 (3) 6 No Male 26 6.3 (4.0) 4 45 (5) 10 No SD = standard deviation. 08 Subclinical hypothyroidism.pmd 12/20/2005, 2:35 PM331 332 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Since in the normal population the preva- lence of subclinical hypothyroidism ranges from 1% to12% [19,20], our results indi- cate lithium is an important risk factor for the development of subclinical hypothy- roidism. Studies have reported the incidence of lithium-induced hypothyroidism increases with the duration of treatment [21]. The mean duration of lithium treatment for our subclinical hypothyroid patients was rela- tively long at 6 years. Duration of lithium treatment was significantly associated with the risk of increased TSH levels and the development of subclinical hypothyroidism (P < 0.001). In addition, researchers have shown that the incidence of subclinical hy- pothyroidism increases with age, especially in women [22]. Our preliminary data indi- cate a correlation (P < 0.01) between TSH level in subclinical hypothyroid patients and age, but sex was not a significant factor (P > 0.05). Previous studies have indicated the presence of thyroid autoantibodies (24%) in patients receiving lithium is high [9,10]. Furthermore, it has been shown that lithi- um may accelerate the production of the thyroid autoantibodies that may be present prior to lithium treatment [9]. Singer re- ported that overt hypothyroidism develops in a high proportion of subclinical cases who have positive thyroid autoantibodies [14]. In our study, 5 subclinical patients (31%) displayed high thyroid autoantibody titres. Our data also indicate that lithium- induced subclinical hypothyroidism may occur in the absence of thyroid autoanti- bodies (11 patients) although subclinical patients are at higher risk of developing overt hypothyroidism. Because of our small sample size, antibody positivity by it- self was not a good marker for predicting conversion of subclinical hypothyroidism to overt hypothyroidism. Our results indicate that duration of lith- ium use and age regardless of sex could be a reasonable indicator for screening asymptomatic patients for subclinical hy- pothyroidism after starting lithium treat- ment. In addition, TSH appears to be a more suitable parameter for assessing sub- clinical hypothyroidism than the presence of positive thyroid autoantibodies. Acknowledgements Special thanks go to co-workers in the RIA laboratory, especially Mrs Farahnaz Mehra- bi and Dr Gasem Khosravanie for all their help during thyroid panel tests. Table 2 Pearson product-moment correlations between variables and thyroid stimulating hormone (TSH) level in the 16 subclinical patients Variable r P-value TSH & age 0.71 0.01 TSH & lithium concentration 0.29 0.28 TSH & lithium duration 0.76 0.001 Table 3 Positivity of thyroid autoantibodies in lithium-induced subclinical hypothyroidism and euthyroid cases Thyroid Euthyroid Subclinical autoantibodies in a (n = 22) (n = 16) Females 5 3 Males 1 2 Total (%) 6 (27) 5 (31) aAnti-thyroid peroxidase & anti-thyroglobulin. 08 Subclinical hypothyroidism.pmd 12/20/2005, 2:35 PM332 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 333 References 1. Birch NJ. Lithium. In: Seiler HG, Sigel H, eds. Handbook on toxicity of inorganic compounds. New York, Marcel Dekker Inc., 1988:383–93. 2. Kocsis JH et al. Neuropsychologic ef- fects of lithium discontinuation. Journal of clinical psychopharmacology, 1993, 13:268–76. 3. Lenox RH et al., eds. American psychiat- ric press textbook of psychopharmacol- ogy, 2nd ed. Washington DC, American Psychiatric Press, 1998:379–429. 4. Ozpoyraz N, Tamam L, Kulan E. Thyroid abnormalities in lithum-treated patients. Advances in therapy, 2002, 19(4):176– 84. 5. Bocchetta A et al. Thyroid abnormalities during lithium treatment. Acta psychiatri- ca Scandinavica, 1991, 83(3): 193–8. 6. Smigan L et al. Lithium therapy and thy- roid function tests. A prospective study. Neuropsychobiology, 1984, 11(1):39– 43. 7. Salata R, Klein I. Effects of lithium on the endocrine system: a review. Journal of laboratory and clinical medicine, 1987, 110:130–6. 8. Surks MI, Sievert R. Drugs and thyroid function. New England journal of medi- cine, 1995, 333:1688–94. 9. Lazarus JH. Endocrine and metabolic ef- fects of lithium. New York, Plenum, 1986: 99–124. 10. Deniker P et al. Thyroid autoantibody levels during lithium therapy. Neuropsy- chobiology, 1978, 4(5):270–5. 11. Bensenor I. Screening for thyroid disor- ders in asymptomatic adults from Brazil- ian populations. São Paulo medical journal, 2002, 120(5):146–51. 12. Kleiner J, Altshuler L, Hendrick V, Hershman JM. Lithium–induced sub- clinical hypothyroidism: review of the lit- erature and guidelines for treatment. Journal of clinical psychiatry, 1999, 60(4):249–55. 13. Perrild H et al. Thyriod function and ultra- sonically determined thyroid size in pa- tients receiving long-term lithium treatment. American journal of psychia- try, 1990, 147:1518–21. 14. Singer PA et al. Treatment guidelines for patients with hyperthyroidism and hy- pothyroidism. JAMA: the journal of the American Medical Association, 1995, 273:808–12. 15. Murray RL. Lithium. In: Pesce A, Kaplan LA, eds. Methods in clinical chemistry . St Louis, CV Mosby Co. 1987. 16. Analytical methods (F-AAS). St Helens, Australia, Varian Australia Pty Ltd, 1989 (Publication # 85 –10000900). 17. Surks MI, Ocampo E. Subclinical thyroid disease. American journal of medicine, 1996, 100(2):217–23. 18. Cooper DS. Subclinical hypothyroidism. New England journal of medicine, 2001, 345(4):260–5. 19. Tunbridge WMG et al. The spectrum of thyroid disease in a community; the Whickham survey. Clinical endocrinol- ogy, 1977, 7:481–93. 20. Canaris GJ et al. The Colorado thyroid disease prevalence study. Archives of in- ternal medicine, 2000, 160:526–34. 21. Transbol I, Christiansen C, Baastrup PC. Endocrine effects of lithium: hypothyroid- ism, its prevalence in long-term treated patients. Acta endocrinologica, 1978, 87:759–67. 22. Helfand M, Redfern CC. Screening for thyroid disease; an update. Clinical Guideline Part 2. Annals of internal medicine, 1998, 129:144–58. 08 Subclinical hypothyroidism.pmd 12/20/2005, 2:35 PM333 334 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 1Faculty of Medicine, Mu’tah University, Karak, Jordan (Correspondence to R.M. Youssef: email: randayoussef@link.net). Received: 23/08/04; accepted: 18/01/05 ABSTRACT The health status, mental well-being and functional capacity of senior citizens was assessed in a community-based survey of people ≥ 60 years in 2004. Analysis was performed using logistic and linear regression analyses. Of the 300 subjects enrolled (53.3% women), 74.4% were affected by chronic di- seases, 24.3% were classified as depressed and 44.0% had a negative health perception. Dependence in instrumental activities of daily living (92.0%) was more frequent than dependence in basic activities (28.0%). Women were more likely to be depressed, and suffer memory impairment and limitation of functional capa- city. Increase in depressive symptoms was independently predicted by increased age, living alone, poor functional capacity, memory impairment and negative perception of health. Low summary performance in instrumental and basic activities was independently predicted by increased age, lack of education, high number of reported symptoms, depression and memory impairment. Depression, poor functional capacity and memory impairment reinforced each other resulting in a state of dependency. Évaluation complète de la santé des citoyens âgés dans le Gouvernorat d’El Karak (Jordanie) RÉSUMÉ L’état de santé, le bien-être mental et la capacité fonctionnelle des citoyens âgés ont été évalués dans le cadre d’ une étude communautaire des personnes âgées de 60 ans et plus réalisée en 2004. L’étude a utilisé les analyses de régression logistique et linéaire. Sur les 300 sujets admis dans l’étude (53,3 % de femmes), 74,4 % souffraient de maladies chroniques, 24,3 % étaient classés comme déprimés et 44,0 % avaient une perception négative de la santé. La dépendance pour les activités instru- mentales de la vie quotidienne (92,0 %) était plus fréquente que la dépendance pour les activités de base (28,0 %). Les femmes étaient plus susceptibles d’être déprimées, et de souffrir de troubles de mémoire et d’une limitation de la capacité fonctionnelle. L’augmentation des symptômes dépressifs était prédite in- dépendamment par un âge plus avancé, le fait de vivre seul, une mauvaise capacité fonctionnelle, des troubles de mémoire et une perception négative de la santé. Une mauvaise performance sommaire dans les activités instrumentales et de base était prédite indépendamment par un âge plus avancé, le manque d’instruction, le nombre élevé de symptômes signalés, la dépression et les troubles de mémoire. La dépression, la mauvaise capacité fonctionnelle et les troubles de mémoire cumulaient leurs effets, en- traînant un état de dépendance. Comprehensive health assessment of senior citizens in Al-Karak governorate, Jordan R.M. Youssef1 09 Comprehensive health assessment.pmd 12/20/2005, 2:35 PM334 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 335 Introduction Globally, the proportion of people aged 60 years and over is growing faster than any other age group. By the year 2025, people in this age group will reach a total of 1.2 billion and this will rise to 2 billion in the year 2050 with 80% of them living in devel- oping countries [1]. In Jordan, there has been a considerable change in the age structure of the population since 1979 as the number of people above the age of 60 years has been increasing steadily and reached 5.4% of the population in 2002 [2]. Projections indicate that this proportion will represent 9.2% of the population in 2025 with a further increase to reach 22.1% in 2050 [3]. As the ageing process continues, dis- eases and impairment become common. Chronic noncommunicable diseases are characteristic of old age and the prime causes of deterioration of physical health [1]. Psychological problems, mainly de- pression [4] and cognitive impairment [1] are also common among the elderly and pose major threats to their mental and physical well-being. Depression in this age group is characterized by physical co-mor- bidity and significant disability [5]. With the continuing increase in life expectancy [2] and the multitude of ailments afflicting old people, the capability of the elderly to main- tain independence becomes a focus of at- tention. Health and functional ability are crucially important to the quality of life of old people as they determine the extent to which they can cope independently with the demands of everyday living [6]. In view of the vulnerability of elderly people and their growing number in the Jordan, assessment of the capability of se- nior citizens to maintain physical and men- tal well-being and a state of independence is of much relevance. Therefore a compre- hensive assessment of the elderly popula- tion was conducted in Al-Karak governor- ate, south Jordan to determine current needs and to set future plans. Methods A community-based survey was conducted in Al-Karak governorate, south Jordan be- tween January and April 2004. A cluster sample survey was used to identify sub- jects aged 60 years and over. A total of 30 clusters, the usual chosen number, were identified and from each cluster 10 subjects in the target age group were enrolled on the basis of a house-to-house survey. There were no refusals to participate. A compre- hensive assessment was performed by the researcher in the homes of elderly using a pre-tested questionnaire, which had been piloted tested to check the appropriateness of the questions and the participants to un- derstand them. The questionnaire consisted of the following sections. • Section I: Sociodemographic character- istics, namely age, sex, marital status, educational attainment, employment, source of income and living conditions. • Section II: Chronic health problems en- dured based on previous diagnosis, number of medications prescribed, pat- tern of health services utilization and participants’ opinion of the health care services provided in the area. • Section III: Self-reported health status which included 14 questions pertinent to symptoms and complaints common to old age experienced in the 6 months prior to the survey. Perceived health status was rated by giving a score of 1 for each complaint or symptom report- ed and a score of 0 if the participants did not express any complaint. Scores 09 Comprehensive health assessment.pmd 12/20/2005, 2:35 PM335 336 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 ranged from zero to 14. Higher scores reflect poor self-rated physical health status. • Section IV: Instrumental activities of daily living (IADL) [7] which reflects the capability of an elderly person to live independently in his/her own home. Ac- tivities include subjects’ capability of getting around, using the telephone, shopping, preparing meals, performing household chores, taking medications and managing finances. Three possibili- ties were given for each area: indepen- dent (scored 0), variable degree of assistance required (scored 1), depen- dent and full assistance required (scored 2). Scores ranged from zero to 14. Higher scores indicate impaired ac- tivities of daily living. • Section V: Basic activities of daily living (ADL) [8] which comprises survival tasks, namely eating, bathing, using the toilet, dressing and moving inside the house. For each task, 2 possibilities were given: capable without assistance (scored 0), not capable and assistance required (scored 1). As regards using the toilet and moving inside the house, a third possibility, which scored 2, was considered that included going to the toilet in bed and not being able to move or being paralyzed. Scores ranged from zero to 7. Higher scores indicate im- paired basic activities. • Section VI: Limitation of basic move- ment using the Nagi physical disability scale [9] which covers subjects’ ability to squat/bend, carry 5 kg, walk for a distance of 200–300 metres, get up 3–5 stairs and use fingers for holding. Re- sponses were dichotomized into not limited (scored 0) and limited (scored 1). Scores ranged from zero to 5. High- er scores indicate limitation of move- ment. • Section VII: Evaluation of the mental status using a set of 10 questions on the participants’ capabilities to know the date and day of the week, his age and year of birth, telephone number, name of the area, family name, name of the late and current king. Correct answers scored zero while wrong answer scored 1. Scores ranged from zero to 10. Higher scores indicate memory im- pairment. • Section VIII: Screening for depression was performed using the short version of the geriatric depression scale [10] which includes 15 questions – 10 posi- tive questions answered by “yes” and 5 negative questions answered by “no”. Total score ranged from zero to 15. Higher scores reflect an increase in de- pressive symptoms and a score of e” 5 was set to indicate depression. Data were analysed using the SPSS, version 10. The mean, standard deviation (SD), odds ratio (OR) and the 95% confi- dence interval (CI) were computed. Signif- icance of the results was tested using the chi-squared test, logistic and linear regres- sion analyses and judged at the 5% level. Results This study included 300 participants be- tween the age of 60 and 100 years. More than half of the participants (53.3%) were women and below the age of 70 years (57.0%). The mean age of the women [68.93 (SD 8.577) years; 95% CI: 67.59– 70.27] was comparable to that of the men [70.44 (SD 9.090) years; 95% CI: 68.92– 71.960). The majority of participants (81.3%) had not received any formal edu- cation. Lack of a formal education was en- countered among a significantly higher proportion of women than men (94.4% 09 Comprehensive health assessment.pmd 12/20/2005, 2:35 PM336 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 337 compared to 66.4%, c21 = 83.410, P = 0.001). At the time of the survey, 67.0% of the participants were married and only 10.7% were living alone in their home. Nearly a quarter (23.7%) of the partici- pants reported caring for themselves while for the majority, care was provided either by the spouse (34.0%) or children and grandchildren (39.0%) (Table 1). Retirement benefit was the source of income for 57.7% of participants while a proportion of them were supported by their children (28.7%) or social security servic- es (14.7%). Only 12.7% of the participants were dependent on income generated from current employment. The per-capita monthly income ranged from 20 to 500 Jordanian dinars and it was less than 50 Jordanian dinars for 78.0% of the partici- pants (1 Jordanian dinar = US$ 1.28 at the time of the study) (Table 1). Chronic health problems were reported by 74.4% of the participants; this was re- ported significantly more among women (81.3% compared to 65.7%; c21 = 9.367, P = 0.002). Diagnosed conditions included hypertension (41.7%), rheumatic pains (36.0%), diabetes mellitus (23.3%), is- chaemic heart disease (15.7%), renal prob- lems (5.0%), cerebrovascular disorders (4.0%) and bronchial asthma (3.7%). In the previous 6 months, 18.0% of the par- ticipants had experienced falls, mostly in- side the house. The use of medications was stated by 73.3% of the participants; 75.9% (n = 167) were receiving 1–3 medications while 24.1% (n = 53) were using 4 or more drugs. Table 2 shows the self-rated health sta- tus in the 6 months prior to the survey based on stated symptoms. The majority of participants (84.0%) reported generalized body aches and a substantial proportion re- ported loss of teeth to the extent of inter- ference with chewing of food (61.3%), dyspepsia (59.7%), loss of weight Table 1 General characteristics of the participants General characteristics No. % of participants (n = 300) Sex Male 140 46.7 Female 160 53.3 Age group (years) 60– 171 57.0 70– 80 26.7 80– 37 12.3 90–100 12 4.0 Educational attainment No formal education 244 81.3 Formal education 56 18.7 Marital status Married 201 67.0 Singlea 99 33.0 Living arrangements Living alone 32 10.7 Living with spouse 82 27.3 Living with spouse, children & grandchildren 119 39.7 Living with children & grandchildren 67 22.3 Care provider Self 71 23.7 Spouse 102 34.0 Children & grandchildren 117 39.0 Othersb 10 3.3 Source of incomec Current work 38 12.7 Pension 173 57.7 Support from children 86 28.7 Social security 44 14.7 Private property 3 1.0 Per capita monthly income (Jordanian dinar) < 50 234 78.0 50–100 54 18.0 > 100 12 4.0 One Jordanian dinar = US$ 1.28. aSingle includes never married, widowed and divorced. bOthers include relatives, neighbours and helpers. cCategories are not mutually exclusive. 09 Comprehensive health assessment.pmd 12/20/2005, 2:35 PM337 338 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 (55.0%), insomnia (52.3%), lack of energy to perform daily activities (51.3%), fre- quency (49.0%), shortness of breath (46.7%) and constipation (46.0%). Less frequently, poor appetite (30.0%), impaired vision (30.0%), dizziness (27.0%), inconti- nence (26.0%) and hearing impairment (22.0%) were reported. Relative to men, women were 2.0 times more likely to have a poor self-perception of health. Among them, all symptoms were 1.5 to 2.5 times more likely to be reported except for the lack of teeth and impaired hearing, which were reported by a nearly equal proportions of men and women. The mean number of symptoms reported by women was 1.18 times higher than that of men (OR = 1.18, 95% CI: 1.09–1.27). Table 3 shows health care coverage and utilization by elderly people. Health insur- ance – military or governmental – was held by 90.0% of the participants. In the 6 months prior to the survey, 58.7% of the participants reported using governmental health services with an average of 4.5 visits while only 19.0% sought the private sector for medical care. In the 12 months prior to the survey, 24.7% had been hospitalized for a mean duration of 8 days (95% CI: 5.53– 10.85). For 80.7%, the health services pro- Table 2 Health status based on reported symptoms by elderly subjects Reported Men (n = 140) Women (n = 160) Total (n = 300) OR (95%CI) symptoms No. % No. % No. % Body aches 111 79.3 141 88.1 252 84.0 1.94 (1.03–3.64) Loss of teeth 83 59.3 101 63.1 184 61.3 1.18 (0.74–1.87) Dyspepsia 76 54.3 103 64.4 179 59.7 1.52 (1.00–2.42) Loss of weight 68 48.6 97 60.6 165 55.0 1.63 (1.03–2.58) Insomnia 64 45.7 93 58.1 157 52.3 1.65 (1.04–2.60) Lack of energy 58 41.4 96 60.0 154 51.3 2.12 (1.34–3.37) Frequency 60 42.9 87 54.4 147 49.0 1.59 (1.01–2.51) Shortness of breath 57 40.7 83 51.9 140 46.7 1.57 (1.00–2.48) Constipation 53 37.9 85 53.1 138 46.0 1.86 (1.17–2.95) Impaired vision 30 21.4 60 37.5 90 30.0 2.20 (1.31–3.68) Poor appetite 30 21.4 60 37.5 90 30.0 2.20 (1.31–3.68) Dizziness 29 20.7 52 32.5 81 27.0 1.84 (1.09–3.12) Incontinence 25 17.9 53 33.1 78 26.0 2.28 (1.32–3.92) Impaired hearing 34 24.3 32 20.0 66 22.0 0.78 (0.45–1.35) Poor self-perception of health 49 35.0 83 51.9 132 44.0 2.00 (1.26–3.19) Mean (SD) no. of symptoms reported 5.56 (3.266) 7.14 (2.96) 6.40 (3.202) 1.18 95% CI 5.01–6.10 6.68–7.61 6.04–6.77 1.09–1.27 OR = odds ratio. CI = confidence interval. SD = standard deviation. 09 Comprehensive health assessment.pmd 12/20/2005, 2:35 PM338 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 339 vided were affordable; however, about 40% reported that the type of service re- quired was neither available in their area of residence nor accessible. Half (51.7%) the participants complained of long waiting times but 67.3% rated the services as being good and 72.3% were satisfied with the services provided. Capability to perform instrumental ac- tivities of daily living is shown in Table 4. Only 8.0% of the elderly participants were able to carry out all these activities indepen- dently; a substantial proportion reported the need for partial or full assistance in per- forming household chores (70.6%) and managing their own finances (68.3%). Nearly half were in need of assistance or were totally dependent on others for shop- ping (56.7%), making telephone calls (55.0%), getting around (52.7%) and pre- paring their own meals (49.3%), while 26.6% required assistance or were totally dependent in taking their medications. Im- pairment of activities of daily living was more pronounced among women as their score was 1.09 times higher than that of men (OR = 1.09; 95% CI: 1.03–1.16). Women were 3–5 times more likely to need partial or full assistance in using the tele- phone, getting around and shopping, while they were 14 times more likely to be partial- ly or fully dependent on others in managing their own finances. In contrast, they were less likely to be dependent in performing household chores and preparing meals. More than a quarter of the participants (28.0%) reported impairment of activities of daily living. Dependence in moving in- side the house was encountered among 20.3% of the elderly participants. Lower proportions were dependent on others for bathing and shaving (16.3%) as well as dressing (14.7%). Few required assistance for going to the toilet (8.0%) and eating (5.0%). Women were significantly more in need for assistance in bathing (OR = 3.21) while no gender difference was observed in other activities. Overall scores for activities of daily living showed women were more impaired than men although the difference was not statistically significant (OR = 1.14, 95% CI: 1.00–1.35) (Table 5). Table 3 Utilization of health services by participants and their opinion of the services provided Participants use and No. % opinion (n = 300) Health services utilization Coverage by health insurance 270 90.0 Use of governmental health services a 176 58.7 Use of private health servicesa 57 19.0 Hospitalizationb 74 24.7 Opinion about governmental health services in the area Affordability Affordable 242 80.7 Not affordable 58 19.3 Availability Available 163 54.3 Not available 137 45.7 Accessibility Accessible 179 59.7 Not accessible 121 40.3 Waiting time Short 145 48.3 Long 155 51.7 Quality of services provided Good 202 67.3 Poor 98 32.7 Satisfaction with the services provided Satisfied 217 72.3 Dissatisfied 83 27.7 ain the past 6 months. bin the past 12 months. 09 Comprehensive health assessment.pmd 12/20/2005, 2:35 PM339 340 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Table 4 State of dependence assessed using instrumental activities of daily living (IADL) IADL area Men (n = 140) Women (n = 160) Total (n = 300) OR (95%CI) No. % No. % No. % Using the phone Independent 89 63.6 46 28.8 135 45.0 4.32 (2.66–7.03) Assistance required 39 27.9 88 55.0 127 42.3 Dependent 12 8.5 26 16.2 38 12.7 Getting around Independent 88 62.9 54 33.8 142 47.3 3.32 (2.07–5.34) Assistance required 41 29.3 73 45.6 114 38.0 Dependent 11 7.8 33 20.6 44 14.7 Shopping Independent 88 62.9 42 26.2 130 43.3 4.76 (2.91–7.77) Assistance required 29 20.7 48 30.0 77 25.7 Dependent 23 16.4 70 43.8 93 31.0 Preparing meals Independent 49 35.0 103 64.4 152 50.7 0.29 (0.19–0.48) Assistance required 20 14.3 17 10.6 37 12.3 Dependent 71 50.7 40 25.0 111 37.0 Doing household chores Independent 22 15.7 66 41.3 88 29.3 0.27 (0.15–0.46) Assistance required 39 27.9 52 32.5 91 30.3 Dependent 79 56.4 42 26.2 121 40.3 Taking medications Independent 98 70.0 122 76.3 220 73.3 0.73 (0.44–1.21) Assistance required 34 24.3 21 13.1 55 18.3 Dependent 8 5.7 17 10.6 25 8.3 Managing finances Independent 81 57.9 14 8.8 95 31.7 14.32 (7.53–27.23) Assistance required 18 12.9 28 17.4 46 15.3 Dependent 41 29.2 118 73.8 159 53.0 Considering all IADL items Independent 12 8.6 12 7.5 24 8.0 1.16 (0.50–2.66) Assistance required 30 21.4 18 11.3 48 16.0 Dependent 98 70.0 130 81.2 228 76.0 Mean (SD) IADL score of impairment 5.07 (3.731) 6.37 (3.852) 5.76 (3.845) 1.09 95% CI (4.44, 5.70) (5.77,6.97) (5.33, 6.20) (1.03–1.16) OR was computed by summing the last two categories (assistance required and dependent). OR = odds ratio. CI = confidence interval. SD = standard deviation. 09 Comprehensive health assessment.pmd 12/20/2005, 2:35 PM340 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 341 Limitation of one or more of the basic movements was encountered by 59.7% of the participants. The movement most com- monly limited was squatting/bending (47.0%), followed by carrying a weight of 5 kg (41.7%), walking for a distance of 200–300 metres (40.0%) and getting up 3– 5 stairs (32.7%). These movements were 2–3.5 times more likely to be limited among women. Restriction of basic activities was more pronounced among women as their mean score on the Nagi scale of physical disability was 1.33 times higher than that of men (OR = 1.33, 95% CI: 1.16–1.53) (Ta- ble 6). The mean scores of elderly subjects on the mental status scale was 1.92 (SD 2.12). Women exhibited marked memory deterio- ration as their mean score [2.67 (SD 2.15), 95% CI: 2.33–3.01] was significantly high- er than that of men [1.07 (SD 1.73), 95% CI: 0.78–1.36] (OR= 1.55, 95% CI: 1.34– 1.78). On the depression scale, the mean scores of the participants was 3.16 (SD 2.64). A quarter (24.3%) scored 5 or more points which indicates the presence of de- pression. Table 7 shows factors contribut- ing to depression among the participants. Depression was significantly more likely Table 5 Functional capacity of the participants assessed by activities of daily living (ADL) ADL area Men (n = 140) Women (n = 160) Total (n = 300) OR (95% CI) No. % No. % No. % Eating Independent 134 95.7 151 94.4 285 95.0 1.33 (0.46–3.84) Dependent 6 4.3 9 5.6 15 5.0 Going to the toilet Independent 132 94.3 144 90.0 276 92.0 1.83 (0.76–4.42) Dependent 8 5.7 16 10.0 24 8.0 Bathing/shaving Independent 128 91.4 123 76.9 251 83.7 3.21 (1.60–6.44) Dependent 12 8.6 37 23.1 49 16.3 Getting dressed Independent 121 86.4 135 84.4 256 85.3 1.18 (0.62–2.25) Dependent 19 13.6 25 15.6 44 14.7 Moving around the house Independent 112 80.0 127 79.4 239 79.7 1.04 (0.59–1,83) Dependent 28 20.0 35 20.6 61 20.3 Considering all ADL items Independent 102 72.9 114 71.3 216 72.0 1.08 (0.65–1.80) Dependent 38 27.1 46 28.7 84 28.0 Mean (SD) ADL scores of impairment 0.55 (1.237) 0.82 (1.586) 0.69 (1.437) 1.14 95% CI (0.34–0.76) (0.60–1.07) (0.53–0.86) (1.00–1.35) OR = odds ratio. CI = confidence interval. SD = standard deviation. 09 Comprehensive health assessment.pmd 12/20/2005, 2:35 PM341 342 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 among women (OR = 2.13), the uneducat- ed (OR = 5.12), those who were single (OR = 4.04) and those who lived alone (OR = 3.70). Increase in age was associated with a significantly higher risk of depres- sion. The mean score of subjects aged 70 years and over [5.17 (SD 2.40), 95% CI: 4.75–5.59] was significantly higher than those less than 70 years [1.64 (SD 1.59), 95% CI: 1.40–1.89]. A higher risk of de- pression was associated with deterioration of mental status (OR = 1.72), negative self- reported health status (greater number of reported symptoms) (OR = 1.24), poor self perception of health (OR = 2.22), limitation of movement (OR = 2.51) as well as deteri- oration of instrumental (OR = 1.86) and basic activities of daily living (OR = 3.63) (Table 7). Independent predictors of instrumental and basic activities of daily living are given in Table 8. Lower performance in instru- mental and basic activities of daily living was independently predicted by increase in age, high frequency of physical and de- pressive symptoms, memory impairment and lack of formal education. In addition, lower performance in basic activities of Table 6 Limitation of basic movements based on the Nagi physical disability scale Basic movement items Men (n = 140) Women (n = 160) Total (n = 300) OR (95%CI) No. % No. % No. % Squatting/bending Not limited 97 69.3 62 38.8 159 53.0 3.57 (2.21–5.76) Limited 43 30.7 98 61.3 141 47.0 Carrying 5 kg Not limited 96 68.6 79 49.4 175 58.3 2.24 (1.40–3.59) Limited 44 31.4 81 50.6 125 41.7 Walking 200–300 m Not limited 99 70.7 81 50.6 180 60.0 2.36 (1.46–3.80) Limited 41 29.3 79 49.4 120 40.0 Getting up 3–5 stairs Not limited 106 75.7 96 60.0 202 67.3 2.08 (1.26–3.42) Limited 34 24.3 64 40.0 98 32.7 Using fingers Not limited 126 90.0 146 91.3 272 90.7 0.86 (0.40–1.88) Limited 14 10.0 14 8.7 28 9.3 Considering all Nagi items Not limited 79 56.4 42 26.3 121 40.3 3.63 (2.24–5.92) Limited in one or more functions 61 43.6 118 73.8 179 59.7 Mean (SD) Nagi scores 1.26 (1.723) 2.10 (1.702) 1.71 (1.76) 1.33 95% CI (0.97–1.55) (1.83–2.36) (1.51–1.91) (1.16–1.53) OR = odds ratio. CI = confidence interval. SD = standard deviation. 09 Comprehensive health assessment.pmd 12/20/2005, 2:35 PM342 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 343 daily living was independently predicted by the presence of chronic health problemI:s, while living alone independently predicted a better performance in instrumental activi- ties of daily living. Increase in depressive symptoms was independently predicted by living alone, negative perception of health, deterioration of mental status, impairment in instrumental and basic activities of daily living and limitation of movement. These variables contributed to 69.6% of the vari- ance in the depression score (Table 8). Table 7 Univariate logistic regression of predictors of depression status among the participants Predictor Depressed (n = 73) Not depressed (n = 227) OR (95% CI) No. % No. % Sex Malea 24 32.9 116 51.1 2.13 Female 49 67.1 111 48.9 (1.22–3.71) Marital status Marrieda 31 42.5 170 74.9 4.04 Single 42 57.5 57 25.1 (2.32–7.02) Educational status Formal educationa 4 5.5 52 22.9 5.12 No formal education 69 94.5 175 77.1 (1.79–14.71) Living arrangements With family membersa 57 78.1 211 93.0 3.70 Alone 16 21.9 16 7.0 (1.74–7.85) Chronic diseases Absenta 15 20.5 63 27.8 1.49 Present 58 79.5 164 72.2 (0.79–2.81) Self-perceived health Gooda 30 41.1 138 60.8 2.22 Poor 43 58.9 89 39.2 (1.29–3.80) Mean (SD) CI Mean (SD) CI Age in years 66.87 (6.783) 65.98–67.76 78.24 (8.993) 76.13–80.33 1.18 (1.13–1.23) Self-reported health statusb 7.96 (2.965) 7.96–7.27 5.90 (3.119) 5.49–6.31 1.24 (1.13–1.37) Instrumental activities of daily living 10.21 (2.619) 9.59–10.82 4.33 (2.991) 3.94–4.72 1.86 (1.61–2.16) Activities of daily living 2.25 (2.107) 1.76–2.74 0.19 (0.538) 0.12–0.26 3.63 (2.53–5.22) Basic movement 3.51 (1.324) 3.20–3.82 1.13 (1.468) 0.94–1.32 2.51 (2.02–3.11) Mental status 3.68 (2.390) 3.13–4.25 1.36 (1.672) 1.14–1.58 1.72 (1.48–2.00) aBaseline category. bScores based on the number of symptoms or complaints reported. OR = odds ratio. CI = confidence interval. SD = standard deviation. 09 Comprehensive health assessment.pmd 12/20/2005, 2:35 PM343 344 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Discussion Traditionally, research on ageing has been concerned with health but recently the con- cept of functional capacity has been at- tracting growing attention [6]. This study explored the health status of senior citizens based on medical diagnosis, symptoms of diseases and capacity to perform activities necessary for independent living. Among the studied population, body movement and basic activities were less likely to be restricted compared to instru- mental activities of daily living as only 8% were capable of performing activities that enabled them to live independently. Lauk- kanen et al. [11] and Sauvaget et al. [12] indicated that independent life in instrumen- tal activities of daily living is much shorter than independent life in basic activities or mobility. Among the studied population, whatever the pattern or extent of assistance required, none of the elderly had an unmet Table 8 Independent predictors of instrumental and basic activities of daily living and depression among the participants Independent predictors Standardized Computed P-value coefficient (β) t for β Instrumental activities of daily living Age (years) 0.145 3.166 0.002 Self-reported health status 0.132 3.534 0.000 Mental status 0.153 3.519 0.001 Depression 0.503 10.210 0.000 Living arrangements (with family = 1; living alone = 2 ) –0.074 2.078 0.039 Education (educated = 1; uneducated = 2) 0.162 4.296 < 0.001 Adjusted R 2 = 0.642 Basic activities of daily living Age (years) 0.183 3.373 0.001 Self-reported health status 0.140 2.877 0.004 Mental status 0.200 3.870 < 0.001 Depression 0.394 6.787 < 0.001 Chronic health problems (absent = 1; present = 2) 0.147 3.224 0.001 Education (educated = 1; uneducated = 2) –0.141 3.137 0.002 Adjusted R 2 = 0.499 Depression Self-perceived health (good =1; poor = 2) 0.318 8.240 < 0.001 Basic movements 0.222 5.090 < 0.001 Instrumental activities of daily living 0.268 4.893 < 0.001 Basic activities of daily living 0.135 3.038 0.003 Mental status 0.091 2.268 0.024 Living arrangements (with family = 1; alone = 2) 0.077 2.353 0.019 Adjusted R 2 = 0.696 09 Comprehensive health assessment.pmd 12/20/2005, 2:35 PM344 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 345 need (R. Youssef, unpublished data, 2004). This is a characteristic of Arab society, es- pecially small rural and Bedouin communi- ties where children assume full responsibility of their elderly parents whether they are living with them or in their own home. In the Bedouin community in this study, almost two-thirds of the elderly were living with their children and grand- children. A longitudinal study from Japan provides evidence of the progressive dete- rioration in instrumental activities of daily living among the elderly living with their children simply because they rely on them even if they have no strong need for their help [13]. In this respect, the role of chil- dren should be strengthened and oriented toward encouraging old people to maintain independence. Such a role could alleviate much of the social burden on the govern- ment and contribute to the mental well-be- ing of old people. It is worth mentioning that this study as well as that of Schulman et al. [5] demonstrated that old people liv- ing alone were at a much higher risk of de- pression although they are able to maintain independence. A large body of literature has document- ed that women are more likely to report deterioration and the use of assistance in instrumental [14,15] and basic [16] activi- ties of daily living as well as impairment of body movement [14,16]. This was also shown in the present study, except for household chores and meal preparation, which was explained by Laukkanen et al. [11] as cultural differences in coping with everyday activities. Previous studies have attributed the apparent greater impairment and disability among women to the fact that men experience disability at a younger age with rapid deterioration [12] while women live longer to very old age when multiple health problems and disability are common [1,17]. Consequently, they spend more time in a disabled state [17]. In some soci- eties, the lower status of women, lack of education and their restricted role as family caregiver contribute to their increased risk of ill-health and disability [1,13,18]. In this study, men and women were of compara- ble age yet the women had received less formal education. It was among the wom- en that chronic diseases, multitude of symptoms and negative perception of health were encountered more often, which could explain the gender variation in the state of dependence and limited func- tional capacity. In this study, chronic dis- eases predicted dependence in basic activities of daily living and high frequency of symptoms predicted low summary per- formance in both instrumental and basic activities of daily living. It is thus the bur- den experienced of disturbance of body systems, and the pain and discomfort rath- er than the disease status that predict de- pendence and the need for assistance, as suggested by Kiosses et al. [15]. Each symptom should therefore be taken seri- ously rather than considering it a feature of old age. A quarter of the participants in this study were classified as depressed, which is higher than the 17.5% reported from Abha, Saudi Arabia [19]. Such a consider- able difference could be attributed to the variation in the socioeconomic level be- tween the two populations as well as the high prevalence of chronic diseases and the over-presentation of women in this study. This study and others point to the higher likelihood of depression among elderly women [18–20]. Among the studied popu- lation, depression was not found to be characteristic of female gender but a con- sequence of their restricted capacity in coping with the demands of daily living and the negative perception of health. Previous studies have underscored the impact of 09 Comprehensive health assessment.pmd 12/20/2005, 2:35 PM345 346 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 limitation and disability [5,15,20–22] and poor self-rated health status [19,23] on the increase in depressive symptoms that is of- ten attributed to to gender variation. In- deed, each depressive symptom increases impairment and disability and reduces the chance of recovery [21]. This study dem- onstrated that depression and the limited coping capacities with everyday demands reinforce each other. Such information is important for primary care physicians who tend to focus on the medical diagnosis of physical ailments. In old age, depression is often overlooked as a clinical entity as it is assumed to be the normal response to age- ing and physical loss [24]. Public health strategies should be directed toward com- prehensive assessment of geriatric popula- tions giving equal attention to the effective management of depression, co-morbidity and functional limitation. Primary care phy- sicians should be particularly vigilant for functional decline in old people who per- ceive their health negatively as further de- cline within a year would be expected [25]. During natural ageing, some cognitive capacities, particularly memory, decline [1]. This study demonstrated that for each depressive symptom, memory declined by 9%. Also, memory impairment was associ- ated with a low summary performance on instrumental and basic activities of daily liv- ing. This was in accordance with other studies that have reported the role of cogni- tive impairment in accelerating depression [1] and generating a state of dependence [26,27]. Continuous mental stimulation can result in a deceleration of the memory loss that accompanies the ageing process with a subsequent positive impact on physical and psychological well-being of old people [4]. Increased life-span and the resulting ageing of populations is one of the greatest achievements of our time. Successful age- ing, which implies adding life to years, can only be realized by addressing all the nega- tive impacts of growing old collectively and simultaneously. Inevitably, this will result in a significant increase in economic and so- cial demands [1]. In Jordan, rapid popula- tion growth and the growing number of old people will put considerable pressures on the health care system. Expenditure on health currently represents 9.12% of its gross domestic product [28]. In the future, Jordan will have to increase this expendi- ture. The per capita utilization of health ser- vices is 2.41 visits with much higher rate of utilization by old people [28]. Among the elderly of this study, more than half had paid on average 4.5 visits to health care fa- cilities in the past 6 months. As 90% of the studied population were covered by gov- ernmental health insurance , demands are higher on the services provided by the Min- istry of Health which are spread all over the country and the Royal Medical Services which are centred in the capital and provide specialized care. There are few private health care facilities in this remote area and they are used when there is a real or per- ceived problem with the quality of govern- mental services [28]. In this regard, it is worth mentioning that almost a third of the participants were dissatisfied with the ser- vices provided and over a quarter consid- ered the quality was poor. Problems with the availability and accessibility of services were also reported in view of the scattered dwelling, lack of a transportation system and the need to seek services in the capital. Research is needed to assess the quality of services and patterns of delivery to old peo- ple. As old people live in the community, health care should be provided by primary health care centres as well as by outreach services which focus on the highly vulner- able, such as women and the very old. In this area, the newly launched community- oriented school of medicine ay Mu’tah Uni- 09 Comprehensive health assessment.pmd 12/20/2005, 2:35 PM346 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 347 versitycould take part as well. In Jordan, medical schools should define “ageing” as a community problem and allot more time for geriatric medicine on both preventive and curative levels. In addition, community surveys of geriatric populations for health assessment and ensuring strong links to the available medical and social services could be part of the practical training of medical students. References 1. Active ageing: a policy framework. Geneva, World Health Organization, 2002. 2002. 2. Jordan Department of Statistics and ORC Macro. Jordan Population and Family Health Survey 2002. Calverton, Maryland, Department of Statistics and ORC Macro, 2003. 3. United States Bureau of Census. Inter- national data base. IDB population pyra- mids. Available at: http:// w w w . c e n s u s . g o v / c g i - b i n / i p c / idbpyrs.pl?cty=JO&out=s&ymax=250 (accessed 25/05.05). 4. Thapar GD. Approach to the problems of the aged. Journal of the Indian Medical Association, 2004, 102(2):93–6. 5. Schulman E et al. Depression and asso- ciated characteristics among commu- nity-based elderly people. Journal of allied health, 2002, 31(3):140–6. 6. Growing older – staying well: ageing and physical activity in everyday life. Geneva, World Health Organization, 1998. 7. Lawton MP, Brody EM. Assessment of older people: Self-maintaining and in- strumental activities of daily living. Ger- ontologist, 1996, 9:179–86. 8. Katz S et al. Studies of illness in the aged. The index of ADL: A standardized mea- sure of biological and psychosocial func- tion. Journal of the American Medical Association, 1963, 85:914–9. 9. Nagi SZ. An epidemiology of disability among adults in the United States. Milbank Memorial Fund quarterly, 1976, 54:439–67. 10. Van Marwijk HW et al. Evaluation of the feasibility, reliability and diagnostic value of shortened versions of the geriat- ric depression scale. British journal of general practice, 1995, 45(393):195–9. 11. Laukkanen P et al. Factors related to car- rying out everyday activities among eld- erly people aged 80. Aging, 1994, 6(6):433–43. 12. Sauvaget C et al. Health-life expectancy according to various functional levels. Journal of the American Geriatrics Soci- ety, 1999, 47(11):1326–31. 13. Ishizaki T, Kobayashi Y, Kai I. Functional transitions in instrumental activities of daily living among older Japanese. Jour- nal of epidemiology, 2000, 10(4):249– 54. 14. Murtagh KN, Hubert HB. Gender differ- ence in physical disability among an eld- erly cohort. American journal of public health, 2004, 94(8):1406–11. 15. Kiosses DN et al. Executive dysfunction and disability in elderly patients with major depression. American journal of geriatric psychiatry, 2001, 9(3):269–74. 16. Beckett LA et al. Analysis of change in self–reported physical function among older persons in four population studies. American journal of epidemiology, 1996, 143(8):766–78. 17. Dunlop DD, Hughes SL, Manheim LM. Disabilities in activities of daily living: patterns of change and a hierarchy of 09 Comprehensive health assessment.pmd 12/20/2005, 2:35 PM347 348 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 disability. American journal of public health, 1997, 87(3):378–83. 18. Bonita R. Women, ageing and health: achieving health across the life span. Geneva, World Health Organization, 1998. 19. Abolfotouh MA et al. Psychosocial as- sessment of geriatric subjects in Abha city, Saudi Arabia. Eastern Mediterra- nean health journal, 2001, 7(3):481–91. 20. Hybels CF, Blazer DG, Pieper CF. Toward a threshold for subthreshold depression: an analysis of correlates of depression by severity of symptoms using data from an elderly community sample. Geron- tologist, 2001, 41(3):357–65. 21. Cronin-Stubbs D et al. Six-year effect of depressive symptoms on the course of physical disability in community-living older adults. Archives of internal medi- cine, 2000, 160(20):3074–80. 22. Dunlop DD et al. Arthritis and heart dis- ease as risk factor for major depression: The role of functional limitation. Medical care, 2004, 42(6):502–11. 23. Smits CH, Deeg DJ, Jonker C. Cognitive and emotional predictors of disablement in older adults. Journal of aging and health, 1997, 9(2):204–21. 24. Raj A. Depression in the elderly. Tailoring medical therapy to their special needs. Postgraduate medicine, 2004, 115(6):26–8, 37–42. 25. Lyness JM et al. Self-rated health, de- pression, and one-year health outcome in older primary care patients. American journal of geriatric psychiatry, 2004, 12(1):110–3. 26. Nourhashemi F et al. Instrumental activi- ties of daily living as potential marker of frailty: A study of 7364 community-dwell- ing elderly women (the EPIDOS study). Journals of gerontology. Series A, Bio- logical sciences and medical sciences, 2001, 56(7):M448–53. 27. Ishizaki T et al. Predictors for functional decline among non-disabled older Japanese living in a community during a 3-year follow-up. Journal of the Ameri- can Geriatrics Society, 2000, 48(11):1424–9. 28. Brosk H et al. Jordan National Health Accounts. Technical Report No. 49. Bethesda, Maryland, Partnerships for Health Reform Project, Abt Associates Inc., 2000. 09 Comprehensive health assessment.pmd 12/20/2005, 2:35 PM348 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 349 Validity and reliability of the Persian translation of the SF-36 version 2 questionnaire N. Motamed,1 A.R. Ayatollahi,2 N. Zare2 and A. Sadeghi-Hassanabadi 3 1Department of Community Medicine, Sari Medical School, Sari, Islamic Republic of Iran (Correspondence to N. Motamed: nilo1351@yahoo.com). 2Department of Biostatistics, Shiraz School of Public Health, Shiraz, Islamic Republic of Iran. 3Department of Community Medicine, Shiraz Medical School, Shiraz, Islamic Republic of Iran. Received: 24/03/03; accepted: 18/02/04 ABSTRACT To assess the validity and reliability of a Persian translation of the SF-36 questionnaire (version 2) as a tool for use in the general population and outpatients, we carried out a cross-sectional study of the staff of Shiraz Medical School during November 2001. Response rate was 83.6%. Internal consistency for 8 scales using Cronbach’s α was 0.87 (minimum recommended 0.85). Construct validity was acceptable as correlation between scales and some sociodemographic characteristics was significant. The correlation coefficient between 8 scales and related principal components was also acceptable. This translation is an appropriate tool for assessing health perceptions of the population. Validité et fiabilité de la traduction en langue perse du questionnaire SF-36 version 2 RÉSUMÉ Afin d’évaluer la validité et la fiabilité d’une traduction en langue perse du questionnaire SF-36 version 2 en tant qu’instrument pour utilisation dans la population générale et chez les patients des consu- ltations externes, nous avons réalisé une étude transversale auprès du personnel de l’École de médecine de Chiraz durant le mois de novembre 2001. Le taux de réponse était de 83,6 %. L’homogénéité interne des 8 échelles en utilisant l’α de Cronbach était de 0,87 (minimum recommandé de 0,85). La validité du construit était acceptable, la corrélation entre les échelles et certaines caractéristiques sociodémographiques étant significatives. Le coefficient de corrélation entre les 8 échelles et les principales composantes du construit était également acceptable. Cette traduction est un outil approprié pour l’évaluation des perceptions de la population à l’égard de la santé. 10 Validity and reliability.pmd 12/20/2005, 2:35 PM349 350 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Introduction It is important to be able to measure the health perceptions of the population to as- sess the benefit of health care interventions and to target services. Researchers have developed measures to assess the health of people with specific diseases or disabilities, but these are of limited application when studying people with more than 1 condition or comparing perceived health across dif- ferent groups [1]. The work of a health professional in- cludes determining effective treatment re- gimes to help patients live better and longer lives. Perhaps more important, it also in- volves giving patients the tools to make in- formed decisions to improve their own quality of life. What is required is a com- prehensive, and at the same time sensitive, measure of the full range of illness [1]. The SF-36 is specifically designed for use in the general population and among ambulatory patients. This questionnaire is sensitive to differences in health status among patients who seek care in various outpatient settings [2]. The original general health measure was lengthy, containing 108 items. In an attempt to develop a general health survey that was comprehensive and psychometri- cally sound, yet short enough to be practi- cal for use in large-scale studies of patients in practice settings, the authors of the SF- 36 experimented with several shortened versions. Version 2 is the substantially re- vised form. Population reference values for this instrument are available from the Unit- ed States of America and England and it is also being adapted into languages other than English by the International Quality of Life Assessment project [2]. The SF-36 is a self-administered ques- tionnaire containing 36 items. It takes about 5 minutes to complete. It measures health on 8 multi-item scales, covering functional status, well-being and overall evaluation of health (Table 1) [1]. The measurement model underlying the construction of the SF-36 scales and sum- Table 1 Dimensions of the SF-36 health survey questionnaire, adapted from Brazier et al. and Ware et al. [1,3] Summary measure Scale No. of questions Physical functioning Physical functioning 10 Role limitations (physical) 4 Pain 2 General healtha 5 Health changeb (1) Mental functioning Social functioninga 2 Role limitations (emotional) 3 Mental health 5 Vitalitya 4 Total 36 aHas strong correlation with both components. bThis item was not included in the 8 scales, nor was it scored. 10 Validity and reliability.pmd 12/20/2005, 2:35 PM350 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 351 mary measures has 3 levels: items; 8 scales that aggregate 2–10 items each; and 2 sum- mary measures, the physical component summary (PCS) and the mental component summary (MCS) measures, that aggregate scales (using factor analysis) (Table 1). All but 1 of the 36 items (self-reported health change) are used to score the 8 SF-36 scales. Each item is used in scoring only 1 scale. Three scales, physical functioning, role limitations (physical problems) and bodily pain, correlate most highly with the physical component and contribute most to the scoring of the PCS measure. The men- tal component correlates most highly with the mental health, role limitations (emotion- al problems) and social functioning scales, which also contribute most to the scoring of the MCS measure. Three of the scales (vitality, general health and social function- ing) have significant correlation with both components [3]. It has been claimed that items in the SF- 36 questionnaire detect positive as well as negative states of health [1]. In 6 of the 8 scales, patients are asked to rate their re- sponses on 3-point or 6-point gradings rather than simply responding yes or no. For each scale, item scores are coded, computed, and transformed on to a scale from 0 (worst health) to 100 (best health). Internal consistency is the extent to which items within a scale are correlated with each other. This can be examined by several methods but the most widely used is Cronbach’s α, based on correlation be- tween items [1]. Validity of a health measure is conceptu- ally difficult to prove without a standard. Types of validity include face validity, where the relevance of the measurement appear obvious to the investigator and the questions supply the information the in- vestigator needs; content validity, if the variable to be measured is a composite one and all the component elements of the vari- able (as conceived) are measured; and con- struct validity, the extent to which a particular measure relates to measures con- sistent with theoretically derived hypo- theses concerning the concepts, or con- structs, that are being measured. One way of checking construct validity is to see how well the measure discriminates between groups that there is reason to believe should differ in the characteristics under study [5]. For example women, older people, and people in semi-skilled and unskilled occu- pations might be expected to perceive their health as being relatively poor; people mak- ing use of health services might also be ex- pected to perceive their health as poorer than non-users [1]. While information is rapidly becoming available on the properties of the American and British versions in ambulatory patients, less is known regarding its performance in healthy people, and even less about the per- formance of adaptations into other lan- guages. Our purpose was to examine the validity and reliability of the Persian lan- guage adaptation of SF-36 in the staff of Shiraz Medical School as a particularly healthy group. Methods In our study, we used the Persian language version of the SF-36 on staff of Shiraz Medical School. The wording of 4 items was altered slightly due to cultural differ- ences (gardening and ping-pong instead of bowling or playing golf, 1 kilometre instead of 1 mile, and several hundred metres in- stead of several hundred yards). Translated questionnaires, together with questions on sociodemographic characteristics, includ- ing age, sex, marital status, years of mar- riage, number of children, occupation, years of work, education and salary, were 10 Validity and reliability.pmd 12/20/2005, 2:35 PM351 352 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 distributed to all 134 members of staff working in the wards of Shiraz Medical School in November 2001 and collected a few days later. Crude scores of the 8 scales (0 to 100) were standardized based on mean = 50 and standard deviation (SD) = 10 and mean and SD for the general population [4]. [(x – µx)/δx] = [(ys – 50)/10] = > ys = 50 + 10 [(x – µs)/δs] where: µs = 50 = general population mean δs = 10 = general population SD ys = standard score of each person x = crude score (0–100) of each person µx = norm-based score of each item (for example physical functioning) δx = norm based SD of each item (for ex- ample physical functioning). Thus, scores of different scales are not crude scores of 0–100. This allows us to compare scores across the various scales and summary measures [4]. The statistical package used was SPSS, version 10. The responses to the questionnaire were subjected to Chronbach’s α to test reliability. To assess face validity and con- tent validity, after translation of the ques- tionnaire into Persian by an experienced translator, we discussed it with health spe- cialists who were fluent in Persian and En- glish and then incorporated the changes in 2 steps (for example changes in wording) to develop the final questionnaire. We used the Student t-test and the Kruskal–Wallis 1-way analysis of variance to test whether the SF-36 scores differed significantly between groups. In addition we used linear regression and Spearman’s rank correlation to assess the association between variables and different scales. We expected that the correlation between each scale and its related summary measure would be greater than its correlation with the other summary measure. The level of the variables were: • age (years): < 25; 25–34; 35–44; 45– 55; > 55 • sex: female; male • marital status: single; married; divorced • years of marriage: < 10; 10–19; ≥ 20 • number of children: ≤ 3; > 3 • education: did not complete high school; high-school diploma; university degree • occupation: high; middle; low • years of work: < 10; ≥ 10. Results We received completed questionnaires from 112 of 134 individuals, giving a re- sponse rate of 83.6%. On 74 question- naires (66.1%) all 36 items were complete, on 20 (17.9%) at least 1 item was missing and on 18 (16.1%) at least 2 items were missing. The most frequent item that was missing was, “During the past 4 weeks, have you been very nervous?” Only 1 paper had so much data missing that it could not be processed. Therefore, in all, data from 111 questionnaires were analysed. The age of the participants ranged from 19 years to 57 years, mean 36.08 years (SD 9.57). Fifty per cent were female. Marital status was: single 34 (30.6%), mar- ried 76 (68.5%) and divorced 1 (0.9%). The monthly salary was: 78 (70.3%) earned less than 800 000 rials (US $1 = 1755 rials, November 2001) and 33 (29.7%) earned 800 000–2 000 000 rials. Education was: 62 (55.9%) of the partici- pants had high-school diploma or less and 49 (44.1%) had university education. Fam- ily size was 0–3 children for 88.7% of the participants. 10 Validity and reliability.pmd 12/20/2005, 2:35 PM352 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 353 We found that internal consistency us- ing Cronbach’s α at the level of 8 scales was acceptable; it exceeded the recom- mended minimum of 0.85 as it was 0.87 in 103 cases [1]. For the 4 scales of the PCS measure (general health, bodily pain, role limitations, physical functioning) Cron- bach’s α was 0.79 in 106 cases and for the 4 scales of the MCS measure (mental health, role limitations, social functioning, vitality) was 0.79 in 105 cases. Table 2 shows Cronbach’s α according to sex, oc- cupation and education level. For mental health, Cronbach’s α was higher for wom- en than for men, but for physical health it was higher for men. The comparatively low Cronbach’s α scores seen here may in part be related to the small number of scales (4 scales). Concerning construct validity, and con- sidering the different levels of variables (Table 3), we found there was a significant linear correlation between physical func- tioning, PCS measure and age (Spearman’s rank correlation coefficient 0.27 and 0.18 respectively) (P < 0.05). Physical func- tioning and the PCS measure decreased with increasing age. Even after the intro- duction of sex into the age equation, the re- lationship between physical functioning and the PCS measure was significant. There was also a significant linear correla- tion between sex and role limitations for physical problems and role limitations for emotional problems (Spearman’s rank cor- relation coefficient 0.21 and 0.23 respec- tively). Women had significantly higher scores than men in role limitations for physical problems and in role limitations for emotional problems (P < 0.005). After ad- justing for sex according to age, this was still true. Table 2 Cronbach’s α for the different scales according to sex, occupation and education level Characteristic 4 scales, mental 4 scales, physical 8 scales health health Cronbach’s α No. Cronbach’s α No. Cronbach’s α No. Sex Female 0.82 52 0.77 53 0.86 51 Male 0.79 53 0.81 53 0.88 52 Occupationa Low 0.76 18 0.70 18 0.85 18 Middle 0.86 25 0.82 25 0.90 23 High 0.77 61 0.81 62 0.86 61 Education Did not complete high school 0.79 19 0.67 19 0.85 19 High-school diploma 0.83 38 0.84 37 0.91 36 University 0.74 47 0.78 49 0.81 47 aOccupation levels: high = permanent staff; medium = temporary staff on yearly contract; low = hired (and paid) by the day. 10 Validity and reliability.pmd 12/20/2005, 2:35 PM353 354 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 There was also a significant linear rela- tionship between years of marriage and physical functioning and the PCS measure (Spearman’s rank correlation coefficient 0.25 and 0.26 respectively), as the longer the participant had been married, the lower the score for physical functioning and the PCS measure, although after the introduc- tion of age that was not significant. Kruskal–Wallis 1-way analysis of variance revealed a significant relationship between marital status and physical functioning and the PCS measure (P < 0.05) as married in- dividuals had lower scores in these. Individuals with a maximum of 3 chil- dren had significantly higher scores in so- cial functioning and general health (P < 0.05), although Spearman’s rank correla- tion coefficient for number of children and physical functioning was not significant (P = 0.16). The higher the occupational level, the lower the role limitations (physical prob- lems) score (Spearman’s rank correlation coefficient 0.23; P = 0.017). The only scale that was different for the 2 income levels was social functioning. In other words, in the lower income group mean score was significantly lower (P < 0.05). Also, people who had worked for a maximum of 10 years had significantly lower scores in physical functioning (P < 0.05): the more years a participant had worked, the lower the score of physical functioning (Spearman’s rank correlation coefficient 0.21; P = 0.026). There was a significant positive rela- tionship between education level and the PCS measure and role limitations (physical problems) (Spearman’s rank correlation coefficient 0.22 (P = 0.026) and 0.20 (P = 0.036) respectively. Table 3 shows correlation coefficients between 8 scales and 2 summary measures for our study (Islamic Republic of Iran) Table 3 Correlation coefficient of the 8 scales and 2 summary measures: comparison of values obtained in Iran (present study), France [2] and the United States of America [3] Scales Correlation coefficient Iran France USA PCS MCS PCS MCS PCS MCS Physical functioning 0.71 0.09 0.74 –0.06 0.85 0.12 Role limitations, physical problems 0.67 0.22 0.63 0.33 0.81 0.27 Bodily pain 0.82 0.20 0.72 0.27 0.76 0.28 General health 0.60 0.41 0.63 0.40 0.69 0.37 Vitality 0.42 0.60 0.41 0.75 0.47 0.65 Social functioning 0.29 0.72 0.36 0.73 0.42 0.67 Role limitations, emotional problems 0.27 0.58 0.01 0.82 0.16 0.78 Mental health –0.002 0.85 0.2 0.84 0.17 0.87 PCS = physical component summary measure. MCS = mental component summary measure. 10 Validity and reliability.pmd 12/20/2005, 2:35 PM354 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 355 along with data from France and the United States of America [2,3] for comparison. Of the 8 scales, physical functioning and bodi- ly pain had the strongest correlation with perception of physical health, similar to val- ues recorded for the United States of Amer- ica and France. Furthermore, in all 3 studies mental health had the strongest cor- relation with the MCS measure. Discussion We achieved a response rate of 83.6%, suggesting that the SF-36 questionnaire is an acceptable tool for measuring health perception in the general population. Either the present Persian translation of the ques- tionnaire is successful or the properties of the original instrument are robust enough to withstand our translation into Persian. Non-response (attrition) rates could per- haps be improved by a better translation. Our findings support the developers’ claims of internal consistency [3] for the SF-36. With rare exceptions, published reliabili- ty statistics have exceeded the minimum standard of 0.7 recommended for mea- sures using in-group comparisons; most have exceeded 0.8 [3]. In our study, Cron- bach’s α for each summary measure was 0.79. In addition, Cronbach’s α for the 8 scales was 0.87. While studies of subgroups indicated slight declines in reliability for the more- disadvantaged respondents, reliability coef- ficients consistently exceeded recommen- ded standards for group level analysis [3]. In our study, Cronbach’s α exceeded 0.85 for the 8 scales (except university educa- tion) and 0.70 for each of the 4 mental and 4 physical scales (except for those who did not complete high school in physical health) (Table 2). It is noteworthy that in both occupation and education, people in the middle groups had the highest Cron- bach’s α. Cronbach’s α for the PCS mea- sure was higher in men and for the MCS measure it was higher in women. This may be the result of differing concerns men and women have regarding their physical and mental health. The evidence for construct validity of the SF-36 was substantial. The distribution of scores observed by sociodemographic characteristic showed that the question- naire is more valid for physical than for mental health, although we did not expect validity for mental health to be significantly different between the sexes or with in- creasing age (borne out by our findings). McCabe et al. found a significant differ- ence in the mean scores for men and wom- en, but did not find any correlation between age and scores [6]. Although in the study by Brazier et al., men perceived themselves to be signifi- cantly healthier than women, except on the general health scales [1], in our study, only in 2 scales (role limitations physical prob- lems and role limitations emotional prob- lems) were men and women different, with the women’s score being higher. Brazier et al. found a significant correlation between age and physical functioning and bodily pain [1]. In our study we also found a cor- relation between age and physical function- ing and the PCS measure (decreasing significantly with increasing age, even after controlling for sex). In addition, on exam- ining sociodemographic characteristics such as education, occupation, number of children, income and marital status, we saw considerable differences in scores. For example, as in the study of Regidor et al., where perceived health status declined with decreasing educational level [7], in our study, scores for role limitations (physical 10 Validity and reliability.pmd 12/20/2005, 2:35 PM355 356 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 problems) and for the PCS measure in- creased with increasing level of education. Another way to assess the construct validity is to assess correlation between in- dividual scales and principal components because if we find a meaningful correla- tion, we can be sure that the summarization of those scales to components is correct. In the American and French studies, physi- cal functioning correlated most strongly with the PCS measure and mental health with the MCS measure (Table 4). The only difference in our study was that bodily pain was more strongly correlated with the PCS measure than physical functioning. There is no obvious explanation for this at present. Also, as in the other 2 studies, vi- tality and general health correlated moder- ately with both principal components. In attempting to be comprehensive, ex- isting general health questionnaires may be too long or require interviews, or both. In primary care or community settings, con- tact time with the patient is often short and thus, to be practical and acceptable to the population, the questionnaire should be brief, easy to use and preferably self- administered. These features are also im- portant for researchers. The SF-36 ques- tionnaire seems to meet these criteria, even when used in a particularly healthy popula- tion and after translation into Persian. In summary, we believe that this ver- sion of SF-36 is suitable for its intended purpose, i.e. to measure health in the gener- al population, although some items that are related to mental health require more revi- sion. We consider that with further work to normalize this instrument, and after con- ducting research with a larger sample size, we could use the Persian language version of the SF-36 questionnaire as an inexpen- sive and rapid tool for the population of the Islamic Republic of Iran. Table 4 Spearman’s linear regression coefficient (rs) between scales and characteristics of participants Characteristic Scale rs P-value Age Physical functioning –0.27 0.004 Physical component summary measure –0.18 0.064 Sex Role limitations, physical problems 0.21 0.024 Role limitations, emotional problems 0.23 0.016 Years of marriage Physical functioning –0.25 0.011 Physical component summary measure –0.26 0.010 No. of children Physical functioning –0.15 0.160 Occupation level Role limitations, physical problems –0.23 0.017 Marital status Physical functioning –0.21 0.023 Years of work Physical functioning –0.21 0.026 Education level Physical component summary measure 0.22 0.026 Role limitations, physical problems 0.2 0.036 10 Validity and reliability.pmd 12/20/2005, 2:35 PM356 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 357 References 1. Brazier JE et al. Validating the SF-36 health survey questionnaire: new out- come measure for primary care. British medical journal, 1992, 305(6846):160– 4. 2. Perneger TV et al. Validation of a French- language version of the MOS 36-Item Short Form Health Survey (SF-36) in young healthy adults. Journal of clinical epidemiology, 1995, 48(8):1051–60. 3. Ware JE Jr, Gandek B. Overview of the SF-36 health survey and the Interna- tional Quality of Life Assessment (IQOLA) project. Journal of clinical epi- demiology, 1998, 51(11):903–12. 4. Ware JE Jr. SF-36® health survey up- date. Quality Metric Incorporated (http:// w w w. s f - 3 6 . o r g / t o o l s / s f 3 6 . s h t m l , accessed 29 December 2004). 5. Abramson JH, Abramson ZH. Survey methods in community medicine, 5th ed. Philadelphia, Churchill Livingstone, 1999:171–204. 6. McCabe CJ et al. Measuring the mental health status of a population: a compari- son of the GHQ-12 and the SF-36 (MHI- 5). British journal of psychiatry, 1996, 169(4):517–21. 7. Regidor E et al. Association between educational level and health related quality of life in Spanish adults. Journal of epidemiology and community health, 1999, 53(2):75–82. Third Regional Conference of Medical Journals in the Eastern Medi- terranean Region The Third Regional Conference of Medical Journals in the Eastern Mediterranean Region will be held in Shiraz, Islamic Republic of Iran from 26 to 29 January 2006. More information and the Conference website will be announced soon. 10 Validity and reliability.pmd 12/20/2005, 2:35 PM357 358 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Screening for microalbuminuria by use of microproteinuria Z. Khatami,1 D.W. McIlveen,2 S.G. Nesbitt 2 and I.S. Young 2 1Biochemistry Department, Queen Elizabeth Hospital, Sheriff Hill, Gateshead, United Kingdom (Correspondence to Z. Khatami: zahra.khatami@ghnt.nhs.uk). 2Biochemistry Department, Kelvin Building, Royal Victoria Hospital, Belfast, United Kingdom Received: 12/03/01; accepted: 25/04/04 ABSTRACT We aimed to develop a reliable, low cost method to assess the early stages of renal impairment in diabetes, for use in high-risk populations in countries with limited resources. We evaluated a trichloroacetic acid (TCA) turbidimetric method for microproteinuria screening in patients with diabetes. The method was compared with an immunoturbidimetric procedure for the detection of microalbumuniuria. Both methods performed within limits of allowable uncertainty based on inter- and intra-individual variation. A urinary albu- min/creatinine ratio of 3.0 g/mol, assumed as diagnostic of microalbuminuria, was found to correlate with a cut-off value of 24 mg/L for microproteinuria. The clinical sensitivity and specificity of the TCA method determined against this ratio were 86% and 90% respectively. The reliability and practicability of the TCA method renders it suitable for the detection of early stage renal damage, with emphasis on screening high- risk populations in countries with limited resources. Dépistage de la microalbuminurie au moyen de la microprotéinurie RÉSUMÉ Notre objectif était de mettre au point une méthode, fiable et peu coûteuse pour évaluer les premiers stades de l’insuffisance rénale dans le diabète afin de l’utiliser dans les populations à haut risque des pays qui ont des ressources limitées. Nous avons évalué une méthode turbidimétrique utilisant l’acide trichloroacétique pour le dépistage de la microprotéinurie chez des patients diabétiques. Celle-ci a été comparée à la méthode immunoturbidimétrique pour le dépistage de la microalbuminurie. Les deux mé- thodes ont produit des résultats se situant dans les limites de l’incertitude permise sur la base de la variation inter- et intra-individuelle. On a trouvé une corrélation entre un rapport albumine/créatinine urinaire de 3,0 g/mol, sur la base duquel est diagnostiquée une microalbuminurie, et une valeur seuil de 24 mg/L pour la microprotéinurie. La sensibilité et la spécificité cliniques de la méthode à l‘acide trichloroacétique déterminées en fonction de ce rapport étaient de 86 % et 90 % respectivement. La fiabilité et la praticabilité de la méthode à l’acide trichloroacétique la rendent appropriée pour le dépistage d’une atteinte rénale précoce, en particu- lier le dépistage dans les populations à haut risque des pays qui ont des ressources limitées. 11 Screening for microalbuminuria.pmd 12/20/2005, 2:35 PM358 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 359 Introduction There is currently a global epidemic of type 2 diabetes. At present an estimated 150 mil- lion people worldwide have diabetes, a fig- ure that is predicted to rise to 300 million by 2025 [1]. Surveys have indicated that while diabetes is widely recognized in North American and western European countries, there is also a high prevalence in the developing world and type 2 diabetes has reached epidemic proportions in many such countries [2]. Hence diabetes is a ma- jor concern for the developing as well as more developed countries. Diabetes is the most common single cause of end-stage renal disease (ESRD) in the United States (US) and Europe [3]. About 20%–30% of patients with type 1 or 2 diabetes develop evidence of nephropa- thy. The onset of diabetic nephropathy can be significantly ameliorated by intervention, instituted early in the course of the devel- opment of this complication, at a stage when microalbuminuria is present [3]. Microalbuminuria is one of the earliest markers of renal disease in diabetes [4], and is a sensitive marker for the develop- ment of diabetic nephropathy. A strong cor- relation has been reported between renal function, the degree of albuminuria and structural renal change [5]. Small increases in the urinary albumin excretion rate in pa- tients with diabetes may be associated with subsequent development of clinical nephr- opathy. It has been reported that the pres- ence of microalbuminuria increases the risk of development of diabetic nephropathy 10-fold [4]. Also in patients with type 2 di- abetes, nephropathy is loosely associated with large vessel disease. Early intervention to reduce coronary events, control of hy- pertension and the introduction of angio- tensin-converting enzyme inhibitors can greatly improve the prognosis [6]. There- fore the determination of the albumin ex- cretion rate has found extensive use in monitoring diabetic populations for early detection of renal damage and in this con- text an active approach to screening for di- abetic nephropathy and its management is recommended [6]. Despite widespread acceptance of the value of screening for microalbuminuria, there is no consensus as to the most appro- priate urine sample for screening of mi- croalbuminuria [7]. Various procedures have been proposed, including 24-hour urine collection, overnight collection and random urine collection (preferably early morning/first voided) [8,9], the latter being most convenient for the patient. Determi- nation of the albumin/creatinine ratio in a urine sample further increases the diagnos- tic usefulness of random urine measure- ment [7]. However, again there is no agreement on the reference albumin/creati- nine ratio cut-off. Different values have been proposed; while some studies suggest a need for age and sex discriminator values [10], others propose a single cut-off of 2.5 g/mol for both sexes [11]. In our laboratory a cut-off value of 3.0 g/mol is used to dis- tinguish microalbuminuria from normoal- buminuria. Many methods have been developed for the measurement of albumin in urine, near- ly all based on immunoassay principles [12–15]. This technology, although reliable in terms of performance characteristics, includes reagents with relatively short bio- logical half-lives, which results in greater cost. It may also require specialized instru- mentation and technical skill. In this con- text, although assessment of renal function is of utmost importance in monitoring pa- tients with diabetes, the standard immu- noassays are unsuitable for use in some developing countries where resources may be limited. 11 Screening for microalbuminuria.pmd 12/20/2005, 2:35 PM359 360 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 The purpose of this study was to find a simple, cost-effective and yet reliable method to detect microalbuminuria that could be recommended to laboratories with limited resources. In this context a manual trichloroacetic acid (TCA) turbidimetric protein assay [16] with reliable perfor- mance specifications in the desired range was evaluated. Comparison of the TCA method for microproteinuria and an immu- noassay procedure for microalbuminuria (expressed in terms of albumin/creatinine ratio) was carried out Methods Materials Albumin immunoassay kits including ap- propriate calibrators were obtained from Randox Laboratories (Crumlin, United Kingdom). A creatinine kit based on the Jaffé reaction was also obtained from Ran- dox Laboratories. Creatinine powder (Sig- ma–Aldrich cat. no. C-4255, Dorset, UK), was used for in house preparation of cali- bration standards. Trichloroacetic acid was obtained from BDH (Leicestershire, UK). Saline (9 g/L) for dilution of calibrator and controls was purchased commercially from Baxter (Mallusk, UK). Instruments Albumin and creatinine were assayed on a Cobas Fara analyser (Roche Diagnostics). Both assays were carried out according to the manufacturer’s protocol. Imprecision in terms of coefficient of variation was es- timated at 1.6% and 5% respectively. Bias for both assays, validated on the basis of the Bias Index Score (BIS) in the UK Na- tional External Quality Assessment Scheme (UK NEQAS), was found to be less than 50, which is indicative of appropriate per- formance. Urinary protein concentrations as determined by the TCA method were examined by both the Cobas Fara analyser and manually. In the manual measurement a Wallac Biochrom 4060 photometer was used to record the absorbance. Samples Samples sent for assessment of microalbu- minuria from primary and secondary care diabetes clinics were used. Over 300 spec- imens, from both sexes and covering the whole spectrum of the analytical interval, were examined. Samples were stored either at room temperature or at 4 °C until analy- sis. The period of storage varied from overnight to over the weekend. Calibrators and controls Randox urinary protein calibrator (Cat. No. ST 1568) was used to calibrate the TCA method. As the concentration of this cali- brator (1000 mg/L) exceeded the concen- trations of interest for the linearity limit of the assay [16], a working calibrator (500 mg/L) was prepared by dilution in saline. Albumin calibrators were provided with the kit. Creatinine calibrator (2 mmol/L) was prepared by dissolving 0.2262 g of creati- nine powder in 1 L of water. This standard was aliquoted and stored at –20 °C. Biorad Liquicheck Urine chemistry con- trols low and high (cat. nos.397 and 398 respectively) were used to assess impreci- sion. SPS-01 human serum liquid calibrator obtained from the Supra-regional Assay Service (Protein Reference Unit) with as- signed values for albumin and a mixture of globulins was used to assess bias. The lat- ter was diluted to appropriate concentra- tions with saline. Assays The TCA assay was carried out as de- scribed by Shahangian et al. [16]. A solu- tion of trichloroacetic acid (765 mmol/L; 11 Screening for microalbuminuria.pmd 12/20/2005, 2:35 PM360 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 361 12.5 g/100 mL) was prepared in deionized water. To 1.6 mL of control, calibrator or sample, 0.4 mL of TCA reagent was added. Control and calibrators were examined in every run. For each test a complementary blank tube was set up containing the sam- ple and TCA reagent in the same propor- tions as the test. The blank underwent the same procedure as the test. All tubes were mixed and allowed to stand at room tem- perature for 35 minutes. After this time all blank tubes were centrifuged for 10 min- utes at 2000 × g. The photometer was set at zero with deionized water and each test read at 420 nm against the appropriate blank (supernatant of TCA treated urine). A modification of the manual method was set up on the Roche Cobas Fara analyser using the same reagent/sample ratios, but an in- cubation time of 15 minutes was used without affecting the assay results. The between assay imprecision was calculated through repeated measurements of control samples at different concentra- tions. It was observed that the repeatability of the test decreased with decreasing con- centrations; however the maximum coeffi- cient of variation, relating to the lowest concentration of about 20 mg/L, was esti- mated at about 9%. The systematic error (bias) of the method as evaluated using the SPS-01 reference material was estimated at 6%. The total uncertainty was therefore well within the allowable range of 46.1% (P < 0.05) [17]. All samples were analysed for protein, albumin and creatinine on the same day. It was observed that on days when the sam- ples had been kept for longer periods before analysis, there was greater discor- dance between the protein and the albumin values. Examination of the samples con- cerned showed that the difference was due to the presence of bacterial contamination, the frequency of which increased with pro- longed storage of samples, especially at room temperature prior to examination. Bacterial contamination does not interfere with the immunoassay for albumin [18], but it does interfere with the turbidimetric method used for the measurement of pro- tein [19]. Therefore on the basis of this finding all samples were examined micro- scopically for the presence of microorgan- isms prior to analysis; if positive, they were excluded from the study. It was also ob- served that rapid transfer of samples to 4 °C for storage, and analysis on the day of receipt reduced the incidence of in vitro contamination to insignificant levels. Results The results obtained from the determina- tion of urinary protein and albumin/creati- nine ratio were compared and the diagnostic effectiveness of the TCA meth- od was evaluated. An albumin/creatinine ra- tio cut-off value of equal to or greater than 3.0 g/mol was used to identify the diseased population as indicated by the presence of microalbuminuria. Table 1 shows the clini- cal sensitivity and specificity along with the positive and negative predictive values cal- culated for different cut-off values of uri- nary protein. It should be noted that the predictive value is valid only for the popula- tion under investigation. From the table, it can be seen that the value of 24 mg/L of urinary protein gives the most appropriate clinical sensitivity and specificity. Hence this cut-off is recommended as the value above which microalbuminuria is most likely to be present. Based on the above delineation of dis- eased and non-diseased states, a receiver operator characteristic (ROC) curve was constructed as shown in Figure 1. This curve shows the performance of the uri- nary protein method through the entire range of decision levels. 11 Screening for microalbuminuria.pmd 12/20/2005, 2:35 PM361 362 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 The distribution of urinary albumin re- sults in urinary protein positive and nega- tive groups is shown in Figure 2. One extreme outlier has been excluded. The fig- ure shows the non-parametric nature of the distribution of results, especially in the pos- itive population. Discussion and conclusions This study was carried out in an attempt to find a simple, cost-effective and at the same time reliable method for the assess- ment of microalbuminuria, which could be used in laboratories at the intermediate and peripheral level in developing countries with limited resources. The TCA method is based on the pre- cipitation of proteins in the urine sample followed by turbidimetric quantitation. The turbidimetric methods for measurement of urinary proteins in general do not have as good quality specifications as the immu- nochemical methods used for the examina- tion of urinary albumin. However, at a cut-off of 24 mg/L the clinical sensitivity of this modification of the TCA method, when Table 1 Results of comparison of albumin/creatinine ratio and microproteinuria Urinary protein Sensitivity Specificity (+) predictive (–) predictive concentration (%) (%) value (%) value (%) (mg/L) 18.7 86.7 85.5 60 96 20.0 86.7 86.3 61 96 20.2 86.7 87.2 63 96 22.0 86.7 88.0 64 96 22.2 86.7 88.9 66 96 23.0 86.7 89.7 68 96 23.9 86.7 90.6 69 96 24.1 83.3 90.6 69 95 24.5 80.0 91.5 70 95 The predictive value of the test is calculated according to the prevalence seen in the specific population under investigation, which was estimated at 20%. Figure 1 Receiver operating characteristic (ROC) curve; use of urinary protein to predict an albumin/creatinine ratio of 3.0 g/ mol 11 Screening for microalbuminuria.pmd 12/20/2005, 2:35 PM362 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 363 compared against the urinary albumin/crea- tinine ratio of 3.0 g/mol, shows that over 86% of the diseased population can be identified. At the same cut-off, the speci- ficity is calculated at 90% bearing in mind that this is only valid in the absence of bac- teriuria and contamination with microor- ganisms. In addition, the inexpensive and rapid nature of the test renders it affordable for regular monitoring of patients. Hence those patients falsely identified as negative are likely to be detected in further regular investigations. It should be noted that the reagent cost per test for the TCA method (~1.2 English pence per test) is approxi- mately 30 times less than that estimated for the immunoassay determination of albumin (~36 English pence per test). Furthmore, the TCA test can be performed using sim- ple analytical instrumentation. The high biological variation inherent in the albumin excretion rate calls for at least 2 consecutive examinations of positive samples. We recommend that samples are analysed on the day of collection to avoid risk of contamination by microorganisms. If possible, we recommend storage at 4 °C prior to analysis. Measurement of micro- proteinuria on a contaminated sample will result in a false positive result. If this possi- bility is suspected, an additional examina- tion should be carried on a fresh sample. The protocol recommended for the perfor- mance of the urinary protein test is shown in Figure 3. - Samples should not be collected from patient with known urinary tract infection. - Collect random urine sample. Preference should be given to the first sample voided in the morning. - No preservatives should be added to samples. - Store all samples at 4 °C until analysis. - All samples should be examined on the day of collection. - A patient with urinary protein > 24 mg/L should be asked to provide a repeat sample in 2 weeks. - In case of discordance between the 2 results a third sample should be analysed. - Drugs taken by the patient should be noted as some can lead to false positive results [20]. - Patients found to have 2 urinary protein values above 24 mg/L should be selected for further investigation and management. Figure 2 Box and whisker plot of the results of the urinary protein versus urinary albumin concentrations. The dashed line indicates the mean and the solid line the median. The box covers the central 50% of results, the whiskers extend to include the 5th and the 95th centile. Figure 3 An outline of the procedure for the measurement of urinary protein 11 Screening for microalbuminuria.pmd 12/20/2005, 2:35 PM363 364 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 References 1. Zimmet P, Alberti KGMM, Shaw J. Global and societal implications of the diabetic epidemic. Nature, 2001, 414:782–7. 2. Zimmet PZ. Diabetes epidemiology as a tool to trigger diabetes research and care. Diabetologia, 1999, 42:499–518. 3. American Diabetes Association. Diabe- tes nephropathy. Diabetes care, 2002, 25(suppl. 1):S85–9. 4. Almdal T et al. The predictive value of microalbuminuria in IDDM. A five-year follow-up study. Diabetes care, 1994, 17:120–5. 5. Mauer SM et al. Structural–functional re- lationships in diabetic nephropathy. Journal of clinical investigation, 1984, 74:1143–55. 6. Harvey JN. Diabetic nephropathy. British medical journal, 2002, 325:50–60. 7. Mogensen CE. Prevention of diabetic re- nal disease with specific reference to microalbuminuria. Lancet, 1995, 346: 1080–4. 8. Watts GF. Urinary albumin excretion in healthy adult subjects: reference values and some factors affecting their interpre- tation. Clinica chimica acta, 1988, 172: 191–8. 9. Nathan DM, Rosebaum C, Protasowicki VD. Single-void urine samples can be used to estimate quantitative microalbu- minuria. Diabetes care, 1987, 10:414–8. 10. Bakker AJ. Detection of microalbu- minuria: receiver operating characteris- tic curve analysis favors albumin-to- creatinine ratio over albumin concentra- tion. Diabetes care, 1999, 22:307–13 11. Mogenson CE et al. Randomised con- trolled trial of dual blockade of rennin- angiotensin system in patients with hypertension, microalbuminuria, and non-insulin dependant diabetes: the CALM study. British medical journal, 2000, 321:1440–4. 12. Keen H, Chlouverakis C. An immunoas- say method for urinary albumin at low concentrations. Lancet, 1963, 186:913– 4. 13. Feldt-Rasmussen B, Dinesen B, Deckert M. Enzyme immunoassay: an improved determination of urinary albumin in dia- betics with incipient nephropathy. Scan- dinavian journal of clinical and laboratory investigation, 1985:45:539– 44 14. Fielding BA, Price DA, Houlton CA. En- zyme immunoassay for urinary albumin. Clinical chemistry, 1983, 29:355–7. 15. Townsend JC. A competitive immu- noenzymometric assay for albumin in urine. Clinical chemistry, 1986, 32: 1372–4. 16. Shahangian S, Brown, PI, Ash KO. Turbi- dimetric measurement of total urinary proteins: A revised method. American The TCA method as described satisfies many of the requirements for the screening of patients’ samples at early stages of renal disease. It is simple, inexpensive and rela- tively rapid with little need for specialized skills. The reagent has a long shelf life and can be prepared in-house. In countries with limited resources we recommend the use of this method for screening of patients with diabetes at high risk of developing re- nal damage. Where circumstances allow, this test can be used to identify patients re- quiring referral for quantification of urinary albumin by an immunochemical method. 11 Screening for microalbuminuria.pmd 12/20/2005, 2:35 PM364 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 365 journal of clinical pathology, 1984, 81: 651–4. 17. Ricos C et al. Current databases on bio- logical variation: pros, cons and progress. Scandinavian journal of clini- cal and laboratory investigation, 1999, 59:491–500. 18. Rowe DJF, Dawnay A, Watts GF. Microalbuminuria in diabetes mellitus: review and recommendations for the measurement of albumin in urine. An- nals of clinical biochemistry, 1990, 27: 297–312 19. Davidson I, Henry JB, eds. Todd- Stanford clinical diagnosis and manage- ment by laboratory methods, 16th ed. Philadelphia, WB Saunders, 1979. 20. Burtis CA, Ashwood ER, eds. Tietz text- book of clinical chemistry, 3rd ed. Phila- delphia, WB Saunders, 1999. Fluorescence microscopy for disease diagnosis and environmen- tal monitoring In countries with limited resources, simple, rapid and sensitive di- agnostic techniques are key elements for the performance of medi- cal and public health laboratories at all levels. Bright-field microscopy has been the primary diagnostic technique of laborato- ries at peripheral and district levels. Fluorescence microscopy is al- most as simple to do, and most often it is more specific. In the past, the high cost of fluorescence microscopes prevented the wider application of this method. More recently, less expensive fluores- cence microscopes have been developed, and accessories are now available that convert a bright-field microscope into a fluores- cence microscope. This development places fluorescence microscopy in a favourable position as a method that can be used by laboratories to enhance their effectiveness at affordable cost. Laboratories should be more aware of the advantages of using fluo- rescence microscopy. This manual, Fluorescence microscopy for disease diagnosis and environmental monitoring provides informa- tion on the principles of fluorescence microscopy and practical ad- vice on the preparation of samples for many simple applications for diagnosing disease and monitoring environmental contamination using a fluorescence microscope. The publication puts emphasis on procedures for direct, rapid identification of microorganisms causing a disease. The practical steps of indirect immunofluores- cence microscopy for the diagnosis of noncommunicable diseases are also considered. The manual can be obtained from: World Health Organization Regional Office for the Eastern Mediterranean, Abdul Razzak Al Sanhouri Street, PO Box 7608, Nasr City, Cairo 11371, Egypt. Email: emr:dsa@ emro.who.int. 11 Screening for microalbuminuria.pmd 12/20/2005, 2:35 PM365 366 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Identification of Candida dubliniensis in a diagnostic microbiology laboratory O.M. Momani1 and A. Qaddoomi2 1Department of Microbiology, Omar Medical Laboratories, Amman, Jordan (Correspondence to O.M. Momani: momaniomar@hotmail.com). 2King Hussein Medical Centre, Amman, Jordan. Received: 09/10/02; accepted: 14/09/03 ABSTRACT Candida dubliniensis is an emerging yeast pathogen isolated mainly from immunocompromised patients. As molecular tests are currently unsuitable for use in routine diagnostic laboratories, we compared a variety of phenotypic techniques for differentiating C. albicans and C. dubliniensis. The tests included: colony colour on CHROMagar™ Candida medium; growth at 37 °C and 45 °C; ability to produce germ tubes and chlamydospores; and the Auxacolor® system. The organisms included 105 isolates previously identified as C. albicans, 10 reference strains of C. albicans, 2 reference strains of C. dubliniensis and 102 fresh clinical isolates identified as C. albicans. None of the tests alone was satisfactory but a combination of 3 tests may be suitable for presumptive identification of C. dubliniensis. Identification de Candida dubliniensis dans un laboratoire de microbiologie diagnostique RÉSUMÉ Candida dubliniensis est un nouvelle levure pathogène isolée principalement chez des patients immunodéprimés. Les tests moléculaires ne convenant pas actuellement pour être utilisés dans les labora- toires de diagnostic de routine, nous avons comparé diverses techniques phénotypiques pour différencier C. albicans et C. dubliniensis, dont : la couleur des colonies sur milieu CHROMagarTM Candida ; la culture à 37 °C et 45 °C ; la capacité de produire des tubes de germes et des chlamydospores ; et le système Auxacolor®. Les micro-organismes comprenaient 105 isolats identifiés auparavant comme C. albicans, 10 souches de référence de C. albicans, 2 souches de référence de C. dubliniensis et 102 isolats cliniques frais identifiés comme C. albicans. Aucun des tests seul n’était satisfaisant mais une association de trois tests peut convenir pour une identification présomptive de C. dubliniensis. 12 Identification of Candida.pmd 12/20/2005, 2:35 PM366 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 367 Introduction Over the past 10 years there have been fre- quent reports describing “atypical” isolates of Candida albicans [1–5]. In 1995, Sulli- van et al. suggested that these isolates com- prised a new species and named it C. dubliniensis, after Dublin, the capital city of Ireland, where the new species was first identified [2,6]. The organism has been recovered from the oral cavity in healthy people, HIV- infected patients and AIDS patients [7,8]. The greatest concern about C. dubliniensis is the potential for development of antifun- gal drug resistance, especially in HIV- infected patients [9,10]. Candida dubliniensis shares many phe- notypic similarities with C. albicans, result- ing in significant problems in differentiation between the 2 species. Currently, there is a real need for a rapid and simple test for use in routine clinical laboratories to distinguish isolates of C. dubliniensis and C. albicans. Definitive identification of C. dubliniensis currently relies on molecular methods; however, these have high running costs and require well-trained laboratory person- nel, who might not be available in routine diagnostic laboratories. Several methods have been proposed for differentiating between the strains. The use of CHROMagar™ chromogenic culture medium has been proposed as a means of recognizing colonies of C. dubliniensis [11,12]. Several researchers have conduct- ed carbohydrate assimilation studies [2,13, 14]; some kits contain individual tests that can be utilized for C. dubliniensis. Pinjon et al. suggested culture at 45 °C, concluding that C. dubliniensis is unable to grow at such high temperatures while C. albicans grows well [5]. A recent study however showed that some C. albicans strains are unable to grow at 45 °C [11]. The aim of this study was to assess the usefulness of different phenotypic tech- niques for differentiating between isolates of the 2 species, C. albicans and C. dublin- iensis. Methods Organisms used Two C. dubliniensis reference strains were used in this study. One strain (coded NCPF3108) was originally identified as C. stellatoidea and deposited in the British collection of pathogenic fungi; it has since been re-identified as C. dubliniensis [6]. The other strain (the Bristol strain) was supplied by Dr Colin Campbell from the Bristol Mycology Reference Laboratory at the Public Health Laboratory Service, Bris- tol, United Kingdom. The 2 strains were identified by molecular methods. Ten C. albicans reference strains were used: 324/94RA, WK1, 122/94Rgl, 684/93, 455/94rgh, 455/94sm, ATCC 3516, Y01.544, LSHTM3153 and 91L. Stock Candida spp. strains supplied by the mycology laboratory of the Department of Microbiolgy, University of Wales College of Medicine were used (n = 105). These had previously been identified as C. albi- cans using the germ tube test, chlamy- dospore formation and Auxacolor® yeast identification system. A total of 158 clinical specimens were studied, comprising 91 urine samples, 66 genital samples (high vaginal, penile and vulvo-vaginal swabs) and 1 blood culture sample, resulting in the identification of 102 isolates of C. albicans by the germ tube and Auxacolor® tests. Culture media CHROMagar™ Candida medium (CHRO- Magar Microbiology, Paris, France) was 12 Identification of Candida.pmd 12/20/2005, 2:35 PM367 368 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 prepared according to the manufacturer’s instructions. Sabouraud agar + chloramphenicol and corn meal agar were supplied by the De- partment of Microbiolgy, University of Wales College of Medicine. Sabouraud agar + chloramphenicol was supplied as ready- to-use agar plates in packs of 15 plates. Corn meal agar was supplied ready made as 100 mL prepared solid medium in a flask. The agar was melted and distributed into 5 sterile Petri dishes. The germ tube test for the production of germ tubes was done using horse serum (TCS Biologicals Ltd, Buckingham, UK). Organisms were cultured on CHROM- agar™ Candida medium and Sabouraud agar + chloramphenicol for 48 hrs at 37 °C and 45 °C to determine the ability to grow at both temperatures and to study colony appearance and colour. To study the for- mation of chlamydospores, cultures on corn meal agar were incubated at room temperature. Identification was carried out using the Auxacolor® system (Sanofi Diag- nostics Pasteur, Marnes La Coquette, France). The Auxacolor® test procedures were done in accordance with the manu- facturer’s instructions. Results Six of the 105 stock strains of C. albicans failed to grow on both the Sabouraud and the CHROMagar™ media and 1 showed mould contamination. These were exclud- ed from the study. The remaining 98 grew equally well on both media. On CHROMagar™ Candida medium, all 10 references strains of C. albicans grew well, forming medium-sized, 3–5 mm smooth, entire colonies. The colony colour ranged from light green to green, most of- ten with lighter edges, sometimes with the intense colour at the edge of the colony and sometimes in the centre. The 2 C. dublin- iensis reference strains formed colonies similar to those of C. albicans; 1 was light green and the other dark green. The conventional method using the germ tube test and chlamydospore forma- tion identified 9 of the 10 C. albicans refer- ence strains, and misidentified the C. dubliniensis strains as C. albicans; these were also misidentified by the Auxacolor® system. The Auxacolor® system correctly identified the 1 strain of C. albicans that was germ-tube negative. Of the 219 cultures investigated, all re- maining 212 which were viable grew well at 37 °C on both CHROMagar™ and Sab- ouraud agar + chloramphenicol. Growth was consistent at 48 hours. Table 1 shows a summary of colony colours and growth at 45 °C of the strains investigated. None of the C. albicans stock culture strains produced a deep green co- lour on CHROMagar™ but 15 clinical iso- lates as well as the Bristol strain of C. dubliniensis produced colonies of a deep bluish-green colour. The germ tube test was positive for these strains, as was chlamydospore formation. The deep green clinical isolates were identified by the Auxacolor® system as C. albicans with biocodes of 7145207 or 7143207. The Bristol reference strain had a biocode of 7141207 and the NCPF3108 reference strain had a biocode of 7143207. All reference strains of C. albicans grew at 45 °C. The 2 C. dubliniensis strains did not grow at all at 45 °C on either CHROMagar™ or Sabouraud agar + chloramphenicol. Among the stock culture strains, only 2 failed to grow at 45 °C. Among the clinical isolates, 1 from a urine sample and 1 from a genital sample did not grow at this temperature. Specimens that did not grow at 45 °C and/or showed dark green colonies on 12 Identification of Candida.pmd 12/20/2005, 2:35 PM368 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 369 CHROMagar™ and showed a biocode of 7143207 or 714107 were re-identified as C. dubliniensis. Discussion Over the past 2 decades, there has been a rapid growth in clinical microbiology tech- nology. In the past, test results were avail- able only after several days owing to the labour-intensive methods. With the emer- gence of new pathogens, especially the drug-resistant Candida spp., it has become necessary for laboratories to seek more ef- ficient and cost-effective methods of iden- tification. Therefore, it is essential that the available diagnostic tools be constantly modified to keep abreast of the ever- changing spectrum of pathogens. A new emerging yeast pathogen, C. dubliniensis, was the focus of this study. It is closely related to C. albicans. Microbio- logical information about C. dubliniensis shows similarities with C. albicans; using the currently available conventional meth- ods it is difficult to discriminate between them. The colony appearance on CHROM- agar™ Candida medium makes it extremely useful for rapid presumptive identification of the most common Candida spp. As re- ported originally by Odds and Bernaert, the green colour is unique to C. albicans [15]. Casal described the deep green colour as characteristic of C. zeylanoides, a species not mentioned in the study of Odds and Bernaert [16]. In our study, C. dubliniensis could show the same colour as C. albicans. However, the colour was less intense in older cultures and in freshly cultured clini- cal specimens it appeared dark green. Sulli- van and Coleman described this pheno- menon as a way of differentiating between the 2 species on primary culture. In our study, 2 of the stock cultures and 1 C. dubliniensis reference strain appeared light green and failed to grow at 45 °C while 2 of the fresh clinical isolates and the Bristol ref- erence strain (all dark green) failed to grow at 45 °C. Thirteen C. albicans strains ap- peared deep green on primary isolation. Table 1 Summary of colony colour and growth at 45 °C of Candida albicans and Candida dubliniensis strains Strain Total Colour on CHROMagara Growth at 45 °C tested Dark Green Yes No green or light green C. albicans reference strains 10 0 10 10 0 C. dubliniensis reference strains 2 1 1 0 2 Stock C. albicans 98b 0 98 96 2 Clinical isolates (C. albicans) 102 15 87 100 2 aCHROMagar™ Candida medium. bTotal viable. 12 Identification of Candida.pmd 12/20/2005, 2:35 PM369 370 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 These findings suggest that CHROMagar™ medium alone is not suitable for discrimi- nating between C. albicans and C. dublin- iensis on primary isolation. Pinjon et al. reported that C. dublinien- sis could be discriminated from C. albicans by the ability of the latter to grow at 45 °C [5]. This test is unsatisfactory because some C. albicans strains are unable to grow at this elevated temperature and this cannot, therefore, be used as a discrimina- tory test [11]. The results obtained in our study are consistent with Pinjon’s criterion, as the 2 reference strains of C. dubliniensis failed to grow at 45 °C. Two stock culture strains and 2 clinical isolates failed to grow at this temperature, although all 10 C. albi- cans reference strains did grow. Those strains are considered suspect for C. dubliniensis. Gales et al. studied the use of xylose and α-methyl-glucose as determined with the API 20C AUX and Vitek YBC systems for the identification of C. dubliniensis [17], The Auxacolor® system used in this study contained xylose. The inability of C. dubliniensis to utilize xylose is reflected by specific biocodes. Conclusion None of the tests alone—CHROMagar™ Candida media, growth at 45 °C or Auxa- color®—was a satisfactory discriminatory test. A combination of the 3 tests used in this study may be a good tool for presump- tive identification of C. dubliniensis. The 4 strains (2 stock culture strains and 2 clini- cal isolates) that appeared to be presump- tive of C. dubliniensis need to be confirmed by further testing. Recognition of C. dubliniensis provides valuable information regarding its epidemiology to help establish its clinical significance. Acknowledgements This work was done at the Department of Microbiology, University of Wales College of Medicine, United Kingdom, as a part of an MSc degree project. This study was sponsored by the Royal Jordanian Medical Services and supervised by Dr Rosemary Barnes. Special thanks go to Professor Sheikh Mahjoub for revising this manu- script. References 1. Sullivan D, Coleman D. Candida dubliniensis: an emerging pathogen. Current topics in medical mycology, 1997, 8(1–2):15–25. 2. Sullivan D, Coleman D. Candida dubliniensis: characteristics and identifi- cation. Journal of clinical microbiology, 1998, 36(2):329–34. 3. Sullivan D et al. Widespread geographic distribution of oral Candida dubliniensis strains in human immunodeficiency vi- rus-infected individuals. Journal of clini- cal microbiology, 1997, 35(4):960–4. 4. Schoofs A et al. Use of specialised isola- tion media for recognition and identifica- tion of Candida dubliniensis isolates from HIV infected patients. European journal of clinical microbiology and in- fectious diseases, 1997, 16(4):296–300. 5. Pinjon E et al. Simple, inexpensive, reli- able method for differentiation of Can- dida dubliniensis from Candida albicans. Journal of clinical microbiol- ogy, 1998, 36(7):2093–5. 6. Sullivan DJ et al. Candida dubliniensis: phenotypic and molecular character- isation of a novel species associated with oral candidosis in HIV-infected indi- viduals. Microbiology, 1995, 141(7): 1507–21. 12 Identification of Candida.pmd 12/20/2005, 2:35 PM370 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 371 7. Jabra-Rizk MA et al. Identification of Candida dubliniensis in a prospective study of patients in the United States. Journal of clinical microbiology, 1999, 37(2):321–6. 8. Polacheck I et al. Recovery of Candida dubliniensis from non-human immuno- deficiency virus-infected patients in Is- rael. Journal of clinical microbiology, 2000, 38(1):170–4. 9. Rinaldi MG, Patterson TF. Detection of Candida dubliniensis in oropharyngeal samples from human immunodeficiency virus-infected patients in North America by primary CHROMAgar Candida screening and susceptibility testing of isolates. Journal of clinical microbiology, 1998, 36:3007–12. 10. Sandven P. Detection of fluconazole- resistant Candida strains by a disc diffu- sion screening test. Journal of clinical microbiology, 1999, 37(12):3856–9. 11. Kirkpatrick WR et al. Detection of Can- dida dubliniensis in oropharyngeal samples from human immunodeficiency virus-infected patients in North America by primary CHROMAgar Candida screening and susceptibility testing of isolates. Journal of clinical microbiology, 1998, 36:3007–12. 12. Odds FC, Van Nuffel L, Dams G. Preva- lence of Candida dubliniensis isolates in a yeast stock collection. Journal of clini- cal microbiology, 1998, 36:2869–73. 13. Salkin IF et al. Distinctive carbohydrate assimilation profiles used to identify the first clinical isolates of Candida dubliniensis recovered in the United States. Journal of clinical microbiology, 1998, 36(5):1467. 14. Hannula J et al. Phenotypic and geno- typic characterisation of oral yeasts from Finland and the United States. Oral mi- crobiology and immunology, 1997, 12(6):358–65. 15. Odds FC, Bernaerts R. CHROMagar Candida, a new differential isolation me- dium for presumptive identification of clinically important Candida species. Journal of clinical microbiology, 1994, 32(8):1923–9. 16. Casal M et al. Appearance of colonies of Prototheca on CHROMagar Candida medium. Mycopathologia, 1997, 137(2): 79–82. 17. Gales AC et al. Identification of Candida dubliniensis based on temperature and utilization of xylose and a-methyl-D- glucoside as determined with the API 20C AUX and Vitek YBC systems. Jour- nal of clinical microbiology, 1999, 37(12): 3804–8. 12 Identification of Candida.pmd 12/20/2005, 2:35 PM371 372 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Seroprevalence of hepatitis C and risk factors in haemodialysis patients in Guilan, Islamic Republic of Iran Z.M. Amiri,1 A.J. Shakib1 and M. Toorchi1 1Department of Social Medicine, Guilan University of Medical Sciences, Guilan, Islamic Republic of Iran (Correspondence to Z.M. Amiri: mohtashamaz@yahoo.com). Received: 23/09/03; accepted: 17/05/04 ABSTRACT We assessed the prevalence of hepatitis C virus (HCV) infection and associated risk factors for all 298 haemodialysis patients in 7 dialysis units in Guilan province. Serum samples were screened for anti- HCV antibodies using a second generation enzyme-linked immunosorbent assay. Positive samples were confirmed by immunoblot assay. Overall prevalence was 24.8% (range: 9%–40%; 95% CI: 19.9–29.7): 80 patients tested positive and 74 were confirmed positive by immunoblot assay. Length of time on dialysis and history of rejected kidney transplant were statistically significantly associated with HCV infection. Age, sex and previous blood transfusion were not associated. Nosocomial transmission may play a role in the spread of HCV in haemodialysis units. A separate dialysis system should be used for seropositive HCV patients. Séroprévalence de l’hépatite C et facteurs de risque chez les patients hémodialysés à Guilan (République islamique d’Iran) RÉSUMÉ Nous avons évalué la prévalence de l’infection par le virus de l’hépatite C (VHC) et les facteurs de risque associés pour l’ensemble des 298 patients hémodialysés dans sept unités de dialyse de la province de Guilan. Des échantillons de sérum ont fait l’objet d’une recherche d’anticorps anti-VHC utilisant un test immuno-enzymatique de deuxième génération. Les échantillons positifs ont été confirmés par test d’immunotransfert. La prévalence globale s’élevait à 24,8 % (fourchette : 9 % - 40 % ; IC 95 % : 19,9-29,7). Quatre-vingt participants ont eu un test positif et 74 ont été confirmés positifs par test d’immunotransfert. Il y avait une association statistiquement significative entre la durée sous dialyse et les antécédents de rejet de greffe de rein, et l’infection par le VHC. Il n’y avait pas d’association avec l’âge, le sexe et les transfusions sanguines précédentes. La transmission nosocomiale peut jouer un rôle dans la propagation du VHC dans les unités de dialyse. Un système de dialyse séparé devrait être utilisé pour les patients séropositifs au VHC. 14 Seroprevalence of hepatitis.pmd 12/20/2005, 2:35 PM372 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 373 Introduction Hepatitis C virus (HCV) accounts for over 85% of transfusion-associated hepatitis cases. The chance of the infection pro- gressing to a chronic state is more than 50%, leading to cirrhosis or hepatocellular carcinoma in 25% [1,2]. Although there are only a few regions in the world for which data on the prevalence of this virus are available, it is estimated that more 170 million persons are infected worldwide. In industrialized nations, HCV prevalence is typically 1%–2% in the gen- eral population and < 0.5% in blood donors [3]. Haemodialysis patients are more vulner- able to HCV infection than others because of history of blood transfusion, frequent injections, partial immunosuppression and history of kidney transplant [4]. The dura- tion of haemodialysis treatment and noso- comial HCV transmission have also been suggested as contributing factors [5–9]. The prevalence of HCV antibodies (anti- HCV) in dialysis patients has been reported to range from 20.0% to 81.6% in previous studies [10–18]. In the Islamic Republic of Iran, studies have shown that haemodialy- sis patients have high anti-HCV prevalence, ranging from 24.7% to 40.0% [19–22]. The aim of this study was to assess the prevalence of HCV infection in a haemodi- alysis population in Guilan, in northern Is- lamic Republic of Iran, and to determine possible risk factors for HCV infection in this population. Methods Guilan province is in the north of the Islam- ic Republic of Iran and has a population of 2 200 000. It has 14 cities, half of which have dialysis units. Our study was carried out in all 7 dialysis units in Guilan between March and April 2001. All chronic haemo- dialysis patients (n = 298) were inter- viewed to collect data on risk factors for HCV infection. No patients refused to par- ticipate. A form was used to collect data on age, sex, length of time on haemodialysis, previous blood transfusion, intravenous drug abuse and history of kidney trans- plant. The interviews were conducted by the physicians on our team. Blood samples were collected from all patients and sera were screened by stan- dard techniques using a commercial sec- ond generation enzyme-linked immu- nosorbent assay kit (CAPTIA, Trinity Bio- tech, Bray, Ireland) for the presence of anti-HCV antibodies. Positive samples were retested for confirmation using a commer- cial immunoblot assay kit (INNOGENET- ICS® Ghent, Belgium). All techniques were carried out according to the manufactur- er’s instructions. Prevalence and 95% confidence inter- vals (95% CI) were calculated. Chi- squared test or chi-squared for trend test was performed to evaluate risk factors as- sociated with HCV infection. Statistical sig- nificance was assessed at the 0.05 probability level in all analyses. Statistical analysis was performed using Epi-Info, version 6.0. Results The study population ranged in age from 13 to 85 years (mean 52.2 years); 156 were male (52.3%) and 142 were female (47.7%). Using enzyme-linked immunosorbent assay, 80 of the 298 haemodialysis patients were seropositive, and 74 (92.5%) were subsequently confirmed positive by immu- noblot assay resulting in an anti-HCV prev- alence of 24.8% (95% CI: 19.9–29.7). The prevalence of HCV ranged from 9.0% to 14 Seroprevalence of hepatitis.pmd 12/20/2005, 2:35 PM373 374 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 40.0% among the 7 haemodialysis units (Table 1). None of the seropositive patients had a history of blood transfusion before haemodialysis or intravenous drug abuse. Thirty-five male and 39 female participants were seropositive, but the relationship be- tween sex and seropositivity was not sig- nificant (P > 0.05). The relationships between length of time on haemodialysis and history of rejected kidney transplant and seropositivity were statistically signifi- cant (Tables 2 and 3). Of 11 patients who had previous kidney transplants, 7 were HCV seropositive (P = 0.006). Of the 3 patients who had tattoo- ing, all were HCV seropositive. Only 1 of the 298 participants was seropositive be- fore the onset of haemodialysis. Discussion Our study showed that the prevalence of HCV infection in haemodialysis patients from Guilan province is higher than that of blood donors (0.5%) from the same region [23]. This demonstrates that HCV infection is a significant problem in this population. Previous studies have indicated that the du- ration of dialysis treatment is clearly corre- lated with HCV seropositivity [4,7,8, 19–22]. This is in agreement with the re- sults of this study and is indicative of noso- comial transmission of HCV. The relationship between HCV infection and renal rejection is unknown, but many studies shown a relationship [14,17,20], and 7 of the 11 patients in our study who had a history of rejected kidney transplant were HCV seropositive. Reports from the early 1990s (shortly after the introduction of diagnostic testing for HCV) suggested that the course of hepatitis C was unaffect- Table 1 Prevalence of hepatitis C virus (HCV) infection in seven haemodialysis units in Guilan, 2001 Unit No. of HCV positive 95% CI patients No. % A 46 15 32.6 19.1–46.1 B 40 6 15.0 4.1–25.9 C 17 3 17.6 8.4–26.8 D 136 34 25.0 17.7–32.2 E 12 4 33.3 6.7–59.9 F 22 2 9.1 – G 25 10 40.0 20.8–59.2 Total 298 74 24.8 19.9–29.7 CI = confidence interval. Table 2 Prevalence of hepatitis C virus (HCV) infection according to length of time on haemodialysis in haemodialysis patients in Guilan, 2001 Time on No. HCV positive haemodialysis patients (months) No. % < 12 105 2 1.9 12–23 67 10 14.9 24–35 25 7 28.0 36–47 23 9 39.1 ≥ 48 78 46 59.0 χ2 test, P < 0.0001. Table 3 Association between history of rejected kidney transplant and hepatitis C virus (HCV) seropositivity among haemodialysis patients in Guilan, 2001 History of HCV antibody rejected kidney Positive Negative transplant No. No. Yes 7 4 No 67 220 χ² test, P = 0.006. 14 Seroprevalence of hepatitis.pmd 12/20/2005, 2:35 PM374 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 375 ed by renal transplant procedures. More re- cent data, however, indicate that renal transplant recipients with HCV infection have a diminished survival rate that is relat- ed to progression of liver disease. During the recovery period after solid organ graft- ing in a recipient infected with HCV, the vi- ral mass in the liver and circulation rises 10-fold under the influence of corticoster- oids [24,25]. These data emphasize the need for stricter adherence to infection control mea- sures in dialysis units. Measures which should be considered include prevention of patient-to-patient contamination, the non- reuse of dialysers, and separate haemodial- ysis systems for HCV seropositive patients. The data also reinforce the importance of serological screening at the onset of dialy- sis treatment and at regular intervals there- after to identify all HCV-infected patients. Acknowledgements We thank Dr Ali Monfared and the staff of the haemodialysis units in Guilan province for their cooperation. References 1. Dienstage JL, Isselbacher KJ. Acute viral hepatitis. In: Fauci AS et al., eds. Harrison’s principles of internal medi- cine, 14th ed. New York, McGraw–Hill, 1999:1686–90. 2. Fallon MB et al. Acute and chronic hepa- titis. In: Andreoli TE, Bennett JC, Carpen- ter CCJ, eds. Cecil essentials of internal medicine, 5th ed. Philadelphia, WB Saunders, 2000:379–83. 3. Thomas DL, Lemon SM. Hepatitis C. In: Mandell GL, Bennett JE, Dolin R, eds. Mandell’s principles and practice of in- fectious diseases, 5th ed. New York, Churchill Livingstone, 2000:1748. 4. Olmer M et al. Transmission of the hepa- titis C virus in a hemodialysis unit: evi- dence for nosocomial infection. Clinical nephrology, 1997, 47(4):263–70. 5. Masuko K et al. Hepatitis C virus antibod- ies, viral RNA and genotypes in sera from patients on maintenance hemodi- alysis. Journal of viral hepatitis, 1994, 1(1):65–71. 6. Sampietro M et al. High prevalence of a rare hepatitis C virus in patients treated in the same hemodialysis unit: evidence for nosocomial transmission of HCV. Kid- ney international, 1995, 47(3):911–7. 7. Sandhu J et al. Hepatitis C prevalence and risk factors in the northern Alberta dialysis population. American journal of epidemiology, 1999, 150(1):58–66. 8. Scotto G et al. Hepatitis C virus infection in four haemodialysis units of southern Italy: epidemiological report. European journal of epidemiology, 1999, 15(3): 217–23. 9. Grethe S et al. Molecular epidemiology of an outbreak of HCV in a hemodialysis unit: direct sequencing of HCV-HVR1 as an appropriate tool for phylogenetic analysis. Journal of medical virology, 2000, 60(2):152–8. 10. Cialdea L, Sagliocca L, Mele A. Epidemi- ologia e prevenzione dell’infe-zione da HCV nei centri di dialisi [Epidemiology and prevention of HCV infection in dialy- sis centers]. Annali dell’Istituto superiore di sanità, 1995, 31(3):313–5. 11. Pru CE et al. Transmission through dialy- sis machine. ASAIO journal, 1994, 40(3): 889–91. 12. Dos-Santos JP et al. Impact of dialysis room and reuse strategies on the inci- dence of HCV infection in haemodialysis units. Nephrology, dialysis, transplanta- tion, 1996, 11(10):2017–22. 14 Seroprevalence of hepatitis.pmd 12/20/2005, 2:35 PM375 376 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 13. Hou CH et al. Intrafamilial transmission of HCV in hemodialysis patients. Journal of medical virology, 1995, 45(4):381–5. 14. Periera BJ et al. The impact of pretransplantation HCV infection on the outcome of renal transplant. Transplan- tation, 1995, 60(8):799–805. 15. Furusyo N et al. Maintenance hemodi- alysis decreases serum hepatitis C virus (HCV) RNA levels in hemodialysis pa- tients with chronic HCV infection. Ameri- can journal of gastroenterology, 2000, 95(2):490–6. 16. Hayashi J et al. Prevalence of antibody to hepatitis C virus in hemodialysis pa- tients. American journal of epidemiology, 1991, 134(6):651–7. 17. Hung K-Y et al. Hepatitis B and C in he- modialysis patients. Dialysis & trans- plantation, 1995, 24(3):135–8. 18. El Gohary A et al. High prevalence of hepatitis C virus among urban and rural population groups in Egypt. Acta tropica, 1995, 59(2):155–61. 19. Nowbakht-Haghighi A, Nowroozi A. [Hepatitis C prevalence and risk factors in hemodialysis patients in Tehran]. Pejouhandeh quarterly research jour- nal, 1997, 2(2):13–17 [in Farsi]. 20. Saffar MJ, Salehi A. [Prevalence of hepa- titis C virus infection in hemodialysis pa- tients in Mazandaran province (Iran)]. Mazandaran University of Medical Sci- ences journal, 1998, 21:39–49 [in Farsi]. 21. Ebrahim-Poor S, Hashemi SN, Koosha- var H. [Prevalence of hepatitis C virus in- fection in diabetic and hemodialysis patients in Tabriz (Iran)]. Tabriz Univer- sity of Medical Sciences journal, 1999, 33(44):3–8 [in Farsi]. 22. Mecvandi M et al. [Prevalence of anti- HCV antibody in hemodialysis patients in Ahwaz (Iran)]. Ahwaz University of Medical Sciences journal, 2000, 29:1–5 [in Farsi]. 23. Ansar MM, Kooloobandi A. Prevalence of hepatitis C virus infection in thalas- semia and hemodialysis patients in north Iran – Rasht. Journal of viral hepa- titis, 2002, 9(5):390–2. 24. Baid S et al. Renal disease associated with hepatitis C infection after kidney and liver transplantation. Transplanta- tion, 2000, 70(2):255–61. 25. Legendre C et al. Harmful long-term im- pact of hepatitis C virus infection in kid- ney transplant recipients. Transplanta- tion, 1998, 65(5):667–70. 14 Seroprevalence of hepatitis.pmd 12/20/2005, 2:35 PM376 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 377 Prevalence of intestinal parasitic infections in the Islamic Republic of Iran A.A. Sayyari,1 F. Imanzadeh,1 S.A. Bagheri Yazdi,2 H. Karami 1 and M. Yaghoobi 3 1Paediatric Digestive Disease Research Centre, Mofid Children’s Hospital, Tehran, Islamic Republic of Iran (Correspondence to A.A. Sayyari: sayyari@sbmu.ac.ir). 2Department of Disease Control, Ministry of Health and Medical Education, Tehran, Islamic Republic of Iran. 3Department of Gastroenterology, Henri Mondor Hospital, Paris University XII, Creteil, France. Received: 06/05/03; accepted: 15/03/04 ABSTRACT A national survey of the prevalence of intestinal parasitic infections in the Islamic Republic of Iran was made on a random sample of families covered by local health centres affiliated to the medical universities. Out of 53 995 people aged 2+ years, from 12 495 families (0.1% of all families in 1999), 45 128 stool samples were analysed by formalin–ether precipitation. Intestinal parasitic infections were found in 19.3% of the study population (19.7% male, 19.1% female). Giardia lamblia (10.9%), Ascaris lumbricoides (1.5%), Entamoeba histolytica (1.0%) and Enterobius vermicularis (0.5%) were the most common infections. The infection rate was highest in the 2–14 years age group (25.5%) and in rural residents (23.7%). Prévalence des parasitoses intestinales en République islamique d’Iran RÉSUMÉ Une enquête nationale sur la prévalence des parasitoses intestinales en République islamique d’Iran a été réalisée dans un échantillon aléatoire de familles couvertes par des centres de santé locaux affiliés aux universités médicales. Sur les 53 995 sujets âgés de 2 ans et plus, dans 12 495 familles (0,1 % de toutes les familles en 1999), 45 128 échantillons de selles ont été analysés par précipitation (méthode formol-éther). Une parasitose intestinale a été trouvée dans 19,3 % de la population de l’étude (19,7 % de garçons, 19,1 % de filles). Giardia lamblia (10,9 %), Ascaris lumbricoides (1,5 %), Entamoeba histolytica (1,0 %) et Enterobius vermicularis (0,5 %) étaient les parasites les plus courants. Le taux d’infection était le plus élevé dans le groupe d’âge des 2-14 ans (25,5 %) et chez les ruraux (23,7 %). 15 Prevalence of intestinal.pmd 12/20/2005, 2:35 PM377 378 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Introduction Intestinal parasitic infection is a major health problem in many developing coun- tries. By increasing standards of health and controlling the carriers or intermediate hosts, most industrialized countries have successfully decreased the rates of infesta- tion. In developing countries, however, geographic and socioeconomic factors as well as unpredictable factors such as natu- ral disasters contribute to the problem. These countries are mainly located in warm or hot and relatively humid areas that, com- bined with poverty, malnutrition, high pop- ulation density, unavailability of potable water and low health status, provide opti- mum conditions for the growth and trans- mission of intestinal parasites. Insufficient research into infectious and parasitic dis- eases, lack of attention in developing coun- tries to the problem and lack of follow-up treatment are also barriers to decreasing the rates of parasitic infestation. The prevalence of infections varies in different parts of the world. The preva- lence of Entamoeba histolytica, for exam- ple, ranges from 5% to 81% and is estimated to involve around 480 million people worldwide. Giardia lamblia is the most common intestinal parasite in the United States [1]. Of the 3% to 7% of the population with G. lamblia in Australia, 1.6% were asymptomatic [2]. In a study in China, Enterobius vermicularis (47.0%), Trichuris trichiura (18.8%) and Taenia saginata (17.2%) were the most frequent causes of intestinal parasitic infections [3]. Estimates of the global prevalence of intes- tinal nematode infections transmitted through soil are: 1000 million cases of As- caris lumbricoides, 900 million of hook- worms (Ancylostoma duodenale and Necator americanus) and 500 million of Tr. trichiura [4]. Studies have shown that the prevalence of intestinal parasitic infection is higher in younger people, especially children [5]. For example, an extensive survey in Malaysia reported the overall prevalence of intestinal parasitic infections as 39.6% with as many as 89.0% in children between the ages of 2 and 12 years [6]. Data from Burma showed that 1185 of 2057 patients admitted to the surgical wards of a children’s hospital with acute abdominal problems during 1981–83 were suffering from ascariasis [7]. Geographical conditions and poor nutri- tional and socioeconomic status contribute to making the Islamic Republic of Iran a favourable area for parasitic infections. A review of 300 cases of intestinal parasitic infection showed that A. lumbricoides was the most common nematode and G. lam- blia and E. histolytica the most common unicellular microorganisms causing intesti- nal parasitic infections [8]. The prevalence of intestinal parasitic infections has been found to vary in different parts of the coun- try, with 47.2% in Kerman [8], 22.4% in Shahrekord [9], 21.9% in Mazandaran [10], 65.5% in Amol [11] and 32.2% in Tabas [12]. This study of a random sample of the Iranian population aimed to determine the demographic factors favouring infestation. The results of this study could then be used by health authorities to improve environ- mental health throughout the community, thereby preventing the spread of intestinal parasitic infections in the Islamic Republic of Iran. Methods This study was carried out between May 1999 and February 2000. The population sample was chosen by randomized cluster sampling from the normal population cov- 15 Prevalence of intestinal.pmd 12/20/2005, 2:35 PM378 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 379 ered by all health centres affiliated to the medical universities throughout the Islamic Republic of Iran. According to the 1996 census, the total number of families in the country was 12 359 298 (7 948 928 urban and 4 410 370 rural dwellers). In this study 1097 clusters from urban areas (8776 fam- ilies) and 590 from rural areas (4720 fami- lies) were selected, such that each cluster comprised 8 families and the sample con- tained around 0.1% of the total number of families in the country in 1999. Data about age, sex, site of residence (urban/rural) for each person and the state of health of the family were obtained from a questionnaire interview. The interviews were begun simultaneously in 3 centres from each province. Families were inter- viewed at their homes by a trained team consisting of a male and female doctor, a laboratory technician, an interviewer (a medical resident) and a driver. After obtaining personal information, a small container for a stool sample was giv- en to each family member, which was re- turned the following day. After sending the specimens to the laboratory, the stools were tested for parasites by the formalin– ether precipitation method. The presence of Enterobius vermicularis, A. lumbri- coides, G. lamblia, E. histolytica, Ancylos- toma duodenale, Necator americanus, Ta. saginata, Tr. trichiura, Trichostrongylus colubriformis and other types of parasites in the stools were noted. Stool samples were obtained for all family members aged 2 years or over. The data were compiled by a team of 2 general physicians, an interviewer and a laboratory technician. The executive groups in each province were supervised by epidemiologists from the Faculty of Health and Nutrition Research Institute of Tehran University. The data were analysed using Epi-Info and SPSS software. Results Among the 53 995 subjects studied, 45 128 stool specimens (83.6%) were collected. Of these, 20 663 (45.8%) specimens were from males and 24 465 specimens from fe- males. People living in urban areas provided 27 513 specimens and in rural areas 17 615. Table 1 shows that, overall, 19.3% of the subjects had intestinal parasitic infec- tions (19.7% of males, 19.1% of females). G. lamblia (10.9%), A. lumbricoides (1.5%), E. histolytica (1.0%) and Entero- bius vermicularis (0.5%) were the most common infections. Only G. lamblia was more frequent in males than females, while the prevalence of other parasites was high- er in females than males (with the excep- tion of Trichostrongylus colubriformis, which had a similar frequency in both sex- es). The frequency of various intestinal par- asitic infections by age is shown in Table 2. The prevalence of intestinal parasitic infec- tions was higher in the younger than the older age groups, with a quarter of all in- fections (25.5%) in those aged 2–14 years. Table 3 shows the frequency of intesti- nal parasitic infections by area of resi- dence. Intestinal parasitic infections overall were more common in rural (23.7%) than urban (16.2%) dwellers. Discussion This study shows that intestinal parasitic infections are a major public health problem in the Islamic Republic of Iran, affecting one-fifth of the population. The findings of studies performed in other countries such as Brazil, China, Egypt and Pakistan are comparable to the results of this study [13,14]. 15 Prevalence of intestinal.pmd 12/20/2005, 2:35 PM379 380 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Table 1 Prevalence of intestinal parasitic infections in the Islamic Republic of Iran by sex Parasite Male Female Total (n = 20 663) (n = 24 465) (n = 45 128) No. % No. % No. % Giardia lamblia 2387 11.6 2516 10.3 4903 10.9 Ascaris lumbricoides 289 1.4 372 1.5 661 1.5 Entamoeba histolytica 175 0.8 264 1.1 439 1.0 Enterobius vermicularis 102 0.5 109 0.4 211 0.5 Taenia saginata 30 0.1 38 0.2 68 0.2 Trichostrongylus colubriformis 42 0.2 50 0.2 92 0.2 Trichuris trichiura 15 0.1 23 0.1 38 0.1 Ancylostoma duodenale 5 < 0.1 7 < 0.1 12 < 0.1 Other 1023 5.0 1284 5.2 2307 5.1 Total 4068 19.7 4663 19.1 8731 19.3 n = number of participants. Table 2 Prevalence of intestinal parasitic infections in the Islamic Republic of Iran by age Parasite 2–14 years 15–39 years 40–69 years ≥ 70 years (n = 15 613) (n = 19 240) (n = 8758) (n = 1517) No. % No. % No. % No. % Giardia lamblia 2495 16.0 1695 8.8 637 7.3 76 5.0 Ascaris lumbricoides 238 1.5 270 1.4 135 1.5 18 1.2 Entamoeba histolytica 142 0.9 191 1.0 96 1.1 10 0.7 Enterobius vermicularis 99 0.6 89 0.5 21 0.2 2 0.1 Taenia saginata 36 0.2 28 0.1 3 < 0.1 1 0.1 Trichostrongylus colubriformis 34 0.2 34 0.2 20 0.2 4 0.3 Trichuris trichiura 16 0.1 18 0.1 4 < 0.1 0 < 0.1 Ancylostoma duodenale 5 < 0.1 4 < 0.1 3 < 0.1 0 < 0.1 Other 916 5.9 929 4.8 397 4.5 65 4.3 Total 3981 25.5 3258 16.9 1316 15.0 176 11.6 n = number of participants. 15 Prevalence of intestinal.pmd 12/20/2005, 2:35 PM380 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 381 The prevalence of intestinal parasitic in- fections was slightly higher in males than females, which is in agreement with the findings of a study in Nigeria [15] and pre- vious studies in local regions of the Islamic Republic of Iran, in Tabas [12], Kerman [8] and Shahrekord [9]. Population studies of the northern part of the Islamic Republic of Iran in Nour, Beshehr [16] and Amol [11] showed a higher prevalence of infec- tion in females than in males. This may be explained by environmental factors as more women are occupied in farming in this area compared with other parts of the country. G. lamblia (10.9%) was the most com- mon species among unicellular organisms and A. lumbricoides (1.5%) the most com- mon among nematodes causing disease. This finding is in accordance with studies in other parts of the world as well as the Islamic Republic of Iran [16]. In Mexico City, up to 18% of cases of acute diarrhoea and dysentery in children requiring hospi- talization were found to be associated with G. lamblia as well as 10% with E. histolyt- ica and 7% with Blastocystis hominis [17]. In 1983, in the United States of America, G. lamblia was identified as the cause of 68% of waterborne outbreaks of diarrhoea in which an etiologic agent was unknown [18]. We found that intestinal parasitic infec- tions are more common in rural than urban areas. People living in rural areas may lack sanitary water supplies and live close to sources of parasites in social and environ- mental conditions that predispose to intesti- nal parasitic infections. According to this study, the common intestinal parasitic infections were more frequent in children, which suggests that screening tests for this age group may be useful. Because the prevalence of parasitic infections is high in developing countries, it Table 3 Prevalence of intestinal parasitic infections in the Islamic Republic of Iran by area of residence Parasite Urban Rural (n = 27 513) (n = 17 615) No. % No. % Giardia lamblia 2602 9.5 2301 13.1 Ascaris lumbricoides 295 1.1 266 1.5 Entamoeba histolytica 270 1.0 169 1.0 Enterobius vermicularis 113 0.4 98 0.6 Taenia saginata 26 0.1 42 0.2 Trichostrongylus colubriformis 39 0.1 53 0.3 Trichuris trichiura 10 < 0.1 28 0.2 Ancylostoma duodenale 9 < 0.1 3 < 0.1 Other 1088 4.0 1219 6.9 Total 4452 16.2 4179 23.7 n = number of participants. 15 Prevalence of intestinal.pmd 12/20/2005, 2:35 PM381 382 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 is suggested that local or regional research- ers undertake studies on such infections in these countries, perhaps with the technical support of the World Health Organization. In view of the distribution of intestinal parasitic infections in this study, preventive measures should be taken for high-risk populations by increasing their level of knowledge about personal and community health and hygiene. By comparing the find- ings of this study with those of the National Health and Disease Survey of 1993 [19], we conclude there has been a noticeable decline in the prevalence of parasitic dis- eases in the Islamic Republic of Iran, which may be due to improvements in en- vironmental and personal health which have occurred through public education campaigns, improved sanitation facilities, proper waste and wastewater disposal and control of drinking-water and food safety. Acknowledgements This study was financially supported by the National Medical Research Centre of Iran and was performed in cooperation with the Ministry of Health and Medical Education of the medical universities throughout the Islamic Republic of Iran. We would like to thank Dr A.A. Noorbala, Dr K. Moham- mad, Dr H. Malekafzali and all who assisted us in performing this study. References 1. Kappus KD et al. Intestinal parasitism in the United States: update on a continu- ing problem. American journal of hy- giene and tropical medicine, 1994, 50(6):705–13. 2. Lacroix M, Sorensen B. Forekomsten af Enterobius vermicularis hos born indlagt pa et centralsygehus. [Occurrence of En- terobius vermicularis in children hospi- talized at a central hospital]. Ugeskrift for laeger, 2000, 162(9):1236–8. 3. Xu LQ et al. Soil transmitted helminthi- asis: nationwide survey in China. Bulle- tin of the World Health Organization, 1995, 73(4):507–13. 4. Tchuem Tchuente LA et al. Polypara- sitism with Schistosoma haematobium and soil-transmitted helminth infections among school children in Loum, Cameroon. Tropical medicine & interna- tional health, 2003, 8(11):975–86. 5. Hellard ME et al. Prevalence of enteric pathogens among community based as- ymptomatic individuals. Journal of gas- troenterology and hepatology, 2000, 15(3):290–3. 6. Levy J. Epidemiological survey of intesti- nal parasitic infections in children of Sabah, Malaysia. Community medicine, 1988, 10(3):240–9. 7. Chandiwana SK, Makaza D. Some epi- demiological aspects of intestinal helm- inth infections in a farmworker community in Burma Valley. Central Afri- can journal of medicine, 1983, 29(9): 173–7. 8. Naser ZA, Jafar M. Prevalence of intesti- nal parasites in the city of Kerman. Ira- nian journal of parasitology, 1997, 11: 129a. 9. Koroosh MN. Prevalence of intestinal parasitic infestations in patients attend- ing the parasitology laboratory in Shahrekord. Iranian journal of parasitol- ogy, 1997, 11:131a. 10. Azam S. Frequency distribution of intesti- nal worms in Behshar and Norr cities. Tehran, Tehran University School of Health Sciences and Research Institute, 1995. 15 Prevalence of intestinal.pmd 12/20/2005, 2:35 PM382 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 383 11. Ali HZO. Prevalence of intestinal para- sites in Amol city. Tehran, Tehran Univer- sity School of Health Sciences and Research Institute, 1999. 12. Saied HM. Prevalence of intestinal para- sites in Tabas city. Tehran, Tehran Univer- sity School of Health Sciences and Research Institute, 1999. 13. Kobayashi J et al. Prevalence of intesti- nal parasitic infection in five farms in Holambra, Sao Paulo, Brazil. Revista do Instituto de Medicina Tropical de São Paulo, 1995, 37(1):13–8. 14. Murray CJL, Lopez AD, Mathers CD, eds. The global epidemiology of infectious diseases. Global burden of disease and injury series. Volume IV. Geneva, World Health Organization, 2004. 15. Agi PI. Pattern of infection of intestinal parasites in Sagbama community of the Niger Delta, Nigeria. West African jour- nal of medicine, 1995, 14(1):39–42. 16. Ashrat Bighom K, Hossin H, Erag M. As- sessment of human intestinal parasites in Iran during the last 50 years. Abstracts of the 2nd National Congress of Para- sitic Diseases, Tehran, Islamic Republic of Iran, October 19–22, 1997:137. 17. Diaz E et al. Epidemiology and control of intestinal parasites with nitazoxanide in children in Mexico. American journal of hygiene and tropical medicine, 2003, 68(4):384–5. 18. Weniger BG et al. An outbreak of water- borne giardiasis associated with heavy water runoff due to warm weather and volcanic ashfall. American journal of public health, 1983, 73(8):868–72. 19. Deputy of Research of Ministry of Health and Medical Education. Health and dis- ease in Iran. Tehran, Moavenat Pajooheshi Publications, 1993 (Report No. 25). Call for reviewers In order to ensure the high standard of EMHJ, all manuscripts re- ceived are peer reviewed and we greatly appreciate the valuable contribution made by our reviewers in this “quality control process”. We are always looking for experienced manuscript reviewers and would like to extend an invitation to biomedical and public health researchers to join our panel of reviewers. If you feel you have the necessary expertise and would be able to provide comprehensive reviews in a timely manner, please email the Editor-in-Chief (emhj@emro.who.int) giving your name and contact details, and your qualifications and specific areas of expertise. We are particu- larly interested to hear from French speakers and reviewers in the following areas: social sciences, psychology, epidemiology and obstetrics/gynaecology. 15 Prevalence of intestinal.pmd 12/20/2005, 2:35 PM383 384 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Is Q fever an emerging infection in Turkey? A. Gozalan,1 B. Esen,1 J.M. Rolain,2 L. Akin3 and D. Raoult2 1Department of Microbiology and Clinical Microbiology, Refik Saydam Hygiene Centre, Ankara, Turkey. 2WHO Collaborative Centre for Rickettsial Reference and Research, Marseille, France. 3Department of Public Health, Hacettepe University School of Medicine, Ankara, Turkey (Correspondence to L. Akin: ltakin@superonline.com). Received: 12/10/03; accepted: 08/03/04 ABSTRACT Between 4 May and 8 August 2002, 46 cases of acute fever were reported near the Black Sea region in northern Turkey. The infection was treated rapidly and successfully with tetracyclines, so clinical diagnosis of rickettsial or ehrlichial infection was considered. Analysis of serum and blood samples taken from 19 patients identified the causative organism as Coxiella burnetii; 7 cases were reported as acute Q fever and 8 as seropositive for past infection. The most common clinical symptoms among the acute cases were vomiting (100.0%), nausea (85.7%), diarrhoea (57.1%), fever (42.9%), abdominal pain (42.9%) and headache (42.9%). Liver enzymes were elevated in all patients. It is considered that epidemiological inves- tigation for Q fever will be essential in the affected region in future. La fièvre Q est-elle une infection émergente en Turquie? RÉSUMÉ Entre le 4 mai et le 8 août 2002, 46 cas de fièvre aiguë ont été signalés près de la Région de la mer Noire au nord de la Turquie. L’infection a été traitée rapidement et avec succès par des tétracyclines ; aussi un diagnostic clinique de rickettsiose ou d’ehrlichiose a-t-il été retenu. L’analyse des échantillons sériques et sanguins prélevés sur 19 patients a permis d’identifier l’agent causal comme étant Coxiella burnetii. Sept cas ont été rapportés comme étant une fièvre Q aiguë et huit comme ayant une sérologie positive séquellaire d’une infection ancienne. Les symptômes cliniques les plus courants chez les cas aigus étaient les vomisse- ments (100,0 %), les nausées (85,7 %), la diarrhée (57,1 %), la fièvre (42,9 %), les douleurs abdominales (42,9 %) et les céphalées (42,9 %). Les taux d’enzymes hépatiques étaient élevés chez tous les patients. On estime que l’investigation épidémiologique pour la fièvre Q sera essentielle dans la région touchée à l’avenir. 16 Is Q fever an emerging.pmd 12/20/2005, 2:36 PM384 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 385 Introduction Q fever (Query fever) is a zoonosis caused by Coxiella burnetii, which occurs world- wide, except in New Zealand. Sheep, goats and cattle are considered the main reser- voirs for the organism, which is shed in milk, urine, faeces and birth products of infected animals. In humans, the infection is caused mainly via inhalation of contami- nated aerosols from amniotic fluid, placen- ta or wool [1], but oral (contaminated milk) [2,3], percutaneous [4], vertical, and even person-to-person, e.g. during delivery of an infant or via blood transfusion [5,6] or ticks [4], can be considered other modes of transmission. The clinical manifestations of Q fever in humans are highly variable. The most com- mon symptoms are fever, pneumonia, hep- atitis, and meningoencephalitis in the acute form, whereas endocarditis is the main symptom in chronic cases. Nevertheless, asymptomatic infection occurs in about 50% of cases [3,7,8]. Serology is the main tool in the diagno- sis of Q fever since clinical diagnosis, in most instances, lacks specificity. The most reliable methods are indirect immunofluor- escence, complement fixation, enzyme- linked immunosorbent assay and micro- agglutination, all of which are very com- monly used. Currently, immunofluores- cence assay is the reference method for the serodiagnosis of Q fever [1,9]. The aim of this study is to investigate the sociodemographic, clinical and labora- tory features of cases in an outbreak of Q fever to call attention to a newly emergent infection in our country. Methods In June 2002, a number of acute cases of fever were reported from Tokat, near the Black Sea region in northern Turkey. Doc- tors from the General Directorate of Prima- ry Health Care and the Department of Communicable Diseases Research at the Refik Saydam National Hygiene Institute drew up a questionnaire after investigation in the field and analysis of the clinical symptoms of the cases. The questionnaire requested information about the participat- ing health centre; sociodemographic, clini- cal and laboratory features of the infection, including treatment of people suspected of having the infection; and additional epide- miological information. According to the analysis of preliminary hospital based data, the probable case definition was established as: people who were living in or near Tokat and neighbouring provinces with symp- toms of fever, myalgia/arthralgia, nausea, vomiting, abdominal pain, diarrhoea, thrombocytopenia, leukopenia and elevated liver enzyme levels since 4 May 2002. The questionnaire was circulated to only those patients whose symptoms met the case definition and blood samples were collected from them. The epidemiological investigation was carried out from 4 May 2002 until 8 August 2002. The operation was conducted by the health facilities lo- cated in the affected region in order to col- lect the information about the probable cases via questionnaires. The sera and blood samples from a total of 26 convalescent patients and people with acute disease were sent to the Communica- ble Diseases Research Department of Refik Saydam National Hygiene Institute. Serum samples from 19 patients were sent to the WHO Collaborative Centre for Rickettsial Reference and Research, Marseille, France in July 2002. The samples were investigat- ed for rickettsial infection (Rickettsia conorii, R. slovaca, R. helvetica, R. israe- li, R. massiliae, R. mongolotimonae, R. aeschlimanii, R. felis and R. typhi), human 16 Is Q fever an emerging.pmd 12/20/2005, 2:36 PM385 386 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 granulocytic ehrlichiosis, leptospirosis and C. burnetii using indirect immunofluores- cence assay methods. Although an anti- phase II IgG antibody titre of ≥ 1:200 and an anti-phase II IgM titre of ≥ 1:50 are considered significant in serologic diagno- sis of acute Q fever [4,9–11], an anti-phase II IgG titre of ≥ 1:100, and an anti-phase II IgM titre ≥ 1:25 were accepted in our group as cut off values in accordance with the clinical and epidemiological data. Five patients died during the outbreak. It was not possible to get samples or to do any post-mortem evaluation in these cases. This was a descriptive epidemiologic study. Statistical analysis was done using Fisher’s exact test and, owing to the low number of cases, Kruskal–Wallis analysis of variance was implemented to compare the means. Results Between 4 May and 8 August 2002, 46 peo- ple, 38 of them from Tokat, became ill with symptoms that were consistent with those of the case definition. The clinical symp- toms were recorded as: abdominal pain (77.8%), nausea (66.7%), vomiting (66.7%), arthralgia/myalgia (66.7%), head- ache (55.6%), fever (55.6%), diarrhoea (33.3%) and rash (33.3%). The aspartate aminotransferase, alanine aminotransferase and lactate dehydrogenase levels were ele- vated in 77.8%, 66.7% and 55.6% respec- tively. Leukopenia and thrombocytopenia were noted in 66.7% of patients. The interval between the onset and the end of the outbreak was 94 days for con- firmed cases. The mean duration of the ill- ness in those who recovered was 11 days for probable cases. Eleven people suspect- ed of having the infection reported tick contact. Only 2 of the patients had family members with similar symptoms. A total of 19 serum samples were sent to the WHO Collaborative Centre for Rick- ettsial Reference and Research, Marseille, France for further evaluation (the limited facilities available in Turkey precluded test- ing locally). Sera and blood samples were tested for rickettsial infection, ehrlichiosis and leptospirosis using indirect immunoflu- orescence assay methods and all were eval- uated as negative. Acute Q fever was confirmed for 7 samples and 8 others were found to be seropositive. The other 4 were negative for Q fever. The first acute case of Q fever was de- tected on 25 May 2002 and the last was on 17 June 2002, an interval of 23 days. The infection started at the end of the spring season and continued till the beginning of summer. Recovery from the illness was re- ported as minimum 5 days and maximum 10 days. The median age was 48 years in individ- uals with acute infection and 60 years in those with evidence of past infection. The median age was 54 years in people who were identified as seronegative for C. bur- netii. These differences were, however, not statistically significant. In addition, there was no statistical significance be- tween males and females for seropositivity. The number of people seropositive for acute infection was higher in rural areas than in urban areas, but again, the differ- ence was not statistically significant (Table 1). Three of the 7 people who had acute infection reported having had close contact with ticks. The most common clinical symptoms among the acute cases were vomiting, nau- sea and diarrhoea (Table 2). The fever was subfebrile. 16 Is Q fever an emerging.pmd 12/20/2005, 2:36 PM386 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 387 Discussion Communicable diseases continue to be an important cause of morbidity and mortality in Turkey, acute upper respiratory tract in- fections being the most common infec- tions. The control programmes for acute respiratory tract infections, acute diarrhoea and neonatal tetanus; the elimination pro- gramme for measles; and the eradication programme for poliomyelitis have been op- erating since the 1980s. There has also been significant progress towards achieving the goals of measures which were implemented for the control and prevention of re-emerging diseases such as malaria and tuberculosis. Future programmes and activities are planned to target the control of bacterial, parasitic and viral diseases [12]. In May–June 2002, near the Black Sea region of Turkey, an infection character- ized by abdominal pain, nausea, vomiting, myalgia/arthralgia, headache, fever and rash in addition to laboratory findings of leukopenia, thrombocytopenia and elevated liver enzymes was noted. As the infection was treated rapidly and successfully with tetracyclines, rickettsial or ehrlichial infec- Table 1 Characteristics of 19 people with suspected Q fever (Coxiella burnetii) infection Characteristic Seropositive for C. burnetii Seronegative for C. burnetii Acute (n = 7) Past infection (n = 8) (n = 4) No. % No. % No. % Age (years) 10–29 2 28.5 0 – 0 – 30–49 2 28.5 1 12.5 2 50.0 50–69 2 28.5 4 50.0 1 25.0 ≥ 70 1 14.3 3 37.5 1 25.0 Median 48 60 54 Mean (SD) 46.5 (19.1) 62.6 (12.2) 52.0 (16.4) Kruskal–Wallis statistic: 0.351; P > 0.05 Sex Male 4 57.1 6 75.0 1 25.0 Female 3 42.9 2 25.0 3 75.0 Likelihood ratio: 2.807; P > 0.05 Residence Rural 6 85.7 7 87.5 4 100 Urban 1 14.3 1 12.5 0 – Likelihood ratio: 2.807; P > 0.05 Contact with tick Yes 3 42.9 2 25.0 1 No 4 57.1 6 75.0 3 Likelihood ratio: 0.642; P > 0.05 SD = standard deviation. 16 Is Q fever an emerging.pmd 12/20/2005, 2:36 PM387 388 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 tions were considered possible causes, re- sulting in a great interest in the region. The outbreak was confirmed as Q fever by the laboratories of WHO Collaborative Centre for Rickettsial Reference and Research. Al- though it was not possible to take liver bi- opsies, all cases were reported as Q fever according to their clinical and laboratory features [3,7,13,14]. Verification of the cause of the infection did, however, take some time as Q fever was not expected in the country. The first outbreak of Q fever in Turkey was reported from Aksaray province in 1948. That outbreak had also begun in May and ended in August. Twenty-one cases were detected by complement fixation test. The disease showed up among animals, be- ing transported by the stool of infected ticks via inhalation by the lambs [15]. Since then, there has been no outbreak of Q fever identified by epidemiological investigation until our study. Of the 46 probable cases, 38 (82.6%) were living in Tokat, where there are a lot of livestock farms especially for cattle, sheep and goats. Acute Q fever has been reported as oc- curring as either an outbreak or as sporadic cases in spring or early summer in Europe, which was the case in this outbreak in our country. Incidence is most frequent during spring, and is epidemiologically related to the environmental contamination caused during the outside lambing and shearing seasons [13,16–19]. Although the first probable case in this outbreak was detected on 4 May, the first confirmed cases were notified on 25 May. Six of those identified Table 2 Some clinical and laboratory features of the cases Clinical symptoms Seropositive Seronegative Acute cases Past infection (n = 4) (n = 7) (n = 8) No. % No. % No. % Physical Vomiting 7 100.0 0 – 0 – Nausea 6 85.7 3 37.5 0 – Diarrhoea 4 57.1 1 12.5 0 – Fever 3 42.9 6 75.0 1 25.0 Abdominal pain 3 42.9 2 25.0 1 25.0 Headache 3 42.9 3 37.5 1 25.0 Myalgia/arthralgia 2 28.5 2 25.0 4 100 Rash 1 14.3 3 37.5 1 25.0 Biochemistry Elevated AST/ALT 7 100.0 5 62.5 1 25.0 Elevated LDH 4 57.1 4 50.0 1 25.0 Haematology Thrombocytopenia 6 85.7 4 50.0 4 100 Leukopenia 6 85.7 6 75.5 4 100 AST = aspartate aminotrasferase. ALT = alanine aminotrasferase. LDH = lactate dehydrogenase. 16 Is Q fever an emerging.pmd 12/20/2005, 2:36 PM388 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 389 as having the infection (85.7%) were living in rural areas. The onset of the outbreak corresponds to the time that animals such as sheep and goats living in that region have their young. Three of the patients with acute Q fever had close contact with ticks. Although Q fever has also been defined as an arthropod-borne infection, the transmis- sion of C. burnetii from ticks to humans has rarely been reported [4]. The Q fever seropositivity of goats, sheep and cows has been reported ranging from 3% to 30% in different regions of Turkey, and the occurrence of the infection in both humans and animals at the same time suggests that it is endemic [20–23]. The incidence period of the outbreaks and the probable routes of transmission suggest that the diseases may occur again in the same season. We found 8 people with evi- dence of past infection. They had probably been exposed to Q fever infection before. It was not established when. Epidemiologic data, therefore, suggest that Q fever is an endemic infection in the region, and surveillance should be strength- ened for tick-borne and zoonotic infec- tions. Since diagnosing this group of cases and the (probable) outbreak, we have start- ed to establish plans for the diagnosis of Q fever. Indirect immunofluorescence has been accepted as the most reliable method, and we are still continuing laboratory sur- veillance activities in our institute. Q fever infection is often asymptomatic in farm animals, but can cause considerable morbidity in humans. Because Q fever is an occupational hazard [14,18,24], specific epidemiological studies in endemic areas to establish for example, seroprevalence in farm animals and farmers, may help to de- fine the target population, extent of trans- mission and potential need for a vaccine. Control and eradication measures require a multidisciplinary approach and cooperation between farmers, veterinarians and public environmental health workers. Acknowledgements We thank Nilgun Noyan, Vedat Buyurgan, Fatih Sua Tapar, Basak Dokuzoguz, Aysel Kocagül Çelikbas, Hakan Leblebicioglu, Mehmet Bakir, Mehmet Sencan, Nazif Elal- di, Özgür Öncül, Hiroshi Yoshikura, Norio Ohashi, Nurcan Baykam, Belkis Levent, Kikuko Miyamura and Denis Coulombier for their cooperation. Addendum Subsequent to the identification of the cas- es reported in this paper, new cases where reported in the same area the following year accompanied by some different clinical symptoms. The new cases were identified as Crimean–Congo haemorrhagic fever in late 2003. After this was recognized, it was suspected that the cases from the previous year, reported in this paper, might also be related to Crimean–Congo haemorrhagic fever. Samples were therefore sent for re- analysis and it was then found that some of the cases were indeed Crimean–Congo haemorrhagic fever. Details of these cases are published in: [Epidemiological evalua- tion of a possible outbreak in and nearby Tokat province] Gozalan A, Akin L, Rolain JM, Tapar FS, Oncul O, Yoshikura H, Zeller H, Raoult D, Esen B. Mikrobiyoloji bülteni, 2004, 38(1–2):33–44 [Article in Turkish]. It should be noted that a multidisci- plinary approach to the control of tick- borne diseases is being carried out by the involved sectors in the country and general practitioners and clinicians have been alert- ed to the emergence of these pathogens. 16 Is Q fever an emerging.pmd 12/20/2005, 2:36 PM389 390 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 1. Carrieri MP et al. Investigation of a slaughterhouse-related outbreak of Q fever in the French Alps. European jour- nal of clinical microbiology, 2002, 21(1): 17–21. 2. Fishbein DB, Raoult D. A cluster of Cox- iella burnetii infections associated with exposure to vaccinated goats and their unpasteurized dairy products. American journal of tropical medicine and hy- giene, 1992, 47(1):35–40. 3. Tissot-Dupont TH et al. Epidemiologic features and clinical presentation of acute Q fever in hospitalized patients: 323 French cases. American journal of medicine, 1992, 93(4):427–34. 4. Maurin M, Raoult D. Q fever. Clinical mi- crobiology reviews, 1999, 12(4):518– 53. 5. Stein A, Raoult A. Q fever during preg- nancy: a public health problem in south- ern France. Clinical infectious diseases, 1998, 27(3):592–6. 6. Raoult D, Stein A. Q fever during preg- nancy—a risk for women, foetuses, and obstetricians. New England journal of medicine, 1994, 330(5):371. 7. Raoult D et al. Q fever 1985–1998. Clini- cal and epidemiologic features of 1,383 infections. Medicine, 2000, 79(2):109– 23. 8. Brouqui P et al. Chronic Q fever. Ninety- two cases from France, including 27 cases without endocarditis. Archives of internal medicine, 1993, 153(8):642–8. 9. Fournier PE, Marrie TJ, Raoult D. Diag- nosis of Q fever. Journal of clinical micro- biology, 1998, 36(7):1823–34. 10. Parola P, Raoult D. Ticks and tick-borne bacterial diseases in humans: an emerg- ing infectious threat. Clinical infectious diseases, 2001, 32(6):897–928. 11. Levett PN. Leptospirosis. Clinical micro- biology reviews, 2001, 14(2):296–326. 12. Activity report of the Primary Health Care General Directorate, Ministry of Health 1996–1999. Ankara, Prestij Publishing, 2001. 13. Serbezov V et al. Q fever in Bulgaria and Slovakia. Emerging infectious diseases, 1999, 5(3):388–94. 14. Hatchette TF et al. Goat-associated Q fe- ver: a new disease in Newfoundland. Emerging infectious diseases, 2001, 7(3):413–9. 15. Payzin S. Q fever epidemic in Ozancik village. Turkish bulletin of hygiene and experimental biology, 1948, 8(3):116– 24 (in Turkish). 16. Hellenbrand W, Breuer T, Petersen L. Changing epidemiology of Q fever in Germany, 1947–1999. Emerging infec- tious diseases, 2001, 7(5):789–96. 17. Lyytikainen O et al. An outbreak of sheep-associated Q fever in a rural com- munity in Germany. European journal of epidemiology, 1998, 14(2):193–9. 18. Richardus JH et al. Q fever in the Nether- lands: a sero-epidemiological survey among human population groups from 1968 to 1983. Epidemiology and infec- tion, 1987, 98(2):211–9. 19. Smith DL et al. A large Q fever outbreak in the West Midlands: clinical aspects. Respiratory medicine, 1993, 87(7):509– 16. 20. Payzin S. Epidemiology of Q fever in Tur- key. Turkish bulletin of hygiene and ex- perimental biology, 1948, 8(2):101–10. 21. Golem BS. Q fever epidemiology and in- formation on Q fever infection among animals in Turkey. Turkish bulletin of hy- giene and experimental biology, 1951, 11(1):1–21 [in Turkish]. References 16 Is Q fever an emerging.pmd 12/20/2005, 2:36 PM390 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 391 22. Payzin S, Golem BS. Q fever in Turkey. Turkish bulletin of hygiene and experi- mental biology, 1948;8(3):94–113 (in Turkish). 23. Ozyer M, Miriglu M, Koksal F. An investi- gation of Q fever infection using a complement fixation test among human and animal populations in Cukurova re- gion. Journal of Pendik Animal Diseases Central Research Institute, 1990, 21(2):28–39 [in Turkish]. 24. Abe T et al. A seroepidemiological study of the risks of Q fever infection in Japa- nese veterinarians. European journal of epidemiology, 2001, 17(11):1029–32. 16 Is Q fever an emerging.pmd 12/20/2005, 2:36 PM391 392 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 School non-enrolment and its relation with health and schistosomiasis knowledge, attitudes and practices in rural Egypt S.I. Mekheimar1 and M. Talaat1 1Department of Community and Social Medicine, Theodor Bilharz Research Institute, Imbaba, Cairo, Egypt (Correspondence to S.I. Mekheimar: shahinazmekheimer@yahoo.com). Received: 10/10/02; accepted: 13/05/03 ABSTRACT Children who are not enrolled in school are deprived of schistosomiasis treatments delivered through school health programmes. We explored perceptions of barriers to school enrolment and health and schistosomiasis knowledge, attitude and practices among 58 enrolled and 41 non-enrolled children and 80 adults in a village in El-Fayoum governate, Egypt. Economic factors, cultural factors and the school system were perceived as barriers. Maternal education had a significant role in enrolment. Non-enrolled children were more anaemic and had more signs of vitamin deficiency. Enrolled children had better schistosomiasis knowl- edge and attitudes, but not practice; both groups had high Schistosoma mansoni infection rates (51.7% enrolled and 65.8% non-enrolled). Education and health policies should tackle perceived barriers. Introduction La relation entre la non-scolarisation et les connaissances, attitudes et pratiques concernant la schistosomiase dans l’Égypte rurale RÉSUMÉ Les enfants qui ne sont pas scolarisés ne bénéficient pas des traitements de la schistosomiase qui sont administrés dans le cadre des programmes de santé scolaires. Nous avons étudié la perception des obstacles à la scolarisation ainsi que les connaissances, attitudes et pratiques concernant la schistosomiase chez 50 enfants scolarisés, 41 enfants non scolarisés et 80 adultes dans un village du Gouvernorat de Fayoum en Égypte. Les facteurs économiques, les facteurs culturels et le système scolaire étaient perçus comme des obstacles. L’éducation de la mère jouait un rôle significatif dans la scolarisation. Les enfants non scolarisés souffraient plus d’anémie et présentaient plus de signes de carence vitaminique. Les enfants scolarisés avaient de meilleures connaissances et attitudes concernant la schistosomiase, mais pas de meilleures pratiques ; les deux groupes avaient un taux élevé d’infestation par Schistosoma mansoni (51,7 % pour les enfants scolarisés et 65,8 % pour les non-scolarisés). Les politiques d’éducation et de santé devraient s’attaquer aux obstacles perçus. 17 School non-enrolment.pmd 12/20/2005, 2:36 PM392 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 393 School-age children experience a consider- able burden of diseases that may have both immediate and long-term consequences for their health, growth and education [1]. In developing countries, much morbidity has been attributed to infectious and parasitic helminth infections, especially schistoso- miasis [2–4]. Enrolled children usually have better health as they have access to health services through school-based health pro- grammes that provide excellent opportuni- ties for early detection and treatment of infectious and parasitic diseases [3–7]. Following this approach, the Egyptian Ministry of Health and Population delivers schistosomiasis treatment to students. Ef- forts have been jeopardized in areas where non-enrolment rates are high; these areas are also characterized by high prevalence and intensity of infection. Previous studies in Egypt have documented poor school at- tendance with marked variations by gender and location and a high proportion of in- fected non-enrolled children who miss schistosomiasis treatment [7,8]. Few stu- dents have explored the reasons for non- enrolment from economic, cultural and health perspectives, although the impor- tance of these factors in developing coun- tries has been acknowledged [9]. The aim of our study was to explore parental factors and barriers to enrolment and to identify the health status and schis- tosomiasis knowledge, attitude and prac- tice (KAP) among enrolled and non- enrolled children. Methods El-Fayoum governorate, a large oasis in the Egyptian Western Desert that is 90–130 km southwest of Cairo, was selected for our study because of its low rate of school en- rolment (38%) [10,11]. Within the El-Fay- oum districts, El-Bitar, an ezba, or small village, was chosen because it has the low- est rate of enrolment (9.4%) and the high- est Schistosoma mansoni prevalence (26.4%) [10,11]. El-Bitar is a small satellite community with 80 houses and 570 inhab- itants. It is crossed by 2 canals, has no drainage system and has no rural health unit. There is 1 preparatory school with 446 students who primarily come from nearby villages. Both quantitative and qualitative data were collected from the school and the vil- lage. Results are given as frequencies or means (standard deviations). School health survey Children were screened at the only prepara- tory school in the satellite where the district laboratory technician performed stool ex- amination, blood testing and schistosomia- sis treatment. Schoolteachers, the district senior laboratory technician, community leaders and the enrolled children were en- listed to recruit non-enrolled children. Trained social workers interviewed the children and with the help of a public health consultant, focus group discussions and physical examinations were conducted. Using statistical methods, a random sample of 58 children enrolled in the school and aged between 12 and 13 years were se- lected. During a 4-month period 41 non- enrolled children in the same age group complied and were included in our study. Both groups were interviewed in a struc- tured interview to explore their parents’ characteristics and to measure their schis- tosomiasis KAP. They were medically ex- amined for height, weight and signs of vitamin deficiency (angular stomatitis and depigmentation) and blood and stool sam- ples were analysed with parental consent. A total of 5 focus group discussions were conducted (2 among the non-enrolled and 3 among the enrolled) exploring their percep- 17 School non-enrolment.pmd 12/20/2005, 2:36 PM393 394 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 tions of health, education and the link be- tween them. Villager survey The socioeconomic background and rea- sons for not enrolling children in school of all participating villagers were collected with a structured questionnaire. A total of 7 focus group discussions were held (4 mother and 3 father groups) to explore pa- rental perceptions about health and educa- tion. There was no attempt to relate the school health survey and the villager survey as most enrolled children came from near- by villages and the perceptions of close communities might not differ. Results School health interviews Most fathers of enrolled and non-enrolled children were farmers (77.6% enrolled and 87.8% non-enrolled, P > 0.05) and most mothers were housewives (93.1% enrolled and 85.4% non-enrolled). Most non- enrolled children were significantly charac- terized by illiterate mothers compared with the enrolled group (95.1% non-enrolled and 74.1% enrolled, P < 0.05). Family size in both groups was high [enrolled 6.2 (SD 2.4) people and non-enrolled 7.0 (SD 2.3) people, P > 0.05]. Most enrolled (91.4%) and non-enrolled (100.0%) children knew the modes of transmission of schistosomiasis. Less than half of each group (32.7% enrolled and 41.5% non-enrolled) identified blood in the urine as a symptom but blood in the stool was not mentioned even once. Nearly half (43.9%) of the non-enrolled and 24.1% of the enrolled incorrectly identified treatment (P < 0.05). The enrolled differed signifi- cantly in their knowledge of complications from the non-enrolled (44.8% and 17.1% respectively, P < 0.05). Knowledge of complications was considered present if the child could name 1 complication of schistosomiasis. Nearly half of both groups said television was their source of knowl- edge (enrolled 55.2% and non-enrolled 58.5%); only 15 (25.9%) of those enrolled said school was the source of their knowl- edge. Most children in both groups believed that treatment can cure (non-enrolled 95.1% and enrolled 100.0%), but the non- enrolled believed more in re-infection (87.8% and 0.0% respectively, P < 0.05). The enrolled had a stronger sense of re- sponsibility in acquiring the infection whereas the non- enrolled believed that infection was determined by destiny (70.7% and 51.2% respectively, P < 0.05). Most enrolled and non-enrolled children reported using canal water, mostly for farming, playing and fishing. Table 1 shows the health status of the enrolled and the non-enrolled children. The non-enrolled children had a lower mean blood haemoglobin count than the enrolled [11.6 (SD 1.2) and 10.6 (SD 1.1) g/dL re- spectively, P < 0.05], weighed more than the enrolled [42.0 (SD 10) kg and 37.7 (SD 8) kg, P < 0.05] and exhibited more signs of vitamin deficiency. No differences were detected in S. mansoni infection, which was high in both groups (65.8% non- enrolled and 51.7% enrolled, P > 0.05). Most children related health to cleanli- ness and some non-enrolled children identi- fied health as a means of getting a job. Some attributed good health to the in- creased number of hospitals and some oth- ers attributed the deterioration of health conditions to pollution and to the poor qual- ity of health services. Most children from both groups per- ceived those who were enrolled in school as having better health because they were 17 School non-enrolment.pmd 12/20/2005, 2:36 PM394 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 395 cleaner and did not work in the fields as the non-enrolled children did. Only one en- rolled child attributed better health of the enrolled to the health awareness provided in schools. The enrolled children complained about health services and doctors were described as reluctant and often absent. Schistosomi- asis treatment was perceived as useless and “does not cure”. Some non-enrolled chil- dren believed in the cure effect of treat- ment. Both groups perceived the importance of education for the same reason, that is, the ability to read and write and to have a professional career, such as doctor, engi- neer or lawyer. The majority of both groups of children identified school fees as a barrier to enrol- ment. Some enrolled children mentioned as barriers that the child himself might not like education, overcrowding of children in the classroom and parental unawareness of the importance of education. Some non-en- rolled children attributed non-enrolment to their parents’ need of child labour, to the unemployment facing those educated and to the attitudes of teachers who beat chil- dren at school. Most enrolled children mentioned that the decision to enrol is made by fathers and mothers together; the non-enrolled children said the father made the decision. Both groups preferred not to educate girls and the reasons given were primarily due to the perceived cultural role of girls in later life, i.e. marriage and helping her hus- band (but not contributing financially to her family). Very few mentioned that parents protected their girls by not sending them to school. Very few mentioned that educating girls was preferable to educating boys as boys could work even if they are not edu- cated. Most respondents mentioned that those who are educated are cleaner and can take better care of their health. Table 1 Comparison between the health status of the enrolled and the non-enrolled children Variable Enrolled Non-enrolled P-value (n = 58) (n = 41) Mean SD Mean SD Age (years) 13.4 1.5 13.7 1.3 > 0.05 Weight (kg) 37.7 8 42 10 < 0.05 Height (cm) 146 10 148 10.2 > 0.05 Blood haemoglobin (g/dL) 11.6 1.2 10.6 1.1 < 0.05 No. % No. % Depigmentation 19 33.3 33 80.5 < 0.05 Angular stomatitis 20 34.5 24 58.5 < 0.05 Schistosoma mansoni infection 30 51.7 27 65.8 > 0.05 P < 0.05 is significant. ´SD = standard deviation. 17 School non-enrolment.pmd 12/20/2005, 2:36 PM395 396 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Villager interviews The majority of villagers (80% of men and 90% of women) were illiterate. More than two-thirds of the men were farmers who primarily worked as day labourers. Only one-quarter were owners of their land. As for women, 50 (62.5%) were housewives and 35% were agricultural workers. Almost all interviewed (73 or 91.2%) stated economic problems were the main barriers to school enrolment. They report- ed that the average school fees for primary education was 18.2 (SD 6.7) Egyptian pounds (LE) and 28.9 (SD 10.8) LE for secondary school education (US$ 1 was about LE 4–5 at the time of the study). The total average additional yearly cost for school uniforms, books and stationary was 195 (SD 40) LE. Other barriers mentioned were the need for the children to generate income, large family size and parental illit- eracy (18.7%, 12.5% and 8.6% respective- ly). All interviewed stated that education is important. A total of 66.3% mentioned that education of boys and girls are both impor- tant, but 21.3% stated that educating boys is more important. Nearly 80% said that enrolment leads to better health. Most mothers and very few fathers per- ceived cleanliness and proper food to be related to good health. Few mothers ex- pressed education and upbringing as related to good health. Factors affecting child health were dis- cussed and diarrhoea and fever were per- ceived as common problems by most respondents. More fathers than mothers reported being concerned about schistoso- miasis and cholera. Most respondents per- ceived schistosomiasis as a serious problem because it causes renal failure and cancer. When discussing factors leading to ill health, most mothers and fathers men- tioned canal water pollution by sewage dis- posal and pesticides, large family size and high costs of medical care. Most respondents preferred to use pri- vate doctors, then general hospitals and lastly health centres. A few fathers some- times used traditional remedies and bought drugs without prescriptions from pharma- cies. One father mentioned that he did nothing for sick children, but rather, he said: “We leave him to his destiny”. Most respondents agreed that the quali- ty of health care provided is unacceptable. Doctors’ attitudes were the main obstacle. Doctors were described as careless, not regularly available and lacking the skills to perform proper physical examinations. Health centre laboratory investigations were perceived as inaccurate and not seen by doctors. Medications were described as unavailable and unaffordable. Some said that: “Schistosomiasis treatment is not al- ways available in the health centres, so we buy it from pharmacies”. Accessibility was also identified as a barrier because some had to travel to reach health centres. Most respondents, regardless of gen- der, said that education is important as it enables children to read the Quran and to write. Most mothers said that cleanliness was a benefit of education, whereas fathers referred to cleanliness as a way of getting a job. Both groups identified problems they encountered with the education system that made the importance of education ques- tionable. One mother reported: “There is no education in schools”. One father said: “The number of schools is not enough, the level of education is below zero and more than 60 pupils are in one class”. Both groups said costs including school fees, unemployment after graduation and loss of child labour were barriers to enrol- ment. All respondents complained of high school fees and extra expenses for private lessons, books, uniforms and transporta- 17 School non-enrolment.pmd 12/20/2005, 2:36 PM396 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 397 tion. Unemployment shaped parental opin- ion about the enrolment decision as some parents, especially fathers, commented on the absence of an incentive for education. One father said: “Education is expensive and afterwards children will not find jobs with their diplomas”. Another said: “It is better for a child to learn a skill, because there are no jobs after finishing schools. They don’t earn money by education”. Child labour was seen as another barri- er. Most respondents explained that chil- dren who do not attend school help their parents in the fields or that they look for other jobs to support their families. A few mothers said that those who attend school do not help their parents because they be- come exhausted after the long school day. Large family size was the reason behind fi- nancial constraints for some. Cultural barriers were expressed in var- ious ways that included gender discrimina- tion. Preference to enrol boys was clear among both fathers and mothers. Boys were seen as future supporters of the fam- ily, because after marrying, boys still help with family expenses. Girls’ education was seen as useless and some said: “Educating girls is a loss of investment. We also fear mixing our girls with boys at certain age. They should marry early and learn all kinds of domestic work before marriage”. Anoth- er cultural barrier emphasized by parents was the availability of schools only in the afternoon, which was perceived as an in- convenient time for girls to go out. While the minority viewed education of boys and girls as equal, a very few mothers said that girls’ education is more important than boys’ education and the reason given was the perceived higher intelligence of girls than boys. Most respondents across the focus group discussions stated that the de- cision to enrol was made by the father alone. A minority referred to the role of grandfathers as decision-makers. Whereas finances were the main barri- ers expressed by fathers, mothers raised the issue of educational system as the main obstacle to enrolment. Most mothers blamed teachers as the main reason for dropouts. One mother reported: “Our chil- dren are afraid to go to school because they are beaten by their teachers and they hate going to school”. In addition, schools do not accept all children because only limited numbers can be accepted each year and because of age constraints. One mother said: “Because my child was less than 6- years-old by a few days, they postponed him to the following year and then a year later they refused him because he was older than the age required”. Both mothers and fathers found that the new system of the long day of education was inconvenient from many aspects. Fathers expressed: “There is no toilet, no clean drinking-water and no activities during the recreation break which takes almost 1 hour. When the school day is prolonged, children get ex- hausted and they do not help us after re- turning so they become a burden on their families”. Apart from financial barriers and large family size, a few parents mentioned child preference as decision factors. Mothers and fathers saw the relation between health and education differently. Some mothers said that school enrolment and education lead to better health because of the availability of health care at schools, teaching cleanliness so that children can take better care of their health and learning so that children can understand medical prescriptions. Confirming this opinion, mothers stressed that the health of children in schools is better. Very few mothers men- tioned that children who do not attend 17 School non-enrolment.pmd 12/20/2005, 2:36 PM397 398 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 school have better health as they eat better, rest at home and do not travel or walk long distances to school. Some others stated that better health could lead to better edu- cation achievements. Fathers’ perceptions about the relationship between education and health were not as strong as mothers’ perceptions. A minority of fathers stated that health and education are unrelated and that health is in the hands of God. Most fa- thers, however, said: “Healthy brains are in healthy bodies”. Discussion Non-enrolled children are deprived of the educational and the health benefits offered by schools. Many studies have provided evidence that the health of enrolled children is better than the health of the non-enrolled [3,12]. Our findings are similar in that the non-enrolled had a lower mean blood hae- moglobin count and showed more signs of malnutrition. In spite of the schistosomiasis treat- ment provided in schools, both enrolled and non-enrolled children had high levels of S. mansoni infection. Schistosomiasis KAP among enrolled and non-enrolled children varied. Both groups had correct knowledge about modes of transmission but that did not prevent them from coming into contact with canal water. The source of knowledge for both groups was mainly television. Nei- ther group identified blood in the stool as a sign of S. mansoni infection. Non-enrolled children differed from enrolled children in not comprehending its seriousness, in their greater belief in re-infection and in their be- lief that infection was beyond their control. This differed from the enrolled children who acknowledged responsibility in ac- quiring the infection. As documented elsewhere, parents are the decision-makers for children’s health and education. Family background vari- ables such as education and parental occu- pation influence their decisions for their children’s enrolment [13–17]. Similarly, low enrolment rates in our village (10%) were related to a high parental illiteracy rate, large average family size and the high proportion of fathers working as day la- bourers and farmers, which denoted low socioeconomic standards. Among these factors, maternal education was the most significant related to enrolment and denoted the effect of maternal illiteracy in the deci- sion of nonenrolment. This finding is simi- lar to the World Bank report on gender and wealth in that the education of adults in the household significantly affects the enrol- ment of children and that the effect of adult female education is greater than that of males [18]. Although many studies have shown that female education is related to better health, smaller family size and more educated children, female non-enrolment rates are still high [18]. Our study high- lighted the role of cultural barriers to female enrolment as in many cultures where gen- der preference plays a major role [15]. Par- ents perceived boys as economic assets to the family and as contributing to productive labour, but saw girls as a burden and a waste of investment as they marry early and their economic productivity was seen to benefit to their husband’s family rather than their parents. Hence, anticipation of future female gender roles explained to some extent gender differences in educa- tional choices made by parents. Unfortu- nately, gender differences in education continue to be very strong elsewhere as well [3,13,15,19]. Apart from cultural barriers to enrol- ment, our study detected financial and edu- cation system obstacles to child enrolment as perceived by parents and children. The poor financial capacity of families to pay 17 School non-enrolment.pmd 12/20/2005, 2:36 PM398 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 399 for extra school charges such as school fees, textbooks, school uniforms and pri- vate lessons and the high unemployment situation facing graduates were all identi- fied financial barriers. Perceived education- al system barriers included: the limited number of children accepted in schools; negative attitudes of teachers including beating, poor teaching methods and forcing children to take private lessons; over- crowding of children in the classrooms; af- ternoon schooling system that interferes with female enrolment; long distances to walk to schools; and the long exhausting schooldays that interfere with the econom- ic contribution of the children. Mothers and fathers perceived the im- portance of education differently. Mothers wanted their children to go to school to be- come cleaner and healthier. Fathers per- ceived the importance of education in obtaining a job; for fathers, unemployment after schooling was a hindering factor. Most fathers did not perceive the benefit of the school health programmes and they ac- cused the health care delivery system of being reluctant. In contrast to the literature, in our study the health of the non-enrolled children was perceived by some parents to be better than the enrolled. The reasons mentioned were that the non-enrolled have more free time, they do not expend effort studying and travelling to and from school and they eat and rest better [3,5,6,20]. This perceived difference could also be due to the higher weights of the non-enrolled children. Sur- prisingly enough, children’s ill health was an indicator of parental decisions to enrol because of the perception that they were unable to contribute to economic pro- duc- tivity. This opinion contradicted what has previously been documented, i.e. that poor health status of the child delays the deci- sion to enrol [14]. Conclusion For successful implementation of school health programmes, active cooperation of the health and education sectors is re- quired. A partnership is also needed be- tween governments and communities to explore community needs and perceptions. Suggested policies to increase access and attainment of schooling have included in- creasing female access, adopting culturally sensitive strategies, improving the health of school-age children and decreasing the health gap between enrolled and non- enrolled children. Female empowerment and involvement in income generating ac- tivities will benefit the whole of the next generation. Also, because fathers are the main decision-makers for their children’s educational and health choices, it is of ut- most importance to involve them in wom- en’s and children’s health programmes. School health promotion programmes need to be implemented to help change the behaviours of those enrolled. Acknowledgements Many thanks to Mrs Susan Watts, Social Research Center, the American University in Cairo, for her consultancy, support and encouragement and for revising the manu- script. Special thanks and appreciation to the funding organization the Special Pro- gramme for Research and Training, School Task Force, World Health Organization, Geneva. 17 School non-enrolment.pmd 12/20/2005, 2:36 PM399 400 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 1. The Partnership for Child Development. The health and nutritional status of schoolchildren in Africa: evidence from school-based health programmes in Ghana and Tanzania. Transcripts of the Royal Society of Tropical Medicine and Hygiene, 1998, 92:254–61. 2. Beasley NM et al. The impact of popula- tion level deworming on the hemoglobin levels of school children in Tanga, Tan- zania. Tropical medicine and interna- tional health, 1999, l4(11):744–50. 3. Bundy DAP, Guyatt H. Schools for health. Focus on health, education and the school-age child. Parasitology today, 1996, 12(8):1–16. 4. The Partnership for Child Development. Better health, nutrition and education for the school-aged child. Transcripts of the Royal Society of Tropical Medicine and Hygiene, 1997, 91:1–2. 5. Beasley NM et al. The health of enrolled and non-enrolled children of school age in Tanga, Tanzania. Acta tropica, 2000, 76(3):223–39. 6. Husein MH et al. Who misses out with school-based health programmes? A study of schistosomiasis control in Egypt. Transcripts of the Royal Society of Tropi- cal Medicine and Hygiene, 1996, 90: 362–5. 7. Talaat M, Omar M, Evans D. Developing strategies to control schistosomiasis morbidity in non-enrolled school-age children: Experience from Egypt. Tropical medicine and internal health, 1999, 4(8): 551–6. 8. El-Zanaty F, Way A. Egypt demography and health survey 2000. Cairo, Egypt, Ministry of Health and Population, Na- tional Population Council and Washing- ton DC, ORC Macro, 2001. 9. Butterworth AE et al. Immunity and mor- bidity in human schistosomiasis. Tropical and geographical medicine, 1994, 46: 197–208. 10. Abdel-Wahab MF et al. The epidemiol- ogy of schistosomiasis in Egypt: Fayoum Governorate. American journal of tropi- cal medicine and hygiene, 2000, 62(2): 55–64. 11. El-Khoby T et al. The epidemiology of schistosomiasis in Egypt: Summary find- ings in 9 governorates. American journal of tropical medicine and hygiene, 2000, 62(2):88–99. 12. Chitsulo L et al. The global status of schistosomiaisis and its control. Acta tropica, 2000, 77:41–51. 13. Dryler H. Parental role models, gender and educational choice. British journal of sociology, 1998, 49(3):375–98. 14. Fentiman A, Hall A, Bundy D. School en- rolment patterns in rural Ghana: a com- parative study of the impact of location, gender, age and health on children’s ac- cess to basic schooling. Comparative education, 1999, 35(3):331–49. 15. Wallace H, Giri K, Serrano CV. Health care of women and children in develop- ing countries. California, USA, Third Party Publishing Co., 1990. 16. Wharton RH et al. Advance care plan- ning for children with special health care needs: a survey of parental attitudes. Pe- diatrics, 1996, 97(5):682–7. 17. Yawn BP et al. Barriers to seeking care following school vision screening in Rochester, Minnesota. Journal of school health, 1999, 68(8):319–24. 18. Filmer D. The structure of social dispari- ties in education: gender and wealth. Washington, DC, The World Bank Devel- References 17 School non-enrolment.pmd 12/20/2005, 2:36 PM400 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 401 opment Research Group, 1999 (http:// www.worldbank.org/gender/prr, acces- sed 29 June 2004). 19. Lask S. Gender matters. Nursing stan- dards, 1998, 12(51):1–25. Expanding schistosomiasis-free areas Worldwide, over 200 million people are infected with schistosomia- sis. Groups particularly at risk are children and adolescents, and special occupation groups such as fishermen and irrigation work- ers. In the short term, the infection causes general debilitation that can seriously affect children’s school attendance and performance, and the work capacity and productivity of adults. In the Eastern Mediterranean Region, 11 countries are free of schistosomiasis, i.e. no local transmission occurs, and another 7 countries have limited local transmission. The disease is endemic in only 4 countries: Egypt, Somalia, Sudan and Yemen. In our Region, and elsewhere, the introduction of praziquantel in the 1980s as part of integrated control programmes demonstrated the effectiveness of the drug. Our goal is to facilitate this progress, and expand the disease-free areas as far as possible. In addition to the 11 countries that are already disease-free, the 7 countries that have limited transmission at present will soon be disease-free. Many people living in the 4 countries where schistosomiasis is still endemic will also soon en- joy life without this old disease. Source: Annual Report 2004: the way forward: Division of Communicable Disease Control. Cairo, World Health Organization Regional Office for the Eastern Mediterranean, 2005. 20. Hartnett T, Heneveld W. Statistical indica- tors of female participation in education in sub-Saharan Africa. Washington, DC, Africa Technical Department, The World Bank, 1999 (AFTHR Tech Note No. 7). 17 School non-enrolment.pmd 12/20/2005, 2:36 PM401 402 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Active trachoma, face washing (F) and environmental improvement (E) in a high-risk population in Oman R. Khandekar,1 R. Mabry,2 K. Al Hadrami3 and N. Sarvanan3 1Eye and Ear Health Care; 2Directorate-General of Health Affairs, Ministry of Health, Muscat, Oman (Correspondence to R. Khandekar: rajshpp@omantel.net.om). 3Department of Public Health, Dhakhiliya, Ministry of Health, Oman. Received: 10/07/03; accepted: 30/11/03 ABSTRACT Oman aims to eliminate blinding trachoma by 2010. As a part of a study to review the health institution approach of trachoma surveillance, “F” (facial cleanliness) and “E” (environmental improvement) components of the SAFE trachoma control strategy were assessed in a high-risk population in Nizwa wilayat. Thus 50 households with 1 member treated for active trachoma in the last 6 months were evaluated for active trachoma and for water and sanitation facilities. In all, 229 children under 15 years of age were clinically examined; the rate of active trachoma was 3.5% in the children. Clean face was found in 97.8% of the children and 70% of houses had piped water or water supplied by tankers. The status of “F” and “E” in the study area was generally high. Trachome actif, nettoyage du visage (N) et changement de l’environnement (C) dans une population à haut risque à Oman RÉSUMÉ Oman entend éliminer le trachome cécitant d’ici 2010. Dans le cadre d’une étude destinée à examiner l’approche de la surveillance du trachome appliquée par les établissements de santé, les com- posantes « N » (nettoyage du visage) et « C » (changement de l’environnement) de la stratégie de lutte contre le trachome dite stratégie CHANCE ont été évaluées dans une population à haut risque de la Wilaya de Nizwa. Une évaluation du trachome actif ainsi que de l’eau et des installations d’assainissement a donc été réalisée dans cinquante foyers comptant un membre ayant été traité pour un trachome actif au cours des six derniers mois. En tout, 229 enfants de moins de 15 ans ont subi un examen clinique ; le taux de trachome actif était de 3,5 % chez les enfants. Une bonne hygiène faciale a été observée chez 97,8 % des enfants et 70 % des maisons avaient l’eau courante ou de l’eau fournie par des camions-citernes. La situation des composantes « N » et « C » dans la zone étudiée était généralement bonne. 18 Active trachoma.pmd 12/20/2005, 2:36 PM402 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 403 Introduction Trachoma is a communicable eye disease affecting 146 million people worldwide, mainly in developing countries [1]. It is one of the leading infectious causes of blind- ness [2]. To control trachoma, the World Health Organization recommends adopting the SAFE trachoma control strategy in na- tional health programmes [3]. “S” indicates surgery for lid complications and “A” indi- cates antibiotics for active trachoma treat- ment. Both are widely used strategies and adequate information on the outcome of these strategies is available [4]. However, less is known about the contribution of “F” indicating face washing and “E” indicating environmental improvement in the overall impact of the SAFE strategy. Reliable com- munity-based information on these issues is needed so that control of many water- and fly-related diseases can be integrated [5]. Oman was a trachoma-endemic coun- try in the past. Rapid socioeconomic im- provement and implementation of SAFE trachoma control strategies have resulted in the decline of active trachoma from 70%– 80% in the late 1970s [6] to 2.2% in 1997 [7]. Oman aims to eliminate blinding tra- choma by 2010 [8]. To achieve this, tra- choma control initiatives should be further strengthened with special emphasis on the high trachoma-endemic areas. Nizwa wilayat is located in Dhakhiliya region of Oman. The climate is usually hot and dry throughout the year with occasion- al rain once or twice a year. Nizwa wilayat has a population of 61 000 Omanis, of whom 27 500 are children under 15 years of age [9]. Since 1983, Nizwa has been a high tra- choma-endemic area and both antibiotic treatment and lid surgery campaigns have been periodically undertaken and reviewed by the Ministry of Health and the World Health Organization [6]. One diagnostic centre and one regional hospital provide secondary eye care to patients of the Nizwa wilayat and 7 other wilayat of the region. These units are staffed by 6 opthalmolo- gists and 3 refractionists who use modern diagnostic and management tools. Thus fa- cilities for addressing ‘A’ and ‘S’ compo- nents of trachoma control are adequate. However, improvements in the socioeco- nomic situation, although visible, have not been assessed using indicators such as bet- ter water sanitation facilities and changes in community attitude to safe ocular hygiene practices. Hence, as a part of a study to review the health institution approach of trachoma surveillance, the assessment of “F” and “E” components of the trachoma initiative was undertaken in Nizwa. This was a commu- nity-based health assessment study. We present here a part of this study to describe the situation of active trachoma in children < 15 years, their “F” face washing status and “E” environmental improvement status of the houses in the study area. Based on the outcomes, reorganization of trachoma control strategies is proposed. Methods Between June and August 2002, active tra- choma cases were detected and treated in the ophthalmic units. The houses of all 50 active trachoma cases diagnosed during this period and the resident members in these houses were the study population. All the households were visited by a field in- vestigator from October to November 2002, and water and sanitation facilities were assessed as well as facial cleanliness practices. 18 Active trachoma.pmd 12/20/2005, 2:36 PM403 404 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 The house visits were scheduled in the afternoons so that most of the family mem- bers could participate in the study. The family members were not told about the assessment of “F” and “E” component pri- or to the visit. Two regional supervisors with at least 10 years experience of trachoma screening activities were the field investigators. Half of the houses were randomly selected and visited by one supervisor and the rest were visited by the other supervisor. One female nurse assisted the investigators while ex- amining female members of the house- holds. A meeting was held with field staff to discuss the procedures to ensure uniform and standard methodology and data collec- tion. The procedures were tested in 5 hous- es in a village outside the study area. During the pilot, the skills to assess tracho- ma and environmental status of both the field investigators were compared with that of an ophthalmologist and with each other. They were found to be uniform and satis- factory. Children < 15 years of age residing in these houses were examined for active tra- choma. Ophthalmic loupe and torchlight were used for eye examination. To cover those members of the family who were ab- sent (who were mainly schoolchildren), the same team examined them in school. Per- sonal details of family members and tra- choma status [trachomatous follicular (TF) and trachomatous intense (TI) in children, trachomatous scarring (TS) in all ages and trachomatous trichiasis (TT) in 35-year- olds and over] were noted on a standard pre-tested form. Children with active tra- choma were treated with supervised oral azithromycin. All other house members were also given prophylactic azithromycin treatment. The head of the household accompa- nied the field investigator when staff in- spected different parts of the house to note the water and sanitation situation. A closed- ended questionnaire was used to note ob- servations about the different components such as quantity and quality of water, type, availability and cleanliness of latrine, clean- liness of kitchen and house, method of gar- bage disposal, presence of faecal material in and around house, number of flies in a defined area of the kitchen and around the latrine, and the presence of pets and do- mesticated animals and the location and condition of animal pens. The field staff imparted health educa- tion on water, sanitation and trachoma dur- ing the house visit. Trachoma grading recommended by the World Health Organization was used in our study [10]. A clean face for a child was defined as absence of nasal discharge and dirt on the face, and absence of flies on the face. Water supply was considered ade- quate if water was available for 24 hours within the compound and was of a quality the family members considered fit for face washing. Animal keeping practice was con- sidered safe and hygienic if there was no animal pen within 10 metres of where fam- ily members were residing. Sanitation in a house was considered good if a closed la- trine was available for a family within the compound, human or animal faecal materi- al was absent in and around house and the fly density was low (< 5 flies/square metre) in the compound and kitchen or on a child’s face. Garbage disposal habits focused mainly on the disposal of leftover food. Data management system The data collection forms of each house as well as each person were entered using Mi- crosoft Excel spread sheet. A unique identi- fication code was given to the house and each family member and the files were sub- sequently merged. Univariate data analysis 18 Active trachoma.pmd 12/20/2005, 2:36 PM404 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 405 was performed. The frequencies and per- centage proportions were presented with the statistical validation using 95% confi- dence intervals (CI). Results All the 50 houses were visited and as- sessed. Of the 510 residents in these hous- es, 236 children (121 male and 115 female) were under 15 years of age. Seven of them were absent (3 male and 4 female) and could not be examined. Of the remaining 229, 68 (29.7%) were preschoolers, 107 (46.7%) were primary-school students and 38 (16.6%) were preparatory students. Ten children were studying in private schools and 2 were mentally challenged and were not attending school. The information on education of 4 children was missing. The active trachoma rate was 3.5% (95% CI: 1.1 to 5.9), all were cases of TF. The active trachoma rate was 5.4% (95% CI: 1.2 to 9.6) in females and 1.6% in males (95%CI: –0.6 to 4.0). It was less common in children under 5 years. The sample size does not rule out the element of chance in the observed rates of active trachoma in different subgroups (Table 1). Infective conjunctivitis was found in 8 (3.6%) out of 225 children examined for this condition. Table 2 gives information of the water and sanitation situation and facilities. Of 226 children, 221 (97.8%) had clean faces while 5 children (2.2%) had dirty faces. Information on 3 children was missing (Ta- ble 2). The type of water supply in the house is given in Table 2. Almost half of the houses had piped water supply and about a quarter each had water supplied by tanker or from a well. Availability of water for face washing was considered adequate in the study area. Of the 50 houses, 48 had private, personal and functional latrines. In 9 houses (18.0%) the investigators ob- served the presence of solid waste in and Table 1 Active trachoma in children by demographic characteristics Variable No. Active trachoma 95% CI examined No. % Sex Male 118 2 1.7 –0.6 to 4.0 Female 111 6 5.4 1.2 to 9.6 Age group (years) < 5 60 0 0.0 – 5 to 9 81 3 3.7 –0.4 to 7.8 10 to 14 88 5 5.7 0.8 to 10.5 School stage Preschool 68 0 0.0 – Primary school 107 5 4.7 0.7 to 8.7 Preparatory school 38 3 7.9 –0.7 to 16.5 Other 12 0 0.0 – Missing 4 – – – Total 229 8 3.5 1.1 to 5.9 CI = confidence interval. 18 Active trachoma.pmd 12/20/2005, 2:36 PM405 406 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 around house. On inquiry, it was reported that small children do not use latrine facili- ties and delay in disposal of waste resulted in this observation. In all 44 houses had an animal pen but it was more than 10 metres away from the kitchen area. The animals were properly fenced and were not found in the house except in 3 houses. As regards fly density, in 39 (78%) houses, flies were not found in the kitchen and in 10 (20%) houses the fly density was less than 5 flies per square metre area in the kitchen. Infor- mation was missing for 1 house. In 6 (12%) houses garbage was found in the house and in 5 (10%) houses garbage was found dispersed on the streets around the houses. The water and sanitation situation in houses with at least 1 child with active tra- choma was compared to houses without a child with active trachoma (Table 3). No significant differences in sanitation facili- ties were observed between these 2 groups. Discussion Oman is aiming to eliminate blinding tra- choma by 2010. One of the criteria to certi- fy a country as having achieved this goal is to remove any pockets of active trachoma with a rate of more than 5%. Information on the “F” and “E” situation in the high risk group of our study is crucial for the pre- vention of blindness programme so that any needed reorganization can be made in order to achieve this goal. The population we examined was more at risk for active trachoma compared to the general population since in each household 1 family member had active trachoma for which he/she had been treated in last 6 months. However, even in this vulnerable group, the active trachoma rate was less than 5%. The clinical grading of active tra- choma is easy to follow and by using expe- rienced health staff, the chances of misclassification were minimized in our study. Therefore, active trachoma does not seem to be a significant public health prob- lem in the study area. However, studies Table 2 Water and sanitation situation in houses of the study area Indicators of “F” and “E” No. % Facial cleanliness (n = 226 children < 15 years) Clean 221 97.8 Not clean 5 2.2 Water source of the house (n = 50 houses) Pipeline 24 48.0 Tanker 11 22.0 Falaj (underground water/ irrigation system) 1 2.0 Well 14 28.0 Sanitation in and around the house (n = 50 houses) Functional latrine Present 48 96 Absent 2 4 Animal pen 10 metres away from house 44 88 No animal pen or animal in the house 6 12 Fly density (1 missing) < 5 flies/m2 39 80 > 5 flies/m2 10 20 Solid waste in and around house Present 41 82 Absent 9 18 Garbage disposal in house Satisfactory 44 88 Not satisfactory 6 12 Garbage disposal in street Satisfactory 45 90 Not satisfactory 5 10 18 Active trachoma.pmd 12/20/2005, 2:36 PM406 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 407 conducted in different seasons and using larger samples could further confirm this observation. The active trachoma rate in children un- der15 years in Oman was reported to be 1.3% in 1996 [7]. In the same study, it was 2.6% in the Dhakhiliya region. Our study was in one wilayat of the same region and we found a much higher rate (3.5%) com- pared with this study. This could be due to the inclusion of a high-risk group in our study. Sharing the same environment, hav- ing similar hygiene habits and cross infec- tion during the latent period could be the reason for the active trachoma cases found in these households. The annual school screening activity in Nizwa wilayat targeting first primary- school students during 2001–2002 revealed an active trachoma prevalence of 1.1% [9]. Our study also suggests a low rate of active trachoma in primary-school students. Higher rates observed in children attending preparatory school suggest the need for periodic review of active trachoma in this age group. The absence of infection in pre- schoolers in this high-risk population needs further investigation. In view of the low prevalence in the study area, clinical finding of active tracho- ma should be complemented by laboratory tests as proposed by other studies [11]. A high proportion of the children in the study area had clean faces. The arrival of guests may have prompted the mother to make their children presentable and thus may have resulted in an over-estimation of the clean face rate due to social desirability bias. However, some of the children were screened in school and their facial cleanli- ness status was similar to those present in the house. This indicates the probable mar- ginal effect of this bias on the study results. The availability of water and the status of environmental sanitation was also better than that reported in other high trachoma- endemic areas of developing countries [12,13]. Low rates of communicable eye diseases such infective conjunctivitis in our sample confirms the low transmission po- Table 3 Active trachoma in children according to water and sanitation situation (“F” and “E”) Variable Active trachoma Chi-squareda Absent Present (P-value) No. No. Clean face (n = 226) Present 202 19 0.22 (0.64) Absent 5 0 Water source Easy access (pipeline/tanker) 21 12 1.7 (0.2) Far from house (falaj and well) 7 9 Sanitation facilities Excellent 22 3 15.01 (< 0.0001) Less satisfactory 8 16 aMantel–Haenszel method of chi-squared calculation. 18 Active trachoma.pmd 12/20/2005, 2:36 PM407 408 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 tential in these houses. Nevertheless, in spite of all efforts, social desirability bias might have influenced the study results causing an over-estimation of positive san- itation situation. Active trachoma in children was in- versely associated with the good sanitation and water facilities in the houses (Table 3). However, since our study was descriptive and the prevalence of active trachoma was less than 5%, the association between ac- tive trachoma and good water and sanita- tion should be viewed with caution. The ongoing screening of schoolchil- dren and prevention of transmission by prophylactic treatment of the family mem- bers using a supervised azithromycin regi- men could have reduced active trachoma. However, improved sanitation and water facilities could sustain these achievements. The multiple benefits of further strengthen- ing the “F” and “E” strategy in other areas, such as diarrhoea control and reduction in other health problems like contagious skin diseases, should be considered when addi- tional resources are allocated to improve water and sanitation components even in areas with low incidence of trachoma. The Integrated Management of Childhood Ill- nesses (IMCI) in trachoma-endemic areas should include trachoma control as one component as a cost-effective and sustain- able initiative for many health problems in children. The monitoring of such a strategy could also be within IMCI rather than as part of the programme management of tra- choma control. The environmental situation in houses with and without active trachoma cases among children was not significantly dif- ferent. The low prevalence rate of active trachoma and the epidemiological transition being experienced by the population of the study area could be responsible for this ob- servation. The national trachoma control initiative should undertake a similar exercise to eval- uate the status of “F” and “E” components in other areas of the country and accord- ingly review the implementation policies. Acknowledgements We acknowledge the consent and support extended by the regional and national health administrators to the field staff for the study. The active participation of the family members of the active trachoma cases in the study made the study a success. References 1. Thylefors B. The World Health Organiza- tion’s programme for the prevention of blindness. International ophthalmology, 1990, 14:211–9. 2. Taylor KI, Taylor HR. Distribution of azithromycin for the treatment of tra- choma [commentary]. British journal of ophthalmology, 1999, 83:134–5. 3. Bailey R, Lietman T. The SAFE strategy for the elimination of trachoma by 2020: will it work? Bulletin of the World Health Organization, 2001, 79(3):233–6. 4. Thylefors B, Negrel AD. Developments for a global approach to trachoma con- trol. Revue internationale du trachome et de pathologie oculaire tropicale et subtropicale et de santé publique, 1994, 71:63–7, 69–77. 5. Pruss A, Mariotti SP. Preventing tra- choma through environmental sanita- tion: a review of the evidence base. Bulletin of the World Health Organiza- tion, 2000, 78(2):258–66. 18 Active trachoma.pmd 12/20/2005, 2:36 PM408 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 409 6. Thylefors B. Review of Prevention of Blindness Programme in Oman. Geneva, World Health Organization, 1992:4 (Unpublished). 7. Ministry of Health. Oman Blindness Sur- vey Report. Oman Eye Study. Muscat, Oman, Mazoon Printers:28–31. 8. Ministry of Health. Eye health care manual, 2nd ed. Muscat, Oman, Golden Printing Press, 2000:16–9. 9. Ministry of Health. Annual health statisti- cal reports, 2002. Muscat, Oman, Minis- try of Health, 2002:31. 10. Primary health care level management of trachoma. Geneva, World Health Or- ganization, 1993:12–4. 11. Lietman TM et al. Clinically active tra- choma versus actual Chlamydial infec- tion. Medical journal of Australia, 2000, 172:93–4. 12. Emerson PM et al. Review of the evi- dence base for the “F” and “E” compo- nents of the SAFE strategy for trachoma control. Tropical medicine & international health, 2000, 5(8):515–27. 13. West S et al. Impact of face-washing on trachoma in Kongwa, Tanzania. Lancet, 1995, 345(8943):155–8. Alliance for Global Elimination of Trachoma by 2020 The World Health Organization leads an international alliance of in- terested parties to work for the global elimination of trachoma, the Alliance for Global Elimination of Trachoma by 2020 (GET 2020). The Alliance supports and collaborates with WHO in carrying out essen- tial activities such as epidemiological assessment, including rapid assessment and mapping, project implementation, coordination, and monitoring, disease surveillance, project evaluation and re- source mobilization. It is open to all parties – governments, interna- tional organizations and nongovernmental organizations – that are willing and ready to contribute to international efforts. These efforts are based on the WHO-developed strategy – a combination of inter- ventions known by the acronym “SAFE” which stands for Surgery for trichiasis (inturned eyelashes), Antibiotics, Facial cleanliness and Environmental improvement. WHO and its partners are supporting implementation of the SAFE strategy in the endemic countries that decided to eliminate trachoma as a result of political engagement and technical capability. Further information about WHO’s work to eliminate trachoma can be found at: http://www.who.int/pbd/blind- ness/trachoma/en/index.html 18 Active trachoma.pmd 12/20/2005, 2:36 PM409 410 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Risk factors for bacterial vaginosis in women attending a hospital in Kerman, Islamic Republic of Iran T. Ashraf-Ganjoei1 1Department of Obstetrics and Gynaecology, College of Medicine, Kerman University of Medical Sciences, Kerman, Islamic Republic of Iran (Correspondence to T. Ashraf-Ganjoei: tashrafganjoei@yahoo.com). Received: 16/12/03; accepted: 24/04/04 ABSTRACT Bacterial vaginosis (BV) is the most prevalent form of vaginal disturbances in women of child- bearing age. This study examined the prevalence and risk factors of BV of 130 non-pregnant women attending a hospital of Kerman University of Medical Sciences. Prevalence was 37.7%. Patients with BV had significantly lower educational and socioeconomic levels. They were also more likely to be smokers. The prevalence of BV was higher in patients with a history of abortion but this was not statistically significant. Women using oral contraceptive pills were at a decreased risk of BV; this was statistically significant. There were significant positive associations between BV and a history of vaginal infection, preterm delivery and premature rupture of the membranes. As BV appears quite prevalent in our patients, and considering the various complications associated with it, screening and treatment of high-risk women is suggested. Facteurs de risque de vaginose bactérienne chez des femmes consultant dans un hôpital de Kerman (République islamique d’Iran) RÉSUMÉ La vaginose bactérienne est la forme la plus courante de trouble vaginal chez la femme en âge de procréer. Cette étude a examiné la prévalence et les facteurs de risque de la vaginose bactérienne chez 130 femmes non enceintes consultant dans un hôpital de l’Université des Sciences médicales de Kerman. La prévalence s’élevait à 37,7 %. Les patientes présentant une vaginose bactérienne avaient un niveau d’instruction et un niveau socio-économique significativement plus faibles. Elles étaient aussi plus suscepti- bles d’être des fumeuses. La prévalence de la vaginose bactérienne était plus élevée chez les patientes ayant des antécédents d’avortement mais cela n’était pas statistiquement significatif. Les femmes utilisant des pilules contraceptives présentaient un risque réduit de vaginose bactérienne ; ceci était statistiquement significatif. Il y avait des associations positives significatives entre la vaginose bactérienne et des antécé- dents d’infection vaginale, d’accouchement prématuré et de rupture prématurée des membranes. La vaginose bactérienne semblant assez fréquente chez nos patientes, et compte tenu des diverses complica- tions qui y sont associées, le dépistage et le traitement des femmes à haut risque sont recommandés. 19 Risk factors for bacterial.pmd 12/20/2005, 2:36 PM410 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 411 Introduction Bacterial vaginosis (BV) is a change in vag- inal ecosystem where the normally domi- nant lactobacilli are greatly reduced and replaced with a number of other organ- isms, predominantly anaerobic flora, such as Gardnerella vaginalis, Mycoplasma ho- mimis, Mobiluncus spp., Baceteroides spp. and Peptostreptococcus spp. Both anaer- obes and G. vaginalis are normal inhabit- ants of the vagina, but overgrowth of the normal lactobacillus-dominant flora by these bacteria results in bacterial vaginosis [1]. BV is the most prevalent form of vaginal disturbances in women of childbearing age [2]. The average incidence of BV varies and is reported to be 10%–35% in patients visiting gynaecological clinics, 10%–30% in patients visiting obstetric clinics and 20%–60% in patients visiting services of sexually transmitted diseases [3]. BV has been associated with many gynaecological and obstetrical complications such as cer- vicitis, salpingitis, endometritis, postpar- tum infection, urinary tract infection, pelvic inflammatory disease, mild abnormal Pap smear results and possible link with cervical intraepithelial neoplasms, preterm delivery, premature rupture of the mem- branes (PROM), chorioamnionitis and post-partum endometritis [3,4]. A typical clinical symptom of BV is a thin, homogenous, gray, malodorous vagi- nal discharge, without significant pruritus or pain [5]. However more than 50% of all women with BV are asymptomatic. Diag- nosis of BV is established by Amsel’s crite- ria whereby 3 of the 4 following signs are found: homogenous discharge, vaginal flu- id pH > 4.5, positive amine test and micro- scopic analysis of Gram stained smear of vaginal discharge with clue cells seen [6,7]. Vaginal culture of G. vaginalis has no place in the evaluation of asymtomatic women because this organism is carried by 30%– 40% of women who do not have bacterial vaginosis [8]. The treatmant of BV consists of metronidazole or clindamycin orally or intravaginally. A study in southern Islamic Republic of Iran indicated a high prevalence of BV [9]. Thus we decided to determine the preva- lence and risk factors of BV in a university hospital Kerman. Methods In 2002 (January–December) a cross- sectional study was carried out of all non- pregnant patients who presented with a complaint of vaginal discharge to the Gy- naecology Clinic of Bahonar Hospital, Ker- man University of Medical Sciences. A senior resident recorded each patient’s his- tory and made a pelvic examination during which vaginal specimens were collected for laboratory testing. Demographic and clinical data recorded included: age, weight, height, occupation, educational level, current marital status, smoking sta- tus, contraceptive use, parity and obstetric history. Diagnosis of BV was established by Amsel’s criteria (presence of 3 of the following 4 symptoms or signs: homoge- nous vaginal discharge, positive whiff test after addition of 10% potassium hydroxide, vaginal pH > 4.5 and presence of clue cells). Fisher exact test, Student t-test and chi- squared test were calculated in the analysis. P > 0.05 was considered as a limit of sig- nificance. Odds ratios with 95% confi- dence intervals (CI) were also computed. 19 Risk factors for bacterial.pmd 12/20/2005, 2:36 PM411 412 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Results In this study, 130 non-pregnant women visiting the Gynaecology Clinic of Bahonar Hospital in 2002 were examined for BV. BV was diagnosed in 49 women (37.7%) (Ta- ble 1). BV was not correlated with age, body mass index, current marital status and parity (P > 0.05). Patients with BV had sig- nificantly lower educational (P = 0.006) and socioeconomic (P = 0.021) levels (Ta- ble 2). There was a statistically significant association between BV and smoking (P = 0.033). The prevalence of BV was higher in patients who had a history of abortion than in women who did not report such history, but the difference was not statistically significant (P = 0.07). There were positive associations between the oc- currence of BV and history of vaginal in- fection (P < 0.001), history of preterm delivery (P < 0.001) and history of PROM (P = 0.028) (Table 2). There was a statisti- cally significant association between lack of use of contraceptive pills and BV (P = 0.035). BV was more prevalent in women who used intrauterine devices than in other women (71.4% versus 38.3%), al- though this difference was not statistically significance (P = 0.091) (Table 3). Discussion Bacterial vaginosis is the most common cause of vaginal discharge in women of child-bearing age. The prevalence of BV differs in various populations (11%–71%) [10]. In the present study the prevalence of BV was 37.7%. This is less than that re- ported from northern Islamic Republic of Iran (61.7%) [11], but similar to findings of studies in Denmark in 2002 [12], Jordan in 2001 [13] and Indonesia in 2001 [14]. BV has been clearly linked to adverse reproductive and gynaecological outcomes in women. Thus many studies have been conducted on the prevalence of BV in preg- nant women. Various rates have been re- ported: 66% for BV or candidiasis or both in Sofia in 1999 [15], 20.6% in Kenya in 1996 [16], 38.5% in India in 2001 [17] and 44.1% in Jamaica in 2000 [18]. These re- sults on pregnant women are similar to our result on non-pregnant women. Table 1 Clinical and laboratory findings in the women with and without bacterial vaginosis Clinical and laboratory Bacterial vaginosis P-value findings Present Absent (n = 49) (n = 81) % % Specific vaginal discharge 100 34.6 < 0.001 Vaginal burning and pruritus 77.6 79.0 0.505 Urinary symptoms 55.1 23.5 < 0.001 Presence of clue cell 14.3 0.0 < 0.001 Positive whiff test 100 33.3 < 0.001 Vaginal pH [Mean (SD)] 6.73 (1.00) 4.06 (0.91) < 0.001 SD = standard deviation. 19 Risk factors for bacterial.pmd 12/20/2005, 2:36 PM412 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 413 The complexity of the sociodemo- graphic picture challenges the field of pub- lic health to continue to explore the role of BV and its relationship to a whole host of social and biomedical conditions that may contribute to adverse health outcomes among society’s most vulnerable members. These factors include, for example, educa- tion level, socioeconomic status, contra- ceptive use and reproductive history. According to one study in 2001 on African–American women, education and contraceptive use were associated with BV. Adjusted odds ratios for BV remained sig- nificant in the women with 13 or fewer years of education (OR = 5.5, 95% CI: 2.1–14.5) and hormone use within the past 6 months (OR = 0.5, 95% CI: 0.2–0.8) [19]. Systemic contraceptives appeared protective whereas little education was linked to an increase in prevalence; these results are similar to our study. Culhance et al. in 2001 showed sociodemographic vari- ables and behavioural characteristics such as use of illicit drugs were significantly as- sociated with the presence of BV [20]. Hellberg et al. in 2000 showed that, after adjustment for possible confounding fac- tors, smoking was significantly associated with BV [21]. Age-adjusted odds ratio for smokers was 2.3 before and 3.0 (95% CI: 1.3–6.9) after adjustment for sexual risk behaviour, reproductive history and alcohol use. In our study the odds ratio for smok- ers was similar at 3.76 (CI: 1.07–13.22). Baeten et al. showed that users of oral contraceptive pills or depot medroxy- progesterone acetate were at increased risk for acquisition of chlamydia and at de- creased risk for BV [22]. Our results were similar. On the other hand, several studies have suggested that BV is more common among intrauterine device users than non- users [12,23]. Our study demonstrated similar results. Bacterial vaginosis is strongly associat- ed with cervicitis and pelvic inflammatory disease, and upper genital tract infection is an important factor for preterm birth and PROM. Preterm birth is one of the most common causes of prenatal morbidity and mortality. Thus BV is associated with ad- verse pregnancy outcomes across all ges- tational ages, such as first and second trimester fetal loss, chorioamnionitis, low- Table 2 Odds ratio for bacterial vaginosis associated with some related factors Variable Odds 95% ratio confidence interval Education levela 3.80 1.68–8.64 Current smokingb 3.76 1.07–13.22 Single (at present) 0.27 0.06–1.28 History of abortionb 1.83 0.88–3.80 History of vaginitisb 6.27 2.05–19.21 History of preterm birthb 11.18 3.50–35.69 History of premature rupture of the membranesb 2.36 1.06–5.24 aEducation level was considered low if only up to primary school had been completed. bAs reported by the women. P < 0.05 was considered statistically significant. Table 3 Odds ratios for bacterial vaginosis associated with contraceptive method Contraceptive Odds 95% method ratio confidence interval No contraception 1.04 0.50–2.16 Oral contraceptive use 0.37 0.14–0.99 Barrier use 2.07 0.44–9.71 Intrauterine device use 4.03 0.75–21.70 Tubal ligation 0.48 0.16–1.43 19 Risk factors for bacterial.pmd 12/20/2005, 2:36 PM413 414 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 birth weight infants and maternal/neonatal infectious mortality. Liahi Camp et al. con- cluded that in women who had had at least one late miscarriage, BV was twice as common as in women who had had only early losses [24]. In the present study the prevalence of BV in women who had histo- ry of at least one abortion was higher than other women, but the difference was not significant. BV was quite prevalent in our sample of non-pregnant women and there is no rea- son to think it would be less prevalent in pregnant women in our community. Given the complications of BV, especially in preg- nancy, it is a problem which the public health authorities need to address and one which should be brought to the attention of the public. References 1. Smayevsky J et al. Vaginal microflora as- sociated with bacterial vaginosis in non- pregnant women: reliability of sialidase detection. Infectious diseases in obstet- rics and gynecology, 2001, 9(1):17–22. 2. Hay P. Recurrent bacterial vaginosis. Current infectious diseases reports, 2000, 2(6):506–12. 3. Dadhwal V. Bacterial vaginosis. Obstet- rics and gynecology communication, 2000, 2(5):24–8. 4. Purwar M et al. Bacterial vaginosis in early pregnancy and adverse pregnancy outcome. Journal of obstetrics and gynaecology research, 2001, 27(4): 175–81. 5. Eschenbach DA. Pelvic infections and sexually transmitted diseases. In: Scott JS, Disaia PJ, Spellacy WA, eds. Danforth’s obstetrics & gynecology, 8th ed. Philadelphia, Lippincott Williams& Wilkins, 1999. 6. Amsel R et al. Nonspecific vaginitis. Di- agnostic criteria and microbial and epi- demiologic associations. American journal of medicine, 1983, 74(1):14–22. 7. Navarrete P et al. Evaluation of Nugent and Amsel criteria for the diagnosis of bacterial vaginosis. Revista médica de Chile, 2000, 128(7):767–71. 8. Murray PR et al. Medical microbiology. St Louis, CV Mosby, 1990:210. 9. Keshavarz H et al. Risk factors for and relationship between bacterial vaginosis and cervicitis in a high risk population for cervicitis in southern Iran. European journal of epidemiology, 2001, 17(1):89–95 10. Georgijevic A, Cjukic-Ivancevic S, Bujko M. Bakterijska vaginoza. Epidemiologija i faktori rizika. [Bacterial vaginosis. Epi- demiology and risk factors.] Srpski arhiv za celokupno lekarstvo, 2000, 128(1– 2):29–33. 11. Moussavi Z, Behrouzi R. Diagnostic Amsel criteria compared to standardized method of Gram stain for the diagnosis of bacterial vaginosis. International Con- gress Series, 2004, 1271:392–5. 12. Schmidt H, Hansen JG. Bacterial vaginosis in a family practice population. Acta obstetricia et gynecologica scandinavica, 2000, 79(11):999–1005. 13. Abu Shaqra QM. Bacterial vaginosis among a group of married Jordanian women: occurrence and laboratory diag- nosis. Cytobios, 2001, 105(408):35–43. 14. Joesoef MR et al. High rate of bacterial vaginosis among women with intrauter- ine devices in Manado, Indonesia. Con- traception, 2001, 64(3):169–72. 15. Chervenkova A et al. Kliniko-mikrobio- logichno prouchvane na bakterialnite vaginozi i vaginiti pri bremenni zheni. [A 19 Risk factors for bacterial.pmd 12/20/2005, 2:36 PM414 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 415 clinical and microbiological study of bac- terial vaginosis and vaginitis in pregnant women.] Akusherstvo I Ginekologiia, 1999, 38(1):33–6. 16. Thomas T et al. Identifying cervical infec- tion among pregnant women in Nairobi, Kenya: limitations of risk assessment and symptom-based approaches. Geni- tourinary medicine, 1996, 72(5):334–8. 17. Mathew R et al. Prevalence of bacterial vaginosis in antenatal women. Indian journal of pathology and microbiology, 2001, 44(2):113–6. 18. Kamara P et al. Vaginal infections in pregnant women in Jamaica: prevalence and risk factors. International journal of STD and AIDS, 2000, 11(8):516–20. 19. Holzman C et al. Factors linked to bacte- rial vaginosis in nonpregnant women. American journal of public health, 2001, 91(10):1664–70. 20. Culhane JF et al. Maternal stress is asso- ciated with bacterial vaginosis in human pregnancy. Maternal–child health jour- nal, 2001, 5(2):127–34. 21. Hellberg D, Nilsson S, Mardh PA. Bacte- rial vaginosis and smoking. International Journal of STD and AIDS, 2000, 11(9): 603–6. 22. Baeten JM et al. Hormonal contraception and risk of sexually transmitted disease acquisition: results from a prospective study. American journal of obstetrics and gynecology, 2001, 185(2):380–5. 23. Hodoglugil U et al. Intrauterine device use and some issues related to sexually transmitted disease screening and oc- currence. Contraception, 2000, 61(6): 359–64. 24. Llahi-Camp JM et al. Association of bac- terial vaginosis with a history of second trimester miscarriage. Human repro- duction, 1996, 11(7):1575–8. 19 Risk factors for bacterial.pmd 12/20/2005, 2:36 PM415 416 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 La toxicomanie féminine en milieu universitaire S. Kjiri,1 F. Boulayoun,1 I. Rammouz,1 I. Cherkaoui2 et J.E. Ktiouet1 1Service de Psychiatrie B, Hôpital psychiatrique universitaire Arrazi, Salé (Maroc) (Correspondance à adresser à S. Kjiri : swkjiri@yahoo.fr). 2Observatoire régional d’Épidémiologie, Rabat (Maroc). Reçu : 13/07/03 ; accepté : 24/02/04 RÉSUMÉ Dans un pays de tradition arabo-musulmane comme le Maroc, le sujet de la toxicomanie est tabou et l’est encore davantage quand il concerne la population féminine. Les étudiantes marocaines, de par leur niveau intellectuel et leur bonne insertion sociale, nous sont apparues comme pouvant représenter une population cible intéressante pour approcher la problématique de la toxicomanie féminine. C’est à partir d’une enquête basée sur un auto-questionnaire anonyme que nous avons tenté d’aborder cette question. Notre étude concerne un échantillon de 1208 étudiants, dont 744 femmes et 464 hommes. Les résultats montrent que 6,59 % de femmes contre 36,2 % d’hommes consomment des substances toxiques. Par ailleurs, nous avons repéré 2 % de cas d’abus et 1,5 % de cas de dépendance chez les consommatrices, selon les critères diagnostiques du DSM-IV. Enfin, nous nous sommes interrogés sur les éventuels facteurs de risque et il s’est dégagé, entre autres, que le lieu de résidence en cité universitaire et que le niveau de revenus peuvent de façon statistiquement significative avoir un impact sur la consommation de substances toxiques. Female drug abuse in a university setting ABSTRACT Morocco is a country with Muslim-Arab tradition. In this sociocultural context, drug abuse in women is very taboo. We tried to approach the subject of drug use in women in a cross-sectional study of a sample of 1208 students (744 women and 464 men) who completed an anonymous self-administered questionnaire. We found 6.59% of women and 36.2% of men used drugs. Using DSM IV criteria, drug dependence was found in 2% of the subsample of female drug users and drug abuse in 1.5%. Place of residence and level of income were statistically significant risk factors for drug use. 20 La toxicomanie fÈminine.pmd 12/20/2005, 2:36 PM416 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 417 Introduction L’étude que nous avons réalisée s’intéresse à la toxicomanie féminine en milieu univer- sitaire à Rabat au Maroc. Le Maroc est un pays de tradition arabo-musulmane où le thème de la consommation de drogues reste encore un sujet tabou. Cependant nul n’ignore que le cannabis est cultivé dans certaines régions du Nord du pays et ce, malgré le projet actuel d’un programme de substitution de la culture de cannabis par d’autres cultures qui soient aussi rentables. Pour l’instant, cette substance reste encore d’accès facile compte tenu de sa disponibi- lité et de son prix peu élevé. Par ailleurs, l’alcool est un produit librement vendu et consommé bien que sa consommation soit interdite pour les Marocains musulmans. En fait, il semble exister un véritable para- doxe dans l’attitude générale face aux substances dites toxiques. La consomma- tion existe mais on n’en parle pas. Pour la femme arabo-musulmane, de par son statut particulier dans ce type de société, ce thème est encore plus difficile à aborder. Nous avons tenté dans ce travail d’approcher cette problématique de plus près. Par rapport à la question de la popula- tion cible, notre choix s’est porté sur la population estudiantine. L’étudiant se ca- ractérise, de façon générale, par son esprit d’ouverture, sa curiosité, sa coopération dans les travaux de recherche. Par ailleurs, les cités universitaires dans lesquelles nous avons réalisé notre enquête sont fréquentées par des étudiants jeunes, du premier cycle universitaire, venant des di- verses régions du Royaume. Les objectifs de ce travail consistent à comparer le comportement des jeunes étudiantes ma- rocaines face aux drogues avec celui des étudiants masculins et, dans un second temps, de décrire les modalités de consom- mation des drogues au sein de la population estudiantine féminine. Il s’agit d’estimer le taux d’abus et de dépendance aux drogues au sein de cette dernière population et enfin de dégager les éventuels facteurs de risque. Méthodologie Pour la réalisation de cette étude nous avons mis au point un questionnaire. Outre les questions d’ordre général concernant l’âge, la situation matrimoniale, le type d’études, le montant et l’origine des reve- nus, ce questionnaire anonyme visait à re- chercher la notion de consommation de substances toxiques, les modes de con- sommation et à repérer d’éventuels trou- bles en rapport avec la consommation à type de dépendance ou d’abus selon les critères du DSM-IV (Manuel diagnostique et statistique des troubles mentaux) [1].Ce manuel est un outil de référence en psy- chiatrie considéré comme une des classifi- cations internationales des maladies men- tales incontournable pour la recherche. Le questionnaire utilisé dans notre étude a été traduit en langue arabe et était donc dis- ponible en français et arabe. Il a été au préalable validé dans les deux versions au- près de 40 étudiants. Les enquêtes ont été réalisées dans trois cités universitaires différentes, dont une exclusivement pour les étudiantes et deux mixtes. Les questionnaires étaient distribués en- tre 11 h et 15 h dans les réfectoires, et ceci après autorisation des directeurs des cités. Une équipe constituée de psychiatres et de psychologues travaillant dans la même institution psychiatrique s’est mobilisée sur les trois sites, trois jours différents. Les questionnaires étaient remis individuelle- ment aux étudiants avec des explications quant aux objectifs de l’enquête et au strict respect de l’anonymat. Ainsi, après leur remplissage, les questionnaires étaient dé- posés dans un sac disposé au fond de la 20 La toxicomanie fÈminine.pmd 12/20/2005, 2:36 PM417 418 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 salle par les étudiants eux-mêmes. Les don- nées recueillies à partir des questionnaires ont été par la suite encodés sur Epi Info et analysées grâce à des études statistiques : Anova, khi2, test Mann-Whitney ou Wil- coxon. Résultats Population de l’étude Elle comprend 1208 étudiants dont 744 femmes, soit 61,6 %, et 464 hommes, soit 34,8 %. Le niveau moyen d’études est de 3 années universitaires. Comme le montre le tableau 1, l’alcool est la substance la plus consommée chez 12,4 % de notre population d’étude. Il est suivi de près par le cannabis chez 11,6 % des étudiants. Les psychotropes (benzo- diazépines essentiellement prises en au- tomédication pour des problèmes d’insomnie ou d’anxiété), quant à eux, con- cernent 3,6 % des étudiants. Résultats quantitatifs selon le sexe Le tableau 2 montre une consommation plus importante de tabac chez les hommes que chez les femmes. La différence est statistiquement significative. Les substances prises ensemble sont largement plus consommées chez les hom- mes. La différence est statistiquement si- gnificative (Tableau 2). En regardant de plus près, la consom- mation d’alcool et de cannabis est nette- ment plus importante chez les hommes, de façon statistiquement significative. L’inverse s’observe au sujet des psycho- tropes, 42,9 % des femmes contre 11,9 % des hommes, de façon également statis- tiquement significative (Tableau 3). Recherche des critères d’abus et de dépendance de substances toxiques au sein des deux sous- groupes de population Pour rechercher la notion d’abus de substances toxiques, nous avons utilisé les critères diagnostiques du DSM-IV [1]. Nous trouvons ce trouble chez 15,5 % des consommateurs masculins contre 2 % seulement des consommatrices féminines (Tableau 4). Cette différence est hautement significative sur le plan de l’analyse statis- tique. La recherche de troubles à type de dépendance aux substances toxiques re- pose également sur les critères diagnos- tiques du DSM-IV [1]. Les résultats sont proches, quoique légèrement plus faibles, de ceux obtenus pour les abus. Ainsi nous trouvons 10,8 % de situations de dépen- dance chez les hommes contre 1,5 % chez les femmes avec la même signification statistique (Tableau 4). Modalités de consommation des substances toxiques chez les femmes Parmi la population d’étudiantes consom- mant des substances toxiques, 24,5 % esti- ment que c’est par curiosité qu’elles ont rencontré pour la première fois la Tableau 1 Substances toxiques consommées chez l’ensemble des étudiants (tabac exclu) Substances Nombre % par rapport à l’ensemble des étudiants Alcool 151 12,5 Cannabis 141 11,7 Psychotropes 44 3,6 Solvants 12 1,0 Cocaïne 12 1,0 Héroïne 5 0,4 Autres 9 0,7 20 La toxicomanie fÈminine.pmd 12/20/2005, 2:36 PM418 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 419 Tableau 2 Consommation de tabac et de substances toxiques selon le sexe Substance Hommes (n = 464) Femmes (n = 744) Consommation de tabac Nombre % Nombre % Oui 160 34,4 39 5,2 Non 295 63,5 684 91,9 Pas de réponse 9 1,9 21 2,8 Consommation de substances toxiques Oui 168 36,2 49 6,6 Non 286 61,6 631 84,8 Pas de réponse 10 2,2 64 8,6 Test khi2 p < 0,001. Tableau 3 Consommation des différentes substances selon le sexe Substance Hommes (n = 168) Femmes (n = 49) Analyse Nombre % Nombre % statistique Alcool p < 0,001 Oui 123 73,2 24 49 Non 45 26,8 25 51 Cannabis p < 0,001 Oui 120 71,4 15 30,6 Non 48 28,6 34 69,4 Psychotropes p < 0,001 Oui 20 11,9 21 42,9 Non 148 88,1 28 57,1 Solvants p < 0,05 Oui 9 5,4 3 6,1 Non 159 94,6 46 93,9 Cocaïne p = 0,88 Oui 9 5,40 3 6,10 Non 159 94,6 46 93,9 Héroïne p = 0,68 Oui 3 1,8 2 4,1 Non 165 98,2 47 95,9 Autres p < 0,05 20 La toxicomanie fÈminine.pmd 12/20/2005, 2:36 PM419 420 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 substance ; 22,4 % incriminent la survenue de problèmes en précisant qu’il s’agissait soit de problèmes d’ordre familial soit suite à une rupture sentimentale. La recherche de plaisir est évoquée chez 18,4 % des étu- diantes, la lutte contre des sentiments d’anxiété chez 10,2 %. Viennent ensuite la timidité dans 6,1 % des cas, les difficultés relationnelles dans 2 % des cas. Interrogées sur les circonstances dans lesquelles s’est faite la première consom- mation, 36,7 % n’ont pas donné de réponse. Pour 26,5 % des consommatri- ces, elle s’est faite en compagnie d’amis ; pour 16,3 %, elles étaient seules ; pour 14,3 %, la rencontre a eu lieu lors d’une soirée entre amis ; 4,1 % ont goûté à la substance en compagnie d’un membre de leur famille et 2 % en compagnie de leur petit ami. Différents symptômes cliniques ont été recherchés au travers du questionnaire. Il s’agissait de repérer, d’une part, s’il existait une notion de souffrance psychologique avant la rencontre avec la substance toxique et d’autre part, de relever l’éventuel impact de l’usage des substances sur les symptômes qui auraient existé au préalable. Le recours aux substances toxiques est donc pour ces étudiantes une solution pour lutter contre une souffrance comme le montre le tableau 5, d’autant plus qu’il semblerait que certains symptômes régres- sent de façon importante après l’usage des toxiques, en particulier en ce qui concerne les troubles du sommeil et l’anxiété. Tableau 4 Nombre de cas d’abus et de dépendance selon le sexe Hommes Femmes Test khi2 (n = 464) (n = 744) Nombre % Nombre % Abus de toxiques p < 0,001 Oui 72 15,5 15 2,0 Non 392 84,5 729 98,0 Dépendance aux toxiques p < 0,001 Oui 50 10,8 11 1,5 Non 414 89,2 733 98,5 Tableau 5 Symptômes décrits avant et après le début de la consommation chez les étudiantes Symptômes décrits Avant la Après la consommation (%) consommation (%) Insomnie 18 des consommatrices 10,2 Perte d’appétit 14,3 16,2 Dépression 14,3 10,2 Anxiété 12,2 2,0 Tentative de suicide 6,1 2,0 Troubles sexuels 8,2 2,1 20 La toxicomanie fÈminine.pmd 12/20/2005, 2:36 PM420 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 421 Recherche des facteurs de risque de consommation de substances toxiques dans la population féminine Il y a 77,1 % de fumeuses qui consomment également des substances toxiques contre 3,1 % (Tableau 6). Le tabac représente donc un facteur de risque majeur pour la consommation de toxiques. L’accès plus facile au tabac favoriserait la rencontre avec d’autres produits toxiques. La signifi- cation statistique est également retrouvée lorsqu’on analyse les sous-groupes d’étudiantes présentant les critères d’abus et de dépendance aux substances. A noter que 70 étudiantes n’ont pas répondu à l’une ou l’autre question (parfois aux 2) et n’ont donc pas pu être intégrées à l’étude statis- tique. L’analyse du revenu moyen des étu- diantes montre que les consommatrices de substances toxiques ont un revenu moyen supérieur à celui des non-consommatrices (Tableau 7). La différence est statistique- ment significative. Elle l’est également lorsque l’on compare au sein du sous- groupe des consommatrices les étudiantes réunissant les critères d’abus et les autres. Par contre, on ne retrouve pas de dif- férence statistiquement significative quand il s’agit du groupe des étudiantes présentant les critères de dépendance. Le lieu de résidence des étudiantes a également été analysé. Il ressort que les étudiantes habitant en cité universitaire consomment plus de substances toxiques que celles qui résident à l’extérieur du cam- pus, en l’occurrence au sein de leur famille. La différence est statistiquement significa- tive. Elle l’est également concernant les sous-groupes d’étudiantes présentant aussi bien les troubles d’abus que de dépen- dance. Par contre, on ne retrouve pas de différence statistiquement significative lorsque l’on compare les étudiantes qui résident en cité universitaire mixte et celles en cité exclusivement féminine. Par contre, parmi les étudiants masculins, il n’a pas été retrouvé de différence statistiquement significative selon le lieu de résidence, cam- pus universitaire ou domicile familial. A no- ter que 80 étudiantes n’ont pas précisé leur lieu de résidence. Une différence statistique hautement significative est retrouvée entre les con- sommatrices dont des membres de la famille sont également des consommateurs (14,7 %) et celles dont aucun membre de l’entourage familial ne consomme de pro- duits toxiques (Tableau 8). A noter que 64 étudiantes n’ont pas précisé la notion de consommation de toxiques au sein de leur famille. Une corrélation positive a été re- trouvée entre l’usage de substances toxiques et la mauvaise assiduité en cours (p = 0,002). Aucune corrélation n’a pu être établie entre l’usage de toxiques et les échecs universitaires ni avec le type Tableau 6 Relation entre consommation de tabac et consommation des autres substances toxiques Consommation Consommation Pas de consommation de toxiques de tabac (n = 35) de tabac (n = 639) Nombre % Nombre % Oui 27 77,1 20 3,1 Non 8 22,9 619 96,9 Test khi2 p < 0,001. 20 La toxicomanie fÈminine.pmd 12/20/2005, 2:36 PM421 422 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 d’études universitaires. On note que 25,5 % des consommatrices ont eu au moins une fois un contact avec la psychia- trie contre 8,3 % des non-consommatri- ces. Cette différence est statistiquement significative. Discussion Dans notre étude, 6,59 % des étudiantes consomment des produits toxiques (nico- tine non comprise) contre 36,2 % des étu- diants masculins. En ce qui concerne le tabac, 5,2 % des étudiantes fument contre 34,4 % des hommes. Parmi les substances consommées, l’alcool vient en première position pour les sous-groupes féminins (49 %) et masculins (73,2 %). En deuxième position, alors que pour les hommes c’est le cannabis (71,4 %), pour les femmes ce sont les psychotropes (42,9 %). Juste derrière, viennent le cannabis pour les femmes (30,6 %) et les psychotropes pour les hommes avec un pourcentage beaucoup plus faible que celui de la gent féminine (11,9 %). Nos résultats rejoignent les observa- tions décrites en général concernant la dis- tribution des substances toxiques en fonction du sexe, en particulier en ce qui concerne la consommation de psycho- tropes plus élevée chez les femmes. [1] Par comparaison aux études réalisées au sein des populations estudiantines, des dif- férences similaires ont été observées dans d’autres études marocaines s’intéressant à la nicotine, bien que les méthodologies Tableau 7 Relation entre la consommation de substances toxiques et les revenus Consommation de Revenus en euros substances toxiques Oui 100 ± 48,7 Non 74 ± 68,4 Différence 26 Test Mann-Whitney ou Wilcoxon p= 0,003479. Tableau 8 Relation entre la consommation de substances toxiques, le lieu de résidence et l’usage de toxiques dans l’entourage familial Variable Consommation de substances toxiques Oui Non Nombre % Nombre % Résidencea En cité universitaire 38 8,2 428 91,8 En famille 7 3,5 191 96,5 Usage de toxiques dans la familleb Oui 33 14,7 192 85,3 Non 16 3,5 439 96,5 a Test khi2 p = 0,0457. b Test khi2 p = 0,001. 20 La toxicomanie fÈminine.pmd 12/20/2005, 2:36 PM422 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 423 soient différentes. Certaines différences sont à souligner cependant. Une étude ca- sablancaise [2] avait trouvé des chiffres plus élevés – 44 % de consommateurs con- tre 15 % de consommatrices – alors qu’une étude à El Jadida [3] trouvait des chiffres plus bas – 21,7 % d’hommes con- tre 2,3 % de femmes. Casablanca, Rabat et El Jadida sont trois villes dont le niveau d’industrialisation et d’occidentalisation est fort différent. Ceci expliquerait peut-être également les résultats relevés respective- ment dans les milieux universitaires de ces trois villes, Casablanca étant la plus mo- dernisée et El Jadida la plus conservatrice. Dans une étude réalisée dans un autre pays arabo-musulman, en l’occurrence l’Égypte [4], une grande différence est re- trouvée concernant l’usage de tabac entre étudiants hommes (52,2 %) et femmes (9 %). En Occident, des études chez les étu- diants montrent des différences moins marquées comme une étude américaine [5] (37,9 % d’hommes contre 27,7 % de femmes) ou parfois des différences in- versées comme une étude française [6] (24,2 % d’hommes contre 25,5 % de femmes) ou une étude polonaise [7] (23,6 % d’hommes contre 26,3 % de femmes). Concernant les autres substances toxiques, une étude française en milieu uni- versitaire [6] montre une consommation plus importante d’alcool chez les étu- diantes, soit 19,6 % de femmes par rapport à l’ensemble des étudiantes et autant pour les psychotropes. L’incidence générale de la toxicomanie en milieu universitaire dans notre étude est de 21,39 % contre 30 % dans l’étude française [6], 33,1 % dans une étude bri- tannique [8] et 45,9 % dans l’étude polonaise [7]. Les comparaisons entre ces études ne peuvent être que très approximatives du fait des différences méthodologiques et des limites de toutes les études basées sur des enquêtes. De plus, comme cela a été abordé en introduction, le degré de sincérité dans les réponses aux questionnaires peut en ef- fet varier d’une région du monde à l’autre en fonction du contexte socio-culturel dans lequel s’inscrit le rapport à la problématique des drogues. Concernant la consommation de substances comme la cocaïne et l’héroïne, elle intéresse 6,1 % des consommatrices de notre étude pour la cocaïne et 4,1 % pour l’héroïne. Cette dernière substance semble être plus utilisée dans les pays asiatiques auprès de la population féminine comme l’indique l’étude de l’OMS de 1993 [9]. Des différences de disponibilité des substances et de culture explique donc ces résultats. Les résultats de notre étude mettent l’accent sur les symptômes psy- chologiques qui peuvent précéder la con- sommation de toxiques, comme les troubles de la sphère instinctuelle, sommeil, appétit, sexualité ou les troubles anxio- dépressifs. Par ailleurs, les tentatives de suicide sont loin d’être négligeables. Enfin, les taux d’abus et de dépendance aux toxiques indiquent la nécessité de réfléchir à une politique de prévention auprès de ces jeunes adultes. D’autre part, l’analyse des facteurs de risque possibles peut représenter une première piste par rapport à un tel projet. En effet, il ressort très clairement que chez les femmes la consommation de tabac, les revenus élevés, la résidence en cité univer- sitaire et la consommation de toxiques au sein des familles sont des facteurs de risque à la consommation, à l’abus et à la dépen- dance des substances toxiques. 20 La toxicomanie fÈminine.pmd 12/20/2005, 2:36 PM423 424 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Conclusion Dans notre étude, 6,5 % des étudiantes consomment des substances toxiques, tabac exclu. Quant à ce dernier produit, il concerne 5,2 % des étudiantes. Nos résul- tats se rapprochent de ceux observés dans les régions de même tradition socio- culturelle et se distinguent de ceux des pays occidentaux par des taux plus faibles et une prédominance qui reste très nette entre le comportement des hommes et des femmes face aux substances toxiques. Les troubles liés à la consommation des substances toxiques représentent 2 % d’abus et 1,5 % de dépendance au sein de notre population d’étude. Par ailleurs, des symptômes de souffrance psychologique ont été rapportés chez les consommatrices ; aussi ces der- nières consultent-elles plus souvent en psy- chiatrie que les non-consommatrices. Cette étude en milieu universitaire mon- tre que la population féminine est loin d’être épargnée par les avatars de la rencontre avec les drogues. Une politique de préven- tion visant l’information et l’éventuelle orientation des jeunes adultes en souffrance psychologique et des étudiantes à risque serait utile à envisager. Références 1. American Psychiatric Association. DSM- IV. Manuel diagnostique et statistique des troubles mentaux, 4e édition (ver- sion internationale, Washington DC, 1995). Masson, Paris, 1996. 2. Fahmi M. Tabagisme chez les étudiants résidant à la cité universitaire de Casablanca [Thèse de Médecine]. Casablanca, 1993 :96. 3. Bentalha I. Tabagisme en milieu universitaire à El Jadida [Thèse de Médecine]. Casablanca, 2001:242. 4. Gadalla MA, Gabal MS, Khello AK: When and why Ain Shams university students started smoking? Journal of the Egyp- tian Public Health Association, 1992, 67:275–90. 5. Rigotti NA, Lee JE, Wechler H. College students’ use of tobacco products. Re- sults of a national survey. Journal of the American Medical Association, 2000, 284:699–705. 6. Prieto N et al. Santé mentale en milieu étudiant : attitude face aux toxiques. La Revue Française de Psychiatrie et de Psychologie Médicale, 1999, 25:9–11. 7. Chodorowski Z. et al. [The evaluation of smoking and alcohol consumption by university students in Gdansk.] Przeglad lekarski, 2001, 58(4):272–5. 8. Pickard M et al. Alcohol and drug use in second years medical students at the University of Leeds. Medical education, 2000, 34(2):148–50. 9. Hsu LN, du querny J. Towards a gender- sensitive approach to drug demand re- duction: a process within the United Nations system. Bulletin of narcotics, 1995, 47(1–2):1–14. 20 La toxicomanie fÈminine.pmd 12/20/2005, 2:36 PM424 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 425 Chlorine residual efficiency in inactivating bacteria from secondary contamination in Isfahan, 2002 A. Tavakoli,1 R. Yazdani,1 M.R. Shahmansouri2 and B.N. Isfahani 2 1Department of Microbiology; 2Department of Environmental Health, Isfahan University of Medical Sciences, Isfahan, Islamic Republic of Iran (Correspondence to A. Tavakoli: ak_ta_ir@yahoo.com). Received: 05/10/03; accepted: 22/03/04 ABSTRACT To investigate the inactivation of potential pathogens, we evaluated survival rates for Escherichia coli, Enterococcus faecalis, Aeromonas hydrophila and Salmonella typhi in samples taken from the Isfahan drinking water system. Chlorine residual, pH, temperature and total organic carbon levels were measured. The organism most sensitive to chlorine was A. hydrophila. It was inactived in < 100 minutes at chlorine levels of 0.11 mg/L to 0.90 mg/L. The other 3 organisms showed higher resistance. E. coli tolerated 0.30 mg/L chlorine for > 1000 minutes while Ent. faecalis and S. typhi survived at total chlorine concentration of 0.50 mg/L for 100 minutes. We concluded that total chlorine levels of less than 0.71 mg/L in water supply systems cannot provide the recommended safety levels. Efficacité résiduelle du chlore pour inactiver les bactéries dans le cas d’une contamination secondaire à Ispahan, 2002 RÉSUMÉ Afin d’examiner l’inactivation d’agents pathogènes potentiels, nous avons évalué le taux de survie d’Escherichia coli, Enterococcus faecalis, Aeromonas hydrophila et de Salmonella typhi dans des échantillons prélevés dans le système d’eau potable d’Ispahan. On a mesuré le taux de chlore résiduel, le pH, la température et le taux de carbone organique total. A. hydrophila était le micro-organisme le plus sensible au chlore. Il était inactivé en moins de 100 minutes à un taux de chlore de 0,11 mg/L à 0,90 mg/L. Les 3 autres micro- organismes présentaient une plus forte résistance. E. coli tolérait 0,30 mg/L de chlore pendant plus de 1000 minutes tandis que Ent. faecalis et S. typhi survivaient à une concentration totale de chlore de 0,50 mg/L pendant 100 minutes. Nous avons conclu qu’une concentration totale de chlore de moins de 0,71 mg/L dans les systèmes d’approvisionnement en eau ne permet pas d’assurer le niveau de sécurité recommandé. 21 Chlorine residual efficiency.pmd 12/20/2005, 2:36 PM425 426 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Introduction Supplying safe drinking water has a signif- icant effect in protecting and improving public health. Epidemiologic investigations of waterborne gastrointestinal diseases re- veal that the water supply distribution sys- tems can play a decisive role in the health of the inhabitants in a region. The annual risk of enteric illness ranges from 1 per 1000 people to 1 per 100 people, and a sig- nificant proportion of those illnesses may be caused by organisms in water [1]. Worldwide, the risk of illness from water- borne microbial pathogens has been report- ed as being “probably thousands- to millions-fold greater than that from chemi- cal contaminants in drinking water” [2]. A 1993 study indicated that approximately 1000 people (800 children, 200 adults) worldwide were dying every hour from waterborne microbial disease. Most of the increased risk ensues because of localities in which treatment of the domestic water supply is inadequate or non-existent [3]. The protection of water quality in the distribution system has, therefore, received constant emphasis and much attention. Al- though the source water may be safe, once flowing in the distribution mains, it may be detrimentally affected from a variety of sources. Contaminants may enter through sources such as network maintenance and repair operations, illegal connections, shut- ting and/or opening main valves, abrupt changes in consumption levels, power breaks and pumping stoppage, pressure drops and backflows from homes into the distribution mains. It is, therefore, essential to maintain a recommended level of a stable disinfectant, such as chlorine, in the flow in order to provide safety and to prevent secondary biological contamination by agents which enter the system after the water has been disinfected for drinking purposes [4]. It has long been known that a free chlo- rine residual (chlorine that must be present in the distribution system after the disinfec- tion process to prevent secondary contam- ination) of 0.75 mg/L in the distribution system at pH ≤ 8.0 and 1 NTU (nephelom- etric turbidity units) can destroy most gas- trointestinal bacteria in less than 30 minutes [5]. The standard proposed and practised for microbial control of water in the Islam- ic Republic of Iran includes the removal/ absence of faecal coliform bacteria. The presence of these bacteria is, therefore, used as an indicator of water contamination [6]. Secondary contaminants entering the water distribution system are normally controlled through the addition of chlorine and maintaining a certain level of free chlo- rine residual. The effect of the chlorine re- sidual depends on the chemical and physical conditions as well as on the length of contact time. Although chlorine residual greatly contributes to the inactivation of in- dicator bacteria, i.e. faecal coliforms, the question awaiting an answer is the level of inactivation of other potential pathogens such as Enterococcus faecalis, Aeromonas hydrophila and Salmonella typhi at the rec- ommended levels of chlorine residual. The aim of this investigation is to find an answer to this question, focussing on pathogens in the drinking water distribution system in Isfahan, which conveys water from the Isfahan treatment plant to Greater Isfahan [7]. Methods This investigation was carried out during the 10 months from January through Octo- ber 2002. We took samples from 30 differ- ent points along the municipal drinking 21 Chlorine residual efficiency.pmd 12/20/2005, 2:36 PM426 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 427 water distribution system in Isfahan [8]. Two samples were later discarded because of the confusing results obtained from the microbial examination. We took 5 samples from each site in 5-litre Erlenmeyer flasks using the grab sampling method (samples were taken at specified points in time and place and with known volume). Four of the filled containers for each site were tightly capped with aluminium foil and paraffin wax and delivered to the microbiology lab- oratory in Isfahan University of Medical Sciences for testing. The water in the fifth flask was used at the sampling location to determine total chlorine content using the DPD method [9,10]. About 150 mL of this sample was deliv- ered to the laboratory in a small, opaque glass container tightly capped with alumin- ium foil and sterilized in a dry-heat sterilizer oven for 4 hours to determine total organic carbon (TOC). The microorganisms selected for inves- tigation were E. coli, Ent. faecalis, A. hy- drophila, and S. typhi, all strains used in laboratory quality control. They were sup- plied by the Iranian Pasteur Institute, Te- hran. In order to study the survival patterns, several colonies of bacteria re- cently cultured in specially prepared culture media (brain heart infusion agar, Merck) were mixed with sterile saline and stirred using a vortex mixer in order to prepare a homogeneous stock suspension with 1 × 108 cells/mL (determined by spectropho- tometer). Two serial dilutions (1:9) were done, then 1 mL of the suspension was added to 1 of the flasks from each location (i.e. each mL of sample water contained a known number of bacteria). The survival rates of the bacteria in the prepared samples were determined at 1 minute, 10 minutes, 100 minutes, and 1000 minutes using the serial dilution method from 1 mL of the sample water cultured on brain heart infusion agar incubated at room temperature. All tests had 1 positive control sample (sterile water + the microorganism under study) and 1 negative control in which a separate culture of the water sample with- out the addition of the test organism was cultured in an enriched medium (blood agar, Merck) and in differential media (Endo medium and desoxycholate citrate agar, Merck). The data were analysed using SPSS software. Graphs were plotted using Har- vard Graphics. Results Chlorine concentrations of the samples ranged from 0 mg/L to 0.90 mg/L. In 53.58% of the samples tested, total chlo- rine concentrations were ≤ 0.50 mg/L. (Figure 1). The pH of the samples was 8.0– 8.2 and sample temperatures were 10 °C– 29 °C. The TOC of the samples varied from 0.07 mg/L to 3.95 mg/L. The frequency distribution of TOC is shown in Figure 2. In samples with total chlorine concen- trations between 0.51–0.90 mg/L chlorine, E. coli was inactivated in 10–100 minutes. This organism remained active for over 1000 minutes in samples with total chlorine concentration of 0–0.30 mg/L (Figure 3). Survival time for A. hydrophila in all samples was less than 100 minutes. More- over, the organism was completely inacti- vated within 1 minute in samples with total chlorine concentration between 0.71 mg/L and 0.90 mg/L (Figure 4). Ent. faecalis showed resistance for over 1000 minutes in samples with total chlorine concentrations ≥ 0.50 mg/L. Greatest sensitivity was observed in sam- ples with total chlorine concentrations of 0.71–0.90 mg/L (Figure 5). 21 Chlorine residual efficiency.pmd 12/20/2005, 2:36 PM427 428 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 The survival time for S. typhi in samples with total chlorine concentrations ≥ 0.50 mg/L was more than 1000 minutes while in samples with chlorine levels of 0.71–0.90 mg/L, survival time was reduced to less than 100 minutes (Figure 6). In order to characterize and enumerate the original bacteria in the samples, water samples were used for direct bacterial cul- ture. Ten of the samples contained between 1 and 24 colonies of aerobic bacteria, in- cluding Gram-positive cocci, Gram-posi- Figure 1 Distribution of total chlorine concentration in water samplings Figure 2 Distribution of total organic carbon concentration in water samplings 21 Chlorine residual efficiency.pmd 12/20/2005, 2:36 PM428 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 429 Figure 3 Mean survival rates for Escherichia coli in chlorine at five different concentrations (both scales are logarithmic) Figure 4 Mean survival rates for Aeromonas hydrophila in chlorine at five different concentrations (both scales are logarithmic) 21 Chlorine residual efficiency.pmd 12/20/2005, 2:36 PM429 430 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Figure 5 Mean survival rates for Enterococcus faecalis in chlorine at five different concentrations (both scales are logarithmic) Figure 6 Mean survival rates for Salmonella typhi in chlorine at five different concentrations (both scales are logarithmic) 21 Chlorine residual efficiency.pmd 12/20/2005, 2:36 PM430 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 431 tive bacilli, or both. None of the bacteria used in our investigation were found in the raw water samples. The findings were analysed using the chi-squared test and assuming P < 0.05, the possible significant relationships be- tween the dependant variable (microorgan- ism survival) and the independent variable (total chlorine) was determined. The mini- mum concentration of total chlorine which ensured the safety of the drinking water in our samples was 0.71 mg/L. Discussion The recommended total chlorine concen- tration after an exposure time of half an hour under non-emergency conditions and at appropriate pH levels is 0.5–0.8 mg/L at the end-of-line points [9]. In our study, 53.58% of the samples had chlorine con- centrations below this level (under emer- gency conditions such as epidemics of gastrointestinal infection and natural disas- ters the recommended level is 1.0 mg/L, mainly because of the distance water has to travel from its point of origin, the treatment plant, to its point of consumption). Haas found that the chlorine residual in distribution systems should be maintained at no less than 0.5 mg/L since at this chlo- rine concentration secondary contaminants would be eliminated and heterotrophic growth prevented [11]. In our study, the chlorine concentration in half the samplings in the service area was not only below that required for emergency conditions but that also below the level rec- ommended for normal conditions. The pH levels of the samples ranged be- tween 8.0 and 8.2, which is within the rec- ommended range of 7.0–8.5 [9]. In our samples, TOC concentrations were > 2.0 mg/L in 17.9% of cases (Figure 2). Organic carbon can create a nutrient environment favourable to bacterial surviv- al and also contribute to the formation of chlorine by-products. Since the municipal drinking water distribution system is large and covers a vast service area, the effects of TOC will be amplified. It is, therefore, essential to reduce TOC levels to the maxi- mum recommended level of 2.0 mg/L (the recommendation in D/DBPS Act) [12]. The most chlorine-sensitive microor- ganism studied was A. hydrophila, which did not remain stable for more than 100 minutes in any of the samples. Thus, a vast range of chlorine concentrations are effec- tive on this bacteria and the presence of chlorine disinfectant in the water distribu- tion system at concentration 0.11–0.90 mg/L will suffice for the removal of sec- ondary contamination by this microorgan- ism in less than 100 minutes. In a previous study, Massa et al. found that A. hydrophi- la TW11 was never completely inactive at various concentrations (0.11–0.50 mg/L) and strain TW27 was generally more sensi- tive than strain TW11, showing a 99% higher mortality rate after 3 minutes con- tact at 0.5 mg/L chlorine [13]. Ozbas and Aytac found that concentrations up to 2.5 mg/L free chlorine had no effect on A. hy- drophila A306, even after 30 minutes expo- sure [14]. Cattabani registered the survival rate of 4 strains of A. hydrophila after an exposure of 10 minutes at a concentration of 0.3 mg/L free chlorine and inhibition af- ter 5 minutes at a concentration of 2.5 mg/L [15]. As this species belongs to the family Vi- bronaceae, its survival characteristics may be generalized to the genera in the family; however, this generalization must be taken cautiously until corroborated by further studies. E. coli showed a higher resistance to the disinfectant than A. hydrophila. In sam- ples with chlorine concentrations < 0.31 21 Chlorine residual efficiency.pmd 12/20/2005, 2:36 PM431 432 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 mg/L, E. coli survived for over 1000 min- utes. There was a significant correlation between E. coli survival rates and chlorine concentration (P < 0.05). Furthermore, E. coli lost its resistance in less than 1000 minutes at total chlorine concentrations 0.31–0.90 mg/L. In a previous study on 3 selected A. hydrophila strains, Knochel found that A. hydrophila E9 was the most resistant isolate in monochloramine solu- tion, followed by E. coli and the 2 other strains [16]. It has also been reported that E. coli was chlorine sensitive but that microorgan- isms such as Clostridium perfringens and viruses did not become inactive in most to- tal chlorine concentrations [17,18]. Our re- sults for E. coli are in agreement with these. Since in microbiological analyses of water, E. coli is in effect an indicator bac- terium, it is essential to bear in mind the fact that indicator bacteria are rapidly inac- tivated in water distribution systems and this may act as a limiting factor to drawing sound, realistic conclusions regarding the quality of water in epidemic outbreaks. Ent. faecalis also showed considerable resistance to low total chlorine concentra- tions, remaining active for more than 1000 minutes at chlorine concentrations < 0.51 mg/L but survival dropped at 0.51–0.90 mg/L (P < 0.05). It follows that the surviv- al rate for this bacterium over periods up to 1000 minutes depends on chlorine concen- tration. As this species is also considered an indicator of faecal contamination of wa- ter owing to its frequent occurrence in fae- cal waste, Ent. faecalis should be studied with regard to inactivation in terms of con- tact time, inactivation rate and chlorine concentration. Torkian reported that the presence of free chlorine disinfectant in the municipal water distribution system would help to remove secondary contaminants [19]. Our results are in agreement with this. Total chlorine concentration ≤ 0.50 mg/L led to reductions in the S. typhi popu- lation in drinking water but not to its total removal. It becomes inactive in < 1000 minutes at chlorine concentrations 0.51– 0.70 mg/L and in < 100 minutes at chlorine concentrations 0.71–0.90 mg/L. In the chi- squared test performed to determine the correlation between S. typhi survival rates and total chlorine concentration, it was found that P was < 0.05, and consequent- ly, the survival rate of this microorganism depends on total chlorine concentration over 1000 minute periods. An investigation to evaluate the inacti- vating power of chlorine residuals in water distribution systems indicated that E. coli remained active in drinking water samples over periods up to 1000 minutes. All other organisms tested in this study remained re- sistant to the disinfectant [5]. These find- ings conform with our own findings on microorganisms in the Isfahan water distri- bution system. Recommendations • The presence of chlorine disinfectant in water distribution systems is essential but the exposure time and chlorine con- centration are critical. • It is essential to bear in mind that, since E. coli is the basic faecal indicator in microbial investigations of water quali- ty, in waterborne outbreaks of epidem- ics chlorine residuals in water may inactivate indicator bacteria but not eliminate other important bacteria in water. • With regard to the spread of waterborne epidemics in the Islamic Republic of 21 Chlorine residual efficiency.pmd 12/20/2005, 2:36 PM432 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 433 Iran and reduced safety in water distri- bution systems, chlorine concentration must be constantly maintained at levels higher than 0.70 mg/L. References 1. Rose JB, Haas CN, Gerba C. Water- borne pathogens: assessing health risks. Health and environment digest, 1993, 7(3):1–3. 2. Haas CN et al. Risk assessment of virus in drinking water. Risk analysis, 1993, 13(5):545–52. 3. Shuval H. Institutional aspects of the management of water quantity and qual- ity on the shared transboundary water resources of the Jordan river basin. In: Stout GE, Al-Weshah RA, eds. Proceed- ings of the international symposium on water resources in the Middle East: policy and institutional aspects. Urbana, Illinois, International Water Resources Association, 1993. 4. Mäkelä A, Meybeck M. Designing a monitoring programme. In: Bartram J, Ballance R, eds. Water quality monitor- ing: a practical guide to the design and implementation of freshwater quality studies and monitoring programmes. London, E & FN Spon, 1996. 5. Payment P et al. A randomized trial to evaluate the risk of gastrointestinal dis- eases due to consumption of drinking water meeting current microbiological standards. American journal of public health, 1991, 81(6):703–8. 6. Drinking water physical, chemical, and microbial standards, 5th ed. Tehran, Ira- nian Industrial Research and Standards Organization Publications, 1998 (Stan- dard No. 1052) [in Farsi]. 7. Montgomery JM. Water treatment prin- ciples and design. New York, John Wiley & Sons Inc., 1985:262–83. 8. Standard methods for the examination of water and wastewater, 18th ed. Wash- ington DC, American Public Health As- sociation, American Water Works Asso- ciation, Water Environment Federation, 1992:1–21. 9. National drinking water standards. Tehran, Iranian Industrial Research and Standards Organization Publications, 1997 (Standard No. 1053, 5th ed.) [in Farsi]. 10. Standard methods for the examination of water and wastewater, 19th ed. Wash- ington DC, American Public Health As- sociation, American Water Works Association, Water Environment Fed- eration, 1995. 11. Haas CN. Benefits of using a disinfectant residual. Journal of the American Water Works Association, 1999, 91(1):65–9. 12. Kawamura S. Integrated design and op- eration of water treatment facilities, 2nd ed. New York, John Wiley & Sons Inc., 2000:29–30. 13. Massa S et al. Note: susceptibility to chlo- rine of Aeromonas hydrophila strains. Journal of applied microbiology, 1999, 86(1):169–73. 14. Ozbas ZY, Aytaç SA. Effect of chlorine on growth and survival of Aeromonas hydrophila and Yersinia enterocolitica in water. Chemie, Mikrobiologie, Techno- logie der Lebensmittel, 1994, 16:146– 50. 15. Cattabani, F. Susceptibility to disinfec- tants of Aeromonas hydrophila and Vibrio fluvialis. Archivio veterinario italiano, 1986, 37:65–73. 16. Knøchel S. Chlorine resistance of motile Aeromonas spp. Water science and technology, 1991, 24:327–30. 21 Chlorine residual efficiency.pmd 12/20/2005, 2:36 PM433 434 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Joint WHO-CEHA and ISESCO regional training course for trainers for sanitary inspection of public water supply systems The World Health Organization, Regional Centre for Environmental Health Activities (WHO/CEHA) organized a joint WHO-CEHA and Is- lamic Educational, Scientific and Cultural Organization (ISESCO) re- gional training course for trainers for sanitary inspection of public water supply systems in Amman, Jordan from 16 to 18 May 2005. The objectives of the training course were to: • provide training on the methods of sanitary inspections to avoid sources of pollution; • strengthen water quality monitoring; • minimize water pollution; • assess long-term planning for water resources; • minimize waterborne diseases; • highlight the use of sanitary inspection as an integral part of wa- ter quality surveillance. Participants from Egypt, Islamic Republic of Iran, Iraq, Jordan, Leba- non, Libyan Arab Jamahiriya, Morocco, Oman, Pakistan, Palestine, Saudi Arabia, Sudan, Syrian Arab Republic, Tunisia and Yemen par- ticipated in the course. 17. Hydes O. European regulations on re- sidual disinfection. Journal of the Ameri- can Water Works association, 1999, 91(1):70–4. 18. Snead MC et al. The effectiveness of chlorine residuals in inactivation of bac- teria and viruses introduced by post- treatment contamination. Water re- search, 1980, 14:403–8. 19. Torkian A. Environmental engineering, Vol. I. Tehran, Kankash Publications, 1995 [in Farsi]. 21 Chlorine residual efficiency.pmd 12/20/2005, 2:36 PM434 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 435 1Service de médecine préventive et d’épidémiologie, C.H.U. Fattouma Bourguiba, Monastir (Tunisie) (Correspondance à adresser à M. Letaief: mondher.letaief@fmm.rnu.tn). Reçu : 18/05/04, accepté : 19/09/04 RÉSUMÉ Afin d’évaluer la fréquence des admissions non pertinentes et d’identifier les facteurs qui leur sont associés, une étude descriptive a été conduite sur un échantillon aléatoire de 411 hospitalisations dans 3 hôpitaux régionaux. L’évaluation de l’adéquation des admissions a été faite en utilisant le protocole d’évaluation Appropriateness Evaluation Protocol (AEP). Vingt et un pour cent (21 %) [IC 95 % : 17-25] des admissions ont été jugées évitables selon l’AEP. Les hospitalisations non pertinentes étaient associées au lieu d’hospitalisation (p = 0,005), à l’âge du patient (p = 0,003), à la durée du séjour et au diagnostic (p < 0,01). Les principales raisons justifiant les admissions étaient la thérapie parentérale, la présence de troubles sensorimoteurs circulatoires ou respiratoires aigus ou progressifs entraînant une incapacité pour le patient et la présence de troubles électrolytiques sévères. Cette étude souligne le besoin d’une meilleure organisa- tion des hôpitaux. Dans ce cadre, il est nécessaire de développer des alternatives à l’hospitalisation. Analysis of the appropriateness of hospitalizations in Monastir, Tunisia ABSTRACT We aimed to measure the extent of inappropriate hospital admissions and to identify factors associated with inappropriate hospital use. A descriptive study was carried out on a randomized sample of 411 hospitalizations in 3 regional hospitals. The appropriateness of admissions was assessed using the Appropriateness Evaluation Protocol (AEP). We found 21% (95% CI: 17%–25%) of the admissions were avoidable according to the AEP. Inappropriate admissions were associated with the hospital (P = 0.005), patient age (P = 0.003), length of stay and diagnosis (P < 0. 001). The most frequent reasons for appropriate admissions were parenteral therapy, an acute or progressive sensory motor circulatory or respiratory condition sufficient to incapacitate the patient and severe electrolyte or blood gas abnormality. Our study highlights the need to improve hospital management and to develop alternatives to hospitalization. Analyse de la pertinence des hospitalisations dans la région de Monastir (Tunisie) M. Letaief1, K. Ben Salem1 , N. Hergli1, N. Riahi1 et M.S. Soltani1 22 Analyse de la pertinence.pmd 12/20/2005, 2:39 PM435 436 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Introduction L’hôpital occupe une part importante dans les dépenses de soins de santé. L’hospitalisation inappropriée génère un coût inutile se faisant au détriment de l’efficience et de la qualité des soins et pourrait exposer à des soins iatrogènes. La mise en place d’un programme d’amélioration de la qualité des soins à l’hôpital devrait, entre autres, prendre en compte des stratégies permettant la réduc- tion des hospitalisations inappropriées. Ces stratégies seraient développées après étude et analyse du processus d’hospitalisation. L’audit clinique est une méthode d’évaluation fréquemment utilisée dans l’amélioration de la qualité des soins [1,2]. Elle fournit une expertise clinique en com- parant une pratique de soins observée à une pratique de référence. Cette dernière est définie par des critères objectifs permettant de mesurer la conformité des procédures de soins à un référentiel [3]. En Tunisie, très peu d’études se sont intéressées à l’évaluation de la pertinence des hospitalisations. C’est dans ce cadre que s’inscrit cette étude ayant pour objec- tifs d’évaluer la pertinence des admissions et d’identifier les facteurs qui lui sont asso- ciés. Méthodes Nous avons procédé par une étude descrip- tive portant sur 412 hospitalisations dans les services de médecine interne de trois hôpitaux non universitaires de la région de Monastir. La population de l’étude est constituée par toutes les hospitalisations ayant eu lieu dans les hôpitaux non universitaires de la région de Monastir pendant la période allant d’avril 1998 à février 2000. À partir de cette population, nous avons effectué un échantillonnage aléatoire simple. La taille de ce dernier a été préalablement estimée en considérant une proportion d’hospitalisa- tions inadéquates de 30 % en se référant à d’autres travaux [4-6], une précision et un seuil de signification de 5 % [7]. Nous avons utilisé la méthode de l’audit clinique en adoptant comme référentiel l’Appropriateness Evaluation Protocol (AEP) (voir Encadré 1) [8]. Ce dernier est constitué d’un ensemble de critères expli- cites, définissant la pertinence d’une ad- mission par rapport à l’utilisation du plateau technique hospitalier, c’est-à-dire l’exis- tence d’un état clinique ou de procédures diagnostiques ou thérapeutiques lourdes nécessitant une surveillance médicale étroite. La variable dépendante était la perti- nence de l’admission, évaluée par référence aux critères de l’AEP. L’évaluation des hospitalisations était faite a posteriori. Les médecins devaient revoir les informations figurant sur le dossier du malade et ensuite classer chaque admission en pertinente ou non pertinente, selon la présence ou l’absence de critères mentionnés dans l’AEP. L’admission était classée appropriée si au moins l’un des 18 critères était rempli. Dès lors que l’on pourrait être confronté à des situations cliniques non décrites dans le protocole, les médecins pouvaient valider des admissions non pertinentes en utilisant l’override. Les décisions sur la justesse des admissions étaient basées sur les services fournis pendant les premières 24 à 48 heu- res. Les variables indépendantes com- prenaient les caractéristiques socio- démographiques des patients (âge, sexe, lieu d’hospitalisation, couverture sociale et antécédents d’hospitalisation) et les ca- ractéristiques du séjour (pertinence de l’admission, diagnostic et durée du séjour). 22 Analyse de la pertinence.pmd 12/20/2005, 2:39 PM436 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 437 L’analyse statistique a été réalisée en utilisant le logiciel SPSS [9]. Nous avons utilisé les odds ratio (OR) comme indices d’évaluation de l’association entre la va- riable dépendante et les variables indépen- dantes. Par ailleurs, nous avons ajusté les OR pour l’âge et calculé les intervalles de confiance à 95 % pour ces paramètres. Résultats Pertinence globale Parmi les 412 admissions, 88 ont été jugées non pertinentes, ce qui nous donne un taux de non-pertinence globale de 21 % avec un intervalle de confiance (IC) à 95 % de 17- 25. L’override n’a permis de justifier que 12 hospitalisations sur 88, soit 13,5 % de l’ensemble des admissions jugées au départ non pertinentes par l’AEP. Critères justifiant les admissions Le critère le plus fréquent était celui relatif à la thérapie parentérale (critère 12). Il avait, à lui seul, justifié 180 admissions (56 %). Le critère 9 (Troubles sensorimoteurs, cir- culatoires ou respiratoires aigus ou pro- gressifs entraînant une incapacité pour le patient) venait ensuite en justifiant 11 % des admissions. Le critère 12 avait justifié 85 % des ad- missions pour des affections génito- urinaires, 70 % des admissions pour des pathologies infectieuses ou signes anor- maux et la moitié des admissions secon- daires à des pathologies endocriniennes ou respiratoires. Le critère 9 ainsi que le critère 1 (Instal- lation brutale de troubles de la conscience ou désorientation) étaient surtout impliqués dans la justification des admissions secon- Encadré 1 Critères de AEP pour évaluer la pertinence des hospitalisations 1. Installation brutale de troubles de la conscience ou désorientation 2. Pouls < 50 ou > 140/mn 3. PAS < 90 ou > 200 mmHg ou PAD < 60 ou > 120 mmHg 4. Perte brutale de la vue ou de l’audition 5. Perte brutale de la capacité de faire bouger une partie du corps 6. Fièvre persistante ³ 38 °C depuis plus de 5 jours 7. Hémorragie active 8. Troubles électrolytiques (Na+, K+) ou gaz du sang (CO2, pH artériel) anormaux 9. Troubles sensorimoteurs, circulatoires ou respiratoires aigus ou progressifs entraînant un incapacité pour le patient. 10. Ischémie aiguë à l’électrocardiogramme (ECG) 11. Plaie ouverte ou éviscération 12. Thérapie intraveineuse et/ou perfusion 13. Chirurgie nécessitant une anesthésie générale ou régionale ou utilisation d’équipement ou procédures disponibles uniquement à l’hôpital 14. Monitoring des paramètres vitaux toutes les 2 heures ou plus 15. Chimiothérapie nécessitant une surveillance continue pour risque de réactions toxiques dangereuses 16. Traitement dans une unité de soins intensifs 17. Antibiothérapie par voie intramusculaire toutes les 8 heures 18. Assistance respiratoire continue ou intermittente au moins toutes les 8 heures 22 Analyse de la pertinence.pmd 12/20/2005, 2:39 PM437 438 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 daires à des affections neurologiques (34 %). Ces affections correspondaient dans 90 % des cas à des accidents ischémiques transitoires. Le critère 8 (Troubles électrolytiques [Na+, K+]) a justi- fié la moitié des admissions occasionnées par des pathologies endocriniennes, princi- palement le diabète. L’étude des admissions en fonction des critères de l’AEP et du lieu de l’hospitalisation nous a montré que le critère 12 a justifié la majorité des admis- sions dans les 3 hôpitaux et que le recours aux critères 8 et 9 a été surtout noté dans les hôpitaux 1 et 3. Facteurs associés à la non-pertinence Une variabilité dans la fréquence des admis- sions non pertinentes a été notée entre les différents hôpitaux (p = 0,005). En prenant l’hôpital 1 comme référence, la non-perti- nence était 2,78 fois plus importante à l’hôpital 2 (Tableau 1). La distribution des admissions non per- tinentes différait d’un groupe d’âge à l’autre (p = 0,01). Le taux de non-perti- nence le plus élevé (32 %) a été noté pour les patients âgés de moins de 30 ans, chez qui le risque d’avoir une admission non per- tinente était multiplié par 2,43 par com- paraison au groupe de référence (≥ 60 ans) (Tableau 1). En prenant comme référence les patho- logies respiratoires, la non-pertinence était 6,53 fois plus importante aussi bien pour les pathologies traumatiques que pour les signes et symptômes anormaux. Discussion Cette étude nous a permis d’étudier la perti- nence des hospitalisations dans les hôpitaux non universitaires de la région sanitaire de Monastir et d’identifier des facteurs qui lui sont associés. Le recueil des données a été fait de manière rétrospective. Cette méthode a été également adoptée par d’autres équipes [10-15]. Cette approche est plus simple de réalisation, de moindre coût et surtout ré- duit le biais d’observation. L’observation directe induirait un changement des pra- tiques et fausserait leur mesure. Par contre, l’approche rétrospective présente des limi- tes, surtout en relation avec la qualité des informations dans les dossiers médicaux [16]. La comparaison des caractéristiques générales de la population de l’étude avec celles de la population notifiées dans le re- gistre régional de morbidité et de mortalité hospitalière [17], pour la même période et pour les mêmes services, n’a pas montré de différence statistiquement significative, ce qui constitue un argument en faveur de la représentativité de l’échantillon. L’AEP définit la pertinence d’une hospi- talisation exclusivement par rapport à l’utilisation du plateau technique hospitalier, ce qui est en rapport avec l’existence d’un état clinique ou de procédures diagnos- tiques ou thérapeutiques nécessitant une surveillance médicale étroite. Cette défini- tion est assez restrictive et ne tient pas compte de l’ensemble des fonctions ac- tuellement assumées par l’hôpital [13]. Certains actes de soins pourraient en effet être délivrés par d’autres structures sani- taires. En l’absence de ces dernières ou d’autres alternatives, une décision médicale peut être de facto justifiée même si la journée n’est pas techniquement pertinente. De même, une hospitalisation médicale- ment inappropriée pourrait être justifiée et représenter, pour certains patients, la meilleure solution. A l’opposé, une hospita- lisation réellement non pertinente peut être 22 Analyse de la pertinence.pmd 12/20/2005, 2:39 PM438 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 439 Tableau 1 Facteurs associés à l’adéquation des hospitalisations Facteur Non Pertinentes OR IC95 % OR IC95 % ajustés p pertinentes ajustésa Nbre (%) Nbre Hôpital 0,005 1b 18 13,5 116 2 35 30 81 2,78 1,41 - 5,53 2,78 1,44 - 5,26 3 34 21 126 1,75 0,89 - 3,41 1,75 0,93 - 3,24 Sexe 0,646 Masculin 45 20 176 0,89 0,54 - 1,48 0,89 0,55 - 1,44 Fémininb 42 22 147 Âge (ans) 0, 003 ≤ 30 29 32 62 2,43 1,30 - 4,54 31-59 27 21 100 1,40 0,76 - 2,58 ≥ 60b 31 16 161 Antécédents 0,756 Absents 43 22 153 1,11 0,65 - 1,78 1,11 0,68 - 1,83 Présentsb 44 20 169 Couverture sociale 0,350 CNSS-CNRPS 50 19 210 0,70 0,33 - 1,43 0,70 0,35 - 1,41 Indigent 23 25 69 1,06 0,54 - 1,71 1,06 0,48 - 2,34 Payant-Autresb 14 26 40 Durée de séjour 0,000 1-7 j 68 30 159 3,33 2,01 - 6,52 3,33 1,89 - 5,92 8-21 jb 19 10 160 ≥ 21 j 0 0 4 * * Type de journée 0,330 Lun-Mar 29 23,6 88 1,39 0,75 - 2,84 1,39 0,74 - 2,61 Merc-Jeu-Vend 35 21 133 1,21 0,62 - 2,19 1,21 0,67 - 2,19 Sam-Dim + Fériéb 23 18 102 Diagnostic 0,000 Maladies circulatoires 2 6,5 29 0,44 0,07 - 2,15 0,42 0,09 - 1,97 Maladies infectieuses 13 24 40 2,06 0,85 - 4,98 1,76 0,77 - 4,04 Maladies endocriniennes 12 27 32 2,38 0,95 - 5,95 2,32 1 - 5 Maladies du système nerveux 2 7 25 0,51 0,08 - 2,53 0,53 0,26 - 2,45 Maladies respiratoiresb 17 13,5 108 Maladies génito- urinaires 11 18 51 1,37 0,55 - 3,37 1,17 0,50 - 2,73 Symptômes & signes 6 50 6 6,35 1,58 - 25,9 6,53 1,87 - 22 Traumatismes 12 52 11 6,93 2,39 - 20,3 6,53 2,48 - 17 Autres 10 32 21 3,03 1,11 - 8,23 2,68 1,06 - 6,75 aAjustement pour l’âge. bCatégorie de référence. CNSS : Caisse Nationale de Sécurité Sociale. CNRPS : Caisse Nationale de Retraite et de Prévoyance Sociale. 22 Analyse de la pertinence.pmd 12/20/2005, 2:39 PM439 440 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 jugée pertinente par la délivrance de soins non nécessaires mais jugés pertinents par l’AEP. Le taux de non-pertinence globale est de 21 % avec un intervalle de confiance à 95 % de 17-25, avec une variabilité entre les hôpitaux. Bien que la comparaison de nos résultats avec ceux de la littérature soit limitée par de nombreux obstacles (dif- férences dans l’organisation des systèmes de soins de santé, financement et mode de rémunération), nos résultats concordent avec ceux obtenus par d’autres études [18]. Parmi les 18 critères du référentiel, seulement 5 ont permis de justifier quasi- ment toutes les admissions pertinentes. La thérapie intraveineuse a justifié 56 % des admissions pertinentes. Ces résultats sou- lignent l’importance de la mise en place de structures de soins à domicile. Ces der- nières pourraient être organisées par les structures hospitalières et pourraient assu- rer des soins tels qu’une thérapie parentérale ou soins infirmiers, ce qui ré- duirait les dépenses hospitalières et libè- rerait des lits supplémentaires pouvant être utilisés pour d’autres patients. La probabilité d’avoir une admission non pertinente était la plus élevée pour le diagnostic. Ceci a été également rapporté par d’autres montrant que les pathologies les moins sévères engendraient le plus d’admissions non pertinentes [12]. Les admissions non pertinentes pour exploration de symptômes et signes anor- maux engendraient dans 70 % des cas des séjours allant de 8 à 21 jours. Ainsi, la mise en place de structures de soins intermé- diaires offrant la possibilité de faire des ex- plorations en ambulatoire serait d’un grand intérêt. La variabilité de la non-pertinence des admissions selon les hôpitaux pourrait être expliquée par la variabilité des ressources matérielles, humaines et financières et des différences dans les besoins en soins de santé d’une région à l’autre, ce qui suggère l’utilisation des cartes sanitaires nationale et régionale qui serviront de base d’allocation des ressources, selon des normes préa- lablement définies, en prenant en con- sidération les données épidémiologiques de chaque région. Remerciements Les auteurs remercient vivement les mé- decins des hôpitaux régionaux de Monastir pour avoir aidé à la réalisation de l’étude. Références 1. Agence Nationale pour le Développe- ment de l’Évaluation Médicale. Paris, ANDEM, 1994. 2. Overtveit J, Gustafson D. Evaluation of quality improvement programmes. Qual- ity and safety in health care, 2002, 11(3): 270–5. 3. Évaluation des pratiques profes- sionnelles dans les établissements de santé. Paris, Agence Nationale d’Accré- ditation et d’Évaluation en Santé, 1998 : 1152. 4. Davido A et al. Appropriateness of ad- missions in an emergency department: reliability and causes of failure. Quality assurance in health care, 1991, 3:227– 34. 5. Restuccia JD et al. Assessing the appro- priateness of hospital utilisation to im- prove efficiency and competitive 22 Analyse de la pertinence.pmd 12/20/2005, 2:39 PM440 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 441 position. Health care management review, 1987, 12:17–27. 6. Kreger BE, Restruccia J. Assessing the need to hospitalize children: Paediatric Appropriateness. Evaluation Protocol. Pediatrics, 1989, 84:242–7. 7. Lemeshow S et al. Adequacy of sample size in health studies. New York, John Wiley and Sons, 1990. 8. Robain M et al. Reproductibilité et validité de la version française de la première partie de l’Appropriateness Evaluation Protocol (AEPf) : critères de pertinence des journées d’hospitalisa- tion. Revue d’épidémiologie et de santé publique, 1999, 47:139–49. 9. SPSS 10.0 for Windows. Chicago, Illi- nois, SPSS Incorporated, 1999. 10. Ramos-Cuadra A et al. The effect of com- pleteness of medical records on the de- termination of appropriateness of hospi- tal days. International journal for quality in health care, 1995, 7(3):267–75. 11. Attena F et al. Appropriateness of admis- sions et hospitalization days in a spe- cialist hospital. Annali igdi igiene, 2001, 13:121–7. 12. Rodriguez-vera FJ et al. Appropriate- ness of admissions to an internal medi- cine department of a second level hospital. Anales medicana interna, 1999, 16:277–80. 13. Barré ML et al. Appropriateness of ad- missions and hospitalisation days in an acute - care teaching hospital. Revue d’epidémiologie et de santé publique, 1995, 43:328–36. 14. Siu AL, Mannig WG, Benjamin B. Pa- tient, provider and hospital characteris- tics associated with inappropriate hospitalization. American journal of public health, 1990, 80(10):1253–6. 15. Miron Canelo JA et al. Admissions esti- mate and inadequate stay in regional hospital. Revista clínica española, 2000, 200(1):654–8. 16. Schulz KF, Grimes DA. Case–control Studies: research in reverse. Lancet, 2002, 359, 2. 17. Bchir A et al. La morbidité hospitalière dans un district urbain de Tunisie. La Revue de Santé de la Méditerranée orientale, 1998, 4(1) : 94–9. 18. Thierry Lang. La Pertinence Médicale des procédures : mesure et relation avec les besoins et l’accès aux soins. Revue d’épidémiologie et de santé publique, 1998, 46:411–9. 22 Analyse de la pertinence.pmd 12/20/2005, 2:39 PM441 442 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 In vivo antiplatelet effect of intravenous alum in rabbits F.I. Mohammed1 and Y.A. Shafagoj1 1Department of Physiology and Biochemistry, Faculty of Medicine, University of Jordan, Amman, Jordan (Correspondence to F.I. Mohammed: fmmed@ju.edu.jo). Received: 11/01/05; accepted: 12/05/05 ABSTRACT Traditionally known as a haemostatic agent, alum shows a paradoxical effect of increased prothrombin and partial thromboplastin times. This study investigated the in vivo effect of alum on platelet aggregation and bleeding time in rabbits. The collagen-induced platelet aggregation of platelet-rich plasma samples from 14 healthy rabbits was measured turbidometrically using a platelet aggregometer, before and 1 hour after intravenous injection of alum. Collagen-induced platelet aggregation was significantly reduced after alum injection. Bleeding time from an ear puncture in 8 rabbits was also significantly prolonged after intravenous alum injection. These results suggest that the use of alum as an oral antiplatelet drug could be explored further, taking into account possible side-effects especially in renal compromised patients. Effet antiplaquettaire in vivo de l’alun injecté par voie intraveineuse chez des lapins RÉSUMÉ Traditionnellement connu comme agent hémostatique, l’alun montre un effet paradoxal d’allongement des temps de prothrombine et de thromboplastine partielle. La présente étude a examiné l’effet in vivo de l’alun sur l’agrégation plaquettaire et le temps de saignement chez des lapins. L’agrégation plaquettaire induite par le collagène dans des échantillons de plasma riche en plaquettes de 14 lapins sains a été mesurée par turbidimétrie en utilisant un agrégomètre, avant et une heure après une injection intra- veineuse d’alun. L’agrégation plaquettaire induite par le collagène était significativement réduite après l’injection d’alun. Le temps de saignement après incision à l’oreille chez 8 lapins était significativement prolongé après l’injection intraveineuse d’alun. Ces résultats semblent indiquer que l’utilisation de l’alun en tant qu’antiplaquettaire oral pourrait faire l’objet d’études complémentaires, en tenant compte des effets secondaires éventuels notamment chez les patients dont la fonction rénale est altérée. 23 In vivo antiplatelet.pmd 12/20/2005, 2:39 PM442 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 443 Introduction Alum (aluminium potassium sulfate) is a food additive and traditional remedy used to stop superficial bleeding from minor abra- sions or cuts. Its astringent properties con- tribute to its efficacy in the treatment of intravesical haemorrhage caused by pros- tate and bladder cancer [1,2] or hemor- rhagic cystitis [3]. Alum can also control haemorrhage from advanced rectal carci- noma [4] or oesophageal varices [5]. Alum proved to be safe and effective and thus remains the drug of choice for persistent vesical haematuria [6]. In addition, mouth rinses containing alum have been shown to reduce dental plaques and can therefore be useful in preventive dentistry [7,8]. The mechanism by which alum halts bleeding is not clearly understood. Protein precipitation and/or vasoconstriction are proposed mechanisms. Precipitation oc- curs primarily on the cell surface and su- perficial interstitial spaces. This leads to decreased capillary permeability, contrac- tion of intercellular space, vasoconstric- tion, hardening of the capillary endothelium and a reduction in oedema, inflammation and exudate [9,10]. Only 2 studies have followed its effect on the blood. Surprisingly, both reported elevated prothrombin and partial thrombo- plastin times with high serum aluminium levels after either intravesical irrigation [11] or accidental exposure to alum at work [12]. This paradox in the mechanism of action of alum is confirmed by our previ- ous report that human platelet aggregation induced by collagen, epinephrine, adenos- ine diphosphate (ADP) and thrombin were inhibited by in vitro treatment with alum [13]. However, to the best of our knowl- edge, the antiaggregation effect of alum has not yet been investigated in vivo. There- fore, to further investigate this apparent contradiction in the mechanism of action of alum, we evaluated the in vivo effect of alum in terms of collagen-induced platelet aggregation and bleeding time. We aimed to elucidate if the in vitro properties of alum were also observed in vivo. Methods Ethical approval for the study was obtained from the Institutional Animal Care Commit- tee of the University of Jordan, and the Committee guidelines were followed to minimize pain and discomfort during the entire experimental period. Materials Local inbred healthy male adult white rab- bits were used. The animals were main- tained on a regular diet with a locally purchased proprietary rabbit food mix and water ad libitum. The animal treatments were made between 10.00 and 12.00 hours. Collagen from calfskin was purchased from Sigma, St. Louis, Missouri, USA. Aluminium potassium sulfate (> 99% pure) was obtained from local sources. Platelet aggregation study For the platelet aggregation study, 14 rab- bits were used with a mean (standard devi- ation) body weight of 2.2 (0.23) kg, range 1.9–2.7 kg. The rabbits were injected in the margin- al ear vein with alum at 5 mg/kg body weight. Alum was dissolved in normal sa- line in a concentration of 1 mg/mL, and thus 5 mL/kg of this solution was injected. Blood samples were drawn from the cen- tral ear artery of study rabbits before the experiment started and 1 hour after injec- tion of alum. 23 In vivo antiplatelet.pmd 12/20/2005, 2:39 PM443 444 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Blood was collected into trisodium- citrated tubes through citrated syringes and butterflies. Citrated blood was then centri- fuged at 1000 rpm (250 × g) for 10 minutes at room temperature to obtain the platelet- rich plasma (PRP) fraction from the super- natant while excluding the buffy coat layer. The PRP was left for half an hour to pre- vent platelet shock. The remaining blood was centrifuged at 2500 rpm (1500 × g) for 10 minutes to obtain the platelet-poor plasma (PPP) fraction. The PPP was used as a blank to set the baseline for the aggre- gometer. Platelet aggregation was induced by adding 20 µL of collage solution (91 µg/ mL) to 200 µL of PRP. The reaction mix- ture was stirred magnetically at 1000 rpm and percentage aggregation was recorded at 30-second intervals for up to 5 minutes until aggregation was complete. Platelet ag- gregation was monitored turbidometrically with continuous reading of light transmis- sion using an aggregometer (model PAP-4, Bio/Data Corporation, Horsham, Pennsyl- vania, USA) according to the method of Born [14]. Bleeding time study For the bleeding time study, 8 rabbits were used with a mean (SD) body weight of 2.3 (0.24) kg, range 2.0–2.6 kg. The rabbits used for platelet aggregation could not be used at the same time for bleeding time ex- periments because of the stress of drawing blood on the animals and the injection site on the ear. The rabbits were injected with alum at 5 mg/kg body weight through the marginal ear vein and the other ear was used for the measurement of bleeding time. The rabbit’s ear was punctured using a routine automat- ic lancet and filter paper and the same ear was punctured in another place 1 hour after infection. Skin bleeding time was assessed before the experiment started and after in- jection of alum. Bleeding time was mea- sured using a stopwatch from the time of puncture until the time bleeding stopped and was determined by the same person. Follow-up The rabbits were followed for several weeks following the injection. The alum dose used had a mortality rate of zero and none of the animals show any gross visual signs of aluminium toxicity such as sei- zures, anorexia, vomiting or coma. Alum toxicity in terms of renal impairment, how- ever, was not assessed in this study. Controls Two groups of parallel experiments were conducted by injecting rabbits with 5 mL/ kg normal saline. Platelet aggregation in one group, and bleeding time in a second group, were measured before and after injection of normal saline. No changes were noted in the above parameters due to normal saline injection (data not shown). Statistical analysis The data are presented as mean [standard error (SE)]. Comparisons between data from control and alum-treated rabbits were made with the 2-tailed paired Student t-test and a P value < 0.05 was considered statis- tically significant. Results The results showed that alum significantly decreased the collagen-induced platelet ag- gregation of platelet-rich plasma 1 hour af- ter a single intravenous injection from a mean of 68.6% to 26.4% (P < 0.05) (Fig- ure 1). These inhibitory effects were sig- nificant as early as 30 seconds following the addition of collagen and reached a max- imum at 3 minutes (Figure 2). The same 23 In vivo antiplatelet.pmd 12/20/2005, 2:39 PM444 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 445 trend of inhibition 0.5 hour and 2 hours af- ter intravenous alum injection was record- ed (data not shown). Bleeding time from an ear puncture in rabbits was prolonged significantly from a mean of 2.36 minutes to 3.79 minutes (P < 0.05) after single intravenous alum injec- tion (Figure 3). Control rabbits were injected intrave- nously with normal saline alone and did not show any change in platelet aggregation or bleeding time. Discussion Platelets are believed to play a key role in the pathogenesis of atherosclerosis and acute coronary syndromes and therefore antiplatelet therapy may offer clinical bene- fits in the prevention of heart attacks and strokes. The haemostatic platelet function is mediated through their aggregation re- sponse to various endogenous agonists such as collagen, thrombin, adenosine diphosphate (ADP), adrenaline and 5- hydroxtryptamine (5-HT) [15]. In addition, bleeding time has been advocated as a sim- Figure 1 Mean percentage platelet aggregation of platelet-rich plasma 3 minutes after collagen induction, before and 1 hour after intravenous alum injection to rabbits (n = 14). Bars indicate standard error; *P < 0.05 Figure 2 Mean percentage platelet aggregation of platelet-rich plasma 30 seconds to 3 minutes after collagen induction, before and 1 hour after intravenous alum injection to rabbits (n = 14). Bars indicate standard error 23 In vivo antiplatelet.pmd 12/20/2005, 2:39 PM445 446 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 ple test for platelet function [16]. Alum stops superficial bleeding and is used to stop intractable haematuria caused by a multiple array of causes. The mechanism of this action has not yet been properly in- vestigated. The present study focused on the re- sponsiveness to in vivo administration of alum in terms of collagen-induced platelet aggregation and bleeding time. Opposite to what would be predicted, alum inhibited collagen-induced platelet aggregation and prolonged bleeding time when administered parenterally. The question of why a haemo- static agent such as alum inhibits platelet aggregation needs to be resolved. However, the current findings are consistent with our previous in vitro results showing that alum inhibited human platelet aggregation in- duced by collagen, epinephrine, ADP and thrombin in vitro [12]. Collagen, being a physiologically impor- tant activating agent, induces aggregation through forming thromboxanes. Its mech- anism of action involves binding to plasma von Willebrand factor that activates platelet receptors glycoprotein IIb/IIIa (GP IIb/ IIIa) [17–19]. It is conceivable that alum inhibits collagen-induced platelet aggrega- tion through competing for their GP IIb/ IIIa surface receptors. Although collagen and ADP react with the same receptors, collagen forms thromboxane but ADP does not. Alum probably works through a mech- anism different than that of aspirin. For in- stance, aspirin inhibits collagen but not ADP-induced platelet aggregation [20]. However, alum inhibits both collagen and ADP-induced platelet aggregation [13]. The importance of our findings needs to be investigated thoroughly. For instance, inhibitors of the platelet GP IIb/IIIa would be of great clinical value as anti-platelet agents. Alum is absorbed by the gas- trointestinal tract when given orally, which can be enhanced by citric acid [21]. If alum conforms to this principle it would in- triguing to explore its potential use as an an- tiplatelet drug. Extrapolating data from animals to hu- mans should be approached with caution. Systemic toxicity and dose–response stud- ies on alum would be needed before making any conclusions regarding its clinical val- ues [22,23]. Although the safety of alum ir- rigation has been established in many studies [11], its toxicity after intravenous injection has to be addressed thoroughly. For instance, patients with a damaged uri- nary bladder wall or renal impairment are at increased risk of developing aluminium toxicity. Intact renal function is essential for rapid disposal of a parenteral aluminium dose [24]. Therefore, patients with renal impairment are at increased risk of devel- oping encephalopathy, which might be a major drawback against routine use of alum [2,25]. Figure 3 Mean bleeding time of an ear puncture, before and 1 hour after intravenous alum injection to rabbits (n = 8). Bars indicate standard error; *P < 0.05 23 In vivo antiplatelet.pmd 12/20/2005, 2:39 PM446 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 447 References 1. Kavoussi LR, Gelstein LD, Andriole GL. Encephalopathy and an elevated serum aluminum level in a patient receiving intravesical alum irrigation for severe uri- nary hemorrhage. Journal of urology, 1986, 136:665–7. 2. Shoskes DA et al. Aluminum toxicity and death following intravesical alum irriga- tion in patients with renal impairment. Journal of urology, 1992, 147:697–9. 3. Seear MD, Dimmick JE, Rogers PC. Acute aluminum toxicity after continuous intravesical alum irrigation for hemor- rhagic cystitis. Urology, 1990, 36:353–4. 4. Paes TR et al. Alum solution in the con- trol of intractable haemorrhage from ad- vanced rectal carcinoma. British journal of surgery, 1986, 73:192. 5. Nishida R et al. A sclerosant with astrin- gent properties developed in China for oesophageal varices: comparison with ethanolamine oleate and polidocanol. Journal of gastroenterology and hepatology, 1999, 14:481–8. 6. Praveen BV, Sankaranarayanan A, Vaidyanathan S. A comparative study of intravesical instillation of 15(s) 15 Me al- pha and alum in the management of per- sistent hematuria of vesical origin. International journal of clinical pharma- cology, therapy, and toxicology, 1992, 30:7–12. 7. Donovan TE, Gandara BK, Nemetz H. Review and survey of medicaments used with gingival retraction cords. Jour- nal of prosthetic dentistry, 1985, 53:525– 31. 8. Olmez A et al. Effect of alum-containing mouthrinse in children for plaque and salivary levels of selected oral microf- lora. Journal of clinical pediatric den- tistry, 1998, 22:335–41. 9. Arrizabalaga M et al. Treatment of mas- sive haematuria with aluminous salts. British journal of urology, 1987, 60:223– 6. 10. Ostroff EB, Chenault OW Jr. Alum irriga- tion for the control of massive bladder hemorrhage. Journal of urology, 1982, 128:929–30. 11. Goswami AK et al. How safe is 15 alum irrigation in controlling intractable vesi- cal hemorrhage? Journal of urology, 1993, 149:264–7. 12. Waldron-Edward D, Chan P, Skoryna SC. Increased prothrombin time and meta- bolic changes with high serum alumi- num levels following long-term exposure to Bayer-process alumina. Ca- nadian Medical Association journal, 1971, 105:1297–9. 13. Mohammed FI, Shafagoj YA. Antiaggre- gation effect of alum on human platelets. Acknowledgements This work was supported by grant 439/97- 98 from the Scientific Deanship of the Uni- versity of Jordan. The authors thank Professor Abdullah Al-Abbadi for his scien- tific advice. We also thank Mrs Amal Abu- Ragheb for her excellent technical work. Conclusion This study indicates that alum inhibits platelet aggregation and prolongs bleeding time in vivo. Alum could be a cost- effective anti-platelet agent, but its side- effects would need to be addressed care- fully. The dose–response and elimination rate of alum would need to be studied to elucidate the clinical importance of these findings. 23 In vivo antiplatelet.pmd 12/20/2005, 2:39 PM447 448 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 International journal of clinical pharma- cology and therapeutics, 2001, 39:322– 4. 14. Born GVR. Aggregation of blood plate- lets by adenosine diphosphate and its reversal. Nature, 1962, 194:927–9. 15. Blockmans D, Deckmyn H, Vermylen J. Platelet activation. Blood reviews, 1995, 9:143–56. 16. Wagner SJ et al. Assessment of the he- mostatic effectiveness of human plate- lets treated with aminomethyltrimethyl psoralen and UV A light using a rabbit ear bleeding time technique. Blood, 1993, 82:3489–92. 17. Kotite NJ, Staros JV, Cunningham LW. Interaction of specific platelet membrane proteins with collagen: evidence from chemical cross-linking. Biochemistry, 1984, 23:3099–104. 18. Fujimoto T, Hawiger J. Adenosine diphosphate induces binding of von Willebrand factor to human platelets. Nature, 1982, 297:154–6. 19. Gralnick HR, Williams SB, Coller BS. Fi- brinogen competes with von Willebrand factor for binding to the glycoprotein IIb/ IIIa complex when platelets are stimu- lated with thrombin. Blood, 1984, 64: 797–800. 20. Naito J et al. Effects of thromboxane syn- thetase inhibitors on aggregation of rab- bit platelets. European journal of pharmacology, 1983, 91:41–8. 21. Deng Z et al. Effect of oral aluminum and aluminum citrate on blood level and short-term tissue distribution of alumi- num in the rat. Biological trace element research, 1998, 63:139–47. 22. Wilhelm M et al. Single-dose toxi- cokinetics of aluminum in the rat. Ar- chives of toxicology, 1992, 66:700–5. 23. Bertholf RL et al. A long-term intrave- nous model of aluminum maltol toxicity in rabbits: tissue distribution, hepatic, re- nal, and neuronal cytoskeletal changes associated with systemic exposure. Toxi- cology and applied pharmacology, 1989, 98:58–74. 24. Phelps KR et al. Encephalopathy after bladder irrigation with alum: case report and literature review. American journal of the medical sciences, 1999, 318:181– 5. 25. Moreno A et al. High serum aluminium levels and acute reversible encephal- opathy in a 4-year-old boy with acute re- nal failure. European journal of pediatrics, 1991, 150:513–4. 23 In vivo antiplatelet.pmd 12/20/2005, 2:39 PM448 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 449 Review Primary care psychiatry: pertinent Arabian perspectives O.E. El-Rufaie1 1Department of Psychiatry and Behavioural Sciences, Faculty of Medicine and Health Sciences, United Arab Emirates University, Al Ain, United Arab Emirates (Correspondence to O.E. El-Rufaie: elrufaie@uaeu.ac.ae). Received: 05/06/03; accepted: 07/04/04 ABSTRACT There is substantive evidence of significant psychiatric morbidity among primary care patients, mainly in the form of anxiety and depressive disorders. A careful critical approach is essential for ensuring the cultural relevance, validity and reliability of the psychiatric screening instruments used to identify such morbidity. Most psychiatric morbidity among primary care patients passes undetected by the primary care practitioners. This will inevitably lead to unnecessary investigation and medication and the continuation of suffering for patients. Comorbidity and physical presentation in most instances contribute significantly to failure to detect psychiatric disorders. To deal with this problem of hidden psychiatric morbidity, carefully designed educational and training programmes need to be tailored to address the particular weaknesses and needs of primary care doctors. Psychiatrie dans les soins primaires : perspectives pertinentes pour les pays arabes RÉSUMÉ De nombreux éléments attestent l’existence d’une morbidité psychiatrique importante chez les patients des soins primaires, principalement sous la forme d’anxiété et de troubles dépressifs. Une approche critique rigoureuse est indispensable pour garantir la pertinence culturelle, la validité et la fiabilité des instru- ments de dépistage psychiatrique utilisés pour identifier cette morbidité. La plus grande partie de la morbidité psychiatrique chez les patients des soins de santé primaires n’est pas décelée par les praticiens des soins de santé primaires, ce qui entraîne inévitablement des examens et des traitements inutiles et la persistance de la souffrance du patient. Une comorbidité et la symptomatologie dans la plupart des cas contribuent de manière significative à la difficulté de dépister les troubles psychiatriques. Afin de s’attaquer à ce problème de la morbidité psychiatrique cachée, des programmes d’éducation et de formation doivent être conçus avec soin de façon à aborder les faiblesses et les besoins particuliers des médecins des soins de santé primaires. 24 Primary care psychiatry.pmd 12/20/2005, 2:39 PM449 450 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Introduction Only half a century ago or so, psychiatric services were mostly confined to asylums. Seclusion and physical restraint constituted the main methods of treatment. A step for- ward occurred when asylums gradually evolved into mental hospitals, including both locked and unlocked wards. The con- cept of the “therapeutic community” grad- ually permeated into hospitals, allowing for more freedom of movement, group meet- ings and interaction with other patients and staff. The main objective was to create a healthy psychosocial environment condu- cive to positive therapeutic outcomes. Day hospitals, facilities in which chronic psy- chiatric patients spend the whole day and return to their homes late in the afternoon, were also established. Most hospitals pro- vide coach transportation for the morning and afternoon journeys. This was consid- ered a stage in rehabilitating chronically ill patients. Liaison services between mental and lo- cal general hospitals developed and were firmly maintained over the years. Psychia- trists were consulted about patients in gen- eral hospitals, particularly those admitted after an overdose or other suicide attempts; physicians from general hospitals used to attend physically ill patients in mental hos- pitals. Gradually, many mental hospitals be- came bases for psychiatric community services in their catchment areas. Psychiat- ric outpatient clinics outside the mental hospitals were offered as part of the com- munity services. From about the early 1960s a revolu- tionary development occurred in psychiat- ric services. This was the beginning of the era of establishing psychiatric departments in general hospitals. This was the time when psychiatry started to be accepted as an integral discipline of medicine, rather than a detached, queer phenomenon. The marriage between psychiatry and general medicine was facilitated by the introduction of electroconvulsive therapy in the 1940s, and more dramatically by the development of psychopharmacological agents since the late 1950s and early 1960s. The outcome was historical in nature, leading to a revival of hope for patients afflicted with horrible and mysterious illnesses. These develop- ments dissolved most, though not all, the barriers between psychiatrists and other medical and paramedical staff [1]. Recently, more courageous moves were introduced to provide high quality psychiatric services. This was part of the trend towards moving psychiatric care away from institutions into the wider com- munity and as close to the home as possible (World Health Organization Regional Office for Europe, unpublished report, 1980). Nowadays, psychiatrists are moving out to work directly with family practitioners in primary health care (PHC) settings. Concepts of psychiatric caseness and psychiatric measurement in primary care patients The notion of mental disorder may differ widely between cultures [2]. Psychiatric scales devised to assess the clinical charac- teristics of mental disorders have been uni- versally used with little consideration of how they relate to accepted definitions of the disorder. The scales are often used on the assumption that all measure the same construct. The item analysis of a number of the most widely used depression scales reveals a variation in the area of psycho- pathology [3]. Psychiatric screening in the PHC setting is unique because of the suspicion that 24 Primary care psychiatry.pmd 12/20/2005, 2:39 PM450 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 451 some psychiatric disturbances are the re- sult of transient distress rather than psychi- atric illness [4]. This confirms the need for a multiaxial evaluation (Lobo A. et al., un- published report, 1986). Some scales such as the General Health Questionnaire (GHQ) detect current emotional disturbances but can miss chronic illness [5]. It should be determined whether only current distur- bances are to be estimated or whether the aim is to identify all types of psychiatric disorder, in which case a different type of screening instrument would be indicated [6]. Comorbidity of physical and psychiatric illness complicates the criteria of “case- ness” (that which constitutes a psychiatric case) in PHC. There is a group of easily identifiable disorders that fit the known psychiatric classification systems, but there are also disorders that cannot be so easily identified [7–9]. Patients with such disorders usually present with definite symptoms, often mixed anxiety and de- pressive symptoms with some functional impairment. In choosing a screening scale for PHC, the reliability of the positive predictive val- ue of the psychiatric instrument should be taken into consideration. Psychiatric case- ness should not be determined by clinical judgement alone but by using standardized procedures linked to operational criteria [10]. The ICD-10 PHC classification can aid in setting realistic criteria for caseness in PHC [11]. Prevalence and nature of psychiatric morbidity in the primary health care setting A number of psychiatric epidemiological studies in PHC settings have been carried out since the late 1950s. One of the early studies reported that individual general practitioners regarded 26%–27% of their patients as psychiatrically ill [12]. A pio- neering study by Shepherd et al. established that most of variation between studies for estimates of prevalence was related to the characteristics of the doctor rather than different rates of psychiatric morbidity [13]. Around 30% of patients would proba- bly meet the criteria for a psychiatric illness if seen by a psychiatrist [14]. A further sur- vey of psychiatric morbidity in general practice and in the community have con- firmed the overall magnitude of this prob- lem [15]. The World Health Organization collaborative study on psychological prob- lems in 14 countries, using a 2-stage case identification methodology in a huge sam- ple (25 916 adults for the first stage and 5438 for the second) indicated that well- defined psychiatric morbidity was frequent in all the general health-care settings exam- ined (median 24.0%). The commonest problems identified were depression, anxi- ety, alcohol misuse, somatoform disorder and neurasthenia [16,17]. A primary care survey in Nigeria estimated the prevalence of psychiatric morbidity at 21.3%. Depres- sive neurosis (51.7%) and anxiety neurosis (36.3%) were the most common disorders [18]. A Greek study using the 28-item Gen- eral Health Questionnaire claimed an esti- mated 32% probable prevalence of mental health problems among their primary care sample [19]. Another primary care study using the Chinese version of the Beck In- ventory indicated a prevalence of 19.6% for major depression in a sample of Chinese Americans [20]. Epidemiological studies in primary care settings have not been limited to major psy- chiatric disorders e.g. affective and neurot- ic disorders, less-common disorders were also investigated by some researchers. Per- sonality disorders in primary care settings 24 Primary care psychiatry.pmd 12/20/2005, 2:39 PM451 452 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 also proved highly prevalent, and may rep- resent a significant burden on patient, fam- ilies and health services, and probably society as a whole in the case of antisocial personality disorders [21]. Prolonged fa- tigue is a persistent diagnosis in primary care settings and it has been suggested that psychiatric classification systems may be improved by treating prolonged fatigue and psychological distress as independent dis- orders [22]. Chronic fatigue syndrome di- agnosed according to the 1988 Centers for Disease Control criteria isestimated to be rare (3%) and atypical [23]. The authors of this study reported that generally patients mentioned fatigue if asked, but pain, in- somnia and worries were the most trouble- some symptoms. Eagles et al. found high levels of seasonal affective disorder among primary care patients during January [24]. The limitations of the study included the criteria for diagnostic assessment of the disorder. The volume of literature on primary care psychiatry in the Arabic-speaking countries is quite scanty, especially when compared to the wide spectrum of such studies done in the rest of the world. A sig- nificant proportion of studies from Arabic speaking countries in PHC psychiatry has been published in local journals. This may explain the limited amount of pertinent work which can be retrieved electronically. A major problem of the estimates of psy- chiatric morbidity in the Arab world is the choice of research methodology. Most psychiatric screening instruments were de- veloped in countries outside the Arab world. In other words, there is a scarcity of published literature on Arabic psychiatric screening instruments with established va- lidity and reliability which are applicable in the primary care setting. The use of psychiatric screening scales in a language and culture other than that for which it was designed and developed may create considerable problems because of the influence of translation and other socio- cultural factors. Some workers prefer to design new, culture-specific instruments [25,26]. Ideally, a screening scale should be developed in the cultural setting in which it is to be used. It would, however, be irra- tional to ignore well-established instru- ments developed in other cultures since many phenomena and concepts are com- mon between cultures. A critical, careful approach is, therefore, vital when translat- ing, modifying and validating psychiatric instruments before they are used across cultures [27]. Screening of representative samples of PHC patients in Saudi Arabia and the United Arab Emirates revealed psychiatric morbid- ity of 26.0% and 27.6% respectively [28,29]. The Saudi Arabian study, using the Hospital Anxiety and Depression (HAD) scale [30], showed the prevalence of de- pression to be 17% and prevalence of anx- iety 16%, and the total morbidity for depression, anxiety or both was 26%. The United Arab Emirates study, using the Clin- ical Interview Schedule (CIS), identified the 3 commonest diagnoses as neurotic de- pression (55.0%), mixed anxiety and de- pressive disorder (13.3%) and anxiety states (11.7%) [29]. A primary health care study carried out in northern Jordan using an Arabic translated version of the 28-item General Health Questionnaire (GHQ-28) estimated the prevalence of psychiatric morbidity as 61%. The highest prevalence rates were found in females ≥ 40 years, in uneducated and highly educated groups, in unemployed people, and in patients who were perceived to have fair or poor physi- cal health [31]. A Bahraini study in a PHC setting estimated the prevalence of psychi- atric morbidity, using the GHQ, as 45.1% (cut-off ≥ 5) and 27.1% (cut-off ≥ 9). Us- 24 Primary care psychiatry.pmd 12/20/2005, 2:39 PM452 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 453 ing the HAD scale, the prevalence was 44.4% (cut-off ≥ 8) and 23.6% (cut-off ≥ 11). Morbidity was commonest in wom- en aged 50–55 years, in divorcees or wid- ows, and in less-educated patients [32]. In conclusion, epidemiological studies done in various parts of the world and us- ing diverse methods and psychiatric instru- ments have furnished evidence of significant psychiatric morbidity (13%– 60%) among the PHC population. Studies done in the Arabic-speaking countries show a generally similar prevalence for psychiatric morbidity. Psychiatric disorders were mainly af- fective in nature, in the form of anxiety and/or depressive disorder. Other psychiat- ric disorders were also identified but at lower prevalence rates. Although depres- sive disorders are common and are associ- ated with considerable functional and medical morbidity in older primary care pa- tients, there is evidence that many of those with clinically significant depressive symp- toms do not meet the criteria for diagnostic categories. This was designated subsyn- dromal depression of older primary care patients [33]. Snaith also highlighted the concept of mild depression and the impor- tance of differentiating between such things as grief at loss, poor self-esteem and demoralization and genuine depressive ill- ness i.e. melancholic or severe biogenic de- pressive illness. Snaith proposed that anhedonia (loss of pleasure and enjoy- ment), if persistent and not adequately ac- counted for by the circumstances was the nearest guide to biogenic depressive disor- der, a state which has a high chance of im- provement with anti-depressant drug treatment [34]. A World Health Organiza- tion primary care collaborative study also confirmed the significant occurrence of subthreshold psychiatric conditions, which do not meet the formal diagnostic criteria of mental disorder in spite of the presence of significant psychiatric symptoms and functional impairment [16]. Somatization and somatized mental disorder among primary care patients High rates of medically unexplained symp- toms have been observed in many non- western countries and in ethnic minorities of industrialized countries. e.g. the obser- vations of Sethi et al. in India [35]. Howev- er, there is evidence that such symptoms are also a common problem in western so- cieties [36]. The word somatization is used in addition to unexplained symptoms to in- dicate that the physical disease by itself does not account for the symptoms and the clinical presentation, e.g. prolonged, exag- gerated severity. Such poorly defined terms will inevitably lead to further complication of an already difficult area of clinical medi- cine. The term somatization denotes, in general, the experience and communication of psychological distress in the form of physical symptoms. Vague terms are some- times used as synonyms of somatization, e.g. hypochondriacal, psychogenic or functional. The clinical presentation of somatiza- tion varies widely. The association between somatization and organic illness is well es- tablished. Bridges and Goldberg claim that 70% of somatizers suffer from organic ill- ness [36]. The association between somati- zation symptoms and a diagnosable, formal psychiatric disorder is fairly common, generally with depressive and anxiety dis- orders. On the other hand, comorbidity of somatization symptoms and formal psychi- atric illness is not an uncommon presenta- tion among primary care patients. The complexity of such clinical presentations may be further enhanced by additional co- 24 Primary care psychiatry.pmd 12/20/2005, 2:39 PM453 454 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 morbidity with personality disorder or un- adaptive personality traits. Such difficult clinical presentations may lead to undue suffering and confusion of management unless a precise and meaningful diagnosis is reached and a clear and rational strategy drawn up for short-term and long-term management. Educating the patient about his condition, with emphasis on the person- ality aspects, should be a central compo- nent of management. Somatization disorder as a formal diag- nostic entity meeting the diagnostic criteria of the ICD-10 or DSM-IV is not common among primary care patients. A primary care Nigerian study demonstrated that only 1.1% of the sample fulfilled the DSMIII-R criteria for somatization disorder but 4.7% met the criteria for somatoform pain disor- der and 10.8% met the criteria for undiffer- entiated somatoform disorder [37]. A study among a sample of primary care patients was carried out in the United Arab Emirates [38] using the Arabic ver- sion of the 12-item General Health Ques- tionnaire (GHQ-12), the Clinical Interview Schedule (CIS), an enquiry schedule and operational definitions for somatized mental disorder [36] and psychologized mental disorder [39]. The estimated prevalence rate of somatized mental disorder was 12%, constituting 48% of the psychiatric patients identified; the rate for psycholo- gized mental disorder was 12%. Education level was significantly lower in the soma- tized mental disorder group. Headache, backache, and abdominal pain were the most commonly presented somatic symp- toms. The symptoms pursued a chronic and persistent course and most patients ex- perienced multiple symptoms. The most common ICD-10 psychiatric diagnoses in both groups were mixed anxiety and de- pressive disorder, generalized anxiety dis- order, and mood and adjustment disorders. Recurrent depressive disorder and dys- thymia were significantly more prevalent in the psychologized mental disorder group. The severity of psychiatric illness was greater among psychologizers than the so- matizers of mental disorder. It is essential to note that although the Al-Ain study dis- closes many aspects of the phenomenon in Arab culture, it was restricted to somatized mental disorders, therefore caution is nec- essary on trying to compare the results with other forms of the broad phenomena of somatization. It is obvious that the concept of somati- zation involves many clinical presentations, therefore it is crucial that in individual pa- tients the precise nature of somatization should be identified before embarking on long-term management. The causes of so- matization are still obscure and there are many hypotheses in the literature. The eti- ology is probably multifactorial—it can hardly be explained in terms of a single cause. More rigorous research is needed to shed light on this fairly common clinical presentation. Missed (undetected) psychiatric morbidity among primary care patients Recognition of psychiatric disorders is of- ten difficult in primary health care settings owing to the complexities resulting from the comorbidity of physical and psychiatric illness. Personality disorders or maladap- tive personality traits in addition to comor- bidity of physical and/or psychiatric disorders may lead to a more complex clin- ical presentation which could easily be missed by a PHC practitioner who does not have a strong background of psychiatric clinical knowledge and experience. 24 Primary care psychiatry.pmd 12/20/2005, 2:39 PM454 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 455 In many situations, the ambiguous in- teraction between physical illness and the accompanying psychiatric symptoms make it difficult to identify which is prima- ry and which is secondary. Anxiety and de- pressive symptoms commonly accompany physical disease and consequently they are either missed or their contribution to the patients’ sufferings is underestimated. Bridges and Goldberg demonstrated that psychiatric illness occurs in a quarter to a third of all new episodes of illness seen in primary care settings. Most of these ill- nesses occur either in conjunction with known physical disease or as a “somatized” presentation of what is basically a psychiat- ric disorder [36]. Undetected psychiatric morbidity among PHC patients commonly leads to unnecessary investigation, medication and possibly hospitalization, as well as the con- tinued suffering of the patient. This will inevitably lead to impaired family, occupa- tional and social functioning. Among pri- mary care patients, many with personality disorders, especially borderline types, tend to be demanding, manipulative and non- compliant. They find ways to be referred for unnecessary tests; they attend more frequently, and sometimes unjustifiably, as emergencies. This makes it crucial for pri- mary care medical staff to be equipped with the necessary skills for detecting and managing these often difficult patients, as well as having a good grasp of basic psy- chiatric knowledge and skills [40,41]. Identifying and managing those with psychiatric morbidity, including personality disorders, among primary care patients is a serious problem which requires adequate intervention and collaboration. A lot of re- sources can be wasted as a result of unde- tected psychiatric morbidity. There is no single ideal model for addressing this prob- lem. Various psychiatric training schemes have been tested in different parts of the world, but it is crucial to assess the strengths and weaknesses of PHC practi- tioners before designing a training scheme for them. For any training scheme to be ef- fective it should address 2 main points: de- tection and management, and identifying weaknesses. The role of psychiatric screening instruments used in clinical set- tings to alert doctors to psychiatric disor- ders needs careful and critical enquiry [42]. Experience in Saudi Arabia and the Unit- ed Arab Emirates indicates that short, inten- sive, psychiatric courses administered at relatively long intervals are ineffective; continuity of rehabilitation programmes is essential and regular psychiatric clinics should be involved in those carried out in primary care settings. These programmes could be integrated into continuing medical education. In conclusion, it is clear that each group of primary care practitioners would need a programme tailored to their specific requirements. Conclusions and recommendations Worldwide epidemiological studies have produced solid evidence of significant psy- chiatric morbidity, mainly affective in na- ture, among PHC patients. Undetected psychiatric morbidity is a serious health problem at PHC level, often leading to a waste of resources. More work is needed to improve the psychiatric knowledge and skills of primary care doctors to reduce the rate of undetected psychiatric morbidity among their patients. In addition, the con- cept of personality and personality disorder should be fully understood by PHC medical and paramedical staff to enable them to ad- dress difficult clinical situations related to maladaptive personality traits. To this end, 24 Primary care psychiatry.pmd 12/20/2005, 2:39 PM455 456 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 the notion of comorbidity between physical illness, mental illness and personality char- acteristics should routinely be taken into account in the assessment of patients in primary care, as well as in other clinical settings. More research focussing on issues re- lating to primary care psychiatry is crucial to improve the quality of services in this ini- tial platform of the hierarchy of clinical medicine. There are many areas needing in- vestigation and clarification so that policies can be agreed by primary care staff, psy- chiatrists and other health care authorities. For carrying out clinically relevant psy- chiatric research in a PHC setting, culture- oriented, valid and reliable psychiatric instruments should be administered. The impact of using psychiatric instruments in a language and culture different from where they were developed should be fully comprehended by researchers, and precau- tions taken to counteract potential errors. There is a consensus that the application of current classification systems is not suit- able for patients. Perhaps in each research project involving identification of psychiat- ric cases in primary care settings, a realis- tic and meaningful definition of caseness, taking into account the sociocultural back- ground and other relevant factors should be drawn up precisely. Clinicians in PHC and other non- psychiatric settings should be alert to the possibility and the impact of undetected psychiatric morbidity. Rather than setting up barriers between physical, psychologi- cal and social health problems, a holistic approach in dealing with patients’ condi- tions should be a core item in the training of medical students, doctors, nurses and paramedical staff. References 1. El-Rufaie OE. Training general practitio- ners to deal with psychiatric disorders within the primary health care setting. Emirates medical journal, 1991, 9:128– 30. 2. Carstairs GM, Kapur RL. The great uni- verse of Kota. Stress, change and mental disorder in an Indian village. London, Hogarth Press, 1976. 3. Snaith PR. What do depression rating scales measure? British journal of psy- chiatry, 1993, 163:293–8. 4. Lloyd GG, Cawley RH. Distress or ill- ness? A study of psychological symp- toms after myocardial infarction. British journal of psychiatry, 1983, 142:120–5. 5. Goldberg D. Detecting psychiatric illness by questionnaire. Oxford, Oxford Univer- sity Press, 1972 (Maudsley Monograph 22). 6. Hobbs P, Ballinger CB, Smith AH. Factor analysis and validation of the general health questionnaire in women: a gen- eral practice survey. British journal of psychiatry, 1983, 142:257–64. 7. Goldberg D, Huxley P. Mental illness in the community. London, Tavistock Publi- cations, 1980. 8. Barrett JF et al. The prevalence of psychi- atric disorders in a primary care practice. Archives of general psychiatry, 1988, 45:1100–9. 9. Goldberg DP. The concept of a psychiat- ric case in general practice. Social psy- chiatry and psychiatric epidemiology, 1978, 17:61–5. 10. Thompson C. The instruments of psychi- atric research. Chichester, United King- dom, John Wiley & Sons, 1989. 24 Primary care psychiatry.pmd 12/20/2005, 2:39 PM456 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 457 11. The ICD-10 classification of mental and behavioural disorders. Clinical descrip- tions and diagnostic guidelines. Geneva, World Health Organization, 1992. 12. Royal College of General Practitioners Research Committee. The continuing observation and recording of morbidity. Journal of the Royal College of General Practitioners, 1958, 1:107. 13. Shepherd M et al., eds. Psychiatric ill- ness in general practice. London, Oxford University Press, 1966. 14. Dean C. Psychiatry in general practice. In: Forrest AD, Kendell RE, Zealley AK, eds. Companion to psychiatric studies, 4th ed. Edinburgh, Churchill Living- stone, 1988:634–47. 15. Goldberg D, Kay C, Thompson L. Psychi- atric morbidity in general practice and the community. Psychological medicine, 1976, 6(4):565–9. 16. Sartorius N et al. An international study of psychological problems in primary care. Preliminary report from the World Health Organization collaborative project on psychological problems in general health care. Archives of general psychiatry, 1993, 50(10):819–24. 17. Sartorius N et al. Depression comorbid with anxiety: results from the WHO study on psychological disorders in primary health care. British journal of psychiatry, 1996, 168(30):38–43. 18. Abiodun OA. A study of mental morbidity among primary care patients in Nigeria. Comprehensive psychiatry, 1993, 34(1): 10–3. 19. Androutsopoulou C et al. Psychological problems in Christian and Moslem pri- mary care patients in Greece. Interna- tional journal of psychiatry in medicine, 2002, 32(3):285–94. 20. Yeung A et al. Use of the Chinese version of the Beck Depression Inventory for screening depression in primary care. Journal of nervous and mental disease, 2002, 190(2):94–9. 21. Moran P et al. The prevalence of person- ality disorder among UK primary care attenders. Acta psychiatrica scandinavica, 2000, 102(1):52–7. 22. Hickie I et al. The temporal stability and co-morbidity of prolonged fatigue: a lon- gitudinal study in primary care. Psycho- logical medicine, 1999, 29(4):855–61. 23. Lee S et al. Psychiatric morbidity and ill- ness experience in primary care patients with chronic fatigue in Hong Kong. American journal of psychiatry, 2000, 157(3):380–4. 24. Eagles JM et al. Seasonal affective disor- der among primary care consulters in January: prevalence and month by month consultation patterns. Journal of affective disorders, 1998, 49(1):1–8. 25. Carstairs GM. Measuring psychiatric morbidity in a south Indian population. Bulletin of the British Psychological So- ciety, 1975, 28:95–101. 26. Verma SK, Wig NN. Standardization of a neuroticism questionnaire in Hindi. In- dian journal of psychiatry, 1977, 19:67– 72. 27. El-Rufaie OE, Daradkeh TK. Psychiatric screening in primary health care: transcultural application of psychiatric instruments. Primary care psychiatry, 1997, 3:37–43. 28. El-Rufaie OE, Albar AA, Al-Dabal BK. Identifying anxiety and depressive disor- ders among primary care patients: a pilot study. Acta psychiatrica scandinavica, 1988, 77(3):280–2. 29. El-Rufaie OE, Absood GH. Minor psychi- atric morbidity in primary health care: prevalence, nature and severity. Interna- tional journal of social psychiatry, 1993, 39(3):159–66. 24 Primary care psychiatry.pmd 12/20/2005, 2:39 PM457 458 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 30. Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta psychiatrica scandinavica, 1983, 67(6):361–70. 31. Al-Jaddou H, Malkawi A. Prevalence, recognition and management of mental disorders in primary health care in North- ern Jordan. Acta psychiatrica scandinavica, 1997, 96(1):31–5. 32. Al-Haddad MK et al. Psychiatric morbid- ity in primary care. Eastern Mediterra- nean health journal, 1999, 5(1):21–6. 33. Lyness JM et al. The importance of subsyndromal depression in older pri- mary care patients: prevalence and as- sociated functional disability. Journal of the American Geriatrics Society, 1999, 47(6):647–52. 34. Snaith RP. The concepts of mild depres- sion. British journal of psychiatry, 1987, 150:387–93. 35. Sethi BB, Nathawat SS, Gupta SC. De- pression in India. Journal of social psy- chology, 1973, 91(1):3–13. 36. Bridges KW, Goldberg DP. Somatic pre- sentation of DSM III psychiatric disorders in primary care. Journal of psychoso- matic research, 1985, 29(6):563–9. 37. Gureje O, Obikoya B. Somatization in pri- mary care: pattern and correlates in a clinic in Nigeria. Acta psychiatrica scandinavica, 1992, 86(3):223–7. 38. El-Rufaie OE et al. Somatised mental disorder among primary care Arab pa- tients. I: Prevalence and clinical and sociodemographic characteristics. Jour- nal of psychosomatic research, 1999, 46(6):549–55. 39. Bridges K et al. Determinants of somati- zation in primary care. Psychological medicine, 1991, 21(2):473–83. 40. El-Rufaie OE, Daradkeh TK. Detection of anxiety and depression in primary health care: assessment by primary health care physician versus assess- ment by psychiatrist. Primary care psy- chiatry, 1996, 2:189–93. 41. El-Rufaie EF et al. Personality profile among primary care patients: experi- menting with the IPDE ICD-10. Acta psychiatrica scandinavica, 2002, 105(1):37–41. 42. El-Rufaie OE, Absood G, Abou-Saleh MT. The primary care anxiety and de- pression (PCAD) scale: a culture ori- ented screening scale. Acta psychiatrica scandinavica, 1997, 95(2):119–24. 24 Primary care psychiatry.pmd 12/20/2005, 2:39 PM458 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 459 Review History of disorders of thyroid dysfunction A.M. Ahmed1 and N.H. Ahmed 2 1Faculty of Medicine, University of Bahr Elghazal, Khartoum, Sudan (Correspondence to A.M. Ahmed: awad_sd@hotmail.com). 2Ministry of Health, Khartoum, Sudan. Received: 17/08/03; accepted: 22/03/04 ABSTRACT The first description of thyroid diseases as they are known today was that of Graves disease by Caleb Parry in 1786, but the pathogenesis of thyroid disease was not discovered until 1882–86. Thy- roidectomy for hyperthyroidism was first performed in 1880, and antithyroid drugs and radioiodine therapy were developed in the early 1940s. Thomas Curling first described hyopothyroidism (myxoedema) in 1850 and the cause and suitable treatment were established after 1883. This paper reviews the main landmarks in the history of thyroid disease, supplemented by a brief discussion of the historically relevant scientific aspects of the thyroid gland, and the evolution of endocrinology as a formal discipline. Histoire des troubles dysthyroïdiens RÉSUMÉ La première description des pathologies thyroïdiennes, telles qu’on les connaît aujourd’hui, a été celle de la maladie de Graves par Caleb Parry en 1786, mais la pathogenèse de l’affection thyroïdienne n’a pas été découverte avant 1882-1886. Une thyroïdectomie pour hyperthyroïdie a été réalisée pour la première fois en 1880, et les antithyroïdiens et le traitement par l’iode radioactif ont été mis au point au début des années 1940. Thomas Curling fut le premier à décrire l’hypothyroïdie (myxoedème) ainsi que sa cause en 1850 et un traitement adéquat a été établi après 1883. Cet article procède à un examen des principales étapes dans l’histoire de la pathologie de la thyroïde, avec une brève discussion des aspects scientifiques concernant la glande thyroïde qui ont eu une importance historique, et de l’évolution de l’endocrinologie en tant que discipline à part entière. 25 History of disorders.pmd 12/20/2005, 2:39 PM459 460 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Introduction The history of thyroid disease is an exam- ple of the close inter-relationship of basic and clinical sciences. Although we can see a slow advancement in knowledge about the clinical features of thyroid disease since the dawn of the 20th century, the methods of investigation and treatment advanced rapidly. This paper reviews the main landmarks in the history of thyroid disease, starting with a brief outline of the evolution of mod- ern endocrinology. Many endocrine con- cepts such as feedback mechanisms, replacement therapy and the use of iso- topes were first established during studies on the thyroid gland. Although the aim of the paper was to discuss the historical as- pects of thyroid disease, we have supple- mented the review with a description of the historical developments in our knowledge of anatomy and physiology of the thyroid gland. Endocrinology: historical landmarks Some endocrine facts, such as the sequelae of castration, are deeply rooted in the past. Frederick Ruysch (1638–1731), a Dutch anatomist, was the first to speculate, in 1690, that an organ such as the thyroid pours into the blood substances of physio- logical importance [1]. The French scien- tist Theophile de Bordeu (1722–76) had a similar theory [1]. The Ruysch and Bordeu speculations were easily criticized by the lack of any experimental proof. But the real father of endocrinology who set its princi- ples as a formal discipline was French physiologist Claude Bernard (1813–78). In 1855, Bernard introduced the term “internal secretions” when he was studying the role of the liver in sugar metabolism. He stated that the external secretions form the bile and the internal secretions form the blood sugar [2]. From 1855 to 1902 important research in endocrinology was carried out [3]. Add- ison wrote an account on the disease of the suprarenal capsule [1]. Brown-Séquard demonstrated that in animals death from adrenalectomy could be delayed by infu- sion of blood from healthy animals. In 1889 Minkowski and Von Merring performed the experimental production of diabetes [1]. By the end of the 19th century a number of diseases such as Graves disease and ac- romegaly were described along with their probable relation to a glandular dysfunction [3]. English physiologists William Bayliss (1860–1924) and Ernest Starling (1866– 1927) made important discoveries in 1902 [3]. While they were studying the effects of “pancreatic juices” on the duodenal mu- cosa, they isolated a substance, giving it the name “secretin”. When secretin was in- jected into the bloodstream of an experi- mental animal, it resulted in marked increased secretion from the pancreas. This result reinforced Bernard’s concept of internal secretion [3]. In 1905 Starling coined the term “hormone” for this internal secretion (including his secretin). The word hormone was derived from a Greek phrase meaning “to set in motion”. Thus, secretin was the first hormone to be isolat- ed. The term “endocrine” can be traced to Starling who differentiated between endo- crine and exocrine secretions [1]. The history of endocrinology is not complete without a mention of the master of the endocrine glands, the pituitary. The pituitary gland was named by the Belgian anatomist Andreas Vesalius (1514–64) who believed that it secreted mucus through the nose (derived from the Latin word pituita which means mucus) [1]. Later, the Ger- man physician Conrad Victor Schneider 25 History of disorders.pmd 12/20/2005, 2:39 PM460 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 461 (1614–80) cast doubt on the mucus secret- ing function of the pituitary when he first described the cribriform plate of the eth- moid [1]. But this view persisted until Pierre Marie (1853–1940) detected two cases of acromegaly with associated pitu- itary enlargement in 1886 and 1888 [3]. Since then a large body of knowledge has accumulated about the pituitary which led Langdon Brown in 1935 to describe it as the “leader of the endocrine orchestra”. Anatomy of the thyroid The anatomy of the thyroid gland can be traced back as far as the first century AD when Galen, the famous Greek doctor, briefly described it in his treatise De voce [1]. In 1543 Vesalius gave a full description of the organ; he believed that it consisted of two separate parts [1]. Later, the anatomist Bartolomeo Eustachius recognized the isth- mus and considered the thyroid as a single organ [1]. But the controversy about whether the thyroid was a single or double organ was finally resolved by Giovanni Bathista Morgagni (1682–1771) who dem- onstrated that the gland had two lobes con- nected by the isthmus. Thomas Wharton of London (1617–73) named the thyroid in 1656, from the Greek word thyreos mean- ing “oblong shield” [3]. The histology of the thyroid was stud- ied from the start of the microscopy era. It was found to be formed of vesicles con- taining colloid material and lined with cubi- cal cells. The vesicles were thought to be communicating together, but in 1841, Hein- rich Bardeleben of Germany (1817–95) re- futed this idea [1]. Physiology of the thyroid Throughout history, the function of the thyroid has been controversial. Galen sug- gested that it functioned to lubricate the lar- ynx and this view was accepted for a long time. Wharton in 1656 suggested that the purpose of the thyroid was to beautify the neck by giving it a rounded contour through filling the vacant spaces around the larynx [3]. In 1829 Astley Cooper proposed that the thyroid was a lymphatic gland when he noticed lymph passing from it to the thoracic duct [1]. Even up to 1880 the thyroid was proposed as a receptacle of worms or even a vascular shunt to cushion the brain against a sudden increase in blood flow. Thomas King (1809–47) first suggested the concept of an internal secretary func- tion for the thyroid. King’s idea was that the thyroid formed and secreted a vital ma- terial into the circulation at a time of emer- gency [4]. This view was supported a few years later by Felix Semon [1]. In 1856 Moritz Schiff (1823–96) of Berne carried out thyroidectomies on dogs and guinea pigs with fatal results [3]. For 30 years nothing was heard of this work. The thy- roidectomy remained generally fatal until French physiologist Eugene Gley (1857– 1934) described the parathyroid glands in 1891 [5]. In 1882–83, Ludwig Rehn (1849–1930) and Jacques-Louis Reverdin (1841–1917) noticed the appearance of symptoms of hypothyroidism after thyroidectomy opera- tions on patients with Graves disease (see later) [6,7]. In 1884 Schiff remembered his 1856 work and repeated his previous ex- periments on dogs and succeeded in pre- venting the effects of thyroidectomy by grafting the thyroid onto another part of the animal body. Unfortunately, the body soon absorbed the gland. Thereafter, the func- tion of the thyroid gland was thought to neutralize or remove poisons from the body and hence that thyroid deficiency leads to toxaemia. 25 History of disorders.pmd 12/20/2005, 2:39 PM461 462 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 The definite function of the thyroid as a controller of metabolism was studied and confirmed by work carried out between 1886 and 1895 by George Murray (1865– 1939), Hector Mackenzie (1856–1938) and Edward Fox (1856–1938) [8]. These stud- ies were based on following up the suc- cessful effects of giving thyroid extracts by different routes to patients with myx- oedema [8]. From around 1895 thyroid researchers began to study the chemistry of the gland secretions. This field was pioneered by Eu- gen Baumann (1846–97) who was the first to recognize the role of iodine in the work of the thyroid gland when he discovered a high concentration in the gland [9]. In 1896, he isolated a compound containing iodine (iodothyrin) and suggested a relation to iodine metabolism. In the same year, En- glish paediatrician Robert Hutchison (1871–1960) found that iodine was con- centrated in the colloidal material within the glandular follicles [10]. In 1899 Oswald extracted an iodized protein which he termed “thyeroglobin” [11]. The presence of organic iodine in the plasma was indicat- ed by Gley and Bourcet in 1900 whereas in 1905 di-iodotyrosine was prepared by Wheeler and Jamieson [3]. Tri-iodothyro- nine, in plasma and thyroid, was recog- nized in 1952 by Gross and Pitt-Rivers [12]. The big success came in 1914 when Edward Kendall (1886–1970) isolated an iodine-containing crystalline product from alkaline hydrolyses of thyroid tissues and named it “thyroxine” [13]. He described it as a “stirring activator of metabolism and probably the hormone of the thyroid” [13]. By then, the active principle of the thyroid gland had been identified and then synthe- sized in 1927 by Charles Harrington (1878– 1965) and George Barger (1878–1939) [14]. Pituitary and thyroid The association between the pituitary gland and endocrine disease was first recognized by Pierre Marie in 1890 who found that pi- tuitary enlargement was a constant finding in acromegaly patients [3,15]. In 1910 Crowe, Cushing and Homans performed complete hypophysectomies in 100 dogs and reported a characteristic series of symptoms—bradycardia, hypothermia, ataxia and loss of consciousness [16]. The same symptoms followed removal of the anterior pituitary alone. During the 1930s enthusiastic and successful research was carried out to separate and isolate the vari- ous hormones of the anterior pituitary. One of them was thyroid-stimulating hormone (TSH), which was identified, by Collip and Anderson in 1935 [17]. In the 1970s TSH was found to be a glycoprotein and was assigned a molecular weight between 26000–30000 by Pierce and Wynston [5]. In 1938 it was first suggested that the engineering concept of “feedback control” could be applied to biological systems. Soon after, this was applied by Hoskins in endocrinology, who coined the term “thy- rostat” [18]. The phrase “pituitary–thyroid axis” was first used in the 1940s to de- scribe the relationship of the two glands. Hypothalamus and thyroid In 1948 Harris proposed that the hypothala- mus served as the key link between the en- docrine and nervous systems in reacting to the surrounding environment [19]. In 1951, Monte Greer confirmed that induc- ing hypothalamic lesions (between the paraventricular nuclei and the median emi- nence) prevented the usual response to the antithyroid drug thiouracil (i.e. thyroxin in- hibition and TSH overproduction) [20]. In 1955 Saffran, Schally and Benfey postulat- ed that first-order hormones interacted in 25 History of disorders.pmd 12/20/2005, 2:39 PM462 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 463 the hypothalamus to regulate the secretion of the pituitary hormones and they coined the term “releasing factors” for these hor- mones [21]. This was the first convincing report of the role of the hypothalamus in regulating pituitary hormones [21]. Simple goitre and cretinism Descriptions of goitre have been found in Hindu writings earlier than 1500 BC [1]. In the same period the Chinese used burnt sponge and seaweed in the treatment of goitre [1]. Hippocrates was aware of goi- tre. In the Middle Ages goitre was men- tioned in the book Lives of saints and around this time, the word “goitre” was coined [1]. The presence of goitre was considered a sign of beauty in some societ- ies. Also, the high prevalence of goitre in adults made it difficult to associate it with disease. Only from the work of Caleb Parry and Giuseppe Flajani at the beginning of the 19th century was goitre recognized as a source of pathological effects [22]. Hence the term “toxic goitre” was introduced. Chatin in 1853 in France was the first to describe the correlation between the iodine content of water, soil and air and the preva- lence of goitre [23]. This information was neglected until confirmed in 1923 by Von Fellenverg in Switzerland and Orr in En- gland [1] Paracelsus recognized the coexistence of endemic goitre and cretinism in the mid- dle of the 16th century in the Duchy of Salzburg [24]. Also in this century the first explorers of Colombia described “heavy and stupid savage of sluggish rabbits” [1]. The virtual disappearance of cretinism from Switzerland after introduction of io- dized salts in the 1920s gave a convincing proof about the relationship between en- demic goitre and cretinism. But the question was raised whether io- dine deficiency was the only cause for goi- tre. This issue was studied in India by Rob- ert McCarrison who started a laboratory for nutrition research in 1918. He hypothe- sized the presence of “goitrogens” that in- hibit thyroid function in the drinking water as a cause of goitre [25]. In 1928 Chesney and his co-workers were the first to ob- serve that cabbage may contain goitrogens and cause goitre in rabbits (originally they were working on syphilis research and were using cabbage-fed rabbits) [26]. Dur- ing the next decade a large number of veg- etables were found to be goitrogenic e.g. cauliflowers, turnip and Brussels sprouts. Burnt sponge and seaweed have been used in the treatment of goitre since the 12th century. After extensive work with these materials, the chemist Bernard Cour- tois in France extracted a substance in 1812 [1]. This was examined by Humphry Davy, then at Paris who named it iodine. Four years later, the English physician Wil- liam Prout tested potassium iodide on him- self to make sure that it was not toxic in small amounts and then he used it as a rem- edy for goitre [1]. But the first genuine therapeutic trial to use iodine was made by Jean-Francois Coindet of Geneva (1774– 1834) in 1820 [27]. He gave a tincture of iodine to 150 patients with goitre without ill-effects [27]. In 1829, Jean Lugol (1786– 1851) of Paris introduced his well-known iodine solution (originally used to treat tu- berculosis). The use of iodine to prevent goitre was established in 1909–13 when Marine and Lenhart, in the USA, demonstrated the role of iodine deficiency in causing goitre in black trout and other animals [28]. During 1917–20 Marine and Kimball gave sodium iodide (2 g in divided does twice yearly for 3 years) to 2000 schoolgirls in the goitre belt in Ohio, USA [29]. This regimen result- ed in a reduction of cases from 87.6% in 1917 to 13.1% in 1922 [29]. 25 History of disorders.pmd 12/20/2005, 2:39 PM463 464 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 The surgical treatment of goitre was probably first attempted by Albucasis, a prominent Arab surgeon in the Middle Ages. The first surgeon who achieved suc- cessful results with thyroidectomy was Theodor Billroth at Vienna in 1849. But the techniques of this operation were highly improved by Theodor Kocher at Berne [1]. He performed more than 7000 thyroidecto- mies in his clinic with a mortality rate of 4.5% [1]. He was awarded the Noble prize in 1909 for pathology and surgery of thy- roid disease, the first surgeon to win the prize. Hypothyroidism Up to 1850 doctors were familiar with the syndrome of cretinism but only in that year was the occurrence of features similar to cretinism in adults (i.e. myxodema) first reported by Thomas Blizzard Curling (1811–88) [30]. Curling noted the connec- tion of absence of thyroid tissues with symmetrical swelling of fat tissues at the sides of the neck in two of his patients [30]. William Gull (1816–90) who was a prominent pathologist at Guy’s Hospital, London gave a complete description of myxoedema in 1873, describing five cases of cretinism in adult women [31]. But it was William Ord (1834–1902) who in 1878 coined the term “myxoedema” when he found extensive deposits of mucin in the skin of feet of his patients at autopsy [32]. After 1880 researchers started to inves- tigate the exact cause, and then to propose a suitable treatment for myxoedema. In 1882, Reverdin noted the occurrence of symptoms of myxoedema after thyroidec- tomy [6]. Also in 1883, Kocher in his Berne clinic described features of myxoedema in 30 out of 100 patients after thyroidectomy [1]. Kocher attributed these features to chronic asphyxia and termed them “cachexia strumipriva”. Around the same time, Victor Horsley (1857–1919) pro- duced artificial myxoedema in dogs after thyroidectomy [3]. Then in the second half of 1883 a controversy followed between Kocher and Reverdin regarding the nature of the postoperative myxoedema-like syn- drome [3]. Felix Semon (1849–1921) put an end to this controversy by stating that myxoedema and cachexia struamiprivia were one disease and, together with cretin- ism, were due to one cause: loss of thyroid function [3]. Semon’s statement was in- vestigated and confirmed by a committee appointed by the Clinical Society of London (included, among others, Victor Horsley and Moritz Schiff). The road was now paved to propose a suitable treatment for hypothyroidism. In 1886 Horsley and Schiff tried to transplant the thyroid but achieved only temporary success as the body soon absorbed the transplanted gland [1]. George Murray in 1891 and Howitz in 1892 tried to treat myx- oedema with injection of thyroid extracts with successful results [33]. One of Mur- ray’s patients lived for a further 28 years after he was treated with hypodermic injec- tions of glycerine extract of sheep thyroid tissue for 6 months [33]. At the same time, Hector Mackenzie was successful in giving patients fresh thyroid extract by mouth [8]. But most patients had to wait until thyrox- ine was isolated and then synthesized 35 years later [14]. An interesting historical point is that af- ter myxoedema was fully described be- tween 1873 and 1883 we notice a slow advancement in knowledge about it. This may be attributed to the early discovery of an extremely simple and efficient therapy that reduced the impetus for exhaustive re- search. But associations of myxoedema with certain clinical problems were de- scribed between 1949 and 1963. Examples 25 History of disorders.pmd 12/20/2005, 2:39 PM464 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 465 include pericardial effusion (1949) by R.A. Kern, myxoedema coma (1953) by Le Marquand, cerebellar disorder (1960) by Jellinek and Kelly, pernicious anaemia (1960) by Wilson and Tudhope and ascites (1963) by Atkinson. The historical aspects of hypothyroid diseases other than myxoedema deserve to be mentioned in this article. Riedel in 1896 described a chronic non-malignant involve- ment of the thyroid gland (chronic thyroidi- tis) [34]. In 1912, Hashimoto first described the disease that now bears his name [35]. He described four cases of goi- tre whose histological features included atrophic epithelium, prominent lymphoid follicles and new abundant connective tis- sues [35]. At first there was argument about whether the disease was a separate entity or an early stage of Riedel disease. This controversy was not settled until the demonstration of serum antibodies in Hash- imoto disease in 1956, which stimulated extensive investigations of this phenome- non. In fact, the first evidence of a possible role for autoimmunity in thyroid disease was the tendency of Hashimoto disease to pass sooner or later into hypothyroidism or even an early phase of thyrotoxicosis. In 1936 a third variety of thyroiditis (sub- acute and non-suppurative) was described by De Quervain and now bears his name [36]. Hyperthyroidism The first person to describe the features of hyperthyroidism (later known as Graves disease) was Caleb Hillier Parry (1755– 1822) in 1786 [22]. He was a highly es- teemed practitioner at Bath, England. Parry saw five cases of exophthalmic goitre, pal- pitation and anxiety [22]. He described his patients as follows: “eyes were protruded from their sockets, faces exhibited an ap- pearance of agitation and distress, the heart beat was so violent that each systole of the heart shook the whole thorax…” [22]. Par- ry attributed these symptoms to alterations in the cardiac function. Also in a paper pub- lished after his death, Parry described a case of exophthalmos associated with goi- tre and ‘non-organic heart disease’ (to ex- plain palpitation). Although Parry was the first to recognize hyperthyroidism, the first published work about it was by Giuseppe Flajani in 1802 [3]. In 1833 Robert James Graves (1796– 1853) of Dublin (who was one of the founders of the Park Street School of Med- icine) published a description of exoph- thalmic goitre so admirable that the disease still goes by his name (although Parry rec- ognized it 47 years earlier) [37]. Like Parry, Graves erroneously thought that the dis- ease was produced by alterations in cardiac function. To quote, “the sudden manner in which the thyroid used to increase and then diminish in size and the connection of this with the state of the heart’s action are cir- cumstances which may be considered as indicating that the thyroid is slightly analo- gous in structure to the tissues properly called erectile” [37]. In 1840, Graves disease was described by Adolph von Basedow (1779–1854) of Germany who completed the picture of toxic goitre [38]. Exophthalmos was noted in only one of the three cases described by Graves but was described in all Basedow’s four cases [38]. Because Basedow had practised in Merseburg, the characteristic signs of Graves (goitre, exophthalmos and palpitation) were frequently called “Meres- berg’s triad”. Basedow also mentioned oth- er thyrotoxic features such as restlessness, wasting and diarrhoea [38]. At different times, many causes and pathogenetic mechanisms for Graves dis- ease have been suggested. Parry and 25 History of disorders.pmd 12/20/2005, 2:39 PM465 466 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Graves, as explained above, suggested a cardiac lesion. Up to 1880 the etiology of hyperthyroidism was thought to be a lesion of the nervous system. In 1880 Ludwig Rehn of Frankfurt performed the first thy- roidectomy for a patient with Graves dis- ease [7]. The improvement in symptoms and signs lead Rehn to propose that the clinical features of this disease were due to hyperactivity of the thyroid gland [7]. In 1886 Rehn’s view was supported by Paul Julius Mobius of Leipzig who suggested an abnormal activity of the thyroid that “poi- soned” the body. The final word on this is- sue was from William Osler who speculated in 1909 that the symptoms of Graves disease were “due to disturbed function of the thyroid gland, probably a hypersecretion of certain materials which induce a sort of chronic toxaemia” [1]. The term hyperthyroidism was suggested by Charles Mayo in 1907. In 1924 Henry Plummer and Walter Boothby of the Mayo Clinic advanced the view that in Graves disease the thyroid produces excessive amounts of thyroxine (which had been iso- lated nine years earlier). The persistence of exophthalmos after thyroidectomy led the researchers to conceive of an exophthal- mos-producing substance, produced out- side the thyroid. During the 1930s the pituitary gland was suspected to be the source of exophthalmos-producing sub- stance. For this reason Graves disease was treated in some centres at that time by pitu- itary irradiation. In 1952 Marcus and his co-workers reinvestigated exophthalmos- producing substance, which was later called long-acting thyroid simulator (LATS) by Adams and co-workers [39]. Therapy By the time therapeutic trials were begin- ning to test a drug to treat hyperthyroidism there was an established association be- tween endemic goitre and hypothyroidism. Thus the development of these drugs was based on the theory that any agent that can cause goitre might be useful in treating hy- perthyroidism. Since 1928 some plants in- cluding cabbage and turnip were found to be goitrogenic [26]. Thus the idea of devel- oping anti-thyroid drugs was borne [1]. In this context, cyanide radicals which were once thought to be goitrogenic were tried as anti-thyroid agents but without success. In 1942 goitre was noticed as a side effect in hypertensive patients using thiocyanates and these drugs were tested in the treat- ment of hypothyroidism but with little suc- cess. In early 1940s it was reported that the administration of sulfonamides in rats re- sulted in thyroid enlargement and a de- crease in metabolic activity. These effects were thought to be due to inhibition of syn- thesis of thyroid hormones. But sulfona- mides proved little value in the treatment of hyperthyroidism [1]. At the same period Astwood in the USA achieved promising results with thiouracil. Thereafter thiouracil compounds were de- veloped and after 1943 used clinically to treat hyperthyroidism [40]. Although they initially gained a bad reputation due to toxic side-effects, this problem was resolved by decreasing the dose. Carbimazole was in- troduced in 1952, and still remains the most widely used anti-thyroid drug. Before the performance of thyroidecto- my to treat hyperthyroidism there were two types of surgical operations that were tried without success. In 1854 the ligation of the inferior or superior thyroidal artery was described [1]. William Moore in 1865 suggested that palpitations in Graves dis- ease were due to increased sympathetic ac- tivity and several surgeons suggested resection of the cervical sympathetic chain [1]. As mentioned above, Rehn was the 25 History of disorders.pmd 12/20/2005, 2:39 PM466 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 467 first to perform subtotal thyroidectomy for Graves disease but for a few years earlier thyroidectomy had been introduced to treat simple goitre [7]. Safer thyroidectomy was made possible by the introduction of pre- operative use of iodine in 1923. Although Saul Hertz and Howard Means realized that radioiodine could be made and used as a tracer, the idea of use of radioiod- ine in endocrine research was first suggest- ed in 1937 in a lecture given by President Compton of the Massachusetts Institute of Technology [41]. In 1938 Robley Evans and Arthur Roberts made the short-lived I128 and studied its physiology in rabbits. This was the first practical use of a labelled substance to study its uptake in animals af- ter its intravenous injection [41]. The only drawback of I128 that hindered its use as a therapy was its very short half-life (25 min- utes). Then in 1939 Joseph Hamilton and Mayo Soley at California made two radio- iodines, I130 and I131, with half-lives of 12 hours and 8 days respectively. They were the first radioiodines used to study iodine physiology in humans. The next step was to use these radioiodines in treatment of hyperthyroidism. This was first performed by Hertz and Roberts in 1941 at Harvard General Hospital, and soon the effective- ness of this therapy was proved [41]. Conclusion As we have seen, the history of thyroid dis- ease gives an interesting insight into the inter-relationship of basic and clinical sci- ences. The understanding of the pathogen- esis and therapeutic trials of thyroid disease went hand in hand with the discovery of the physiological aspects of the thyroid gland. From the middle of the nineteenth century onwards, researchers gradually elucidated the anatomy, physiology and function of the thyroid which has led to the current diagnostic techniques and therapies available today. And the work continues in order to further our understanding of this important gland and improve and refine our diagnosis and treatment of thyroid dys- function. References 1. Singer C, Underwood EA. A short history of medicine, 2nd ed. Oxford, Oxford Uni- versity Press, 1962:519–34. 2. Bernard C. Piqure diabetes. Memorna societa de biologica, 1849, 1:80–92. 3. Hughes AF. A history of endocrinology. Journal of the history of medicine and allied sciences, 1977, 32:292–313. 4. King TW. Observations on the thyroid gland. Guy’s Hospital report, 1856, 1: 429–47. 5. Gley E. Sur les fonctions du corps thyroide. CR Società Biologica, 1891, 48:429–35. 6. Reverdin JL. Accidents consécutifs à l’ablation totale du goitre. Reviera medica suisse rom, 1882, 2:539. 7. Rehn L. Uber die extirpation des Kropfs bei morbus Basedowii. Berlin Klinica Wescher, 1884, 21:163–6. 8. Mackenzie HWG. A case of myxoedema treated with great benefit by feeding with fresh thyroid glands. British medical jour- nal, 1892, 2:940–1. 9. Baumann E. Uber das normale Vorkommen Von Jod in Thierkorper. Hoppe-Seyler’s Zeitschrift fur physio- logische Chemie, 1895–1896, 21:319– 30. 10. Hutchison R. The chemistry of the thyroid gland and the nature of its active con- stituents. Journal of physiology, 1896, 20:474–96. 25 History of disorders.pmd 12/20/2005, 2:39 PM467 468 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 11. Oswald A. Die Eiweisskorpor der Schilddruse. Hoppe-Seyler’s Zeitschrift fur physiologische Chemie, 1899, 27: 14–21. 12. Gross J, Pitt-Rivers R. Tri-iodothyronine in relation to thyroid physiology. Recent progress in hormone research, 1954, 10:109–24. 13. Kendall EC. The isolation of a compound containing iodine in the thyroid. Journal of the American Medical Association, 1915, 64:2042–3. 14. Harrington CR, Barger G. Constitution and synthesis of thyroxine. Biochemistry journal, 1927, 21:169–81. 15. Marie P. Acromegaly. Brain, 1890, 12:59– 81. 16. Crowe SJ, Cushing H, Homans J. Ex- perimental hypophysectomy. Bulletin of the Johns Hopkins Hospital, 1910, 21: 121–69. 17. Collip JB, Anderson EM. Studies on the thyrotropic hormone of the anterior pitu- itary. Journal of the American Medical Association, 1935, 104:965–6. 18. Hoskins RG. The thyroid–pituitary appa- ratus as a servo (feed-back) mechanism. Journal of clinical endocrinology, 1949, 9:1429–31. 19. Harris GW. Natural control of the pituitary gland. Physiological review, 1948, 28: 139–79. 20. Greer M. Evidence of a hypothalamic control of pituitary release of thyrotro- phin. Proceedings of the Society of Ex- perimental and Biological Medicine, 1951, 77:603–8. 21. Saffran M, Schally AV, Benfey BG. Stimu- lation of the release of corticotropin from the adenohypophysis by a neurohypo- physial factor. Endocrinology, 1955, 57: 439–49. 22. Parry CH. Collections from the unpub- lished medical writings of the late Caleb Hillier Parry (volume 2). London, Underwoods, 1825:111. 23. Chatin M. Presence de iode dans les eaux pluviales, les plantes des Antilles et des cotes de la Mediterranee. CR Academa Scienca, 1853, 37:723–4. 24. Paracelsus M. De generative stultorum. Liber theophrasti. In: Omnia opera tractatus. Rome, 1603, volume 1. 25. McCarrison R. Observations on endemic goitre. Lancet, 1908, 2:1275–80. 26. Chisney AM, Clawson TA, Webster B. Endemic goiter in rabbits. Bulletin of the Johns Hopkins Hospital, 1928, 43:261– 75. 27. Coindet JF. Iodine: on its application as a medicine. Quarterly journal of science, 1821, 11:408. 28. Marine D, Lenhart CH. Observations and experiments. Journal of experimental medicine, 1910, 12:311–20. 29. Marine D, Kimball OP. The prevention of simple goitre in man. Archives of internal medicine, 1920, 25:661–74. 30. Curling TB. Two cases of absence of the thyroid body, and symmetrical swellings of fat tissue at the sides of the neck, con- nected with defective cerebral develop- ment. Medico-chirurgical transactions, 1850, 33:303–6. 31. Gull WO. On a cretinoid state superven- ing in adult life in women. Transactions of the Clinical Society of London, 1873, 7:180–5. 32. Ord WM. On myxoedema, a term pro- posed to be applied to an essential con- dition in the cretinoid affection occasionally observed in the middle- aged women. Medico-chirurgical trans- actions, 1878, 61:57–78. 33. Murray G. Notes on the treatment of myxoedema. British medical journal, 1891, 2:796–7. 25 History of disorders.pmd 12/20/2005, 2:39 PM468 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 469 34. Riedel B. Vorstellung eines Kraken mit chronischer Strumitis. Ver Deutsch Ges Chirurgica, 1897, 26:127–31. 35. Hashimoto H. Zur Kenntniss der lympho- matosen Veranderung der Schilddruse (struma lymphomatosa). Archiv fur klinische Chirurgie, 1912, 97:219–25. 36. De Quervain F, Giordanengo G. Die akute and subacute nicht Eiterige. Thy- roiditis. Mitt Grenzgeb Medica Chi- rurgica, 1936, 44:588–93. 37. Graves RJ. Newly observed affection of the thyroid gland. London medical and surgical journal, 1835, 7:516–7. 38. Von Basedow CA. Exophthalmus durch Hypertrophie des Zellgewebes in der Augenhöhle. Wochenschrift für die gesammte Heilkunde, 1840, 6:197–204. 39. Adam DD, Beavan DW. The presence of abnormal thyroid stimulator in the blood of thyrotoxic patients. Journal of clinical endocrinology and metabolism, 1962, 68: 154–65. 40. Astwood EB. Treatment of hyperthyroid- ism with thiourea and thiouracil. Journal of the American Medical Association, 1943, 122:78–85. 41. Sawin CT, Becker DV. Radioiodine and the treatment of hyperthyroidism: the early history. Thyroid, 1997, 7:163–76. Corrections Schoolteachers’ knowledge of common health problems in Bahrain. F.A. Alnasir and J.H. Skerman. Eastern Mediterranean Health Journal, 2004, Vol. 10 Nos 4/5, pages 537–546. In the abstract: The name in Arabic of the first author should read: Frequency of the CCR5-delta 32 chemokine receptor gene mutation in the Lebanese population. W. Karam, R. Jurjus, N. Khoury, H. Khansa, C. Assad, P. Zalloua and A. Jurjus. Eastern Mediterranean Health Journal, 2004, Vol. 10 Nos 4/5, pages 671–675. The chemokine receptor gene in the title of the article on page 671 was not printed correctly and should read: CCR5-delta 32. 25 History of disorders.pmd 12/20/2005, 2:39 PM469 470 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Rapport Aspects épidémiologiques et cliniques de la neurosyphilis au Maroc M. Yahyaoui,1 S. Serragui2 W. Regragui,1 L. Errguig,1 O. Mouti,1 A. Benomar1 et T. Chkili1 Service de Neurologie, Hôpital des Spécialités, Rabat (Maroc) (correspondance à adresser à M. Yahyaoui : medyahaoui@hotmail.com). 2Faculté des Sciences, Marrakech (Maroc). Reçu : 06/04/03 ; accepté : 25/07/04 RÉSUMÉ La neurosyphilis représente 56 à 70 % de l’ensemble des syphilis viscérales et complique 5 à 10 % des syphilis non traitées. Le but de ce travail est une évaluation de l’aspect épidémiologique et de la présentation clinique de la neurosyphilis au Maroc à travers une série de 201 cas colligés au Service de Neurologie du Centre hospitalier universitaire de Rabat entre 1986 et 1997. L’âge moyen des patients est de 41,26 (E.T. 9,23) ans (extrêmes : 17-70 ans). Le sexe masculin est majoritaire (91 %). L’incidence de la neurosyphilis est élevée au Maroc. De 31 cas par an en 1985, elle a diminué à partir de 1990 pour atteindre 10 cas en 1997. Parmi les différents tableaux cliniques enregistrés, la méningo-encéphalite chronique vient au premier plan, suivie par la méningo-vascularite, le tabès et l’atrophie optique. Epidemiological and clinical aspects of neurosyphilis in Morocco SUMMARY Neurosyphilis accounts for 56%–70% of all visceral syphilis and is a complication in 5%–10% of cases of untreated syphilis. The aim of this study was to evaluate the epidemiological aspects and clinical presentations of neurosyphilis in Morocco through a series of 201 patients attending the Centre for Neuro- logical Services at the university hospital in Rabat between 1986 and 1997. The mean age of the patients was 41.26 (SD 9.23) years (range: 17–70 years); the majority (91%) were male. The incidence of neurosyphilis in Morocco is high. From 31 cases per year in 1985, it has fallen since 1990 to reach 10 cases in 1997. Among the different clinical presentations recorded, chronic meningoencepahalitis was the commonest, followed by meningovasculitis, tabes dorsalis and optic atrophy. 26 Aspects ÈpidÈmiologiques.pmd 12/20/2005, 2:39 PM470 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 471 Introduction La neurosyphilis représente 56 à 70 % de l’ensemble des syphilis viscérales diagnos- tiquées [1] et complique 5 à 10 % des syphilis non traitées [2]. Sa fréquence a re- marquablement diminué parallèlement à celle de la syphilis primaire grâce à l’introduction de la pénicilline. Cependant, à partir de 1985, un certain nombre de pub- lications attirent l’attention sur la résur- gence de la syphilis, en particulier chez les patients positifs pour le VIH, de même qu’une modification de la présentation clin- ique de la neurosyphilis. On note, en effet, une diminution des formes parenchyma- teuses au profit des manifestations pré- coces telles que les méningites et les méningo-vascularites [3]. Le but de ce travail est une évaluation de l’aspect épidémiologique et de la présenta- tion clinique de la neurosyphilis au Maroc à travers une série de 201 cas colligés au Ser- vice de Neurologie entre 1986 et 1997. Méthodes 201 patients présentant une neurosyphilis ont été colligés en 10 ans. Les critères d’inclusion ont été la positivité des sérolo- gies syphilitiques (VDRL et TPHA) dans le liquide céphalorachidien (LCR) réalisées chez tous les patients présentant des mani- festations neurologiques et une sérologie positive dans le sang. Cette dernière est de- mandée systématiquement chez tous les patients hospitalisés au Service de Neurolo- gie du Centre hospitalier universitaire (C.H.U.) de Rabat qui est le premier centre de référence en neurologie au Maroc. Les paramètres étudiés sont l’âge, le sexe, les antécédents de syphilis primo-secondaire, lh’incidence et la forme clinique. Résultats L’âge moyen de nos patients est de 41,26 ± 9,23 ans avec des extrêmes allant de 17 à 70 ans (Figure 1). Plus de la moitié des pa- tients (60 %) sont âgés de moins de 40 ans. Le sexe masculin est majoritaire avec 183 cas, soit 91 %. Le délai moyen de consultation est de 1,4 ± 2,6 ans ; 94 % des patients ont con- sulté avant 3 ans d’évolution. Les antécédents de chancre syphilitique ont été retrouvés chez 61 patients parmi 149, chez lesquels cette notion a pu être précisée (soit 41 %) ; douze d’entre eux seulement ont reçu un traitement adéquat (19,7 %). Les autres formes de syphilis pri- mo-secondaire ont été notés dans 7,5 % des cas (n = 15). Elles étaient dominées par la roséole (8 cas). Les résultats de la sérol- Figure 1 Répartition de la neurosyphilis en fonction de l’âge chez 201 patients Âge (ans) 26 Aspects ÈpidÈmiologiques.pmd 12/20/2005, 2:39 PM471 472 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 ogie du VIH n’ont pu être enregistrés que chez 29 patients ; ils étaient positifs chez 2 patients seulement. Le délai moyen d’apparition de la neuro- syphilis après le chancre est de 12,7 ± 8,6 ans [0-38] toutes formes confondues (Tab- leau 2). Ce délai est de 13 ans pour la mén- ingo-encéphalite chronique, de 7 ans pour la méningo-vascularite et de 15 ans pour le tabès. Enfin, la plupart des patients (43,7 %) ont été adressés par un psychiatre ; 18 % ont été amenés par leur famille, alors que 12,5 % ont été orientés par un médecin de ville et 25,1 % se sont présentés d’eux- mêmes à la consultation. Le nombre de nouveaux cas par an est de 17 ± 6,5 avec des extrêmes de 10 et de 31 cas. Toutefois, à partir de 1990, ce taux a commencé à diminuer pour atteindre 10 cas en 1997 (Figure 2). Différents tableaux cliniques ont été en- registrés avec, au premier plan, la méningo- encéphalite chronique avec 145 patients, soit 72,1 %, suivie par la méningo-vascu- larite, le tabès et l’atrophie optique. D’autres manifestations plus rares ont été notées et sont détaillées dans le tableau 3. Quarante patients (20 %) présentaient plu- sieurs formes cliniques associées dont la plus remarquable est l’association d’une méningo-encéphalite chronique avec un tabès, une atrophie optique et une gomme cérébrale. Tableau 1 Délai du diagnostic de la neurosyphilis chez 201 patients Délai du Nbre de cas % diagnostic (ans) < 1 77 38,3 1 62 30,8 2 33 16,4 3 17 8,5 4 1 0,5 5 3 1,5 6 2 1,0 7 1 0,5 8 2 1,0 10 1 0,5 11 1 0,5 30 1 0,5 Tableau 2 Délai de l’atteinte neurologique Délai (ans) Nbre % % validé de cas (201) (46) Validé <1 2 1,0 4,3 1 4 2,0 8,7 2 1 0,5 2,2 4 1 0,5 2,2 5 2 1,0 4,3 7 1 0,5 2,2 8 5 2,5 10,9 10 3 1,5 6,5 11 2 1,0 4,3 13 3 1,5 6,5 14 5 2,5 10,9 15 4 2,0 8,7 16 2 1,0 4,3 20 7 3,5 15,2 25 1 0,5 2,2 28 1 0,5 2,2 36 1 0,5 2,2 38 1 0,5 2,2 Total 46 22,9 100,0 Valeurs manquantes 155 77,1 Total 201 100,0 26 Aspects ÈpidÈmiologiques.pmd 12/20/2005, 2:39 PM472 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 473 Discussion Notre série de 201 malades se singularise par l’âge jeune des patients estimé à 41,2 ans. Le délai moyen de consultation de 1,4 an ramène l’âge de début à 39,86 ans en moyenne contre 56 ans dans la série de Pavlovic [4] et 50 ans dans celle de Rodg- ers [5]. L’âge moyen dans la série de Flood [3] portant sur 117 patients était de 39 ans. Toutefois, 64 % de ces patients étaient séropositifs pour le VIH. En effet, le délai entre le chancre et la neurosyphilis chez les sidéens se trouve réduit à 1 an en moyenne [6,7]. Dans certains cas, l’atteinte neu- rologique peut même être contemporaine du chancre [8,9,10,11], ce qui explique l’âge jeune des patients. Dans notre série, seuls deux patients sur 29 étaient co-in- fectés par le VIH. Ils étaient âgés de 30 et de 41 ans ; le délai d’apparition des mani- festations neurologiques était inférieur à un an chez les deux patients et leur évolution était fatale en moins d’un mois. La sérolo- gie du VIH était négative chez 27 patients et inconnue chez les autres ; 2 patients seule- ment sont décédés à la phase aiguë, les au- tres cas ayant été suivis pendant au moins 10 mois (période correspondant à 4 cures de pénicilline G à fortes doses à 3 mois d’intervalle). La prédominance masculine pourrait être expliquée par la plus grande fréquence de la méningo-encéphalite chronique chez l’homme ; cette dernière est en effet 4 à 7 fois plus fréquente chez l’homme que chez la femme [12]. Figure 2 Incidence de la neurosyphilis entre 1986 et 1997 Tableau 3 Différents syndromes cliniques de la neurosyphilis et leur répartition Syndromes cliniques Nbre % de cas Méningo-encéphalite chronique 145 72,1 Méningo-vascularite 42 20,9 Accidents vasculaires ischémiques 24 11,9 Striatite 12 6,0 Cérébellite 6 3,0 Tabès 20 9,9 Atrophie optique 13 6,4 Myélite d’Erb 6 3,0 Myélite aiguë 1 0,5 Sclérose latérale amyotrophique 4 2,0 Radiculite 3 1,5 Myoclonies 2 1,0 Gomme 3 1,5 Latente 2 1,0 26 Aspects ÈpidÈmiologiques.pmd 12/20/2005, 2:39 PM473 474 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 La faible proportion de patients chez qui les antécédents de chancre ont pu être re- trouvés est due au tabou qui entoure l’activité sexuelle dans notre culture. Le chancre est retrouvé dans moins de 50 % des cas. Cette sous-estimation est due probablement à l’absence de douleur mais surtout à la difficulté d’accès aux soins et à l’ignorance. Les délais moyens d’apparition de la neurosyphilis après le chancre sont compa- rables à ceux de la littérature [13,14], sauf pour la méningo-encéphalite chronique où l’on note un rajeunissement de 7 ans, ce qui explique en partie l’âge jeune de nos pa- tients. L’importance de l’atteinte psy- chologique et/ou intellectuelle est à l’origine d’une perte d’autonomie, et de désinsertion sociale et professionnelle chez des patients de plus en plus jeunes. Nous sommes par ailleurs frappés par l’incidence élevée de la neurosyphilis dans notre pays. Cette incidence a atteint 31 cas pas an en 1989. En effet, l’ignorance et la sous-médicalisation des patients font que peu d’entre eux traitent leur chancre. Le tréponème pâle (TP) envahit le système nerveux central (SNC) dès les premiers stades de l’infection. Non traités, 25 % des patients n’arrivent pas à l’éliminer de leur SNC et un tiers parmi eux va développer une neurosyphilis, alors que sous pénicilli- nothérapie, très peu de personnes immuno- compétentes échouent dans leur lutte contre le TP et sont exposées au risque de complications neurologiques [15]. Cette in- cidence a commencé à diminuer à partir de 1990 pour atteindre les 10 cas par an en 1997. Ceci pourrait être le résultat des cam- pagnes de sensibilisation établies par le min- istère de la santé publique pour la lutte contre le SIDA et les infections sexuelle- ment transmissibles (IST), dont la syphilis. Cette diminution de l’incidence pourrait également être le fruit de la constitution d’une consultation spécialisée des maladies vénériennes au sein des dispensaires. Alors que ces dernières années ont vu une raréfaction de la neurosyphilis tardive (méningo-encéphalite et tabès) au profit des formes précoces méningo-vasculaires et atteintes oculaires, notamment avec la recrudescence de l’infection par le VIH [16,3,7,17], nous enregistrons toujours une fréquence élevée de la méningo- encéphalite chronique. En effet, au Maroc, cette dernière demeure l’atteinte la plus fréquente et constitue 82 % des cas de la série de Chkili [1] alors que Hooshmand (1972), dans sa série de 241 cas, n’en a enregistré que 12 (soit 5 %) [15]. Certains auteurs rapportent une relation probable entre le système HLA et la surve- nue de la méningo-encéphalite et en parti- culier l’HLA Aw32, A2, B7 et B18 [19,20,21]. Cette relation indiquerait que des mécanismes déterminés génétiquement contribuent au développement de la méningo-encéphalite syphilitique et font que certains patients, en dépit d’un traite- ment bien conduit à la phase primo- secondaire, développent des années plus tard une neurosyphilis. La méningo-vascularite n’est pas rare dans notre série (11,9 %). Sa fréquence a été diversement appréciée : 94 % pour Laplane [22], 11 % pour Hooshmand [18] et 10 % pour Burne [23]. Dans notre série, il s’agissait d’accidents vasculaires ischémiques intéressant le territoire de l’artère sylvienne ou de l’une de ses branches. Une seule thrombophlébite a été notée dans notre série. Par ailleurs, nous avons été frappés par la fréquence particu- lière de formes rares, notamment la striatite et la cérébellite. En effet, 12 cas de striatite ont été enregistrés, entrant dans le cadre de l’artérite de Nissl Alzheimer [24,25,26,14]. Elles sont exceptionnelles et se manifestent 26 Aspects ÈpidÈmiologiques.pmd 12/20/2005, 2:39 PM474 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 475 par un syndrome akinéto-hypertonique, rarement par un tremblement [1,27]. La cérébellite est considérée comme rare et se traduit par un syndrome cérébelleux aigu qui peut être d’origine vasculaire, ischémique surtout, faisant évoquer une artérite oblitérante vertébrobasilaire rarement rapportée dans la littérature [1,27,28,29,30,14]. L’atteinte cérébelleuse peut aussi s’installer progressivement ; elle entre alors dans le cadre de la méningo- encéphalite et s’explique par l’atrophie cérébelleuse secondaire à l’altération de l’ensemble des structures cérébrales [1]. En effet, 3 de nos patients avaient une méningo-encéphalite associée. Le tabès occupe le 3e rang des formes cliniques de la neurosyphilis, ce qui con- firme encore la fréquence des formes tardi- ves chez nous. Sa fréquence est très variable selon les auteurs ; Hooshmand l’estime à 5 % [18]. Il complique 10 % des syphilis non ou mal traitées [31,14] ou con- stitue une manifestation tardive de la syphi- lis congénitale [29]. Il est plus fréquent chez les hommes et survient 10 à 20 ans après le chancre [14]. L’atrophie optique est une complication de la névrite optique qui évolue en 5 à 10 ans vers la cécité [14] ; malheureusement, elle n’est pas rare et aggrave le pronostic fonctionnel. Elle a été relevée dans 13 cas, soit 6,46 %, dont 4 avaient également un signe d’Argyll Robertson et 11 cas étaient associés à une méningo-encéphalite ou à un tabès. La plupart ont gardé des séquelles visuelles, certains se sont même aggravés. En effet, la gravité de cette atteinte vient de son évolution possible vers la cécité en l’absence de traitement [28,14]. Les myélites syphilitiques sont assez rares. Aucun cas n’a été rapporté dans la série de Hooshmand de 241 patients [18] ni dans la série d’El Alaoui Faris de 53 cas [27]. Luxon estime leur fréquence entre 5 et 30 % [32], et Oubejji dans sa série de 7 patients a enregistré 3 cas, soit 42,85 % [31]. Nous en avons enregistré 6 dont 5 ont gardé des séquelles motrices. En effet, la dégénérescence des faisceaux pyramidaux fait que l’évolution se fait lentement vers un état grabataire en 10 à 15 ans [33]. La sclérose latérale amyotrophique (SLA) syphilitique est exceptionnelle. Hoo- shmand en a rapporté un cas dans sa série de 241 patients et il y avait 5 cas sur 52 dans la série d’El Alaoui Faris [27]. Elle tou- che des sujets plus jeunes et l’évolution est meilleure que dans la forme idiopathique [34]. En effet, notre série en comporte 4 cas, dont 3 ont relativement bien évolué avec même une guérison totale pour l’un d’eux. Sa traduction clinique était clas- sique, associant une paraparésie spastique avec amyotrophie, crampes et fascicula- tions. La gomme syphilitique est une entité ex- ceptionnelle [14]. Trois cas figurent dans notre série, dont une médullaire et deux cérébrales. Ces dernières ont bien évolué alors que la gomme médullaire a laissé des séquelles sous forme de paraparésie après traitement chirurgical. Les myoclonies comme seule manifes- tation épileptique sont exceptionnelles. En effet, un seul cas a été rapporté dans la lit- térature par G. Heide et al à propos d’une femme de 40 ans qui a présenté des myocl- onies généralisées à prédominance distale comme seule manifestation de la neuro- syphilis [35]. Dans notre série, les myoclo- nies n’étaient pas isolées et s’associaient à une cérébellite dans un cas, à une cérébel- lite et une méningo-encéphalite dans le deuxième cas. Elles ont disparu sous traite- ment antibiotique. L’atteinte polyradiculaire, en dehors du tabès, a été observée dans 3 cas : une asso- ciée à une névrite crânienne, l’autre à une striatite, et la dernière réalisant un syn- 26 Aspects ÈpidÈmiologiques.pmd 12/20/2005, 2:39 PM475 476 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 drome de la queue de cheval dans le cadre d’une ostéo-arthropathie syphilitique. En effet, l’atteinte périphérique peut se mani- fester par une sciatique atypique, un syn- drome de la queue de cheval [14] ou exceptionnellement une polyradiculonévrite [36] et une atteinte de la corne antérieure [26]. Enfin, deux cas de syphilis latente ont été notés. Les patients se plaignaient uniquement de céphalées chroniques, et le diagnostic a été posé fortuitement à l’occasion d’un don de sang pour l’un et d’une sérologie systématique pour l’autre. Malheureusement, l’évolution était péjora- tive pour le second, avec installation d’un tableau typique de méningo-encéphalite sous traitement faisant soulever l’hypothèse d’une résistance à la pénicil- line. Références 1. Chkili T et al. Manifestations neuro- logiques et psychiatriques de la syphilis cérébrale. Encyclopédie Médicale Chirurgicale (Paris-France) Psychiatrie, 37620 A10, 10-1989, 10 p. 2. Uemura K et al. Cerebral gumma mim- icking glioblastoma on magnetic reso- nance images—case report. Neurologia medico-chirurgica, 1995, 35(7):462–6. 3. Flood JM et al. Neurosyphilis during the AIDS epidemic, San Francisco, 1985– 92. Journal of infectious disease, 1998, 177(4):931–40. 4. Pavlovic DM, Milovic AM. [Clinical char- acteristics and therapy of neurosyphilis in patients who are negative for human immunodeficiency virus]. Srpski arhiv za celokupno lekarstvo, 1999, 127(7– 8):236–40. 5. Rodgers CA, Murphy S. Diagnosis of neurosyphilis: appraisal of clinical caseload. Genitourinary medicine, 1997, 73(6):528–32. 6. Hsieh SM et al. Concomitant human im- munodeficiency virus infection and syphilitic meningitis. Journal of the Formosan Medical Association, 1996, 95(2):166–9. 7. Katlama C. Manifestations neurolo- giques de l’infection à VIH. Editions Techniques Encyclopédie Médicale Chirurgicale (Paris-France) Neurologie, 17-051-B-10, 1993, 8p. 8. Guerrier Y, Guerrier B, Romieu M. La syphilis de l’oreille. Etude clinique et thérapeutique. Minerva otorinolaringo- logica, 1970, 20(6):226–30. 9. Johns DR, Tierney M, Felsenstein D. Al- teration in the natural history of neuro- syphilis by concurrent infection with the human immunodeficiency virus. New England journal of medicine, 1987, 316: 1569–72. 10. Quinn P, Weisberg L. Cerebral syphilitic gumma. New England journal of medi- cine, 1997, 336(14):1027; 1027–8. 11. Schofer H et al. Active syphilis in HIV in- fection: a multicentre retrospective sur- vey. The German AIDS Study Group (GASG). Genitourinary medicine, 1996, 72(3):176–81. 12. Simon RP. Neurosyphilis. Archives of neurology, 1985, 42:606–13. 13. Gallego J et al. Magnetic resonance an- giography in meningovascular syphilis. Neuroradiology, 1994, 36(3):208–9. 14. Uldry P-A et Regli F. Neurosyphilis. - Edi- tions Techniques Encyclopédie Médi- cale Chirurgicale (Paris-France) Neuro- logie, 17-055-A-10, 1994, 4p. 26 Aspects ÈpidÈmiologiques.pmd 12/20/2005, 2:39 PM476 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 477 15. Golden MR, Marra CM, Holmes KK. Up- date on syphilis. Resurgence of an old problem. Journal of the American Medi- cal Association, 2003, 290:1510–4. 16. De Souza MC, Nitrini R. Effects of human immunodeficiency virus infection on the manifestations of neurosyphilis. Neurol- ogy, 1997, 49(3):893–4. 17. Podwinska J. Syphilis and AIDS. Archivum immunologiae et therapiae experimentalis, 1996, 44(5–6):329–33. 18. Hooshmand H, Escobar MR, Kopf SW. Neurosyphilis. A study of 241 patients. Journal of the American Medical Asso- ciation, 1972, 219(6):726–29. 19. Strandberg Pedersen N, Axelsen NH, Svejgaard A. General paralysis of the in- sane associated with HLA-Aw32. Acta pathologica et microbiologica scandinavica. Section C, Immunology [C], 1981, 89(3):181–4 20. Wheeler JS et al. Bladder dysfunction and neurosyphilis. Journal of urology, 1986, 136(4):903–5. 21. Whitsett C et al. HLA antigen frequencies in paralytic demencia. Journal of the neurological sciences, 1976, 30(2–3): 417–20. 22. Laplane D, Ricou PH. La neurosyphilis. La Revue du Praticien, 1976, 26:4133– 6. 23. Byrne RE et al. Evaluation of a Tre- ponema pallidum western immunoblot assay as a confirmatory test for syphilis. Journal of clinical microbiology, 1992, 30(1):115–22. 24. Aupy M et al. Neuro-vascularites syphilitiques. Données cliniques, paracliniques et thérapeutiques. A propos de six observations. La Semaine des hôpitaux de Paris, 1982, 58(18): 1101–6. 25. Harris DE et al. Neurosyphilis in patients with AIDS. Neuroimaging clinics of North America, 1997, 7(2):215–21. 26. Nakane H et al. Brain infarction caused by syphilitic aortic aneurysm. A case re- port. Angiology, 1996, 47(9):911–7. 27. El Alaoui Faris M et al. Les complications neurologiques de la syphilis. Etude de cinquante-trois cas. La Semaine des hôpitaux de Paris, 1986, 62(44):3511–5. 28. Hajjout K. La méningo-encéphalite syphilitique (à propos de 48 cas) [Thèse de médecine]. Rabat, 1990 nº 187. 29. Malvoisier JF. Etude biologique, scanographique et psychométrique de la neurosyphilis traitée à propos de 16 observations [Thèse de médecine]. Sainte-Etienne, 1984 nº 8424. 30. Nader J et al. Vasculites cérébrales infectieuses. Encyclopédie Médicale Chirurgicale (Ed Elsevier, Paris) Neurologie, 17046-B-05, 1996, 6p. 31. Oubajji M. Neurosyphilis à Kénitra (à propos de 07 cas) et revue de la littérature [Thèse de médecine]. Rabat, 1998 nº 162. 32. Luxon L, Lees AJ, Greenwood RJ. Neu- rosyphilis today. Lancet, 1979, 1(8107): 90–3. 33. Jourani S. Les myélopathies d’origine syphilitique. Etude de 20 cas [Thèse de médecine]. Rabat, 1990 nº 160. 34. El Alaoui Faris M et al. Le syndrome de sclérose latérale amyotrophique d’origine syphilitique. Etude de 5 cas. Revue neurologique (Paris), 1990, 146:41–4. 35. Heide G, Lindemuth R, Schimrigk K. [Generalized myoclonus as the only symptom of neurosyphilis]. Nervenarzt, 1997, 68(10):845–7. 36. Caponnetto C et al. Late symptomatic neurosyphilis presenting as a motor polyradiculoneuropathy. Italian journal of neurological science, 1997, 18(1):62. 26 Aspects ÈpidÈmiologiques.pmd 12/20/2005, 2:39 PM477 478 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Report Nosocomial blood stream infections in Imam Khomeini Hospital, Urmia, Islamic Republic of Iran, 1999–2001 M. Rahbar,1 R. Gra-Agaji 2 and S. Hashemi 2 1Department of Microbiology, Reference Laboratories of Iran, Tehran, Islamic Republic of Iran. 2Faculty of Medicine, University of Medical Sciences, Urmia, Islamic Republic of Iran (Correspondence to M. Rahbar: mhhf_rz@yahoo.com). Received: 02/04/03; accepted: 18/12/03 ABSTRACT In a 2-year retrospective study, the database of the microbiology laboratory of the Imam Khomeini Hospital was reviewed to identify patients who had nosocomial bacteraemia between 1 May 1999 and 31 May 2001 and identify the pathogen responsible and its resisitance to antibiotics. Of 6492 patients in various wards, 593 (9.1%) had positive blood cultures; 85 of those (14.3%) had signs of potential skin contamination. Gram-positive cocci, including coagulase-negative staphylococci, Staphylococcus aureus, Streptococcus pneumoniae and other Gram-positive cocci, accounted for 42.3% of isolates. Gram-negative bacilli were responsible for another 42.3% of isolates; Pseudomonas aeruginosa was the predominant isolate. Patterns of drug resistance varied according to species of bacteria but were generally quite high. Infections nosocomiales sanguines à l’hôpital Imam Khomeini, Urmia (République islamique d’Iran), 1999-2001 RÉSUMÉ Lors d’une étude rétrospective sur deux ans, la base de données du laboratoire de microbiologie de l’hôpital Imam Khomeini a été examinée pour identifier les patients qui avaient eu une bactériémie noso- comiale entre le 1er mai 1999 et le 31 mai 2001 et déterminer les agents pathogènes responsables et leur résistance aux antibiotiques. Sur 6492 patients hospitalisés dans divers services, 593 (9,1 %) avaient une hémoculture positive ; 85 % de ces derniers (14,3 %) présentaient des signes pouvant indiquer une conta- mination cutanée. Les cocci à Gram positif, y compris les staphylocoques à coagulase négative, Staphylo- coccus aureus, Streptococcus pneumoniae et autres cocci à Gram positif représentaient 42,3 % des isolats. Des bacilles à Gram négatif étaient impliqués dans 42,3 autres pour cent des isolats, Pseudomonas aerug- inos étant l’isolat prédominant. La pharmacorésistance variait selon les espèces de bactéries mais elle était généralement assez importante. 27 Nosocomial blood.pmd 12/20/2005, 2:39 PM478 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 479 Introduction Bacteraemia is the presence of bacteria in the bloodstream and may be transient, in- termittent or continuous. Bloodstream in- fections are important causes of morbidity and mortality and many are acquired in hospital. Not only are nosocomial infec- tions increasing, they are more frequently being caused by pathogenic bacteria that are resistant to antibiotics [1–3]. The spectrum of microorganisms that invade the bloodstream has been systemi- cally evaluated in several studies. The or- ganisms most commonly isolated from blood cultures are Gram-positive cocci in- cluding coagulase-negative staphylococci, Staphylococcus aureus, Enterococcus spp. and other organisms likely to inhabit the hospital environment and the gastrointesti- nal tracts of hospitalized patients [4,5]. Clinical symptoms and signs in patients are not sufficiently reliable to predict bacter- aemia with sepsis but early diagnosis and treatment of bloodstream infections are im- portant clinical concerns in order to reduce frequency and increase survival of patients. Epidemiological studies of etiologic agents of bacteraemia and their susceptibility to antimicrobial agents are commonly based on international collaborative studies. How- ever, due to selective pressure from antibi- otic use, especially in developing countries, there is a need for local surveillance [6]. Our objective, therefore, was to review the microbiology database of the Imam Khomeini Hospital in Urmia from 1 May 1999–31 May 2001 for cases of bacter- aemia. Methods Imam Khomeini Hospital is a university af- filiated hospital with more than 300 beds. At this hospital blood culture results have been stored in the central laboratory infor- mation system. Each data set includes the following information for each patient: family name, first name, age, ward, sam- pling date, organism identified and results of antimicrobial susceptibility testing. As per our standard hospital procedure, nurses disinfect skin with 70% alcohol fol- lowed by 2% povidone-iodine before col- lecting blood samples. Typically 10 mL of blood are collected from adults and approx- imately 1–5 mL collected from children at each phlebotomy. Blood samples are trans- ferred to blood culture media and immedi- ately transported to the hospital’s microbiology laboratory. All blood cultures are incubated at 35 °C for at least 2 weeks. In addition to daily visual examination, blood subcultures are performed 6 to 12 hours after the first incubation on well- mixed media and the inoculum is spread onto chocolate agar, blood agar and eosin methylene blue agar plates. All blood agar and chocolate agar plates are incubated in 5%–10% CO2 at 35 °C for 48 hours. Neg- ative bottle cultures are re-incubated for an additional 2 weeks. Isolates are identified by Gram staining and conventional biochemical methods [3,5]. Methods used for identification of Gram positive bacteria are Gram-stain, cat- alase, coagulase, DNase, bile esculin hy- drolysis, growth on NaCl, susceptibility to optochin, colonial morphology and haemol- ysis. Methods used for identification of Gram-negative rods are catalase, oxidase, indole, motility, H2S production, fermenta- tion of sugars and other biochemical tests. Susceptibility is tested using the disk diffu- sion method described by the National Committee for Clinical Laboratory Stan- dards (NCCLS) [7]. A nosocomial bloodstream infection was diagnosed in a clinically ill patient (with fever, chills, hyperventilation and hy- 27 Nosocomial blood.pmd 12/20/2005, 2:39 PM479 480 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 potension) when 1 or more cultures of blood drawn at least 48 hours after admis- sion yielded a pathogenic organism. If the isolated bacteria was a potential skin con- taminant, e.g. Diphtheroid spp., Propioni- bacterium spp., Bacillus spp., coagulase- negative staphylococci or micrococci, all of the following additional criteria were re- quired for diagnosis of nosocomial bacter- aemia: the presence of an intravascular catheter, the initiation of antimicrobial ther- apy, and either temperature > 38 °C or sys- tolic blood pressure < 90 mmHg. Multiple positive blood cultures yielding the same organism were considered a single infec- tion [4,8–11]. Frequencies were calculated using de- scriptive statistics and expressed as per- centages. Results Among the various wards of the Imam Khomeini Hospital, 593 (9.1%) of 6492 pa- tients had positive blood cultures over the 2-year period and were considered to have a nosocomial infection; 85 (14.3%) of these had evidence of potential skin con- tamination. The predominant contaminant bacteria included Diphtheroid spp., Bacil- lus spp., coagulase-negative staphylococci and Propionibacterium spp. Of the other bacteria isolated, Gram-positive cocci and Gram-negative bacilli (42.3%) were iso- loated in equal proportions (42.3% each). Table 1 shows that coagulase-negative staphylococci (isolted from 111 patients or 18.7%) and Staph. aureus (from 107 pa- tients or 18.0%) were the most common causes of nosocomial bacteraemia. Strepto- coccus pneumoniae and other gram-posi- tive cocci were isolated in 4.0% and 1.6% of patients respectively. The most com- monly isolated Gram-negative organisms were Pseudomonas aeruginosa (from 104 patients or 17.5%), Enterobacter spp. (from 48 patients or 8.1%) and Esherichia coli (from 41 patients or 6.9%) (Table 1). In addition to bacteria, we isolated Candida albicans in 5 patients. The frequency of nosocomial blood- stream infections varied with wards. The highest frequency of positive cultures was in the neonatal ward (for neonates aged un- der 45 days). In that ward, 323 (54.5%) patients had positive blood cultures. Indi- vidual pathogens varied slightly in various wards, e.g. P. aeruginosa was most com- mon in the burn (52.8%), haematology (30.5%) and paediatric (18.4%) wards, whereas in the neonatal ward, coagulase- negative staphylococci were predominant (21.9%). The crude mortality rate of bacteraemia due to P. aeruginosa was very high; for example, the mortality rates in the burn, haematology and neonatal wards Table 1 Bacteria most commonly isolated from 6492 blood cultures at Imam Khomeini Hospital, 1 May 1999–31 May 2001 Bacteria No. % Total Gram-positive cocci 251 42.3 Coagulase-negative staphylococci 111 18.7 Staphylococcus aureus 107 18.0 Streptococcus pneumoniae 24 4.0 Other gram-positive cocci 9 1.5 Total Gram-negative bacilli 251 42.3 Pseudomonas aeruginosa 104 17.5 Enterobacter spp. 48 8.1 Esherichia coli 41 6.9 Klebsiella pneumoniae 19 3.2 Citrobacter spp. 19 3.2 Other Gram-negative rods 20 3.4 Candida spp. 6 1.0 Contaminants 85 14.3 Total 593 100 27 Nosocomial blood.pmd 12/20/2005, 2:39 PM480 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 481 were 86.6%, 75.0% and 48.7% respective- ly. Gram-positive bacteria exhibited a sig- nificant rate of antibiotic resistance (Table 2). The antimicrobial resistance rates of Staph. aureus were: penicillin 82.6%; ox- acillin 40.7%; gentamicin 36.5%; trimetho- prim–sulfamethoxazole 7.6% and clinda- mycin 1.9%. No Staph. aureus isolates were resistant to vancomycin or ciproflox- acin. The resistance rates of coagulase- negative staphylococci to the antibiotics were: penicillin 76.8%; oxacillin 61.5%; gentamicin 25.0%; clindamycin 3.2%; and trimethoprim–sulfamethoxazole 6.4%. No coagulase-negative staphylococci were re- sistant to ciprofloxacin or vancomycin (Ta- ble 2). Antimicrobial resistance levels for the Gram-negative organisms most commonly causing nosocomial blood stream infec- tions were relatively high (Table 3). Resis- tances rates of P. aeruginosa to the antibiotics were: gentamicin 32.2%; tobra- mycin 21.5%; and amikacin 21.4%. This organism was not resistant to ciprofloxa- cin. Enterobacter spp. resistance rates to the antibiotics were: ampicillin 47.8%; cephalotin 65.2%; ceftizoxim 30.4%; gen- tamicin 71.7%; tobramycin 56.5%; ciprof- loxacin 0%; and trimethoprim–sulfa- methoxazole 8.6%. Resistance rates of E. coli were: ampicillin 57.4%; cephalotin 62.5%; ceftizoxim 12.5%; gentamicin 22.5%; tobramycin 22.5%; ciprofloxacin 0%; and trimethoprim–sulfamethoxazole 7.5% (Table 3). Discussion The rate of contamination of blood cultures in the microbiology laboratory of Imam Khomeini Hospital was high (14.3%), espe- cially in comparison with other hospitals in other countries [6]. One important source of contamination of blood cultures with bacteria such as Staph. epidermidis, Diph- theroid spp., Propionibacterium spp. and Bacillus spp. is the insufficient disinfecting of skin during blood sampling. Further complicating matters, in our hospital physi- cians often submit 1 instead of 3 samples, making the interpretation of blood culture results difficult [9,12,13]. The spectrum of microorganisms that invade the bloodstream has been systemati- cally evaluated in several studies. In a sur- veillance study of nosocomial bloodstream Table 2 Antimicrobial resistance of Gram-positive cocci isolated from blood cultures of 251 patients at Imam Khomeini Hospital, 1 May 1999–31 May 2001 Antibiotic Staphylococcus Coagulase-negative Streptococcus aureus (%) staphylcocci (%) pneumoniae (%) Penicillin 82.6 76.8 31.8 Oxacillin 40.7 61.5 22.7 Vancomycin 0 0 0 Gentamicin 36.5 25.0 Not tested Ciprofloxacin 0 0 0 Clindamycin 1.9 3.2 0 Trimethoprim– sulfamethoxazole 7.6 6.4 0 27 Nosocomial blood.pmd 12/20/2005, 2:39 PM481 482 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 infection at 49 hospitals in the United States of America over a 3-year period, over 10 000 bloodstream infections were de- tected [8]. Gram-positive organisms were responsible for 64% and Gram-negative or- ganisms were responsible for 27%; 8% were caused by fungi [8]. In our study, 42.3% of infections were caused by Gram- positive cocci and 42.3% by Gram-nega- tive bacilli. The rate of isolation of fungi (1.0%) in our study was low in comparison with other studies [6]. The role of coagulase-negative staphy- lococci in bacteraemia continues to be con- troversial. Until the 1970s, coagulase- negative staphylococci were mainly recog- nized as a contaminant, i.e. as part of the skin flora. Since then, several studies have reported increasing incidence of infections due to coagulase-negative staphylococci [14,15]. In one study, coagulase-negative staphylococci and Staph. aureus were the first and second most common etiologic agents respectively causing nosocomial bloodstream infections in the USA [16]. In another study, Staph. aureus and E.coli were the most commonly isolated blood- stream infections, followed by coagulase- negative staphylococci and enterococci [17]. Klebsiella spp., Enterobacter spp., P. aeruginosa, Strep. pneumoniae and β- haemolytic streptococci were also among the 10 most frequently reported species in both the USA and Canada [17]. In our study coagulase-negative staphy- lococci and Staph. aureus were isolated in almost half the total cases of bacteraemia. Other recent studies report similar results [7,18]. In a general hospital in Kerman in the southern part of our country, Staph. aureus and coagulase-negative staphylo- cocci were isolated from 179 blood speci- mens (45.7%) [18]. This is similar to our study (36.8% for all specimens). Strep. pneumoniae (4.0%) was isolated from blood cultures of 24 patients. This organ- ism is a major and well-known cause of community-acquired infections, but there is increasing interest in its role in the epide- miology of hospital-acquired infection [19]. Table 3 Antimicrobial resistance of Gram-negative bacilli isolated from blood cultures of 251 patients at Imam Khomeini Hospital, 1 May 1999–31 May 2001 Antibiotic Pseudomonas Enterobacter Klebsiella Esherichia aeruginosa spp. (%) pneumoniae coli (%) (%) (%) Ampicillin (not tested) 47.8 72.2 57.4 Cephalotin (not tested) 65.2 27.3 62.5 Ceftizoxim 72.3 30.4 16.6 12.5 Gentamicin 32.2 71.7 27.7 22.5 Tobramycin 21.5 56.5 11.1 22.5 Amikacin 21.4 2.1 0 0 Ciprofloxacin 0 0 15.5 0 Trimethoprim– sulfamethoxazole 90.1 8.6 16.6 7.5 27 Nosocomial blood.pmd 12/20/2005, 2:39 PM482 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 483 The resistance of Staph. aureus and coagulase-negative staphylococci to antibi- otics such as penicillin and oxacillin was high. Nearly 50% of isolates were resistant to oxacillin. Fortunately we did not observe any resistance of Staph. aureus and coagu- lase-negative staphylococci to vancomy- cin. Gram-negative bacilli accounted for ap- proximately half of the nosocomial blood- stream infections and their resistance to commonly used antibiotics was also rela- tively high. Nosocomial bloodstream infec- tions caused by P. aeruginosa were prevalent in the burn ward. The high mor- tality rate from septicaemia caused by P. aeruginosa may be due to resistance of this organism to numerous drugs and also to the poor management of patients [4,5] by physicians. Conclusion Our study provides data on the rate of noscomial infection in our hospital and the organisms involved as well as the antibiotic resistance. This could be of value in the proper antimicrobial therapy of bacter- aemia in our hospital. In addition, this on- going surveillance project will continue to track nosocomial bloodstream infections in Imam Khomeini Hospital and emerging trends in antibiotic resistance. Plans should be made to evaluate the relationship be- tween antibiotic use in the hospital and pat- terns of antimicrobial resistance. References 1. Reimer LG, Wilson ML, Weinstein MP. Update on detection of bacteremia and fungemia. Clinical microbiology reviews, 1997, 10:444–65. 2. Yokoe D et al. Simplified surveillance for nosocomial blood-stream infections. In- fection control and hospital epidemiol- ogy, 1998, 19:657–60. 3. Mylotte JM, Tayara A. Blood cultures: Clinical aspects and controversies. Eu- ropean journal of clinical microbiology and infectious diseases, 2000, 19:157– 63. 4. Fluckiger U et al. Clinical impact of an infectious disease service on the man- agement of bloodstream infection. Euro- pean journal of clinical microbiology and infectious diseases, 2000, 19:493–500. 5. Forbes BA et al. Bailey & Scott’s diagnos- tic microbiology, 10th ed. St. Louis, Mis- souri, Mosby, 1998:283–304. 6. Luzzaro F et al. Prevalence and suscep- tibility of pathogens causing blood- stream infections in Northern Italy: A two- year study in 16 hospitals. European journal of clinical microbiology and in- fectious diseases, 2002, 21:849–55. 7. Performance standards for antimicrobial susceptibility testing. Supplemental tables, M100-S12. Wayne, Pennsylva- nia, National Committee for Clinical Laboratory Standards, 2002. 8. Edmond MB et al. Nosocomial blood- stream infections in United States hospi- tals: a 3-year analysis. Clinical infectious diseases, 1999, 29:239–44. 9. Chandrasekar PH, Brown WJ. Clinical issues of blood cultures. Archives of in- ternal medicine, 1994, 145:841–9. 10. Bryan CS. Clinical implications of posi- tive blood cultures. Clinical microbiology reviews, 1989, 2:329–53. 11. Lundberg G. Rapid classification of posi- tive blood cultures. Journal of the Ameri- can Medical Association, 1992, 267: 1962–6. 27 Nosocomial blood.pmd 12/20/2005, 2:39 PM483 484 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 12. Mermel L, Maki D. Detection of bacter- emia in adults: Consequences of cultur- ing an inadequate volume of blood. Annals of internal medicine, 1993, 119: 270–2. 13. Schifman RB, Bacher P, Howanitz PJ. Blood culture quality improvement. A College of American Pathologists Q- Probes study involving 909 institutions and 289 572 blood cultures sets. Ar- chives of pathology and laboratory medi- cine, 1996, 120:999–1002. 14. Boisson K et al. Characterisation of co- agulase-negative staphylococci iso- lated from blood infections: incidence, susceptibility to glycopeptides and mo- lecular epidemiology. European journal of clinical microbiology and infectious diseases, 2002, 21:660–5. 15. Thylefors JD, Harbarth S, Pittet D. In- creasing bacteremia due to coagulase- negative staphylococci: Fiction or reality? Infection control and hospital epidemiology, 1998, 19:581–9. 16. Marshall SA et al. Staphylococcus aureus and coagulase-negative staphy- lococci from blood stream infections: fre- quency of occurrence, antimicrobial sus- ceptibility and molecular (mecA) characterization of oxacillin resistance in the SCOPE program. Diagnostic mi- crobiology and infectious disease, 1996, 30:205–14. 17. Pfaller MA et al. Bacterial pathogens iso- lated from patients with bloodstream in- fection: frequencies of occurrence and antimicrobial susceptibility patterns from the SENTRY antimicrobial surveillance program (United States of America and Canada, 1997). Antimicrobial agents and chemotherapy, 1998, 42:1762–70. 18. Moshtaghi-Kashani GR. Bacteremia in a general hospital in Kerman: An analysis of 376 isolated bacteria and their antibiogram. Archives of Iranian medi- cine, 2000, 30:178–9. 19. Paradisi F, Corti G. Is Streptococcus pneumoniae a nosocomially acquired pathogen? Infection control and hospital epidemiology, 1998, 19:578–80. 27 Nosocomial blood.pmd 12/20/2005, 2:39 PM484 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 485 Short communication Risk factors for severe hypoglycaemia in type 2 diabetic patients admitted to hospital in Piraeus, Greece A. Sotiropoulos,1 E.A. Skliros,2 C. Tountas,3 U. Apostolou,1 T.A. Peppas1 and S.I. Pappas1 1Third Medical Department and Diabetes Clinic, General Hospital of Nikaea, Piraeus, Greece (Correspondence to A. Sotiropoulos: eskliros@otenet.gr). 2Clinic of Social and Family Medicine, University of Crete, Crete, Greece. 3Faculty of Medicine, University of Athens, Athens, Greece. Received: 14/04/03; accepted: 25/04/04 ABSTRACT A 3-year study assessed the prevalence and causes of severe treatment-related hypoglycae- mia in type 2 diabetes mellitus patients admitted to hospital in Piraeus. Out of the 2858 patients admitted, 207 (7.2%) had severe hypoglycaemia: 72 were being managed with insulin, 132 oral hypoglycaemic drugs and 3 combined insulin/oral drugs. Only 28.5% of patients were attending a diabetes clinic. The cause of the hypoglycaemic attack could be determined for 86.1% of cases; 30.8% were due to a missed meal. Interviews showed that education and level of knowledge about diabetes mellitus, and particularly hypoglycaemia symptoms, was inadequate. Logistic regression analysis showed that knowledge about diabetes mellitus correlated with educational status and with follow-up in a diabetes clinic. Facteurs de risque d’hypoglycémie sévère chez des patients diabétiques de type 2 hospitalisés au Pirée (Grèce) RÉSUMÉ Une étude sur trois ans a permis d’évaluer la prévalence et les causes de l’hypoglycémie sévère associée au traitement chez des patients atteints de diabète sucré de type 2 hospitalisés au Pirée. Parmi les 2858 patients hospitalisés, 207 (7,2 %) avaient une hypoglycémie sévère ; 72 étaient traités par insuline, 132 par hypoglycémiants oraux et 3 par l’association insuline antidiabétiques oraux. Seuls 28,5 % des patients consultaient dans une clinique du diabète. La cause de la crise d’hypoglycémie a pu être déterminée pour 86,1 % des cas, 30,8 % étant dus à un repas sauté. Les entretiens ont montré que l’éducation et le niveau de connaissances sur le diabète sucré, et en particulier les symptômes d’hypoglycémie, étaient insuffisants. L’analyse de régression logistique a indiqué qu’il y avait une corrélation entre les connaissances sur le diabète sucré, le niveau d’instruction et le suivi dans une clinique du diabète. 28 Risk factors for severe.pmd 12/20/2005, 2:39 PM485 486 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Introduction Severe hypoglycaemia is an important ad- verse effect of diabetes mellitus treatment and is accompanied by considerable mor- bidity and mortality [1,2]. In the study of Miller et al., treatment-related severe hy- poglycaemia in a specialty diabetes centre occurred in only 3 patients (0.5%), all due to insulin treatment [3]. In such a setting, patients have strict goals for achieving op- timal glycaemic control, are closely fol- lowed for disease complications and drug side-effects and receive intensive education about diabetes mellitus and the signs and symptoms of hypoglycaemia. Without such care, all these factors are unlikely to be achieved in everyday practice [4,5]. We performed a 3-year study to assess the prevalence and problems of severe treatment-related hypoglycaemia in type 2 diabetes mellitus patients admitted to a hos- pital medical department in Piraeus, Greece. We studied the risk factors for hy- poglycaemia in terms of management of di- abetes and patients’ knowledge about diabetes mellitus and hypoglycaemia. Methods This study included patients with type 2 di- abetes mellitus admitted to the medical de- partment of the General Hospital of Nikaea, Piraeus due to severe hypoglycaemia. Dur- ing this 3-year period (November 1996 to November 1999) a total of 3767 cases of diabetes mellitus were admitted for various reasons; 3631 of these were type 2 diabetes mellitus. Some cases were admitted more than once, so the total number of different type 2 diabetes mellitus patients was 2858. The criteria for diagnosis of severe hy- poglycaemia were: comatose or pre-coma- tose status (according to the Glasgow coma scale) on arrival at the emergency ward, serum glucose level < 2.8 mmol/L (50 mg/dL) and necessity for intravenous glucose administration for resuscitation. Glycosylated hemoglobin (HbA1c) levels were also measured. Structured interviews were performed with patients and their close relatives after recovery to ascertain patients’ demograph- ic details, including education status, the management of their condition (current treatment for diabetes and follow-up atten- dance at a diabetic clinic) and the cause of the hypoglycaemic attack. A previously de- signed questionnaire [6] assessed patients’ level of knowledge about diabetes mellitus. The questionnaire had 33 items in 5 sec- tions: demographic data and history of dia- betes (8 items, not scored); symptoms and complications (8 items); treatment and management (6 items); risk factors (5 items) and monitoring (6 items). Scoring was 1 point per each correct response; maximum score 25. The cut-off score for a good level of knowledge was 65% items correct. The data were analysed using SPSS, version 8.0 for Windows. Means and stan- dard deviations (SD) were computed. Sta- tistical analysis was performed using Student’s t-test for intra- and inter-group comparisons of continuous variables. A lo- gistic regression analysis in a backwards stepwise approach was performed using the knowledge level about diabetes mellitus as the dependent variable and age, sex, edu- cational status and follow-up in diabetes clinics as independent variables. Results Out of the 2858 type 2 diabetes mellitus patients admitted for any reason, 207 (7.2%) were admitted due to severe hy- poglycaemia (85 males, 122 females). The demographic and clinical characteristics of 28 Risk factors for severe.pmd 12/20/2005, 2:39 PM486 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 487 the patients are shown in Table 1. The mean (SD) age was 62.1 (8.7) years and the mean duration of diabetes was 7.4 (2.8) years. Out of these 207 patients, 72 (34.8%) were being managed with insulin, 132 (63.8%) were on oral hypoglycaemic drugs and 3 (1.4%) were being treated with a combination of insulin and oral hypogly- caemics. Only 28.5% of patients were at- tending a diabetes clinic. Of the patients, 140 (70.5%) presented comatose. Of the 61 non-comatose pa- tients, 28 (46.0%) reported adrenergic symptoms and 60 (98.3%) had neurogly- copaenia. The cause of the hypoglycaemic attack was determined as a missed meal for 30.8% of patients and chronic renal failure for 21.9% (Table 1). No cause could be established for 13.9%. The mean (SD) HbA1c level of the pa- tients overall was 6.8% (1.3%). In the pa- tients who were taking oral hypogly- caemics the mean (SD) HbA1c was 6.6% (1.2%), while in the insulin-treated group it was 7.1% (1.5%) (P > 0.05). The mean (SD) venous serum glucose level of patients on arrival at the emergency department was 30 (9) mg/dL for insulin- treated and 33 (12) mg/dL for those on oral drugs (P < 0.05). All patients recovered after administra- tion of intravenous glucose and no fatalities occurred. Cerebrovascular ischaemic stroke was established in 2 patients, while 2 more patients suffered a transient ischaem- ic attack and 3 presented with convulsions which were not repeated or did not need anticonvulsant treatment. The questionnaire was completed by 205 of the 207 patients and showed that the knowledge level of the patients about diabe- tes mellitus, and particularly hypoglycae- mia symptoms, was poor. Knowledge was judged to be “good” in only 14.6% of the sample. In the logistic regression analysis, knowledge about diabetes mellitus correlat- ed with educational status (β = 0.499, SE = Table 1 Demographic and clinical data of 207 diabetic patients Variable Mean (SD) Range Age (years) 62.1 (8.7) 45–88 Duration of diabetes (years) 7.4 (2.8) 1–14 HbA1c level (%) 6.8 (1.3) No. % Sex Male 85 41.1 Female 122 58.9 Presentation Coma 146 70.5 Semi-coma 61 29.5 Usual treatment Insulin 72 34.8 Sulfonylureas 132 63.8 Insulin and sulfonylureas 3 1.4 Follow-up in diabetes clinic Yes 59 28.5 No 148 71.5 Educational status Illiterate 28 13.5 Elementary 117 56.5 Middle 47 22.7 Higher 15 7.3 Diabetes knowledge Poor 175 85.4 Good 30 14.6 Causes of hypoglycaemia Missed meal 76 30.8 Chronic renal failure 54 21.9 Exercise 28 11.4 Alcohol 20 8.2 Dosage error 16 6.5 Unknown 34 13.9 SD = standard deviation. HbA1c = glycosylated haemoglobin. 28 Risk factors for severe.pmd 12/20/2005, 2:39 PM487 488 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 0.05, P < 0.0001) and with follow-up in a diabetes clinic (β = 1.055, SE = 0.08, P < 0.0001). Discussion In our study, out of the total of type 2 dia- betes mellitus patients admitted to the med- ical department, 7.2% were admitted due to severe hypoglycaemia. The percentage of cases where no cause could be established was almost 14%, which is relatively low compared with other similar studies show- ing the percentage where a clear cause for severe hypoglycaemia could be established ranges from 11%–94% [7–10]. We found in our study, as in similar ones, that the causes of hypoglycaemia, especially the most frequent ones (meal omission, alcohol intake, co-existing dis- ease with reduced food intake), respond to prompt intervention and can be managed to prevent hypoglycaemia episodes [11,12]. In our study about two-thirds of hospital- ized severe hypoglycaemia cases were in patients being managed with oral hypogly- caemic drugs. This percentage was less in other studies: about 33% [10,13,14]. The high rate of hypoglycaemia related to oral hypoglycaemic treatment in our study is possibly explained by the fact that cases of insulin-related hypoglycaemia are managed without referral to the emergency depart- ment or are not admitted after prompt intra- venous administration of 10% glucose and recovery in the emergency department. Only 28.5% of the patients were attend- ing a diabetes clinic, a low rate, and, as would be expected, attending a diabetes clinic correlated positively with the level of knowledge about diabetes mellitus. Fur- thermore, the majority of the sample (85.4%) had a poor level of knowledge about diabetes mellitus and hypoglycaemia. The low rate of diabetes clinic attendance and poor level of knowledge about diabetes suggests that primary care physicians in Greece devote little time to diabetic patients and their management and confirms the need for better diabetes care in primary care [15]. As the causes of hypoglycaemia are preventable in the majority of cases, efforts are needed to improve patients’ level of ed- ucation about diabetes mellitus to prevent hypoglycaemia. Such efforts should be the development and implementation of a pro- gramme focussed on the treatment and prevention of diabetes complications, to- gether with continuing education pro- grammes for primary health care professionals. References 1. Seltzer HS. Drug-induced hypoglycemia. A review of 1418 cases. Endocrinology and metabolism clinics of North America, 1989, 18:163–83. 2. Fischer KF, Lees JA, Newman JH. Hy- poglycemia in hospitalized patients. Causes and outcomes. New England journal of medicine, 1986, 315:1245– 50. 3. Miller CD et al. Hypoglycemia in patients with type 2 diabetes mellitus. Archives of internal medicine, 2001, 161:1653–9. 4. MacLeod KM, Hepburn DA, Frier BM. Frequency and morbidity of severe hy- poglycemia in insulin-treated diabetic patients. Diabetic medicine, 1993, 10: 238–45. 5. Gold AE, MacLeod KM, Frier BM. Fre- quency of severe hypoglycemia in pa- tients with type I diabetes with impaired awareness of hypoglycemia. Diabetes care, 1994, 17:697–703. 6. Preventive-care knowledge and prac- tices among persons with diabetes mel- 28 Risk factors for severe.pmd 12/20/2005, 2:39 PM488 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 489 litus—North Carolina, Behavioral Risk Factor Surveillance System, 1994–1995. Morbidity and mortality weekly report, 1997, 46(43):1023–7. 7. Potter J et al. Insulin-induced hypogly- caemia in an accident and emergency department: the tip of an iceberg? British medical journal, 1982, 285:1180–2. 8. Goldgewicht C et al. Hypoglycaemic re- actions in 172 type I (insulin-dependent) diabetic patients. Diabetologia, 1983, 24:95–9. 9. Casparie AF, Elving LD. Severe hypogly- cemia in diabetic patients: frequency, causes, prevention. Diabetes care, 1985, 8:141–5. 10. Gill GV, Huddle KR. Hypoglycaemic ad- missions among diabetic patients in Soweto, South Africa. Diabetic medicine, 1993, 10:181–3. 11. The DCCT Research Group. Epidemiol- ogy of severe hypoglycemia in the dia- betes control and complications trial. American journal of medicine, 1991, 90:450–9. 12. Ter Braak EW et al. Clinical characteris- tics of type 1 diabetic patients with and without severe hypoglycemia. Diabetes care, 2000, 23:1467–71. 13. Leese GP et al. Frequency of severe hy- poglycemia requiring emergency treat- ment in type 1 and type 2 diabetes: a population-based study of health ser- vice resource use. Diabetes care, 2003, 26(4):1176–80. 14. Shorr RI et al. Incidence and risk factors for serious hypoglycemia in older per- sons using insulin or sulfonylureas. Ar- chives of internal medicine, 1997, 157: 1681–6. 15. Barnes CS et al. Little time for diabetes management in the primary care setting. Diabetes education, 2004, 30(1):126– 35. 28 Risk factors for severe.pmd 12/20/2005, 2:39 PM489 490 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Short communication Management of diabetic foot in a Jordanian hospital A.D. Al-Ebous,1 B. Hiasat,1 M. Sarayrah,1 M. Al-Jahmi 1 and A.N. Al-Zuriqat1 1Royal Medical Services, Amman, Jordan (Correspondence to A.D. Al-Ebous: aalebous@hotmail.com). Received: 23/05/01; accepted: 24/03/02 ABSTRACT The effect of different antibiotics on the outcome of surgical care in the management of diabetic foot was investigated. We randomly allocated 100 patients with diabetic foot into one of four groups. Each patient’s infection was graded (Wagner classification). All patients were offered the same surgical care but each group was assigned a different antibiotic. Hospital stay for the four groups ranged from 7 to 14 days. Five patients experienced complications from septicaemia; 15 patients underwent amputation; and five patients experienced temporary renal impairment. Careful consideration to the type of antibiotic used is essential. Prise en charge du pied diabétique dans un hôpital jordanien RÉSUMÉ Nous avons examiné l’effet de différents antibiotiques sur l’issue des soins chirurgicaux dans la prise en charge du pied diabétique. Nous avons réparti de manière aléatoire 100 patients ayant un pied diabétique dans quatre groupes. L’infection de chaque patient a été évaluée (classification de Wagner).Tous les patients ont reçu les mêmes soins chirurgicaux mais chaque groupe s’est vu attribué un antibiotique différent. La durée du séjour hospitalier pour les quatre groupes était comprise entre 7 et 14 jours. Cinq patients ont connu les complications d’une septicémie ; 15 patients ont subi une amputation et cinq patients ont eu une insuffisance rénale passagère. Il est donc essentiel de choisir soigneusement le type d’antibiotique utilisé. 29 Management of diabetic.pmd 12/20/2005, 2:40 PM490 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 491 Introduction The skin and mucous membrane are the first line of defence against invading micro- organisms, and therefore, the integrity of the skin is of paramount importance in pro- tecting against infection. Diabetes mellitus is a common systemic disease associated with a high incidence of infection, especial- ly in the foot. The reasons for this are mul- tifactorial with the most important factor being neuropathy. Angiopathy and poor host defence mechanisms also play a vital role [1,2]. The initial lesions of diabetic foot are usually superficial and caused by a single pathogen such as Staphylococcus aureus or Streptococcus species [3], although an in- creasing number of patients have been en- countered for whom the infecting agent has been Enterococcus species and Pseudomonus aeruginosa [4,5]. We aimed to study treatment outcomes in patients with diabetic foot, including the effect of different antibiotics on the outcome of sur- gical care. Methods The study was conducted at the Prince Hashem Hospital, Jordan between August 1999 and August 2001. All patients admit- ted to the surgical ward diagnosed with di- abetic foot were included in the study (n = 100). The 100 patients were randomly allo- cated to one of 4 groups (A, B, C or D). Treatment consisted of surgical care for all with the type and extent of the surgical in- tervention for the individual patient deter- mined by the grade of infection according to the Wagner classification [6]. Surgical procedures ranged from simple debride- ment to above the knee amputation if con- servative treatment failed. Table 1 shows the distribution among the groups of infec- tion as determined by the Wagner classifi- cation. In addition to surgical procedures, each group had its own antibiotic regime to compare the effectiveness of different anti- biotic treatments on patient outcome. These antibiotics used were the available broad-spectrum antibiotics. Group A re- ceived piperacillin sodium/tazobactam so- dium (4.5 g 3 times daily; group B received ceftazidime (1 g 3 times daily)/metronida- zole (500 mg 3 times daily); group C re- ceived lincomycin hydrochloride (600 mg 2 times daily)/amikacin (500 mg 2 times daily and group D received imipenem. Therapy commenced prior to isolation of the etiologic agent due to the need for ur- gent treatment of infections requiring im- mediate intervention. Results The mean healing time in the different groups is shown in Table 2 together with the number who experienced septicaemia, amputation and renal impairment. The heal- ing time was measured from the start of treatment until the wound was clean and healthy looking with no evidence if dis- Table 1 Number of diabetic foot patients by Wagner grade of infection Grade Group Group Group Group A B C D Grade 0 1 0 1 0 Grade 1 4 3 5 2 Grade 2 10 11 9 12 Grade 3 2 3 3 2 Grade 4 3 2 3 2 Grade 5 5 6 4 7 Total 25 25 25 25 29 Management of diabetic.pmd 12/20/2005, 2:40 PM491 492 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 charge. The mean healing time was 7 days for group A; 10 days for group B; and 14 days for groups C and D. There were 2 patients who experienced complications from septicaemia in each of groups B and D, one in group C, and none in group A. Amputation was performed on 3 patients in group A, 5 patients in groups B and D each, and 2 patients in group C. Renal impairment was observed in 4 patients in group A and 1 patient in group C (Table 2). Discussion Clinical illness of patients with skin and soft tissue infections should be evaluated to determine whether hospital admission is re- quired. In this study, the extent of infec- tion, the presence of devitalized tissue, the state of the vascular supply and the pa- tient’s general condition were assessed. The option of timely surgical interven- tion was considered for patients with deep infection, especially grades 3, 4 and 5 ac- cording to the Wagner classification. Suc- cessful treatment of these grades was highly dependent upon early debridement or, when appropriate, amputation. Such ac- tion resulted in rapid resolution of the prob- lem and a consequent reduction in hospital stay. Control of blood glucose levels did not present any problems as long as the infec- tion was kept under control. Our results suggest that antibiotics con- tributed little to the outcome of patients with these deep tissue infections. However, the results demonstrated the importance of antimicrobial therapy in the treatment of skin and soft tissue infection [7,8] during the management of grade 1 and 2 infec- tions. A comparison between types of anti- microbial agents found that patients using piperacillin sodium/tazobactam sodium (group A) had a shorter hospital stay and a lower incidence of septicaemia. Routine testing of patients’ liver and re- nal functions helped in the detection of ear- ly deterioration in renal function in 5 patients, 4 of whom were in group A. For- tunately, for each of these patients renal impairment was reversible with renal func- tion parameters returning to normal follow- ing discontinuation of antibiotic therapy. Conclusion Management of diabetic foot requires a team approach. Grade of infection deter- Table 2 Antibiotic regime, healing time and number of patients with septicaemia, amputation and/or renal impairment Variable Group A Group B Group C Group D Antibiotic regime Piperacillin Ceftazidime/ Lincomycin Imipenem sodium/tazobactam metronidazole hydrochloride/ sodium amikacin Mean healing time (days) 7 10 14 14 Septicaemia (No.) 0 2 1 2 Amputation (No.) 3 5 2 5 Renal impairment (No.) 4 0 1 0 29 Management of diabetic.pmd 12/20/2005, 2:40 PM492 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 493 mines treatment modality, whether surgical or medical, the former being of paramount importance in deep infection. Antimicrobial therapy was effective in mild infection. Careful consideration of the type of antimi- crobial therapy is required for optimal man- agement of diabetic foot. References 1. Levin ME, O’Neal LW, eds. Diabetic foot, 3rd ed. St Louis, Mosby, 1982. 2. Wheat LJ. Infection and diabetes melli- tus. Diabetes care, 1980, 3:187–97. 3. Lipsky BA, Pecoraro RE, Wheat LJ. The diabetic foot. Soft tissue and bone infec- tion. Infectious disease clinics of North America, 1990, 4:409–32. 4. Tan JS, File TM, Salstrom SJ. Timentin versus moxalactam in the treatment of skin and soft tissue infections. American journal of medicine, 1985, 79:130–3. 5. File TM Jr, Tan JS. Ticarcillin–clavulanate therapy for bacterial skin and soft tissue infections. Reviews of infectious dis- eases, 1991, 13(suppl. 9):S733–6. 6. Wagner FW Jr. The dysvascular foot: a system for diagnosis and treatment. Foot & ankle, 1981, 2:64–122. 7. File TM Jr, Tan JS. Treatment of bacterial skin and soft tissue infections. Sur- gery, gynecology & obstetrics, 1991, 172(suppl.):17–24. 8. Sexton DJ et al. Twice daily intramuscu- lar imipenem/cilastatin in the treatment of skin and soft tissue infections. Chemo- therapy, 1991, 37(suppl. 2):26–30. 29 Management of diabetic.pmd 12/20/2005, 2:40 PM493 494 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Short communication Hepatitis B surface antibody response of household contacts of hepatitis B virus carriers in Palestine K. Adwan,1 N. Abu-Hasan,1 G. Adwan1 and K. Abu-Khater1 1Department of Biological Sciences, An-Najah Nablus University, Nablus, Palestine (Correspondence to K. Adwan: adwank@yahoo.com). Received: 23/02/03; accepted: 22/03/04 ABSTRACT To evaluate the effectiveness of hepatitis B virus (HBV) vaccination of household contacts of HBV carriers in Tulkarm district, Palestine, quantitative hepatitis B surface (anti-HBs) antibody response in 161 household contacts was measured after vaccination. A seroprotective anti-HBs response (titre ≥ 10 IU/L) was elicited in all vaccinated subjects. Of these 2.5% had titres of 10–99 IU/L, 61.5% 100–999 IU/L and 36.0% ≥ 1000 IU/L. The number of vaccination doses had no effect on the achievement of seroprotection. HBV infection was demonstrated in 13 cases and their anti-HBV titres were in the range 25–350 IU/L. Réponse en anticorps dirigés contre l’antigène de surface de l’hépatite B des contacts familiaux de porteurs du virus de l’hépatite B en Palestine RÉSUMÉ Afin d’évaluer l’efficacité de la vaccination contre le virus de l’hépatite B (VHB) des contacts familiaux de porteurs du virus de l’hépatite B dans le district de Tulkarem (Palestine), on a mesuré la réponse quantitative en anticorps dirigés contre l’antigène de surface du virus de l’hépatite B (anti-HBs) chez 161 contacts familiaux après vaccination. Tous les sujets vaccinés présentaient des titres séroprotecteurs (titres d’anti-HBs ≥ 10 UI/L). Parmi ceux-ci, 2,5 % avaient un titre de 10-99 UI/L, 61,5 % de 100-999 UI/L et 36,0 % supérieur ou égal à 1000 UI/L. Le nombre de doses de vaccin reçues n’avait aucun effet sur l’apparition de la séroprotection. Une infection par le VHB a été démontrée dans 13 cas et le titre d’anticorps anti-VHB chez ces derniers était compris entre 25 et 350 UI/L. 30 Hepatitis B surface.pmd 12/20/2005, 2:40 PM494 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 495 Introduction The hepatitis B virus (HBV) is one of the most common chronic pathogens in the world. Over 2 billion of the world’s popula- tion has been exposed to this virus. About 350 million of these, 5% of the world’s population, are chronic carriers [1–4]. An- nually up to 1 million of this population dies due to the consequences of this infection, such as cirrhosis and hepatocellular carci- noma [5]. More than three-quarters of HBV infections occur in Asia, the Middle East and Africa [1]. The risk of HBV infection varies among different groups. Besides medical profes- sionals, drug users and people with multiple sexual partners, other high-risk groups in- clude household contacts of HBV infection carriers [6]. Without ignoring current trends and recommendations for universal vaccination against HBV, the Immunization Practices Advisory Committee (ACIP) of the Centers for Disease Control and Prevention has in- sisted on the importance of vaccination of all persons belonging to identified risk groups, including household contacts of HBV carriers [7–10]. In Palestine, an area of high endemicity of HBV carriers [1], the Ministry of Health established an obligatory hepatitis B vacci- nation of household contacts of HBV carri- ers and of other high-risk groups in 1994. All household contacts of HBV carriers have to be vaccinated against HBV with 3 intramuscular vaccine doses according to the schedule 0 (initial), 1 and 6 months [11]. The objective of this study was to evaluate the hepatitis B surface antibody response to vaccination of household contacts of HBV carriers in northern Palestine. Methods The subjects of this study were all house- hold contacts of confirmed HBV chronic carriers (n = 161) of 50 families in Tulkarm district, Palestine. The median age was 23.6 years (range 3 to 55 years). Data was collected from records of the epidemiolog- ical service of the Ministry of Health of Pal- estine about the HBV status of subjects according to 3 markers: hepatitis B surface antigen (HBsAg), antibody to hepatitis B surface antigen (anti-HBs) and antibody to hepatitis B core antigen (anti-HBc). Only those with all 3 HBV markers negative were enrolled in the vaccination programme. For adults, 20 µg of recombinant, mammalian cell derived HBsAg (Bio-Hep- B™/Sci-B-Vac™, Bio-Technology General, Israel) was administered in the deltoid re- gion at 0, 1 and 6 months. The dose for individuals less than 11 years of age was 10 µg. The subjects who received 3 doses of vaccine, regardless of the between-dose in- terval, were considered completely vacci- nated. Blood samples were collected from all subjects at 6 months after enrolment in the vaccination programme. The serological anti-HBs antibody titrations using an inter- national standard were carried out at An- Najah National University, Palestine during the period 1999 to 2000 using a commer- cial kit (Bioelisa, Biokit, Barcelona, Spain). Anti-HBs antibody titres greater than or equal to 10 IU/L were considered to be se- roprotective against HBV disease [12–14]. In addition, HBsAg was determined using Auszyme® Monoclonal (Abbott Laborato- ries, Abbott Park, Illinois) as a marker for recent (< 3 months) or chronic (carrier) infection. 30 Hepatitis B surface.pmd 12/20/2005, 2:40 PM495 496 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Statistical analysis was performed using the SPSS program. Unpaired Student t-test was used to compare mean values between different independent groups. P < 0.05 was considered statistically significant. Results Of the 161 participants, 90 (55.9%) were male and 71 (44.1%) were female. The mean age of the study group was 23.6 years (range 3–55 years). Two-thirds of the participants (67.1%) completed the vaccination series (3 doses of the vaccine), 31.1% received 2 doses and 1.9% received 1 dose (Table 1). Re- gardless of the dosing schedule, all the sub- jects achieved seroprotection. Of the participants, 2.5% presented anti-HBV titres of 10–99 IU/L, 61.5% 100– 999 IU/L and 36.0% ≥ 1000 IU/L. The mean (standard deviation) value of anti- HBV titres were higher in the group receiv- ing 3 doses [1843 (585) IU/L] than in the group receiving 1 or 2 doses [1455 (415) IU/L]. However, no statistically significant difference was observed between the groups completing and not completing the vaccination schedules (P > 0.05). Testing of the study group for the pres- ence of HBsAg revealed that 13 (8.1%) of the 161 subjects had HBV infection. None of these subjects was haemophiliac or had a history of blood transfusion or haemodi- alysis. Of them, 4 had anti-HBV titres in the range 25–99 IU/L and the remaining sub- jects had anti-HBV titres in the range 110– 350 IU/L. Discussion Systematic vaccination of individuals at risk of exposure to the virus has been the main method of controlling the morbidity and mortality associated with hepatitis B. All studies of the antibody response to currently licensed hepatitis B vaccines have shown that between 5% and 10% or more of healthy immunocompetent subjects do not mount an antibody response to the sur- face antigen component present in these preparations (non-responders) or that they respond poorly (hyporesponders) [12]. The minimal protective titre has been assumed almost universally to be 10 IU/L, and immunological memory is thought to ensure protection even after circulating an- tibody becomes undetectable [12–14]. In the United Kingdom, a healthy vaccinee who develops an anti-HBs titre of < 100 IU/ mL is considered to be unprotected and at risk of HBV [15]. In our study, regardless of dosing schedule, protective anti-HBs were ob- served in all vaccinated subjects. Despite the extremely high rate of response among the subjects, which is higher than reported in numerous studies [12–14], HBV infec- tion was demonstrated in 13 (8.1%) of the studied group. Of these, 4 subjects were Table 1 Serological data for 161 household contacts of hepatitis B virus (HBV) carriers Variable No. % No. of vaccinations 1 3 1.9 2 50 31.1 3 108 67.1 Anti-HBV (IU/L)a 10–99 4 2.5 100–999 99 61.5 > 1000 58 36.0 HBV infection Yes 13 8.1 No 148 91.9 aTitres ≥ 10 IU/L were considered seroprotective. 30 Hepatitis B surface.pmd 12/20/2005, 2:40 PM496 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 497 hyporesponders. The remaining subjects had anti-HBs titres in the range 110–350 IU/ L. It is worth noting that only 3 of the 13 subjects with verified HBV infection did not complete the vaccination series. This find- ing is consistent with the fact that no em- pirical data are available for the hepatitis B surface antibody titre required for protec- tion against particular routes of infection or the size of the infectious inoculum at the time of exposure. Our study also looked at the effect of the number of vaccination doses, and found that this variable had no effect on the achievement of seroprotection. We conclude that people at risk of a poor response should be tested after com- pletion of HBV vaccination and, if neces- sary, offered additional doses of vaccine. References 1. Andre F. Hepatitis B epidemiology in Asia, the Middle East and Africa. Vac- cine, 2000, 18(suppl. 1):S20–2. 2. Kane M. Global status of hepatitis B immunisation. Lancet, 1996, 348:696. 3. Koziol DE, Henderson DK. Risk analysis and occupational exposure to HIV and HBV. Current opinion in infectious dis- eases, 1993, 6:506–10. 4. Michel AL et al. Immunotherapy of chronic hepatitis B by anti HBV vaccine: from present to future. Vaccine, 2001, 19:2395–9. 5. Mast EE, Alter MJ, Margolis HS. Strate- gies to prevent and control hepatitis B and C virus infections: a global perspec- tive. Vaccine, 1999, 17:1730–3. 6. Milas J et al. Hepatitis B in the family. Eu- ropean journal of epidemiology, 2000, 16:203–8. 7. Centers for Disease Control and Pre- vention. Hepatitis B virus: a comprehen- sive strategy for eliminating trans- mission in the United States through uni- versal childhood immunization: recom- mendations of the Advisory Committee on Immunization Practices (ACIP). Mor- bidity and mortality weekly report, 1991, 40(RR-13):1–19. 8. Centers for Disease Control and Pre- vention. Recommendations to prevent hepatitis B virus transmission—United States. Morbidity and mortality weekly report, 1995, 44:574–5. 9. Centers for Disease Control and Pre- vention. Recommendations to prevent hepatitis B virus transmission—United States. Morbidity and mortality weekly report, 1999, 48:33–4. 10. Bonanni P. Universal hepatitis B immuni- zation: infant, and infant plus adolescent immunization. Vaccine, 1998, 16(suppl.): S17–22. 11. Woodruff BA et al. Progress toward inte- grating hepatitis B vaccine into routine infant immunization schedules in the United States, 1991 through 1994. Con- necticut Hepatitis B Project Group. Pedi- atrics, 1996, 97:798–803. 12. Westmoreland D et al. Immunization against hepatitis B—what can we ex- pect? Results of a survey of antibody re- sponse to immunization in persons “at risk” of occupational exposure to hepati- tis B. Epidemiology and infection, 1990, 104:499–509. 13. Duval B et al. Preadolescent non- and hyporesponders following three doses of hepatitis B vaccine need only one more dose. Vaccine, 2002, 20:3632–4. 14. Jacques P et al. The immunogenicity and reactogenicity profile of a candidate 30 Hepatitis B surface.pmd 12/20/2005, 2:40 PM497 498 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 hepatitis B vaccine in an adult vaccine non-responder population. Vaccine, 2002, 20:3644–9. 15. Zuckerman JN et al. Immune response to a new hepatitis B vaccine in healthcare workers who had not responded to stan- dard vaccine: randomised double blind dose–response study. British medical journal, 1997, 314:329–33. Hepatitis B surface antigen Assays: operational characteristics. Phase I. Report 2 In 1998, the World Health Organization, Blood Safety and Clinical Technology Department, conscious of the need to advise Member States on laboratory aspects associated with Hepatitis B and Hepati- tis C testing for blood transfusion safety, initiated a project to pro- vide objective assessments of commercially available assays for detection of Hepatitis B surface antigen (HBsAg) and Hepatitis C (HCV) antibodies, similar to that which has existed for HIV since 1988. This second report presents the findings of the Phase I evalu- ations of 5 HBsAg assays conducted between September 2001 and January 2004. The HBsAg assays evaluated included: • Enzygnost HBsAg 5.0 (Dade Behring Inc) • Equipar HBsAg One Step (Equipar Diagnostici) • Genedia HBsAg ELISA 3.0 (Green Cross Life Science Corp) • HEPALISA (J Mitra & Co) • Murex HBsAg Version 3 (Abbott-Murex). Copies of these reports are available on request from the Depart- ment of Essential Health Technologies (EHT), World Health Organiza- tion, 1211 Geneva 27, Switzerland. 30 Hepatitis B surface.pmd 12/20/2005, 2:40 PM498 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 499 Short communication Epidemiology of malaria in New Halfa, an irrigated area in eastern Sudan Y.El-S. Himeidan,1 M.I. Elbashir,2 El-A. El-Rayah 2 and I. Adam3 1University of Kassala, Kassala, Sudan. 2University of Khartoum, Khartoum, Sudan. 3New Halfa Hospital, New Halfa, Sudan (Correspondence to I. Adam: ishagadam@hotmail.com). Received: 12/10/03; accepted: 17/05/04 ABSTRACT A prospective study investigated the epidemiology of malaria in an agricultural area in eastern Sudan from November 1999 to June 2000 (1 irrigation and 1 dry season). In monthly parasite surveys, 99/ 1539 blood films from 190 individuals were positive for malaria: 95% for Plasmodium falciparum, 3% P. vivax and 2% P. ovale. The slide positive rate (SPR) of malaria episodes ranged from 3.7% to 12.8% in different months of the survey, with a peak in January. There was no significant difference in SPR between irrigation and dry seasons (7.2% versus 5.1%). SPR differed significantly by age group and was highest in under 5- year-olds. However, there was no significant difference in SPR between males and females. Transmission and intensity of malaria in this area is perennial and moderate rather than low. Épidémiologie du paludisme à New Halfa, une zone irriguée du Soudan oriental RÉSUMÉ Une étude prospective a examiné l’épidémiologie du paludisme dans une zone agricole du Soudan oriental de novembre 1999 à juin 2000 (1 saison d’irrigation et 1 saison sèche). Dans les enquêtes parasi- taires mensuelles, 99 des 1539 étalements de sang prélevé sur 190 personnes étaient positifs pour le paludisme : 95 % pour Plasmodium falciparum, 3 % pour P. vivax et 2 % pour P. ovale. Le taux de lames positives pour les épisodes de paludisme se situait entre 3,7 % et 12,8 % pour les différents mois de l’étude, avec un pic en janvier. Il n’y avait pas de différence significative dans le taux de lames positives entre la saison d’irrigation et la saison sèche (7,2 % contre 5,1 %). Le taux de lames positives différait significative- ment selon le groupe d’âge et était le plus élevé chez les moins de 5 ans. Toutefois, il n’y avait pas de différence significative selon le sexe dans le taux de lames positives. La transmission et l’intensité du paludisme dans cette zone perdurent et sont modérées plutôt que faibles. 31 Epidemiology of malaria.pmd 12/20/2005, 2:40 PM499 500 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Introduction Malaria accounts for 10% of Africa’s total disease burden [1]. The epidemiology of malaria is important in terms of under- standing the basic immunological process- es, as well as in deciding the control strategies to apply. It has been suggested that the clinical course of malaria is likely to differ according to the transmission in the area. Therefore, in areas where malaria transmission is seasonal and unstable, the disease burden is confined to a wide age range, and adults as well as children suffer severe morbidity. In areas with intense transmission, however, the burden of dis- ease is confined to the youngest age groups, as adults would have already devel- oped immunity and the highest incidence would be among children under 5 years old [2–4]. While the epidemiology of malaria has been studied in other areas of Sudan [5–7], no data is available from the New Halfa area of eastern Sudan, which has the second largest permanent irrigation scheme in Sudan. Permanent irrigation has been re- ported to influence the transmission of ma- laria adversely [8,9]. The present study was conducted to continue our previous entomological work in this area of eastern Sudan [10], to investigate the morbidity pattern of malaria in the area and to estab- lish adequate baseline data for evaluation of the effectiveness of various preventive measures, including future vaccines. Methods Study area The study was carried out in 2 localities (Heielmasakine and Dibaira camp), which are surrounded by free-hold farms in the north and south of New Halfa town, re- spectively. The study area has been de- scribed in detail elsewhere [10]. In summa- ry, New Halfa area lies in the semi-arid belt of Sudan approximately 500 km east of Khartoum in the middle of an agricultural scheme. Cotton, wheat, sorghum and vari- ous vegetables are cultivated in area of 400 000 feddans. The total rainfall and the average temperature during the study peri- od were 431.6 mm and 30 °C respectively. Plasmodium falciparum is the predominant malaria parasite species, and has been shown to be 75.0% and 9.6% resistant to chloroquine and quinine respectively [11]. Parasitology survey A cohort of 195 individuals from all age groups was randomly selected, which rep- resented 14% of the population in both lo- calities (95 from Heielmasakine and 100 from Dibaira camp). Of these, 5 cases were excluded initially from the study as they were positive for P. falciparum. Blood samples were taken from patients every month during the period November 1999 to June 2000 to determine the rate of new cases of malaria parasite infection. The first 4 months (1 November to 29 Febru- ary) were considered an irrigation season and the remaining 4 months (1 March to 30 June) were considered the dry season. After obtaining verbal consent from pa- tients, a closed-ended questionnaire was completed to determine their sociodemo- graphic characteristics. A blood sample was taken by finger prick for preparation of thick and thin smears. Blood films were Giemsa-stained and the thick films were examined for malaria parasites. If these were positive, the thin films were read to determine the parasite species. A slide was considered negative after examination of 200 fields of thick blood films. Clinical ma- laria was defined as a positive slide and a body temperature ≥ 37.5 °C. 31 Epidemiology of malaria.pmd 12/20/2005, 2:40 PM500 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 501 Patients who had blood films positive in 2 consecutive months were classified and reported as treatment failure rather than re- infection. The slide positive rate (SPR) (%) was expressed as the number of positive slides/total number of slides examined As a high level of chloroquine-resistant P. falciparum species is present in the area [11], all patients who had clinical symp- toms confirmed by positive blood films were treated with pyrimethamine/sulfadox- ine (Fansidar) followed by quinine if neces- sary. Statistical analysis The data were analysed using SPSS, ver- sion 10. Percentages, means and standard deviations (SD) were calculated. The dif- ference in the rate of parasitaemia between different age groups, seasons, sites and sexes were compared using chi-squared tests. P < 0.05 was considered significant. Results Of the final cohort of 190 patients, 93 (48.9%) were males and 97 (51.1%) fe- males. The mean (SD) age was 15.4 (13.9) years and mean (SD) weight 35.0 (20.4) kg. During the 8 monthly surveys, 1542 blood films for malaria were prepared for the 190 patients, and 102 slides were found to be positive. Three positive slides were excluded and considered treatment failure as they were found positive from the same individuals in 2 consecutive months. Thus 99/1539 slides were positive, an overall SPR of 6.4% (95% CI: 3.9–9.0). Among the 99 positive slides, 95% were positive for P. falciparum, 3% for P. vivax and 2% for P. ovale. The SPR of parasitaemia was 7.3% in Dibaira camp and 5.7% in Heielmasakine (P > 0.05). The SPR ranged from 3.7% to 12.8% during the different periods of the study. The peak was in January (12.8%), gradually decreasing in the following dry 4 months and starting to rise at the beginning of the rainy season in June (Figure 1). There was no statistically significant differ- ence in the overall SPR of malaria parasi- taemia between the irrigation and the dry seasons: 7.2% versus 5.1% (P > 0.05). The SPR in different age groups in the 2 seasons (Table 1) was significantly higher in children < 5 years (10.4%) and 5–9 years (9.7%) than in those aged ≥ 20 years (1.9%) (P < 0.0001). The SPR was higher in males (50/646, 7.7%) than in females (49/893, 5.5%), but this was not statistical- ly significant. Discussion This is the first epidemiological survey to investigate malarial morbidity in an agricul- tural area in eastern Sudan. The area has been characterized by a high level of chlo- roquine resistant falciparum malaria and a relatively high human biting rate of the prin- ciple vector Anopheles arabiensis [10,11]. P. falciparum was the predominant species detected during the survey. This agrees with our previous study of drug resistance in the same area [11]. In a nearby area, El Gadaref, recent reports showed very simi- lar findings, and the presence of P. vivax was attributed to population influx from neighbouring Ethiopia [5,6]. However, in the last parasitology survey in the capital, Khartoum, relatively high percentages of P. ovale and P. vivax (8.2% and 6.9% respec- tively) were reported and this was ex- plained by the mixture of inhabitants in Khartoum, many of whom are internally displaced people from south Sudan [7]. 31 Epidemiology of malaria.pmd 12/20/2005, 2:40 PM501 502 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 The study showed that the monthly SPR of malaria ranged from 3.7% to 12.8% during the study period, with the peak during January, 3 months following the rainy season. This rate is much higher than that estimated in a nearby village (Asar) locate about 150 km from New Hal- fa town where there is no permanent irriga- Figure 1 Seasonal variation of Plasmodium falciparum infection in the New Halfa area of Sudan during the study period (November 1999 to June 2000) Table 1 Slide positive rate (SPR) of Plasmodium falciparum infection in the New Halfa area of Sudan by age and season for 190 patients during the study period (November 1999 to June 2000) Age Irrigation season Dry season Total P-valueb (years) No. +ve/ SPR No. +ve/ SPR No. +ve/ SPR tested (%) tested (%) tested (%) < 5 22/188 11.7 10/121 8.3 32/309 10.4 0.333 5–9 23/220 10.5 10/120 8.3 33/340 9.7 0.528 10–14 10/151 6.6 5/95 5.3 15/246 6.1 0.664 15–19 9/93 9.7 1/76 1.3 10/169 5.9 0.022 ≥ 20 6/320 1.9 3/155 1.9 9/475 1.9 0.964 Total 70/972 7.2 29/567 5.1 99/1539 6.4 0.130 P-valuea < 0.001 0.032 aComparing across age groups. bComparing irrigation and dry seasons. 31 Epidemiology of malaria.pmd 12/20/2005, 2:40 PM502 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 503 tion system [12]. We have previously ob- served this peak in an earlier survey in this area and we proposed that it might be the result of irrigation of the main crops during this period [10]. Similar findings of this peak during the irrigation of crops has been reported from the area around the Gezira irrigation scheme in central Sudan [13] as well as in other African countries. For ex- ample, in the Tigray region of Ethiopia the overall incidence of malaria was 7 times higher in villages close to micro-dams compared with control villages [9], and in the Benoue River valley of northern Came- roon, malaria incidence increased after the large-scale irrigation development follow- ing completion of the Lagdo dam [8,9]. In the present study, the peak of infection dur- ing January and the presence of malaria cases during all periods of the study—with no statistically significant difference in SPR between the dry and irrigation sea- sons—was different from the incidence of malaria in eastern Sudan, which varies con- siderably from year to year depending on rainfall [14,15]. Thus, the permanent irri- gation in this area has influenced the pat- tern of the disease and the transmission is perennial rather than seasonal. Although malaria episodes were ob- served in all age groups, the SPR varied considerably between different age groups. The most noticeable observation was the consistently reduced risk of having clinical malaria after the age of 20 years compared with infants and young children (0–9 years of age). The latter represents 66.7% of all positive cases detected in the area and the difference was highly significant (P < 0.0001). Recent work on the age-specific risk for malaria in eastern Sudan showed that the prevalence of malaria was high up to the age of 19 years [5,15], whereas in the capital Khartoum, the susceptibility to malaria was higher in the age group 2–9 years compared with those aged over 15 years [7]. However, in areas of intense transmission the main burden of the disease is usually confined to children under 5 years old [3]. Nevertheless, our findings do not rule out an intrinsic role for age in the development of natural immunity against P. falciparum malaria, but it does indicate that malaria transmission in this area is more likely to be moderate or high because the transmission intensity is roughly reflected in the point prevalence of parasitaemia in children [16]. The study showed that the SPR of ma- laria was not statistically different between males and females. The sex variation on malaria risk in an area of low and seasonal transmission in eastern Sudan varied from one season to another [5]. However, in a population close to Lake Zwai in central Ethiopia the malarial episodes were signifi- cantly higher in males than females. The phenomenon was explained by the obser- vation that males are more likely to spend the evening hours outdoors, a favourable time for vector biting [17]. In conclusion, the findings of the study suggest that preventive measures against malaria such as chemoprophylaxis and bednets, should be employed in the New Halfa area for all age groups and at all times of the year. References 1. Roll back malaria: advocacy brochure. Geneva, World Health Organization, 1999 (http://www.emro.who.int/rbm/ AMD03/Advocacy%20Eng.pdf, acce- ssed 18 May 2005). 31 Epidemiology of malaria.pmd 12/20/2005, 2:40 PM503 504 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 2. Snow RW et al. Relation between severe malaria morbidity in children and level of Plasmodium falciparum transmission in Africa. Lancet, 1997, 349:1650–4. 3. Snow RW et al. Models to predict the in- tensity of Plasmodium falciparum trans- mission: applications to the burden of disease in Kenya. Transactions of the Royal Society of Tropical Medicine and Hygiene, 1998, 92:601–6. 4. Trape JF et al. The Dielmo project: a lon- gitudinal study of natural malaria infec- tion and the mechanism of protective immunity in a community living in a holoendemic area of Senegal. American journal of tropical medicine and hy- giene, 1994, 5:123–37. 5. Giha HA et al. The epidemiology of fe- brile malaria episodes in an area of un- stable and seasonal transmission. Transactions of the Royal Society of Tropical Medicine and Hygiene, 2000, 94:645–51. 6. Hamad AA et al. A marked seasonality of malaria transmission in two rural sites in eastern Sudan. Acta tropica, 2002, 83: 71–82. 7. El Sayed BB et al. Study of the urban malaria transmission problem in Khartoum. Acta tropica, 2000, 25:163– 71. 8. Robert V et al. Mosquitoes and malaria transmission in irrigated rice-fields in the Benoue valley of northern Cameroon. Acta tropica, 1992, 52:201–4. 9. Ghebreyesus TA et al. Incidence of ma- laria among children living near dams in northern Ethiopia: community based in- cidence survey. British medical journal, 1999, 319:663–6. 10. Himiedan YE et al. Anopheles ara- biensis: abundance and insecticide re- sistance in an irrigated area of eastern Sudan. Eastern Mediterranean health journal, 2003, 2004, 10(1/2):167–74. 11. Adam I et al. In the Sudan: chloroquine resistance is worsening and quinine re- sistance is emerging. Sudan medical journal, 2001, 39:5–11. 12. Babiker HA. Unstable malaria in Sudan: the influence of the dry season. Plasmo- dium falciparum population in the un- stable malaria area of eastern Sudan is stable and genetically complex. Transac- tions of the Royal Society of Tropical Medicine and Hygiene, 1998, 92:585–9. 13. El Gaddal AA et al. Malaria control in the Gezira-Managil Irrigated Scheme of the Sudan. Journal of tropical medicine and hygiene, 1985, 88:153–9. 14. Roper C et al. Seasonal changes in the Plasmodium falciparum population in in- dividuals and their relationship to clini- cal malaria: a longitudinal study in a Sudanese village. Parasitology, 1998, 116:501–10. 15. Theander TG. Unstable malaria in Sudan: the influence of the dry season. Malaria in areas of unstable and sea- sonal transmission. Lessons from Daraweesh. Transactions of Royal Soci- ety of Tropical Medicine and Hygiene, 1998, 92:589–92. 16. Gilles HM. Epidemiology of malaria. In: Gilles HM, Warrell DA, eds. Bruce– Chwatt’s essential malariology, 3rd ed. London, Edward Arnold, 1993:124–64. 17. Abose T et al. Re-orientation and defini- tion of the role of malaria vector control in Ethiopia. Geneva, World Health Organi- zation, 1998 (WHO/MAL/98.1085). 31 Epidemiology of malaria.pmd 12/20/2005, 2:40 PM504 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 505 Short communication Urinary tract infection among pregnant women in Al-Mukalla district, Yemen A.M. Al-Haddad Department of Microbiology, Hadhramout University, College of Medicine (HUCOM), Al-Mukalla, Yemen (Correspondence to A.M. Al-Haddad: ahmed_al_haddad@yahoo.com). Received: 27/01/03; accepted: 14/04/04 ABSTRACT Urinary tract infection (UTI) is a common problem in pregnant woman. This study examined the frequency of UTI in 137 pregnant women attending Al Mukalla maternity hospital from January to June 2002. Urine samples were examined for UTI microscopically and by culture, and sensitivity tests were done for the organisms isolated using a range of antibiotics. Information on age, trimester, parity and number of pregnan- cies were also collected for each woman. This study showed that 30% of the women suffered from UTI; Escherichia coli was the most frequently isolated organism (41.5%), and it was highly susceptible to chloram- phenicol, ciprofloxacin, ceftizoxime and amikacin. Of the variables examined, 53.7% of the infected women were in the age group 15–24 years, 48.8% were in their 3rd trimester and 75.6% had 1–3 children. L’infection urinaire chez la femme enceinte dans le district d’Al Mukalla (Yémen) RÉSUMÉ L’infection urinaire constitue un problème courant chez la femme enceinte. La présente étude a examiné la fréquence de l’infection urinaire dans un échantillon de 137 femmes enceintes consultant à la maternité d’Al Mukalla de janvier à juin 2002. Des prélèvements d’urine ont fait l’objet d’un examen microbi- ologique (examen au microscope et mise en culture) à la recherche d’une infection urinaire, et des tests de sensibilité ont été réalisés avec divers antibiotiques pour les germes isolés. Des informations sur l’âge, le trimestre, la parité et le nombre de grossesses ont été également recueillies pour chaque femme. Cette étude a montré que 30 % des femmes de l’échantillon souffraient d’une infection urinaire ; Escherichia coli était le germe le plus fréquemment isolé (41,5 %), et il était très sensible au chloramphénicol, à la ciproflox- acine, à la ceftizoxime et à l’amikacine. Parmi les variables examinées, 53,7 % des femmes ayant une infection urinaire appartenaient au groupe d’âge des 15-24 ans, 48,8 % étaient au troisième trimestre de leur grossesse et 75,6 % avaient 1-3 enfants. 31 Urinary tract infection.pmd 12/20/2005, 2:40 PM505 506 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Introduction Urinary tract infection (UTI) is an extreme- ly common clinical problem. It is important because it may involve the urethra, bladder, uterus, and kidney [1]. UTI affects all age groups, but women are more susceptible than men, due to short urethra, absence of prostatic secretion, pregnancy and easy contamination of the urinary tract with fae- cal flora [2]. Additionally, the physiological increase in plasma volume during pregnan- cy decreases urine concentration and up to 70% pregnant women develop glucosurea, which encourages bacterial growth in the urine [3,4]. Infections, particularly in pregnancy and in the elderly, can be asymptomatic [1], but asymptomatic bacteriurea is associated with an increased risk of intrauterine growth retardation and low birth weight [5]. Furthermore, untreated asymptomatic bacteriurea leads to the development of cystitis in approximately 30% of cases, and can lead to the development of pyelonephri- tis in about 50% of cases [6]. In addition acute pyelonephritis has been associated with anaemia [7]. Thus it is important to identify and treat UTI to avoid such com- plications. Sheikh et al. demonstrated that a history of past urological problems was associated with an increased incidence of UTI in preg- nancy [8]. Another study showed that oral hormone replacement therapy did not re- duce the frequency of UTI and the risk fac- tors in postmenopausal women are different from those in younger women and include diabetes, vaginal symptoms, and urge incontinence [9]. The organisms that cause UTIs during pregnancy are the same as those found in non-pregnant patients; Escherichia coli ac- count for 80%–90% of infection [3,10,11] and the rest are caused by Proteus mirabi- lis, Klebsiella aerogenes, Pseudomonas spp. and Streptococcus spp. [12]. The aim of this study was to determine if UTI is a widespread problem among pregnant woman in Al-Mukalla district and to assess its relation with some possible risk factors such as age, duration of preg- nancy and parity. The sensitivity of the or- ganisms isolated to various antibiotics was also studied. Methods This was a cross-sectional study of all pregnant woman attending Al Mukalla ma- ternity hospital for the period from January to June 2002. This hospital serves the whole of Al-Mukalla district which has a population of about 500 000. Midstream urine was collected from 137 pregnant women in sterile bottles; 10 mL were transferred to sterile centrifuge tubes and then centrifuged at 3000 rpm for 10–15 minutes. The supernatant was dis- carded and 1 mL of the precipitate was re- suspended in residual urine by shaking vigorously. Wet mount preparation for general urine examination was performed. Plates of blood agar and MacConkey medium were aseptically incubated with 2–3 drops of the suspended precipitate and then incubated at 37 °C for 24–48 hours or until visible growth appeared. The isolated pathogens were identified using the Cowan and Steel method [13]. Antibiotic sensitivity testing was performed using the Kirby–Bauer disc diffusion method [14]. The media used were Muller–Hinton agar (Oxoid) or nutrient agar. The antibiotic contents of the multidiscs were ampicillin/ sulbactam (20 µg), co-trimoxazole (25µg), cephalexin (30 µg), tetracycline (30 µg), cefotaxime (30 µg), ciprofloxacin (5 µg), 31 Urinary tract infection.pmd 12/20/2005, 2:40 PM506 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 507 Table 1 Frequency distribution of microorganism and its relation with urine analysis Microorganism Culture Microscopic urine analysis (No.) No. % Pus Epithelial RBC Crystals Bacteria cells cells seen Escherichia coli 17 41.5 14 15 8 15 13 Klebsiella spp. 4 9.8 4 4 4 3 4 Neisseria gonorrhoae 2 4.9 2 2 1 1 1 Proteus spp. 5 12.2 3 3 2 3 2 Staphylococcus aureus 8 19.5 8 6 6 5 6 Streptococcus spp. 2 4.9 2 2 1 1 1 Candida albicans 3 7.3 2 3 1 1 1 Total 41 100 35 35 23 29 28 RBC = red blood cells. pefloxacin (10 µg), ofloxacin (5 µg), clox- acillin (1 µg), roxithromycin (15 µg), linco- mycin (2 µg), and gentamicin (10 µg) for Gram-positive bacteria. The antibiotic con- tents of multidiscs for Gram-negative bacteria were piperacillin (100 µg), chloramphenicol (30 µg), ceftizoxime (30 µg) and amikacin (30 µg) in addition to ampicillin/sulbactam, co-trimoxazole, ce- fotaxime, ciprofloxacin, tetracycline, pe- floxacin, ofloxacin and gentamicin at the same concentrations as for the Gram-posi- tive bacteria. Information on age, period of gestation, parity, and number of pregnancies was col- lected by face-to-face interview with the pregnant women. The socioeconomic sta- tus was not included in this study. Results Of the 137 women tested, 41 (30%) were positive for UTI, while 96 (70%) were neg- ative. The distribution of the different iso- lates is shown in Table 1; E. coli (41.5%) was the most frequently isolated strain, fol- lowed by Staphylococcus aureus (19.5%). Most of the Gram-positive bacteria were highly susceptible to ceftizoxime and the Gram-negative to gentamicin and cefo- taxime (Table 2). E. coli, was susceptible to ampicillin/sulbactam (71% susceptibility), cefotaxime (71%), ciprofloxacin (94%), chloramphenicol (94%), ceftizoxime (100%), ofloxacin (82%), and amikacin (100%). S. aureus was susceptible to ampi- cillin/sulbactam (75%), cephalexin (75%), cefotaxime (100%), roxithromycin (75%), lincomycin (75%), and gentamicin (100%). The distribution of UTI among the in- fected women according to age group, tri- mester, parity and number of pregnancies is shown in Table 3. Bacteriuria was com- monest in the age group 15–24 years (53.7%). There was an increase in frequen- cy of bacteriurea with progress of pregnan- cy, with 48.8% of infections in the women in the 3rd trimester. Of the infected women 75.6% had 1–3 children. 31 Urinary tract infection.pmd 12/20/2005, 2:40 PM507 508 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Discussion UTIs are the commonest infections seen in hospital settings, and the second common- est infections seen in the general population [15]. UTIs are especially problematic for women and up to a third of all women will experience a UTI at some point in their life [15]. Appropriate treatment requires accu- rate classification of infection site, com- plexity of the infection and the likelihood of recurrence. In this study the prevalence of UTI in our sample of pregnant women was 30% and the predominant pathogens were pathogenic E. coli and S. aureus. Our re- sults agree with those of research workers in other countries, with minor differences, which could be due to differences in the environment, social habits of the communi- ty, the standard of personal hygiene and differences in education [15–17]. In many studies of UTIs account for about 10% of office visits by women, and one-third of women will have a UTI at some time during their life [15]. In preg- nant women, the incidence of UTI can be as high as 8% [18,19], while Baleiras et al. reported that UTI is the commonest type of infection in pregnant women, with preva- lence ranging from 5% to 10% [20]. In similar studies in our Region, the preva- lence was 38% in Iraq [1], 28.5% in Paki- stan [8], 14.2% in Saudi Arabia [21] and 10.6% in Turkey [22]. Sheikh et al. demonstrated that socio- economic status, personal hygiene, educa- tion level, pregnancy duration, post-coital washing, contraceptive use, and use of un- derclothing had no significant association Table 2 Susceptibility of isolated Gram-negative and Gram-positive bacteria Bacteria Total no. Susceptibility to antimicrobial drugs (%) isolated Gram-negative AS BA CF PC CH CP CI TE OF GM AK PF bacteria Escherichia coli 17 71 6 71 41 94 94 100 24 82 71 100 0 Klebsiella spp. 4 75 0 50 50 100 100 100 0 50 75 75 0 Neisseria gonorrhoae 2 50 0 100 100 100 0 50 0 0 50 100 0 Proteus spp. 5 0 20 80 0 0 80 100 80 0 100 0 20 Total 28 11 2 14 8 15 17 20 6 10 16 15 1 Gram-positive bacteria AS BA PR TE CF CP PF OF CX RF LM GM Stapylococcus aureus 8 75 0 75 0 100 13 0 0 50 75 75 100 Streptococcus spp. 2 100 0 50 50 100 0 0 0 50 50 50 100 Total 10 8 0 7 7 10 1 0 0 5 7 7 10 AS = ampicillin/sulbactam (20 µg); BA = co-trimoxazole (25 µg); CF = cefotaxime (30 µg); PC = piperacillin (100 µg); CH = chloramphenicol (30 µg); CP = ciprofloxacin (5 µg); CI = ceftizoxime (30 µg); TE = tetracycline (30 µg); OF = ofloxacin (5 µg); GM = gentamicin (10 µg); AK = amikacin (30 µg); PF = pefloxacin (10 µg); PR = cephalexin (30 µg); CX = cloxacillin (1 µg), RF = roxithromycin (15 µg); LM = lincomycin (2 µg). 31 Urinary tract infection.pmd 12/20/2005, 2:40 PM508 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 509 Table 3 Distribution of urinary tract infection by age group, trimester, parity and number of pregnancies Variable Infected No. % Age group (years) 15–24 22 53.7 25–34 16 39.0 35–45 3 7.3 Trimester 1st 7 17.1 2nd 14 34.1 3rd 20 48.8 Parity Nullipara 15 36.6 1 13 31.7 Multipara 13 31.7 Number of pregnancies 1–3 31 75.6 4–6 8 19.5 7–9 2 4.9 Total 41 100 with UTI occurrence [8]. Krcmery et al. demonstrated that the risk factors for UTI in women include sexual intercourse, hav- ing a first UTI at an early age, and having a maternal history of UTIs [23]. Our results regarding the age of infection, gestation and parity concur with this study [23]. Women with recurrent UTI have an in- creased susceptibility to vaginal coloniza- tion and uropathogens, which is due to a greater propensity for uropathogenics coliforms to adhere to uroepithelial cells [23]. Ceftizoxime and amikacin were the most effective antibiotics against E. coli. However, Skerk et al. reported that urine culture should be performed prior to the initiation of antimicrobial therapy of UTIs [24]. It is important that antibiotics used for treating UTIs in pregnant women are carefully chosen, either in the case of monotherapy or combination therapy using more than one antibiotic to give a synergis- tic effect so as to decrease the dose need- ed, and accordingly the side-effects and toxicity of using a single high-dose drug. Also care must be taken when using potent antibiotics which may harm both the fetus as well as the mother, such as chloram- phenicol or tetracycline which are totally contraindicated in pregnancy due to their harmful effects on the fetus [25]. Further studies should be performed on a larger sample of pregnant women, where the type of infection, past history of uro- logical problems, recurrence of UTI and the relation between possible risk factors, such as socioeconomic status, personal hygiene, education level, frequency of sex- ual intercourse, could be examined. References 1. Al-Dujiaily AA et al. Urinary tract infection during pregnancy in Tikrit. Medical Jour- nal of Tikrit, 2000, 6:220–4. 2. Awaness AM, Al-Saadi MG, Aadoas SA. Antibiotics resistance in recurrent uri- nary tract infection. Kufa medical journal, 2000, 3:159. 3. Patterson TF, Andrriole VT. Bacteriuria in pregnancy. Infectious disease clinics of North America, 1987, 1:807–22. 4. Lucas MJ, Cunningharm FG. Urinary tract infection in pregnancy. Clinical ob- stetrics and gynecology, 1993, 36:855– 68. 5. Harris RE, Thomas VL, Shelokor A. Asymptomatic bacteriuria in pregnancy: antibody coated bacteria, renal function and intrauterine growth retardation. American journal of obstetrics and gyne- cology, 1976, 126:20–5. 31 Urinary tract infection.pmd 12/20/2005, 2:40 PM509 510 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 6. Kass EH. Pregnancy, pyelonephritis and prematurity. Clinical obstetrics and gyne- cology, 1970, 13:239–54. 7. Gilstreap LC III et al. Renal infection and pregnancy outcome. American journal of obstetrics and gynecology, 1981, 141: 709–16. 8. Sheikh MA et al. Incidence of urinary tract infection during pregnancy. Eastern Mediterranean health journal, 2000, 6(2–3):265–71. 9. Brown JS et al. Urinary tract infections in postmenopausal women: effect of hor- mone therapy and risk factors. Obstetrics and gynecology, 2001, 98(6):1045–52. 10. Barr JG et al. Microaerophilic anaerobic bacteria as a cause of UTI in pregnancy. British journal of obstetrics and gynecol- ogy, 1985, 92:506–10. 11. McDowall DR et al. Anaerobic and other fastidious microorganisms in asymptom- atic bacteriuria in pregnant woman. Jour- nal of infectious diseases, 1981, 144: 114–22. 12. Chamberlin GVP. Gynecology by ten teachers. London, Edward Arnold, 1995. 13. Cowan SF, Steel KJ. Manual for identifi- cation of medical bacteria, 3rd ed. Cam- bridge, Cambridge University Press, 1993. 14. Bauer AW et al. Antibiotic susceptibility testing by standardized single disk method. American journal of clinical pa- thology, 1966, 45:493–6. 15. Valiquette L. Urinary tract infections in women. Canadian journal of urology, 2001, 8(1):6–12. 16. Coxe E. Comparison of intravenous fleroxacin with ceftazidimine for treat- ment of complicated UTI. American jour- nal of medicine, 1991, 94(3):118–25. 17. Pummer K. Fleroxacin versus norfloxa- cin in the treatment of urinary tract infec- tions: a multicenter, double-blind, pro- spective, randomized, comparative study. American journal of medicine, 1993, 94(3A):108S–113S. 18. Patterson TF, Andriole VT. Bacteriuria in pregnancy. Infectious disease clinics of North America, 1987, 1:807–22. 19. Mikhail MS, Anyaegbunam A. Lower uri- nary tract dysfunction in pregnancy: a re- view. Obstetrical & gynecological survey, 1995, 50:675–83. 20. Baleiras C et al. Infecções urinarias e gravidez. [Urinary tract infections and pregnancy.] Acta médica portuguesa, 1998, 11(10):839–6. 21. Al-Sibai MH, Saha A, Rasheed P. Socio- biological correlates of bacteriuria in Saudi pregnant women. Public health, 1989, 103(2):113–21. 22. Kutlay S et al. Prevalence, detection and treatment of asymptomatic bacteriuria in a Turkish obstetric population. Journal of reproductive medicine, 2003, 48(8): 627–30. 23. Krcmery S, Hromec J, Demesova D. Treatment of lower urinary tract infection in pregnancy. International journal of an- timicrobial agents, 2001, 17(4):279–82. 24. Hooton TM. Recurrent urinary tract infec- tion in women. International journal of antimicrobial agents, 2001, 17(4):259– 68. 25. Skerk V et al. Antimikrobno lijecenje infekcija mokracnog sustava. [Antimicro- bial therapy of urinary tract infection.] Lijecnicki vjesnik, 2001, 123(1–2):16– 25. 26. Laurence DR, Bennett PN, Brown MJ. Clinical pharmacology, 8th ed. Edinburgh, Churchill Livingstone, 1997: 207–10. 31 Urinary tract infection.pmd 12/20/2005, 2:40 PM510 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 511 Short communication Prevalence of Pityrosporum orbiculare on normal skin of Iraqi children K.A. Sharquie,1 W.A. Al-Kubaisy2 and M.C. Al-Rubaey3 1Department of Dermatology, College of Medicine, University of Baghdad, Baghdad, Iraq. 2Department of Community Medicine, Saddam Medical College, Saddam University, Baghdad, Iraq (Correspondence to W. A. Al-Kubaisy: waqar_abd@yahoo.com). 3Department of Community Medicine, College of Medicine, Al-Mustansiriyah University, Baghdad, Iraq. Received: 13/09/01; accepted: 07/04/04 ABSTRACT To determine the age incidence and prevalence of Pityrosporum orbiculare on the normal skin of healthy Iraqi children, we carried out a survey of clinically normal skin of 110 healthy children in Baghdad during April 1998–October 1998. We isolated P. orbiculare on a peptone–glucose–yeast extract medium containing chloramphenicol and cycloheximide, and overlaid with olive oil. The organism was present on the trunk in 77.5% of children 10–14 years, 27.5% of children 5–9 years, and 6.6% of children < 5 years. Prévalence de Pityrosporum orbiculare sur la peau normale d’enfants iraquiens RÉSUMÉ Afin de déterminer l’incidence selon l’âge et la prévalence de Pityrosporum orbiculare sur la peau normale d’enfants iraquiens en bonne santé, nous avons réalisé une étude de la peau cliniquement normale de 110 enfants en bonne santé à Bagdad d’avril à octobre 1998 . Nous avons isolé P. orbiculare dans un milieu peptone-glucose-extrait de levure contenant du chloramphénicol et de la cycloheximide, et enrichi en lipides (huile d’olive). Le micro-organisme était présent sur le tronc chez 77,5 % des enfants âgés de 10- 14 ans, chez 27,5 % des enfants âgés de 5-9 ans et chez 6,6 % des enfants de moins de 5 ans. 32 Prevalence of Pityrosporum.pmd 12/20/2005, 2:40 PM511 512 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 Introduction Pityrosporum orbiculare was the name giv- en by Gordon to the spherical lipophilic yeast which he cultured from the lesions of pityriasis versicolor [1]. Malassezia furfur has been recorded as the causative organ- ism of pityriasis versicolor. The yeast phase of this organism is classified as P. orbiculare [2,3]. P. orbiculare can be isolated not only from pityriasis versicolor scales, but also from normal skin. Roberts found P. orbic- ulare to be present on the trunk in more than 92% of the participants in his study [4]. Faergemann and Bernarder cultured P. orbiculare from the back in 90% of healthy individuals in a study done in Sweden [5]. Also, Sharquie, Al-Rubyae and Al-Rawi cultured P. orbiculare from the back in 92% of healthy Iraqis in a Baghdad study [6]. Several studies have been conducted to detect the age incidence of this yeast. In a survey of children, from newborn to age 15 years, P. orbiculare was not cultured from normal-looking skin on the back be- fore the age of 5 years, but was found in 10% of 5-year-old children, 23% of 10- year-old children and 93% of 15-year-old children [7]. In Caracas, P. orbiculare was present in 45% of children 8–10 years old and in 70% of children 13–15 years old [8]. It was subsequently suggested that coloni- zation of normal skin with P. orbiculare begins with the increase in sebum excretion in pre-puberty and puberty [7,8]. Therefore, we conducted this study to identify the age incidence of this yeast on the skin of healthy Iraqi children. Methods The study was conducted at the Depart- ment of Dermatology, Baghdad Teaching Hospital from April 1998 to October 1998. We studied 110 normal children. They were all living in Baghdad. Ages ranged from 1 to 14 years. All were healthy chil- dren with normal looking skin. And all the skin at the sample sites was clinically nor- mal and had not been affected in the past as far as their parents could recall. Any child on systemic therapy with steroids or antibi- otics was excluded as well as those treated with topical applications. We used peptone–glucose–yeast extract medium containing peptone, 10 g/L, Bacto- agar, 18 g/L, glucose, 40 g/L and yeast ex- tract, 0.1 g/L. The pH was adjusted to 5–6. After autoclave sterilization, chlorampheni- col 50 mg/L, gentamycin 100 mg/L and cycloheximide 0.5 g/L were added to the medium. From each child, skin scrapings were taken from the back in the interscapular re- gion. Scrapings were taken using a fresh, sterile scalpel blade for each site sampled. The specimens were transferred directly to the culture plate and distributed evenly over the entire surface of the plate. Cultures were incubated at 37 °C. Seven days later, the cultures were examined macroscopi- cally and microscopically; if no growth had occurred at that time, they were retained and checked 14 days later. P. orbiculare was recorded as positive only if thick-walled, spherical budding forms were seen. The presence of clusters was also noted. Results The prevalence of P. orbiculare increased with age: 6.7% of samples from children < 5 years, 27.5% of those from children 5– 9 years and 77.5% of those from children 10–14 years were positive (Table 1). On the medium, P. orbiculare grew in a confluent manner with a slight tendency to solitary colonies, white to cream in colour 32 Prevalence of Pityrosporum.pmd 12/20/2005, 2:40 PM512 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 513 Table 1 Incidence of Pityrosporum orbiculare on the skin of healthy children in 3 age groups Pityrosporum Age (years) Total orbiculare < 5 5–9 10–14 No. % No. % No. % Positive 2 6.7 11 27.5 31 77.5 44 Negative 28 93.3 29 72.5 9 22.5 66 Total 30 27.3 40 36.4 40 36.4 110 (Figure 1). Microscopically, clusters of round, budding cells were seen (Figure 2). Discussion Interestingly, this is the first time P. orbicu- lare has been isolated from the skin of healthy children under 5 years old—2 of the positive cultures were from 4-year-old children. In contrast, Faergemann and Fre- driksson did not detect the organism in children < 5 years old [7]. In agreement with the results of other studies we found that the prevalence of P. orbiculare increased with increasing age [7,8]. Our results indicate that P. orbiculare (the causative agent of pityriasis versicol- or) is part of the normal skin flora in Iraqi children and colonization starts during the period when the sebaceous glands become active. This explains why it is difficult to eradicate this infection and at the same time indicates that excessive washing and steril- ization of the clothes of people suffering from the condition is not necessary. Our study was the first in which P. or- biculare was detected on the skin of Figure 1 Colonies of Pityrosporum orbiculare on Sabouraud’s dextrose agar after three days incubation at 37 °C Figure 2 Microscopic view of Pityrosporum orbiculare 32 Prevalence of Pityrosporum.pmd 12/20/2005, 2:40 PM513 514 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 healthy children in our country. We con- clude that P. orbiculare can be detected on the normal skin of healthy children, and is even found in some children < 5 years old. This could possibly be attributed to early activation of sebaceous glands in addition to other factors, but further studies would be needed to determine this. References 1. Gordon MA. The lipophilic microflora of the skin. In vitro culture of Pityrosporum orbiculare. Mycologia, 1951, 43:524–35. 2. Odom RB, James WD, Berger TG. Dis- eases resulting from fungi and yeasts. In: Odom RB, James WD, Berger TG, eds. Andrews’ diseases of the skin, 9th ed. Philadelphia, WB Saunders Company, 2000:358–416. 3. Faergemann J. Pityrosporum species as a cause of allergy and infection. Allergy, 1999, 54(5):413–9. 4. Roberts SO. Pityrosporum orbiculare: in- cidence and distribution on clinically normal skin. British journal of dermatol- ogy, 1969, 8(4)1:264–9. 5. Faergemann J, Bernander S. Tinea ver- sicolor and Pityrosporum orbiculare: a mycological investigation. Sabouraudia, 1979, 17(3):171–9. 6. Sharquie KA, Al-Rubyae MG, Al-Rawi JR. Prevalence of Pityrosporum orbiculare on normal skin of Iraqi healthy people. Iraqi journal of community medi- cine, 2001, 14(1):11–3. 7. Faergemann J, Fredriksson T. Age inci- dence of Pityrosporum orbiculare on hu- man skin. Acta dermato-venereologica, 1980, 60(6):531–3. 8. Borelli D. Epidemiology, ecology and treatment of pityriasis versicolor in Latin America (Venezuela). In: Oral therapy in dermatomycoses: a step forward: pro- ceedings of a symposium, Frankfur t, February 1985. Oxford, Medicine Pub- lishing Foundation, 1985:111–7. 32 Prevalence of Pityrosporum.pmd 12/20/2005, 2:40 PM514 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 515 Current topics 1Cairo, WHO Regional Office for the Eastern Mediterranean, 2004 (WHO Regional Publications, Eastern Mediterranean Series; 30):44–9. The full text of this publication is available free on line at: http:// www.emro.who.int/publications/pdf/HealthResearchers_Guide.pdf Types of research design The study type may dictate certain re- search designs. More commonly, the study objectives can be achieved through a num- ber of alternative designs. The investigators have to select the most appropriate and most feasible design. Generally, there are two main categories of research design: observational study, and experimental or in- tervention study. In the observational study, the investigators stand apart from events taking place in the study. They sim- ply observe and record. In the experimental or intervention study, the investigators in- troduce an intervention and observe the events which take place in the study. Observational studies An observational study may be descriptive or analytical. A descriptive study is an ob- servational study that simply describes the distribution of a characteristic. An analyti- cal study is an observational study that de- scribes associations and analyses them for possible cause and effect. An observational study may be cross- sectional or longitudinal. In a cross-sec- tional study, measurements are made on a single occasion. In a longitudinal study, measurements are made over a period of time. A longitudinal observational study may be retrospective or prospective. In a retro- spective study, the investigators study present and past events. In a longitudinal prospective study, the investigators follow subjects for future events. Case–control studies are a type of ob- servational-analytical-retrospective studies over time in which a group of subjects with a specified outcome (cases) and a group without that outcome (controls) are identi- fied. Investigators then compare the extent to which each subject was previously ex- posed to the variable of interest, such as a risk factor, a treatment or an intervention. Case–control studies are useful for study- ing rare conditions and conditions with long intervals between exposure and out- come such as, for example, risk of devel- oping neoplasia. In such situations, a prospective study will be difficult. Case– control studies can be efficient and eco- nomical, but do not have the strength of evidence of a prospective study. In clinical and epidemiological research, a longitudinal observational study is usually called a cohort study. The word cohort was the ancient Roman term for a group of soldiers who marched together into battle. The prospective cohort design is generally considered to be the “crème de la crème” For the interest and benefit of our readers, we are pleased to reproduce a second extract on research topics from the EMRO publication, A practical guide for health researchers1 by Mahmoud F. Fathalla and Mohamed M.F. Fathalla. The first extract on the criteria for a good research topic was published in Volume 11 Nos 1/2 (2005). 33 Current topics.pmd 12/20/2005, 2:40 PM515 516 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 of observational methodologies for the fol- lowing reasons. Data are gathered prospectively. • Recall bias is not a problem (research subjects are not asked to recall past events). • Time–order relationships are clear (it is easy to decide that an outcome fol- lowed, rather than preceded, a possible cause). • Investigators have much more control on the quality of the data. There are, however, some drawbacks. • The biggest single problem of these fol- low-up design investigations is the loss of valuable information through attri- tion, due to loss to follow-up, or sub- jects opting out of the study. • Subjects may change their behaviour over time. • A bias can occur if there is unequal sur- veillance of subjects in the two com- pared groups, during follow-up. One of the best examples of a prospec- tive cohort study was initiated by Austin Bradford Hill and Richard Doll, to investi- gate the relationship between smoking and lung cancer. They followed up 40 000 Brit- ish doctors who were divided into four co- horts: non-smokers, and light, moderate and heavy smokers. Death was the out- come they recorded. They used both all cause death (any death) and cause specific death (death from a particular disease). Publication of their interim 10 year results in 1964, showed a substantial excess in both mortality from lung cancer and all cause mortality in smokers, with a “dose- response” relation (that is, the more the subjects smoked the greater were their chances of getting lung cancer). The study went a long way in demonstrating that the link between smoking and ill-health was causal rather than coincidental. The 20 year and 40 year results of this momentous study (which achieved 94% follow-up of those recruited in 1951 and not known to have died) illustrate the strength of evi- dence that can be obtained from a properly conducted cohort study (Doll and Hill, 1964; Doll and Peto, 1976; Doll et al., 1994). Experimental or intervention studies In the experimental or intervention study, the investigators test the effect of an inter- vention on the events taking place in the study. An experimental or intervention study may be controlled or non-controlled. Giving a treatment to a patient or group of patients and finding that the treatment works gives only preliminary and non-de- finitive information. We do not know what would have happened if no treatment or a different treatment was given. For a more definitive answer, we need a “control” group of patients who do not get the treat- ment under study. Hawthorne effect: In the late 1920s, a group of researchers at the Western Elec- tric Hawthorne Works in Chicago were in- vestigating the effects of lighting, heating and other physical conditions upon the pro- ductivity of workers. Much to the surprise of the researchers, the productivity of the workers kept improving even when the ac- tual physical conditions were not im- proved. The Hawthorne effect can be manifested in clinical research settings. Even “inert” treatments might result in sig- nificant improvements in the patient’s con- dition (Polgar and Thomas, 2000). A controlled experimental study may be randomized or non-randomized. In testing the outcome in a group of patients who re- ceive the treatment and another group who do not, we are still not sure whether any 33 Current topics.pmd 12/20/2005, 2:40 PM516 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 517 difference observed is because of the treat- ment or because the characteristics of the patients in the two groups were different. The best way to be sure is to randomize the allocation of patients to either treatment or to no treatment. Randomized controlled trials are inter- vention studies characterized by the pro- spective assignment of subjects, through a random method, into an experimental group and a control group. In a clinical tri- al, the experimental group receives the drug or treatment to be evaluated, while the con- trol group receives a placebo, no treatment, or the standard of care. Both groups are followed for the outcome(s) of interest. Randomization is the most reliable method to ensure that the participants in both groups are similar as far as possible with respect to all known or unknown factors that might affect the outcome. With ran- domization, only chance determines the as- signment of subjects to study groups. Random allocation does not mean haphaz- ard allocation. It is a carefully planned method of assigning subjects to similar groups. If important risk factors can be identified at the outset, subjects may be grouped or stratified prior to assignment. Whenever it is ethical and practical, a ran- domized design should be considered in controlled intervention studies. Controlled trials without randomization are intervention studies in which allocation to either experimental or control group is not based on randomization, making as- signment subject to possible biases that may influence study results. A crossover study is a special design of controlled intervention study that is some- times used in drug trials. In this design, half of the participants are randomly assigned to start with the placebo and then switch to active treatment, while the other half does the opposite. It has the advantage of reduc- ing the number of subjects required, since each subject serves as both an experimental subject and a control. It also decreases the biological variability inherent in comparing different subjects by comparing each sub- ject with himself or herself. It has the dis- advantage of increasing the duration of the study. There will also be a problem if the treatment has a carry-over effect after it is stopped. A before-and-after study is a method of control in which results from experimental subjects are compared with outcomes from patients treated before the new inter- vention was available. These are called his- toric controls. A randomized controlled trial may be blinded if participants in the trial are likely to change their behaviour in a systematic way that may influence the outcome of the study when they are aware of which inter- vention they receive. (Ophthalmologists prefer the term “masking” to the term “blinding”.) Blinding can take place at a number of levels. At one level, those responsible for assigning the subjects to groups do not know to which group the next subject will be assigned. In another level, research sub- jects are also not aware of which interven- tion they are receiving. Then, health workers who take care of patients in the study may not be allowed to know what treatment the different patients are receiv- ing. Lastly, researchers who assess the outcome are also not able to distinguish the subjects in the different groups. The term double-blind is used when nei- ther researchers not subjects are aware of the type of intervention. A trial in which there is no attempt at blinding may be called open or open label. The Rosenthal effect: Rosenthal and his colleagues in 1976 performed an experi- ment involving the training of two groups 33 Current topics.pmd 12/20/2005, 2:40 PM517 518 La Revue de Santé de la Méditerranée orientale, Vol. 11, No 3, 2005 of rats in a maze learning task. A bright strain and a dull strain of rats especially bred for the purpose were trained by un- dergraduate student experimenters to nego- tiate the maze. After a suitable training interval, the relative performances of the groups were compared. Not surprisingly, the bright strain significantly outperformed the dull strain. What was surprising, how- ever, was that the two strains were actually not different. The two groups of rats were actually genetically identical. The research- ers had deceived the student experimenters for the purposes of the study, and the stu- dents’ expectations of the rats had resulted in different methods of treatment, which had affected the rats’ learning ability. These results have been confirmed time and time again in a variety of experimental settings, and with a variety of subjects. They con- firm the need for blinding (Polgar and Tho- mas, 2000). Selecting a research design A research question may be answered by more than one research design. The re- searcher has to select the appropriate de- sign for the particular study. All types of research design have a place, and all have advantages and disadvantages. But not all types of design are always possible for a particular study. For example, the investigators may want to study if there is a relationship be- tween post-menopausal hormone replace- ment therapy and subsequent development of uterine endometrial carcinoma. The in- vestigators can design an observational study or an experimental study. If the deci- sion was for an observational study, the in- vestigators may do a descriptive study or an analytical study. For a descriptive study, they will review the clinical records of all patients diagnosed as having endometrial carcinoma. They will look for a history of post-menopausal hor- monal therapy. This study will be useful but cannot be definitive. It shows whether fur- ther study is needed to confirm or refute the impression gained from the descriptive study. The information about the strength of the association will also help in the de- sign of further analytical studies. The find- ing that many of the women who developed endometrial carcinoma had a history of homonal therapy cannot lead to any conclusion. It may simply mean that this therapy is widely used in the communi- ty, both by women who develop and who do not develop endometrial carcinoma. This shows the need for further studies. For an analytical study, the investigators may do a cross-sectional study or a longi- tudinal study. In a cross-sectional study, the investigators may study all postmeno- pausal women admitted to hospital over a defined time period. For each woman, they record whether she received or did not re- ceive hormonal therapy, and whether she had or did not have endometrial cancer. The advantage of this study is that it can be done rapidly. It gives more evidence than the simple descriptive study. However, the two groups of patients may not be compa- rable. In a longitudinal observational study, the investigators may do a prospective study or a retrospective study. For a prospective study, a cohort of two groups of post- menopausal women is followed up: one group already receiving hormone replace- ment therapy and another matched group not receiving this therapy. For a retrospec- tive study, a case–control design can be se- lected. A group of women who have recently developed endometrial cancer (cases) and a group of women with similar characteristics and did not develop en- dometrial cancer (controls) are identified. 33 Current topics.pmd 12/20/2005, 2:40 PM518 Eastern Mediterranean Health Journal, Vol. 11, No. 3, 2005 519 The use of hormone replacement therapy in each woman in the case group and in the control group is determined to assess expo- sure history. The advantage is that the study can be done relatively quickly. The disadvantage is that the two groups may still not be completely similar. Other vari- ables may influence the outcome and may be difficult to exclude. If the investigators decide on an experi- mental or intervention study, they may se- lect a randomized or a non-randomized design. In a randomized controlled study, post-menopausal women identified from a population are randomly assigned either to a study group that will receive hormone re- placement therapy or to a control group that will be prescribed a placebo. Both groups will then be followed prospectively to determine how many in each group will develop endometrial cancer. This study, if successfully conducted, will provide a more definitive answer to the research question. However, it will raise ethical con- cerns. Additional difficulties are the large sample size needed because of the relatively low incidence of the disease, the long fol- low-up because of the long latent period before the development of the disease and the possibility of poor compliance or loss to follow-up. Alternatively, a non-randomized controlled design may be considered. This may be easier, will allow women to make an informed choice but there will be a need to consider other possible variables that may influence the outcome, since the two groups may not be similar. Different types of research design are not considered equal in the strength of evi- dence they provide. In the traditional hier- archy of evidence, randomized controlled studies are generally ranked high, followed by cohort and case–control studies, while observational descriptive studies are ranked at a lower level. The investigators may, however, not be able to select the design that gives a high level of evidence, because it will not be feasible to do, or will not be ethical to do. In this case, their selection of another design will be acceptable and justi- fied. References and additional sources of information 1. Doll R, Hill AB. Mortality in relation to smoking: ten years’ observation of British doctors. British Medical Journal, 1964, 1:1399–1414; 1460–1467. 2. Doll R, Peto R. Mortality in relation to smoking: 20 years’ observation on Brit- ish doctors. British Medical Journal, 1976, 2:1525–1536. 3. Doll R, Peto R, Wheatley K et al. Mortality in relation to smoking: 40 years’ observa- tions on male British doctors. British Medical Journal, 1994, 309:901–911. 4. Polgar S, Thomas SA. Introduction to re- search in the health sciences, 4th edi- tion. London, New York, Churchill Livingstone, 2000:62; 63; 107–114. 33 Current topics.pmd 12/20/2005, 2:40 PM519