Eastern Mediterranean Health Journal [2010; Vol.16, Issue 9]

طسوتلما قشرل ةيلماعلا ةحصلا ةمظنلم ةيميلقلإا ةنجللا ءاضعأ نادلبلا ةيملاسلإا ناريإ ةيروهجم . ةيبيللا ةيبرعلا ةييرهمالجا . سنوت . نيرحبلا . ناتسكاب . ةدحتلما ةيبرعلا تاراملإا . ناتسناغفأ . ندرلأا صرم . نانبل . تيوكلا . رطق . ينطسلف . نماُع . قارعلا . لاموصلا . نادوسلا . تيوبيج . ةينميلا ةيروهملجا . ةيروسلا ةيبرعلا ةيروهملجا ةيدوعسلا ةيبرعلا ةكلملما . برغلما Subscriptions and Distribution Enquiries regarding subscriptions and distribution of the print edition of EMHJ should be addressed to: Printing and Marketing of Publications at: email: pam@emro.who.int; tel: (+202) 2276 5000; fax: (+202) 2670 2492 or 2670 2494 Permissions Requests for permission to reproduce or translate articles, whether for sale or non-commercial distribution should be addressed to EMHJ at: emhj@emro.who.int Members of the WHO Regional Committee for the Eastern Mediterranean Afghanistan . Bahrain . Djibouti . Egypt . Islamic Republic of Iran . Iraq . Jordan . Kuwait . Lebanon Libyan Arab Jamahiriya . Morocco . Oman . Pakistan . Palestine . Qatar . Saudi Arabia . Somalia Sudan . Syrian Arab Republic . Tunisia . United Arab Emirates . Republic of Yemen Membres du Comité régional de l’OMS pour la Méditerranée orientale Afghanistan . Arabie saoudite . Bahreïn . Djibouti . Égypte . Émirats arabes unis . République islamique d’Iran Iraq . Jamahiriya arabe libyenne . Jordanie . Koweït . Liban . Maroc . Oman . Pakistan . Palestine . Qatar République arabe syrienne . Somalie . Soudan . Tunisie . République du Yémen Correspondence Editor-in-chief EMHJ WHO Regional Office for the Eastern Mediterranean P.O. Box 7608 Nasr City, Cairo 11371 Egypt Tel: (+202) 2276 5000 Fax: (+202) 2670 2492/(+202) 2670 2494 Email: sabrib@emro.who.int/emhj@emro.who.int EASTERN 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, con- cepts, 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 Col- laborating Centres and individuals within and outside the Region. LA REVUE DE SANTÉ DE LA MÉDITERRANÉE ORIENTALE EST une revue de santé officielle publiée par le Bureau régional de l’Organisation mondiale de la Santé pour la Méditerranée orientale. Elle offre une tribune pour la présentation et la promotion de nouvelles politiques et initiatives dans le domaine des ser-vices de santé ainsi qu’à l’échange d’idées, de concepts, de données épidémiologiques, de résultats de recherches et d’autres informations, se rapportant plus particulièrement à la Région de la Méditerranée orientale. Elle s’adresse à tous les professionnels de la santé, aux membres des instituts médicaux et autres instituts de formation médico-sanitaire, aux ONG, Centres collabora- teurs de l’OMS et personnes concernés au sein et hors de la Région. EMHJ is a trilingual, peer reviewed, open access journal and the full contents are freely available at its website: http://www/emro.who.int/emhj.htm EMHJ is abstracted/indexed in the Index Medicus and MEDLINE (Medical Literature Analysis and Retrieval Systems on Line) and the ExtraMed-Full text on CD-ROM, the Cumulative Index to Nursing and Allied Health Literature (CINAHL), CAB International, Lexis Nexis, Scopus and the Index Medicus for the WHO Eastern Mediterranean Region (IMEMR). ©World Health Organization 2010 All rights reserved Disclaimer 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 World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. Dotted lines on maps represent approximate border lines for which there may not yet be full agreement. 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. All reasonable precautions have been taken by the World Health Organization to verify the information contained in this publication. However, the published material is being distributed without warranty of any kind, either express or implied. The responsibility for the interpretation and use of the material lies with the reader. In no event shall the World Health Organization be liable for damages arising from its use. The named authors alone are responsible for the views expressed in this publication. ISSN 1020-3397 Cover designed by Diana Tawadros Internal layout designed by Emad Marji and Diana Tawadros Printed by WHO Regional Office for the Eastern Mediterranean ميدقتل برنم ىهو .ةيلماعلا ةحصلا ةمظنمب طسوتلما قشرل ىميلقلإا بتكلما نع ردصت ىتلا ةيمسرلا ةلجلما ىه ةيئابولا تايطعلماو ميهافلماو ءارلآا لدابتلو ،اله جيوترلاو ةيحصلا تامدلخا فى ةديدلجا تاردابلماو تاسايسلا لك لىإ ةهجوم ىهو .طسوتلما قشر ميلقإب اهنم قلعتي ام ةصاخو ،تامولعلما نم كلذ يرغو ثاحبلأا جئاتنو زكارلماو ،ةينعلما ةيموكلحا يرغ تماظنلما اذكو ،ةيميلعتلا دهاعلما رئاسو ةيبطلا تايلكلاو ،ةيحصلا نهلما ءاضعأ .هجراخو ميلقلإا فى ةحصلاب ينمتهلما دارفلأاو ةيلماعلا ةحصلا ةمظنم عم ةنواعتلما طسوتلما قشرل ةيحصلا ةلجلما Cover 6.indd 2 5/24/2010 11:46:47 AM EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 920 Isolation and identification of Helicobacter pylori from drinking water in Basra governorate, Iraq A.A. Al-Sulami,1 A.M. R. Al-Taee 2 and M.G. Juma’a 3 ABSTRACT The mode of the transmission of Helicobacter pylori infection remains poorly understood. A total of 198 samples of drinking water from 22 districts of Basra governorate were collected during the period October 2006 to July 2007. The concentration of residual chlorine was measured and the numbers of total and faecal coliforms were counted. On modified Columbia urea agar, 469 bacterial cultures were obtained, of which 173 isolates were identified. Only 14 isolates were Helicobacter spp., of which 10 were H. pylori (2.0% of the total isolates). These isolates were tested for antibiotic susceptibility as well as ability to tolerate chlorine at 0.5 mg/L. To our knowledge, this is the first report of the occurrence of H. pylori in treated municipal drinking water. 1Department of Biology, College of Education, University of Basra, Basra, Iraq (Correspondence to Al-Sulami: Aminabdulah@yahoo.com). 2Department of Marine Environmental Chemistry, Marine Science Centre, Basra, Iraq. 3College of Medicine, University of Missan, Missan, Iraq. Received: 30/10/08; accepted: 16/03/09 قارعلاب ةصربلا ةظفامح في بشرلا هايم في ةيباّوبلا ةَّيوْلَمـلا ةموثرج صيخشتو دارفتسا ةعجم يزاغ ءاسيم ،يئاطلا اضر دممح دعسأ ،يمَل ُّسلا رابلجا دبع ينمأ ةرـتفلا للاخ ةصربلا ةظفامح في ةقطنم 22 نم بشرلا هايلم ةنيع 198 تع ُمجم دقو .ةحضاو يرغ تلازام ةيباوبلا ةَّيوللماب ىودعلا لاقتنا زرط نإ :ةصلالخا دقو .ةيزابرلاو ةيلاجملإا تاينولوقلا ميثارج ددع كلذكو يقبتلما رولكلا زيكرت َسيِقو .2007 ويلوي/زوتم ىتح 2006 ربوتكأ/لولأا نيشرت نم ًةدرفتسم 173 لىع اهيف ف ُّرعتلا نكمأو ،ر َّوحلما مدلاب راغلأا ايبمولوك تَبْنَتْس ُمم لىع تانيعلا عارزتسا دعب ،ةيموثرج ةعرزم 469 لىع نوثحابلا لصح نم تادَرفتسلما هذه رابتخا متو .)تادَرفتسلما لياجمإ نم %2.0( ةيباّوبلا ةّيوْلَملل طقف ةشرع اهنم ،ةَّيوْلَمـلا عاونلأ تناك اهنم ةشرع عبرأ ،ةيموثرج غلابلإا متي ةرم لوأ يه – ينثحابلا تامولعم بسح - هذهو .ل/غم 0.5 زيكرـتب رولكلا لمتح لىع اتهردقو ةيويلحا تاداضملل اهتيساسح ثيح .ةَلجاعلما بشرلا هايم في ةيباوبلا ةيوللما ةموثرج دوجو نع اهيف Isolation et identification d’Helicobacter pylori dans l’eau potable du gouvernorat de Bassora (Iraq) RÉSUMÉ Le mode de transmission d’une infection à Helicobacter pylori reste mal connu. Cent quatre-vingt-dix- huit échantillons d’eau potable provenant de 22 districts du gouvernorat de Bassora ont été recueillis entre octobre 2006 et juillet 2007. La concentration de chlore résiduel a été mesurée et le nombre de coliformes totaux et fécaux ont été comptés. Sur les 469 cultures bactériennes obtenues à partir de la gélose Columbia modifiée à l’urée, 173 isolats ont été identifiés. Seuls 14 d’entre eux correspondaient à Helicobacter spp., dont 10 à H. pylori (2 % du total des isolats). La sensibilité aux antibiotiques de ces isolats a été testée, de même que leur tolérance au chlore à une concentration de 0,5 mg/l. À notre connaissance, il s’agit du premier signalement d’H. pylori dans de l’eau potable municipale traitée. طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 921 Introduction Helicobacter pylori,  originally  classified  as Campylobacter pylori, is a Gram- negative, microaerophilic, spiral-shaped,  motile bacterium associated with gas- tritis, peptic ulcer, duodenal ulcer and  chronic  gastritis.  It  is  also  implicated  in  the development of  gastric  cancer  [1–3]. The mode  of  transmission  of  H. pylori  remains poorly understood.  It has been suggested that the housefly  has  the potential  to  transmit  the bac- terium, especially  in areas of  the world  with poor  sanitation  [4]. Other  likely  transmission  routes  are  faecal–oral,  iatrogenic or oral–oral [5]. While drink- ing water contaminated with faeces has  been proposed as a source of infection,  H. pylori  has  not  been  isolated  from  water except in some instances in which  it was detected using polymerase chain  reaction (PCR) on samples  from Co- lombia, Lima and Peru [6,7].  Three  epidemiological  studies  in  South America have  linked  transmis- sion  to  food and water.  In Chile, more  than 60% of  1815 Chileans  younger  than 35 years old and of lower socioeco- nomic groups were found to be H. pylori seropositive [8]. A study of 407 children  aged 2 months to 12 years in Peru also  concluded  that water was  the  vehicle  of infection, because children who used  the municipal water supply had a higher  prevalence of H. pylori  infection  than  children who used private wells [9]. Fur- thermore, an increased risk of infection  was  observed  in  children who  swam  in  rivers  and  streams  in  the  southern  Colombian Andes [10]. While all these  studies  confirmed  the  possibility  of  H. pylori  transmission via water, efforts  to isolate the bacterium from water have  been unsuccessful [11,12].  This  study  in Basra,  Iraq,  aimed  to  isolate H. pylori  from  treated drinking  water  and  investigate  the  relationship  of H. pylori to total and faecal coliforms  as well  as  its  susceptibility  to  several  antibiotics and to chlorine. Methods Enumeration and identification of H. pylori A total of 198 samples of drinking water  from 22 districts  in Basra governorate  were collected during the period Octo- ber 2006 to July 2007.  The concentration of residual chlo- rine  for each sample was measured us- ing a chlorine meter (Lovibond 2000).  Aliquots of 250 mL  from each sample  were filtered by  the membrane filtra- tion  technique using 47 mm cellulose  acetate filters with a nominal pore size  of 0.22 µm (Sartorius). The filter papers  were cultured on modified Columbia  urea  agar medium [13]  consisting of  Columbia agar supplemented with 1%  haemin, 5% urea solution, 4 µg of van- comycin and 0.12 mg of phenol red and  incubated at 37 ºC for 5–7 days under  microaerophilic  conditions (5% CO 2 , 10% H 2 , 85% N 2 )  for  the  isolation of  H. pylori. H. pylori   was  identified  using  biochemical  tests which  included:  the  catalase, oxidase and urease  tests,  tests  for hydrogen  sulphide (H 2 S) produc- tion,  nitrate  reduction,  growth  with  3.5% NaCl,  growth with  1%  glycine,  growth at varying  temperatures (25 °C  and 42 °C), growth on peptone-starch- dextrose agar and sensitivity  to cepha- lothin and nalidixic acid.  H. pylori isolates were tested for their  antibiotic  susceptibility  according  to  the method of Piddock  [14] using 7  antibiotics disks  including  tetracycline,  ampicillin,  amoxicillin,  erythromycin,  kanamycin, gentamicin and rifampicin  (Bioanalyze). The  isolates  of H. pylori  were  ex- posed  to 0.5 mg/L concentrations of  chlorine  for  4  different  time  periods  (10s, 20s, 40s and 60s) [15]. Enumeration of coliform bacteria The concentrations of coliform bacteria  were determined by filtering 2 × 10 mL  volumes of each sample using 47 mm  cellulose acetate filters with a nominal  pore  size of 0.45 µm (Sartorius). The  filters were cultured on m-FC agar and  m-Endo agar for detection of faecal and  total coliforms respectively. Results A total of 198 water samples were col- lected from 22 different districts during  the period of the study which extended  over the winter and summer seasons.  Measurement of  residual  chlorine  concentration  in  the  water  samples  showed that 41.3% of samples were free  of chlorine. Figure 1 shows the residual  chlorine concentrations  in  the districts  in  the winter  and  summer  seasons.  In  some of the districts the concentrations  varied greatly between the seasons and  were  usually  higher  in  summer  than  winter.  Only  14.1%  of  the  samples  con- formed to World Health Organization  criteria for water quality of zero fecal and  total  coliforms  [16]  (Table 1). Only  80/198  isolates were positive  for  the 3  biochemical  tests  for H. pylori positiv- ity  (urease,  catalase  and oxidase). On  completing the other biochemical tests  only 14  isolates were characterized as  Helicobacter spp.: 4 H. mustelae  and 10  H. pylori (Table 2). Therefore isolates of  H. pylori comprised 2.0% of the total iso- lates and 5.0% of the total samples. The  presence of these isolates was higher in  certain districts than others, especially in  Old Basra, Al-Jamiayat, Al-Zubair and  Al-Ma’aqal. Antibiotic  susceptibility  tests  showed  that 80% of H. pylori  isolates  were  susceptible  to  tetracycline,  50%  to  ampicillin  and  amoxicillin,  40%  to  kanamycin, gentamicin and rifampicin  and 30% to erythromycin.  Inactivation  of H. pylori isolates  by  chlorination  showed that H. pylori were not sensitive  to chlorine,  since  the final numbers of  bacterial colonies were high after each  period of chlorine exposure, i.e. less than  1-log reduction (Table 3).  EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 922 Discussion Culture  is  considered  the gold  stand- ard  for detection of  bacteria,  but  the  method  is not  sensitive, and  is  specific  only  if  additional  testing  is performed  on  the  isolates. However,  the applica- tion of molecular biology identification  methods has potential drawbacks [16].  The method of  choice  involves PCR  amplification of specific H. pylori genes  such as ureA and ureC. Although  this  technique  appears  to  be  sensitive,  it  lacks  specificity  [17]. Therefore other  approaches have been reported involv- ing the use of H. pylori 16s rRNA [18,19]  or different sets of primers [16]. The present study was intended as a  preliminary probing for the presence of  H. pylori in drinking water. A more com- prehensive  research  project  on  both  drinking water and sewage is underway,  for which a combination of biochemical  and PCR using ureA primer gene and  rRNA gene primer is applied.  The  mode  of  transmission  of  H. pylori  remains an area of discussion.  Increased  risk  of  infection  has  been  associated with  contaminated drink- ing water  [9] and  the consumption of  uncooked vegetables irrigated with un- treated sewage [8]. Currently the role of  water in dissemination of this pathogen  remains problematic  since H. pylori is a fastidious organism and has been dif- ficult  to  isolate  from  environmental  sources such as water [15]. Studies of the presence of H. pylori in the aquatic environment have relied  on  molecular  methods  using  PCR,  immunomagnetic  separation  and  au- toradiography  [20,21]. These  studies  suggest  that  the organism may survive  in water for an extended period of time  and that H. pylori infection is spread by  contaminated water  [15]. The present  study  succeeded  in  isolating H. pylori from chlorine-treated drinking water  by using culture methods. We can link  this  to  the  low concentration of  chlo- rine in the water samples, permitting it  to grow and survive  in  large numbers,  Table 1 Mean concentrations of total and faecal coliforms in water samples isolated from different districts of Basra (n = 198) District Total coliforms (CFU/100 mL) Faecal coliforms (CFU/100 mL) Abu-Al-Khaseeb 1000 1300 Al-Ashar 415 393 Al-Tuwaisa 322 302 Al-Zubair 208 254 Al-Eskan 197 207 Al-Jazair 380 200 Old Basra 236 177 Al-Hussain Q 106 125 Al-Hakeemya 60 109 Al-Qibla 1500 100 Five Miles 5 92 Shatt-Al-Alarab 113 76 Al-Guzaiza 187 57 Al-Maaqal 81 48 Al-Abela 7 30 Al-Jamiayat 16 24 Al-Jubaila 525 20 Garmat-Ali 45 14 Al-Asmaaee 14 10 Al-Qurna 0 2 Al-Junaina 10 0 CFU = colony-forming units. Figure 1 Residual chlorine concentrations of water samples from Basra during winter and summer seasons (n = 198) طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 923 and to the water distribution system, since this  bacterium has the ability to form biofilms in wa- terpipes  [22,23]. Furthermore,  the waterpipes  suffer from breaks and corrosion at many sites,  which cause drinking water contamination by  sewage  infiltration  and  rain  leakage  into  the  system [24,25].  The higher  isolation  rate of  this bacterium  in some districts than others may be due to the  increased breaks and corrosion sites in the water  supply systems which raise the rate of contami- nation.  In some of  these samples H. pylori were  detected in the absence of coliforms, suggesting  the shortcomings of  these  indicators  for patho- genic bacteria.  Another explanation  for  the high  isolation  rate of H. pylori  is  the use of modified Colum- bia urea agar medium, which may enhance  the  growth of this bacterium. Isolates of H. pylori were 80%  tetracycline  sensitive. This  resistance may  be  a  result  of  mutation  in  the 16rRNA  gene,  which  is  the  target  for  this  antibiotic  [26]. For kanamycin,  gentamicin,  erythromycin  and  rifampicin  the  resistance could be  a  result of  gene mutation  such as mutation of  the rpoB gene which  is  the  target  for rifampicin. Generally antibiotic resist- ance is considered the main problem associated  with H. pylori treatment since antibiotic-resistant  strains  have  become  prevalent  throughout  the world and are  the main cause of  failure  in  H. pylori treatment. Increased resistance of H. pylori  to chlorine  may  be  attributed  to  the  growth  of  bacterial  cells in the form of biofilms that make these cells  acquire greater  resistance  to disinfectants  than  free  cells  [27]. However,  some  studies  found  that H. pylori  isolates were  resistant  to chlorine  and ozone but  sensitive  to monochloramine  disinfection [28].  To our knowledge,  this  is  the first  report of  the  isolation of H. pylori  in municipal  treated  drinking water and  this could be of epidemio- logical  significance. Further  research  is needed  to  establish which  factors  affect  the  ability of  H. pylori to survive in distribution systems and be  isolated from drinking water, such as the bacte- rial strains, density of bacteria in the distribution  systems,  type of waterpipe materials,  efficiency  of disinfection process and  the  techniques and  materials used for culture.Ta bl e 2 Bi oc he m ic al te st s of 1 4 is ol at es o f H el ic ob ac te r s pp . Is ol at e no . C at al as e O xi da se U re as e N itr at e re du ct io n H 2S G ro w th w it h 3. 5% N aC l G ro w th o n PS D a ga r G ro w th o n 1% g ly ci ne G ro w th a t 42 ° C G ro w th a t 25 ° C C ep ha lo th in re si st an t N al id ix ic a ci d re si st an t H . p yl or i 1 + + + – – – + – + – + + 2 + + + + – – + – + – + + 3 + + + – – – + – – – + – 4 + + + – – – + – – – + – 5 + + + – – – + – – – + + 6 + + + – – – + – – – – + 7 + + + – – – + – + – – + 8 + + + – – – + – – – + + 9 + + + + – – + – – – + + 10 + + + – – – + – – – + + H . m us te la e 1 + + + + – – – + + – + – 2 + + + + – – – + + – + – 3 + + + + – – – + + – + – 4 + + + + – – – + + – – – PS D = p ep to ne -s ta rc h- de xt ro se . EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 924 References Table 3 Chlorine inactivation of Helicobacter pylori isolates in 0.5 mg/L chlorine at different exposure times Isolate no. Exposure time Initial 10 s 20 s 40 s 60 s (log10 CFU/mL) (log10 CFU/mL) (log10 CFU/mL) (log10 CFU/mL) (log10 CFU/mL) 1 3.07 2.93 2.51 2.13 1.85 2 4.15 1.01 3.13 1.09 0.95 3 5.16 4.95 4.81 4.57 3.91 4 2.11 2.09 1.33 1.15 1.85 5 5.61 2.15 3.88 3.60 3.47 6 4.81 4.67 2.56 2.31 0.02 7 3.75 3.51 3.55 3.19 3.65 8 UC UC 2.16 1.25 0.91 9 2.15 1.35 1.17 0.57 1.01 10 4.95 4.88 3.11 2.15 1.01 CFU = colony-forming units; UC = uncountable. Goodwin CS. 1. Campylobacter pylori, detection and culture. In: Rathbone BJ, Healthy RV, eds. Campylobacter pylori and gastro-duodenal disease. Oxford, Blackwell Scientific Publica- tions, 1989:60–62. Guidelines for drinking-water quality2. , 3rd ed. Volume 1: Recom- mendations. Geneva, World Health Organization, 2004. Lavigne A, de Reuse H. Determination of 3. Helicobacter pylori pathogenicity. Infectious Agents and Disease, 1996, 5:191–202. Grubel P et al. Vector potential of houseflies (4. Musca domestica) for Helicobacter pylori. Journal of Clinical Microbiology, 1997, 35:1300–1303. Dunn BE et al. 5. Helicobacter pylori. Clinical Microbiology Re- views, 1997, 10:720–741. Hulten K et al. 6. Helicobacter pylori in the drinking water in Peru. Gasroenterology, 1996, 110:1031–1035. Schauer DB et al. Detection of 7. Helicobacter pylori in drinking water using polymerase chain reaction amplification. Gut, 1995, 37:A27. Hopkins RJ et al. Seroprevalence of 8. Helicobacter pylori in Chile: vegetables may serve as one route of transmission. Journal of Infectious Diseases, 1993, 168:222–226. Klein PD et al. Water source as risk factor for 9. Helicobacter py- lori infection in Peruvian children. Gastrointestinal Physiology Working Group. Lancet, 1991, 337:1503–1506. Goodman KJ et al. 10. Helicobacter pylori infection in the Colombi- an Andes: a population-based study of transmission pathways. American Journal of Epidemiology, 1996, 144:290–299. Adams BL et al. Survival of 11. Helicobacter pylori in a natural fresh- water environment. Applied and Environmental Microbiology, 2003, 69:7462–7466. Engstrand L. 12. Helicobacter in water and waterborne routes of transmission. Journal of Applied Microbiology, 2001, 90:80S– 84S. Al-Sulami A et al. Primary isolation and detection of 13. Helico- bacter pylori from dyspeptic patients: a simple, rapid method. Eastern Mediterranean Health Journal, 2008, 14 (2):268–276. Piddock LJJ. Techniques use for the determination of antimi-14. crobial resistance and sensitivity in bacteria. Antimicrobial Agents Research Group. Journal of Applied Microbiology, 1990, 68:307–318. Johnson CH et al. Inactivation of 15. Helicobacter pylori by chlorin- ation. Applied and Environmental Microbiology, 1997, 63:4969– 4970. Liu H et al. Specific and sensitive detection of 16. H. pylori by real time RT-PCR and in situ hybridization. PLoS ONE, 2008, 3(7):e2689.doi:10.1371/J. Camorlinga-Ponce M et al. Topographical localisation of 17. cagA positive and cagA negative Helicobacter pylori strains in the gas- tric mucosa; an in situ hybridisation study; an in situ hybridiza- tion study. Journal of Clinical Pathology, 2004, 57:822–828. Kolbert CP, Persing DH. Ribosomal DNA sequencing as a tool 18. for identification of bacterial pathogens. Current Opinion in Microbiology, 1999, 2:299–305. Smith SI et al. Comparison of three PCR methods for detec-19. tion of Helicobacter pylori DNA and detection of cagA gene in gastric biopsy specimens. World Journal of Gastroenterology, 2004, 10:1958–1960. West AP, Millar MR, Tompkins DS. Effect of physical environ-20. mental on survival of Helicobacter pylori. Journal of Clinical Pathology, 1992, 45:228–231. Shahamat M et al. Use of autoradiography to assess viability of 21. Helicobacter pylori in water. Applied and Environmental Micro- biology, 1993, 59:1231–1235. Al-Taee, M.R. Assessment of water quality due to microbial 22. growth in drinking water distribution systems in Basrah city. Marina Mesopotamica, 2001, 16(1):37–46. Momba MNB, Makala N. Comparing the effect of various 23. pipe materials on biofilm formation in chlorinated and com- bined chlorine-chloraminated water systems. Water SA, 2003, 30(2):175–182. Geldreich EE et al. Searching for a water supply connection 24. in the Cabool, Missouri disease outbreak of Escherichia coli 0157:H7. Water Research, 1992, 26(8):1127–1137. Sartory PD, Holmes P. Chlorine sensitivity of environmental, 25. distribution system and biofilm coliforms. Water Science and Technology, 1997, 35(11–12):289–292. Ribeiro QML et al. Detection of high-level tetracycline resist-26. ance in clinical isolates of Helicobacter pylori using PCR-RFLP. FEMS Immunology and Medical Microbiology, 2004, 40:57–61. طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 925 Progress on sanitation and drinking-water: 2010 update The above-mentioned report describes the status and trends with respect to the use of safe drinking-water and basic  sanitation and progress made towards the Millennium Development Goals’ drinking-water and sanitation target. As the world approaches 2015, it becomes increasingly important to identify who are left behind and to focus on the  challenges of addressing their needs. This report therefore shows disparities: the gap between progress in providing  access to drinking-water versus sanitation; the divide between urban and rural populations in terms of the services  provided; differences in the way different regions are performing, bearing in mind that they started from different  baselines; and disparities between different economic strata in society. The information presented in this report includes data from household surveys and censuses completed during the  period 2007–2008. It also contains datasets from earlier surveys and censuses that have become available to the Joint  Monitoring Programme (JMP) since the publication of the previous JMP report in 2008. Further information about this and other WHO publication is available at: http://www.who.int/publications/en/  Ford TE. The microbial ecology of water distribution and outfall 27. systems. In: Ford TE, ed. Aquatic microbiology: an ecological ap- proach. London, Blackwell Scientific, 1993:455–482. Baker KH, Hegarty JP. Presence of 28. Helicobacter pylori in drink- ing water is associated with clinical infection. Scandinavian Journal of Infectious Diseases, 2001, 33:744–746. EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 926 Microbiological and chemical profile of Lebanese qishta (heat-coagulated milk) Z.G. Kassaify,1 M. Najjar,1 I. Toufeili 1 and A. Malek 1 ABSTRACT Qishta is a popular Middle Eastern coagulated cream product, prepared using a traditional heating and skimming process. This study in Lebanon aimed to assess the microbiological and chemical profile of the product. Samples selected from 31 different manufacturers and outlets were analysed using standard methods. The plate counts for the various microorganisms were either of borderline acceptability or unacceptable for Escherichia coli, Salmonella spp. and Listeria monocytogenes (detected in 32%, 7% and 42% of the analysed samples respectively). Chemically, the mean moisture content [67.5 (SD 2.6) g/100 g] and pH (6.53) were relatively high. High counts of spoilage and pathogenic microorganisms and the nature of its chemical composition make qishta highly perishable. 1Department of Nutrition and Food Sciences, Faculty of Agricultural and Food Sciences, American University of Beirut, Beirut, Lebanon (Correspondence to Z.G. Kassaify: zk18@aub.edu.lb). Received: 10/01/09; accepted: 08/03/09 )ةرارلحاب رثختلما نبللا( ةينانبللا ةدشقلل يئايميكلاو يجولويبوركلما مسترلما كلام لمأ ،لييفوط دماع ،راجن دممح ،يفيصقلا ةنيز ةساردلا هذه ىعستو .ةداتعلما دشقلاو ينخستلا ةيلمعب َّرضتح يهو ،طسولأا قشرلا في ةجئارلا نبللا رثتخ تاجتنم دحأ يه ةدشقلا نإ :ةصلالخا قرطلاب تانيعلا ليلتح ىرجو ،عيبلاو عينصتلل ًافلتمخ ًاعقوم 31 نم تانيع تيرتخا دقو .جتنلما اذله يئايميكلاو يجولويبوركلما مسترلما مييقت لىإ ةينانبللا ةدِحْوَتْسُمـلا ةييرتسللاو ،ةلينولماسلاو ،ةينولوقلا ةيكيشرلإل ةبسنلاب داكي وأ لوبقم يرغ ةيموثرلجا عرازلما في ةفلتخلما تابوركلما ددع ناكو .ةيرايعلما فارحناب( 67.5[ ةبوطرلا ىوتمح طسوتم ناك دقف ،ةيئايميكلا ةيحانلا نم امأ .)بيترـتلا لىع ةصوحفلما تانيعلا نم %42و ،%7و ،%32 في تفشتكا ذإ( بيكرـتلا ةعيبطو ،داسفلل ةثدحلماو ةضرملما تابوركلما نم يربكلا ددعلا نإ .ًايبسن ْينَعفترم )6.53( ءاهابلاو ]مارغ 100/مارغ )2.6 :هُرْدَق يرايعم .بطعلا ةعيسر اهلعيج ةدشقلل ةيئايميكلا Profil microbiologique et chimique de la qishta libanaise (lait coagulé par traitement thermique) RÉSUMÉ La qishta est un produit crémeux coagulé très populaire au Moyen-Orient, préparé selon un procédé traditionnel de chauffage et d’écrémage. Cette étude réalisée au Liban visait à évaluer le profil microbiologique et chimique de ce produit. Les échantillons choisis auprès de 31 fabricants et points de ventes ont été analysés à l’aide de méthodes standardisées. Le comptage sur plaque des divers micro-organismes a fourni des valeurs soit à la limite de l’acceptabilité, soit non acceptables pour Escherichia coli, Salmonella spp. et Listeria monocytogenes (détectées dans 32 %, 7 % et 42 % des échantillons analysés, respectivement). Sur le plan chimique, la teneur moyenne en eau (67,5 g/100g [ET 2,6]) et le pH (6,53) sont apparus relativement élevés. Compte tenu du nombre important de micro-organismes pathogènes et provoquant l’autolyse qu’elle contient et de la nature de sa composition chimique, la qishta est une denrée très périssable. طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 927 Introduction Outbreaks of disease  in humans have  been  traced  to  the  consumption  of  milk  and  milk  products  from  both  unpasteurized  and  pasteurized milk.  Pathogens such as Listeria monocytogenes can survive and multiply after pasteuri- zation,  leading  to  recontamination of  dairy  products  [1,2]. Although  strict  microbiological  standards have been  set  for milk and dairy products,  tradi- tional dairy products in many countries  are still produced under poor hygienic  conditions with different manufactur- ing  technologies  that  require extensive  handling of perishable ingredients [3,4].  Identifying the reservoirs of a pathogen  is  vital  for  the control of  isolated out- breaks and epidemics. Qishta is  a Middle Eastern hand- made heat-coagulated  cream product  that  resembles  the  clotted or  scalded  cream  that  is made principally  in  the  west  of  England.  It  is  prepared  from  powdered or  pasteurized  liquid milk  in  small  dairy plants or  in  large-scale  bakeries to be consumed fresh as a des- sert or used as a filling  in a number of  traditional  sweets. Unlike many other  coagulated dairy products, qishta  is not  fermented or coagulated by chemical or  microbial methods and, despite  recent  increases  in  the volume of production,  the process  remains quite  traditional.  Milk  is heated  in a  large  tilted shallow  pan until  it boils. This allows  the pro- teins  to  coagulate  and  entrap  the  fat  particles while  they float  to  the surface  and move  towards  the  colder part of  the pan at temperatures around 60 ºC.  The process  takes over 3 hours  for  the  collection of  the final product. Qishta is then cooled in bulk to 45 ºC at room  temperature  for  2  hours  before  it  is  stored under  refrigeration (process as  described  by  Refaat Hallab,  Tripoli,  Lebanon). The shelf-life of qishta is only  24 hours  at  room  temperature  and 4  days at 2–5 ºC.  Personal contacts and  information  from hospitals suggest a high incidence  of  foodborne diseases associated with  qishta  in  various  regions of Lebanon.  Since  to our  knowledge  there  are no  published studies on qishta, the present  study  aimed  to  assess  the microbial  quality  and  chemical  constituents  of  the product. It was hoped that the data  would be applicable to other coagulated  milk products in the region, with a view  to enhancing the process, extending the  shelf-life of  the product and  improving  public safety.  Methods Samples Triplicate  samples  of  qishta  (250  g  each) were collected  from 31 different  manufacturers and outlets  in the north  and the south of Lebanon and the Bei- rut area between April and November  2007. The samples were transported in  sterile plastic bags to the laboratory un- der aseptic and refrigerated conditions.  Microbiological and chemical analyses  were performed within 1–2 hours after  purchasing.  Microbiological analysis Product samples were analysed micro- biologically and identified according to  standard methods  for  total mesophilic  bacteria,  total coliforms and  faecal col- iforms, Enterobacteriaceae, Salmonella spp., L. monocytogenes, Staphylococcus aureus and yeasts and moulds [5]. In  the  procedure,  10  g  of  qishta were  homogenized  with  90  mL  of  sterile 0.1% peptone water (356-4684)  in  a  stomacher  (Seward 400, Seward,  London)  for at  least 2 minutes. Serial  dilutions were prepared  in 0.1% pep- tone water  (356-4684)  and  samples  of 0.1 mL of each of the 10–2, 10–4 and 10–6 dilutions were  spread on appro- priate media  in duplicates  and 1 mL  was used  for  the pour-plate  technique.  Total mesophilic bacteria were enumer- ated on plate-count agar (356-4475) at  30 ºC for 48 hours while total coliforms  and Escherichia coli were differentiated  and  enumerated  on RAPID’E.coli  2  agar  (356-4024)  by  the  pour-plate  technique  and plates were  incubated  at 37 ºC  for 24 h. Total Enterobacte- riaceae and Sta. aureus were respectively  detected  on  violet-red  bile  dextrose  agar (256-4584) and Baird-Parker agar  supplemented with egg-yolk–tellurite  emulsion  (356-4814)  and  incubated  at 37 ºC  for 24 h. Sta. aureus were  fur- ther  confirmed  biochemically  using  rabbit plasma (355-6352). Yeast  and  moulds were identified on yeast glucose  chloramphenicol agar (256-4104) with  plates incubated at 25 ºC for 5 days.  For   the   i solat ion   of   S a l m o - nella  spp.  and  L. monocytogenes,  the pre-enrichment/enrichment  selec- tive plating method was used [5]. For  Salmonella  spp.,  selective  enrichment  was performed in rappaport-vassiliadis- soya broth (256-4324) to be incubated  at 41.5 ºC. After 24 h of  incubation,  a  0.1 mL sample was plated on RAPID  Salmonella  agar (356-4705) and plates  were incubated at 37 °C for 24 h (± 2 h)  and  another  0.1 mL was  transferred  onto xylose–lysine–desoxycholate agar  plates (356-9124) that were  incubated  at 37 ºC for 24 h. Salmonella spp. colonies were identi- fied biochemically by the lysine iron agar  (B211363) and tryptic sugar  iron agar  (D4402) 2 slants biotyping  technique.  Additional  confirmation  for  positive  Salmonella  spp.  colonies was done by  the API 20E bacterial identification test  strip (Biomérieux, France). For L. mono- cytogenes,  Fraser ½ broth  (356-4616)  was used  in  the  selective  enrichment  and after incubation for 1 h at 20 ºC, 0.1  mL of the homogenate was transferred  onto RAPID’L. mono agar (356-3694)  plates to be incubated at 37 ºC for 24– 48 h. Typical L. monocytogenes colonies  were  afterwards  selectively  identified  and enumerated.  All media were supplied by Bio-Rad  Laboratorios, California, USA. Micro- bial counts were reported as geometric  means of colony-forming units (CFU)  per  g  of qishta,  except  for Salmonella EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 928 spp.  and L. monocytogenes which were  reported as present or absent. Chemical analysis The pH of qishta was measured using a  pH meter with a glass electrode (Orion,  USA). Moisture content,  salt, proteins,  carbohydrates  and  fat  contents were  determined according  to  the Associa- tion of Official Agricultural Chemists  (AOAC) standard methods  for proxi- mate  analysis  of  dairy  products  [6].  Samples were analysed in duplicate and  each proximate analysis was  repeated  twice. Statistical analysis Data were analysed by analysis of vari- ance using SPSS,  version 8.0  for Win- dows and means were separated by the  Duncan multiple range test. Significance  was defined at P < 0.05 and numerical  results are given as means and standard  deviations (SD). Results The mean plate  counts of  the various  microorganisms  in  the analysed qishta samples  are  shown  in  Table  1. The  mean total aerobic count was 4.42 (SD  0.10)  log 10  CFU/g and total coliforms  were 3.23 (SD 0.09) log 10  CFU/g. The  mean plate counts of E. coli, Enterobac- teriaceae  and Sta. aureus also  ranged  from 3–4  log 10  CFU/g (Table 1). Of  the  tested qishta  samples 9/31 (29%)  had mesophilic bacteria counts  in  the  range of 3–4 log 10  CFU/g while 15/31  (48%)  had  yeast  and  mould  levels  within  that  range. Salmonella  spp.  and  L. monocytogenes were present  in 2/31  (7%) and 13/31 (42%) of the samples  respectively (detailed  identification of  Salmonella  spp. will  be  reported  in  a  future study). The  results  of  chemical  analyses  (moisture, acidity and pH and fat, pro- teins  and  ash  content)  are  shown  in  Table 2. There was a high mean mois- ture content at 67.5 (SD 2.6) g/100 g.  The mean fat content was 13.0 (SD 2.4)  g/100 g (Table 2). The mean pH of the  samples was 6.53. Table 3 compares the mean micro- bial counts of qishta samples from large- scale manufacturers  and  small-scale  manufacturers/retailers. There were no  significant differences between  the 2  categories of suppliers in any of the pa- rameters studied (P > 0.05) (Table 3). Discussion The chemical composition of qishta as determined  in  this  study,  along with  other procedural aspects  in  the manu- facturing process, may have  contrib- uted  to  the high microbial  incidence  encountered  in  the  analysed  samples  and  the  likelihood  of  rapid multipli- cation  if  contamination  did  occur.  Most  of  the microbial  counts  in  our  samples of qishta were  relatively high  when compared with  the  international  microbiological criteria [7], on  the up- per margin  for  rejecting products. The  high incidence of pathogenic and spoil- age microorganisms  in qishta may be  accounted for by contamination during  manufacturing or post-processing and  cross-contamination  in plants  and  in  refrigerators or retail stores.  Because the total mesophilic bacte- ria and yeasts and moulds are indicators  of  spoilage,  their numbers are essential  in deciding the shelf-life of the product.  This may  explain  the observed  short  shelf-life of qishta  in the Lebanese mar- ket. In addition, the bitter taste in qishta which develops during prolonged refrig- erated storage may be attributed to the  high mould counts. Hence,  the  results  of  this  study  agree with  similar work  performed on dairy products in Turkey  [8]. Milk-coagulated products  such as  qishta are rich in nutrients and have high  moisture content (in  this study almost  70%), which is conducive to the growth  of spoilage and pathogenic microorgan- isms [9].  The  presence  of  coliforms  in  the  various samples  is  indicative of  insuffi- cient sanitary conditions during manu- facturing and  storage. Coliforms  such  as E. coli, Enterobacteriaceae  and Sta. aureus are markers for unsafe foods, due  to  their pathogenic nature. Hence  the  high mean microbial  counts obtained  Table 1 Microbial plate counts of 31 qishta samples from various manufacturers in Lebanon Microorganism No. of positive samples/total no. of samples analysed Mean (SD) plate counts (log10 CFU/g) Minimum Maximum Total aerobic count 31/31 4.42 (0.10) 2.41 7.01 Total coliforms 31/31 3.23 (0.09) 1.39 5.43 Escherichia coli 10/31 3.01 (0.60) 2.30 4.03 Enterobacteriaceae 31/31 3.67 (0.06) 1.74 5.66 Yeasts & moulds 31/31 4.54 (0.08) 2.19 7.27 Staphylococcus aureus 30/31 3.03 (0.10) 0.05 4.63 Salmonella spp. 2/31 –a – – Listeria monocytogenes 13/31 –a – – aSalmonella spp and L. monocytogenes are reported as present or absent. SD = standard deviation; CFU = colony-forming units. طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 929 in  this  study  suggest  that  the product  may be hazardous for human consump- tion. This  result  agrees with work  in  other  countries on different products  that have similar characteristics to qishta [10–13]. Mandokhot et al. [12] and Gill  et al. [13] in studies from different parts  of  India  indicated  that khoa  (partially  desiccated milk) is often contaminated  with pathogens such as Sta. aureus and Bacillus cereus.  Furthermore, Rajorhia  et  al.  [10]  and Kumar  et  al.  [11]  also  reported that paneer (a coagulated milk  product) is often contaminated with Sta. aureus and coliforms. The high counts of  S. aureus  in  particular may be due  to  extensive handling by personnel dur- ing  the elaborate multi-stage process- ing. Dairy products are a  rich medium  and can enable  the growth of  this and  many other pathogens. Hence, Halpin- Dohnalek et al. reported that both sweet  and neutralized  sour  cream products  support  the growth of all strains of Sta. aureus [14].  The  qishta   samples  also  tested  positive  for Salmonella  spp.,  while L. monocyto genes was detected  in 45% of  samples. According to the microbiologi- cal  criteria  for milk-based products  in  the European Union  those 2 high-risk  pathogens should be absent for a prod- uct  to be  considered  safe  for  human  consumption  [15]. There  are  several  explanations  for  the  results confirming  the presence of those 2 pathogens in the  final product. The heating  step during  processing or  the contact  time at  that  boiling temperature may be insufficient  to  eliminate  the  pathogens. Another  possibility is the existence of heat-resist- ant Salmonella and Listeria strains and this hypothesis will be  investigated  in  another  study. On the other hand,  the  high microbial and pathogen counts ob- served in qishta are indicative of possible  contamination before, during and after  processing.  Several  researchers  have  noted the prevalence of these pathogens  in milk-based products. For  instance,  Szwarcbort de Tamsut et  al.  reported  the presence of Sal. typhimurium and other pathogens such as Sta. aureus and Shigella spp. in pasteurized milk creams  in Venezuela [16]. As  for L. monocytogenes,  several  re- searchers  in the Middle Eastern region  reported the presence of this pathogen in  raw milks and pasteurized milk products  [17–20]. L. monocytogenes was  isolated  from cheese products  in Turkey, some  of which were either  ripened or brine- salted [17,21]. Rudolf et al.  also  found  L. monocytogenes in European red smear  cheese [22], while Cordano et al. found  Listeria spp. in soft cheeses samples [23].  Contamination might also occur post- processing from environmental sources  and cross-contamination  in  the dairy  plant and/or retail stores or inadequate  processing [22,24,25]. It has been noted  by Sergelidis et al.  that contamination  can also occur because of colonization  of L. monocytogenes  in  refrigerators  in  retail  stores  [26]. L. monocytogenes  can  survive a number of processes and can  remain viable  in  the final product  for  a considerable  length of  time [19,27].  Therefore, in most cases the contamina- tion sources are likely to be insufficient  hygiene during  the milking and manu- facturing process. The microbial  counts were mean  values of the total samples collected and  analysed  from the various manufactur- ers and retailers. Theoretically, the range  of microbial counts should be the result  of differences between major manufac- turers who apply hazard analysis critical  control points (HACCP) procedures or  have a controlled sanitation programme  and the small retail stores and/or manu- facturers who do not abide by any strict  sanitation procedures. However,  there  were no significant differences in the mi- crobial counts between large and small  suppliers. This could be attributed to the  fact that the larger manufacturers supply  the small retailers with the product and  the shelf-life of the product is short. Table 2 Chemical composition, pH and acidity of the 31 samples of qishta analysed Parameter Mean (SD) Minimum Maximum Moisture (g/100 g) 67.5 (2.6) 62.4 72.0 Fat (g/100 g) 13.0 (2.4) 9.2 18.3 Protein (g/100 g) 12.9 (1.9) 9.8 16.2 Ash (g/100 g) 1.8 (0.14) 1.6 2.2 Aciditya (g/100 g) 0.3 (0.001) 0.11 0.5 pH 6.53 (0.14) 6.28 6.82 aAs lactic acid. SD = standard deviation. Table 3 Comparison of microbial plate counts between large-scale manufacturers and small-scale manufacturers/retailers in Lebanon Microorganism Mean (SD) plate counts (log10 CFU/g) Large-scale producers (n = 13 samples) Small-scale producers/retailers (n = 18 samples) Total aerobic count 4.36 (1.34) 4.48 (1.02) Total coliforms 3.14 (1.19) 3.29 (0.85) Enterobacteriaceae 3.41 (1.16) 3.85 (0.99) Yeasts & moulds 4.21 (1.70) 4.78 (1.14) Staphylococcus aureus 2.89 (1.24) 3.13 (0.66) No significant difference between large-scale and small-scale producers, P > 0.05. SD = standard deviation; CFU = colony-forming units. EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 930 Such  chemical  characteristics  of  qishta  are essential  in determining  the  hazards associated with the product. The  mean values show that qishta has a high  moisture content which  is comparable  to  thickened yoghurts and other  types  of clotted creams. Furthermore, the pH  of qishta is higher than many other dairy  products such as whipped creams, clot- ted  creams  and  yoghurts  [28].  Such  findings explain the short shelf-life of the  product. Furthermore, the combination  of high moisture content and pH  is  a  major  factor  in  rendering  the product  susceptible  to high microbial contami- nation and growth. The proximate analysis results  indi- cate  that  the product  is nutritious and  not excessively high  in  its  fat composi- tion, with a  fat content (13.0 g/100 g)  close to but less than sour cream (18%)  or  light  creams (20%) and much  less  than whipping cream (30%) and other  types of clotted cream (55%).  The  labour-intensive  and  lengthy  preparation,  storage  and  cooling  procedures  and  the  poor  hygienic  environment observed  in many of  the  facilities visited are concerning. An ad- ditional  safety concern  is  that  in  some  Cullor, J.S. Risks and prevention of contamination of dairy 1. products. Revue Scientifique et Technique, 1997, 16:472–481. Oliver SP et al. Foodborne pathogens in milk and dairy farm 2. environment: Food Safety and public health implication. Food- borne Pathogens and Disease, 2005, 2(2):115–129. Rampling A. Raw milk cheeses and Salmonella. 3. British Medical Journal, 1996, 312:67–68. Freitas AC et al. Technological optimization of Picante cheese 4. using microbiological, chemical and physical criteria. Journal of Food Engineering, 1999, 41:163–175. Bacteriological analytical manual5. , 8th ed. Silver Spring, Mary- land, US Food and Drug Administration, 1995:5.01–5.20. AOAC official methods of analysis6. , 17th ed. Washington DC, As- sociation of Official Analytical Chemists, 2003. Principles for the establishment and application of microbiologi-7. cal criteria for foods. Rome, Codex Alimentarius Commission, 1997 (CAC/GL 21). Osman A et al. A survey on the microbiological quality of carra, 8. a traditional Turkish cheese. Journal of Food Engineering, 2004, 66:401–404. Sawhney IK et al. Evaluation of glycerol as a water activity 9. modifier in khoa. Journal of Food Science and Technology, 1994, 31:252. References Rajorhia GS et al. Quality of paneer marketed in Karnal and 10. Delhi. Indian Journal of Dairy Science, 1984, 37:274–275. Kumar V et al. Incidence of coliforms in indigenous milk prod-11. ucts. Indian Journal of Dairy Science, 1989, 42:579–580. MandokhotVV, Garg SR. Market quality of khoa, burfi and 12. pera: A critical review. Journal of Food Science and Technology, 1986, 22:299–304. Gill JPS, Joshi DV, Kwatra MS. Qualitative bacteriological 13. survey of milk and milk products with special reference to Staphylococcus aureus. Indian Journal of Dairy Science, 1994, 47:680–682. Halpin-Dohnalek MI et al. Growth and production of entero-14. toxin A by Staphylococcus aureus in cream. Journal of Dairy Sci- ence, 1989, 72(9):2266–2275. Council of the European Communities. Council Directive 15. 92/46/EEC of 16 June 1992 laying down the health rules for the production and placing on the market of raw milk, heat- treated milk and milk-based products. Official Journal L 268, 14/09/1992:0001–0031. Szwarcbort de Tamsut L et al. Calidad microbiologica de las 16. cremas de leche pasteurizadas elaboradas en Venezuela [Microbiological quality of pasteurized milk creams manufac- tured in Venezuela]. Archivos Latinoamericanos de Nutrición, 1999, 49 (1):76–80. manufacturing  locations,  in an attempt  to  cool  the  product  faster  before  re- frigeration,  fans without  air  filters  are  used. Furthermore qishta in many retail  outlets, whether sold bulk or in desserts,  is unpackaged and unprocessed. This  amplifies the problem of contamination  and cross-contamination. Furthermore,  the problem of microbial  safety  is  of  course exacerbated during  spring and  summer months,  when  the  ambient  temperatures  in  Lebanon  rise  above  25 ºC and most of the facilities have no  or insufficient cooling systems. Concerning the contamination with  moulds, an entry airlock to avoid direct  contact with  the outside air or an area  with positive air pressure where major  parts of the process are carried out could  alleviate this problem [29]. Conclusions The current study shows that the perish- able coagulated dairy product qishta is a high-risk product  since  its  chemical  composition makes  it  susceptible  to  growth of microorganisms and because  most of  the  tested  samples  showed a  high frequency of pathogenic and spoil- age microbial  contamination. Because  qishta  is  consumed  with  no  further  processing to reduce any microbial load,  the results are concerning. This study is  important because qishta  is consumed  widely  not  only  in Lebanon  and  the  region  but  is  becoming  increasingly  popular elsewhere.  In order  to  improve  the safety and  quality of  the product, processing and  storage needs  to be carried out under  good hygienic conditions and HACCP  systems  should be  implemented.  It  is  also imperative that the product should  be stored continuously under  refriger- ated conditions  to avoid post-heating  contamination  during  cooling  and  packing procedures.  Acknowledgements This study was supported by the Ameri- can University of Beirut Research Board  Grant. Great  appreciation  is  expressed  from  the  authors  to  the personnel  at  the Department of Nutrition and Food  Sciences at the University.  طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 931 FAO/WHO Expert Meeting on the Application of Nanotechnologies in the Food and Agriculture Sectors: Potential Food Safety Implications: meeting report The advent of nanotechnology has unleashed enormous prospects for the development of new products and  applications for a wide range of industrial and consumer sectors. Many countries have identified its potential in the  food and agriculture sectors and are investing significantly in its applications to food production. However, owing to  our limited knowledge of the human health effects of these applications, many countries recognize the need for early  consideration of the food safety implications of the technology. In response to this request, FAO and WHO convened  an Expert Meeting on the topic in order to identify further work that may be required to address the issue at a global  level. Seventeen experts from relevant disciplines, such as food technology, toxicology and communication, met to  discuss three main areas: the use of nanotechnology in food production and processing; the potential human health  risks associated with this use; the elements of transparent and constructive dialogues on nanotechnology among  stakeholders. The above-mentioned publication reports the outcome of the meeting and is available online at: http://whqlibdoc.who. int/publications/2010/9789241563932_eng.pdf   Sagun E et al. [The presence and prevalence of 17. Listeria species in milk and herby cheese in and around Van]. Turkish Journal of Veterinary and Animal Sciences, 2001, 25:15–19 [in Turkish]. Soyutemiz E et al. Presence of 18. Listeria monocytogenes in raw milk samples from West Anatolia. Turkish Journal of Infection, 2001, 15(1):5–9. Erol I et al. 19. Occurrence and contamination levels of Listeria spp. in milk and dairy products in Ankara. Paper presented at the FEMS Symposium on the versatility of Listeria species, 10–11 October, Izmir, Turkey, 2002. Osman A et al. 20. Listeria spp. in the raw milk and dairy products in Antakya, Turkey. Food Control, 2005, 17:676–679. Gulmez M et al. [Investigation of 21. Campylobacter, Salmonella and Listeria spp. from Turkish white and cecil cheese]. Kafkas Univer- sitesi Veteriner Fakultesi Dergisi, 2001, 7(2):155–161 [in Turkish]. Rudolf M et al. High incidence of 22. Listeria monocytogenes in European red smear cheese. International Journal of Food Microbiology, 2001, 63:91–98. Cordano AM et al. Occurrence of 23. Listeria monocytogenes in food in Chile, International Journal of Food Microbiology, 2001, 70:175–178. Pak SI et al. Risk factors for 24. Listeria monocytogenes contamina- tion of dairy products in Switzerland, 1990–1999. Preventive Veterinary Medicine, 2002, 53(1–2):55–65. Risk assessment of 25. Listeria monocytogenes in ready-to-eat foods. Rome, Food and Agriculture Organization/Geneva, World Health Organization, 2004 (Microbiological Risk Assessment Series 5). Sergelidis D et al Temperature distribution and prevalence 26. of Listeria spp. in domestic, retail and industrial refrigerators in Greece. International Journal of Food Microbiology, 1997, 34:171–177. Griffiths MW. 27. Listeria monocytogenes: its importance in dairy industry. Journal of the Science of Food and Agriculture, 1989, 47:133–158. Dairy science and technology education series.28. University of Guelph, Ontario [website] (http://www.foodsci.uoguelph.ca/ dairyedu/home.html, accessed 2 June 2010). Jolley S. Clotted cream. In: Robert M, Robinson RK, Sadler MJ. 29. Encyclopedia of Food Science, Food Technology and Nutrition. New York, Academic Press, 1993:1692–1697. EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 932 Ineffectiveness of myrrh-derivative Mirazid against schistosomiasis and fascioliasis in humans M.M. Osman,1 H.A. El-Taweel,1 A.Y. Shehab 1 and H.F. Farag 1 ABSTRACT The present study assessed the schistosomicidal and fasciolicidal actions of the myrrh-derivative Mirazid® in an area of low schistosomiasis transmission. A total of 27 patients infected with Schistosoma mansoni and 16 with Fasciola spp. received the maximum recommended dose of Mirazid. Pretreatment egg counts in 4 Kato–Katz slides were compared with similar counts in stool samples collected 1 and 2 months after treatment. Standard procedures and quality control measures were followed. The results revealed that Mirazid used as schistosomicidal or fasciolicidal agent in the maximum recommended dose has a low cure rate and produced a negligible reduction in egg counts. Prescribing such an ineffective drug in Egypt might endanger the achievements of the schistosomiasis control strategy. 1Department of Parasitology, Medical Research Institute, University of Alexandria, Alexandria, Egypt (Correspondence to H.A. El-Taweel: Hend_omn@ hotmail.com). Received: 01/02/09; accepted: 23/03/09 ناسنلإا في تاق ِّروتلما ءادو تايسراهلبلا ءاد دض رلما تاقتشم ةيلاعف جرف يمهف ىده ،باهش فسوي لمأ ،ليوطلا ليع دنه ،نماثع ىفطصم تفيرم ةيا ِرس اهيف ضفخنت ةقطنم في تاق ِّروتلماو تايسراهلبلا لىع )ديزايرلما وه يراجتلا اهمساو( رلما تاقتشلم داضلما يرثأتلا ةساردلا هذه مِّيقت :ةصلالخا دادعت نروقو .ديزايرلما نم ابه صىوم ةعرج صىقأ تاقروتلما عاونأب ًاباصم ًاضيرم 16و ةينوسنلما ايسراهلبلاب ًاباصم ًاضيرم 27 ىقلت دقو .تايسراهلبلا يعورو .جلاعلا ءاطعإ نم نيرهشو رهش دعب تَذخُأ ةيزارب تانيع في اهدادعت عم Kato-Katz زتاك-وتاك حئاشر عبرأ في جلاعلا ءاطعإ لبق ضويبلا تاق ِّروتلماو تايسراهلبلا جلاعل ابه صىوم ةعرج لىعأب مدختسلما ديزايرلما نأ جئاتنلا ترهظأ دقو .ةدولجا ةبقارم يربادتو ةيرايعلما تاءارجلإا عابِّتا دق صرم في لاعفلا يرغ ءاودلا اذه فصو نأ لىإ ةساردلا تصلخو .ضويبلا دادْعَت في رَكْذُي ضافخنا هنع ْمُجني لمو ضفخنم يجلاع لدعم هل ناك .تايسراهلبلا ءاد ةحفاكم ةيجيتارـتسا تازاجنإ رطخلل ض ِّرعي Inefficacité du Mirazid (à base de myrrhe) dans le traitement de la schistosomiase et de la fasciolase chez l’homme RÉSUMÉ La présente étude visait à évaluer l’action schistosomicide et fasciolicide du Mirazid®, à base de myrrhe, dans une zone de faible transmission de la schistosomiase. Au total, 27 patients infectés par Schistosoma mansoni et 16 patients infectés par Fasciola spp. ont reçu la dose maximale recommandée de Mirazid. Le nombre d’œufs comptés avant traitement sur quatre lames Kato–Katz a été comparé au nombre d’œufs identifié dans les échantillons de selles recueillis un mois et deux mois après le traitement. Les résultats ont révélé que le Mirazid, employé en tant qu’agent schistosomicide ou fasciolicide à la dose maximale recommandée, est associé à un faible taux de guérison et induit une diminution négligeable du nombre d’œufs. La prescription d’un médicament aussi inefficace en Égypte peut compromettre les avancées de la stratégie de lutte contre la schistosomiase. طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 933 Introduction An anti-schistosomal action of myrrh,  derived from a natural botanical source,  was  first  described  by Massoud  and  colleagues  in 1996 [1]. They  reported  that  a  combination of  resin, obtained  from the plant by alcohol extraction and  volatile oil, separated by water and steam  distillation, produced complete elimina- tion of worms in hamsters infected with Schistosoma mansoni.  Subsequently,  a  combination of  resin  and  volatile  oil  was  used  to  treat  62  schistosomiasis  patients  in  a dose of 11.5 mg/kg/day  for 3 days. The number of eggs in stool  samples decreased  from 250 eggs per  gram (epg) before  treatment  to 3 epg  8 weeks post treatment [2]. In another  study,  204  schistosomiasis  patients  treated with myrrh in a dose of 10 mg/ kg/day for 3 days showed an initial cure  rate of 91.7%. Re-treatment of uncured  cases with a second dose of 10 mg/kg / day for 6 days gave a cure rate of 76.5%,  increasing the overall cure rate to 98.1%  [3]. Massoud et  al.  also  studied  the ef- ficacy of myrrh  in  the  treatment of  7  cases of human  fascioliasis;  eggs were  not detected in the stool 3 weeks post- treatment and remained so throughout  a  follow-up  period  of  3 months  [4].  Further clinical  trials and experimental  animal studies conducted  in Egypt ap- proved  these  anti-helminthic  actions  of myrrh [5–9]. The Egyptian Ministry  of Health has licensed it as a treatment  for human schistosomiasis and  fascio- liasis.  In 2001, a pharmaceutical-grade  myrrh product known as Mirazid® has  been  launched  for clinical use. Further  human  studies  have  emphasized  the  efficacy  of  the  commercial  form  of  myrrh [10–15]. Successful advertising  campaigns encouraged its use by physi- cians  and private  clinicians. Being of  natural origin without overt side-effects,  it gained great acceptance by the villag- ers who often preferred  it  over other  effective anti-parasitic drugs.  In contrast to the previous findings,  one experimental  study  failed  to dem- onstrate a satisfactory anti-schistosomal  action of different derivatives of myrrh,  including  the commercial preparation,  tested at different doses against many  strains of S. mansoni [16]. Furthermore,  2  randomized  controlled  clinical  tri- als  published  in  2005  demonstrated  negligible cure rates of Mirazid in schis- tosomiasis mansoni [17,18].  In  view  of  the  great  reliance  on  chemotherapy as the mainstay of schis- tosomiasis control strategy in Egypt, the  present study assessed cure rates of Mi- razid in human cases of schistosomiasis  and fascioliasis. Methods Study subjects The study  included 31 schistosomiasis  and 18  fascioliasis cases  identified  in a  survey conducted  in 2008  in Abis,  an  Egyptian rural area with a  low schisto- somiasis transmission rate [19]. The age  of patients ranged from 15 to 30 years;  53% of them were males.  Pretreatment stool examination Initial diagnosis of  infection was based  on egg detection by examination of 2  Kato–Katz  slides  (of  41.7 mg  each)  prepared  from  a  single  stool  sample  [20]. On the 2 days preceding  the day  of treatment, patients were asked to pro- vide  stool  samples. Eggs  in 4  slides (2  slides from each sample) were counted  and egg counts in stools were recorded.  Patients who were negative for parasites  on these days were excluded. Treatment Patients and parents of  infected chil- dren were  informed about the nature  of Mirazid  and  the  study  protocol.  After getting  their consent  to partici- pate, Mirazid capsules were given  in  a dose of 600 mg  for 6  consecutive  days,  regardless of  the patient’s  age  or weight,  as  recommended by  the  manufacturer. The  drug  was  given  on  an  empty  stomach  and patients  continued  fasting  for  1  hour  post- treatment. Side-effects were recorded  if reported by the patient. Cure evaluation Follow-up  was  performed  1  and  2  months  post-Mirazid  treatment.  For  each  subject,  2  stool  samples  were  collected on 2 consecutive days and 2  Kato–Katz  slides were  examined per  sample. Schistosoma  spp.  and Fasciola spp. eggs were counted when detected.  Viability of Schistosoma  spp.  eggs was  assessed in the second follow-up period.  Fresh  stool  samples were  suspended  in  saline,  sieved  and  left  to  sediment.  Dechlorinated water was added  to  the  sediment and  left  for 1 h at  room tem- perature. The supernatant was exposed  to light and examined with a hand lens  for  hatched miracidia. The  sediment  was  examined  microscopically  for  hatching miracidia,  viable miracidia  within unhatched eggs and  for empty  shells [21]. Quality control All  steps of  slide preparation  followed  standard procedures under the authors’  supervision. Slides were  examined by  well-trained  technicians  and  the  re- corded egg counts were  checked and  revised by the authors. Patients received  Mirazid capsules under the supervision  of one of the investigators. Statistical analysis S. mansoni and Fasciola  spp. egg counts  were expressed  in epg using  the arith- metic mean  of  egg  counts  obtained  from 4 slides, multiplied by 24 [20]. The  geometric mean egg count  (GMEC)  was calculated in each follow-up period  and compared with  the pretreatment  GMEC  to  calculate  the  percentage  change  in  egg  counts  using  the  for- mula (1 – [GMEC/g after  treatment/ GMEC/g before  treatment]) ×  100.  Statistical  comparison  of  mean  egg  EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 934 counts prior to and post-treatment were  evaluated by paired sample t-test on egg  counts. Results The day before Mirazid  treatment,  4  schistosomiasis  and 2  fascioliasis  pa- tients were found to be parasitologically  negative and were not  included  in  the  study. Treated patients  tolerated  the  drug well with no recorded side-effects.  Among  the  schistosomiasis  patients,  only 4 out of  27  cases  (14.8%) were  cured 1 month post-treatment and 3 of  them re-excreted Schistosoma spp. eggs 1  month later, decreasing the cure rate to  just 1 case (3.7%).  All  fascioliasis  patients  continued  passing eggs in the 2 follow-up periods  (Table  1).  Among  uncured  schisto- somiasis  patients, GMEC  rose  from  63.5  [standard deviation (SD) 163.5]  pre-treatment  to  74.2  (SD  216.2)  1  month  post-treatment,  a  16.8%  rise.  One month  later,  a 3.5%  reduction  in  GMEC was detected as compared with  the  pretreatment  count.  Statistically,  these  changes  in GMEC were  non- significant. Testing  for  viability  in  the  second  follow-up period  revealed  that  all uncured subjects were passing viable  ova.  In  fascioliasis  cases,  the  GMEC  showed  29.4%  reduction  in  the  first  follow-up period (P > 0.05) and only  0.1%  reduction  in  the  second period  (Table 2). Discussion According  to Southgate,  controversy  surrounds  the  schistosomicidal  action  of Mirazid  [22].  In  the present  study,  schistosomiasis cure  rates evaluated 1  and 2 months after Mirazid  treatment  were only 14.8% and 3.7% respectively.  Furthermore, a negligible  reduction  in  GMEC was observed among uncured  individuals  assessed  2  months  after  receiving Mirazid. This  denotes  the  very low efficacy of the drug in schisto- somiasis. On  the other hand, Mirazid  was completely  ineffective  in patients  infected with Fasciola spp., who showed  only an  initial,  transient nonsignificant  reduction in GMEC.  When  drug  efficacy  is  evaluated  with  insensitive  parasitological  tech- niques,  the  cure  rate  is  likely  to  be  overestimated and the number of eggs  passed  in  the  faeces  underestimated  [23]. The quality of slide preparation, as  well as technicians’ expertise, influences  the accuracy of the parasitology results.  This  is  especially  important  in  light of  infections that can be easily missed even  if more  than one slide  is examined.  In  the present  study, we maximized  the  sensitivity of  the Kato–Katz  technique  by  examination  of  4  slides  prepared  from stool samples collected on 2 con- secutive days [24,25]. Technicians were  of high professional calibre and quality  control measures were followed in both  slide preparation  and examination  to  ensure credibility and enhance the reli- ability of the parasitological diagnosis. In the present study, 4 schistosomia- sis and 2  fascioliasis cases  identified  in  the initial survey were found to be para- sitologically negative when re-examined  1 day before  treatment although  they  did not  receive any antiparasitic drugs.  This may be due to spontaneous cure or,  in the case of fascioliasis, false infection.  Absence of eggs was reported in 26% of  schistosomiasis patients  re-examined  within 5–6 weeks [17]. Inclusion of these  cases would result in over-estimation of  Table 1 Schistosomiasis and fascioliasis cure rates 1 and 2 months after treatment with Mirazid Type of infection After 1 month After 2 months No. cured % cured No. cured % cured Schistosomiasis cases (n = 27) 4 14.8 1 3.7 Fascioliasis cases (n = 16) 0 0.0 0 0.0 Table 2 Eggs counts in stool samples of uncured schistosomiasis and fascioliasis patients 1 and 2 months after treatment with Mirazid Type of infection/time after treatment No. of cases Egg count: % change t-valuea P-value Range (epg) GMEC (SD) Schistosomiasis cases Before treatment 27 6–708 63.5 (163.5) – – – After 1 month 23 6–906 74.2 (216.2) +16.8 0.127 0.900 After 2 months 26 6–768 61.3 (196.4) –3.5 0.954 0.354 Fascioliasis cases Before treatment 16 12–606 77.9 (171.9) – – – After 1 month 16 12–180 55.0 (49.4) –29.4 1.187 0.254 After 2 months 16 12–420 77.8 (121.9) –0.1 0.004 0.997 aVersus pre-treatment GMEC. egp = eggs per gram; GMEC = geometric mean egg count; SD = standard deviation. طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 935 the cure rate, especially in studies involv- ing longer follow-up periods.  It  is unlikely  that variation  in strain  susceptibility among different  localities  can account for the discrepancy among  studies. Reports from rural areas of near- by governorates (El-Fayoum and Giza)  showed discrepant results [11,17]. Authors  recommending  the use of  Mirazid have  reported  that  its chemo- therapeutic action was  less evident on  immature worms  [14]. Therefore,  in  areas where  individuals are  simultane- ously  infected with adult  schistosomes  and  the  juvenile  stages of  the parasite,  parasitological  cure  rates may be un- derestimated  to  some extent. New  in- fection  following drug  intake may also  result in apparently low response to the  drug. Yet it is unlikely that these are the  only explanations  for  the observed  low  cure  rate,  since  the  transmission  rate  is not high  in  the study area  in view of  the low prevalence and light intensity of  infection [19]. On  the other hand,  improper  ad- ministration  of  the  drug,  inadequate  doses and/or poor patient compliance  may result in treatment failure. The low  schistosomiasis  cure  rates  (9.1% and  8.9%)  reported by Botros  et  al. were  attributed  to employing a  suboptimal  dose  of Mirazid  (300 mg/day  for  3  days)  [17].  In  the  present  study, we  used the maximum recommended dose  of Mirazid (600 mg/day for 6 days) and  proper intake of the full drug course was  assured by strict supervision of capsule  administration. Vomiting was not  re- ported among  the  treated cases. Nev- ertheless, we did not detect satisfactory  cure or a significant change in GMEC of  uncured patients, and eggs were still vi- able after treatment. This observed weak  anti-schistosomal  activity  of Mirazid  was supported by Barakat et al. [18]. In  conclusion,  the  present  study  did not  reveal  any  fasciolicidal  activ- ity  of  myrrh  and  demonstrated  its  very week  anti-schistosomal  action.  The use of  such an  ineffective drug  is  likely  to  increase patients’ morbidity  and might endanger the achievements  of  the schistosomiasis control strategy  in Egypt. Acknowledgements This work was supported by the Medi- cal Research Institute, Alexandria Uni- versity. Egypt. The authors are grateful  to Dr Omima Gaber Yassein, lecturer in  medical biostatistics, Medical Research  Institute, Alexandria University  for  sta- tistical analysis.  Massoud AM, Galal MM, Bennett JL. Experimental studies 1. demonstrating the anti-schistosomal activity of myrrh, cam- miphora molmol. Proceedings of the 45th Annual Meeting of the American Society of Tropical Medicine and Hygiene, 1996, Baltimore, Maryland. Supplement to American Journal of Tropi- cal Medicine and Hygiene, 55(2 suppl.):233–234. Massoud AM, Salama O, Bennett JL. Efficacy, toxicity and side 2. effects of a new antischistosomal drug derived from myrrh. Proceedings of the 46th Annual Meeting of the American Society of tropical Medicine and Hygiene.1997, Florida. Sup- plement to American Journal of Tropical Medicine and Hygiene, 57(3 Suppl.):244. Sheir Z et al. A safe, effective, herbal antischistosomal therapy 3. derived from myrrh. American Journal of Tropical Medicine and Hygiene, 2001, 65:700–704. Massoud A et al. Preliminary study of therapeutic efficacy of 4. a new fasciolicidal drug derived from Commiphora molmol (myrrh). American Journal of Tropical Medicine and Hygiene, 2001, 65:96–99. Badria F et al. Mirazid: a new schistosomicidal drug. 5. Pharma- ceutical Biology, 2001, 39:127–131. Gaballah M et al. Control of schistosomiasis in a rural area using 6. a new safe effective herbal treatment. Journal of Environmental Science, 2001, 21:63–84. Motawea SM et al. Control of 7. Fasciola in an Egyptian endemic rural area by a safe, effective fasciolicidal herbal drug. Journal of Environmental Science, 2001, 21:85–104. Hamed MA, Hetta MH. Efficacy of 8. Citrus reticulata and Mirazid in treatment of Schistosoma mansoni. Memórias do Instituto Oswaldo Cruz, 2005, 100:771–778 Mahmoud M S, Abou-El Dobal S, Soliman K. Immune Re-9. sponse in Fasciola gigantica experimentally infected rabbits treated with either carnosine or Mirazid®. Research Journal of Parasitology, 2008, 3:40–49 References El Baz MA et al. Clinical and parasitological studies on the ef-10. ficacy of Mirazid in treatment of schistosomiasis haematobium in Tatoon, Etsa Center, El Fayoum Governorate. Journal of the Egyptian Society of Parasitology, 2003, 33:761–767. Abo-Madyan AA, Morsy TA, Motawea SM. Efficacy of myrrh 11. in the treatment of schistosomiasis (haematobium and man- soni) in Ezbet El-Bakly, Tamyia Center, El-Fayoum Governorate, Egypt. Journal of the Egyptian Society of Parasitology, 2004, 34:423–446. Abo-Madyan AA et al. Clinical trial of Mirazid in treatment of 12. human fascioliasis, Ezbet El-Bakly (Tamyia Center) Al-Fayoum Governorate. Journal of the Egyptian Society of Parasitology, 2004, 34:807–818. Soliman OE et al. Evaluation of myrrh (Mirazid) therapy in 13. fascioliasis and intestinal schistosomiasis in children: immuno- logical and parasitological study. Journal of the Egyptian Society of Parasitology, 2004, 34:941–966. Massoud AM et al. Evaluation of schistosomicidal activity of myrrh 14. extract: parasitological and histological study. Journal of the Egyp- tian Society of Parasitology, 2004, 34(3 Suppl.):1051–1076. El-Mathal EM, Fouad MA. Human fascioliasis among immigrant 15. workers in Saudi Arabia. Journal of the Egyptian Society of Parasi- tology, 2005, 35(3 Suppl.):1199–1207. Botros S et al. Lack of evidence for an antischistosomal activity 16. of myrrh in experimental animals. American Journal of Tropical Medicine and Hygiene, 2004, 71:206–210. Botros S et al. Efficacy of Mirazid in comparison with praziquan-17. tel in Egyptian Schistosoma mansoni-infected school children and households. American Journal of Tropical Medicine and Hygiene, 2005, 72:119–123. Barakat R, Elmorshedy H, Fenwick A. Efficacy of myrrh in the 18. treatment of human schistosomiasis mansoni. American Journal of Tropical Medicine and Hygiene, 2005, 73:365–367. EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 936 World Health Statistics 2010 World Health Statistics 2010 contains WHO’s annual compilation of data from its 193 Member States, and includes a  summary of progress towards the health-related Millennium Development Goals and targets. It provides a comprehensive summary of the current status of national health and health systems including; mortality  and burden of disease, causes of death, reported infectious diseases, health service coverage, risk factors, health systems  resources, health expenditures, inequities and demographic and socioeconomic statistics. The publication is available in 6 languages including Arabic (http://www.who.int/whosis/whostat/AR_WHS10_Full.pdf) Allam AF et al. Assessing the marginal error in diagnosis and 19. cure of Schistosoma mansoni in areas of low endemicity using Percoll and PCR techniques. Tropical Medicine and Interna- tional Health, 2009, 14(3):316–321. Katz N, Chaves A, Pellegrino J. A simple device for quantita-20. tive stool thick smear technique in schistosomiasis mansoni. Revista do Instituto de Medicina Tropical de São Paulo, 1972, 14:397–400. Additional techniques for stool examination. In: Garcia LS, 21. ed. Diagnostic medical parasitology, 4th ed. Washington DC, American Society for Microbiology Press, 2001:786–801. Southgate VR et al. Towards control of schistosomiasis in sub-22. Saharan Africa. Journal of Helminthology, 2005, 79:181–185. Utzinger J et al. Efficacy of praziquantel against 23. Schistosoma mansoni with particular consideration for intensity of in- fection. Tropical Medicine and International Health, 2000, 5:771–778. Utzinger J et al. Relative contribution of day-to-day and intra-24. specimen variation in faecal egg counts of Schistosoma man- soni before and after treatment with praziquantel. Parasitology, 2001, 122:537–544. El-Morshedy H et al. Intra-specimen and day-to-day variations 25. of Fasciola egg counts in human stools. Eastern Mediterranean Health Journal, 2002, 8:619–625. طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 937 Retrospective survey of human hydatidosis in Bahir Dar, north-western Ethiopia N. Kebede,1 A. Mitiku2 and G. Tilahun1 ABSTRACT To determine the prevalence of human hydatidosis in Bahir Dar and the sex and age distribution and awareness, we used a questionnaire and retrospective analysis covering 5 years of the case reports at public health institutions in Bahir Dar town and the surrounding areas. We interviewed 244 individuals: 150 households, 32 abattoir workers, 30 butchers and 32 health professionals. None of the householders or butchers knew about cystic echinococcosis, nor did 25% of the health professionals; however, 100% of the abattoir workers and 75% of the health professionals recognized it. The householders owned on average 3 livestock and 1 dog and during holidays, ceremonies and other feasts, the community mostly practised backyard slaughter. Case book analysis between January 2002 and December 2006 showed that out of the total of 36 402 patients admitted for ultrasound examination, 24 hydatidosis cases were registered, giving a mean annual incidence rate of approximately 2.3 cases per 100 000 per year. Ultrasound was the most frequently used diagnostic method. 1Aklilu Lemma Institute of Pathobiology, Addis Ababa University, Addis Ababa, Ethiopia (Correspondence to N. Kebede: knigatu2003@yahoo. com, nigatukebede@yahoo.com). 2Amhara National Regional State Bureau of Agriculture and Rural Development, Bahir Dar Regional Veterinary Laboratory, Bahir Dar, Ethiopia. Received: 05/11/08; accepted: 19/02/09 ايبويثأ برغ لماش في راد رحب ةطقنم في يشربلا تاكِوْشُمـلا ءادل يداَِعتْسا حسم نوهلايت ويشاتيغ ،وكيتيم يبيبآ ،يديبيك وتاغين نوثحابلا مدختسا ،ةيعوتلا ىوتسمو ،رمعلاو سنلجا بسحب ِهِع ُّزوتو راد رحب ةقطنم في يشربلا تاكِوْشُمـلا ءاد راشتنا لّدعم ديدتح َةيغُب :ةصلالخا نوثحابلا ىرجأو .اله ةرواجلما قطانلماو راد رحب ةدلب في ةيمومعلا ةحصلا قفارم في تلاالحا نع غيلبتلا نم تاونس 5 ى َّطغ ًايداعتسا ًلايلتحو ًانايبتسا دارفأ نم ٍيأ ىدل ْنكت لمو .ينّـ يحصلا ينلماعلا نم 32و ،ًارازج 30و ،خلاسلما لماع نم 32و ،تلائاعلا دارفأ نم 150 :ًادرف 244 عم تلاباقم نم %75و خلاسلما لماع نم %100 نأ لاإ ؛ًائيش ينيحصلا ينلماعلا نم %25 هنع فرع لاو ،سييكلا تاكوشلما ءادب ةفرعم ةّيأ نيرازلجا وأ تلائاعلا دايعلأا ءانثأ ّلحلما عمتجلما دارفأ ةيبلاغ سراتمو ،ًادحاو ًابلكو ةيشالما نم ةثلاث طسوتلما في تلائاعلا كلتتمو .ءادلا لىع اوف َّرعت ينيحصلا ينلماعلا /لولأا نوناك ىتح 2002 رياني/نياثلا نوناك نم ةلجسلما تلاالحا لجس ليلتح رهظأ دقو .لزنلما ءانف في حبذلا ةيلمع تازاجلإاو تلاافتحلااو طسوتم غلبي كلذبو ،ةيتوصلا َقوف تاجولماب صحفلل اولخدُأ ًاضيرم 36402 لياجمإ نم ،تاكوشلما ءاد نم ةلاح 24 تلجس دق هنأ 2006 برمسيد .ًامادختسا صيخشتلا لئاسو رثكأ وه ةيتوصلا َقوف تاجولماب صحفلا ناكو ،ًايونس فلأ ةئم لكل ةلاح 2.3 لياوح يونسلا عوقولا لدعم Étude rétrospective sur l’hydatidose humaine à Bahir Dar, au nord-ouest de l’Éthiopie RÉSUMÉ Afin de déterminer la prévalence de l’hydatidose humaine à Bahir Dar, sa répartition en fonction du sexe et de l’âge, ainsi que le degré de connaissance de la maladie, nous avons utilisé un questionnaire et une analyse rétrospective portant sur cinq années de rapports de cas recensés dans cette ville et ses environs. Nous avons interrogé 244 personnes : 150 chefs de famille, 32 ouvriers d’abattoir, 30 bouchers et 32 professionnels de santé. Aucun chef de famille, ni aucun boucher n’avait entendu parler de l’échinococcose kystique, mais 100 % des ouvriers d’abattoir et 75 % des professionnels de santé la connaissaient. Les chefs de famille possédaient en moyenne trois animaux d’élevage et un chien. Durant les jours de congés, les cérémonies et autres fêtes, la plupart des membres de la communauté pratiquaient l’abattage domestique. L’analyse des registres entre janvier 2002 et décembre 2006 a révélé que, sur un total de 36 402 patients admis pour une échographie, 24 cas d’hydatidose avaient été enregistrés, portant le taux d’incidence annuelle moyen à environ 2,3 cas pour 100 000. L’échographie était la méthode de diagnostic la plus fréquemment employée. EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 938 Introduction Cystic  echinococcosis  (CE)  (hydati- dosis),  caused  by  the  larval  stage  of  Echinococcus granulosus,  is recognized as  being one of the world’s major zoonoses  [1,2]. The distribution of hydatidosis  is  normally  associated with underdevel- oped countries, especially in rural com- munities where humans maintain close  contact with dogs and various domestic  animals, which may act as intermediate  hosts [2].  Dogs  are  the  primary  definitive  hosts for the parasite, with livestock act- ing as  intermediate hosts and humans  as aberrant intermediate hosts. Humans  become  infected  by  the  ingestion  of  eggs passed  in  the  faeces of dogs  [3].  The outcome of  infection  in humans  is  hydatid  cyst  development  in  the  lung, liver or other organs [4]. This caus- es severe disease and death  in humans  and results in economic loss from treat- ment  costs,  lost wages  and  livestock- associated production losses [4].  E. granulosus has a cosmopolitan dis- tribution [5] and CE is a public health  problem  in Asia,  the Mediterranean,  South America and Africa. Previous re- ports describe the endemic occurrence  of human cases of CE in North African  countries bordering the Mediterranean  [6] and in sub-Saharan Africa [7].  In Ethiopia,  the  situation of CE  in  humans  is not well  documented  and  explored.  In southern parts of Ethiopia  prevalence rates of 1.6% and 0.5% have  been  reported  [5]. Hence,  the  aim of  the current study was to determine the  prevalence,  sex  and  age  distribution  and awareness  information of human  hydatidosis in Bahir Dar. Methods Study area The study was conducted in Bahir Dar  town  in  Amhara  National  Regional  State  in north-western Ethiopia. Bahir  Dar  is 565 km  from Addis Ababa and  has 1 public hospital,  1 health  centre  and 3 higher private clinics. The popula- tion of the town is about 205 000 [8].  Study design A questionnaire and retrospective hos- pital/clinical data survey were the tools  used in the study. A structured question- naire was prepared in the local language  and pilot tested on 10 individuals from  the same population to assess perform- ance. The questionnaire gathered infor- mation on  the extent of  awareness of  CE, control measures taken and related  factors. The participants were  selected  based on their occupation using simple  random sampling. All 32 meat  inspec- tors were  selected,  plus  50%  (30) of  the  butchers,  50%  (total  32)  of  the  health professionals per  institute  and  10% (150) of the householders. We col- lected retrospective data on ultrasound  examinations which had been carried  out  for different purposes during  the  period January 2002–December 2006  from Felegehiowt Hospital and Kidane  Mehiret, Gambi and Eyerusalem health  services. Data on human hydatid  in- fection were  identified,  and  the mean  annual incidence of human hydatidosis  was determined.  Data analysis Percentages prevalence was determined  and  analysis  of  variance  (ANOVA)  carried out  to measure  the differences  by  age-group  and  the odds  ratios  for  differences according  to  sex.  In all  the  analyses, confidence level was 95% and  significance level was 5%. Ethical approval The  study  was  conducted  after  the  objectives were explained to those com- pleting  the questionnaire  and written  agreement was obtained from them.  Results Questionnaire survey None of  the households  interviewed  in the study area knew about CE. They  owned on  average  3  livestock  and  1  dog. The animals were kept for draught  power and as a  source of  income. The  animals  grazed  communally  where  livestock  and  humans  had  frequent  contact with  dogs  and  their  excreta.  During holidays, ceremonies and other  feasts, the community mostly practised  homestead/backyard  slaughter;  dur- ing  slaughter, meat  inspection  is  not  conducted and the offal is often given to  pets or disposed of  in  the backyard.  In  10.7% of the households, the dogs were  treated with  traditional  preparations  at  their puppy  stage only; other  than  this,  treatment with commercial drugs  was not practised at all. The dogs were  kept  for  guarding  the homestead and  hunting. All (100%) of the 32 abattoir workers  we interviewed in Bahir Dar recognized  hydatid cyst and  the organs  frequently  affected without knowing  the name of  the disease. When an organ affected by  hydatid  cyst was  encountered,  it was  totally condemned.  Only 75% of  the 32 health profes- sionals working in public health institu- tions were aware of CE.  Among  the 60 butchers present  in  Bahir Dar  town, 30 were  interviewed;  none (0%) of whom were aware of the  disease. Ultrasound was the most frequently  used diagnostic method  and up  to 2  cases were encountered per year. How- ever, for various reasons the profession- als had never tried to educate people or  create awareness about the disease. Retrospective hospital and clinical case-book survey From a  total of 36 402 patients admit- ted  for ultrasound examination, about  24  human CE  cases were  registered  between  January 2002 and December  2006, 8 were  from Felegehiwot Hos- pital,  12  from Gambi Clinic,  3  from  Kidanemeheret Clinic and 1 from Eye- rusalem Clinic. Gambi Clinic had  the  highest rate for hydatidosis cases among  the 19 160 patients admitted (52.6% of  طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 939 the total examined) with 12 cases, about  50% of  the  total  recorded  in  this study  (Table 1).  Age and  sex groups of human hy- datid  cases  are described  in Tables 2  and 3. Females were much more  likely  to be infected than males, and the high- est  prevalence was  in  the  age  group  21–30 years. Hydatid  cyst  was  encountered  most  commonly  in  the  liver 79.2%  (19 of the 24 cases registered during  ultrasound examination) and less fre- quently in the spleen 20.8% (5 cases).  Only 2 patients had  surgical  inter- vention and none  received chemo- therapy. During the last 5 observation  years, major surgical operations were  only performed at Felegehiwot Hos- pital.  According  to  the  surgical  case  recording book at  the hospital, 3203  patients  (2229  males  and  974  fe- males) with a variety of health prob- lems underwent surgical intervention.  Only 2 of these patients (0.06%, both  females) were  recorded  as  hydatid  cyst cases. Discussion The culture and tradition of inhabitants  in  the study area  favour  the keeping of  dogs in periurban areas (urban and rural  households) often  in close association  with  the  family  and  farm animals. Al- most all  cattle owners,  shepherds  and  urban dwellers  keep  at  least 1 dog  to  safeguard  their  properties  from wild  carnivores and thieves. These and other  socioeconomic  realities  in  the  areas  studied are considered to be conducive  to  the maintenance and  further propa- gation of hydatid disease.  CE  is not  currently  a well  known  medical condition, nor it considered to  be of public health importance; it is not  a notifiable disease and there are limited  surveillance data available in humans in  the country. However,  in  the  residents  in  the study area,  the annual  incidence  rate was  approximately  2.3  cases per  100 000 per year,  calculated based on  the  catchments  of  the  hospital.  In  2  community-based  surveys  conducted  in southern Ethiopia in 1987 and 1996,  1997 and 3224 participants respectively  Table 1 Total number of admitted patients for ultrasound and registered hydatid (cystic echinicoccosus) cases in the public hospital and private clinics in Bahir Dar, January 2002–December 2006 Hospital/clinic & patient type No. patients admitted 2002 2003 2004 2005 2006 Total % Felegehiwot Hospital Admitted 624 2 687 2 822 680 793 7 606 20.9 Cases – 4 4 – – 8 33.3 Gambi Clinic Admitted – 2 803 3 780 8 919 3 658 19 160 52.6 Cases 2 2 6 2 12 50.0 Kidanemehiret C linic Admitted – – 712 3 637 1 441 5 790 15.9 Cases – – – 3 – 3 12.5 Eyerusalem Clinic Admitted – – 875 2 280 763 3 846 10.6 Cases – – – 1 – 1 4.2 Total Admitted 624 5 490 8 189 15 444 6 655 36 402 100.0 Cases 0 6 6 10 2 24 100.0 Prevalence (%) 0 0.11 0.07 0.06 0.03 0.066 0.066a a0.066 = 2.3 cases per 100 000 population of the study area Table 2 Age distribution of human hydatid cases in public health institutions in Bahir Dar, January 2002–December 2006 Age (years) Total examined Infected No. % < 10 6 524 – 11—20 7 332 2 0.03 21—30 6 898 13 0.19 31—40 6 912 5 0.07 > 40 8 736 4 0.05 Total 36 402 24 0.07 F = 5.540, degrees of freedom = 4, P = 0.001. EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 940 were screened with ultrasound: preva- lence  rates were 1.6% (31 cases)  and  0.5% (16 cases)  respectively  [5]. The  average annual incidence in humans per  100 000 has been reported in a number  of  studies,  e.g. 2.2  in Portugal  [9], 3.3  in Bulgaria  [10], 3.4  in Greece  [11],  3.6–15.8 in Morocco [12], and 3.4–4.6  in Algeria  [13]. Our  findings 2.3 per  100 000 per year, was comparable with  these studies, and showed that CE  is a  zoonosis of  importance  in Bahir Dar.  This may be because of  the  low public  awareness, backyard slaughtering prac- tices, poor control measures,  the pres- ence of a large number of stray dogs and  poor  infrastructure  that contributed to  human infection in this area.  In  the  study  area,  the  number  of  patients  admitted and  the number of  hydatid cases were unexpectedly high  at Gambi private  clinic. This may be  related to the preference of patients for  that clinic due to the relatively superior  disease diagnosis. The finding that all hy- datid cases were in the age groups over  10 years implies that both young people  and adults are  susceptible  to  infection.  A  higher  incidence  rate  was  seen  in  females than males; this may be due to  the  fact  that  females had more contact  time with dogs  than males around  the  home. Retrospective  hospital  and  clinic  survey data on human CE gave an esti- mate of the prevalence of hydatidosis in  the study area. The  infection  is asymp- tomatic with  an extended  incubation  period,  and  there  is  a  clinical  similar- ity with other diseases. ths study area is  lacking  in modern diagnostic  facilities,  and  there  is  inability  to  afford mod- ern  treatment by  the most  vulnerable  section  of  society. However,  despite  such  limitations,  careful  examinations  of hospital/clinic  records provided  a  useful indication of infection rates. This,  when  done  continuously  over many  years, may detect changes  in  the  trend  and incidence of infection.  It is therefore concluded that CE is  a disease of considerable  importance  for  public  health.  In  view of  the  re- sults of  this  retrospective  survey,  and  considering  the nature of  the disease,  the actual natural,  social,  cultural and  behavioural situation in Bahir Dar and  the  surrounding areas are conducive  to  the maintenance of a high  level of  infection  and  spread  of  the  disease  between  animals  and  humans.  Im- proving public awareness (education),  avoidance of  stray dogs, meat control  legislation and proper disposal of con- demned organs are  recommended  to  mitigate  the  effect of  the disease on  public health. Acknowledgements The authors would like to acknowledge  staff members of Felegehiowt Hospital  and Kidane Mehiret, Gambi and Eyeru- salem health services  in Bahir Dar, and  the abattoir workers and butchers. Financial  support  was  obtained  from the Research and Graduate Stud- ies Department of Addis Ababa Univer- sity and Amhara Regional Agricultural  Research Institute (ARARI). Table 3 Sex distribution of human hydatid cases in public health institutions in Bahir Dar, January 2002 to December 2006 Sex Total no. examined Infected Odds ratio 95% CI P-value No. % Female 19 298 20 0.10 4.435 1.516–12.978 0.02 Male 17 104 4 0.02 0.226 0.077–0.659 0.2 Total 36 402 24 0.07 CI = confidence interval. Eckert J, Deplazes P. Biological, epidemiological, and clinical 1. aspects of echinococcosis, a zoonosis of increasing concern. Clinical Microbiology Reviews, 2004, 17(1):107–135. Torgerson PR, Budke C. Echinococcosis – an international 2. public health challenge. Research in Veterinary Science, 2003, 74:191–202. Budke CM, Deplazes P, Torgerson PR. Global socioeconomic 3. impact of cystic echinococcosis. Emerging Infectious Diseases, 2006, 12(2):296–303. Jenkins DJ, Romig T, Thompson RCA. Emergence/ re-emer-4. gence of Echinococcus spp.a global update. International Journal for Parasitology, 2005, 35:1205–1219. Eckert J et al. 5. WHO/OIE manual in echinococcosis in humans and animals. Geographic distribution and prevalence. Paris, World References Health Organization and World Organisation for Animal Health, 2002:101–143. Magambo J, Njoroge E, Zeyhle E. Epidemiology and control 6. of echinococcosis in sub-Saharan Africa. Parasitology Interna- tional, 2006, 155:S193–S195. Shambesh MA et al. An extensive ultrasound and serologic 7. study to investigate the prevalence of human cystic echinococ- cosis in northern Libya. American Journal of Tropical Medicine and Hygiene, 1999, 60:462–468. Livestock resource development and animal health department 8. annual report. Bahir Dar, Ethiopia, Bureau of Agriculture and Rural Development (BoARD), 2006:1–8. Battelli G. Socioeconomic impact of CE. 9. WHO Mediterranean Zoonoses Control Centre Information Circular, 2003, 57:1020–1378. طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 941 Global Early Warning System for Major Animal Diseases, including Zoonoses (GLEWS) Disease emergence and spread do not respect geographical boundaries, and animals are often implicated as the source  of human infection. Zoonotic disease management therefore requires an integrated approach that involves different  sectors; mainly human, animal and food. Efficient early warning and forecasting of zoonotic disease trends through  functional surveillance systems is key to effective containment and control. Early intervention during a disease epidemic  often leads to better outcomes with reduced disease burden and associated economic impact. GLEWS is a joint system that builds on the added value of combining & coordinating alert mechanisms of the World  Health Organization (WHO), the Food and Agriculture Organization of the United Nations (FAO) and the World  Organisation for Animal Health (OIE), while linking networks from the international community and stakeholders, to  assist in prediction, prevention and control of zoonotic disease threats through sharing of information, epidemiological  and risk analysis. More information about GLEWS can be found at:  http://www.glews.net Todorov T, Boeva V. Epidemiology of echinococcosis in Bul-10. garia—a comparative study. Inernational Archive for Hidatidosis, 1997, 32:232–233. Economides P, Thrasou K. Echinococcosis/hydatidosis and 11. program for its control in the Mediterranean countries. Inter- national Archive for Hidatidosis, 1999, 33:63–83. El Idrissi et al. 12. Retrospective survey for surgical cases of CE in Morocco (1980–1992). Provo, Utah, United States of America, Brigham Young University, 1997:194–222. Shambesh MK. Human CE in North Africa. In: Andersen FL, 13. Ouhelli H, Kachani M, eds. Compendium on CE in Africa and in Middle Eastern countries with special reference to Morocco. Provo, Utah, United States of America, Brigham Young Univer- sity, 1997:223–244. EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 942 Causes of blindness in people aged 50 years and over: community-based versus hospital-based study S.A. Al-Akily,1,2 M.A. Bamashmus 1,2 and K.A. Al-Mohammadi 3 ABSTRACT The causes of blindness in Yemen were determined in people aged 50+ years in both a community- and hospital-based study and documented using WHO/PBL criteria. In the community sample of 707 individuals in a rural area of Taiz governorate the prevalence of bilateral blindness was 7.9% and the main causes were cataract (71.4%) and age-related macular degeneration (ARMD) (14.3%). Corneal opacities and uncorrected aphakia were rare (1 case each) and there were no cases of diabetic retinopathy. Unilateral blindness was found in 8.6% of the community sample. In a case-notes review of 1320 new patients attending an eye clinic in Sana’a, bilateral blindness was documented in 26.5% and unilateral blindness in 9.0% (main causes: cataract, glaucoma, ARMD, diabetic retinopathy, corneal opacities and trauma). 1Eye Department, Faculty of Medicine and Health Sciences, Sana’a University, Sana’a, Yemen (Correspondence to M.A. Bamashmus: bamashmus@y.net.ye). 2Ibn Al-Haitham Eye Centre, University of Science and Technology, Sana’a, Yemen. 3Eye Department, Skin and Venereal Diseases Hospital, National Leprosy Elimination Programme, Taiz, Yemen. Received: 02/01/09; accepted: 16/03/09 تايفشتسلما في ةساردلاو ةيعمتجلما ةساردلا ينب ةنراقم :رثكأ وأ ينسملخا نس في مه نلم ىمعلا بابسأ يدمحلما للها دبع دلاخ ،سومشماب للها دبع ظوفمح ،لييقعلا دحمأ حلاص كلذ قيثوت متو ،ىفشتسلما في ىرخأو ةيعمتمج ةسارد للاخ نم رثكأ وأ ةنس ينسملخا نس في مه نلم نميلا في ىمعلا بابسأ ديدتح ىرج :ةصلالخا ةيفيرلا قطانلما في نوشيعي ًاصخش 707 نم ةفّلؤم ةيعمتمج ةنيع فيو .ةيلماعلا ةحصلا ةمظنلم صربلا فعضو ىمعلا ةحفاكم جمانرب يرياعم مادختساب سكنت مث ،)%71.4( )تكاراتاكلا( ّداسلا وه كلذ في سييئرلا ببسلا ناكو %7.9 يننيعلا لاك في ىمعلاب ةباصلإا راشتنا لدعم غلب ،زعت ةظفامح نم تلااح دهاشُت لمو ،ححصلما يرغ ةسدعلا مادعنا نم ةدحاو ةلاحو ،ةينرقلا ةماتع نم ةدحاو ةلاح تدجُوو .)%14.3( رمعلاب طبترلما ءارفصلا ةعقبلا ددلجا ضىرلماب ةصالخا تاركذملل ةعجارم فيو .ةيعمتجلما ةنيعلا نم %8.6 في ةدحاو ينع في ىمعلاب ةباصلإا تفشتكاو .يركسلا ةيكبشلا للاتعا نم ىمعلاو ،مهنم %26.5 ىدل يننيعلا لاك في ىمعلاب ةباصلإا تل ِّجُس ،ًاضيرم 1320 مهددع غلبو ءاعنص في ينعلا بط تادايع ىدحإ في اولجوع نيذلا ةيكبشلا للاتعاو ،رمعلاب طبترلما ءارفصلا ةعقبلا سكنتو ،قَر َّزلاو ،)تكاراتاكلا( ّداسلا :يه ةيسيئرلا بابسلأا تناكو( %9.0 ىدل ةدحاو ينع في .)حوضرلاو ،ةينرقلا ةماتعو ،يركسلا Causes de cécité chez les personnes âgées de 50 ans et plus : étude comparant des données recueillies au sein d’une communauté et en milieu hospitalier RÉSUMÉ Au Yémen, les causes de cécité chez les sujets âgés de 50 ans et plus ont été déterminées dans le cadre d’une étude réalisée à la fois au sein d’une communauté et en milieu hospitalier. Elles ont ensuite été documentées selon les critères du Programme OMS de Prévention de la cécité. Pour l’échantillon de 707 individus issus d’une région rurale du gouvernorat de Taïz sur lequel portait l’étude communautaire, la prévalence de la cécité bilatérale était de 7,9 %, avec pour principales causes la cataracte (71,4 %) et la dégénérescence maculaire liée à l’âge (14,3 %). Les opacités cornéennes et les aphakies non corrigées se sont révélées rares (un cas pour chaque) et aucun cas de rétinopathie diabétique n’a été recensé. En outre, 8,6 % de l’échantillon communautaire présentait une cécité unilatérale. Dans le cadre de l’analyse des notes de cas qui portait sur 1 320 patients se présentant pour la première fois à une consultation ophtalmologique à Sanaa, une cécité bilatérale a été documentée chez 26,5 % des patients et une cécité unilatérale chez 9 % d’entre eux (principales causes : cataracte, glaucome, dégénérescence maculaire liée à l’âge, rétinopathie diabétique, opacités cornéennes et traumatisme). طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 943 Introduction Blindness constitutes a significant pub- lic health problem  in many countries.  The available data  suggest  that  in  the  year  2002  the  estimated  number  of  people with visual  impairment globally  was  in  excess of 161 million: 37 mil- lion were blind  and 124 million had  low  vision  [1,2] The majority  live  in  Asia and Africa. Of the 37 million blind  people in the world, it is estimated that  approximately half of cases are due  to  cataract  (47.8%), with  the  remainder  mainly due  to glaucoma (12.3%), age- related macular degeneration (ARMD)  (8.7%),  corneal opacities  (5.1%)  and  diabetic  retinopathy (4.8%) [1,2]. Un- corrected refractive error is also a cause  of  visual  impairment and  is  estimated  to  affect  another  153 million people  [2]. Uncorrected refractive errors  then  become  the main cause of  low vision  and the second cause of blindness [3].  Due  to  the  increasing populations  and rising  life expectancies blindness  is  a growing problem worldwide, causing  loss of quality of  life  to  the  individual  and an economic burden on  the  indi- vidual,  family  and  society  in  general.  Yemen is a poor country with a human  development  index  of  0.508,  human  poverty  index of 38.0,  life  expectancy  of 61.5 years and literacy rate of 54.1%.  The family health survey conducted by  the Ministry of Health in 2003 revealed  that 2.9% of  the population had some  kind  of  self-reported  handicap,  with  visual  handicap  representing  36% of  the  total  identified disabled  individuals  in  the country  [4]. To our knowledge  there  are  no  published  community- based studies on the causes of blindness  in Yemen. The  available  data  on  the  frequency  and  causes of blindness  in  Yemen were all obtained from hospital  settings  [3,5,6]. This  study  therefore  aimed to compare the causes of bilateral  and unilateral blindness in people aged  50 years and over in both a community-  and hospital-based study. Methods Sample For  the community-based  study a  to- tal of  707  individuals  aged 50+ years  were examined in a non-interventional,  cross-sectional descriptive population  study  in Hethran  subdistrict  of Taiz  governorate, which is a peri-urban area  8 km west of Taiz city. It is divided into  7 villages, which have 43 hamlets and  994 houses with a  total population of  5427  individuals, 2798 males (51.6%)  and 2629 females (48.4%). During July  to  August  2002  all  individuals  aged  50+ years living in the district and who  had  resided  there  for 6+ months were  approached  in a door-to-door  survey.  Those who refused to be examined were  coded as absentees,  then after 3 visits,  they were coded as blind or not blind  according  the  reply of his/her  family  member.  The sample  for  the hospital-based  study  was  the  records  of  1320  con- secutive new patients  aged 50+ years  attending Ibn Al-Haitham eye centre (af- filiated to the University of Science and  Technology) in the capital city, Sana’a,  from January to December 2002.  Verbal  consent was  taken  from all  eligible individuals. All records and per- sonal  information were kept confiden- tial. Both studies had approval from the  local ethics committees. Data collection For  the  community-based  study  the  individuals were given an eye examina- tion at home by  the  same consultant  ophthalmologist (K.M.).  In the hospital based-study the cases  notes of the patients were retrieved and  analysed. All patients had been exam- ined by  the  same consultant ophthal- mologist (M.B). The data collected  for both studies  included age,  sex,  chief complaint and  complete eye examination. The medical  history,  family history and ophthalmo- logical  evaluation were  reviewed  and  recorded. Visual  acuity was measured  using  the Snellen chart. Eye examina- tion  included  slitlamp,  fundoscopy,  retinoscopy and tonometry, as required.  Goldmann applanation was used in the  hospital and Schiotz  tonometer  in  the  community study. All findings were recorded according  to the World Health Organization Pro- gramme for the Prevention of Blindness  (WHO/PBL) eye examination record  (version  III)  [7]. Unilateral blindness  was defined as a  corrected visual  acu- ity 3/60 or below  in one eye. Bilateral  blindness was defined  as  a  corrected  visual acuity of 3/60 or below  in both  eyes  [8].  In most  cases  the  cause  of  blindness was a  single disorder. When  there was more  than  one  pathology  in a patient,  the WHO recommenda- tion was adhered  to: namely  the most  avoidable or preventable pathology was  chosen as the cause of blindness or, al- ternatively, the cause that led to the last  event rendering the individual sightless  [2]. This principle was also adhered  to  while classifying the causes of blindness  in unilateral cases when multiple causes  were present in one eye. After establish- ing  the diagnosis, medicines or glasses  were prescribed as appropriate.  Analysis Data were entered into Excel 2003 pro- gramme and analysed. The chi-squared  test was used  to compare  the propor- tions  in  the  community  and hospital  studies.  Results The  community  study  included  707  individuals: 256 (36.2%) males and 451  females (63.8%), a male:female ratio of  0.57:1. The hospital  study was of 1320  individuals: 806 males (61.1%) and 514  females (38.9%), a male to female ratio  of 1.57:1. The difference in the propor- tion of the sexes sampled was significant  in both groups (P < 0.001). EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 944 Table 1  shows  the  rate of bilateral  and unilateral blindness  for males and  females in the community and hospital  samples.  In  the community  study  the  prevalence of bilateral blindness (best  corrected visual acuity < 3/60 in the bet- ter eye) was 7.9% (56 patients) and the  main causes were cataract (71.4%) and  ARMD (14.3%) (Table 2). The preva- lence of unilateral blindness in the com- munity  study was 8.6% (61 patients)  (Table 1) and the  leading causes were  cataract (62.3%) and corneal opacities  (9.8%) (Table 2).  In  the hospital  study 26.5%  (350  patients) were bilaterally blind and the  main  causes  were  cataract  (47.7%),  glaucoma (11.4%), ARMD (9.7%), dia- betic  retinopathy (9.4%), uncorrected  aphakia  (6.3%) and corneal opacities  (5.4%). The rate of unilateral blindness  was 9.0% (119 patients) and the leading  causes were cataract (27.7%), glaucoma  (13.4%), trauma-related ocular compli- cations (11.8%) and corneal opacities  (11.8%). Discussion The prevalence and causes of blindness  have so  far not been published  from a  community-based study in Yemen, but  several  hospital-based  studies  about  the causes of blindness and visual  im- pairment  in  adults  and children have  been published recently [3,5,6]. Reports  showed that the prevalence of blindness  in  Yemen  was  around  1.0%–2.0  %  [6,9].  In our  community-based  study  the prevalence of blindness  in  the age  group 50+ years was found to be 7.9%. A  nationwide rapid assessment should be  carried out to estimate the prevalence of  avoidable blindness  in  the community  in all ages in Yemen. The proportion of females examined  in the community study was 63.8% ver- sus 36.2%. The difference could be ex- plained by sampling bias as more males  to females refused the examination. It is  also  likely  that  the population of  rural  areas is skewed as there is a higher rate of  internal and external migration of males  to other parts of the country or abroad.  In  contrast,  in  the hospital  study  the  proportion of females was 38.9% and of  males was 61.1%. The high male:female  ratio  in  the hospital-based  study  can  be explained by the higher rate of men  seeking medical  advice  in major cities  and  the difficulty  in access  to medical  care for women.  Throughout  the world 64% of  the  people who are blind are  females. This  may be because women live longer than  men and are at greater risk of blindness  from age-related causes—some blind- ing  conditions  such  as  trachoma and  cataract are more likely to affect women  than men—and because women and  girls do not have easy access to eye care  services [9,10].  In previous hospital-based  studies  in Yemen  the most  common  causes  of blindness,  in decreasing order, were  cataract,  glaucoma, diabetic  retinopa- thy, ARMD and corneal opacities [3,5].  Cataract was  the major cause of both  bilateral and unilateral blindness in both  our community and hospital  samples.  The  prevalence  of  cataract  bilateral  blindness  in  the  rural community area  was 71.4% versus 47.7% in the city hos- pital sample. Glaucoma was found to be the sec- ond major  cause of bilateral  (11.4%)  and  unilateral  (13.4%)  blindness  in  the hospital-based  study, while  in  the  community-based  study, where  glau- coma was defined as C/D ratio > 0.5  and  intraocular pressure > 21 mmHg,  only 1 case was  found  in  the unilateral  blindness  group. Glaucoma  patients  who  visit  hospitals may  be  detected  earlier because of better  facilities  such  as  applanation  tonometry,  visual field  assessment,  gonioscopy  and  dilated  fundoscopy.  In  the community-based  study most of  these  facilities were not  available. This  is a drawback of a com- munity-based study. ARMD is becoming more prevalent  in  the developed world due  to  the  ris- ing  life expectancy of  the populations.  ARMD is also increasing in some devel- oping countries following improvements  in  health  care.  Both  our  community  and hospital-based  studies  showed  a  surprisingly high prevalence of ARMD  as a cause of bilateral blindness (14.3%  and 9.6% respectively). This may be due  to  some  special  habits of  the  general  Yemeni population  in both sexes such  Table 1 Prevalence of bilateral and unilateral blindness in community and hospital studies by sex Type/Sex Community study Hospital study χ2-value P-value No. examined No. affected % affected No. examined No. affected % affected Bilateral blindness Male 256 19 7.4 806 210 26.1 39.9 < 0.001 Female 451 37 8.2 514 140 27.2 58.1 < 0.001 Total 707 56 7.9 1320 350 26.5 99.4 < 0.001 Unilateral blindness Male 256 25 9.8 806 77 9.6 0.01 0.917 Female 451 36 8.0 514 42 8.2 0.01 0.914 Total 707 61 8.6 1320 119 9.0 0.09 0.769 طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 945 as chewing qat  leaves, which  is usually  accompanied by cigarettes or waterpipe  smoking. Qat  contains  chemical  sub- stances  such  as  cathinone,  ampheta- mines,  esters,  citrals  ethereal oils  and  tannins [10,11]. Cigarette smokers have  3  to 5  times higher  risk of developing  ARMD than nonsmokers [12]. ARMD  is  the  leading cause of blindness  in de- veloped countries [11,13] since nearly 2  out of 3 people will have developed early  ARMD and 1  in 4 will have  lost vision  from ARMD [13,14]. Diabetic  retinopathy was not de- tected in the community-based study as  only a  few cases had diabetes mellitus,  while  in  the hospital-based  study dia- betic  retinopathy was  the  fourth cause  of bilateral blindness (9.4%).  There was  only  1  case  of  corneal  opacities  (1.8%)  in  the  community- based study while in the hospital-based  study  it  accounted  for  5.4% of  cases  and  this  can be  attributed  to  trauma,  microbial corneal ulceration and use of  traditional eye medicines.  Preventable  causes  of  blindness,  such as  trauma-related complications,  can be partially reduced at  the primary  levels of service delivery, whereas treat- able causes, such as cataract, glaucoma,  amblyopia and refractive errors, require  Table 2 Causes of bilateral and unilateral blindness in community and hospital studies Type/Cause Community study (n = 56) Hospital study (n = 350) χ2-value P-value No. % No. % Bilateral blindness Cataract 40 71.4 167 47.7 10.85 < 0.001 Glaucoma 0 0.0 40 11.4 – – Age-related macular degeneration 8 14.3 34 9.7 1.10 0.294 Corneal opacities 1 1.8 19 5.4 1.34 0.247 Diabetic retinopathy 0 0.0 33 9.4 – – Uncorrected aphakia 1 1.8 22 6.3 1.82 0.177 Optic atrophy 1 1.8 0 0.0 – – Other 5 8.9 35 10.0 0.07 0.80 Unilateral blindness Cataract 38 62.3 33 27.7 19.20 < 0.001 Glaucoma 1 1.6 16 13.4 6.10 0.014 Age-related macular degeneration 2 3.3 1 0.8 1.51 0.219 Corneal opacities 6 9.8 14 11.8 0.15 0.695 Diabetic retinopathy 0 0.0 5 4.2 – – Trauma-related ocular complications 2 3.3 14 11.8 3.33 0.068 Optic atrophy 1 1.6 1 0.8 0.23 0.629 Other 11 18.0 35 29.4 2.59 0.108 specialized ophthalmology units,  sys- tems  for early  identifications and refer- rals, as well as increased public awareness  of the possibility of treatment.  Acknowledgements We extend out thanks to all the staff of  Ibn Al-Haitham eye  centre  especially  the manager Mr Abdallah Al-Ghubary.  Also we appreciate the assistance of Dr  Babar Qureshi, Director  Academics  of PICO, Pakistan  for his  supervision  of  the community-based  study and  to  Dr Ahmed Mousa, Magrabi Hospital,  Cairo for his help in statistical analysis.  Foster A, Gilbert C, Johnson G. Changing patterns in global blind-1. ness 1988–2008. Community Eye Health, 2008, 21(67):37–39. Resnikoff S et al. Global data on visual impairment in the 2. year 2002. Bulletin of the World Health Organization, 2004, 82:844–851. References Resnikoff S et al. Global magnitude of visual impairment 3. caused by uncorrected refractive errors in 2004. Bulletin of the World Health Organization, 2008, 86(1):63–70. Family health survey 20034. . Sana’a, Yemen, Ministry of Health, 2003. EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 946 Causes of blindness and visual impairment In spite of the progress made in surgical techniques in many countries during the last 10 years, cataract (47.9%) remains  the leading cause of visual impairment in all areas of the world, except for developed countries. Other main causes of visual impairment in 2002 are glaucoma (12.3%), age-related macular degeneration (AMD)  (8.7%), corneal opacities (5.1%), diabetic retinopathy (4.8%), childhood blindness (3.9%), trachoma (3.6%), and  onchocerciasis (0.8%). The causes of avoidable visual impairment worldwide are all the above except for AMD. In the  least-developed countries, and in particular Sub-Saharan Africa, the causes of avoidable blindness are primarily, cataract  (50%), glaucoma (15%), corneal opacities (10%), trachoma (6.8%), childhood blindness (5.3%) and onchocerciasis  (4%).  Looking at the global distribution of avoidable blindness based on the population in each of the WHO regions, we see  the following: South-East Asian 28%, Western Pacific 26%, African 16.6%, Eastern Mediterranean 10%, the American  9.6%, and European 9.6%. Source: http://www.who.int/blindness/causes/en/ Al-Akily SA, Bamashmus MA. Causes of blindness among adult 5. Yemenis: a hospital-based study. Middle East Journal of Oph- thalmology, 2008, 15(1):3–6. Bamashmus M, Al-Akily S. Pattern of childhood blindness and 6. low vision in Yemen—hospital based study. Eastern Mediterra- nean Health Journal, 2010, 16(4):425–428. Coding instructions for the WHO/PBL eye examination record (ver-7. sion III). Geneva, World Health Organization, 1988 (PBL/88.1). International statistical classification of diseases and related 8. health problems, tenth revision (ICD-10). Geneva, World Health Organization, 1992. Resnikoff S et al. Global data on visual impairment in the 9. year 2002. Bulletin of the World Health Organization, 2004, 82(11):844–851. Courtright P, Lewallen S. Improving gender equity in eye care: 10. advocating for the needs of women. Community Eye Health Journal, 2007, 20(64):68–69. Feyissa AM, Kelly JP. A review of the neuropharmacological 11. properties of khat. Progress in neuro-psychopharmacology & biological psychiatry, 2008, 32(5):1147–1166. Chan D. Cigarette smoking and age-related macular degenera-12. tion. Optometry and Vision Science, 1998, 75(7):476 –484. Bamashmus M, Matlhaga B, Dutton G. Causes of blindness 13. and visual impairment in the West of Scotland. Eye, 2004, 18(3):257–261. Taylor HR, Keeffe JE. World blindness: a 21st century perspec-14. tive. British Journal of Ophthalmology, 2001, 85:261–266. طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 947 Diabetes mellitus and sensorineural hearing loss among non-elderly people M. Mozaffari,1 A. Tajik,2 N. Ariaei,1 F. Ali-Ehyaii 3 and H. Behnam 3 ABSTRACT One of the known complications of diabetes is hearing impairment. This comparative study in Tehran, Islamic Republic of Iran, aimed to evaluate the association of diabetes mellitus and sensorineural hearing loss (SNHL) among a non-elderly population. Among 160 subjects aged < 60 years with no history of occupational noise exposure (80 diabetics and 80 age- and sex-matched non-diabetic controls), 45% of diabetic patients and 20% of controls had SNHL (OR 3.5, 95% CI: 1.6–6.6). Age at onset and duration of diabetes were associated with SNHL. Diabetes mellitus may be a risk factor for hearing loss regardless of age and smoking. Determining the cause of SNHL in diabetic patients may lead to development of better treatment options. 1Department of Metabolism and Endocrinology; 3Department of Ear, Throat and Nose Surgery, Boo-Ali Hospital, Tehran Unit of Azad University, Medical School, Tehran, Islamic Republic of Iran. 2Department of Community Medicine, Faculty of Medicine, Tehran University of Medical Sciences, Tehran, Islamic Republic of Iran (Correspondence to A. Tajik: dralitajik@yahoo.com). Received: 27/03/08; accepted: 02/07/08 يننسلما يرغ نم ناكسلا ينب يبصعلا سيلحا عمسلا نادقفو يركسلا مانبه دبيره ،يئايحأ ليع ديرف ،يئايرآ همغن ،كيجات ليع ،يرفظم ميرم ناريإ ةيروهجم في نارهط ةنيدم في تيرجأ يتلا ةنراقلما ةساردلا هذه ىعستو .ةفورعلما يركسلا تافعاضم دحأ عمسلا فعض لّـ ثمي :ةصلالخا نم لقأ مهرماعأ تناك ًاصخش 160 ينب نمو .ناكسلا نم يننسلما يرغ في يبصعلا سيلحا عمسلا نادقف ينبو يركسلا ينب طُبارـتلا مييقت لىإ ةيملاسلإا ،)سنلجاو رمعلا في منهولثماي مهو يركسلاب ينباصم يرغ نونماثو ،يركسلاب نوباصم مهنم نونماث( ةينهم ءاضوضل ضرعتلا مله قبسي لمو ةنس 60 ُّنِس تقفارـتو .)6.6-1.6 :95% ةقثلا ةلصاف ،3.5 ةيحجرلأا ةبسن( يبصع سيح عمس نادقف دهاوشلا نم %20و ينـيركسلا نم %45 ىدل دهوش ضغب عمسلا نادقفل راطتخا لماع نوكي نأ نكمي يركسلا نأ لىع ةساردلا ُّلدتو .يبصعلا سيلحا عمسلا نادقفب ضرلما ةدم عم يركسلا ءاد روهظ .لضفأ ةيجلاع تارايخ راكتبا لىإ ينـيركسلا في يبصعلا سيلحا عمسلا نادقف ببس ديدتح يدؤي نأ نكميو .ينخدتلاو رمعلا نع رظنلا Diabète sucré et perte auditive neurosensorielle chez le sujet non âgé RÉSUMÉ La déficience auditive constitue l’une des complications connues du diabète. Cette étude comparative réalisée à Téhéran, en République islamique d’Iran, visait à évaluer l’association du diabète sucré et de la perte auditive neurosensorielle au sein de la population non âgée. Parmi les 160 sujets âgés de moins de 60 ans ne présentant aucun antécédent d’exposition professionnelle au bruit (80 diabétiques et 80 sujets témoins non diabétiques appariés selon l’âge et le sexe), 45 % des patients diabétiques et 20 % des témoins étaient atteints de perte auditive neurosensorielle (OR 3,5 ; 95 % IC : 1,6 – 6,6). Il est apparu que l’âge au moment de la survenue du diabète et la durée de celui-ci sont associés à la perte auditive neurosensorielle. Le diabète sucré peut constituer un facteur de risque de perte d’audition, quel que soit l’âge ou le statut tabagique. La détermination de la cause de la perte auditive neurosensorielle chez les patients diabétiques peut permettre le développement d’options thérapeutiques plus efficaces. EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 948 Introduction Diabetes mellitus (DM) is a noncom- municable chronic disease with numer- ous cardiovascular  [1,2], neurological  [3], infectious [4] and other complica- tions. One of the known complications  of DM  is  hearing  impairment,  espe- cially hearing  loss and  tinnitus  [5–7],  which  leads  to  a decreased quality of  life  among  those  affected  [8]. There- fore,  prevention  and  treatment  of  sensorineural  hearing  loss  (SNHL)  among diabetic patients  is  important  [9,10].  Some of  the most  important  etiological hypotheses are neuropathy  [11], microangiopathy  [12]  and  an  inevitable consequence of  the ageing  process  (presbycusis)  [13].  In previ- ous  reports evaluating  the association  between DM and SNHL, elderly age  and other confounding factors such as  smoking, occupational noise exposure,  sex and ethnicity were significantly dif- ferent among case and control groups  [14–16].  In  the Islamic Republic of  Iran, de- spite the existence of a national diabetes  prevention  and  control  programme  [17–19],  we  are  nevertheless  chal- lenged with a  large burden of diabetes.  According  to existing records,  the cur- rent prevalence of DM in  the country  is 7.8%–14.5% [20–22]. We designed  this study to evaluate the correlation of  DM and sensorineural hearing loss in a  non-elderly population with no positive  history of occupational noise exposure  or smoking. Methods The current  study was  a  comparative  cross-sectional survey of known cases of  DM and non-diabetic healthy subjects.  The medical ethics committee of Azad  University of Medical Sciences, Tehran,  Islamic Republic of  Iran approved  the  survey.  Sample The care of  all diabetic patients  in  Is- lamic Republic of Iran is coordinated by  the Iranian Diabetes Association (IDA)  and their medical history is recorded in  a central database  in Tehran. Patients  from Tehran  aged 20–60  years with  DM according to existing medical docu- ments in the IDA registry were included  in  the  sample. Using  random number  tables  and  the  IDS  list we  randomly  selected  130  patients  from  among  a  larger  population  attending  diabetes  clinics and  invited  them by  telephone  to participate in our study; 114 subjects  (87.7%)  accepted. All  these  patients  were asked to attend with at least 2 close  relatives aged < 60 years  to be used as  control  subjects  if  they  fulfilled  the  in- clusion criteria.  Our main inclusion criteria were age  < 60  years  and never having  smoked  tobacco.  Insulin  and/or oral  glucose- lowering  agents were used  to  control  diabetes  in  the  case  group. Exclusion  criteria were: older  than 60 years,  cur- rent and/or previous smoking, alcohol  consumption, using any ototoxic drugs,  current  and/or previous work  in  jobs  or  situations with noise exposure,  and  history of hearing disorders such as ana- tomical inner and middle ear disorders  and unilateral conductive deafness. Oc- cupational noise exposure was defined  as  a  self-report  of  holding  a  job  that  required them to speak in a raised voice  to be heard or working without a cap in  jobs usually needing to use an acoustic  cap apart from seasonal and occasional  jobs and activities. Confounding back- ground  diseases  such  as  anatomical  inner  and middle-ear  disorders  and  unilateral conductive deafness were as- sessed by a question about the hearing  status of  the patient  as  a  preliminary  self-report.  Simple  random sampling was used  to  select  a  control  group with  similar  demographic characteristics who met  all the inclusion and exclusion criteria.  A  total of 80 DM patients and 80  age- and sex-matched control  subjects  who met the  inclusion criteria were re- cruited to the study.  Data collection The key variables analysed in relation to  DM were: age,  sex, presence of SNHL  and the type (1- or 2- sided) and sever- ity of SNHL.  Patients’  age,  sex,  smoking history,  history  of  alcohol  consumption  and  previous medical  history were  avail- able from the IDA registry. Similar data  were obtained from control subjects by  questionnaire. Other variables that were  evaluated specifically for this study only  in diabetic patients included duration of  DM, age at onset, type of diabetes (type  1 or 2), glycaemic control  [defined as  glycosylated haemoglobin (HbA1C) <  8%], mean fasting blood glucose (FBG)  level from at least 2 measurements and  presence of DM complications. FBG  was determined by the glucose oxidase- peroxidase aminophenazone phenol en- zymatic colorimetric  test using venous  blood samples obtained after 12 hours  of  fasting. Data on patients’  type  and  duration of DM and diabetes complica- tions were taken from the IDA registry  and were based on clinical examinations  performed by general practitioners and  specialists. These  included presence of  nephropathy,  retinopathy, neuropathy,  cerebrovascular disease, cardiovascular  disease  and other  reported complica- tions. The association of  these  factors  with SNHL and  its  type  and  severity  were also  separately evaluated among  the 80 DM patients. An otoscopic evaluation of patients  and controls was made using pure-tone  air- and bone- conduction audiometry  by 2  experienced otorhinolaryngolo- gists. We used the same AC9 2-channel  clinical audiometer (Weltone, Tehran)  and  the  same  examiner  to  reduce  confounding  factors.  Pure-tone  air- conduction  thresholds were obtained  for  each  ear  at  250,  500,  1000, 2000,  طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 949 3000, 4000, 6000 and 8000 Hz. Bone- conduction threshold was measured at  500 and 4000 Hz. We defined hearing  loss as having pure-tone average (PTA)  thresholds  greater  than 25 dB  in  the  worse ear at 0.5, 1, 2 and 4 kHz frequen- cies.  Severity of SNHL was  classified  from 1 to 5 (Table 1) [23].  Analysis Data were analysed using SPSS, version  13.0  software. Differences were  tested  by analysis of variance (ANOVA), chi- squared, Fisher exact, independent sam- ples  t-test,  logistic  regression  analysis  and Mann-Whitney U-tests  and were  considered  statistically  significant at P values < 0.05. Results The mean age of our subjects was 45.0  [standard  deviation  (SD 9.9)]  years  in  the DM group  and 45.1  (SD 9.8)  years in the healthy control group (P = 0.990), range 21–59 years. In both DM  and control groups 51 subjects (63.8%)  were female and 29 (36.3%) were male  (P = 1.00). Of  the diabetic patients, 9  (11.3%)  had  type  1  and  71  (88.8%)  type 2 DM. According  to  the  PTA  readings,  SNHL was present  in 36 DM patients  (45.0%) and 16 non-diabetics (20.0%)  (P < 0.001). The odds ratio of DM for  the  presence  of  hearing  loss was  3.5  (95% confidence  interval 1.6–6.6, P <  0.001). However,  the  type of  involve- ment  (1-  or  2-sided)  and  severity  of  SNHL were not related to the presence  of  SNHL  (P  =  0.771  and P  =  0.644  respectively) (Table 2). The mean age of diabetic patients  with SNHL was 47.7 (SD 8.07) years  and in diabetic patients without SNHL  was 42.3 (SD 10.12) years. There was a  borderline  statistically  significant asso- ciation between presence of SNHL and  age in DM patients (P < 0.05). However,  the type and severity of SNHL were not  related  to patient’s age (P = 0.804 and P = 0.217 respectively).  The  mean  duration  of  DM  was  significantly  longer  among  diabetic  patients with  SNHL  [11.7  (SD 7.6)  years]  than  those without SNHL [7.3  (SD  5.4)  years]  (P  =  0.001)  (Table  3). Age at onset of DM and FBG level,  however, were not associated with pres- ence of SNHL. Mean FBG was higher  in diabetic patients with SNHL than in  those without SNHL [175.3 (SD 83.3)  mg/dL  versus  157.7  (54.9) mg/dL]  but  the difference was not  statistically  significant (P = 0.247). The FBG  level  was not significantly related to severity  of SNHL, but  lower age at DM onset  and  longer duration of diabetes were  related  to  higher  severity  of  SNHL (P = 0.042 and P = 0.007 respectively).  Of  the  34  patients  uncontrolled  DM 19  (55.9%)  had  SNHL  and  15  did not (44.1%) but the difference was  not statistically  significant (P = 0.110).  Type and  severity of SNHL were not  associated with  glycaemic  control  in  diabetics. SNHL frequency, severity or  type also showed no statistically signifi- cant associations with  the presence of  DM complications  (Table 4). SNHL  severity  was  associated with  type  of  DM, with 1/4 (25%) of type 1 patients  with SNHL having grade 5 SNHL com- pared with  0/12  (0%)  of  the  type  2  patients with SNHL (P = 0.032) Discussion In the current survey the rate of SNHL  was compared in a case group of patients  Table 1 Classification of severity of sensorineural hearing loss Category Hearing loss (dB) Severity Normal 0–15 – Slight 16–25 – Mild 26–40 1 Moderate 41–55 2 Moderate to severe 56–70 3 Severe 71–90 4 Profound > 90 5 Table 2 Comparison of sensorineural hearing loss (SNHL) type and severity between diabetic patients and non-diabetic healthy controls Variable Diabetic (n = 80) Non-diabetic (n = 80) No. % No. % Presence of SNHL Yes 36 45.0 16 20.0 No 44 55.0 64 80.0 SNHL type 1-sided 12 33.3 6 37.5 2-sided 24 66.7 10 62.5 SNHL severity 1 14 39.0 9 56.3 2 17 47.2 6 37.5 3 3 8.3 1 6.3 4 1 2.8 0 0.0 5 1 2.8 0 0.0 EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 950 with DM and a healthy control group.  Our findings showed a relationship be- tween some aspects of SNHL and DM.  This  is  similar  to findings  reported by  Kakarlapudi et al.  in  the United States  [14]. DM had no statistically significant  correlation with  the severity of SNHL,  suggesting that DM only may act as an  initiating  factor  and  that  the progres- sion of hearing  loss  is  related  to other  features. Neither FBG nor  glycaemic  control were associated with the occur- rence or  severity of SNHL. While  the  FBG level was higher in diabetic patients  with SNHL (175.3 versus 157.7 mg/ Table 3 Association of sensorineural hearing loss (SNHL) type and severity with diabetes mellitus (DM)-related characteristics among patients with DM Variable No. of patients DM duration (years) Age at onset (years) FBG level (mg/dL) Mean (SD) Mean (SD) Mean (SD) Presence of SNHL Yes 36 11.7 (7.6) 35.6 (8.9) 175.3 (83.3) No 44 7.3 (5.4) 35.8 (12.3) 157.7 (54.9) P < 0.001 P = 0.946 P = 0.247 SNHL type 1-sided 12 9.3 (6.6) 39.3 (11.2) 161.7 (56.2) 2-sided 24 12.9 (7.9) 33.8 (10.4) 182.1 (85.3) P = 0.064 P = 0.203 P = 0.343 SNHL severity 1 14 10.1 (7.1) 35.1 (8.4) 169.1 (78.4) 2 17 10.7 (7.2) 39.0 (11.1) 173.9 (60.2) 3 3 24.0 (12.9) 29.3 (6.2) 227.3 (97.2) 4 1 19.0 (–) 3.0 (–) 115.0 (–) 5 1 7.0 (–) 38.0 (–) 190.0 (–) P = 0.007 P = 0.042 P = 0.496 FBG = fasting blood glucose; SD = standard deviation; – = not applicable. Table 4 Comparison of sensorineural hearing loss (SNHL) type and severity among patients with diabetes mellitus (DM) by type, complications and glycaemic control Variable DM type DM complicationsa Glycaemic controlb Type 1 Type 2 Yes No Yes No No. % No. % No. % No. % No. % No. % Presence of SNHL Yes 4 44.4 32 45.1 9 64.3 27 40.9 17 37.0 19 55.9 No 5 55.6 39 54.9 5 35.7 39 59.1 29 63.0 15 44.1 P = 1.0 P = 0.093 P = 0.11 SNHL type 1-sided 0 0.0 12 37.5 2 22.2 10 37.0 6 35.3 6 31.6 2-sided 4 100.0 20 62.5 7 77.8 17 63.0 11 64.7 13 68.4 P = 0.278 P = 0.813 P = 0.414 SNHL severity 1 3 75.0 11 34.4 2 22.2 12 44.4 9 52.9 5 26.3 2 0 0.0 17 53.1 5 55.6 12 44.4 6 35.3 11 57.9 3 0 0.0 3 9.4 2 22.2 1 3.7 1 5.9 2 10.5 4 1 25.0 0 0.0 0 0.0 1 3.7 1 5.9 0 0.0 5 0 0.0 1 3.1 0 0.0 1 3.7 0 0.0 1 5.3 P = 0.032 P = 0.227 P = 0.332 aNephropathy, retinopathy, neuropathy, cerebrovascular disease, cardiovascular disease and other reported complications. bGlycosylated haemoglobin (HbA1C) < 8%. طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 951 dL) and  the proportion with SNHL  was higher  among  subjects with un- controlled DM (55.9% versus 44.1%),  these differences were not statistically  significant. This demonstrates that glu- cose metabolism may not be the most  important  issue  in  the development  of SNHL and perhaps only acts as an  aggravating  factor. We did not meas- ure  the  insulin  level of patients, but  it  has been reported that neither  insulin  resistance nor decreased insulin secre- tion are  association with SNHL [5].  Despite  the small number of patients  with  type 1 DM in  the current study,  these patients were significantly more  likely to have a severe grade of SNHL  than patients with  type 2 DM. How- ever,  there was no significant correla- tion between type of DM and presence  of SNHL. Most  previous  surveys  on  this  subject have been carried out among  patients of all ages, whereas our study  was  performed  only  in  non-elderly  subjects aged < 60 years. Sakuta et al.  reported a statistically significant higher  Janghorbani M, Amini M, Tavassoli A. Coronary heart disease 1. in type 2 diabetes mellitus in Isfahan, Iran: prevalence and risk factors. Acta Cardiologica, 2006, 61:13–20. Esteghamati A et al. H. Prevalence of diabetes and other car-2. diovascular risk factors in an Iranian population with acute coronary syndrome. Cardiovascular Diabetology, 2006, 5:15. Arabshahi KS, Koohpayezade J. Investigation of risk factors for 3. surgical wound infection among teaching hospitals in Tehran. International Wound Journal, 2006, 3:59–62. Booya F et al. Potential risk factors for diabetic neuropathy: a 4. case control study. BMC Neurology, 2005, 5:24. Maia CA, Campos CA. Diabetes mellitus as etiological factor of 5. hearing loss. Revista Brasileira de Oto-Rino-Laringologia. 2005, 71:208–214. Diaz de Leon-Morales LV et al. Auditory impairment in patients 6. with type 2 diabetes mellitus. Archives of Medical Research, 2005, 36:507–510. Kazmierczak H, Doroszewska G. Metabolic disorders in ver-7. tigo, tinnitus, and hearing loss. International Tinnitus Journal, 2001, 7:54–58. Dalton DS et al. The impact of hearing loss on quality of life in 8. older adults. Gerontologist, 2003, 43:661–668. Cruickshanks KJ et al. The 5-year incidence and progres-9. sion of hearing loss: the epidemiology of hearing loss study. Archives of Otolaryngology—Head and Neck Surgery, 2003, 129:1041–1046. References Kakehata S et al. Comparison of intratympanic and intrave-10. nous dexamethasone treatment on sudden sensorineural hearing loss with diabetes. Otology and Neurotology, 2006, 27:604–608. Toth F et al. Investigation of auditory brainstem function in dia-11. betic patients. International Tinnitus Journal, 2003, 9:84–86. Shikowitz MJ. Sudden sensorineural hearing loss. 12. Medical Clin- ics of North America, 1991, 75:1239–1250. Frisina ST et al. Characterization of hearing loss in aged type II 13. diabetics. Hearing Research, 2006, 21:1103–1113. Kakarlapudi V, Sawyer R, Staecker H. The effect of diabetes 14. on sensorineural hearing loss. Otology and Neurotology, 2003, 24:382–386. Sakuta H et al. Type 2 diabetes and hearing loss in personnel of 15. the self-defense forces. Diabetes Research and Clinical Practice, 2007, 75:229–234. Dalton DS et al. Association of NIDDM and hearing loss. 16. Dia- betes Care, 1998, 21:1540–1544. Ghanbari A et al. Assessment of factors affecting quality of 17. life in diabetic patients in Iran. Public Health Nursing, 2005, 22:311–322. Azizi F et al. The diabetes prevention and control programme 18. of the Islamic Republic of Iran. Eastern Mediterranean Health Journal, 2003, 91:114–121. Russel M. Applying DALY to assessing national health insur-19. ance performance: the relationship between the national prevalence of hearing loss among dia- betic  and non-diabetic middle-aged  men (60.2% and 45.2% respectively)  [15]. Dalton  et  al.  showed  a  higher  incidence of hearing  loss among dia- betic subjects compared with a control  group, but they reported no significant  association between hearing  loss and  DM type 2 [16].  We also found that the age of onset  and duration of DM were  associated  with occurrence of SNHL. Therefore,  the role of DM progression and ageing  should be  considered more  carefully  [14,24].  In  the  current  study  the  age  of diabetic patients had only a border- line association with severity of SNHL  (P  = 0.042)  suggesting  that  ageing  is  not a  factor  in SNHL in  these patients  and that the role of disease progression  should be  investigated more precisely.  Patients’ sex was matched with controls  in the current study to diminish its con- founding  role  [25]. Also, none of our  patients were ever  smokers  [26] and/ or  had  previous  or  current  exposure  to noise pollution. All of  these  factors  allowed  us  to  eliminate  some  of  the  possible confounding factors in the role  of DM in SNHL development.  Since many people worldwide are  living  in communities with a high  rate  of  undiagnosed DM  [27]  and  since  hearing  loss can be considered  to be a  consequence of diabetes,  a metabolic  assessment may be useful  for patients  presenting with hearing  loss. On  the  other hand, routine screening  for hear- ing  loss  in diabetic patients may  also  be helpful  to diminish  comorbidities  among  these  patients,  with  a  conse- quent  improvement  in  their quality of  life. Determining the cause of SNHL in  diabetic patients may  lead  to develop- ment of  better  treatment options  for  both conditions [28].  Acknowledgements We are indebted to M. Amoorvand for  her helpful  advice during  this project  and also to the patients and physicians  who participated in this survey. EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 952 WHO activities to prevent and control diabetes WHO aims to stimulate and support the adoption of effective measures for the surveillance, prevention and control of  diabetes and its complications, particularly in low- and middle-income countries. To this end, WHO: provides scientific guidelines for diabetes prevention; • develops norms and standards for diabetes care; • builds awareness on the global epidemic of diabetes; including partnership with the International Diabetes • Federation in the celebration of World Diabetes Day (14 November);  conducts surveillance of diabetes and its risk factors. • The WHO Global Strategy on Diet, Physical Activity and Health complements WHO’s diabetes work by focusing on  population-wide approaches to promote healthy diet and regular physical activity, thereby reducing the growing global  problem of overweight and obesity.  Source: WHO Fact sheet, No. 312 November 2009 (http://www.who.int/mediacentre/factsheets/fs312/en/index.html) health insurance expenditures and the burden of disease measures in Iran. International Journal of Health Planning and Management, 2005, 20:89–98. Azizi F. Diabetes mellitus in the Islamic Republic of Iran. 20. IDF Bulletin, 1996, 41:38–39. Azizi F et al. Screening for type 2 diabetes in the Iranian na-21. tional programme: a preliminary report. Eastern Mediterranean Health Journal, 2003, 9:1122–1127. Amini M et al. Prevalence and risk factors of diabetes mellitus in 22. the Isfahan city population (aged 40 or over) in 1993. Diabetes Research and Clinical Practice, 1997, 38:185–190. Harrell RW. Pure tone evaluation. In: Katz J, ed. 23. Handbook of clinical audiology. New York, Williams and Wilkins, 2002:82. Rózańska-Kudelska M et al. [Hearing loss in patients with 24. diabetes mellitus type II.] Zaburzenia sluchu u chorych na cukrzyce 2 typu. Otolaryngologia Polska, 2002, 56:607–610. Helzner EP et al. Race and sex differences in age-related hear-25. ing loss: the Health, Aging and Body Composition Study. Jour- nal of the American Geriatric Society, 2005, 53:2119–2127. Cruickshanks KJ et al. Cigarette smoking and hearing loss: the 26. epidemiology of hearing loss study. Journal of the American Medical Association, 1998, 279:1715–1719. Wierusz-Wysocka B et al. Wystepowanie cukrzycy nieznanej w 27. populacji czynnych zawodowo osob w srodowisku miejskim [Appearance of undiagnosed diabetes mellitus in the popula- tion of professionally active people in the urban areas]. Polskie Archiwum Medycyny Wewnetrznej, 2001, 106:815–821. Syal R, Tyagi I, Goyal A. Bilateral Ramsay Hunt syndrome in a 28. diabetic patient. BMC Ear, Nose and Throat Disorders, 2004, 4:3 (doi:10.1186/1472-6815-4-3). طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 953 Radiological changes in rheumatoid arthritis patients at a teaching hospital in Saudi Arabia S.M. Attar 1 and A. Al-Ghamdi 2 ABSTRACT The frequency of radiological changes of the hands and the feet were investigated in a well-defined hospital population of patients with rheumatoid arthritis in Jeddah, Saudi Arabia. A total of 57 patients who fulfilled the American College of Rheumatology 1987 criteria for the diagnosis of rheumatoid arthritis were randomly chosen from the rheumatology outpatient clinic at King AbdulAziz University Hospital. Erosions were seen in 60% of the patients and periarticular osteopenia in 34%; deformity and soft tissue swelling were present in 26% and 14% of patients respectively. The proportion with erosions is lower than data reported from Western European and North American populations but higher than previous data from the Central region of Saudi Arabia. 1Department of Rheumatology; 2Department of Internal Medicine, King Abdul-Aziz University Hospital, Jeddah, Saudi Arabia (Correspondence to S.M. Attar: suzan_attar@hotmail.com). Received: 26/12/08; accepted: 17/03/09 ةيدوعسلا ةيميلعتلا تايفشتسلما دحأ في يديوتامورلا لصافلما باهتلا ضىرم في ةيعاعشلا تا ُّريغتلا يدماغلا ةشئاع ،راطع روصنم نازوس ضىرم نم ةد َّدمح ةيناّكس ةرَه َْج لىع لمتشي ىفشتسم في ينمدقلاو نيديلا في ةيعاعشلا تا ُّريغتلا ثودح ر ُـ تاوت لوح ِّيصقتلا اذه يرجُأ :ةصلالخا باهتلا صيخشت يياعم نوفْوَتسي ًاضيرم نوسخمو ةعبس ،ًايئاوشع ،يتخاو .ةيدوعسلا ةيبرعلا ةكلملماب ةدج ةنيدم في يديوتامورلا لصافلما باهتلا ىدل ظحولو .زيزعلا دبع كللما ةعماج ىفشتسم في مزيتامورلا بطل ةيجرالخا ةدايعلا نم ،1987 مزتامورلا بطل ةيكيرملأا ةيلكلاب ةصالخا لصافلما ةبسن تناكو .مهنم %14 في ةوخرلا ةجسنلأا م ُّرروتو ،%26 في ه ُّروشتلاو ؛مهنم %34 في ةيمظعلا ةدالما ةّلق تظحول ماك ،%60 في لكآتلا دوجو ضىرلما ةكلملماب ىطسولا ةقطنلما في ةل َّجسلما ةقباسلا تايطعلما نم لىعأ اهنكلو ،ةيلماشلا اكيرمأو ةيقشرلا ابوروأ ناكس في ةل َّجسلما تايطعلما نم لقأ لكآتلا .ةيدوعسلا ةيبرعلا Modifications radiologiques chez des patients atteints de polyarthrite rhumatoïde dans un hôpital universitaire en Arabie saoudite RÉSUMÉ La fréquence des modifications radiologiques au niveau des mains et des pieds a été étudiée au sein d’une population hospitalière bien définie constituée de patients atteints de polyarthrite rhumatoïde, à Jeddah en Arabie saoudite. Au total, 57 patients répondant aux critères de l’American College of Rheumatology 1987 pour le diagnostic de la polyarthrite rhumatoïde ont été choisis au hasard dans le service de consultation rhumatologique externe de l’hôpital universitaire Roi Abdul Aziz. Des érosions ont été observées chez 60 % des patients et une ostéopénie péri-articulaire chez 34 % ; des déformations et des œdèmes au niveau des tissus mous étaient présents chez 26 % et 14 % des patients, respectivement. La proportion de patients présentant des érosions s’avère plus faible que les données rapportées sur les populations d’Europe occidentale et d’Amérique du Nord, mais plus élevée que les données précédemment recueillies dans le centre de l’Arabie saoudite. EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 954 Introduction Rheumatoid arthritis (RA) is a chronic  autoimmune disease characterized by  inflammatory erosive polyarthritis  and  systemic features. It leads to irreversible  joint damage, disability  and even pre- mature mortality [1–3]. Work disability  occurs in 20%–30% of patients with RA  if left untreated [4]. It is well know that  the progression of radiological changes  is slowed by the use of disease-modify- ing  antirheumatic  agents  (DMARD)  [5] and can be ameliorated by  the use  of biologic agents  [6–8]. Therefore  to  improve the outcome, it is important to  diagnose and treat the disease early. In  clinical  trials,  the  diagnosis  of  RA  is usually based on  the American  College of Rheumatology (ACR) clas- sification criteria  that  includes clinical  manifestations,  laboratory  investiga- tions  and  radiographic bone erosions  [9].  In  typical  outpatient  practice,  a  definitive diagnosis using  these criteria  may be difficult  to obtain early  in  the  disease process  [10]. Reports  suggest  that erosions at  the metacarpophalan- geal and proximal interphalangeal joints  can  be  seen  on  plain  radiography  in  15%–30% of patients during  the first  year of disease and  in patients who do  not  respond  to  therapy  the  incidence  rises  to 90% by  the end of  the  second  year [11,12]. In some patients, erosions  occurs first  in  the ulnar  styloid or  the  5th metatarsophalangeal joint, whereas  in other cases erosion at the feet occurs  earlier  and  faster over  the  years  than  erosions at the hand [13,14]. It is there- fore worth evaluating both  the hands  (including the wrists) and the feet in all  patients  in whom a diagnosis of RA is  suspected [15].  A  number  of markers  have  been  suggested as predictors of erosive dis- ease  in RA [16–21]  including  levels of  the  autoantibody  rheumatoid  factor  (RF). Seropositive patients have more  aggressive disease and more common  extra-articular manifestations  than  se- ronegative patients  and  radiographic  progression  is more  rapid  among RA  patients with positive  levels  of RF  at  initial evaluation [16,17].  The prevalence of RA in  the Saudi  Arabian  population was  reported  to  be about 0.02%  in a  study  in  the Cen- tral  region [22]. Another  study  in our  country showed that the proportion of  RA patients with  radiological  changes  in the form of erosions at the hands and  the feet was 39% and that feet erosions  were  less  frequent  than hand erosions  (6% versus 39% respectively)  [23].  In  the present  study we  investigated  the  frequency of radiological changes of the  hands and the feet at a teaching hospital  in  Jedda, Saudi Arabia  and compared  the findings with those from other stud- ies here and elsewhere. We also studied  the  relationship between  radiological  changes and the levels of the prognostic  marker RF. Methods Study design and setting A prospective study was conducted at  King Abdulaziz University Hospital,  the only government  teaching hospital  in Jedda in the Western region of Saudi  Arabia. With a bed capacity around 700  it provides health care to a multinational  population  of mixed  socioeconomic  status.  Sample All RA patients attending the outpatient  clinic of our hospital over  the period  June to December 2007 (n = 57) were  enrolled in the study. All had been diag- nosed with RA according  to  the 1987  ACR classification criteria [9].  Data collection The  following  data  were  collected  from the clinical records: demographic  features; clinical findings, e.g. duration  of disease at  the  time of  the study and  activity of the disease based on the 28- item disease  activity  score  (DAS28)  [24];  serum  level of RF; and smoking  history (defined as current  smoker or  nonsmoker). A DAS  score ≥ 5.1 was  considered  as  active disease. RF was  measured  by  immunonephelometry  with the quantitative N Latex RF system  (Dade Behring, Germany), with a nor- mal upper  limit of 20 IU/L, according  to the manufacturers’ instructions.   Radiological  changes were  evalu- ated by taking X-rays of  the hands and  wrists  (anteroposterior,  lateral  and  semi supine) and  the  feet  (anteropos- terior and lateral). These were reviewed  by 2  radiologists and  the authors. The  radiological  changes noted  included:  soft-tissue  swelling,  periarticular  os- teopenia (defined as a localized area of  decreased bone density with loss of the  trabecular pattern  at  the periarticular  area);  joint  space narrowing; erosions  (defined as erosion  through  the cortex  of  the bone around the margins of  the  joint  “bare area” where  the bone  is not  protected by overlying cartilage);  and  subluxation and deformity.  Statistical analysis Data  analysis  was  done  using  SPSS,  version 16  software. Mean and stand- ard deviation (SD) was calculated  for  quantitative data  and proportions  for  categorical variables. Student  t-test was  used  for comparing means of continu- ous variables. The chi-squared  test was  used  to  analyse  group differences  for  categorical variables and a P value of <  0.05 was considered significant. Results Radiological  changes were  reviewed  in 57 patients  fulfilling  the 1987 ACR  classification criteria  for  the diagnosis  of RA: 44 (77%) women and 13 (23%)  men, a female to male ratio of 3:1. Their  mean age at  the  time of  the study was  45.3 (SD 11.8) years. The majority of  the patients (61%) were Saudi Arabian  nationality while 39% were non-Saudi  Arabianss. All patients were Muslims ex- طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 955 cept for 3 patients (5%). Only 1 patient  was a current smoker. The mean disease duration  at  the  time of the study was 7.3 (SD 4.2) years  (range 1–13 years).  There were 22 patients (39%) with  active disease according to their DAS28  scores. Radiological  changes were de- tected  in 38 patients (67%). However,  none of  them was disabled or using a  wheelchair. All of  the patients were  re- ceiving DMARD  treatment. Table 1  defines the joint involvement in the 57  RA patients according to the radiologi- cal  changes;  some patients had more  than  1  finding. The most  frequently  involved joints were the proximal inter- phalangeal  joint, ulnar  styloid and  the  metatarsophalangeal joint in 13 patients  each (23%),  followed by  the metacar- pophalangeal joint in 12 patients (21%),  the  carpal bones  in 7 patients  (12%)  and the distal interphalangeal joint in 3  patients (5%). Radiological  evaluation  revealed  erosions in 34 patients (60%), followed  by periarticular  osteopenia  in 21 pa- tients  (37%), deformity with  subluxa- tion in 15 patients (26%) and soft-tissue  swelling in 8 patients (14%).  RF was classified as positive  in 29  patients (51%). There was no correla- tion between radiological changes and  disease activity (r = 0.38, P = 0.53) or RF  positivity (r = 2.12, P = 0.145). Discussion Our data  showed  that 60% of RA  patients developed erosions over  the  course of  disease  and  the number of  erosions were similar at  the hands and  the  feet. Our  data  are  different  from  the radiological changes of RA in some  industrialized countries. A Swedish 10- year follow-up study of 181 patients with  early RA showed  that 90% developed  erosion by year 2 and 96% by year 10.  Feet were more involved than the hands  at the early stages in these patients, but  later the hands and the feet were equally  affected [16].  In  the Netherlands 147  patients with RA were  followed  for 3  years;  70% of  the patients developed  erosion by the end of the study and foot  involvement  were  higher  than  hand  involvement  [11]. A  study  of  58 RA  patients in the United States who were  followed up  for  2  years  showed  that  erosions started early in the feet, and by  the end of the 2-year study feet erosions  were greater  than hand erosions  [25].  All  the above  studies  suggest  that RA  is a more aggressive disease in Western  European and North American popu- lations,  that  foot  involvement occurs  earlier than hand involvement and that  these differences may persist until late in  the disease. Based on a literature search  there was only one other study conduct- ed in our country, in the Central region,  showing that radiological changes were  less  severe  than our data and  that  feet  were  less  involved  than the hands (6%  versus 39%) [23].  The finding that RA in our country  is milder  than  in  industrialized popu- lations  has  also  been  reported  from  Kuwait  [26], Oman  [27], UAE  [28],  Iraq [29] and Egypt [30]. Radiological  erosions developed  in 42% of RA pa- tients in Kuwait, 45.2% in Oman, 55.2%  in UAE, 48% in Iraq and 75% in Egypt.  Even  the highest figure,  from Egypt,  is  still lower than the figures of 90%–96%  from Sweden and 93%–96%  from the  United States [11,12].  Although seropositive RF has been  suggested  to  be  an  independent  risk  factor for the development of radiologi- cal changes [16,17], no association with  erosions was found in our study. While smoking has also been linked  to  joint  damage  in RA  [18–21],  we  could not to determine any association  of  smoking with erosions as  there was  only 1  smoker  in our  sample, possibly  because most of our RA patients were  women (77%) and the rate of smoking  among women in Saudi Arabia is tradi- tionally very low (0.9% in 1999) [31]. The  difference  between  our  data  and  those  from patients  in  industrial- ized  countries  could be  explained by  different associations of human  leuko- cyte  antigen  (HLA)  alleles  between  Saudi Arabian and other populations.  RA in  the Saudi Arabian population  is  associated with HLA-DR10, whereas in  Caucasians  it  is  associated with HLA- DR4 and HLA-DR1 [32]. Other pos- sible factors are the differences in social  habits:  slippers  rather  than  shoes  are  more commonly worn due  to  the hot  climate;  the preferred  sitting position  is on the floor, which may improve the  blood circulation at  the  feet compared  with sitting on a chair; and the 5 times  daily  ablutions  and Muslim  religious  prayers  involve many actions  that flex  and exercise  the  joints [23,33]. Studies  have been published  from other  cul- tures that support our hypothesis about  the possible  influence of  lifestyle  and  cultural activities on the range of move- ments. A Japanese study examined the  effect of the Japanese way of sitting after  Table 1 Joint involvement in 57 rheumatoid arthritis patients according to radiological changes Joint involvement Radiological changes No radiological changes No. % No. % Proximal interphalangeal joint 13 23 44 77 Ulnar styloid 13 23 44 77 Metatarsophalangeal joint 13 23 44 77 Metacarpophalangeal joint 12 21 45 79 Carpal bones 7 12 50 88 Distal interphalangeal joint 3 5 54 95 EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 956 total knee replacement in patients with  RA, and found that this had beneficial  postoperative  results  [34].  Another  study conducted  in  the United Arab  Emirates  (UAE)  indicated  that Mus- lim patients with  knee osteoarthritis  were more mobile  than non-Muslim  patients [35]. At the same time, studies  from Norway  and  the Netherlands  clearly demonstrated  the positive ef- fect of hand exercises  in RA patients  on  handgrip  strength  and  erosions  [36,37]. Conclusions Our findings  showed a higher  rate of  radiological  changes  in RA  patients  in  the Western  region  of  Saudi Ara- bia  than  in  the Central  region  but  a  lower  rate  than  among patients  from  industrialized countries. The hands and  the  feet were  equally  involved. There  was no significant association between  the  radiological  findings  and  RF  or  smoking  history. We  recommend  a  large prospective  study  to document  the predictors  for  radiological progres- sion, as well evaluations of the effect of  cultural  factors  such as Muslim prayer  on erosions of the hands and feet in RA  patients. Mimori T. Clinical significance of anti-CCP antibodies in 1. rheumatoid arthritis. Internal Medicine (Tokyo, Japan), 2005, 44:1122–1126. Boire G et al. Anti-Sa antibodies and antibodies against cy-2. clic citrullinated peptide are not equivalent as predictors of severe outcomes in patients with recent–onset polyarthri- tis. Arthritis Research and Therapy, 2005, 7:592–603. Gerard A et al. The diagnostic properties of rheumatoid ar-3. thritis antibodies recognizing a cyclic citrullinated peptide. Arthritis and Rheumatism, 2000, 43:155–163. Sokka T. Work disability in early rheumatoid arthritis. 4. Clinical and Experimental Rheumatology, 2003, 21(5 Suppl. 31):S71–S74. Valiukienė K, Butrimienė I, Venalis A. The dynamic of ra-5. diological changes in early rheumatoid arthritis patients treated with methotrexate and sulphasalazine. Acta Medica Lituanica, 2001, 8(1):52–56. Nurmohamed MT, Dijkmans BA. Are biologics more effec-6. tive than classical disease-modifying antirheumatic drugs? Arthritis Research and Therapy, 2008, 10(5):118. Nurmohamed MT, Dijkmans BA. Efficacy, tolerability and 7. cost effectiveness of disease–modifying antirheumatic drugs and biologic agents in rheumatoid arthritis. Drugs, 2005, 65(5):661–694. Lorenz HM. Biological agents: a novel approach to the 8. therapy of rheumatoid arthritis. Expert Opinion on Investiga- tional Drugs, 2000, 9(7):1479–1490. Arnett FC et al. The American Rheumatism Association 9. 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis and Rheumatism, 1988, 31(3):315–324. Visser H. Early diagnosis of rheumatoid arthritis. 10. Best Prac- tice and Research. Clinical Rheumatology, 2004, 19(1):55–72. Van der Heijde DM et al. Biannual radiographic assess-11. ments of hands and feet in a three-year prospective follow up of patients with early rheumatoid arthritis. Arthritis and Rheumatism, 1992, 35(1):26–34. Fuchs HA et al. Evidence of significant radiographic damage 12. in rheumatoid arthritis with in the first 2 years of disease. Journal of Rheumatology, 1989, 16(5):585–591. Van der Heijde DM et al. Radiographic progression on 13. radiographs of hands and feet during the first 3 years of rheumatoid arthritis measured according to Sharp’s meth- od (van der Heijde modification Journal of Rheumatology, 1995, 22(9):1792–1796. Hulsmans HM et al. The course of radiologic damage dur-14. ing the first six years of rheumatoid arthritis. Arthritis and Rheumatism, 2000, 43(9):1927–1940. References Listing J et al. HLA-DRB1 genes, rheumatoid factor, and 15. elevated C-reactive protein: independent risk factors of radiographic progression in early rheumatoid arthritis. Ber- lin Collaborating Rheumatological Study Group. Journal of Rheumatology, 2000, 27(9):2100–2109. Lindqvist E et al. Course of radiographic damage over 10 16. years in a cohort with early rheumatoid arthritis. Annals of the Rheumatic Diseases, 2003, 62(7):611–616. Flatø B et al. The influence of patient characteristics, dis-17. ease variables, and HLA alleles on the development of radiographically evident sacroiliitis in juvenile idiopathic arthritis. Arthritis and Rheumatism, 2002, 46(4):986–994. Vencovský J et al. Autoantibodies can be prognostic mark-18. ers of an erosive disease in early rheumatoid arthritis. An- nals of the Rheumatic Diseases, 2003, 62(5):427–430. Massardo L et al. The presence of the HLA-DRB1 shared 19. epitope correlates with erosive disease in Chilean patients with rheumatoid arthritis. Rheumatology (Oxford), 2002, 41(2):153–156. Combe B et al. Prognostic factors for radiographic 20. damage in early rheumatoid arthritis: a multiparam- eter prospective study. Arthritis and Rheumatism, 2001, 44(8):1736–1743. Lindqvist E et al. Prognostic laboratory markers of joint 21. damage in rheumatoid arthritis. Annals of the Rheumatic Diseases, 2005, 64(2):196–201. Al-Dalaan A et al. The prevalence of rheumatoid arthritis 22. in the Qassim region of Saudi Arabia. Annals of Saudi Medi- cine, 1998, 18(5):396–397. Al-Boukai AA, Al-Arfaj AS. Patterns of radiographic changes 23. in hands and feet of rheumatoid arthritis in Saudi Arabia. Saudi Medical Journal, 2003, 24(4):396–399. Prevoo ML et al. Modified disease activity scores that 24. include twenty-eight-joint counts. Development and validation in a prospective longitudinal study of patients with rheumatoid arthritis. Arthritis and Rheumatism, 1995, 38(1):44–48. Möttönen TT. Prediction of erosiveness and rate of devel-25. opment of new erosions in early rheumatoid arthritis. An- nals of the Rheumatic Diseases, 1988, 47(8):648–653. Al-Salem IH, Al-Awadhi AM. The expression of rheuma-26. toid arthritis in Kuwaiti patients in an outpatient hospital- based practice. Medical Principles and Practice, 2004, 13(1):47–50. Al-Attia HM et al. Rheumatoid arthritis in a population 27. sample in the Gulf: clinical observations. Clinical Rheuma- tology, 1993, 12(4):506–510. طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 957 CD ROM international statistical classification of diseases and health related problems (The) ICD- 10. 2008 Edition This new edition of  WHO’s International Statistical Classification of Diseases and Related Health Problems, 10th Revi- sion (ICD-10) has been fully updated. Originally published in the early 1990s, ICD-10 now incorporates all updates  and other changes to this core health classification since 1996 up to the end of 2008. This electronic version of the three printed volumes of ICD-10, 2008 edition is a self-contained electronic publication  using the powerful eComPress® technology which offers users the several powerful and valuable features. This product is available for use on a single terminal based computer and/or a local area network (LAN). More infor- mation on obtaining this product is available from: http://www.who.int/publications/en/ Badsha H, Kong KO, Tak PP. Rheumatoid arthritis in the 28. United Arab Emirates. Clinical Rheumatology, 2008, 27(6):739–742. Al-Rawi ZS et al. Rheumatoid arthritis in population 29. samples in Iraq. Annals of the Rheumatic Diseases, 1978, 37(1):73–75. Zayed A et al. Some biochemical changes in serum and 30. synovial fluid in patients with rheumatoid arthritis. Journal of Medical Science, 2007, 7(4):526–535. Jarallah JS et al. Prevalence and determinants of smoking 31. in three regions of Saudi Arabia. Tobacco control, 1999, 8:53–56. Al-Arfaj AS. Characteristics of rheumatoid arthritis relative 32. to HLA-DR in Saudi Arabia. Saudi Medical Journal, 2001, 22(7):595–598. Coruh B et al. Does religious activity improve health out-33. comes? A critical review of the recent literature. Explore (NY), 2005, 1(3):186–191. Yoshino S, Shoji H, Komagamine M. Full flexion after total 34. knee replacement in rheumatoid arthritis. International Orthopaedics, 1990, 14(1):13–16. Szabó G et al. A prospective comparative analysis of mobil-35. ity in osteoarthritic knees. Journal of Bone and Joint Surgery. British Volume, 2000, 82(8):1167–1169. Rønningen A, Kjeken I .Effect of an intensive hand exercise 36. programme in patients with rheumatoid arthritis. Scandina- vian Journal of Occupational Therapy, 2008, 7:1–11. De Jong Z et al. Long term high intensity exercise and dam-37. age of small joints in rheumatoid arthritis. Annals of the Rheumatic Diseases, 2004, 63(11):1399–1405. EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 958 Clinico-pathological profile of acute promyelocytic leukaemia at Al-Amal oncology-haematology centre, Qatar F.A. Ibrahim,1 M.A.Yassin,2 H.R. El-Ayoubi,2 I.A. Alhijji,2 A.S. Albinali,1 S.M. Almansour 1 and F.M. Qafoud 1 ABSTRACT This cases series describes the profile of adult patients with acute promyelocytic leukaemia (APL) at a referral hospital in Qatar. Of 34 acute myeloid leukaemia (AML) cases diagnosed, 11 (32%) were classified as APL. Disseminated intravascular coagulation was common at presentation (91%). Severe thrombocytopenia was seen in 73%, leukocytosis in 55% and severe anaemia in 45%. Only 2 patients were of the classic hypergranular type. In the remaining 9 patients, 3 morphological subtypes were recognized: microgranular variant (6 patients), hyperbasophilic (2 patients) and regular nuclear outline M3r (1 patient). Translocation t(15;17) was detected in 63% of cases. APL constitutes a high proportion of AML cases in Qatar, with considerable morphological heterogeneity and a predominance of APL variants with unfavourable presenting features. 1Department of Laboratory Medicine and Pathology; 2Department of Haematology and Bone Marrow Transplant, Al-Amal Hospital, Qatar (Correspondence to M.A. Yassin: yassin160@yahoo.co.uk). Received: 13/01/09; accepted: 12/03/09 رطق ةلودب تايومدلاو - مارولأا ملعل لملأا زكرم في دالحا ةّيوقنلا فئلاسلا ضاضيبلا ضيرلما - يريسرلا مسترلما دوفاك ةمطاف ،يروصنلما ةراس ،ليعنبلا ءماسأ ،يجلحا دحمأ ميهاربإ ،بيويلأا يعفار يدانه ،ينساي مئادلا دبع دممح ،ميهاربإ لايرف نمو .رطق ةلود في يعجرم ىفشتسم في دالحا ةّيوقنلا فئلاسلا ضاضيباب ينباصلما ينغلابلا ضىرلما مسترم تلاالحا نم ةلسلسلا هذه فصت :ةصلالخا ةيعولأا لخاد رثتنلما رثختلا ناكو .دالحا ةيوقنلا فئلاسلا ضاضيباب نوباصم منهأ لىع )%32( ًاضيرم 11 فِّنـُص ،داح يوقن ضاضيباب ًاضيرم 34 ينب كانه ناكو .مهنم %45 في ميخولا مدلا رقفو ،%55 في ضيبلا تايركلا ةرثكو ،%73 في تاحْيَف ُّصلا في ةديدش ةلق تدهوشو .)%91( ثودلحا عئاش تابيبلحا قيقدلا ب ْ َّرضلا :يه ةيعرف ةيجولوفروم طمانأ ةثلاث ينقابلا ةعستلا ضىرلما في تفشُتكاو .دوهعلما بُّبحتلا طرف طمنب ناباصم طقف ناضيرم .تلاالحا نم %63 في )17؛15( ءافزإ فشتكاو .)دحاو ضيرم( M3r يوونلا راطلإا ماظتناو ،)ناضيرم( تايولقلاب غابطصلاا طرف ،)ضىرم 6( بوضر ةبَلَغ ظوحلم يجولوفروم رياغت عم ،رطق في دالحا يوقنلا ضاضيبلاا تلااح نم ةيربك ةبسن دالحا ةيوقنلا فئلاسلا ضاضيبا لكشيو .ةبوغرم يرغ حملام دوجو عم دالحا ةيوقنلا فئلاسلا ضاضيبا Profil anatomoclinique de la leucémie promyélocytaire aiguë établi au centre d’oncologie et d’hématologie d’Al-Amal (Qatar) RÉSUMÉ Cette série de cas décrit le profil des patients adultes atteints de leucémie promyélocytaire aiguë et traités dans un hôpital de recours au Qatar. Sur 34 cas diagnostiqués de leucémie myéloïde aiguë, 11 d’entre eux (soit 32 %) ont été classés comme leucémie promyélocytaire aiguë. Une coagulation intravasculaire disséminée a été fréquemment constatée au moment de la présentation du patient (91 %). Une thrombopénie sévère a été observée dans 73 % des cas, une leucocytose dans 55 % des cas et une anémie sévère dans 45 % des cas. Seuls deux patients présentaient une forme classique hypergranulaire. Chez les neuf autres patients, trois sous-types morphologiques ont été identifiés : une variante microgranulaire (six patients), une variante hyperbasophile (deux patients) et une variante M3 avec contour nucléaire régulier (un patient). Une translocation t (15 ; 17) a été détectée dans 63 % des cas. La leucémie promyélocytaire aiguë constitue une proportion importante des cas de leucémie myéloïde aiguë au Qatar, et se caractérise par une hétérogénéité morphologique importante et une prédominance des variantes de leucémie promyélocytaire aiguë avec signes d’appel défavorables. طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 959 Introduction Acute promyelocytic leukaemia (APL)  is a subtype of acute myeloid leukaemia  (AML) that has unique morphological,  cytogenetic and molecular features [1].  These  include  a  potentially  devastat- ing coagulopathy, which carries a high  risk of mortality  [2],  and sensitivity  to  retinoid-differentiating agents including  all-trans-retinoic acid (ATRA) [3]. The  use of ATRA,  together with  conven- tional chemotherapy, has  substantially  increased  the number of patients who  can be  cured of APL. The  impressive  improvement  in  treatment outcomes  observed  with  ATRA  is,  at  least  in  part, due  to  the rapid correction of  the  coagulopathy, implying that the admin- istration of  the drug must occur at  the  beginning of treatment [4]. Thus, rapid  diagnosis of APL is critical for treatment  decisions  since ATRA must be given  promptly. Unequivocal diagnosis of APL can  only be established by cytogenetic or  molecular  genetic  studies. However,  these highly specialized techniques take  some days to be performed and are usu- ally restricted to specialized haematology  laboratories. Thus from a practical point  of view morphology, cytochemistry and  immunophenotyping are still important  tools for rapid recognition of APL. The  great majority  of  APL  cases  with  the  t(15;17)  translocation  are  of  the  classic hypergranular M3  type  and  the microgranular (hypogranular)  variant M3v, as described in the French- American-British  (FAB) classification  [5,6]. Other morphological subtypes of  APL have been  reported:  for example,  hyperbasophilic  [7], M1- and M2-like  [8]  and  regular  nuclear  outline M3r  [1]. While in the majority of classic APL  cases  the  immunophenotypic pattern  is distinctive,  the  features of  the  vari- ant cases are more heterogeneous and  nonspecific  [9]. This case  series  is  the  first  report  from Qatar addressing  the  clinical-pathological profile of patients  with APL. Methods Patients All adult patients admitted  to Al-Amal  Hospital,  the haematology/oncology  centre of Hamad Medical Corporate  in Qatar, and diagnosed with APL over  the period  January 2006  to May 2008  were  included. Diagnosis of AML was  established by combined morphologi- cal examination,  immunophenotyping  and cytogenetic studies. Clinical evaluation The medical  records  of  the  patients  were  reviewed  for  clinical manifesta- tions of AML and results of  laboratory  tests. Complete and differential blood  counts, full baseline biochemical profile,  serum electrolytes, urea and creatinine,  liver profile and coagulation screening  profiles were  done  on  all  patients  at  presentation. Morphologic evaluation In 10 patients both peripheral  smear  and  bone marrow  aspirates  (stained  with Wright stain) were available, while  in 1 patient diagnosis was based on the  peripheral  blood flow cytometry  im- munophenotyping as  the patient died  before there was time to perform bone- marrow aspiration. Differential  counts  of 100 cells on the peripheral smear and  500 cells on  the bone marrow  smear  were performed. Final morphological  subtype was determined by consensus  between 2 haematopathologists. Cytogenetic analysis Cytogenetic  analysis was  performed  on 10 of  the 11 patients. Karyotypes  were classified according to the Interna- tional System for Human Cytogenetic  Nomenclature (ISCN) [10]. Flow cytometry immunophenotypic methods Bone marrow  samples  in 10 patients  and peripheral blood  in 1 patient were  studied by flow cytometry using a panel  of 19 monoclonal  antibodies directly  conjugated with fluorochrome. A 3-colour  flow cytometry  analy- sis  was  performed  on  bone marrow  aspirate/peripheral blood samples col- lected in EDTA tubes. After incubation  of cells with monoclonal antibodies for  15 minutes  at  room  temperature,  the  red blood cells were  lysed with cyclic  amine reagent (VersaLyse) for 10 min- utes,  followed by washing  steps using  phosphate-buffered saline solution. The  cells were  resuspended  in phosphate- buffered saline. The panel of antibodies included the  following mouse monoclonal antibod- ies specific for: CD45 (energy-coupled  dye  [ECD]),  CD34  (fluorescein  isothiocyanate  conjugated  [FITC]),  CD33  (phycoerythrin  [PE]), CD64  (R-phycoerythrin  covalently  link  to  cyanin 5  [PC5]),CD13 (PE), CD14  (FITC),  CD19  (FITC  or  ECD),  CD117 (PE), CD10 (PE), CD7 (PE),  CD3 (ECD), CD15 (FITC), HLA-DR  (FITC), CD5 (FITC), TdT (FITC),  cCD22 (PE), cMPO (FITC), cCD79a  (PE) and CD20 (FITC). All antibodies  were obtained  from Beckman Coulter.  For  each  antibody,  negative  staining  levels were  set by comparison with an  isotype-matched control. All  samples  were  analysed  with  a  FC500  flow  cytometer  (Beckman  Coulter) equipped with an argon-ion  laser with a wavelength of 488 nm, by  collecting at  least 10 000 ungated  list- mode events per  tube. The cytometer  was  set  up  using  standard  operation  procedures,  and  quality  control  was  performed  using  the manufacturer’s  methods for calibration and compensa- tion. Gating was based on the CD45 ex- pression and side scatter. CXP software  (Beckman Coulter) was used  for data  acquisition and analysis.  Cells were classified as positive for a  given marker when the expression (fluo- rescence intensity) was greater than that  of a negative (isotypic) control. Intensity  of fluorescence was semi-quantitatively  estimated as dim, moderate and high,  based on the population position on the  fluorescence scale. Antigen expression  EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 960 was considered  to be homogenous  if  the  distribution of  the  cells  occupied up  to  1  logarithmic decade on  the  scale of  the  fluorescence  intensity,  otherwise  it was  considered heterogeneous [11]. Results Of  the  34  patients  diagnosed  as  AML  between January 2006 and May 2008, 11  (32%) were diagnosed with APL and were  included in the study. Of these, 7 were men  (1 Qatari  and 6 non-Qatari)  and 4 were  women (1 Qatari and 3 non-Qatari). Their  ages ranged from 16 years to 46 years,with  a median age at presentation of 34.5 years. Clinical and laboratory features Of  these   11   pat ients ,   10   (91%)  p r e s e n t e d   w i t h   d i s s em i n a t e d  intravascular coagulation (DIC) according  to the ISCN criteria [12] (Table 1). Gum  bleeding,  ecchymosis  and  symptomatic  anaemia were presenting  symptoms  in 7  (63%) patients,  and haemoptysis, pulmo- nary haemorrhage and infection were found  on diagnosis  in 2 (18%) patients; only 1  patient (9%) had epistaxis on presentation  and 1 presented with abdominal pain. None  of the patients had organomegaly. Severe  thrombocytopenia  (<  30  ×  109/L) was seen in 8 patients (73%). Leu- kocytosis (> 10 × 109/L) was noted  in 6  patients (55%) and 5 (45%) were severely  anaemic at presentation. Pancytopenia was  noted in 4 patients (36%). In all, leukaemia  promyelocytes were present in the periph- eral blood (Table 1). Only 2 patients (18%) were classified  as M3c, classic hypergranular APL, while  9 patients (82%) had APL variants (Table  2).  The classic APL patients (patient nos.  1 and 2) presented with pancytopenia. The  majority of the promyelocytes in the mar- row were hypergranular; however,  some  hypogranular and agranular promyelocytes  were also noted. Nuclear folding or convo- lution were  frequently noted. Occasional  faggot cells were seen in 1 patient (patient  no. 1) (Figure 1A). Ta bl e 1 H ae m at ol og ic al a nd c oa gu la ti on p ro fil e of a cu te p ro m ye lo cy ti c le uk ae m ia (A PL ) c as es C as e no . Se x A ge (y ea rs ) H b (g /d L) Pl at . (× 1 0 9 /L ) W BC (× 1 0 9 /L ) Pr om ye lo cy te s in pe ri ph er al s m ea r (% ) PT a (s ) aP TT b (s ) Fi br in og en c (g /d L) D -d im er (l at ex ag gl ut in at io n) (μ g/ L) d D -d im er (im m un ot ur bi di m et ri c) (μ g/ L) e LD H (U /L ) 1 M 29 11 .2 11 0 .8 10 16 .8 31 .4 1.1 6 ≥ 4 < 8 n/ d 53 6 2 M 41 2. 1 13 1.7 90 19 .4 32 .0 2. 90 n/ d 59 5 26 7 3 M 46 12 .9 23 75 .7 96 16 .2 31 .5 2. 39 > 8 n/ d n/ d 4 F 22 12 .7 16 23 .0 80 20 .0 33 .6 0 .7 2 n/ d n/ d 13 43 5 M 36 6. 2 28 11 3. 6 96 19 .0 30 .0 4. 0 0 > 8 n/ d 24 30 6 F 38 14 .7 22 23 3. 0 92 14 .0 24 .0 1.7 0 n/ d 31 82 n/ d 7 F 30 8. 9 73 39 .0 91 18 .6 28 .0 0 .6 0 > 8 n/ d 72 6 8 M 39 12 .8 88 7.8 61 17 .0 24 .0 0 .8 7 > 8 16 38 44 3 9 M 16 5. 5 11 2. 0 65 14 .9 26 .2 1.6 0 > 8 n/ d 66 6 10 F 40 3. 5 11 3. 0 66 11 .6 26 .6 0 .2 0 n/ d 12 83 48 5 11 M 41 7.1 45 14 .6 39 10 .6 34 .8 7.3 3 > 8 n/ d 69 9 a P T = pr ot hr om bi n tim e, n or m al ra ng e: 11 .7 –1 4. 5 s; b a PT T = ac tiv e pa rt ia l t hr om bo pl as tin ti m e, n or m al ra ng e: 2 5– 33 .6 s. c F ib rin og en , n or m al ra ng e: 2 –4 g /d L; d D -d im er (l at ex a gg lu tin at io n) , n or m al < 0 .5 μ g/ m L. e D -d im er (i m m un ot ur bd im et ric ), no rm al < 3 26 μ g/ L. H b = ha em og lo bi n; P la t. = pl at el et s; W BC = w hi te b lo od ce lls ; L D H = la ct at e de hy dr og en as e. M = m al e; F = fe m al e; n /d = n ot d et er m in ed . طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 961 In  the  9  APL  variant  patients,  3  morphological  categories were  recog- nized. Microgranular variants M3v,  as  described by  the French–American– British Co-operative Group  [6] were  recognized  in  6  patients,  with more  than 50% of  cells having agranular or  hypogranular  cytoplasm, with promi- nent nuclear folding. Bilobed cells were  noted  in 4 (patient nos. 3, 4, 6 and 8).  Faggot cells were seen in 6 patients and  were frequent in 1 of them (patient no.  3) (Figure 1B). In patients no. 9 and 10  the predominant cells were small in size  with high nuclear–cytoplasmic  ratio,  deeply basophilic cytoplasm and show- ing prominent cytoplasmic blebs and  projections, mostly agranular. Nuclear  folding was also prominent with some  bilobed cells and  faggot cells noted  in  patient no. 9. These 2 cases correspond  to  the hyperbasophilic variant M3b of  McKenna et al. [7] (Figure 1C). In the  last patient (patient no. 11), a majority  of  the marrow promyelocytes  (86%)  had  regular,  round or oval nuclei, hy- pergranularity was frequent (40%) and  faggot  cells were occasionally  noted,  and APL was morphologically  similar  to  the M3r regular variant  reported by  Sainty et al. [1] (Figure 1D). Cytogenetic analysis Out of the 10 patients analysed, 5 were  positive  for  the  t(15;17)  translocation,  1  classic  case  (patient no.  2),  2 with  M3-variant (patients no. 4 and 5) and  2 with hyperbasophilic APL. Additional  cytogenetic abnormalities were noted  in 2 cases (patients no. 2 and 5) (Table  2). In 3 cases the translocation was not  detected and reported as normal karyo- type. The marrow failed to culture  in 2  patients. Table 2 Bone marrow differential and cytogenetic findings in 11 acute promyelocytic leukaemia (APL) cases Case no. Promy- elocyte Cytoplasm Nucleus Cytogenetic APL subtypeHypergranular Hypo/ agranular Regular Folded Bilobed % % % % % % 1 90 65 35 14 69 17 Failed M3c 2 85 68 32 59 34 7 46,XY,der(4),t(4;7) (q31;q22),t(15;17) (q22;q21),del(16)(q22) [12] M3c 3 96 3 97 12 57 31 46,XY [16] M3v 4 90 14 86 20 59 21 46,XX,t(15,17)(q22;q12) [15] M3v 5 92 38 62 22 68 10 46,XY,del(3)(q12q23),t(15;17)(q22;q12) [26] M3v 6a n/d 0 100 10 46 44 n/d M3v 7 92 21 79 29 66 5 46,XX [20] M3v 8 83 29 71 7 70 23 Failed M3v 9 78 11 89 15 77 8 46,XY,t(15;17) [14] M3b 10 75 20 80 5 90 5 46,XX,t(15;17)(q22;q11.2) [11] M3b 11 82 35 65 86 12 2 46,XY [27] M3r For descriptive purposes, morphologic variations were grouped into different subtypes: M3c = hypergranular classic; M3v = microgranular variant; M3b = hyperbasophilic variant; M3r = microgranular regular. Within the M3v group cases are arranged chronologically. n/d = not determined. aDone on peripheral blood. Figure 1 Bone-marrow smears of acute promyelocytic leukaemia (APL) cases. (A) Classic hypergranular APL in which hypergranular cells predominate and some cells show irregular nuclear outline. (B) Microgranular APL variant, where almost all the cells look agranular with prominent nuclear convolution. (C) Hyperbasophilic APL, where the majority of cells show hyperbasophilic cytoplasm with cytoplasmic projections. (D) M3r APL where most of the cells have round nuclei, both hypo- and hypergranular cells are seen A B C D EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 962 Immunophenotyping The expression pattern of  the  selected  markers  studied  by  flow  cytometry  analysis  and  their  correlation with  the  morphological  subtypes  are  summa- rized in Table 3. On  side-scatter  versus CD45  leu- kaemic cells in the classic hypergranular  patient displayed primarily high  side- scatter (Figure 2A). In contrast, 8 of the  9 M3-variant patients (89%),  the cells  displayed less side-scatter and generally  fell in the area closer to the blast region  (Figure 2B). Myeloperoxidase (MPO)  protein and CD33 were consistently ex- pressed in all  the cases (100%) regard- less of  the  subtype. CD33 expression  was homogenous in 10 patients (91%)  and heterogeneous in 1 patient (patient  no. 5). CD13 was positive in 10 patients  (91%) and CD117 in 9 patients (82%).  CD13 expression was heterogeneous  in all  the positive patients, CD34 was  positive  in 4 patients  (36%);  all were  of  the variant  type (patients no. 3, 6, 9  and 10). Each of CD14 and CD15 was  positive in 2 patients (18%). CD64 was  positive  in 6 out of  the 8 cases where  it was done (75%); 1 of  them (patient  no. 2) coexpressed CD14 as well. The  human  leukocyte  antigen DR (HLA- DR) molecule was expressed  in 1 case  (patient no. 9) and CD7 in 1 case (pa- tient no. 10). TdT, CD3, CD19, CD10, CD5,  CD20, CD79 and CD22 were negative  in all the cases. Discussion The methods  used  for  the  diagnosis  and classification of  acute  leukaemias  include morphology,  cytochemistry,  immunophenotyping, cytogenetic and  molecular genetics [13]. For treatment  purposes, once the separation between  myeloid and  lymphoid  leukaemias has  been solved, the most important defini- tion  is  if  a given case of acute myeloid  leukaemia  could  be  sub-classified  as A B Fi gu re 2 F lo w -c yt om et ry fi nd in gs o f a cu te p ro m ye lo cy ti c le uk ae m ia (A PL ) c as es . ( A ) A PL M 3 cl as si c le uk ae m ic c el ls d is pl ay h ig h si de -s ca tt er ; e xp re ss io n pr ofi le (c M PO + C D 11 7+ C D 33 + C D 13 + C D 34 – C D 14 – C D 15 – H LA D R– ). (B ) A PL M 3 va ri an t l eu ka em ic c el ls d is pl ay lo w s id e- sc at te r c om pa re d w it h th e hi gh s id e- sc at te r o f c la ss ic M 3; e xp re ss io n pr ofi le (c M PO + C D 11 7+ C D 34 + C D 33 + C D 13 + C D 7+ C D 14 – C D 15 – H LA D R– ). SS C = s id e- sc at te r; F IT C = fl uo re sc ei n is ot hi oc ya na te ; P E = ph yc oe ry th ri n; E C D = R -p hy co er yt hr in -t ex as re d- x. طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 963 Ta bl e 3 Im m un op he no ty pi c ch ar ac te ri st ic s of th e ac ut e pr om ye lo cy ti c le uk ae m ia (A PL ) c as es C as e no A PL s ub ty pe Si de -s ca tt er C el l m ar ke rs M PO C D 33 C D 13 C D 11 7 C D 34 H LA -D R C D 14 C D 15 C D 64 1 M 3c H ig h M od H ig h Lo w H ig h –v e –v e –v e –v e –v e 2 M 3c H ig h M od H ig h M od –v e –v e –v e M od Lo w H ig h 3 M 3v Lo w H ig h H ig h H ig h M od Lo w –v e –v e –v e n/ d 4 M 3v Lo w H ig h H ig h M od H ig h –v e –v e –v e –v e n/ d 5a M 3v H ig h H ig h M od H ig h –v e –v e –v e Lo w –v e n/ d 6 M 3v Lo w M od H ig h H ig h M od M od –v e –v e –v e Lo w 7 M 3v Lo w M od H ig h Lo w Lo w –v e –v e –v e –v e Lo w 8 M 3v Lo w M od H ig h M od M od –v e –v e –v e –v e Lo w 9 M 3b Lo w H ig h H ig h Lo w H ig h M od Lo w –v e –v e –v e 10 b M 3b Lo w H ig h H ig h M od M od M od –v e –v e –v e M od 11 M 3r Lo w H ig h H ig h –v e Lo w –v e –v e –v e Lo w Lo w Lo w = lo w /d im e xp re ss io n; M od = m od er at e ex pr es si on ; H ig h = hi gh /b rig ht e xp re ss io n. a C as e w ith h et er og en eo us C D 33 e xp re ss io n; b C as e w ith C D 7 ex pr es si on . A PL su bt yp es : M 3c = h yp er gr an ul ar c la ss ic ; M 3v = m ic ro gr an ul ar v ar ia nt ; M 3b = h yp er ba so ph ili c va ria nt ; M 3r = m ic ro gr an ul ar re gu la r. n/ d = no t d et er m in ed . APL,  since  this  subtype benefits  from  treat- ment with ATRA. Although the use of morphology for initial  diagnosis  sounds appealing,  there  are  some  problems when  only  this method  is  used.  In  the  recovery phase  from acute agranulo- cytosis,  for  example,  the bone marrow may  be  replenished with promyelocytes, display- ing  a  picture  similar  to  classic  APL  [14].  Microgranular APL may  also  be  confused  with other  subtypes of AML, mainly  acute  monocytic  leukaemias  [15].  Furthermore  some groups have described other morpho- logical  subtypes of APL  in  addition  to  the  classic and microgranular forms—such as the  hyperbasophilic,  the M1-like and M2-like— that  display  a morphological  picture  quite  different  from classic APL and  are difficult  to distinguish  from M7, M1 and M2 AML  respectively [7,8]. Although  our  series  included  a  small  number of cases,  it seems that APL is a com- mon  subtype  (32%)  among AML cases  in  Qatar, higher than the frequency of 5%–10%  reported  from other  countries  [16,17]  and  the  frequency of 17%  from Saudi Arabia  in  a  series  including both Saudi  and nonSaudi  patients [18] and 10% in Omani patients [19].  A higher frequency of APL has been noted in  certain ethnic groups such as Latin American  and Spanish populations, which have a  sig- nificantly higher frequency of the PML/RARα gene [caused by  fusion of  the promyelocytic  leukaemia (PML) gene and  the gene encod- ing the retinoic acid receptor-α (RARα), which  may  suggest  a possible  influence of  genetic  and/or  environmental  factors  [20].  In our  series 82% of patients were nonQatari,  and  were of  several different nationalities, which  makes  the possibility of  a  link  to  a  genetic  factor rather unlikely and this high frequency  needs  to be verified on a  larger  series over a  longer period.  The other interesting finding was the pre- dominance of the variant APL which we found  in 82% of our patients in contrast to the well- known notion that variant APL cases account  for only about 15%–20% of all APL in adults  [1,2]. This would explain  the high  frequency  of high white blood cell count (55%),  severe  thrombo cytopenia  (73%) and DIC (91%),  features  reported  to  be  more  commonly  EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 964 encountered  in variant APL cases and  all of which have been shown to be unfa- vourable prognostic factors [21]. Immunophenotypic  studies  have  unveiled a characteristic pattern in clas- sic APL compared with other subtypes  of AML, which  includes expression of  MPO, CD33, CD13 and lack of HLA- DR  and  CD34  [9]. This  distinctive  pattern was also displayed by  the pro- myelocytes  in our patients with classic  APL morphology. On  the other hand  the pattern in variant APL has been re- ported to be more heterogeneous, with  a  higher  percentage  of  cells  express- ing  the T-cell  antigen CD2,  stem cell  marker CD34 [9], HLA-DR [21] and  CD56 [1]. This heterogeneity was dem- onstrated in our variant APL patients as  4 expressed CD34 (2 M3v and 2 hyper- basophilic subtypes). CD34 expression  was  suggested  as  a  reliable marker  to  distinguish between M3v  and  classic  APL [9]. HLA-DR was expressed  in 1  patient with hyperbasophilic APL. Fenu  et al. reported 2 cases of hyperbasophilic  APL variant with positive expression of  CD33, CD13, HLA-DR, CD34, CD2  and CD9 [22]. CD14  was  expressed  in  1 M3v  patient  and CD64  in 5 out of  the 6  variant patients tested (3 M3v, 1 hyper- basophilic and 1 M3r  types). Both of  these  are monocytic markers which  add to the difficulty in recognizing M3v  APL as  it might  simulate monocytic  leukaemia, not only morphologically  but immunophenotypically as well. The  t(15;17)  translocation  yield- ing  the  PML-RARα   fusion  gene  is  the diagnostic hallmark of APL. This  translocation can be detected by con- ventional  cytogenetic  techniques  in  about 90% of cases.  In  the majority of  cases  lacking  the  t(15;17)  transloca- tion, PML-RARα  gene  rearrangement  can be detected by molecular analysis  created by  insertion or more complex  rearrangements. Such mechanisms oc- cur in approximately 4% and 2% of APL  cases respectively. In a minority of APL  cases  other  rearrangements  of  genes  were reported, whereby the RARα gene  is  fused  to  the promyelo cytic  leukae- mia zinc finger (PLZF) gene  in about  0.8% of  cases  and  less  commonly  to  nucleophosmin (NPM), nuclear matrix  associated  (NuMA)  and  signal  trans- ducer and activator of  transcription 5b  (STAT5b)  genes  [23]. Many  of  the  cases with  the  latter molecular defects  show clinical and cytological differences  from  the M3/M3v subtypes of AML,  and  the  designation  “M3-like”  has  been suggested [24].  In common with  PML-RARa-associated APL,  patients  with  fusion genes  involving NPM and NuMA appear to be sensitive to ATRA.  In contrast, APL with PLZF/RARa or  STAT5b/RARa rearrangements are typ- ified by a lack of response to retinoids.  In  this  series  classic  t(15;17) was  demonstrated  in  5  out  of  the  8 APL  variant cases analysed (63%). Analysis  of 3 cases revealed a normal karyotype.  Whether the lower than expected rate of  t(15;17) expression  is another unique  feature of APL  in Qatar,  like  the mor- phological diversity, or whether  it  is  a  chance finding needs to be verified on a  larger number of cases.  In conclusion, APL seems to consti- tute a major proportion of AML cases  in Qatar, with significant morphological  heterogeneity  and  predominance  of  variant APL cases with unfavourable  presenting features. The impact of these  findings on patient outcomes  is being  evaluated and will be reported. The  heterogeneous  pattern  and  limitations of flow cytometry  in diag- nosing APL variants,  as  confirmed  in  this  study, emphasize  the  importance  of a careful morphological evaluation  with good awareness of  the cytologi- cal spectrum of APL, especially  in this  country  where  t(15;17)  is  not  fre- quently detected. The study also highlighted  the  im- portance of more sensitive approaches  to APL diagnosis,  such  as  interphase  FISH  [fluorescent  in  situ  hybridiza- tion]  reverse  transcriptase (RT)-PCR  molecular  analysis  or  the  interesting  option  for  a  rapid  and  accurate diag- nosis of APL by immunostaining assays  with anti-PML antibodies to detect the  characteristic microparticulate nuclear  pattern of  the PML protein. The  latter  may be an important tool for identifying  cytologically atypical APL cases and  in  monitoring patients for minimal residu- al disease, particularly in our institution  where genetic tests for acute leukaemias  are not yet routinely available.  Acknowledgements We  would  like  to  express  our  deep  thanks to following colleagues for help in  performing this study: Halima El-Omri,  Ruba T. Yassin, Reeham N. Hasssan,  Ussama Al-Homsi, Zainab M. Fawzi,  Munir Jalis, Kamal R. Abboudi.  Sainty D et al. A new morphologic classification system for acute 1. promyelocytic leukemia distinguishes cases with underlying PLZF-RARa gene rearrangements. Blood, 2000, 96:1287–1296. Avvisati G, Lo Coco F, Mandelli F. Acute promyelocytic leuke-2. mia: clinical and morphologic features and prognostic factors. Seminars in Hematology, 2001, 38:4–12. References Fenaux P, Chromienne C, Degos L. All-3. trans retinoic acid and chemotherapy in the treatment of acute promyelocytic leuke- mia. Seminars in Hematology, 2001, 38:13–25. Barbui T, Finazzi G, Falanga A. The impact of all-4. trans-retinoic on the coagulopathy of acute promyelocytic leukemia. Blood, 1998, 91:3093–3102. طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 965 Bennett JM et al. Proposals for the classification of the acute 5. leukemias. FAB Cooperative Group. British Journal of Haema- tology, 1976, 33:451–458. Bennett JM et al. A variant form of hypergranular promy-6. elocytic leukaemia M3. British Journal of Haematology, 1980, 44:169–170. McKenna RW et al. Acute promyelocytic leukaemia: a study of 7. 39 cases with identification of a hyperbasophilic microgranular variant. British Journal of Haematology, 1982, 50:201–214. Neam PB et al. Morphology of acute promyelocytic leukaemia 8. with cytogenetic or molecular evidence for the diagnosis: characterization of additional microgranular variants. American Journal of Hematology, 1997, 56:131–142. Paietta E et al. A surrogate marker profile for PML/RAR. Ex-9. pressing acute promyelocytic leukemia and the association of immunophenotypic markers with morphologic and molecular subtypes. Cytometry Part B Clinical Cytometry. 2004, 59B:1–9. Mittleman F ed. 10. ISCN 1995: An international system for human cytogenetic nomenclature. Basel, Karger, 1995. Gorczyca W, ed. 11. Flow cytometry in neoplastic hematology: morphologic–immunophenotypic correlation. London, Taylor and Francis, 2006:1–20. Taylor FB Jr et al. Towards definition, clinical and laboratory 12. criteria, and a scoring system for disseminated intravascular co- agulation. Thrombosis and Haemostasis, 2001, 86:1327–1330. McKenna RW. Multifaceted approach to the diagnosis and 13. classification of acute leukaemias. Clinical Chemistry, 2000, 46:1252–1259. Rizzatti EG et al. Expression of CD117 and CD11b in bone mar-14. row can differentiate acute promyelocytic leukaemia from recovering benign myeloid proliferation. American Journal of Clinical Pathology, 2002, 118:31–37. Nagendra S et al. Leukemias resembling acute promyelocytic 15. leukemia, microgranular variant. American Journal of Clinical Pathology, 2002, 117:651–657. Ghosh S et al. Haematologic and immunophenotypic profile 16. of acute myeloid leukemia: an experience of Tata Memorial Hospital. Indian Journal of Cancer, 2003, 40 2:71–76. Arber DA et al. Prognostic impact of acute myeloid leukemia 17. classification. Importance of detection of recurring cytogenetic abnormalities and multilineage dysplasia on survival. American Journal of Clinical Pathology, 2003, 119:672–680. Harakati MSE et al. Adult acute myeloblastic leukemia: experi-18. ence at King Khalid University Hospital. Annals of Saudi Medi- cine, 1998, 18:221–225. Udayakumar AM et al. Cytogenetic, morphologic and immu-19. nophenotypic patterns in Omani patients with de novo acute myeloid leukemia. Cancer Genetics and Cytogenetics, 2007, 177:89–94. Ribeiro RC, Rego E. Management of APL in developing coun-20. tries: epidemiology, challenges and opportunities for interna- tional collaboration. Hematology, 2006, 1:162–168. Olga F et al. APL: a classic tale of bench to beside. In: Karp J. 21. Acute myelogenous leukemia. Totowa, New Jersey, Humana Press, 2007:193–231. Fenu S et al. P21. Acute promyelocytic leukemia: two cases with 22. cryptic aspects in morphological diagnosis. Haematologica Reports, 2005, 1(7):46. Grimwade D et al. Characterization of acute promyelocytic 23. leukemia cases lacking the classic t(15;17): results of the Euro- pean Working Party. Blood, 2000, 96(4):1297–1308. Acute leukaemia: immunophenotypic, cytogenetic and 24. molecular genetic analysis in the classification of acute leukae- mia—the EGIL, MIC, MIC-M and WHO classifications. In: Bain BJ, ed. Leukaemia diagnosis, 3rd ed. Oxford, Blackwell Publish- ing, 2003:57–144. EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 966 Comparison of erythrocyte indices to differentiate between iron deficiency and alpha-thalassaemias in children with microcytosis and/or hypochromia H. Narchi 1 and R.B. Basak 2 ABSTRACT Erythrocyte indices used to differentiate between iron deficiency anaemia (IDA) and thalassaemias have been studied mainly in countries with a high prevalence of IDA or beta-thalassaemias. This study was carried out in the United Arab Emirates where alpha-thalassaemias are prevalent. We studied the predictive value and Youden index of several indices in 85 children aged 6 months to 12 years with microcytosis and/or hypochromia, with or without anaemia (determined by serum measurement of iron or therapeutic iron trial). The best discriminatory indices for detecting IDA versus alpha-thalassaemias were a Green –King index > 65 (correctly identified 75.3% of children, Youden index 44.3) and red cell distribution width > 14% (correctly identified 57.8%, Youden index 38.1). 1Department of Paediatrics, Faculty of Medicine and Health Sciences, Al Ain, United Arab Emirates (Correspondence to H. Narchi: hassib.narchi@ uaeu.ac.ae). 2Department of Paediatrics, Al Ain Hospital, Al Ain, United Arab Emirates. Received: 21/01/09; accepted: 17/03/09 تايركلا رغصب ينباصلما لافطلأا ىدل ةيميسلااث-افللأا ينبو ديدلحا زَوَع ينب قيرفتلل ءارملحا تايركلا بساَنَم ينب ةنراقم ماهيلك وأ غابصلا صقنب وأ رملحا كازاب ىَنتر ،شيران بيسح يتلا نادلبلا في ماَّيسلاو ةيميسلااثلاو ديدلحا زوع نع مجانلا مدلا رقف ينب قيرفتلا في ةمدختسلما ءارملحا تايركلا بساَنَمـل ةسارد تيرجأ :ةصلالخا ثيح ةدحتلما ةيبرعلا تاراملإا ةلود في ةساردلا هذه تيرجأ دقو .ةيميسلااث-اتيبلاو ديدلحا زَوَع نع مجانلا مدلا رقف نم ةعفترم تلادعم ايهدل رهشأ ةتس ينب مهرماعأ حوارـتت ًلافط 85 في بسانَمـلا نم ددعل Youden ندوي بسْنَمو ةيؤبنتلا مَيقلا نوثحابلا سردو .ةيميسلااث-افللأا شرتنت مدلا رقفب ينباصم يرغ وأ ينباصم اوناك ءاوس ،ماهيلك وأ ءارملحا تايركلا غابص صقن وأ ءارملحا تايركلا رغصب ينباصم اوناك ،ةنس ةشرع ْيَتنثاو ةيميسلااث-افللأا نع ديدلحا زَوَعب مدلا رقف تزّيم يتلا بسانَمـلا لضفأ تناكو .)ديدلحاب ةيجلاعلا ةبرجتلاب وأ لصلما ديدح سايقب ص ِّخُش يذلا( يه ةحيحصلا هدودح( %14 رملحا تايركلاو )44.3 ندوي بسنم ،%75.3 يه ةحيحصلا هدودح( Green-King index، 65 غنك-نيرغ بسْنَم يه .)38.1 ندوي بسنم ،%57.8 Comparaison des indices érythrocytaires en vue de distinguer les carences en fer des alpha-thalassémies chez les enfants présentant une microcytose et/ou une hypochromie RÉSUMÉ L’utilisation des indices érythrocytaires pour distinguer les anémies ferriprives des thalassémies a été principalement étudiée dans les pays présentant une forte prévalence d’anémie ferriprive ou de bêta- thalassémies. Cette étude a été réalisée aux Émirats arabes unis où les alpha-thalassémies sont répandues. Nous avons étudié la valeur prédictive et l’indice de Youden de plusieurs indices chez 85 enfants âgés de six mois à 12 ans, présentant une microcytose et/ou une hypochromie, avec ou sans anémie (déterminée par la mesure sérique du fer ou l’essai thérapeutique du fer). Les meilleurs indices discriminatoires pour le dépistage de l’anémie ferriprive et des alpha-thalassémies et la distinction entre ces deux affections sont un indice de Green-King supérieur à 65 (75,3 % des enfants correctement dépistés, indice de Youden de 44,3) et un indice de distribution érythrocytaire supérieur à 14 % (57,8 % des enfants correctement dépistés, indice de Youden de 38,1). طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 967 Introduction In the United Arab Emirates (UAE) the  prevalence  of  anaemia  in  children  is  high (36.1%), primarily caused by  iron  depletion (36.0%) and beta-thalassae- mia (8.7%) and  in  the majority of  the  rest by alpha-thalassaemias (AT) [1–3].  As hypochromia and microcytosis are  almost universally present in this hetero- geneous group and cannot differentiate  between these conditions, several com- putational indices have been proposed,  using  the  readily available erythrocyte  results  in  a  routine  blood  count  [4].  These  include  the red blood cell distri- bution width (RDW), the RDW index,  the Mentzer index, the Green and King  index, the England and Fraser index, the  Shine and Lal  index and the Srivastava  index (described later). All these formulae have been tested  with different  cut-off values but none  has been  found  to be  sufficiently  spe- cific or  sensitive  to  obviate  the need  for confirmation of the diagnosis. Con- firmatory tests require additional blood  sampling from these children and incur  non-negligible costs [5,6] and a signifi- cant number of false-negative results for  the detection of beta-thalassaemia  trait  occur with these indices [7]. In addition,  most of these studies have been carried  out  in  adults  and  in  countries with  a  high prevalence of either iron deficiency  anaemia  (IDA) or beta-thalassaemia  haemoglobinopathies.  For any investigation with a known  sensitivity and specificity, the predictive  value (positive or negative) varies  ac- cording to the prevalence of the condi- tion  in  a particular population.  In  the  UAE, AT trait is much more common  (28%–56%)  than  beta-thalassaemia  trait  (8.7%) [8,9]. We  therefore evalu- ated these erythrocyte indices (with the  cut-off values defined in the literature)  to differentiate between  IDA and AT  haemoglobinopathies  in  our  popu- lation.  A  good  discriminating  index  would allow a  selective diagnostic  ap- proach  in children with hypochromia  and/or microcytosis in our population,  with potentially important cost savings  as well  as  a  reduced  need  for  blood  sampling. In addition, a high likelihood  of  haemoglobinopathy based on  the  indices would  alert  physicians  to  the  risks  of  empirical  iron  therapeutic  treatment in a child more likely to have  thalassaemia. Methods This was a  retrospective cohort  study  of  children with microcytosis, with or  without anaemia, managed  in  the pae- diatric  department of  a  large  general  hospital  in  the United Arab Emirates,  over a 4-month period from 1 February  2008 to 30 May 2008. Sample Based on  the published  incidence of  hypochromic microcytic  anaemia  of  35%  in  the  paediatric  population  in  the UAE (regardless of  the  cause),  a  minimum sample size of 85 was needed  to  have  a  5%  significance  level,  10%  precision and 95% confidence to detect  these abnormal indices (Epi-Info statisti- cal software, version 6.04). The  log book  in  the haematology  laboratory of our hospital was reviewed  to identify all children between the ages  of 6 months  and 12 years who were  diagnosed over  the study period with  microcytosis  and/or  hypochromia,  with or without anaemia. The criteria  for microcytosis were mean corpuscu- lar volume (MCV) < 70 fL in children  aged < 2  years,  < 75  fL  for  ages 2–6  years and < 77 fL at any other age. Cri- teria  for hypochromia were mean cell  haemoglobin concentration (MCH)  < 30 g/dL at age < 2 years and < 31g/ dL at any other age. Criteria  for anae- mia were haemoglobin  (Hb) < 10.5  g/dL at age < 2 years and < 11.5 g/dL  otherwise. The exclusion criteria were: children  aged < 6 months (in view of a low preva- lence  of  IDA  and  of  haematological  findings of thalassaemia); those already  diagnosed  prior  to  the  study  with  beta-thalassaemia,  sickle-cell  anaemia,  another  haemoglobinopathy,  iron  deficiency or  lead  intoxication;  those  already on  iron  therapy or  transfused  over the previous 4 months (regardless  of the cause); those with a known alter- native diagnosis (e.g.  aplastic anaemia,  haematological  malignancy);  those  with haemoglobinopathies other  than  AT (e.g. beta-thalassaemias,  sickle-cell  anaemias) as the focus of the study was  IDA versus AT;  those whose  investi- gations  for  iron deficiency or  for hae- moglobinopathy were not performed  or were not  available  for  review;  and  those with haemoglobinopathies with  a coexistent  iron deficiency,  to avoid a  “contamination” effect  in  interpreting  the results. Approval was granted and patient  consent was waived (as  this was a  ret- rospective  case-notes  study)  by  the  institutional  ethics  review committee  (Medical  District Human  Research  Ethics Committee protocol 07/127).  Data collection This  study was  a  retrospective  case- notes review of  the results of  investiga- tions already performed by the treating  physician  on  the  children  identified  from  the  laboratory  log  book. These  included complete red blood cell count  (RBC), Hb, MCV, MCH and RDW  levels obtained with a Coulter Counter  STKS (Coulter), serum iron and serum  iron binding  capacity  (TIBC) deter- mined  calorimetrically,  ferritin  by  an  automated enzyme-linked fluorescent  assay  (Vidas  Ferritin,  bioMérieux- Vitek)  and  high-performance  liquid  chromatography  analysis  for haemo- globinopathies  (Biorad Variant). The  remaining discrimination erythrocyte  indices were  calculated using  the  red  blood cell indices as defined in Table 1. The  data  collected  and  analysed  included: age,  sex,  erythrocyte  indices  values,  results of  investigations  for  iron  deficiency and haemoglobinopathy and  EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 968 response to a therapeutic iron trial when  applicable.  The outcome was the confirmation  of either iron deficiency (serum ferritin  < 10 ng/mL,  serum  iron < 50 µg/dL,  TIBC < 250 µg/dL or correction by iron  therapy) or AT haemoglobinopathy. As  no DNA  testing was  available  for  the  diagnosis of AT trait, this was diagnosed  when Hb electrophoresis was normal,  without any elevation of  fetal haemo- globin or haemoglobinalpha 2, with no  iron deficiency but with a family history  suggestive  of  asymptomatic  chronic  microcytosis without iron deficiency.  For the calculated erythrocyte indi- ces, we used the cut-off values common- ly described  in  the  literature  to  study  their  sensitivity,  specificity, correlation  and  positive  predictive  value  for  the  diagnosis of IDA versus AT. Analysis The chi-squared  test  (or Fisher  exact  test for small numbers) was used for the  study of univariate associations between  each index and the outcome. Continu- ous variables were compared with  the  Student 2-sample t-test. For each index,  we  calculated  its  specificity,  sensitiv- ity and positive and negative predictive  value for IDA, as well as the proportion  of IDA it correctly  identified, using the  statistical package Stata,  version 8. For  all  calculations,  statistical  significance  was defined as P-value < 0.05.  We  also  calculated  the  Youden  index  (sensitivity +  specificity – 100)  for  each of  these  indices  to diagnose  IDA, which, by taking into account both  sensitivity and specificity, measures their  validity  to diagnose  IDA versus AT. A  high Youden value  for  a  test makes  it  a  reliable discrimination  index  for  that  purpose. Results Clinical descriptive data A total of 85 children (50.6% females)  were  included  in  the  study. The mean  age was 5.9 years (range 0.5–15 years).  The majority were of Emirati nationality  (75.4%), with a small proportion  from  the Indian subcontinent (5.0%) and the  remainder from diverse Middle Eastern  countries. Iron  deficiency  was  the  cause  of  microcytosis  and/or hypochromia  in  18 children (21.2%).  It was diagnosed  by  serum measurement of  iron  in 13  (72.0%) and by  therapeutic  iron  trial  in  the remainder. All other 67 children  (78.8%) had AT trait. Haematological data and calculated indices The mean and standard deviation (SD)  values  obtained  from  the  full  blood  count  in  the whole sample were: RBC  5.2  (SD 0.6) × 1012 /L, Hb 10.5  (SD  1.9) g/dL, MCV 64 (SD 7.5) fL, MCH  20.4  (SD 3.0)  pg, MCHC 30.4  (SD  2.3) g/dL and RDW 16.3% (SD 4.2)%.  These values and  the calculated  in- dices were compared between children  with IDA and those with haemoglobin- opathies (Table 2). There was a statisti- cally significant difference between the  2  groups  in  all  parameters  except  for  the MCH, RDW index, Mentzer index,  MCH/RBC ratio and Srivastava index.  When we  analysed  these  indices  using  published  cut-off  values,  all  of  them discriminated very well between  the 2 groups, except  for  the Srivastava  index > 4.4, the RDW index > 220, and  the Shine–Lal index < 1530 (Table 3). The calculated  sensitivity,  specifi- city and positive and negative likelihood  ratios  for  these  tests  to diagnose  IDA  versus AT are  shown  in Table 4. The  calculated Youden index shows that the  best discriminatory indices, in descend- ing order were: Green–King index > 65,  RDW > 14%, England–Fraser index > 0,  RDW/RBC ratio > 3.3 and Srivastava  index > 4.4.  Discussion Differentiating  between  the  causes  of microcytosis  and hypochromia  in  children is important. Not only does an  appropriate diagnosis  allow adequate  management  and  appropriate  family  counselling  (in  case of haemoglobin- opathy) and have important prognostic  implications, but it also prevents unnec- essary iron therapy (as a therapeutic trial  for presumed  IDA)  for  children with  haemoglobinopathies who are already  at increased risk of iron toxicity. Although obtaining a detailed  fam- ily  and nutritional history  and history  of blood  loss are part of  the diagnostic  process  in such children, haematologi- cal parameters  are often measured  to  confirm the etiology.  If not diagnosed  in the early neonatal period by the pres- ence of Bart haemoglobin,  the diagno- sis of AT  in childhood requires DNA  analysis, which is not widely available. The routine blood count  is widely  available and inexpensive to perform. In  conjunction with microcytosis, differ- ent erythrocyte indices have been used  with  variable  success  to  differentiate  between  IDA  and  haemoglobinopa- thies [4]. None is entirely satisfactory in  Table 1 Cut-offs used for the different indices in this study Index Calculation Cut-off value Mentzer index MCV/RBC 13 Green–King formula MCV2 × RDW/(Hb × 100) 65 England–Fraser formula MCV – (5 × Hb) + RBC + 3.4 0 Shine–Lal formula MCV2 × MCH/100 1530 Srivastava formula MCH/RBC 4.4 RDW index MCV × RDW/RBC 220 MCV = mean cell volume; RBC = red blood cell count; Hb = haemoglobin; MCH = mean cell haemoglobin; RDW = red cell distribution width. طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 969 that aspect [5–7]. Most of these studies  have been carried out  in populations  with either a  low risk of haemoglobin- opathies (such as in some industrialized  countries) or a high prevalence of beta- thalassaemia haemoglobinopathies (in  developing countries) [10–13]. As with  any investigation with a known sensitiv- ity and specificity,  the predictive value  (positive or negative) varies according  to  the prevalence of  the condition  in a  particular population,  and  it  is  impor- tant to validate the value of these indices  in a population such as ours where there  is a high  incidence of AT haemoglob- inopathies. Because  the  focus of  this  study was to differentiate IDA from AT  haemoglobinopathy, we  deliberately  excluded  children  with  haemoglob- inopathies who had a  coexistent  iron  deficiency, in order to avoid a “contami- nation” effect in interpreting the results.  We cannot  therefore comment on  the  value of these indices or extrapolate the  study  findings  to  children who  have  such a combination. An  ideal  screening  test would have  a sensitivity of 100% with a specificity of  100%, while a totally useless one would  have 0%  sensitivity  and  specificity.  In  reality,  each  test  has  an  intermediate  value  for  sensitivity and  for  specificity,  with  a high value  for one being often  associated with  a  lower  value  for  the  other measurement. A  test with a very  high sensitivity  is desirable to diagnose  conditions we do not wish to miss, while  a test with a high specificity  is required  when expensive or  invasive confirma- tory investigations are required for posi- tive cases. For the current study, neither  of  these  requirements was a necessity;  Table 2 Haematology results of children with iron deficiency anaemia or alpha-thalassaemias Variable Iron deficiency anaemia (n = 18) Alpha-thalassaemias (n = 67) P-value Mean (SD) Mean (SD) RBC count (1012/L) 4.8 (0.7) 5.3 (0.5) < 0.001 Hb (g/dL) 8.6 (2.7) 11.0 (1.3) < 0.001 MCV (fL) 60.6 (9.5) 65.0 (6.6) 0.03 MCH (pg) 19.2 (4.2) 20.7 (2.6) 0.06 MCHC (g/dL) 29.2 (2.7) 31.3 (1.4) 0.02 RDW (%) 19.2 (4.2) 15.5 (3.9) < 0.001 RDW index 224.8 (69.4) 202.9 (91.4) 0.3 Mentzer index 11.5 (2.4) 12.4 (3.2) 0.2 England–Fraser index 8.1 (8.9) 0.5 (14.2) 0.03 Shine–Lal index 630.3 (314.8) 927.3 (316.8) < 0.001 MCV/MCH ratio 3.3 (0.48) 3.0 (0.50) 0.03 MCH/RBC ratio 11.5 (2.4) 12.5 (3.2) 0.2 RDW/RBC ratio 4.0 (1.3) 3.2 (1.3) 0.02 Green–King index 84.2 (34.2) 62.5 (31.4) 0.01 Srivastava index 4.0 (1.0) 4.2 (2.2) 0.8 SD = standard deviation; RBC = red blood cell count; Hb = haemoglobin; MCV = mean cell volume; MCH = mean cell haemoglobin; MCHC = mean cell haemoglobin concentration; RDW = red cell distribution width. Table 3 Haematology results of children with iron deficiency anaemia or alpha-thalassaemias using cut-off values for indices Index and cut-off Iron deficiency anaemia (n = 18) Alpha-thalassaemias (n = 67) P-value No. of patients % No. of patients % Mentzer index > 13 5 27.8 37 56.0 0.03 Srivastava index > 4.4 8 47.0 14 23.7 0.06 RDW index > 220 6 33.3 21 31.3 0.8 England–Fraser index > 0 17 94.4 45 67.1 0.02 RDW > 14% 16 88.9 33 50.8 0.004 RBC > 5 × 1012/L 6 33.3 50 79.4 < 0.001 RDW/RBC ratio > 3.3 11 61.1 23 34.3 0.04 Green–King index > 65 12 66.7 15 22.4 < 0.001 Shine–Lal index < 1530 18 100.0 64 95.5 0.3 RBC = red blood cell count; RDW = red blood cell distribution width. EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 970 we were only interested in investigations  with the combined maximum values for  sensitivity and  specificity,  as  they cor- rectly identify the highest proportion of  children evaluated. The Youden  index  takes  into account both  the sensitivity  and  specificity of  a  test  and correlates  positively with  the proportion of  chil- dren  correctly diagnosed by  that  test  [14,15]. Using  the Youden  index, we  found  that  a Green–King  index > 65  (which  correctly  identified  75.3% of  children)  and a RDW > 14% (which  correctly  identified 57.8% of children)  were  the best  indices  to differentiate  between IDA and AT.  Miller CJ et al. Factors associated with iron depletion and 1. iron deficiency anemia among Arabic preschool children of the United Arab Emirates. Saudi Medical Journal, 2004, 25:843–847. Miller CJ et al. A hematological survey of preschool children 2. of the United Arab Emirates. Saudi Medical Journal, 2003, 24:609–613. Hossain MM et al. The prevalence and correlates of anaemia 3. among young children and women of childbearing age in Al Ain, United Arab Emirates. Annals of Tropical Paediatrics, 1995, 15:227–235. D’Onofrio G et al. Automated measurement of red blood cell 4. microcytosis and hypochromia in iron deficiency and beta- thalassemia trait. Archives of Pathology & Laboratory Medicine, 1992, 116:84–89. Beyan C, Kaptan K, Ifran A. Predictive value of discrimination 5. indices in differential diagnosis of iron deficiency anemia and beta-thalassemia trait. European Journal of Haematology, 2007, 78:524–526. References Marsh WL Jr, Bishop JW, Darcy TP. Evaluation of red cell vol-6. ume distribution width (RDW). Hematologic Pathology, 1987, 1:117–123. Ntaios G et al. Discrimination indices as screening tests for beta-7. thalassemic trait. Annals of Haematology, 2007, 86:487–491. El-Hazmy MAF, Warsy AS. Hemoglobinopathies in Arab coun-8. tries. In: Teebi AS, ed. Genetic disorders among Arab popula- tions. New York, Oxford University Press, 1997:83–110. El-Kalla S, Baysal E. Alpha-thalassemia in the United Arab Emir-9. ates. Acta Haematological, 1998, 100:49–53. McClung JP et al. Prevalence of iron deficiency and iron de-10. ficiency anemia among three populations of female military personnel in the US army. Journal of the American College of Nutrition, 2006, 25:64–69. Aslan D et al. Importance of RDW value in differential diagnosis 11. of hypochrome anemias. American Journal of Haematology, 2002, 69:31–33. Rathod DA et al. Usefulness of cell counter-based parameters 12. and formulas in detection of beta-thalassemia trait in areas of A  previous  study,  from  Turkey,  looked at the value of erythrocyte indi- ces  to differentiate between  IDA and  beta-thalassaemia  trait,  and also used  the Youden  index  [16]. Contrary  to  our results, they found that RBC count  and RDW were  the most discrimina- tory  for  that purpose. The differences  can be  explained by  the difference  in  study  design.  Beta-thalassaemia  trait  was  the only haemoglobinopathy  they  tested for (constituting 41% of all cases  with microcytosis) and  the prevalence  of  IDA was 58.7%. By contrast,  in our  study, the prevalence of IDA was much  lower  (21.2%)  and  the prevalence of  haemoglobinopathies  much  higher  (78.8%),  confirming previous  studies  in our population.  [1–3].  In addition,  unlike  the  study  from Turkey where  children with haemoglobin  level < 8.7  g/dL were excluded, we did not use a  specific haemoglobin  level as a  thresh- old for exclusion. Conclusions In our population with a high prevalence  of AT haemoglobinopathies, a Green– King index > 65 and a RDW > 14% were  the best indices to diagnose IDA versus  AT in children with microcytosis and/ or hypochromia.  Table 4 Ability of indices to diagnose iron deficiency anaemia in children with iron deficiency anaemia or alpha- thalassaemias, by descending Youden index Index and cut-off Sensitivity (%) Specificity (%) LR (+ve test) LR (–ve test) % correctly classified Youden index Green–King index > 65 66.7 77.6 2.9 0.4 75.3 44.3 RDW > 14% 88.9 49.2 1.7 0.2 57.8 38.1 England–Fraser index > 0 94.4 32.8 1.4 0.1 45.8 27.2 RDW/RBC ratio > 3.3 61.1 65.7 1.7 0.6 64.7 26.8 Srivastava index > 4.4 47.6 76.2 1.9 0.7 69.7 23.8 RDW index > 220 33.3 68.6 1.0 0.9 61.2 1.9 Shine–Lal index < 1530 0.0 95.5 0.0 1.0 75.3 –4.5 Mentzer index > 13 27.8 43.9 0.5 1.6 40.5 –28.3 RBC > 5 × 1012 /L 33.3 20.6 0.4 3.2 23.4 –46.1 RBC = red blood cell count; RDW = red blood cell distribution width; LR = likelihood ratio. طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 971 Thalassaemias The alpha and beta thalassaemias are the most common inherited single-gene disorders in the world with the highest  prevalence in areas where malaria was or still is endemic. The burden of this disorder in many regions is of such a  magnitude that it represents a major public health concern. For example in the Islamic Republic of Iran, it is estimated  that about 8000 pregnancies are at risk each year. In some endemic countries in the Mediterranean region, long- established control programmes have achieved 80%–100% prevention of newly affected births.  Early treatment of thalessaemia has proved to be very effective in improving the quality of life of patients. Currently,  genetic testing and counselling, and prenatal diagnosis play an increasingly important role in informing individual as  well as professional decisions around the prevention, management and treatment of this disease. Source: http://www.who.int/genomics/public/geneticdiseases/en/index2.html high prevalence. American Journal of Clinical Pathology, 2007, 128:585–589. Baqar MS, Khurshid M, Molla A. Does red blood cell distribu-13. tion width (RDW) improve evaluation of microcytic anaemias? Journal of the Pakistan Medical Association, 1993, 43:149–151. Pekkanen J, Pearce N. Defining asthma in epidemiological 14. studies. 2 European Respiratory Journal, 1999, 14:951–957. Burney PG et al. What symptoms predict the bronchial re-15. sponse to histamine? Evaluation in a community survey of the bronchial symptoms questionnaire (1984) of the International Union Against Tuberculosis and Lung Disease. International Journal of Epidemiology, 1989, 18:165–173. Demir A et al. Most reliable indices in differentiation between 16. thalassemia trait and iron deficiency anemia. Pediatrics Interna- tional, 2002, 44:612–616. EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 972 Molecular genetics of beta-thalassaemia syndrome in Pakistan M. Usman,1 M. Moinuddin 1 and R. Ghani 2 ABSTRACT This molecular genetics study was conducted in Karachi, Pakistan from 2004 to 2006 to provide guidelines for prenatal diagnosis programmes in the country. Blood samples of patients with beta-thalassaemia minor (n = 200) and beta-thalassaemia major (n = 150) were collected from hospitals, transfusion centres and diagnostic laboratories from different districts of Karachi, representing 5 major ethnic groups. Molecular analysis revealed 11 genetic mutations of the beta-thalassaemia gene, among which 5 mutations accounted for 88% of the total beta-thalassaemia genes identified [IVS-1-5 (G–C), Fr 8/9 (+G), Fr 41/42 (–TTCT), IVS-1-1 (G–T) and Del 619]. Other mutations identified were: CAP+1, IVS-II-1 (G–A), Cd 5 (–CT), Cd 15 (G–A), Cd 16 and Cd 30. 1Institute of Haematology, Baqai Medical University, Karachi, Pakistan. 2Department of Biochemistry, Baqai Medical University, Karachi, Pakistan (Correspondence to M. Usman: staytune1@hotmail.com). Received: 25/01/09; accepted: 09/03/09 ناتسكاب في ةيميسلااث-اتيبلا ةمزلاتلم ةيئيزلجا تايثارولا ينغ هنيبور ،نيدلا ينعم ،نماثع دممح صيخشتلا جمابرل ةيداشرإ لئلاد ميدقتل 2006و 2004 ْيَماع ينب ام في ،ناتسكابب شيتارك ةنيدم في ةيئيزلجا ةيثارولا ةساردلا هذه تيرجأ :ةصلالخا ىبركلا ةيميسلااث-اتيبلاو )200=مهددع( ىرغصلا ةيميسلااث-اتيبلاب ينباصم ضىرم نم مدلا تانيع تع ُمجو .ناتسكاب في ةدلاولل ةقباسلا ةرـتفلا في .ىبرك ةيقرع تاعوممج سخم لّـِ ثتم يهو ،شيتارك تاعطاقم فلتمخ نم تابرتخلماو ،مدلا لقن زكارمو ،تايفشتسلما في كلذو ،)150=مهددع( يتلا ةيميسلااث-اتيبلا تانيج لياجإ نم %88 اهنم تارفط سخم تلكش ،ةيميسلااث-اتيبلا ينج في ةينيج ةرفط َةْشع ىدحإ يئيزلجا ليلحتلا رهظأو :اهيلع فرعتلا مت يتلا ىرخلأا تارفطلا نمو .[Del619و ،(IVS-1-5(G-C)، Fr8/9(+G)، Fr41/42(-TTCT)، IVS-1-1(G-T يهو] اهيلع فرعتلا مت .Cd 30و CAP+1، IVS-II-1(G-A)، Cd 5(-CT)، Cd 15(G-A)، Cd 16 Génétique moléculaire de la bêta-thalassémie au Pakistan RÉSUMÉ Cette étude de génétique moléculaire a été réalisée à Karachi, au Pakistan, entre 2004 et 2006, dans le but de formuler des recommandations applicables aux programmes de diagnostic prénatal du pays. Les prélèvements sanguins de patients atteints de bêta-thalassémie mineure (n = 200) et de bêta-thalassémie majeure (n = 150) ont été recueillis dans des hôpitaux, des centres de transfusion et des laboratoires de diagnostic de différents districts de Karachi, représentant ainsi cinq grands groupes ethniques. L’analyse moléculaire a révélé 11 mutations du gène responsable de la bêta-thalassémie, dont cinq concernaient 88 % des gènes responsables de la bêta-thalassémie identifiés [IVS-1-5 (G – C), Fr 8/9 (+ G), Fr 41/42 (–TTCT), IVS-1-1 (G – T) et Del 619]. Les autres mutations observées ont été les suivantes : CAP + 1, IVS-II-1 (G – A), Cd 5 (–CT), Cd 15 (G – A), Cd 16 et Cd 30. طسوتلما قشل ةيحصلا ةلجلماشع سداسلا دلجلما عساتلا ددعلا 973 Introduction Thalassaemia  is  the most common ge- netic disorder across  the world [1–4].  Thalassaemia was not  recognized as a  clinical entity until 1925, when Cooley  and Lee described a  syndrome occur- ring early in life that was associated with  splenomegaly  and  bone  deformities  [4,5].  Thalassaemia occurs with a high fre- quency in a broad belt extending from  the Mediterranean basin through to the  Middle East,  Indian subcontinent and  South-East Asia  [6–9]. About 3% of  the world population (150 million peo- ple) are carries of the beta-thalassaemia  gene [10–12]. Beta-thalassaemia is also  the most common genetic disorder in  Pakistan, a country with a population of  around 160 million people. The annual  rate of population growth is 3% and al- most 40% of the population is below 15  years of age [13,14]. There are 5 major  ethnic  groups:  Sindhi, Urdu  speak- ing, Punjabi, Baluchi and Pathan. The  Urdu-speaking group also  includes 2  large  communities,  the Memon and  the Gujarati. The carrier  frequency of  beta-thalassaemia  is  estimated  to be  around  6%  in  Pakistani  population  [15,16]. This  molecular  genetic  study  of  beta-thalassaemia  syndrome  in  5  major  ethnic  groups of Pakistan was  established  to provide guidelines  for  prenatal diagnosis programmes  in  the  country. Methods Sample The  study  group was  a  convenience  sample of 150 patients with beta-tha- lassaemia major and 200 with beta-tha- lassaemia minor. The beta-thalassaemia  major cases were collected mostly from  transfusion centres;  these patients had  transfusion-dependent  thalassaemia  and were registered at different transfu- sion  centres of Karachi. Patients had  been diagnosed on the basis of complete  blood  picture,  haemoglobin  electro- phoresis, peripheral blood morphology  and clinical signs and symptoms at  the  time of  registration.   The beta-thalas- saemia minor cases had been identified  during screening  for beta-thalassaemia  minor  in  the  general  population  and  included all major  ethnic  groups  and  thalassaemia  families  (i.e.  those with  children with  transfusion-dependent  beta-thalassaemia). These samples were  collected  from  different  transfusion  centres, hospitals and diagnostic labora- tories of Karachi.  Patients were selected  to  represent  the different ethnic groups of Pakistan.  The ethnic breakdown was as  follows:  80  Punjabi,  50  Pathan,  70  Sindhi,  50  Baluchi  and  100  Urdu  speaking (Table 1). Laboratory methods Venous blood samples were collected  into  EDTA  tubes.  All  samples  were  tested by  a modified  amplification of  refractory mutation  system (ARMS)  [17,18] for the 11 mutations previously  reported in the population of the Indian  sub-continent. The 11 mutated primer  sequences  which  were  used  during  this study were: IVSI-1 (G–T), IVSI-5  (G–C),  IVSII-1  (G–A), Del  619,  Fr  41–42 (–TTCT), Fr 8–9 (+G), Fr 16  (–C), Cd 5 (–CT), Cd 30 (G–C), Cd  15 (G–A) and Cap +1 (A–C). DNA  was extracted from whole blood by us- ing the Genomic DNA Purification Kit  (Gentra Systems, Minneapolis, USA).  ARMS primers were designed  for de- tection of normal and mutant DNA. A  control pair of primers was included in  each assay. Control primers A, B  and  C were amplified at 861 bp  fragments  from the 3′ end of the β-globin gene.  Polymerase chain  reaction (PCR)  was conducted by a modified method  in a mixture of 10 mmol/L tris (pH 8.3),  50 mmol/L KCl, 1.5 mmol/L MgCl 2 .  500 µM of each dNTP, 0.2 µmol/L of  each primer, 0.5 units of Taq polymer- ase and 0.5  to 1 pg of  genomic DNA  was added to the PCR mixture in a total  volume of 20 µL. The modified cycling  reaction (DNA Thermal Cycler, Per- kin-Elmer/Cetus) was programmed at  94 °C  for 1 min. (denature), 65 °C  for  1 min. (anneal) and 72 °C for 1.5 min.  (extend). After 25 cycles,  the  samples  were incubated for an additional 3 min.  at 66 °C [17,19].  PCR products were  removed and  mixed with 3 μL of a loading buffer and  then  loaded  on  2%  agarose  gel. The  Table 1 Ethnic breakdown of selected patients with beta-thalassaemia syndrome (n = 350) Ethnic group Total no. of patients sampled No. with beta thalassaemia major No. with beta- thalassaemia minor No. of beta- thalassaemia genes identified Punjabi 80 40 40 120 Pathan 50 20 30 70 Sindhi 70 20 50 90 Baluchi 50 20 30 70 Urdu speaking 100 50 50 150 Total 350 150 200 500 EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 974 gel was set at 100 volts  for 1 hour and  then  stained with  ethidium bromide.  After staining, the bands became visible  under ultraviolet light. The different mu- tations were characterized with a 100 bp  DNA ladder. Results Molecular analysis was carried out for  200 patients with beta-thalassaemia  minor and 150 with beta-thalassaemia  major. A  total of 500 beta-thalassae- mia genes were  identified. The eth- nic  breakdown was  as  follows:  120  genes  from Punjabi patients, 70  from  Pathans,  90  from  Sindhi,  70  from  Baluchi and 150 from Urdu-speaking  patients (Table 1).  Figures 1 and 2 show sample DNA  sequences. The most  common muta- tions  were  IVS-1-5  (G–C),  Fr  8/9  (+G),  Fr  41/42  (–TTCT),  IVS-1-1  (G–T) and Del 619 and together these  comprised 88% of the total beta-thalas- saemia genes identified (Table 2). Other  common mutations  that  were identified were CAP+1, IVS-II-1  (G–A), Cd 5 (–CT), Cd 15 (G–A),  Cd 16 and Cd 30 and they were 12%  of  the  total beta-thalassaemia genes  (Table 2).  Table 2 Gene mutations identified among patients with beta-thalassaemia syndrome (n = 300) in Pakistan by patient’s ethnic group Mutation Punjabi Pathan Sindhi Baluchi Urdu speaking All No. % No. % No. % No. % No. % No. % IVS-1-5 (G–C) 50 41.7 22 31.4 30 33.3 25 35.7 73 48.7 200 40.0 Fr-8/9 (+G) 30 25.0 21 30.0 20 22.2 9 12.9 20 13.3 100 20.0 Fr-41/42 (–TTCT) 15 12.5 14 20.0 15 16.7 11 15.7 20 13.3 75 15.0 IVS–1-1 (G–T) 6 5.0 2 2.9 8 8.9 8 11.4 11 7.3 35 7.0 Del 619 6 5.0 2 2.9 12 13.3 2 2.9 8 5.3 30 6.0 Cap+1 (A–C) 2 1.7 2 2.9 1 1.1 2 2.9 3 2.0 10 2.0 IVS-II-1 (G–A) 1 0.8 0 0.0 0 0.0 6 8.6 1 0.7 8 1.6 Cd 5 (–CT) 3 2.5 1 1.4 1 1.1 1 1.4 4 2.7 10 2.0 Cd 1 (G–A) 4 3.3 2 2.9 1 1.1 3 4.3 5 3.3 15 3.0 Cd 16 (–C) 1 0.8 2 2.9 1 1.1 1 1.4 2 1.3 7 1.4 Cd 30 (G–C) 2 1.7 2 2.9 1 1.1 2 2.9 3 2.0 10 2.0 Total 120 100.0 70 100.0 90 100.0 70 100.0 150 100.0 500 100.0 Discussion Beta-thalassaemia is the most common  genetic disorder  in Pakistan. The car- rier  frequency of beta-thalassaemia  is  estimated  to be around 6%  in  the Pa- kistani population [16]. Our  results of  molecular analysis of beta-thalassaemia  syndrome  in  these  5  major  ethnic  groups  revealed 5 common mutations  which comprised 88% of the total beta- thalassaemia genes: IVS-1-5 (G–C), Fr  8/9 (+G), Fr 41/42 (–TTCT), IVS-1-1  (G–T) and Del 619. Other mutations  identified  during  this  study,  which  constituted 12% of  the  total beta-tha- lassaemia genes, were: CAP+1, IVS-II-1  (G–A), Cd 5 (–CT), Cd 15 (G–A), Cd  16 and Cd 30. Other molecular  stud- ies of beta-thalassaemia  syndrome  in  Pakistan also showed the same 5 com- mon mutations:  IVS-1-5  (G–C),  Fr  8/9 (+G), Fr 41/42 (–TTCT), IVS-1-1  (G–T) and Del 619 [13–16]. The spectrum of beta-thalassaemia  mutations that was identified in the In- dian population also showed the same 5  common mutations: IVS-1-5 (G–C), Fr  8/9 (+G), Fr 41/42 (–TTCT), IVS-1-1  (G–T) and Del 619. These accounted  for 93.6% of the total beta-thalassaemia  genes in the Indian population [19–23].  In the Pakistani Gujrati population (an  Urdu-speaking population)  the most  common beta-thalassaemia gene was  Del 619 which  is also common  in  the  Indian Gujrati  population  [20]. The  similar  pattern of molecular  genetics  of beta-thalassaemia syndromes  in  the  Pakistani and  Indian populations may  be explained  in 2 ways. First, Pakistan  and India was one state for several hun- dred years before partition  in August  1947  and  cross-population  and  con- sanguineous marriages were common.  The  second  reason  is  the  large-scale  migration  from  India  to Pakistan and  from Pakistan to India during partition.  There are also close  similarities  in  the  molecular  genetics  of  beta-tha- lassaemia  syndrome  in  the Arab  and  Pakistani populations. The molecular  spectrum  of  beta-thalassaemia  syn- drome in United Arab Emirates (UAE)  nationals revealed that IVS-1-5 (G–C)  was  the  most  common  mutation,  with a  frequency of 66%. Other muta- tions  that were  reported  in  the UAE  included: Fr 8/9 (+G), Cd 5 (–CT),  IVS-II-1 (G–A), Cd 30 (G–C) and Cd  15 (G–A) [24,25]. The molecular spec- trum of beta-thalassaemia  in  the Arab  populations of  Jordan, Egypt,  Syrian  Arab Republic, Lebanon, Yemen and  Saudi Arabia revealed that the most fre- quent mutations were: IVS-1-5 (G–C),  طسوتلما قشل ةيحصلا ةلجلماشع سداسلا دلجلما عساتلا ددعلا 975 Figure 1 Analysis of DNA sample with mutation primers frame 41/42 and Cd 15. Lane 1a and 1b show normal control bands, Lane 2a, 3a and 4a are Fr 41/42; Lane 5b shows Cd 15. IVS-II-1  (G–A),  IVS-1-1,  Fr  8/9, Fr  41/42, Cd 15, Cd 16, Cap +1 (A–C),  IVS-1-110,  IVS-1-3′ end (-25 bp) and  IVS-1-6 [25,26]. These similarities may  be due  to  the  influence of  trading and  population migration by  the Arabs on  the Pakistani populations of Sindh and  Pakistani Punjab. Conclusions The beta-thalassaemia gene is present  in  all  the major ethnic groups  in  the  country  (Punjabi,  Pathan,  Sindhi,  Baluchi and Urdu speaking). Our mo- lecular  analysis of beta-thalassaemia  syndrome  in  these  5  major  ethnic  groups has  shown 5 common muta- tions which comprised 88% of the total  beta-thalassaemia genes in the sample.  These mutations  included:  IVS-1-5  (G–C),  Fr  8/9  (+G),  Fr  41/42  (–TTCT),  IVS-1-1  (G–T) and Del  619. Other mutations we  identified  during  this  study constituted 12% of  Figure 2 Analysis of DNA sample with mutation primers IVS 1-5 and Fr 8/9. Lane 3a, 5a and 9b and 10b show IVS 1-5, Lane 7a, 8a, 9a and 2b are Fr 8/9; Lane 11a and 11b show normal control bands of 50 bp. 1a 2a 3a 4a 5a 6a 7a 8a 9a 10a 11a 1b 2b 3b 4b 5b 6b 7b 8b 9b 10b 11b 1a 2a 3a 4a 5a 1b 2b 3b 4b 5b Cd 15 500bp 861bp 1500bp EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 976 Quek L, Thein SL. Molecular therapies in beta thalassaemia. 1. British Journal of Haematology, 2007, 136(3):353–365. Pan HF et al. Current status of thalassemia in minority popula-2. tion in Guangxi, China. Clinical Genetics, 2007, 71(5):419–426. Bun HF et al. 3. Hemoglobin: molecular genetics and clinical as- pects. Philadelphia, WB Saunders, 1986. Cooley TB et al. A series of cases of splenomegaly in children 4. with anemia and peculiar bones changes. Transactions of the American Pediatric Society, 1925, 37:29. Cooley TB et al. Erythroblastic anemia. 5. American Journal of Diseases of Children, 1932, 43:705. Kulozik et al. The molecular basis of alpha thalassemia in 6. India. Its interaction with the sickle cell gene. Blood, 1988, 71(2):467–472. Livingstone FB. 7. Abnormal hemoglobin in human population. Chicago, Aldine, 1970. Verma IC et al. Multicenter study of the molecular basis of 8. thalassemia intermedia in different ethnic populations. Hemo- globin, 2007, 31(4):439–452. Derakhshandeh-Peykar P et al. Distribution of beta-thalassem-9. ia mutations in Northern Province of Iran. Hemoglobin, 2007, 31(3):351–356. Aksoy M. Thalassemia intermedia; a genetic study in 11 pa-10. tients. Journal of Medical Genetics, 1970, 7(1):47–51. Ibrahim SA et al. Thalassemia and high F-gene in Aleppo. 11. Acta Haematologica, 1970, 44(5):287–291. Weatherall DJ. Molecular biology at the bedside. 12. British Medi- cal Journal, 1986, 292:1505–1508. Ahmed S et al. Molecular genetics of beta thalassaemia in Paki-13. stan: a basis for prenatal diagnosis. British Journal of Haematol- ogy, 1996, 94(3):476–482. Baig SM et al. Prenatal diagnosis of beta-thalassaemia in Southern 14. Punjab, Pakistan. Prenatal Diagnosis, 2006, 26(10):903–905. Ahmed S et al. Prenatal diagnosis of beta-thalassaemia in 15. Pakistan: experience in a Muslim country. Prenatal Diagnosis, 2000, 20(5):378–383. References Hafeez M et al. Regional and ethnic distribution of beta tha-16. lassemia mutations and effect of consanguinity in patients re- ferred for prenatal diagnosis. Journal of the College of Physicians and Surgeons of Pakistan, 2007, 17(3):144–147. Newton CR et al. Analysis of any point mutation in DNA. The 17. amplification refractory mutation system (ARMS). Nucleic Acids Research, 1989, 17(7):2503–2516. Old JM et al. Rapid detection and prenatal diagnosis of beta-18. thalassemia: studies in Indian and Cypriot populations in the UK. Lancet, 1990, 336:834–837. Varawalla NY et al. The spectrum of beta-thalassaemia muta-19. tions on the Indian subcontinent: the basis for prenatal diagno- sis. British Journal of Haematology, 1991, 78(2):242–247. Vaz FE et al. Distribution of beta-thalassemia mutations in the 20. Indian population referred to a diagnostic center. Hemoglobin, 2000, 24(3):181–194. Gupta A et al. Molecular genetic testing of beta-thalassemia 21. patients of Indian origin and a novel 8-bp deletion mutation at codons 36/37/38/39. Genetic Testing, 2003, 7(2):163–168. Colah R et al. Impact of beta globin gene mutations on the 22. clinical phenotype of beta thalassemia in India. Blood Cells, Molecules & Diseases, 2004, 33(2):153–157. Chakrabarti P et al. Spectrum of beta-thalassemia mutations in 23. North Indian states: a beta-thalassemia trait with two mutations in cis. Clinical Biochemistry, 2005, 38(6):576–578. Quaife R et al. The spectrum of beta-thalassaemia mutations in 24. the UAE national population. Journal of Medical Genetics, 1994, 31(1):59–61. El-Hazmi MA, Warsy AS, Al-Swailem AR. The frequency of 14 25. beta-thalassemia mutations in the Arab populations. Hemo- globin, 1995, 19(6):353–360. Baysal E. Molecular heterogeneity of beta-thalassemia in the 26. United Arab Emirates. Community Genetics, 2005, 8(1):35–39. the  total beta-thalassaemia genes and  included CAP+1,  IVS-II-1 (G–A), Cd  5  (–CT), Cd 15  (G–A), Cd 16  and  Cd 30. Beta-thalassaemia is the most com- mon genetic disorder  in Pakistan and  this molecular  analysis of beta-thalas- saemia syndrome  in Pakistan provides  a baseline  to help  in  the organization  of a  large-scale prevention programme  based on prenatal  diagnosis  of  beta- thalassaemia syndrome.  طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 977 Role of dipstick in detection of haeme pigment due to rhabdomyolysis in victims of Bam earthquake M. Amini,1 A. Sharifi,1 I. Najafi,1 P. Eghtesadi-Araghi 2 and M.R. Rasouli 1 ABSTRACT Avoiding life-threatening complications of rhabdomyolysis depends on early diagnosis and prompt management. The aim of this study was to evaluate the role of urinary dipstick test in the detection of haeme pigment in patients who were at risk of acute renal failure (ARF) due to rhabdomyolysis after suffering injury in the Bam earthquake. Serum creatine phosphokinase (CPK) level was used as the gold standard for prediction of ARF. ARF developed in 8 (10%) of 79 patients studied. We found no significant differences in the sensitivity, specificity and accuracy of dipstick urine and serum CPK tests for identifying patients who were at risk of ARF. However, dipstick urine test is an easy test that can be performed quickly at an earthquake site. 1Department of Nephrology, Tehran University of Medical Sciences, Tehran, Islamic Republic of Iran. 2Parsteb Pajouheshyar Medical Sciences Research Institute, Tehran, Islamic Republic of Iran (Correspondence to P. Eghtesadi-Araghi: payam_ eghtesadi@yahoo.com). Received: 02/02/09; accepted: 23/03/09 ماب لازلز اياحض في ةيلضعلا تادْيَب ُّرلا للاحنا نع مجانلا ميلها غابِص فاشتكا في ةَسيِمَغلا رابتخا رود ليوسر اضر دممح ،يقارع يداصتقا مايب ،يفجن جَريإ ،يفيشر ليع ،ينيمأ رْهَشونم هذه فدتهو .ةيروفلا ةلجاعلماو ركبلما صيخشتلا لىع ةيلضعلا تاديبرلا للاحنا نع ةجمانلا ةايحلل ةددهلما تافعاضلما ثودح عنم دمتعي :ةصلالخا تادْيَب ُّرلا للاحنا ببسب دالحا يولكلا لشفلا رَطَخ لىع مه نيذلا ضىرلما في ميلها غابِص فاشتكا في ةيلوبلا ةَسيِمَغلا رابتخا رود مييقت لىإ ةساردلا يذلا دالحا يولكلا لشفلاب ن ُّهكتلل لياثم رايعمك لصلما في ينتايركلا زانيكوفسوف ىوتسم مدخُتساو .ماب لازلز في تاباصلإل مهضرعت دعب ةيلضعلا ةسيمغلا ةقدو ةيعونو ةيساسح في ابه دتعُي تافلاتخا دوجو مدع نوثحابلا فشتكاو .مهتسارد تتم ًاضيرم 79 لصأ نم )%10( ضىرم ةينماث باصأ ةسيمغلا رابتخا نأ لاإ َّمهللا ،دالحا يولكلا لشفلا رطلخ ينضرعلما ضىرلما فاشتكا ثيح نم لصلما في ينتايركلا زانيكوفسوف ىوتسم ينبو ةيلوبلا .لازلزلا عقوم في ةعسرب هؤارجإ نكمي لهس رابتخا ةيلوبلا Rôle des bandelettes réactives dans la détection du pigment hémique lié à la rhabdomyolyse chez les victimes du tremblement de terre de Bam RÉSUMÉ Un diagnostic précoce et une prise en charge rapide sont essentiels pour éviter les complications de la rhabdomyolyse mettant en jeu le pronostic vital. Cette étude visait à évaluer le rôle des bandelettes urinaires dans la détection du pigment hémique chez les patients présentant un risque d’insuffisance rénale aiguë lié à une rhabdomyolyse induite par une blessure lors du séisme survenu à Bam. Le taux de créatine-phosphokinase sérique (CPK) a été employé comme méthode de référence pour prévoir la survenue d’une insuffisance rénale aiguë. Huit des 79 patients étudiés (10 %) ont développé une insuffisance rénale aiguë. Nous n’avons relevé aucune différence significative en termes de sensibilité, spécificité et précision entre les bandelettes urinaires et les dosages de la CPK sérique dans l’identification des patients présentant un risque d’insuffisance rénale aiguë. Les bandelettes urinaires constituent cependant un test rapide et simple à réaliser sur les lieux d’un tremblement de terre. EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 978 Introduction On 26 December 2003, an earthquake  of amplitude 6.3 on the Richter scale hit  the  south-east of  the  Islamic Republic  of Iran, decimating the city of Bam and  killing more  than 26 000 people  and  injuring 30 000 [1]. During catastroph- ic  events  such  as  earthquakes,  crush  syndrome  is  the second most  frequent  cause of mortality  after  the  impact of  trauma  [2].  Crush  injuries  resulting  in  traumatic  rhabdomyolysis  are  an  important  cause of  acute  renal  failure  (ARF) [3]  in 4%–33% of patients, due  to myoglobinuria [4]. This kind of ARF,  requiring  renal  replacement  therapy  support, has  a mortality  rate of  about  14% [5]. Avoiding life-threatening complica- tions of  rhabdomyolysis  strongly de- pends on early diagnosis  and prompt  management  of  the  condition  [6].  Many clinical features of rhabdomyoly- sis are nonspecific, and the course of the  syndrome depends on  the underlying  condition [7]. It has been demonstrated  that the creatine phosphokinase (CPK)  level  in  serum has an association with  rhabdomyolysis [8]. However, it cannot  be measured at the site of earthquake. It  is  therefore  important  to develop new  easy screening tests.  Myoglobinuria  establishes  the di- agnosis of  rhabdomyolysis. Screening  may be performed with a urine dipstick  test  [9]. The ortho-toluidine portion  of  the dipstick  turns blue  in  the pres- ence of haemoglobin or myoglobin.  If  a freshly-spun sediment of urine shows  no red blood cells, positive urine blood  can be used as a  surrogate marker  for  myoglobin  [7],  although  there  is  still  the possibility of haemoglobinuria due  to other  causes  such  as  intravascular  haemolysis. Urine  dipstick  has  been  used  for detection of  rhabdomyolysis  in patients  suffering heat  injury  [10],  physical abuse [11], general convulsions  [12],  immersion  and  near-drowning  [13] and exercise-induced  rhabdomy- olysis [14]. To our knowledge it has not  been not evaluated as a screening test in  patients with rhabdomyolysis following  injury in an earthquake. This  study  evaluated  the  role  of  dipstick urinalysis  in detection of  haeme pigment  in patients  injured  in  the Bam earthquake who were at  risk  of ARF due  to  traumatic  rhabdomy- olysis. The  study  also  compared  the  dipstick urinalysis as possible evidence  of myoglobinuria with the serum CPK  level which is considered to be the gold  standard  test  in  this  setting.  If  suffi- ciently sensitive, urine dipstick analysis  may provide a reliable screening test for  rhabdomyolysis and could then be in- corporated into future rapid screening  protocols,  thus allowing earlier  initia- tion of treatment. Methods In a cross-sectional study, we evaluated  all victims of the Bam earthquake who  were admitted to the tertiary care centre  of Shariati hospital, which is affiliated to  Tehran University of Medical Sciences.  Immediately  after  admission,  a  team  including nephrologists and  internists  evaluated  the patients  for crush  injury  in the emergency department.  Measured parameters In 79 patients admitted during the first  3  days  after  the  earthquake,  a  urine  dipstick  test  (Uriyab-8  tapes, Bakhtar  Chimie) was performed using  the first  voiding  or  catheterization  sample  to  detect  haematuria/myoglobinuria.  The  reactions were evaluated visually  by comparing  the  strip  test  areas with  the colour chart printed on  the bottle  as  either positive  (from 1+  to 4+) or  negative. The first CPK value  that had been  measured during the first 3 days after the  earthquake was obtained. Serum CPK  level was estimated using a commercial  kit (Pars Azmoon) and an autoanalyser  (Technicon RA-1000). Criteria for ARF and haemodialysis ARF was  defined  as  an  acute  loss  of  renal function with a persistent (at least  2 times on 2 different days) elevation of  serum creatinine of ≥ 2.5 mg/dL, which  did not  improve with hydration  [15].  The  criteria  for  haemodialysis  were:  serum  creatinine  >  8 mg/dL,  blood  urea nitrogen  (BUN) > 100 mg/dL,  serum potassium > 7 mEq/L,  serum  bicarbonate < 10 mEq/L, and clinical  symptoms  and findings of ARF  such  as oedema, hypertension, heart  failure,  nausea and vomiting [16]. Statistical analysis Data  were  expressed  as  mean  and  standard deviation (SD) or number of  patients.  Statistical  analysis ware per- formed utilizing SPSS, version 11.5 and  MedCalc,  version 9.2. Mann–Whitney  U test was used to compare CPK levels  in patients with and without ARF. The  best cut-off points of urine dipstick and  serum CPK  level were obtained based  on  receiver  operating  characteristics  (ROC)  analysis  and  area  under  the  curve  (AUC). Then,  comparison  of  AUCs was performed. Finally, sensitiv- ity  and  specificity,  positive predictive  value,  negative  predictive  value  and  accuracy of both  tests  for detection of  occurrence of ARF were calculated. P value < 0.05 was considered statistically  significant. Results A total of 79 patients including 54 males  (68%) were  studied. The age distribu- tion of the patients is shown in Table 1.  Of  these patients, 8 (10%) developed  ARF and 4 of  them needed haemodi- alysis; 2 patients (2%) died, 1 of  them  suffering ARF;  3  fasciotomies  and  2  amputations were performed. In  Table  2,  the  results  of  urine  dipstick are  summarized. There was a  significantly higher occurrence of posi- tive dipstick  test  for blood  in  the urine  طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 979 in  the  group with ARF  than  in  those  without. The degree of  reaction of  the  urine dipstick test was also significantly  higher in this group. When the patients were divided into  2 groups:  low-risk (urine blood: nega- tive,  1+  and 2+)  and high-risk  (urine  blood: 3+ and 4+),  the  sensitivity and  specificity of urine blood  in high-risk  patients for prediction of occurrence of  ARF was 100% (95% CI: 67%–100%)  and 74% (95% CI: 63%–83%) respec- tively. Positive predictive value and neg- ative predictive values were 30% (95%  CI:  16%–49%)  and  100%  (95% CI:  93%–100%) respectively. The accuracy  of the dipstick test in prediction of ARF  was  77%  (95% CI:  68%–83%). The  AUC was 0.90 (95% CI: 0.79–0.96). Mean  serum  CPK  levels  were  13 225  (SD  11 554)  IU/L  (range  2430–34 230  IU/L)  and  1882  (SD  2023) IU/L (range 39–11 424 IU/L)  in  patients  with  and  without  ARF  respectively (P < 0.001). When the pa- tients were divided into low-risk (CPK  level ≤ 2259 IU/L) and high-risk (CPK  level > 2259 IU/L) groups, the sensitiv- ity and specificity of CPK  in high-risk  patients for detection of ARF was 100%  (95% CI: 64%–100%) and 68% (95%  CI:  54%–79%)  respectively. Positive  predictive value and negative predictive  values were 23% (95% CI: 10%–45%)  and 100% (95% CI:  90%–100%)  re- spectively. The accuracy of serum CPK  in prediction of ARF was 71% (95% CI:  60%–80%). AUC was 0.890 (95% CI:  0.77–0.95). Comparison of the AUCs of serum  CPK level and urine blood by dipstick  did not show a significant difference (P = 0.929). Discussion The results of this study showed that the  dipstick urine  test  is  a highly  sensitive  and easy screening tool for identification  of patients who are at risk of developing  ARF due  to  rhabdomyolysis and may  have a role in the early detection of pa- tients at the site of an earthquake. Also  comparison of  the dipstick urine  test  with serum CPK level (gold standard)  did  not  reveal  any  significant  differ- ences.  A   posi t ive   ur ine   myoglobin  test  provides  supportive  evidence  of  rhabdomyolysis  [7]  and  some other  studies have been done on patients with  trauma. In a study of cases of traumatic  rhabdomyolysis, Muckart et al. revealed  that  an  initial  venous  bicarbonate  concentration of  <  17 mmol/L with  myoglobinuria is highly sensitive for pre- dicting ARF [17]. Spicer et al. in a study  of patients with acute renal impairment  due  to  immersion and near-drowning,  found that a dipstick reaction on admis- sion was significantly more common in  patients with  acute  renal  impairment  [13]. Also Melli et al. in a review of 475  patients with  rhabdomyolysis,  showed  that urine myoglobin detected by dip- stick/ultrafiltration was positive in 19%  [18]. However, other  studies  found a  lack of adequate sensitivity [19]. As we wanted to assess the efficacy  of dipstick urine  test  at  the  site of  the  earthquake, we did not  exclude cases  with  haematuria;  however,  some  in- vestigators  have  emphasized  that  for  clinical purposes, myoglobinuria  is  just  demonstrated by a blood-positive dip- stick when  there  is no haematuria or  haemoglobinuria  [20]. However, we  had no patients with overt abdominal– pelvic  injuries  and  likely  traumatic  haematuria. Also we did not perform  dipstick  test  in  the first  urine  sample  after catheterization.  The importance of early initiation of  vigorous fluid  resuscitation and other  medical  treatments  in earthquake vic- tims have been  emphasized.  In mass  disasters,  early  treatment  in  the  field  should be  focused on seriously  injured  persons who  require  immediate  care  [21]. To identify patients at risk of renal  damage,  the quickest and  least expen- sive  screening  test  for  rhabdomyolysis  is the serum CPK level [22]; however, it  has to be performed in a hospital setting  and with the lack of medical equipment  and personnel  in  the field  in an emer- gency,  it  is hard to use it as an effective  diagnostic tool. Increased muscle enzymes particu- larly CPK is a marker of muscle  injury  [23]. Reported thresholds of CPK in the  Table 1 Age distribution of the study patients Age (years) No.a % 0–14 2 13.9 15–34 48 60.8 35–54 10 12.7 > 55 8 10.1 Unknown 2 2.5 aData on age were missing for 9 patients. Table 2 Urine dipstick results of patients with and without acute renal failure in the first 3 days after suffering trauma Variable Acute renal failure P-value Yes (n = 8) No (n = 71) Dipstick Negative 0 34 0.018a Positive 8 37 Intensity of positive samples 0.024b 0 0 34 1+ 0 9 2+ 0 10 3+ 7 13 4+ 1 5 aFisher exact test; bChi-squared test. EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 980 literature vary  from 500 to 3000 IU/L  [23].  In  this  study we  found a  serum  CPK threshold of > 2259 IU/L for pre- diction of ARF. Our results showed that  the dipstick urine  test and serum CPK  level had the same sensitivity for predic- tion of ARF; however, the specificity of  the dipstick urine test was slightly higher  than for CPK. Considering the fact that  our study did not find any significant dif- ferences between these 2 tests, dipstick  urine test has a great advantage as it can  be  easily performed at  the  site of  the  injury. Another  important problem  to be  considered  is  the  time between  injury  and appearance of sufficient amounts of  myoglobin  to be detected  in  the urine  sample  by dipstick. When urine my- oglobin concentration is < 60 000 μg/L,  utilization of conventional dipstick has  the potential  to misdiagnose patients  with myoglobin concentrations associ- ated with  increased risk of  subsequent  renal dysfunction [24]. Then a normal  urine dipstick  test  result does not  rule  out this condition [18]. There were  important  limitations  in this study. First, we assessed dipstick  results  visually. To minimize  the con- founder effects of human error, it would  be better  to  assess  the dipsticks with  an  automated analyser. Furthermore,  it was not possible  to  identify whether  the positive dipstick test was caused by  the presence of  blood, myoglobin or  both  in urine. The presence of blood  in urine causes significant false positive  readings (both for visual and automated  assessments), which has negative effects  on  the  specificity  and,  in  turn, on  the  accuracy of urinary dipstick  in detec- tion  of myoglobinuria. To minimize  this problem, we divided our patients  into 2 groups of high- and  low-risk  for  development of ARF according  to  the  colour chart on  the dipstick bottle.  In  addition, in some of our cases, there was  a 36-hour gap between  the  injury and  dipstick test. Thus, it is possible that the  therapeutic efforts  in  this  time  interval  influenced  the final  results. Finally  the  number of patients who suffered  from  ARF was small (n = 8) and so the results  are preliminary. Therefore,  the  results  need  to  be  tested  on other  data  sets  before clinical application. In  conclusion, our  results did not  reveal  significant differences between  dipstick urine  test  and  serum CPK  in  identifying patients who were  at  risk  of ARF. However,  in view of  the high  sensitivity of urine dipstick  test and  its  ease of use  in  the field, we suggest  that  utilization of this test as an early screen- ing  tool  in detection of  victims prone  to ARF may have benefits and facilitate  triage of high-risk patients. Further stud- ies are needed to evaluate the potential  efficacy of dipstick urine test. Table 3 Serum creatine phosphokinase levels of patients with and without acute renal failure in the first 3 days after suffering trauma Creatine phosphokinase level (IU/L) Acute renal failure Total (n =79) Yes (n = 8) No (n = 71) ≤ 2259 0 35 35 > 2259 5 16 21 Total 5 51 56 P = 0.005. Emami MJ et al. Strategies in evaluation and management of 1. Bam earthquake victims. Prehospital and Disaster Medicine, 2005, 20(5):327–330. Ukai T. The great Hanshin-Awaji earthquake and the problems 2. with emergency medical care. Renal Failure, 1997, 19(5):633– 645. Malinoski DJ, et al. Crush injury and rhabdomyolysis. 3. Critical Care Clinics, 2004, 20(1):171–192. Bagley WH et al. Rhabdomyolysis. 4. Internal and Emergency Medicine, 2007, 2(3):210–218. Atef MR et al. Acute renal failure in earthquake victims in Iran: 5. Epidemiology and management. Quarterly Journal of Medicine, 1994, 87(1):35–40. Poels PJE et al. Rhabdomyolysis: a review of the literature. 6. Clini- cal Neurology and Neurosurgery, 1993, 95(3):175–192. Sauret JM et al. Rhabdomyolysis. 7. American Family Physician, 2002, 65(5):907–912. Lima RS et al. Acute kidney injury due to rhabdomyolysis. 8. Saudi journal of kidney diseases and transplantation, 2008, 19(5):721–729. Line RL et al. Acute exertional rhabdomyolysis. 9. American Fam- ily Physician, 1995, 52(2):502–506. References Young SE et al. Is urine dipstick a reliable screening tool for 10. rhabdomyolysis in the suspected heat injury patient? Annals of Emergency Medicine, 2006, 48(4 Suppl.):90. Peebles J et al. Child physical abuse and rhabdomyolysis: case 11. report and literature review. Pediatric Emergency Care, 2007, 23(7):474–477. Os I et al. General convulsions and rhabdomyolysis. Case re-12. ports. Acta Neurologica Scandinavica, 1989, 79(3):246–248. Spicer ST et al. Acute renal impairment after immersion and 13. near–drowning. Journal of the American Society of Nephrology. 1999, 10(2):382–386. Sinert R et al. Exercise-induced rhabdomyolysis. 14. Annals of Emergency Medicine, 1994, 23(6):1301–1306. Ward MM. Factors predictive of acute renal failure in rhabdomy-15. olysis. Archives of Internal Medicine, 1988, 148(7):1553–1557. Gunal AI et al. Early and vigorous fluid resuscitation prevents 16. acute renal failure in the crush victims of catastrophic earth- quakes. Journal of the American Society of Nephrology, 2004, 15(7):1862–1867. Muckart DJJ et al. Prediction of acute renal failure following 17. soft-tissue injury using the venous bicarbonate concentration. Journal of Trauma, 1992, 33(6):813–817. طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 981 Diagnostics and Laboratory Technology The goal of the World Health Organization’s Diagnostics and Laboratory Technology team is to promote and facilitate  access to safe, reliable and appropriate diagnostic technologies and laboratory services in an equitable manner through:  Prequalification of diagnostics for high burden diseases;• Capacity building of national regulatory authorities and national reference laboratories;• Facilitation of procurement of affordable and appropriate diagnostics;• Policy, guidance and advocacy to Member States; • Provision of quality assurance programmes to countries;• Training and technical support including country projects.• Further information about the work of WHO in Diagnostics and Laboratory Technology is available at: http://www. who.int/topics/diagnostic_techniques_procedures/en/ Melli G et al. Rhabdomyolysis: an evaluation of 475 hospital-18. ized patients. Medicine (Baltimore), 2005, 84(6):377–385. Gabow PA, et al. The spectrum of rhabdomyolysis. 19. Medicine (Baltimore), 1982, 61(3):141–152. Briner V et al. Die akute Rhabdomyolyse [Acute rhabdomy-20. olysis]. Schweizerische Medizinische Wochenschrift, 1986, 116(7):198–208. Pepe PE et al. Field management and critical care in mass disas-21. ters. Critical Care Clinics, 1991, 7(2):401–420. Mote JD et al. Rhabdomyolysis and acute renal failure. 22. Me- dicina Interna de Mexico, 2007, 23(1):47–58. Sever MS et al. The Marmara earthquake: Admission laboratory 23. features of patients with nephrological problems. Nephrology Dialysis Transplantation, 2002, 17(6):1025–1031. Loun B et al. Ultrafiltration discrepancies in recovery of my-24. oglobin from urine. Clinical Chemistry, 1996, 42(6 Suppl.):965– 969. EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 982 Role of shame in the stigmatization of people with human immunodeficiency virus: a survey of female college students in 3 Arab countries A.M. Badahdah 1 and C.E. Foote 2 ABSTRACT AIDS stigma is a challenge to controlling the HIV/AIDS epidemic especially in more conservative cultures. This study explored the impact of knowledge about HIV and AIDS, and the impact of shame, on the stigmatization of people living with HIV/AIDS in the Arab world. Survey data were collected from 277 female college students in 3 Arab countries: Kuwait, Bahrain and Jordan. Only in Bahrain was knowledge about HIV and AIDS inversely related to negative attitudes toward people with HIV/AIDS. AIDS-related shame, however, was a strong predictor of AIDS stigma in all 3 countries. HIV education is needed for young people in Arab countries, especially women, both for their own health and to reduce the problem of AIDS stigma. 1Department of Sociology, University of North Dakota, Grand Forks, North Dakota, United States of America (Correspondence to A.M. Badahdah: abdallah.badahdah@und.nodak.edu). 2Department of Sociology, Indiana University-Purdue University, Indianapolis, Indiana, United States of America. Received: 15/01/09; accepted: 23/03/09 ةيبرع نادلب ةثلاث في تايعمالجا تابلاطلا حسم :يشربلا يعانلما زوعلا سويرفب ينباصلما مصو في راعلاب روعشلا يرثأت توف ييرك ،حدحداب دممح للها دبع هذه تفشكتسا دقو .ًةظفامح رثكأ يه يتلا تافاقثلا في ماّيسلاو هسويرفب ىودعلاو زديلإا ءابو ةحفاكم لقرعي ًايدتح زديلإا ةمصو دعت :ةصلالخا تع ُمج دقو .بيرعلا لماعلا في هسويرفب ىودعلاو زديلإل ينشياعلما مصو في ،راعلاب روعشلا رثأو ،هسويرفب ىودعلاو زديلإا نع فراعلما رثأ ةساردلا نع فراعلما تطبارـت دقف اهدحو نيرحبلا في امأ .ندرلأاو ،نيرحبلاو ،تيوكلا :يه ةيبرع نادلب ةثلاث في ةيعماج ةبلاط 277 نم حسلما تايطعم ناك زديلإاب قلعتلما راعلاب روعشلا نأ ولو .هسويرفب ىودعلا وأ زديلإاب ينباصلما وحن ةيبلسلا فقاولما عم ًايسكع ًاطبارـت هسويرفب ىودعلاو زديلإا ،ةيبرعلا نادلبلا في زديلإا سويرف لوح بابشلا فيقثتل ةجاح كانه نأ لىإ ةساردلا صلتخو .ًاعيج ةثلاثلا نادلبلا في زديلإا ةمصو ثودحب ًايوق ًائبنم .زديلإا ةطبترلما ةمصولا ةلكشم نم دلحاو نهتحص لىع ظافلحا لجأ نم كلذو ،تايتفلا ماَّيسلاو Rôle de la honte dans la stigmatisation des personnes atteintes du virus de l’immunodéficience humaine : étude réalisée sur des lycéennes dans trois pays arabes RÉSUMÉ La stigmatisation du sida constitue un obstacle en matière de lutte contre l’épidémie de VIH/sida, en particulier au sein des cultures conservatrices. Cette étude portait sur les conséquences du niveau de connaissance sur le VIH et le sida et de la honte suscitée par la maladie, sur la stigmatisation des personnes vivant avec le VIH/ sida dans le monde arabe. Les données de cette étude ont été recueillies auprès de 277 étudiantes dans trois pays arabes : Bahreïn, la Jordanie et le Koweït. Bahreïn s’est révélé le seul pays où le niveau de connaissance sur le VIH/sida est inversement proportionnel aux attitudes négatives à l’égard des personnes vivant avec le virus. La honte associée au sida constitue toutefois un facteur prédictif important de la stigmatisation de la maladie dans les trois pays. L’éducation en matière de VIH est nécessaire auprès des jeunes des pays arabes, notamment des femmes, tant pour préserver leur propre santé que pour réduire le problème de la stigmatisation du sida. طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 983 Introduction Deeply held cultural values, such as dis- approval of nonmarital sex and of drug  use, have been accredited with slowing  the  spread  of HIV  infection  in Arab  societies until now. Yet, more recently,  these  same cultural  values have been  blamed  for  the  increase of HIV cases  which  is being  seen  in  these  societies.  This seemingly contradictory view rests  on the idea that the intense disapproval  of such behaviour will discourage people  from being  tested  for HIV or pursuing  adequate care or  treatment when  they  need  it  [1–4]. These  factors have  led  to a growing awareness  that  reducing  AIDS stigma is critical to effective HIV  prevention and treatment [5].  In  the Arab world, women  are  at  heightened  risk  for HIV  infection be- cause of  their  socially  disadvantaged  position in society. Lack of independent  income and  low  literacy curtails wom- en’s ability  to access HIV  information,  to ward off unwanted sexual advances  and  to negotiate  safe  sexual practices  [6,7].  Indeed, Arab women now repre- sent half the total number of people car- rying the virus in the region and nearly  80% of affected women contracted the  virus from their husbands [8].  Although  genuine  attempts  have  been made to raise AIDS awareness by  authorities  in  the Arab region,  they are  comparatively meagre [1,9]. In particu- lar, they fail to address the population’s  continued  lack of access  to good HIV  information. Young adults are  likely  to  be the most negatively affected because  they  are  at  the  life  stage where  risky  behaviour,  such as having unprotected  sex or experimenting with drugs, is most  likely to take place [1]. Studies of college  students  in  Jordan  and United Arab  Emirates (UAE) found alarming gaps in  HIV and AIDS knowledge and negative  attitudes  towards  people with HIV/ AIDS [10–12]. All 3 studies found that  male  students had greater knowledge  than  female  students. However,  these  and  other  studies  in  Saudi  Arabia,  Yemen and Kuwait  [13–16]  focused  largely  on  the  impact  of  knowledge  about HIV and AIDS on AIDS stigma;  none paid close attention to the role of  emotion in AIDS stigma.  Some  researchers have  suggested  that  certain emotions are  triggered at  the  introduction of  the  topic of AIDS,  and  that  these  emotional  responses  remain  among  the most  challenging  barriers to reducing stigmatization [17].  One of these emotions is shame, which  was  found  in  to be a  significant  factor  that  caused HIV-positive  individuals  to hide  their  health  status  from  their  caregivers  and  to withdraw  from  so- cial  interactions [3,18,19]. Two recent  studies  from Saudi Arabia and Yemen  suggest that the emotion of shame may  be particularly salient in understanding  the stigmatization of people with HIV/ AIDS in the Arab world and that feeling  ashamed that a friend or family member  has HIV may be among  the  strongest  predictors of stigmatization [16,18].  In view of the lack of AIDS research  in  the Arab  region and  the  increasing  vulnerability of Arab women  to HIV,  the present study sought to assess gaps  in younger Arab women’s HIV/AIDS  knowledge,  along with  their  concep- tions  of  shame,  and  their  reactions  toward  people  with HIV/AIDS.  To  this  end we  collected  and  compared  data from female college students from  Kuwait, Bahrain  and  Jordan.  Specifi- cally, we wanted to investigate whether  knowledge  about HIV and AIDS, on  the one hand, and shame on the other,  had  any  effect  on  tolerance  towards  persons with HIV/AIDS.  Methods Sample Three convenience  samples of  female  undergraduate  students  from colleges  in Kuwait,  Bahrain  and  Jordan  com- pleted  a  self-administered  question- naire written  in Arabic.  Participation  in  the  research was  voluntary  and no  payment or curricular credit was given  for participation. The samples consisted  of 83 students  from Kuwait, 108  from  Bahrain and 86 from Jordan.  Data collection A questionnaire consisting of 13  items  was used  to gauge participants’ knowl- edge of HIV  transmission  and AIDS  (Table 1). These  items were  adapted  from  the  Attitudes  Towards  AIDS  Knowledge Scale [20] which has been  used  for  college  students. The  items  covered 3 areas: behavioural and situ- ational risk factors (5 items); fluids that  carry HIV and  risk  reduction steps (4  items); and misconceptions about HIV  and AIDS (4  items). Each  item had 3  choices: “true,” “false” and “don’t know”:  a score of 1 was assigned to the correct  answer and 0  to an  incorrect or  “don’t  know” answer. Therefore, the higher the  score a respondent received, the greater  that participant’s knowledge about HIV  and AIDS.  In  the AIDS  literature  there  are 2  items that have been frequently used in  assessing the expression of shame asso- ciated with AIDS [21]: being ashamed  of having a  family member with HIV/ AIDS and  the belief  that people with  HIV/AIDS  should  be  ashamed  of  themselves.  In  this  study we  added  a  third  item  that  assessed  shame at  the  national level by asking the participants  if  they would be ashamed of having a  fellow citizen with HIV/AIDS. Again,  responses were  recorded on a 5-point  Likert  scale and  the higher  the score a  respondent received, the greater the re- spondent’s expression of AIDS-related  shame.  We measured  the dependent vari- able, AIDS stigma, with 7  items  [21].  These included: people with HIV/AIDS  should be fired  from their  jobs; people  with HIV/AIDS  should  be  quaran- tined; I am sympathetic toward people  with HIV/AIDS; I have no objection to  marrying into a family with HIV/AIDS;  I do not mind being touched by some- one with HIV/AIDS;  all  people with  EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 984 HIV are responsible for their infection;  and if a friend of mine got infected with  HIV,  I would continue being a  friend  with him/her.  Responses  were  recorded  on  a  5-point Likert  scale  that  ranged  from  “strongly disagree”  to  “strongly agree”.  After reversing the score values for those  statements  that  read  sympathetically  toward people with AIDS,  the  scores  were  tallied  as before. The higher  the  score,  the more negative attitudes  stu- dents held toward people with AIDS.  Analysis SPSS,  version  17,  was  used  for  both  descriptive and  inferential  statistics. A  P-value of < 0.05 was considered statisti- cally significant. Results Demographic data The ages of the Kuwaiti women (n = 83)  ranged from 18 to 26 years, with a mean  age of 20.6  [standard deviation (SD)  1.4] years, while the Bahraini (n = 108)  and Jordanian women (n = 86) ranged  from 18 to 28 years, with mean ages of  21.4 (SD 1.9) years and 19.8 (SD = 1.9)  years  respectively. Most of  the women  were single: 90% of the Kuwaitis, 79% of  the Bahrainis and 97% of the Jordanians.  Overall a total of 33 women were mar- ried and 1 was divorced.  HIV and AIDS knowledge Table 1  reports  the  frequency of  cor- rect  answers  to  the  HIV  and  AIDS  knowledge  items of  the  respondents  by country. Nearly all the students cor- rectly  identified  the main behaviours  or  situations  that place people  at  risk  for HIV,  i.e.  sex with an HIV-positive  person,  intravenous  drug  use with  a  contaminated needle and being carried  to term by an infected pregnant woman  (range 88%–100% answered correctly).  However, while most  students  could  correctly  identify  the behavioural  risks  of  transmission,  they were much  less  knowledgeable  about  the  fluids  that  carry HIV. A  large proportion of  the  women were unaware  that HIV could  be  transmitted  through  semen  (only  23% of Bahraini,  27% of Kuwaiti  and  51%  of  Jordanian  women  answered  correctly) or  through mother’s breast  milk (32% of Bahraini, 52% of Kuwaiti  and 51% of Jordanian women answered  correctly).  A number of other misconceptions  about HIV  and AIDS were  also  evi- dent. About 30% of the women in each  country believed  there was  a  cure  for  AIDS, 63%–73% believed one could get  HIV from a mosquito bite and a striking  78%–83% were unaware that condoms  could help  reduce  the chance of HIV  infection.  Assigning a “correct score” for each  question answered correctly by at  least  half  of  participants  in  each  sample  yielded the following pattern. The sam- ple of Bahraini women answered 9/13  questions correctly (69%), the Kuwaitis  7/13 (54%) and  the  Jordanians 6/13  Table 1 Frequency of correct answers on AIDS-knowledge index of female college students in 3 Arab countries Knowledge item Kuwait (n = 83) Bahrain (n = 108) Jordan (n = 86) No. correct % No. correct % No. correct % One can get HIV by having sex with an HIV-positive person 83 100 108 100 83 97 One can get HIV by sharing HIV-contaminated needles 81 98 105 97 80 93 An infected mother can transmit HIV during pregnancy 73 88 98 91 77 90 One can get HIV by touching someone with AIDS 59 71 92 85 41 48 Only homosexuals get AIDS 62 75 82 76 51 59 HIV can live in the human body for years before symptoms appear 54 65 75 69 30 35 There is a cure for AIDS 53 64 74 69 53 62 You can tell if someone has AIDS by looking at him/her 49 59 71 66 40 47 Unsafe anal sex is the most risky sexual act for contracting HIV 37 45 59 55 45 52 An infected mother can transmit HIV through breastfeeding 43 52 34 31 44 51 A mosquito bite can transmit HIV 13 16 29 27 15 17 HIV can be transmitted through semen 22 27 24 22 44 51 Condoms help reduce the probability of HIV transmission 18 22 20 19 15 17 طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 985 (46%). These  variations  in HIV  and  AIDS knowledge among  the  samples  were statistically  significant (ANOVA,  F (2, 278) = 4.78, P = 0.009). The Tukey  test revealed that Bahraini students had  significantly higher knowledge  [mean  score 8.06 (SD 1.88)] than the Kuwaiti  and  Jordanian  students  [mean  scores  7.79  (SD  2.20)  and  7.18  (SD  1.89)  respectively].  HIV shame and stigma measures Table  2  shows  the means  scores  for  the shame and stigma items from the 3  countries. Looking at  the stigma  items  the highest mean scores for the Kuwaiti  and Bahraini students was  for  the  item  about  isolating HIV-positive  people  from society. For the Jordanian students,  the highest mean score was  for having  no objection to marrying  into a  family  with HIV/AIDS. For  the shame  items,  the highest mean score for all 3 groups  was  agreement with  the  statement  “I  would  be  ashamed of  having  people  with HIV/AIDS in my country”.  Cronbach  alpha  values  for  the  shame measure for the 3 countries were  acceptable (0.74, 0.62 and 0.77  for  the  Kuwaiti,  Bahraini  and  Jordanian  stu- dents  respectively). There were no sta- tistically significant differences between  the 3  samples  in  the  shame measure  [ANOVA, F (2, 274) = 1.30, P = 0.23].  For the stigma measure, the reliabili- ties for all samples were acceptable and  comparable  (Cronbach  alpha  values  were 0.72,  0.72  and 0.75  for  the Ku- waitis, the Bahrainis and the Jordanians  respectively). There were no statistically  significant differences among  the sam- ples [F (2, 274) = 0.92, P = 0.40].  Correlations between measures For  the Bahraini  students,  the knowl- edge  index was negatively  correlated  with the AIDS stigma scale (r = –0.38)  but was unrelated  to  the AIDS-related  shame items (r = –0.08). In the women  from Kuwait  and  Jordan,  the  knowl- edge  index was uncorrelated with  the  AIDS stigma scale (r = –0.09, r = –0.15  respectively) and AIDS-related shame  scale (r = – 0.08, r = –0.12 respectively).  However,  the AIDS  stigma  scale  and  the  AIDS-related  shame  scale  were  positively  correlated  for  all  3  samples  (Bahrainis  r = 0.49, Kuwaitis  r = 0.59,  Jordanians r = 0.59). Multiple regression analysis Multiple  regression  analysis was per- formed  to predict participants’  level of  tolerance  towards people with HIV/ AIDS based on their knowledge of HIV  and AIDS and AIDS-related shame for  each sample (Table 3).  For  the  Kuwaiti  students  AIDS- related  shame was  the only predictor  of  stigmatization of people with HIV/ AIDS  (β =  0.59, P  <  0.001).  For  the  Bahraini women AIDS-related  shame  was  the best predictor of AIDS stigma  (β = 0.46, P < 0.001). Students who per- ceived AIDS to be a shameful disease ex- pressed more negative attitudes toward  people with HIV/AIDS. Knowledge of  HIV and AIDS was also a good predic- tor  (β = – 0.34, P  < 0.001):  students  Table 2 Mean scores of stigma and shame items of female college students in 3 Arab countries Item Kuwait (n = 83) Bahrain (n = 108) Jordan (n = 86) Mean score SD Mean score SD Mean score SD Stigma People with HIV/AIDS should be fired from their jobs 3.53 1.27 3.24 1.18 3.15 1.28 People with HIV/AIDS should be quarantined 4.09 1.03 3.88 1.07 3.52 1.29 I am sympathetic toward people with HIV/AIDS 2.54 1.15 3.02 1.16 2.82 1.19 I have no objection to marrying into a family with HIV/AIDS 3.81 1.22 3.58 1.29 4.29 1.01 I do not mind being touched by someone with HIV/AIDS 3.97 1.21 3.73 1.23 4.17 1.07 All people with HIV are responsible for their infection 2.50 1.36 2.30 1.27) 2.08 1.40 If a friend of mine got infected with HIV I would being a friend with him/her 2.89 1.33 3.23 1.19 3.18 1.21 Shame Ashamed of having a relative with HIV/AIDS 3.61 1.18 3.53 1.19 3.26 1.44 People with HIV/AIDS should be ashamed of themselves 3.32 1.31 3.23 1.18 3.17 1.31 Ashamed of having people with HIV/AIDS in my country 3.69 1.18 3.58 1.17 3.46 1.41 SD = standard deviation. EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 986 with greater knowledge about HIV and  AIDS  tended  to view  individuals with  HIV/AIDS less negatively. In Jordanian  students,  AIDS-related  shame was  a  significant predictor of AIDS stigma (β  = 0.58, P < 0.001) (Table 3).  Discussion Our  results  suggest  that  participants  had  inaccurate knowledge of  the kinds  of behaviours that could transmit HIV.  Bahraini participants were more knowl- edgeable  than  the  participants  from  both  Kuwait  and  Jordan.  However,  the majority of participants  from all 3  countries were unable  to  identify  the  correct answers  to  several  items. Their  lack  of  knowledge  about  2  items  in  particular  is alarming. All 3 groups had  trouble providing  the correct  answers  to  the  items on  condom use  and  the  possible transmission of HIV in semen.  More  specifically, more  than 75% of  the participants in each sample did not  know or gave  the wrong answer about  the role of condoms in minimizing the  risk of HIV infection. Similarly, 73% of  Kuwaiti,  78% of Bahraini  and 49% of  Jordanian female students did not know  that HIV can be carried in semen.  One possible explanation  for  these  results  is  the  taboos  regarding  the  topic  of  sexuality  in  the Arab world.  Women are discouraged  from seeking  information  about  sex  and  about  the  proper way to protect themselves from  sexually  transmitted  infections  [7].  In  addition,  there  is almost no education  about  sex  in  the  schools  [9]. Another  plausible  reason  for  these specific gaps  in knowledge about HIV and AIDS  is  that HIV prevention efforts in the Arab  world focus strongly on abstinence, and  place little if any emphasis on the role of  condoms [22].  Knowledge  about HIV and AIDS  played an influential role in the Bahraini  women’s  reactions  toward people with  HIV/AIDS, but not in those of the other  2  groups. This  is  not  surprising  since  the Bahraini women  in  this  study had  greater knowledge about HIV and AIDS.  The findings  therefore  indicate  that ac- curate knowledge about HIV and AIDS  can play a powerful  role  in minimizing  the stigmatization of people with HIV/ AIDS. Therefore,  as  shown elsewhere  [23,24],  improving HIV and AIDS-re- lated knowledge  is one way to decrease  AIDS stigma in the Arab region.  The most  intriguing  and  unique  feature  of  this  study was  the  role  of  shame  in predicting  the negative  atti- tudes towards people with HIV/AIDS.  Whereas  the HIV  and AIDS-related  knowledge was predictive of negative  attitudes  toward  persons with HIV/ AIDS only among the Bahraini women,  shame related  to AIDS was predictive  across all 3  samples. The  link between  shame  and AIDS may be due  to  the  association between HIV/AIDS and  such socially undesirable activities such  as unlawful  sex and drugs  in  the Arab  world.  If  so,  people with HIV/AIDS  are considered a  source of  shame and  should be avoided for the sake of one’s  reputation.  Studies  about AIDS  stigma  have  reported  that  family  and caregivers of  people  with HIV/AIDS  suffer  from  courtesy stigma (stigma by association)  and have been the target of ridicule and  discrimination [25,26].  Courtesy  stigma may  be  particu- larly detrimental in Arab culture because  shame is an especially powerful emotion  that plays  a major  role  in  this  society.  Indeed, Arab culture has been labelled a  shame-oriented culture because shame is  an intensely feared emotion and often re- sults in the social isolation of the shamed  individual, if not their murder [27]. Thus,  it seems that fear of shame, not for one’s  wrongdoing,  but  for being  associated  with  someone with HIV/AIDS,  is  the  main reason women in this study stigma- tized people with HIV/AIDS.  Because  our  study  focused  on  young college women with small con- venience samples, The ability to gener- alize  from  it  is  limited. The findings of  this study, however, point to a need for  more  studies  that examine  the  role of  shame  in  the  stigmatization of people  with HIV/AIDS. Studies with bigger  samples of both males and females and  from other countries in the Arab world  are needed.  Conclusion This study provides evidence of an ur- gent need  to provide HIV education  to young people  in Arab countries, es- pecially women,  for their own sake and  to  reduce AIDS stigma. Reducing  the  shame associated with HIV/AIDS may  prove to be one of the key strategies to- wards improving the situation of people  with HIV/AIDS in the Arab world.  Table 3 Multiple regression analysis of AIDS stigma of female college students in 3 Arab countries Predictor B SE B β Kuwait AIDS-related shame 0.432 0.067 0.587** Knowledge of AIDS –0.014 0.030 –0.043 Bahrain AIDS-related shame 0.380 0.064 0.464** Knowledge of AIDS –0.133 0.031 –0.341** Jordan AIDS-related shame 0.385 0.060 0.576** Knowledge of AIDS –0.034 0.036 –0.083 R2 = 0.35 for Kuwait, 0.36 for Bahrain and 0.35 for Jordan. ***P < 0.001. SE = standard error. طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 987 Fortenberry JD et al. Relationships of stigma and shame to gon-1. orrhea and HIV screening. American Journal of Public Health, 2002, 92(3):378–381. Kalichman SC, Simbayi LC. HIV testing attitudes, AIDS stigma, 2. and voluntary HIV counselling and testing in a black township in Cape Town, South Africa. Sexually Transmitted Infections, 2003, 79(6):442–447. Rintamaki LS et al. Social stigma concerns and HIV medication 3. adherence. AIDS Patient Care and STDs, 2006, 20(5):359–368. Roudi-Fahimi F. 4. Time to intervene: preventing the spread of HIV/ AIDS in the Middle East and North Africa. Washington DC, Popu- lation Reference Bureau, 2007:1–8. Holzemer WL, Uys LR. Managing AIDS stigma. 5. Journal of Social Aspects of HIV/AIDS, 2004, 1(3):165–174. Akala F, Jenkins C. 6. Preventing HIV/AIDS in the Middle East and North Africa: a window of opportunity to act (Orientations in De- velopment Series). Washington DC, World Bank, 2005:1–110. DeJong J et al. The sexual and reproductive health of young 7. people in the Arab countries and Iran. Reproductive Health Mat- ters, 2005, 13(25):49–59. Arab human development report: challenges to human security 8. in the Arab countries. New York, United Nations Development Programme, Regional Bureau for Arab States, 2009. El Feki S. Middle-Eastern AIDS efforts are starting to tackle ta-9. boos. Lancet, 2006, 367:975–996. Petro-Nustas W. University students’ knowledge of AIDS. 10. Inter- national Journal of Nursing Studies, 2000, 37(5):423–433. Petro-Nustas W, Kulwicki A, Zumout AF. Students’ knowledge, 11. attitudes, and beliefs about AIDS: a cross-cultural study. Jour- nal of Transcultural Nursing, 2002, 13(2):118–125. Gańczak M et al. Break the silence: HIV/AIDS knowledge, at-12. titudes, and educational needs among Arab university students in United Arab Emirates. Journal of Adolescent Health, 2007, 40(6):572.e1–572.e8. Al-Ghanim SA. Exploring public knowledge and attitudes to-13. wards HIV/AIDS in Saudi Arabia. A survey of primary health care users. Saudi Medical Journal, 2005, 26(5):812–818. Al-Owaish R et al. Knowledge, attitudes, beliefs, and prac-14. tices about HIV/AIDS in Kuwait. AIDS Education and Prevention, 1999, 11(2):163–173. References Al-Serouri AW et al. Knowledge, attitudes and beliefs about 15. HIV/AIDS in Sana’a, Yemen. Eastern Mediterranean Health Journal, 2002, 8(6):706–715. Badahdah A. Saudi attitudes towards people living with HIV/16. AIDS. International Journal of STD and AIDS, 2005, 16(12):837– 838. Link BG, Phelan JC. Conceptualizing stigma. 17. Annual Review of Sociology, 2001, 27:363–385. Badhadah A, Sayem N, Foote C. Development of a Yemeni 18. AIDS Stigma Scale. AIDS Care, 2009, 21(6):754–759. Cunningham SD et al. Attitudes about sexual disclosure and 19. perceptions of stigma and shame. Sexually Transmitted Infec- tions, 2002, 78(5):334–338. Goh D. 20. The development and reliability of the Attitudes Towards AIDS Scale. Paper presented at the Annual Meeting of the American Psychiatric Association, San Francisco, May 16–21, 2009. Nyblade LC. Measuring HIV stigma: existing knowledge and 21. gaps. Psychology, Health and Medicine, 2006, 11(3):335–345. Madani TA et al. Epidemiology of the human immunodefi-22. ciency virus in Saudi Arabia: 18-year surveillance results and prevention from an Islamic perspective. BMC Infectious Dis- eases, 2004, 4:25. Chen J et al. The effects of individual- and community-level 23. knowledge, beliefs, and fear on stigmatization of people living with HIV/AIDS in China. AIDS Care, 2007, 19(5):666–673. Brown L, Trujillo L, Macintyre K. Interventions to reduce HIV/24. AIDS stigma: what have we learned? AIDS Education and Pre- vention, 2001, 15(1):49–69. Mwinituo PP, Mill JE. Stigma associated with Ghanaian caregiv-25. ers of AIDS patients. Western Journal of Nursing Research, 2006, 28(4):369–382. Greene K, Banerjee SC. Disease-related stigma: comparing 26. predictors of AIDS and cancer stigma. Journal of Homosexual- ity, 2006, 50(4):185–209. Glidden HW. The Arab world. 27. American Journal of Psychiatry, 1972, 128(8):984–988. EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 988 Community pharmacists’ knowledge, attitudes and practices towards herbal remedies in Riyadh, Saudi Arabia K.M. Alkharfy1 ABSTRACT There is an increasing trend towards consumption of complementary and alternative herbal products in many parts of the world. A cross-sectional sample of 115 community pharmacists in Riyadh, Saudi Arabia was visited and information on knowledge, attitudes and practices towards herbal remedies was collected using a structured questionnaire. All pharmacists acknowledged dispensing herbal products through their pharmacies. Ginseng was the most widely used product (47%), followed by ginkgo (23%), valerian (17%) and St John’s wort (3.5%). In general, pharmacists had poor awareness about potential herb–drug interactions. While 56% of participating pharmacists expressed concerns about the safety of herbal remedies, 30% considered them to be harmless. Community pharmacists need to be better informed about herbal products. 1Department of Clinical Pharmacy, College of Pharmacy, King Saud University, Riyadh, Saudi Arabia (Correspondence to K.M. Alkharfy: alkharfy@ ksu.edu.sa). Received: 30/01/09; accepted: 05/04/09 ةيدوعسلاب ضايرلا ةنيدم في ةيبشعلا تاجلاعلا هاتج ينيعمتجلما ةلدايصلا تاسراممو فقاومو فراعم فيرلخا دممح نب دلاخ 115 نم ةفّلؤم ةيضْرَع ةيعطقم ةنيع تيرتخا دقو .لماعلا ءاحنأ نم يرثك في ةليدبلا وأ ةيليمكتلا ةيبشعلا تاجتنلما كلاهتسا في ةدايز كانه :ةصلالخا نع ةيبشعلا تافصولا هاتج متهاسراممو مهفقاومو مهفراعم لوح تامولعلما عملج مله تارايز تيرجأو ،ةيدوعسلاب ضايرلا ةنيدم في ًايعمتمج ًايلديص هيلي ،)%47( ًلاوادت رثكلأا جتنلما وه غنسنلجا ناكو .متهايلديص في ةيبشع تاجتنم نوفصري منهأب ةلدايصلا عيجم فرـتعاو .يجهنم نايبتسا قيرط باشعلأا ينب ثدتح دق يتلا ةلدابتلما تلاعافتلاب ةليئض ةلدايصلا ةيارد تناكو .)%3.5( ترو نوج تنسو ،)%17( نايرلافلا مث ،)%23( وغكنلجا .رضرلا ةميدع ةيبشعلا تافصولا نأ مهنم %30 برتعا ،ةيبشعلا تافصولا ةملاس لوح قَلَقلا ضعب نع ةلدايصلا نم %56 برعأ ينح فيو .ةيودلأاو .ةيبشعلا تاجتنلما لوح مهفراعم ينستح لىإ ينّـِ يعمتجلما ةلدايصلا ةجاح لىع ةساردلا ُّلدتو Connaissances, attitudes et pratiques des pharmaciens communautaires vis-à-vis des médicaments à base de plantes à Riyad (Arabie saoudite) RÉSUMÉ Dans de nombreuses parties du monde, on observe une tendance croissante à la consommation de médicaments complémentaires et alternatifs à base de plantes. Des visites réalisées auprès d’un échantillon transversal de 115 pharmaciens communautaires de Riyad, en Arabie saoudite, et la distribution à ces derniers d’un questionnaire structuré ont permis de recueillir des informations sur leurs connaissances, leurs attitudes et leurs pratiques vis-à-vis des médicaments à base de plantes. Tous ont reconnu délivrer des produits à base de plantes dans leur officine. Le ginseng s’est avéré le plus utilisé (47 %), suivi par le ginkgo (23 %), la valériane (17 %) et le millepertuis (3,5 %). En général, les pharmaciens étaient peu conscients des éventuelles interactions entre plantes et médicaments. Alors que 56 % des pharmaciens participants se sont montrés préoccupés par la sécurité des médicaments à base de plantes, 30 % considèrent ceux-ci comme sans danger. Les pharmaciens communautaires doivent être mieux informés sur les produits à base de plantes. طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 989 Introduction In many  parts  of  the world  the  con- sumption of herbal products is increas- ing,  especially  for  the management of  chronic diseases  [1,2].  Studies  in  the  United States reported a 380% increase  in the use of herbal remedies by patients  between 1990 and 1997,  and a 350%  increase  in  the  use  of  herbal  supple- ments between 1996 and 1999  [3,4].  Furthermore,  it  is now estimated  that  30%–50% of  the population of devel- oped countries are using complemen- tary and alternative medicine (CAM)  products, and nearly US$ 4 billion are  spent on purchasing herbal products  every year [5–7].  Despite  this high  rate of use,  fewer  than  half  of  patients who  use  herbal  products discuss  the use of  them with  their clinicians [8,9]. Patients often do  not consider  these agents  to be medi- cines because of  their  natural  source  or  they may believe  that  their  health  care providers  lack  an understanding  of, or have negative attitudes  towards,  herbal  agents.  In addition, health care  providers may neglect to ask about use  of  herbal  products  [10].  In  a  review  of 67 patients hospitalized  for asthma  treatment, none of the patients’ records  documented the use of CAM, although  interviews  indicated  that over 40% of  the patients had used herbal remedies as  a treatment for asthma [10,11].  The  health  care  system  in  Saudi  Arabia,  as  in most Middle East  coun- tries, is primarily based on conventional  medicine. However,  traditional  local  remedies  continue  to  be  very  popu- lar among Saudis  [12,13].  In addition,  there  is  an  increasing  trend  towards  the use of new,  imported herbal prod- ucts such as St John’s wort, valerian and  many others. This  is  a  concern  for  lo- cal health care professionals. One clear  difference  between  traditional  herbs  and these newer remedies is their avail- ability  in pharmaceutical dosage  forms  such as  tablets,  capsules  and extracts.  This  has  translated  into  provision  of  herbal  products  through  community  pharmacies  rather  than  folk medicine  shops where traditional herbal remedies  are usually  sold  in  their original crude  forms.  Community  pharmacists,  as  the  most  accessible health  care providers  [14],  are  receiving more queries  from  patients  about  herbal  products  than  ever before  [15]. Unfortunately,  infor- mation about community pharmacist’s  knowledge of herbal products  in Saudi  Arabia  is  lacking. Therefore,  the  aim  of  the current  study  in Riyadh was  to  assess community pharmacists’ knowl- edge, attitudes and dispensing practices  concerning herbal products.  Methods A cross-sectional survey of community  pharmacists in Riyadh city was conduct- ed over a period of 3 months between  January and March 2008. Riyadh is the  capital and biggest metropolitan city of  Saudi Arabia with a population of about  5 million people.  Sample A total of 115 community pharmacies in  Riyadh city were randomly selected for  visits based on their geographical distri- bution (i.e., north, south, east, and west).  They  represented about 15%–20% of  all community pharmacies in the city. A  typical community pharmacy  in Saudi  Arabia is usually operated by 1 licensed  pharmacist and an assistant who work  for about 8–12 hours per day, 6 days a  week. Data collection A  structured  questionnaire  was  de- signed by  the author and consisted of  3 main sections: demographic informa- tion, pharmacist’s beliefs and dispensing  practices regarding herbal products and  a knowledge assessment section. A pilot  testing was conducted  to assure ques- tionnaire’s validity.  Questions on  the  types of  herbal  products  available  in  the  pharmacy,  reasons why consumers acquire  these  products  and  the  rate  of  dispensing  them were  included  in  the question- naire. The demographic characteristics  (sex  and age group) of  consumers of  herbal medicines were  estimated  by  pharmacists.  The pharmacist’s  knowledge  sec- tion  contained  several measures. The  first was a 5-point  scale  in which  they  rated  their  own  knowledge  about  herbal products  from 1 (excellent)  to  5  (poor). Questions  on  their  aware- ness to possible adverse effects of herbal  products  and herb–drug  interactions  was  assessed  by  3  options  (yes/no/ don’t  know). The questionnaire  also  included questions  about  the  source  of  information  that pharmacists usu- ally utilize to update their knowledge of  herbal products. The knowledge com- ponent was further tested using a set of  closed-book multiple choice questions  relating to assumed indications, adverse  reactions,  precautions  and  important  herb–drug  interactions. For  example:  “St John’s wort may decrease the blood  level of which of the following drugs: a)  aspirin; b) digoxin; c) ceftriaxone; d)  ranitidine?” Finally,  professional  attitudes  to- wards  these  remedies were  assessed  with a 5-point  scaling  ranging  from 1  (very effective)  to 5 (not  sure).  In ad- dition,  the pharmacists’ personal views  about  the  safety of herbal products as  therapeutic alternatives were  included  in the attitude assessment.  The selected pharmacies were vis- ited  and  the questionnaire was  filled  during a 10-minute interview. Participa- tion was anonymous.  Analysis The data was entered  into a Microsoft  Access database and  the  frequencies of  consumption of herbal products  and  consumers’ characteristics and reasons  for use of products were tabulated. The  knowledge  section was  scored  as  the  EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 990 percentage of correct answers  for each  respondent. The chi-squared  test was  used  for 2-way comparisons  for nomi- nal  and  categorical  data which were  analysed using SPSS, version 11. Results Background characteristics A total of 115 interviews were conduct- ed, a 100% response rate. The mean age  of participants was 33 (standard devia- tion 7) years, with a median duration of  7 years in practice.  All  pharmacists  who  were  inter- viewed  acknowledged  that  they  dis- pensed herbal products  to consumers.  Almost  two-thirds  of  pharmacists  (63%)  reported  receiving  requests  for  herbal products daily, 32% weekly and  5% monthly. Commonly dispensed herbal remedies Ginseng was the most widely dispensed,  with 47% of pharmacists reporting that  they dispensed this product, followed by  ginkgo (23%) and valerian (17%) (Ta- ble 1). The main reasons for consumers  to  use  herbal  remedies  according  to  pharmacists was  for  “boosting energy”  or  “enhancing physical performance”  (reported by 49% of pharmacists). Oth- er health-related problems  for which  herbal products were  sought  included  poor mental alertness (19%), insomnia  (17%),  anxiety  (5%),  and  low mood  (1%).  The majority of those seeking herbal  products  were  reported  by  pharma- cists  to be males, while  females were  only 24%. Pharmacists  reported  that  the majority of consumers were middle  aged  (around 47% of herbal product  users),  followed by  the elderly  (40%)  and young people (13%).  Pharmacists’ views on herbal prod- ucts  as  alternative  remedies  showed  that a considerable proportion (53%)  believed them to be effective or very ef- fective (Figure 1). Despite this attitude,  about half  of  the pharmacists  (56%)  expressed concerns about  the safety of  herbal remedies, 30% thought they were  harmless and 14% were not sure about  their safety.  Knowledge about herbal remedies When participating pharmacists were  asked to self-rate their knowledge about  herbs, 12% reported it as excellent, 34%  as very good, 35% as good and 19% as  fair. While  the majority believed  that  their knowledge was good to excellent,  the mean score of  the multiple-choice  questions used  to assess acquaintance  was 68% out of 100% (95% confidence  interval: 63%–73%), with possible ad- verse effects being the most well-known  (Table 2). The majority of pharmacists  (73%) knew about possible herb–drug  interactions,  whereas  14%  reported  no known  interactions and 13% were  uncertain. Although many of  the phar- macists acknowledged the existence of  herb–drug  interactions,  they did not  demonstrate a strong knowledge of the  well-established  interactions  reported  in the literature between selected herbs  and common drugs. A  cross-tabulation  using  the  chi- squared  test was used  to assess differ- ences  in  the  frequency of pharmacists  who had the most positive attitudes to- wards herbal products effectiveness (i.e.  very effective and effective) versus those  who were  less persuaded  about  their  effectiveness according  to pharmacists’  self-rated knowledge. Similarly, concern  about herb–drug  interactions and self- perceived knowledge was  tested using  the same approach. Those pharmacists  who considered themselves to be well- informed  about herbal  products had  more positive attitudes  towards  the ef- fectiveness of these alternative remedies  as  therapeutic options  than  those with  less positive attitudes (33% versus 13%,  P < 0.001). Nevertheless, pharmacists’  self-rated knowledge did not affect their  concerns about herb–drug interactions  (38% versus 46%, P = 0.276). Product  package  instructions  were  the most  commonly used source of  information  by community pharmacists (40%), fol- lowed by  information obtained  from  product  representatives (33%),  formal  pharmacy education (21%) and in a few  cases via the Internet (6%).  Discussion It has been reported that 24% of patients  attending a health centre in Saudi Arabia  had used a local alternative remedy [16].  Table 1 Pharmacists’ reports of herbal products commonly dispensed through community pharmacists and reasons why consumers use them Item % of respondents (n = 115) Herbal products commonly dispensed Ginseng 47 Ginkgo 23 Valerian 17 St John’s wort 4 Other (e.g. Echinacea, devil’s claw, horse chestnut, passionflower, guarana) 10 Reasons why consumers use herbal products Vitality 49 Mental alertness 19 Insomnia 17 Anxiety 5 Low mood 1 Other (e.g. boosting immunity in children, arthritic pain, varicose veins, haemorrhoids, weight loss) 9 طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 991 Furthermore,  the  use  of  local  herbal  products  reached  17%–33%  among  Saudi diabetics, and up to 73% of users  did not inform their health care provid- ers regarding their use of herbs [17]. As  there  is a wide belief  in CAM in Saudi  Arabia, and the community pharmacist  is  generally  the  first  health  care  pro- vider from whom patients seek recom- mendations  on  new  products, more  information about CAM is needed by  community  pharmacists.  To  the  au- thor’s knowledge,  this  is  the first  study  to  examine  community  pharmacists’  attitudes,  knowledge  and prescribing  practices concerning herbal remedies in  Saudi Arabia.  The current study revealed that about  half of  the community pharmacists  in  this  sample  in Riyadh (57%) consider  herbal products  as potentially unsafe,  whereas  a  considerable  proportion  (30%) deemed  them  to be harmless.  This was found despite the finding that  53%  of  pharmacists  believed  herbal  remedies were effective or very effective.  This generally positive attitude towards  herbal  remedies  as  safe  and  effective  may  influence patients, who may  feel  more  comfortable  using  these  prod- ucts with  their prescribed medications.  Although  it  is  likely  that  community  pharmacists  in Saudi Arabia have suffi- cient information about the indications  for use of commonly dispensed herbs,  they are  less  likely  to be well-informed  about potential herb–drug interactions.  These results are similar to those found  among pharmacists in the United States  [18–20].  The  recent  trend  towards  greater  use of herbal products and the attitudes  of pharmacists  towards  this  trend have  been evaluated by  some  investigators  [20,21]. While some health care provid- ers acknowledge  that CAM may be of  use, most know little about it [22]. Aba- hussain et  al. have examined pharma- cists’  attitudes and awareness  towards  the use  and  safety of herbal products  in Kuwait and found that about 31% of  the pharmacists did not have enough  information  about  potential  interac- tions between herbs and conventional  medicines  [23]. Similarly,  the present  study revealed that about 27% of com- munity pharmacists in Saudi Arabia had  little or no knowledge about important  herb–drug interactions.  The study also  found  that ginseng  was  the most commonly dispensed al- ternative remedy by many community  pharmacies in Riyadh, being dispensed  daily  in  many  instances.  Ginseng  is  claimed  to  improve overall well-being  and  enhance  stress  resistance  [24]  and was mostly used  to boost  energy  according  to  the  interviewed pharma- cists.  Several  adverse  reactions  have  been  reported with ginseng,  including  insomnia,  hypoglycaemia,  hyperten- sion and  skin  reactions  [25]. Further- more, many  prescription  drugs  have  been reported to  interact with ginseng,  including oral hypo glycaemic  agents,  warfarin, nifedipine and  loop diuretics  [26]. Several herbal products  found  to  be widely used in the current study have  also  demonstrated  significant  herb– drug  interactions,  including  increased  risk of bleeding with warfarin  therapy  in  combination with  ginkgo  and  in- creased digoxin  levels with St.  John’s  wort [26–28].  Finally,  it  is worth noting  that  the  prevalence of chronic  illnesses  such as  diabetes and cardiovascular diseases  is  rapidly  increasing  in Saudi Arabia due  to  a number of  factors  including  life- style changes [29,30]. This has resulted  Table 2 Pharmacists’ knowledge of indications, precautions and interactions of selected herbs Statement % of respondents (n = 115) Correct Incorrect Echinacea is used to boost immunity 91 9 Ginseng may increase blood pressure 90 10 Ginseng should be avoided in diabetic patients 90 10 St John’s wort is commonly used for mild to moderate depression 82 18 Ginkgo is claimed to delay dementia 78 23 Ginkgo can increase the risk of bleeding when combined with warfarin 60 40 St John’s wort may increase blood digoxin level 30 70 Valerian should be used cautiously in patients using benzodiazepines 28 73 % of pharmacists 0 10 20 30 40 50 Uncertain Very effective Effective Somewhat effective Not effective 0 38 15 45 2 Figure 1 Pharmacists’ attitudes toward herbal products as alternative therapeutic options (n = 115) EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 992 in  an  expanding  number  of  patients  taking multiple  medications,  with  a  consequently  greater  potential  for  clinically  significant  interactions when  herbal  remedies  are  also  consumed.  Therefore,  this  potential  problem  in  drug therapy needs further attention by  pharmacists  if  they are  recommending  these products to patients. In particular,  community pharmacists  in Saudi Ara- bia need to be vigilant while establishing  a dialogue with their patients regarding  the use of herbal products with prescrip- tion medications. Patients  should be  informed  about  possible  effects  and  closely monitored.  In order  to achieve  these goals, continuing education pro- grammes with more  information  on  the safety and potential harmful effects  of some CAM products should be pro- vided to practising pharmacists.  Furthermore,   a l l   community  pharmacies  in Saudi Arabia  should be  equipped with  evidence-based  refer- ences on herb–drug  interactions,  such  as a  recent herbal  textbook and/or an  updated computer database on natural  products.  Some  important  limitations should  be  considered when  interpreting  the  results of  the  current work. First,  in  a  cross-sectional  study, data collection  is  limited to a single time point, so changes  over  time were not assessed. Secondly,  there  may  have  been  errors  by  the  pharmacists in recollecting the type and  frequency of dispensed herbs as well as  consumers’  characteristics.  In  spite of  these limitations, the present study lays  a foundation for further work and guides  the selection of appropriate measures to  assess pharmacists’ knowledge and pat- terns of practice of herbal products  in  Saudi Arabia in the future.  Astin JA. Why patients use alternative medicine: results of a na-1. tional study. Journal of the American Medical Association, 1998, 279(19):1548–1553. O’Hara M et al. A review of 12 commonly used medicinal herbs. 2. Archives of Family Medicine, 1998, 7(6):523–536. Eisenberg DM et al. Trends in alternative medicine use in 3. the United States, 1990–1997: results of a follow–up national survey. Journal of the American Medical Association, 1998, 280(18):1569–1575. Gordon NP, Lin TY. Use of complementary and alternative 4. medicine by the adult membership of a large northern Califor- nia health maintenance organization, 1999. Journal of Ambula- tory Care Management, 2004, 27(1):12–24. Brevoort P. The booming US botanical market: a new over-5. view. HerbalGram, 1998, 44:33–46. Canedy D. Real medicine or medicine show? Growth of herbal 6. remedy sales raises issues about value. New York Times, 1998, July 23:C1. Johnston B. One-third of nation’s adults use herbal remedies: 7. market estimated at 3.24 billion. Herbalgram, 1997, 40:49. Eisenberg DM et al. Unconventional medicine in the United 8. States. Prevalence, costs, and patterns of use. New England Journal of Medicine, 1993, 328(4):246–252. Martin-Facklam M et al. Undeclared exposure to St. John’s wort 9. in hospitalized patients. British Journal of Clinical Pharmacol- ogy, 2004, 58(4):437–441. Cockayne NL et al. Health professionals rarely record history 10. of complementary and alternative medicines. British Journal of Clinical Pharmacology, 2005, 59(2):254–258. References Rivera JO et al. Herbals and asthma: usage patterns among 11. a border population. Annals of Pharmacotherapy, 2004, 38(2):220–225. Al-Awamy BH. Evaluation of commonly used tribal and tradi-12. tional remedies in Saudi Arabia. Saudi Medical Journal, 2001, 22(12):1065–1068. Al-Saeedi M et al. Patterns of belief and use of traditional 13. remedies by diabetic patients in Mecca, Saudi Arabia. Eastern Mediterranean Health Journal, 2003, 9(1–2):99–107. Your pharmacist serves you and your community14. . Alexan- dria, Virginia, Alliance for Pharmaceutical Care, 2003 [online brochure] (http://www.pharmacist.com/AM/ Template.cfm?Section=Search1&section=Alliance_For_ Pharmaceutical_Care&template=/CM/ContentDisplay. cfm&ContentFileID=2696). Clauson KA et al. Knowledge and attitudes of pharmacists in 15. Missouri regarding natural products. American Journal of Phar- maceutical Education, 2003, 67(2):301–309. Ajaji N et al. Prevalence of utilization of native medicine 16. among primary care consumers. Saudi Medical Journal, 1998, 19:551–554. Al-Rowais NA. Herbal medicine in the treatment of diabetes 17. mellitus. Saudi Medical Journal, 2002, 23(11):1327–1331. Bouldin AS et al. Pharmacy and herbal medicine in the US. 18. Social Science & Medicine, 1999, 49(2):279–289. Rickert K et al. Pharmacist knowledge of common herbal 19. preparations. Proceedings of the Western Pharmacology Society, 1999, 42:1–2. Conclusions Despite  the high provision of  herbal  remedies through community pharma- cies  in Saudi Arabia,  the knowledge of  pharmacists about herb–drug  interac- tions was only moderate. Many com- munity pharmacists in Riyadh believed  that herbal products were an effective  therapeutic option and a considerable  proportion of pharmacists thought they  were  safe.  Community  pharmacists  need to be better informed about herbal  products and better able  to advise and  identify  patients who  are  seeking  al- ternative  remedies  for  chronic health  problems  so as  to  refer  them to other  health care providers if necessary.  Acknowledgements The author would like to thanks all phar- macists who participated in the study. طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 993 Practical aspects of signal detection in pharmacovigilance In recent years public expectations for rapid identification and prompt management of emerging drug safety issues  have grown swiftly. Over a similar timeframe, the move from paper-based adverse event reporting systems to electronic  capture and rapid transmission of data has resulted in the accrual of substantial datasets capable of complex analysis and  querying by industry, regulators and other public health organizations. Practical aspects of signal detection in pharmacovigilance aims primarily to provide a comprehensive resource for those  considering how to strengthen their pharmacovigilance systems and practices, and to give practical advice. But the  report does not specify instant solutions. These will inevitably be situation-specific and require careful consideration  taking into account local needs.  The report also looks ahead and anticipates a number of ongoing developments,  including techniques with wider applicability to other data forms than individual case reports. Further information about this and other WHO publication is available at: http://www.who.int/publications/en/  Chang ZG et al. Pharmacists’ knowledge and attitudes to-20. ward herbal medicine. Annals of Pharmacotherapy, 2000, 34(6):710–715. Koh HL et al. Pharmacists’ patterns of use, knowledge, and 21. attitudes toward complementary and alternative medicine. Journal of Alternative and Complementary Medicine, 2003, 9(1):51–63. Nelson MV et al. Pharmacists’ perceptions of alternative health 22. approaches––a comparison between US and British phar- macists. Journal of Clinical Pharmacy and Therapeutics, 1990, 15(2):141–146. Abahussain NA et al. Pharmacists’ attitudes and awareness to-23. wards the use and safety of herbs in Kuwait. Pharmacy Practice, 2007, 5(3):125–129. Kitts D, Hu C. Efficacy and safety of ginseng. 24. Public Health Nutri- tion, 2000, 3(4A):473–85. Coon JT, Ernst E. Panax ginseng: a systematic review of adverse 25. effects and drug interactions. Drug Safety, 2002, 25(5):323– 344. Izzo AA et al. Cardiovascular pharmacotherapy and herbal 26. medicines: the risk of drug interaction. International Journal of Cardiology, 2005, 98(1):1–14. Zhou S et al. Interactions of herbs with cytochrome P450. 27. Drug Metabolism Reviews, 2003, 35(1):35–98. Zhou S et al. Herbal modulation of P28. -glycoprotein. Drug Me- tabolism Reviews, 2004, 36(1):57–104. Al29. -Nozha MM et al. Hypertension in Saudi Arabia. Saudi Medi- cal Journal, 2007, 28(1):77–84. Al30. -Nozha MM et al. Diabetes mellitus in Saudi Arabia. Saudi Medical Journal, 2004, 25(11):1603–1610. EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 994 Congé maternité et vécu des mères qui travaillent au Liban N. Saadé,1 B. Barbour 1 et P. Salameh 1 RÉSUMÉ Une étude transversale a été menée auprès de 802 mères libanaises, afin d’évaluer l’effet du retour rapide au travail sur leur santé et celle de leur enfant. Les pratiques d’allaitement maternel ont aussi été évaluées. La durée du congé maternité est considérée comme insuffisante pour 72,8 % des femmes. Le retour rapide de la mère au travail provoquerait plusieurs problèmes physiques et psychiques, selon le secteur de travail. Le temps moyen de l’allaitement est de 4,7 mois et la moyenne de l’allaitement souhaité va jusqu’à 10,9 mois. L’allaitement dépend de la durée du congé de maternité, de la possibilité de pauses d’allaitement et de la présence de garderies sur le lieu du travail. Des interventions urgentes sont nécessaires pour prolonger la durée du congé de maternité et promouvoir l’allaitement chez les femmes qui travaillent. 1Université Libanaise, Faculté de Santé publique, Section II, Fanar, Beyrouth (Liban) (Correspondance à adresser à P. Salameh : psalameh@ul.edu.lb). Reçu : 09/11/09; accepté : 23/12/09 نانبل في تلاماعلا تادلاولل ةموملأا تازاجإب ةقلعتلما تابرلخا ةملاس لاكساب ،روبرب تيدانرب ،هداعس انين تمَّيق ماك .نلهافطأ ةحصو نهتحص لىع لمعلا لىإ نتهدوع ةعسر رثأ مييقتل ةينانبل ًةدلاو 802 لىع ةيضْرَع ةسارد تاثحابلا ترجأ :ةصلالخا لمعلا لىإ تادلاولا ةدوع ةعسر تدأو .ءاسنلا نم %72.8 في ةيفاك يرغ ةموملأا ةزاجإ ةدم ْتَبرُتعا دقو .يدثلا نم عاضرلإا تاسرامم تاثحابلا دمتعاو .ًارهش 10.9 ةبوغرلما ةدلما طسوتم ناك مانيب ًارهش 4.7 يدثلا نم عاضرلإا ةدم طسوتم غلبو .لمعلا عون بسح ةيسفنو ةيندب لكاشم لىإ .لمعلا عقاوم في ةناضح روُد دوجو لىعو ،عاضرلإل ةحارلا تارـتف لىع لوصلحا ةيناكمإ لىعو ،ةموملأا ةزاجإ لوط لىع يدثلا نم عاضرلإا ءاسنلا ينب يدثلا نم عاضرلإا ةسرامم زيزعتو ةموملأا ةزاجإ ةدم ةلاطإ لجأ نم ةلجاع تلاخدتب مايقلل ةّحلم ةروضر كانه نأ لىإ ةساردلا تصلخو .تلاماعلا Maternity leave and experience of working mothers in Lebanon ABSTRACT We conducted a cross-sectional study of 802 Lebanese mothers to evaluate the effect of rapid return to work on their health and that of their child. Breastfeeding practices were also assessed. The duration of maternity leave was insufficient for 72.8% of the women. Rapid return to work could cause physical and psychological problems depending on the type of work. The average length of breastfeeding was 4.7 months and while the average desired length was 10.9 months. Breastfeeding depended on the duration of the maternity leave, the possibility of breaks for breastfeeding and the presence of nurseries at work. Urgent interventions are necessary to prolong maternity leave and promote breastfeeding among working women. طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 995 Introduction Depuis quelques décennies, les femmes  travaillent  et  veulent  de plus  en plus  concilier  leur  maternité  et  leur  vie  sociale,  voire  professionnelle.  Après  l’accouchement,  cette  conciliation  devient  plus  difficile  si  elles  doivent  reprendre  leur vie professionnelle alors  que  le  bébé  a  encore  besoin  de  leur  présence pendant de  longues heures.  Par  rapport  à  la mère qui  travaille,  les  changements physiques et psychiques  qui ont lieu en post-partum peuvent se  compliquer et entraîner des  situations  graves, en particulier si elle est fatiguée ;  sa santé mentale en est affectée. Il a ainsi  été démontré que le retour rapide au tra- vail augmente  le  risque de symptômes  dépressifs [1,2].  Les  mères  qui  travaillent  sont  ainsi  amenées  à  confier  la  garde  de  leur  enfant  assez  tôt,  selon  la  durée  du  congé  maternité.  Souvent,  les  bébés qui  arrivent  en  crèche ou chez  l’assistante maternelle n’ont pas atteint  la maturité  requise  pour  affronter  la  séparation précoce de  la mère  [3]. En  fait, la diffusion des connaissances sur le  développement psychologique du petit  enfant a révélé que la période d’éveil et  de découverte qui se situe de la naissance  à l’âge de trois ans est capitale [4] ; ainsi,  la  carence maternelle  sévère, précoce  (avant deux ans) et prolongée, pourrait  être psychopathogène ou génératrice  d’inaffectivité [3].  Des  études  ont même  démontré  qu ’un   congé   matern i té   rédu i t  augmenterait  le  risque  de mortalité  périnatale, néonatale, post-néonatale et  infantile [1,5], ce qui pourrait s’expliquer  en partie par le fait que le retour rapide au  travail contribue à diminuer ou arrêter  l’allaitement  de  façon  notable [6].  Ceci aurait des conséquences néfastes  pour  le  bébé,  puisque  l’allaitement  protège contre  les  infections aiguës,  a  des effets bénéfiques sur la santé à long  terme et assure une meilleure  relation  psychologique avec  la mère  ainsi que  des économies au niveau des  soins de  santé  [7,8].  L’allaitement maternel  a  aussi des effets bénéfiques sur la santé de  la mère, en réduisant le risque ultérieur  de diabète de type 2, de cancer du sein et  de l’ovaire. L’arrêt précoce ou l’absence  de  l’allaitement  est  aussi  associé  avec  un risque accru de dépression en post- partum [8]. D’autre part, du fait de l’insuffisance  des  garanties  qui  lui  sont  offertes  en  matière d’emploi dans quelques pays,  la mère  qui  travaille  s’expose  aussi  à  plusieurs risques économiques, tels que  la perte de son emploi et la suspension de  ses gains [9]. Les tendances mondiales  actuelles tendent à répondre aux besoins  des  familles  ayant des  enfants  en bas  âge ;  un des plus  importants  besoins  dans ce domaine est le congé maternité  [10].  L’Organisation  internationale  du  travail  (OIT) a adopté en 1919  le  premier  instrument mondial destiné à  protéger  les  travailleuses avant et après  la naissance d’un enfant : la convention  sur  la protection de  la maternité. Cette  norme  a  été  révisée  en  1952  et  en  2000, et prévoit un congé d’une durée  minimum de quatorze semaines. Pour  les pays qui accordent des prestations  en espèces dans  le cadre de  la  sécurité  sociale, la convention recommande que  la  rémunération ne  soit pas  inférieure  à  deux  tiers  du  revenu  antérieur  et  que  les  prestations médicales  soient  intégralement assurées [9,11]. De plus,  cette convention interdit le travail de nuit,  les heures supplémentaires et les tâches  préjudiciables à la santé de la mère et de  l’enfant. Ces mesures visent à  limiter  la  fatigue, à réduire la tension physique et le  stress ainsi qu’à épargner aux femmes les  tâches dangereuses et insalubres [9,11].  En outre,  selon  la nouvelle convention  de  l’OIT (Article 10),  les  travailleuses  liées à leur employeur par un contrat de  travail ont le droit de prendre des pauses  d’allaitement  quotidiennes  pendant  leurs heures de travail [11]. Ces pauses  d’allaitement  visent  à  permettre  à  la  mère d’allaiter son enfant et/ou de tirer  son  lait  pendant  les heures de  travail  [12]. La  travailleuse dont  la durée de  travail  effective  est  d’au moins  sept  heures et demie par jour a droit à deux  pauses d’une demi-heure par  jour. Les  travailleuses qui  effectuent  au moins  quatre  heures  de  travail  par  jour  ont  droit  à  une  pause  d’une  demi-heure  [12].  Les pays qui garantissent un congé  de maternité généreux sont nombreux :  par exemple,  le Danemark,  la Norvège  et  la Suède  accordent des  congés  ré- munérés de longue durée dont une par- tie est réservée à la mère, l’autre pouvant  être prise par  l’un ou l’autre parent [5].  Ainsi,  les gouvernements de plusieurs  pays envisagent d’augmenter  la durée  du congé maternité et d’introduire des  congés  supplémentaires,  comme  le  congé de paternité, le congé parental, le  congé familial, etc. [5,9]. Au  Liban,  en  2005,  une  étude  réalisée par l’Administration centrale de  la Statistique (ACS), sous le patronage  de  l’UNICEF,  a montré que 20,57 %  des  femmes  libanaises  sont employées  dans  les différents  secteurs de  travail  au Liban, et que 43,74 % cherchent du  travail  pour pallier  au  revenu  familial  insuffisant  (92,44 %)  [13]. On  peut  donc s’attendre au Liban, dans les années  à venir,  à des pourcentages progressifs  de  femmes  sur  le marché  du  travail.  Cependant,  la  loi  libanaise prévoit un  congé maternité de sept semaines dans  le  secteur privé et de 60  jours dans  le  secteur public, englobant la période qui  précède  et  qui  suit  l’accouchement ;  l’allaitement  n’y  est  pas  mentionné  [14]. Dans un article précédent, nous  avons montré que  le  congé maternité  et  les droits de  la mère qui  travaille ne  sont  pas  respectés  au Liban,  surtout  dans le secteur privé : 60 % des femmes  prennent un congé maternité de durée  inférieure à celle préconisée par  la  loi,  20 % des  femmes  exercent un  travail  de nuit que ce  soit durant  la grossesse  ou après  le  retour du congé maternité,  et  certaines  sont même  licenciées de  leur travail dès l’annonce de la grossesse  (2 %) [15].  EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 996 Méthodes Type et objectif de l’étude Il s’agit d’une étude transversale auprès  des mères  libanaises,  visant  à  évaluer  les perceptions des femmes concernant  l’effet du retour au travail sur leur santé  et  celle de  leur  enfant  au moment de  leur dernier accouchement (expérience  passée).  Population de l’étude La population cible est constituée par les  femmes mariées ayant vécu une fois ou  plus la période de grossesse et la période  qui suit l’accouchement, au moment de  l’exercice d’un travail professionnel dans  une institution donnée. Méthodes La  liste des hôpitaux a été obtenue au- près du ministère de la Santé : la division  des hôpitaux a  été  faite  selon  les  cinq  mohafazat.  Les  hôpitaux  participant  à  l’étude ont  été  choisis par  tirage  au  sort systématique à partir de la liste des  hôpitaux,  la décision ayant été prise de  prendre le 1/6 des hôpitaux de chaque  mohafazah ,   ce  qui  correspond  à  5  hôpitaux de Beyrouth administratif,  8  hôpitaux du Mont-Liban,  4 hôpitaux  de la Bekaa, 4 hôpitaux du Nord et 4 du  Sud-Liban.  L’enquête  s’est déroulée de début  juillet à fin octobre 2007. Après contact  des  responsables  présents  dans  ces  hôpitaux, des  rendez-vous ont été pris  pour  les  rencontrer  et  leur  expliquer  les objectifs de l’étude, ses modalités et  son déroulement. Les questionnaires  ont été distribués dans  les  services de  maternité et de pédiatrie dans chacun  des hôpitaux choisis. Les  femmes sont  présentes dans  ces  services  au moins  pour un jour complet et peuvent remplir  le  questionnaire  calmement.  Dans  chaque service, une personne désignée  par  le  ou  la  responsable  de  l’hôpital  (sage-femme,  infirmière ou secrétaire)  a pris  en charge  le questionnaire. Les  directives  données  à  ces  personnes  étaient  que  les  femmes  devaient  le  remplir  seules,  après  consentement  oral.  Puisqu’il  s’agissait  d’une  étude  descriptive  observationnelle,  aucune  autorisation  éthique  n’a  été  jugée  nécessaire.    Cent   questionnaires   ont   été  distribués dans chacun des hôpitaux de  Beyrouth et du Mont-Liban,  sauf pour  la région du Chouf et  les districts de  la  Bekaa, du Nord et du Sud. Le nombre  final de questionnaires distribués a été  de 1350.  Outil utilisé Les données utilisées proviennent d’un  questionnaire  standardisé,  rédigé  en  arabe, formé de quarante-cinq questions  à  éventail  :  fermées,  ouvertes,  semi- ouvertes et pré-codées semi-ouvertes.  Les  variables dépendantes  sont  la  prise du congé maternité,  la durée du  congé maternité, les causes d’insuffisance  de  la  durée  du  congé  maternité,  la  récupération  de  l’état  de  santé  de  la  mère,  la  cause  d’insatisfaction  de  la  garde du bébé, l’allaitement du bébé, la  durée et  le  rythme de  l’allaitement,  la  durée  souhaitée  de  l’allaitement,  les  changements   psychosomatiques  du bébé et  les  accidents du bébé. Les  variables  indépendantes  sont  l’âge,  l’adresse, le nombre d’enfants, le niveau  d’études,  le  lieu de  travail,  la  région de  travail, le nombre d’années de travail, la  catégorie professionnelle et l’horaire de  travail. Le bien-être physique a été défini  comme « la sensation d’être totalement  rétablie après l’accouchement du point  de  vue physique »,  alors que  le bien- être mental  a  été  défini  comme « la  sensation  d’être  totalement  rétablie  après  l’accouchement du point de vue  psychique ». Calcul de la taille de l’échantillon Puisque nous n’avons pas de connais- sance  préalable  concernant  le  congé  maternel  chez  les  femmes  libanaises,  nous  avons  supposé  que  la majorité  était insatisfaite de la durée de ce congé.  Ainsi,  nous  avons  estimé  à  75 %  le  pourcentage d’insatisfaction ; si la valeur  acceptée  la plus éloignée serait à ± 3 %  de ce pourcentage, un minimum de 800  femmes était nécessaire pour avoir un  niveau de confiance à 95 %.  Analyse statistique Le logiciel de statistique de cette étude  SPSS 12.0 a été utilisé pour la saisie des  données et les analyses statistiques. Pour  étudier  les  relations  entre différentes  variables, on a appliqué le test de χ2 avec  une  erreur  de  premier  degré  de  5 %.  De plus, une analyse multivariée a été  effectuée, en appliquant une régression  logistique  de  la  variable  dépendante  prenant  les  variables  indépendantes  comme tierces variables : relation entre  la durée du congé maternité pris  et  la  fatigue,  la  récupération  du  bien-être  physique,  la  récupération du bien-être  mental,  le  lieu  de  travail,  le  niveau  d’études,  le nombre d’enfants,  l’horaire  de  travail,  l’allaitement maternel  et  le  sentiment  d’être  maltraitée  par  son  employeur. Résultats Sur  1350  questionnaires  mis  à  la  disposition des personnes responsables  de  l’étude  dans  les  hôpitaux,  huit  cent  deux  femmes  (52,4 %)  ont  été  interrogées.  La  durée  du  congé  maternité pris par  les  femmes de notre  échantillon est  jugée  insuffisante par  la  plupart d’entre elles  (72,8 %) ; 21,7 %  la  trouvent moyennement  suffisante,  contre  3,2 %  seulement  qui  disent  qu’elle est  suffisante. Dans  la Figure 1,  nous rapportons les difficultés majeures  rencontrées par  les  femmes dues  à  la  courte durée du congé maternité :  ce  sont pour  66,6 %  la  séparation  avec  le  bébé,  57,9 %  trouvent  la  période  d’allaitement  insuffisante,  42,2 %  et  43,1 % de  ces  femmes  parce  qu’elles  n’avaient pas encore récupéré leur bien- être physique et mental  ; ceci explique  que 43,4 % soient encore  fatiguées au  retour au  travail. Dans 41 % des cas,  la  طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 997 maman n’a pas  eu un  temps  suffisant  pour assurer une bonne prise en charge  d’elle-même et du bébé, et dans 37,7 %  des  cas,  le  bébé  a  encore  des  réveils  fréquents pendant la nuit. En réponse à la question concernant  le  temps  qu’ont  mis  les  mères  à  récupérer  leur  bien-être  physique,  66,9 %  des  femmes  disent  qu’elles  l’ont  récupéré de  trois à  six mois après  l’accouchement  et  29,3 %  après un  à  deux mois.  Pour  le  bien-être mental,  50 % des  femmes  l’ont  récupéré huit  mois après l’accouchement. La  garde  du  bébé,  en  l’absence  de  sa mère,  est  surtout  assurée par  sa  grand-mère maternelle  (42,9 %) ;  suit  en deuxième  la grand-mère paternelle  (27,6 %).  La  garderie  constitue  le  troisième choix,  avec un pourcentage  de  25,1 %.  À  noter  que  1,2 %  des  femmes  de  notre  échantillon  ont  dû  avoir recours à leur voisine pour garder  le bébé pendant leurs heures de travail.  Pour  la  satisfaction  des  mères  par  rapport  à  la garde de  leurs bébés, 346  (43,1 %) sont totalement satisfaites, 274  (34,2 %) sont moyennement satisfaites  et 161 (20,1 %) sont  insatisfaites. Pour  celles qui étaient  insatisfaites (20,1 %),  les causes sont diversifiées (Figure 2) : le  manque d’information en cas d’accident  ou de problème arrivant au bébé est  le  plus  fréquent (43,4 %), puis  le manque  d’information  sur  le  comportement  du  bébé  en  général  (32,6 %).  De  plus,  le  fait  que  cette personne n’agit  pas  avec  le bébé comme  la maman  le  veut  est  de  27,8 %,  surtout  du  point  de vue alimentaire (24,8 %). Parmi  les  maladies des bébés qui  sont en garde  en dehors de  la maison,  les maladies  du système respiratoire (28 %) sont les  plus  fréquentes, avec celles du système  digestif (17 %). Pour  la décision d’allaiter  le bébé,  on  a  trouvé  que  85,4 %  des  femmes  de  cette  étude  ont  décidé  d’allaiter  leur bébé. Celles qui ne  l’ont pas  fait  (14,5 %) avaient  surtout un problème  de  santé  (39,6 %)  ou  un  horaire  de  travail inadéquat (36,2 %). Au retour au  travail après  le congé maternité, 50,4 %  des femmes ont continué l’allaitement :  à  raison de deux  fois pendant  le  jour  (70,8 %)  et deux  fois pendant  la nuit  (68,6 %), ceci pour une durée de deux  et  quatre  mois  (71,6 %).  De  plus,  46,3 % des mères n’ont pas  continué  l’allaitement. La cause  la plus citée est  l’horaire de travail qui est incompatible  (63,4 %) ;  44 % des  femmes  écrivent  que l’allaitement est une charge difficile  à  concilier  avec  le  travail.  Si  elles  ne  travaillaient  pas,  54,3 %  des  femmes  auraient  continué  à  allaiter de  sept  à  douze mois,  alors que 31 %  l’auraient  fait pour trois à six mois. Figure 1 Causes de l’insuffisance du congé maternité Anémie Charge d’autres enfants Problème de garde du bébé Réveils pendant la nuit Pas de temps pour une bonne prise en charge pour elle et le bébé Bien-être mental non récupéré Bien-être physique non récupéré Fatiguée Période d’allaitement insuffisante Difficulté de séparation du bébé 0 10 20 30 40 50 60 70 14.5 18.8 23.6 37.7 41.0 42.2 43.1 43.4 57.9 66.6 Figure 2 Causes de l’insatisfaction de la mère à l’égard de la garde des bébés Ne sait pas agir avec les bébés Maladie par contagion ou par négligence Manque de propreté Manque d’information sur le comportement de la personne Insatisfaite du point de vue alimentaire N’agit pas comme la maman veut Maladie due au fait de sortir le bébé par mauvais temps Manque d’information sur le comportement du bébé Ne raconte pas en cas d’accident 0 5 10 15 20 25 30 35 40 45 50 7.5 17.0 17.9 22.5 24.8 27.8 28.0 32.6 43.4 % % EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 998 Dans  le  tableau 1,  l’augmentation  des pleurs (35,7 %) et  le  rythme veille- sommeil  perturbé  (35,4 %)  sont  les  deux changements psychosomatiques  des bébés pendant  la période de garde  les plus cités dans notre étude. Suit en  deuxième lieu la diminution du sourire  (22,3 %)  et  le  refus de  l’alimentation  (19,1 %), avec l’état général du bébé qui  est perturbé (20 %). De plus, 25,8 % des  mères déclarent que  leurs enfants ont  eu divers  accidents durant  la période  de garde, que ce soit à la garderie ou à la  maison. Les blessures et les brûlures sont  les accidents  les plus  fréquents (56 %).  Les  chutes  aussi  sont  à  considérer  (51,2 %).  Dans  le  tableau 2, on retrouve une  forte relation entre la durée insuffisante  du  congé maternité  et  les  causes  de  non-satisfaction de cette durée décrites  par  les mères  (p < 0,001) :  par ordre  d’importance, nous avons la séparation  avec  le bébé,  l’allaitement,  la  fatigue,  le  bien-être physique,  le bien-être mental,  la bonne prise en charge et l’anémie. Les  mères qui ont  considéré que  la durée  du  congé  maternité  est  totalement  suffisante ont récupéré leur état de santé  physique entre un et deux mois (25 %)  et leur état de santé mental à 10 mois ou  plus (10,5 %) (p < 0,001). Celles qui ont  considéré  la durée du congé maternité  moyennement  suffisante ont  récupéré  leur  état de  santé physique vers deux  mois (50 %) et leur état de santé mental  et social entre six et sept mois (50,7 %)  (p < 0,001). Celles qui ont  considéré  la durée de  ce  congé  insuffisante ont  récupéré  leur  état  de  santé  physique  entre sept et neuf mois (78,7 %) et leur  état de santé mental et social entre huit  et neuf mois (78,9 %) (p < 0,001). Dans  le  tableau 3,  la  relation entre  l’âge  effectif  de  l’allaitement  est  non  significative pour  le niveau d’études, ni  pour  le  lieu de  travail, ni pour  l’horaire  de  travail  par  jour.  La  relation  avec  les  catégories  professionnelles  est  significative : ce sont surtout les ouvrières  qui  allaitent  de  deux  à  quatre  mois  (96,2 %), le personnel paramédical pour  la période de cinq à huit mois (22,6 %), et  ce sont les cadres et les enseignantes qui  ont le pourcentage le plus élevé pour la  période de neuf à douze mois (15,5 %).  La majorité des  femmes qui ont allaité  entre deux et quatre mois auraient aimé  continuer jusqu’à treize à dix-huit mois  (86,2 %) (p < 0,001). La majorité des  femmes qui ont allaité entre cinq à huit  mois  auraient  aimé continuer  jusqu’à  sept à douze mois (24,1 %) (p < 0,001).  Celles qui ont allaité entre neuf et douze  mois  auraient  aimé continuer  jusqu’à  plus  de  dix-huit mois  (20,8 %)  (p  <  0,001). La  comparaison des  femmes  primipares avec  les  femmes multipares  n’a pas donné de  résultats  significatifs  pour les facteurs étudiés (p > 0,05 pour  toutes les comparaisons).   Dans  le  tableau 4, nous rapportons  les  résultats  de  l’analyse multivariée.  La  récupération du bien-être physique  dépend de la durée du congé maternité  seulement, alors que celle du bien-être  mental dépend de  la durée du  congé  maternité,  du  niveau  d’études  et  du  travail dans les cliniques et les magasins.  L’insuffisance de l’allaitement maternel  dépend  uniquement  de  la  durée  du  congé  maternité.  Le  sentiment  de  séparation du bébé dépend du congé  maternité et du niveau d’études. Discussion Le congé maternité est insuffisant pour  72,8 %  des  femmes  de  cette  étude,  et  21,7 %  le  trouvent moyennement  suffisant.   De  l ’avis   de  97 %  des  femmes  de  l’étude,  il  y  a  nécessité  urgente  d’augmenter  la  durée  du  congé  maternité.  Une  étude  faite  à  Minneapolis a montré que  la durée de  11,1  semaines  est  considérée  comme  courte  par  les  femmes  de  l’étude,  la  période de 8 mois est  idéale pour elles  [16]. Plusieurs  études  rapportent des  causes  de  l’insuffisance  de  la  durée  du congé maternité  similaires  à  celles  Tableau 1 Changements psychosomatiques et accidents du bébé notifiés durant la période de garde Changements Nombre (n = 802) % Accidents durant la garde du bébé 207 25,8 Si oui, quels accidents : Chute 106 51,2 Blessure-brûlure 116 56,0 Fausse route 72 34,7 Augmentation des pleurs 286 35,7 Rythme veille-sommeil changé 284 35,4 Diminution du sourire 179 22,3 Bébé perturbé 160 20,0 Refus de l’alimentation 153 19,1 Changement de réaction aux câlins de la maman 130 16,2 Fièvre sans cause 39 4,9 Pas de réponse 27 3,4 طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 999 rapportées  dans  notre  étude  [10,16- 20] : fatigue (43,4 %), non-récupération  du bien-être physique (43,1 %),  et du  bien-être mental  (42,2 %),  et manque  de bonne prise en charge d’elle-même  et du bébé (41 %).  Les  raisons  citées par  les  femmes  de notre étude ont été démontrées par  d’autres  études  indépendantes. Une  étude  au Minnesota  a démontré que  la durée du congé maternité a un effet  complexe sur la santé maternelle à sept  mois de post-partum et que les femmes  qui ont pris  six mois de congé et plus  ont un meilleur état de  santé mentale  à neuf et douze mois de post-partum.  Les femmes signalent qu’elles dorment  mieux,  elles  sont  plus  sociables,  ont  moins de difficultés à assurer  la charge  de  leur  enfant  et  atteignent  de hauts  niveaux de satisfaction dans leur travail  [21].  Ceci  est  comparable  avec  une  étude  faite  à Minneapolis montrant  que  l’effet des problèmes potentiels du  post-partum sur  l’état de bien-être des  femmes qui  travaillent  est  en  relation  avec  la durée de  leur congé maternité :  Gjerdingen  et  al.   rapportent  que  recouvrir  un  état  de  santé  équilibré  après  l’accouchement  demande plus  que les six semaines de congé données  tradit ionnel lement   aux   femmes  aux  États-Unis.  Les  complications  de  l’allaitement,  les  hémorroïdes,  les  symptômes  respiratoires  et  la  fatigue  persistent  jusqu’au moins  trois mois  après  l’accouchement. La constipation  et  l ’ inconfort  vaginal  continuent  jusqu’à neuf mois de post-partum. Les  changements de la santé mentale durent  au moins vingt-quatre  semaines  [22].  De même,  la  femme qui prend plus de  douze  semaines  de  congé maternité  ressent  un  effet  positif  sur  sa  vitalité  physique, alors que celle qui prend plus  de  quinze  semaines  le  ressent  sur  sa  vitalité mentale ; pour celle qui prend  plus de  vingt  semaines,  il  y  a un  effet  positif sur son rôle [22]. En France et en  Italie, une étude comparative des deux  pays  a  trouvé  que  80 %  des  femmes  retournent  au  travail  un  an  après  la  naissance  sans  aucun  problème  de  santé ni physique ni mental  [23]. Une  autre étude faite à ce sujet au Wisconsin  montre  que  le  long  congé maternité  contribue en grande partie à  l’équilibre  personnel et marital [17]. Au  Liban,  la  garde  des  bébés  est  surtout  assurée par  les  grands-mères,  ce qui permet aux femmes de continuer  plus  facilement à  travailler : 42,9 % des  bébés de cette étude sont gardés par  la  grand-mère maternelle  et  27,6 % par  la  grand-mère paternelle,  alors  qu’au  Japon et  aux États-Unis par  exemple,  les grands- parents  sont une  ressource  secondaire  après  les   centres  de  protection  [10]. Ceci  est un avantage  pour  les bébés  libanais, puisqu’il  vaut  mieux    garder  le bébé  en  famille que  l’envoyer à la garderie [24]. Le fait qu’il  y  ait des bébés gardés avec  la  voisine,  même en  faible pourcentage (1,2 %),  constituerait  cependant un problème.  Ainsi  25,1 % des  bébés  sont mis  à  la  garderie. Une  étude  aux États-Unis  a  montré  que  26,8 %  des  enfants  sont  soumis  à  la protection non parentale,  avec  31,8 %  en  Suède  mis  dans  les  maisons  de  protection  familiale  et  centres publics [10].  L’insatisfaction à l’égard de la garde  du bébé concerne surtout  la garderie :  Tableau 2 Relation entre la durée du congé de maternité pris par la mère, les causes de l’insuffisance de ce congé et la récupération de l’état de santé physique et mental Causes Opinion de la mère sur la durée du congé maternité p Totalement suffisante Moyennement suffisante Insuffisante Nombre (%) Nombre (%) Nombre (%) Fatigue 1 (3,8) 14 (8,0) 314 (53,8) < 0,001 Anémie 2 (7,7) 6 (3,4) 102 (17,5) < 0,001 Bien-être physique 1 (0,3) 16 (4,9) 310 (53,1) < 0,001 Bien-être mental 1 (3,8) 19 (10,9) 300 (51,4) < 0,001 Bonne prise en charge 1 (3,8) 16 (9,2) 294 (50,3) < 0,001 Allaitement 1 (3,8) 31 (17,8) 385 (65,9) < 0,001 Séparation du bébé 2 (0,4) 41 (8,4) 445 (76,2) < 0,001 Santé physique récupérée (mois) 1-2 2 (25,0) 4 (50,0) 2 (25,0) < 0,001 3-6 10 (3,1) 87 (26,7) 229 (70,2) 7-9 12 (2,8) 81 (18,6) 343 (78,7) > 9 2 (14,3) 2 (14,3) 10 (71,4) Santé mentale récupérée (mois) 6-7 4 (5,6) 36 (50,7) 31 (43,7) < 0,001 8-9 2 (2,2) 17 (19,1) 70 (78,9) ≥ 10 2 (10,5) 5 (26,3) 12 (63,2) EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 1000 manque  de  pédagogie,  manque  de  propreté,   alimentation  inadaptée  et  surtout,  contagion  des  maladies  et  trajet  à  faire,  de bon matin, même  par mauvais  temps.  Les maladies  du  système  respiratoire  étaient  les  plus  fréquemment citées dans notre étude ;  elles  sont  les plus  fréquentes  chez  les  enfants de zéro à sept ans à cause de leur  vulnérabilité  aux  agents bactériens  et  viraux. Une accréditation des garderies  au Liban selon des standards de qualité  est donc recommandée.  Concernant la durée de l’allaitement,  nos  résultats  sont  comparables  à  ceux d’une étude descriptive qui a été  conduite  en Turquie  auprès  de  trois  cent une mères qui  travaillaient, dont  77 % allaitaient leur enfant quatre mois  et  plus.  Cependant,  les  conditions  de  l’allaitement  au  travail,  le  congé  maternité,  les  habitudes  de  vie  de  la  mère (tabac, etc.) et le tirage du lait sont  des facteurs indépendants influençant la  durée de l’allaitement maternel par ordre  décroissant [25]. Le  retour précoce au  travail a un effet plus négatif sur la durée  de l’allaitement que le nombre d’heures  de travail par  jour. Les  femmes qui ont  un haut niveau d’éducation débutent  plus  tard  le  substitut du  lait maternel.  Celles qui ont un emploi  indépendant  avec un horaire flexible et celles qui ont  un congé maternel plus  long, un  taux  élevé d’emploi à temps partiel et un taux  élevé d’allaitement sur  le  lieu du travail  allaitent leurs enfants pour une période  plus  longue  que  les  autres  groupes.  De plus,  la  reprise de  leur  rôle  social  dans  leur travail et  leur environnement  diminue l’allaitement maternel [25]. Le  travail   de  la  mère  n’a  pas  d’influence  sur  la décision  initiale de  l’allaitement,  car  celles  qui  n’ont  pas  allaité l’ont fait surtout pour un problème  de santé (39,6 %). Celles qui ont arrêté  d’allaiter après  le  retour au  travail  l’ont  fait  à  cause de  l’horaire de  travail  qui  Tableau 3 Relation entre la durée de l’allaitement et le niveau d’études, le lieu de travail, l’horaire de travail, les catégories professionnelles et l’âge souhaité d’allaitement Variable Durée de l’allaitement (mois) P 2-4 5-8 9-12 Nombre (%) Nombre (%) Nombre (%) Niveau d’études Primaire 8 (100 ) 0 0 0,153 Complémentaire 23 (85,2 ) 2 (7,4 ) 2 (7,4 ) Secondaire 81 (75,0 ) 17 (15.7 ) 10 (9,3 ) Universitaire 149 (66,2 ) 48 (21,3 ) 28 (12,4 ) Études supérieures 29 (78,4 ) 7 (18,9 ) 1 (2,7 ) Secteur de travail Santé 63 (69,2 ) 16 (17,6 ) 12 (13,2 ) 0,705 Enseignement 71 (74,0 ) 19 (19,8 ) 6 (6,3 ) Services 149 (71,3 ) 37 (17,7 ) 23 (11,0 ) Commerce 7 (77,8 ) 2 (22,2 ) 0 Horaires (nombre d’heures) 2-5 34 (72,3 ) 7 (14,9 ) 6 (12,8 ) 0,619 6-9 244 (70,9 ) 65 (18,9 ) 35 (10,2 ) 10-12 12 (85,7 ) 2 (14,3 ) 0 Catégorie professionnelle Ouvrière 25 (96,2 ) 0 1 (3,8 ) 0,035 Employée 128 (74,9 ) 31 (18,1 ) 12 (7,0 ) Cadre 37 (63,8 ) 12 (20,7 ) 9 (15,5 ) Enseignante 63 (64,9 ) 19 (19,6 ) 15 (15,5 ) Personnel paramédical 37 (69,8 ) 12 (22,6 ) 4 (7,5 ) Aurait voulu continuer l’allaitement (mois) 3-6 83 (84,7 ) 15 (15,3 ) 0 < 0,001 7-12 142 (63,4 ) 54 (24,1 ) 25 (12,5 ) 13-18 25 (86,2 ) 2 (6,9 ) 6 (6,9 ) > 18 40 (75,5 ) 2 (3,8 ) 11 (20,8 ) طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 1001 Tableau 4 Analyse multivariée Variable dépendante Variables indépendantes P ORa (IC à 95 %) Bien-être physique non récupéré au retour au travail Durée du congé de maternité < 0,001 0.96 (0,95-0,98) Autres variables Non significatif Bien-être mental non récupéré au retour au travail Durée du congé de maternité < 0,001 0,98 (0,96-0,99) Niveau d’études < 0,001 1,40 (1,13-1,73) Travail dans une clinique 0,012 10,75 (1,70-68,11) Travail dans un magasin 0,004 13,32 (2,26-78,35) Autres variables Non significatif Allaitement insuffisant Durée du congé de maternité < 0,001 0,97 (0,96-0,99) Autres variables Non significatif Séparation du bébé Durée du congé de maternité < 0,001 0,97 (0,96-0,99) Niveau d’études 0,076 1,20 (0,98-1,45) Autres variables Non significatif est  incompatible  avec  l’allaitement  (ouvrière,  employée  et  personnel  paramédical). Ceci  est  comparable  à  une  étude  faite  en Turquie  [26]  où  le  travail n’a pas  influencé  la décision  de  l’allaitement,  mais  la  durée  de  l’allaitement  était  courte  à  cause  du  retour  au  travail  après huit  semaines.  Ainsi, pour faciliter l’allaitement après le  retour au travail, nous recommandons  des changements d’horaires de  travail,  des  garderies  sur  le  lieu de  travail  et  des pauses allaitement [27],  similaires  à  celles  qui  existent  dans  les  pays  développés :  en Suède,  par  exemple,  le  secteur  d’emploi  des  femmes  est  caractérisé  par  la  flexibilité  et  les  dispositions du  temps partiel  (46 %)  qui  facilitent  l’éducation des  enfants.  Au  Japon,  la  loi exige que  les  sociétés  permettent  aux  employées  ayant des  enfants de moins d’un  an de  choisir  des  heures  de  travail  courtes,  des  Staehelin K, Bertea PC, Stutz EZ. Length of maternity leave and 1. health of mother and child--a review. International Journal of Public Health, 2007, 52(4):202–209. Chatterji, P. Markowtiz, S. Does the length of maternity leave 2. affect maternal health? Southern Economic Journal, 2005, 72:16–41. Ferrari P 3. et al. Actualités en psychiatrie de l’enfant et de l’adoles- cent. Paris, Flammarion Médecine-Sciences, 2001:104–151. Références Bee H, Boyd D. 4. Psychologie du développement - Les âges de la vie, 2e édition. Bruxelles, de Boeck, Collection Ouvertures psy- chologiques, 2003:186–189. Ruhm C. Parental leave and child health. 5. Journal of Health Eco- nomics, 2000, 19:931–960. Chuang CH et al. Maternal return to work and breastfeeding: a 6. population-based cohort study. International Journal of Nurs- ing Studies, 2010, 47(4):461–474. heures flexibles ou de choisir  le congé  de  la protection de  l’enfant  [10]. Les  dispositions relatives à la maternité dans  la  loi du  travail du Liban doivent être  révisées. Les municipalités pourraient  prendre  la  relève  de  l’État  dans  ce  domaine,  comme elles ont  réussi  à  le  faire dans d’autres pays, tels que la Suède  [10].  Concernant  la  validité  de  notre  étude, plusieurs points sont à noter : un  biais de sélection est possible, puisque  l’échantillon n’est pas de type aléatoire.  De  plus,  un  biais  d’information  est  possible puisque  le bien-être physique  et  le  bien-être mental  n’ont  pas  été  évalués par des  instruments de mesure  appropriés  mais  par  des  questions  d’ordre subjectif. Il n’y a pas non plus de  distinction entre les notions d’allaitement  maternel  et  d’allaitement  maternel  exclusif ;  les  résultats pourraient donc  indiquer une  représentation excessive  du pourcentage de nourrissons  ayant  bénéficié du  lait maternel. Cependant,  nous n’avons pas de  raison de  croire  que  l’utilisation  d’autres  méthodes  pour mener cette étude donnerait des  résultats essentiellement différents.   Conclusion Le  congé  maternité  au  Liban  est  insuffisant de  l’avis de  la majorité des  femmes interrogées. Le retour rapide de  la mère au travail provoquerait plusieurs  problèmes physiques et mentaux, selon  le  secteur de  travail. Le  temps moyen  de  l’allaitement  dans  la  population  étudiée est de 4,7 mois et  la moyenne  de l’allaitement souhaité va jusqu’à 10,9  mois. L’allaitement dépend de la durée  du congé maternité, de  l’existence des  pauses d’allaitement et de  la présence  de garderies sur le lieu de travail. EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 1002 Lawrence RA. Breastfeeding: benefits, risks and alternatives.7. Current Opinions in Obstetrics & Gynecology, 2000, 12(6):519– 524. Ip S et al. A summary of the Agency for Healthcare Research 8. and Quality’s evidence report on breastfeeding in developed countries. Breastfeeding Medicine, 2009, 4(Suppl. 1):S17–30. Le congé de maternité est rémunéré dans plus de 120 pays. 9. L’inégalité de traitement des hommes et des femmes dans l’emploi perdure. Genève, Organisation internationale du travail, 1998 (communiqué de presse BIT/ 98/7, 16 février 1998). http:// www.ilo.org/global/About_the_ILO/Media_and_public_ information/Press_releases/lang--fr/WCMS_071342/index. htm Allen S. Working parents’ young children: cross national 10. comparisons, policies, programs in three countries. International Journal of Social Welfare, 2003, 12(4):261–273. Convention11. sur la protection de la maternité, 2000. Genève, Organisation internationale du travail, 2000 (http://www.ilo. org/ilolex/cgi-lex/convdf.pl?C183, consulté le 29 avril 2010). Raibant G. 12. La Charte des droits fondamentaux de l’Union européenne. Paris, Éditions du Seuil, 2001:186–189. Administration centrale de la Statistique et Fonds des Nations 13. Unies pour l’enfance. La situation des enfants au Liban en l’an 2000. Beyrouth, Administration centrale de la Statistique, 2001. Tremblay DG, De Sève M. 14. La conciliation emploi- famille et le temps de travail : analyse de cas dans les secteurs de l’éducation, de la santé et des services sociaux. Rapport de recherche. Montréal, Télé-Université, Université du Québec, 2002:10–25. Saadé N, Salameh P, Barbour B. Congé maternel et mères 15. travailleuses au Liban : circonstances professionnelles. Sciences sociales et santé, 2010, 28(2):115–128. Gjerdingen Dk, McGovern PM. Women’s postpartum ma-16. ternity benefits and work experience. Family Medicne, 1995, 45(8):580–592. Hyde JS, Essex MJ Maternity leave, women’s employment, and 17. marital incompatibility. Journal of Family Psychology, 2001, 15(3):476–491. Scarr Sandra. Working mothers and their families. 18. American Psychologist, 1989, 44(11):1402–1409. Tassel V. L’allaitement maternel sur le temps de travail. 19. Soins Pédiatrie/Puériculture, 2003, 213:28–29. Yilma G. Factors influencing breastfeeding for working mothers. 20. Turkish Journal of Pediatrics, 2002, 44(1):30–34. Gjerdingen D, Chaloner K. The relationship of women’s 21. postpartum mental health to employment, childbirth and social support. Journal of Family Practice, 1994, 38(5):465–472. McGovern P et al. Time off work and the post-partum health of 22. employed women. Medical Care, 2002, 35(5):507–521. Romito P, Saurel-Cubizolles MJ, Escriba-Aguir V. Maternity 23. rights, work and health in France and Italy. Journal of the American Medical Women’s Association, 2002, 57(1): 47–48. Lebovici S, Soulé M, Diatkine R. L’expression manifeste des 24. troubles et leur compréhension. In : Lebovivi s, Soulé M, Diaktine R, eds. Nouveau Traité de psychiatrie de l’enfant et de l’adolescent. Paris, Presses Universitaires de France, Collection Quadrige, 1995:1535–1771. Eydal GB. Politiques de la petite enfance dans les pays 25. nordiques. Lien Social et Politiques, Revue internationale d’Action communautaire, 2003, 50:165–184. Robert J, Oberdoff H. 26. Libertés fondamentales et droits de l’homme. Paris, Montchrestien, 1989:164–177. Johnston ML, Esposito N. Barriers and facilitators for breast-27. feeding among working women in the United States. Jour- nal of Obstetric, Gynecologic and Neonatal Nursing, 2007, 36(1):9–20. طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 1003 Obesity in Saudi children: a dangerous reality S.S. Al-Dossary,1 P.E. Sarkis,1 A. Hassan,2 M. Ezz El Regal3 and A.E. Fouda3 ABSTRACT Obesity among children is an increasing concern. This cross-sectional study in 2006 determined the prevalence and demographic characteristics of overweight and obesity in children in the Eastern province of Saudi Arabia. A total of 7056 children (aged 2–18 years) were selected from schools and the outpatient department of a hospital. The overall prevalence of overweight was 19.0% and of obesity was 23.3%. More than 50% of children between 14 and 18 years had weight above the 85th percentile. More males than females were obese by ages 14–18 years (35.6% versus 19.2%). Saudi and non-Saudi nationalities had the same distribution of body mass index. Interventions to encourage healthier lifestyles for children are needed at the national level. 1Department of Paediatrics, Saad Specialist Hospital, Al-Khobar, Saudi Arabia. 2Department of Bone Marrow Transplant/Immunology, Great Ormond Street Hospital, London, United Kingdom. 3Department of Paediatrics, Faculty of Medicine, University of Mansoura, Mansoura, Egypt (Correspondence to A.E. Fouda: Ashraf_foda@mans. edu.eg). Received: 01/02/09; accepted: 24/03/09 ةيرطخ ةقيقح :ينيدوعسلا لافطلأا في ةنم ِّسلا هدوف فشرأ ،لاجرلا زع دممح ،نسح لمأ ،سيكسر ويرب ،يسرودلا رحس صئاصلخاو راشتنلاا لدعم ددتح 2006 ماع تيرجأ يتلاو اهددصب نحن يتلا ةيضرعلا ةساردلا هذهو .ًاديازتم ًاقلق لافطلأا في ةنم ِّسلا يرثت :ةصلالخا )ةنس 18و 2 ينب مهرماعأ حوارـتت( ًلافط 7056 يرتخا دقو .ةيدوعسلا ةيبرعلا ةكلملما في ةيقشرلا ةقطنلما لافطأ في ةنم ِّسلاو نزولا طرفل ةيفارغوميدلا نم رثكأ في نزولا ناكو .%23.3 ةنادبلاو %19.0 نزولا طرف راشتنا لدعم لياجمإ غلب دقو .تايفشتسلما ىدحإ في ةيجرالخا ةدايعلا نمو سرادلما نم ثانلإا ينب اهنم رثكأ روكذلا ينب ةنْمِسلا تناكو .يننماثلاو ةسمالخا ةيوئلما ةيحشرلا نم لىعأ ةنس 18و 14 ينب مهرماعأ حوارـتت نيذلا لافطلأا نم %50 ةجالحا لىع ةساردلا ُّلدتو .ينيدوعسلا يرغو ينيدوعسلا ينب ًايواستم مسلجا ةلتك بَسْنَم ناكو .)%19.2 لباقم %35.6( ةنس 18-14 ةيرمعلا ةرـتفلا في .ينطولا ديعصلا لىع لافطلأا في ةيحصلا ةايلحا طمانأ عيجشتل تلاخدت لىإ Obésité chez les enfants saoudiens : une réalité dangereuse RÉSUMÉ L’obésité chez les enfants suscite de plus en plus d’inquiétudes. Cette étude transversale réalisée en 2006 visait à déterminer la prévalence et les caractéristiques démographiques du surpoids et de l’obésité chez les enfants de la province orientale d’Arabie saoudite. Au total, 7056 enfants (âgés de 2 à 18 ans) ont été sélectionnés dans des établissements scolaires ou au service de consultation externe d’un hôpital. La prévalence globale du surpoids était de 19 % et celle de l’obésité de 23,3 %. Plus de 50 % des enfants âgés de 14 à 18 ans avaient un poids supérieur au 85e percentile. Entre 14 et 18 ans, le nombre de garçons obèses était supérieur à celui des filles (35,6 % contre 19,2 %). La répartition de l’indice de masse corporelle était identique pour les Saoudiens et les non Saoudiens. Il est nécessaire de mettre en place au niveau national des interventions visant à encourager des modes de vie plus sains. EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 1004 Introduction Towards  the end of  the 20th century,  obesity was  identified as  a worldwide  health care problem affecting  the well- being of populations. Previously identi- fied only as a problem of adult health,  obesity among children  is  increasingly  becoming a concern [1]. The Gulf  re- gion is not exempt. Surveys in a number  of different areas and provinces have re- ported a high prevalence of overweight  and obesity  in Saudi children  in all age  groups [2,3]. Government and local au- thorities have implemented educational  programmes to help weight reduction or  prevention of obesity. Abnormal weight  in children is still considered by experts  to be caused by an imbalance between  diet  and habit,  although  a  hormonal  etiology  is a diagnosis  that needs  to be  excluded in children. In order  to  add valid  information  about  the weight  status of  children  in  our  region,  this  study was  conducted  in the Eastern province of Saudi Arabia  to determine  the prevalence of over- weight and obesity in children from the  Eastern province of Saudi Arabia and to  compare  these prevalences with  those  of non-Saudi  children  living  in Saudi  Arabia. Methods This was  a  cross-sectional  study con- ducted in Al-Khobar city, in the Eastern  province of Saudi Arabia. Data collec- tion started on January 2006 and con- tinued for 6 months. Sample Our  sample  included  7056  children  aged 2–18 years of Saudi or non-Saudi  nationality enrolled  from a school and  from  the outpatient department of  a  hospital.  Saad specialist hospital  is a  tertiary  centre  in  the Eastern province.  It  is  a  private  institution with  a  capacity  of  600 beds,  admitting patients  from all  socioeconomic  classes. All  paediatric  consultations done during the period of  study in the outpatients department (n = 9249) were collected  from the elec- tronic database. After excluding follow- up consultations and  incomplete files,  a total of 6237 files were evaluated and  after  excluding  those  aged  <  2  years  (n  =  441  cases),  5796  children were  enrolled. Saad schools  for boys and girls are  private  institutions  in  the Al-Khobar  region, Dammam,  in the Eastern prov- ince. All 1260 students aged 6–17 years  enrolled  in  the  schools were  included  in  the  study, which  is  10% of  the  to- tal number of  students  in  the private  educational  sector and 3% of  the  total  (public and private) students of the Al- Khobar–Dammam region [4].  Prior  to enrolment a consent  form  was signed by the child’s parent after an  explanation of the aims and methods of  the study.  Data collection The  internal  validity of  the  study was  ensured by  the data collectors  in both  Saad  schools  and Saad  specialist hos- pital. These were nurses who  received  training prior  to data  collection with  special  emphasis on  standardizing  the  methods of measurement. Body weight  and height of children were measured  using a digital scanner. The instruments  used were calibrated daily. A data collection form was designed  to gather data on: age, sex and national- ity as well as measures of body weight  and height.  Body mass  index  (BMI)  was calculated for each child according  to  the  formula  adopted  internation- ally: BMI = weight  (kg)/height  (m)2 [5]. We used  the Centers  for Disease  Control and Prevention (CDC) 2000  growth charts  [6]. The children were  classified into 3 weight categories: nor- mal weight (BMI < 85th percentile  for  age and sex), overweight (BMI between  85th–95th  percentiles)  and  obese  (BMI > 95th percentile)  [7,8]. Then  percentile weight categories among the  sample were studied by nationality, sex  and age groups (2–4, 5–9, 10–13 and  14–18 years). To ensure better generalization of  the  results,  the data  from patients pre- senting  to  the outpatients department  of the hospital were compared with the  data  from school  students  to confirm  that  the  distribution  by  age,  sex  and  nationality was  similar before pooling  the data. Statistical analysis Data analysis was done using SPSS soft- ware,  version 12,  and  the chi-squared  test, Student  t-test and analysis of vari- ance (ANOVA). Statistical significance  was set at P < 0.05. Results The  sample  included  7056  children,  mean age 8.7 (standard deviation 4.9)  years. Males were 55.7% of  the  study  children and females were 44.3%. Most  of the children were of Saudi nationality  (79.4%) while 20.6% were various non- Saudi nationalities.  Classification of  the enrolled chil- dren  according  to BMI  revealed  that  only 57.7% of them had normal weight  for age and sex, while 19.0% were over- weight and 23.3% obese (none of    the  children were underweight). The rates of overweight and obesity  increased progressively with  age. The  overall  rate of obesity was significantly  higher  than  the  rate of overweight  in  all age groups (P < 0.05). A progressive  rise  in  the  rates  of  obesity  and over- weight were  found with age  from ages  5–9 years  to 14–18 years, peaking at  age 10–13 years (20.2% and 28.0% for  obesity  and overweight  respectively)  (Table 2). At  age 2–4 years  the pro- portions of children who were normal  weight,  overweight  and  obese  were  62.7%, 18.1% and 19.2%  respectively.  By  age  14–18  years  the  proportions  were 53.0%, 20.0% and 27.0%. طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 1005 Regarding  distribution  of weight  categories  by  sex,  our  study  showed  that  the overall  rate of obesity among  males was significantly higher than that  among  females  (26.4% versus  19.3%  respectively) (P < 0.05). We also found  that  the  percentage  of  obese  males  was  significantly  higher  than  that  of  overweight ones (26.4% versus 18.0%  respectively). The percentage of over- weight  females was significantly higher  than  the percentage of obese  females  (20.3% versus 19.3% respectively). Comparison of the 3 weight catego- ries by nationality  showed no  statisti- cally significant difference in the rates of  normal weight, overweight and obesity  between Saudi and non-Saudi children.  Studying  the distribution of weight  categories  among  the children by age  groups and sex revealed that more males  than females were overweight in the age  group 2–4 years (19.6% versus 16.3%  respectively). This was  reversed  in  the  age group 14–18 years, by which age  the proportion of overweight  females  was higher than overweight males (23%  versus 16.7% respectively) (Figure 1). The percentage of obese males was  significantly  higher  than  that  among  females  in  all  age  groups  (P  <  0.05).  Obesity  in males  showed a  steady  in- crease with age  from 20% at ages 2–4  years up to 35.6% at 14–18 years, while  in females the rate of obesity rose from  18.1%  ages  2–4  and  peaked  at  ages  10–13 years at 22.7% before declining  at  ages 14–18 years  to 19.2% (Figure  2). Thus  by  ages  14–18  years  35.6%  of males were obese  versus 19.2% of  females (P < 0.05). Discussion Obesity  is  one of  the main  concerns  for health care worldwide. In 2001, the  World Health Organization (WHO)  announced  that  10%  of  the  world’s  children were obese and  that  the  rate  was rising in developing countries: 155  million children at school age were over- weight, while 22 million under 5 years  were overweight [9]. According to 2002  WHO statistics, there has been a broad  shift in disease burden with the majority  of deaths worldwide now being related  to noncommunicable diseases, many  of which can be  linked  to  imbalances  of nutrition, diet  and physical  activity  [10]. In Saudi Arabia, many studies have  been done  to evaluate  the magnitude  of overweight and obesity among Saudi  children (Table 3) [2,3,11–21].  It was  found  that overweight and obesity oc- curs in all provinces [21]. A recent study  found that the Eastern province has the  highest rates and the Southern province  the lowest rates [3]. According  to our study about 50%  of Saudi children  in  this Eastern prov- ince sample have a BMI above the 85%  percentile.  It  was  evident  that  Saudi  children started developing overweight  when  they  are 5–9  years of  age—by  which age 21% of children were over- weight  and  21%  obese—and  their  weight continued  to  increase  into  the  adolescent years. This could be attrib- uted to the fact that children start going  to school at  that age, and hence,  there  is  a  less  control on  their  eating habits  and nutrition at  this  stage. Moreover,  children  in our  country have become  less active;  few or none walk to school,  spending more time in sedentary enter- tainment activities, such as viewing TV,  computer and video games. On average,  a child  in Saudi Arabia spends 6 hours  per day in front of screens [22]. Recent  studies have  found  that a 2%  increase  in  the prevalence of obesity has been  documented for each extra hour per day  Table 1 Distribution of body mass index (BMI) categories by sex, age and nationality Categories Normal weight (BMI < 85%) Overweight (BMI 85%–95%) Obese (BMI > 95%) Total P-value No. % No. % No. % No. % in category Sex Male 2185 55.6 708 18.0 1040 26.4* 3933 55.7 *P < 0.05 Female 1886 60.4 633 20.3 604 19.3 3123 44.3 Age (years) 2–4 1164 62.7 337 18.1 356 19.2 1857 26.3 *P < 0.05 5–9 1346 60.6 406 18.3 469 21.1 2221 31.5 10–13 709 51.8 276 20.2 384 28.0* 1369 19.4 14–18 852 53.0 322 20.0 435 27.0* 1609 22.8 Nationality Saudi 3222 57.5 1048 18.7 1329 23.7 5599 79.4 NS Non-Saudi 849 58.3 293 20.1 315 21.6 1457 20.6 Total 4071 57.7 1341 19.0 1644 23.3 7056 100.0 NS = not significant. EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 1006 of screen viewing by those aged 12–17  years [22]. Our  results agree with a previously  published study in the country that obes- ity in both sexes is low among preschool  children (31%) and highest among ado- lescents (50%–76%) [2,20,21].  In our  sample,  the peak of obesity was at age  10–13 years (28.0%) and stayed at the  same high  rate until  age 14–18 years.  Studies have shown that 80% of obese  adolescents become obese adults  [8].  Adolescence has been described as the  “critical period for the development for  adult obesity” [23]. Hence, intervention  before this age or is vital for both future  health and  the ability  to  sustain  long- term weight control [22]. In our study, males showed a higher  prevalence of obesity  in all age groups,  with  a  sharp  rise  at  age 14–18 years.  Females in this age group may be more  self-conscious about  their weight  and  avoid progressing into the obesity range.  This is in accordance with a recent study  conducted in Saudi Arabia which found  that obesity was more common among  Males Females 25 20 15 10 5 2-4 5-9 10-13 14-18 19.6 16.3 16.4 16.7 19.8 21.0 20.7 23.0 Figure 1 Distribution of overweight by age group and sex Figure 2 Distribution of obesity by age group and sex Males Females 40 35 30 25 20 15 10 5 18.1 20 22.9 35.6 31.6 18.6 22.7 19.2 2-4 5-9 10-13 14-18 Age group (years) % Age group (years) % طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 1007 adolescent boys  than was overweight.  This was attributed to the lack of physi- cal activity [2,21] and to their ability to  drive  in our country which gives males  easy access  to unhealthy diets (e.g.  fast  foods  that  contain 40%–45%  fat  and  soda drinks  instead of water) and  less  time to eat at home where meals would  be more nutritional  [9,20]. With more  mothers employed outside  the home,  traditional  foods  are  replaced by  fast  foods and  typically one-third of meals  are eaten outside of the home environ- ment, often at fast-food restaurants [22].  Studies have  shown  that  the  average  consumption of  sugar-sweetened bev- erages and  fruit  juices  increased  from  13 g/week  in 1950  to 446 g/week  in  1992/3 [24]. As recently as 1989–1991  to 1994–1995,  the consumption  rose  by 65% [22]. There was no significant difference  in the prevalence of obesity among Sau- di and non-Saudi children in our study,  which draws attention towards the role  of the environment, lifestyle and lack of  physical activity as contributing  factors  over genetic  factors  in  influencing  the  pattern of obesity [21]. A study of Saudi  children  in the Eastern province  found  that they were not engaged in sporting  activities  as much  as  their American  counterparts  [21,25]. Daily participa- tion in physical education in highschool  in the United States dropped from 42%  in 1991  to 21%  in 1999. Furthermore,  60%–70% of  urban  inhabitants were  living sedentary lives [22]. It  is  difficult  to  reduce  excessive  weight  once  it  becomes  established.  Prevention of obesity in children should  therefore  start  from birth  by  putting  more emphasis on exclusive breastfeed- ing  for  the first  6 months of  life.  It  is  becoming  a priority  to  establish pre- school,  school  and adolescent health  programmes,  with  the  emphasis  on  increasing  physical  education  hours  and consumption of healthy  food, by  incorporating health messages  into the  school curricula [26]. Ta bl e 2 Re vi ew o f s tu di es fr om S au di A ra bi a ev al ua ti ng th e pr ev al en ce o f o ve rw ei gh t a nd o be si ty a m on g ch ild re n Re fe re nc e Re gi on Ta rg et c hi ld re n N o. A ge s (y ea rs ) Se x O ve rw ei gh t pr ev al en ce (% ) O be si ty pr ev al en ce (% ) Pr es en t s tu dy Ea st er n pr ov in ce Sc ho ol ch ild re n, P riv at e ho sp ita l 70 56 2– 18 M ,F 19 23 .3 A la m , 2 0 0 8 [1 1] W es t R iy ad h El em en ta ry sc ho ol 10 72 8– 12 F n/ d 14 .9 A m in , 2 0 0 8 [1 2] A l H as sa Pr im ar y sc ho ol s 11 39 10 –1 4 M 14 .2 9. 7 A l-H az za a, 2 0 0 7 [1 3] Ri ya dh Pr im ar y sc ho ol s 10 82 (1 98 8) 70 2 (2 0 0 5) 6– 14 M 3. 4 (1 98 8) 24 .5 (2 0 0 5) A l T ur ki , 2 0 0 7 [1 4] Ri ya dh Pr im ar y ca re c lin ic s 26 7 12 –2 0 n/ d 18 .7 21 .0 M ah fo uz , 2 0 0 7 [1 5] A bh a ci ty Sc ho ol s 26 96 11 –1 9 M 11 5. 0 Fa rg ha l, 20 0 7 [1 6] A bh a ci ty Sc ho ol s 76 7 7– 20 M ,F 11 15 .9 A l-A lm ai e, 2 0 0 5 [1 7] A l-K ho ba r In te rm ed ia te a nd a ll 3 gr ad es o f se co nd ar y sc ho ol 17 66 14 –1 9 M ,F 19 .3 (M ) 11 .8 (F ) 17 .2 (M ) 10 .2 (F ) A l-R uk ba n, 2 0 0 3 [2 ] Ri ya dh In te rm ed ia te a nd se co nd ar y sc ho ol s 89 4 12 –2 0 M 13 .8 20 .5 A l-S ae ed , 2 0 0 3 [1 8] D iff er en t p ro vi nc es Pr im ar y an d pr ep ar at or y sc ho ol s 22 39 6– 17 F 20 11 .0 El -H az m i & W ar sy , 2 0 0 2 [3 ] D iff er en t p ro vi nc es H ou se ho ld sc re en in g pr og ra m m e 12 7 0 1 1– 18 M ,F 23 .4 12 .7 El -H az m i, 20 0 2 [1 9] D iff er en t p ro vi nc es N at io na l e pi de m io lo gi ca l h ou se ho ld su rv ey 12 0 71 1– 18 M ,F 10 .7 (M ) 12 .7 (F ) 6. 0 (M ) 6. 7 (F ) A ba hu ss ai n, 19 99 [2 0 ] A l-K ho ba r c ity A do le sc en t g irl s 67 6 12 –1 9 F 28 (o ve rw ei gh t o r o be se ) A l-N ua im , 1 99 6 [2 1] D iff er en t p ro vi nc es Sc ho ol ch ild re n 90 61 6– 18 M 11 .7 15 .8 M = m al e; F = fe m al e; n /d = n ot d et er m in ed . EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 1008 Conclusions In our  sample 50% of  children  in  the  Eastern province of Saudi Arabia were  overweight  or  obese  (BMI  >  85th  percentile). Our  study  showed  that  obesity started early in life (ages 10–14  years) and continued  throughout  the  Reilly JJ. Descriptive epidemiology and health consequences 1. of childhood obesity. Best Practice and Research Clinical En- docrinology and Metabolism, 2005, 19:327–341. Al-Rukban MO. Obesity among Saudi male adolescents in 2. Riyadh, Saudi Arabia. Saudi Medical Journal, 2003, 24:27–33. El-Hazmi MA, Warsy AS. A comparative study of prevalence 3. of overweight and obesity in children in different provinces of Saudi Arabia. Journal of Tropical Pediatrics, 2002, 48:172–177. Statistics. Ministry of Education, Saudi Arabia [website] (http://4. www.moe.gov.sa/openshare/englishcon/Statistics/Statistics- for-the-year-1426H_1427H.htm_cvt.html, accessed 5 July 2010). Pinhas-Hamiel O et al. Type 2 diabetes in childhood: not 5. just for grown-ups anymore. Contemporary Pediatrics, 2001, 18:102–125. Kuczmarski RJ et al. 2000 CDC growth charts for the United 6. States: methods and development. Vital Health Statistics, 2002, 11:1–190. Dehghan M, Akhtar-Danesh N, Merchant AT. Childhood obes-7. ity, prevalence and prevention. Nutrition Journal, 2005, 4:24. Daniels SR et al. Overweight in children and adolescents: 8. pathophysiology, consequences, prevention, and treatment. Circulation, 2005, 19, 111(15):1999–2012. WHO global strategy on diet, physical activity and health. 9. Resolution WHA55.23. Geneva, World Health Organization, 2004. Barlow SE, Dietz WH. Obesity evaluation and treatment: Expert 10. Committee recommendations. The Maternal and Child Health Bureau, Health Resources and Services Administration and the Department of Health and Human Services. Pediatrics, 1998, 102:E29. Alam AA. Obesity among female school children in North West 11. Riyadh in relation to affluent lifestyle. Saudi Medical Journal, 2008, 29:1139–1144. Amin TT, Al-Sultan AI, Ali A. Overweight and obesity and 12. their relation to dietary habits and socio-demographic char- acteristics among male primary school children in Al-Hassa, Kingdom of Saudi Arabia. European Journal of Nutrition, 2008, 47:310–318. Al-Hazzaa HM. Prevalence and trends in obesity among school 13. boys in Central Saudi Arabia between 1988 and 2005. Saudi Medical Journal, 2007, 28:1569–1574. References Al Turki YA. Overweight and obesity among attendees of 14. primary care clinics in a university hospital. Annals of Saudi Medicine, 2007, 27:459–460. Mahfouz AA et al. Obesity and related behaviors among ado-15. lescent school boys in Abha city, southwestern Saudi Arabia. Journal of Tropical Pediatrics, 2008, 54:120–124. Farghaly NF et al. Lifestyle and nutrition and their impact on 16. health of Saudi school students in Abha, Southwestern region of Saudi Arabia. Saudi Medical Journal, 2007, 28:415–421. Al-Almaie SM. Prevalence of obesity and overweight among 17. Saudi adolescents in Eastern Saudi Arabia. Saudi Medical Jour- nal, 2005, 26:607–611. Al-Saeed WY et al. Prevalence and socioeconomic risk factors 18. of obesity among urban female students in Al-Khobar city, East- ern Saudi Arabia, 2003, Obesity Reviews, 2007, 8:93–99. El-Hazmi MA, Warsy AS. A comparative study of prevalence 19. of overweight and obesity in children in different provinces of Saudi Arabia. Journal of Tropical Pediatrics, 2002, 48:172–177. Abahussain NA et al. Nutritional status of adolescent girls in the 20. eastern province of Saudi Arabia. Nutrition and Health, 1999, 13:171–177. Al-Nuaim AR, Bamgboye EA, Al-Herbish A. The pattern of 21. growth and obesity in Saudi Arabian male school children. International Journal of Obesity and Related Metabolic Disor- ders, 1996, 20:1000–1005. Baker S et al. Overweight children and adolescents: a clinical 22. report of the North American Society for Pediatric Gastroen- terology, Hepatology and Nutrition. Journal of Pediatric Gas- troenterology and Nutrition, 2005, 40:533–543. Steinbeck K. Childhood obesity. Treatment options. Best Prac-23. tice and Research Clinical Endocrinology and Metabolism, 2005, 19:455–469. Rennie KL, Johnson L, Jebb SA. Behavioural determinants of 24. obesity. Best Practice and Research Clinical Endocrinology and Metabolism, 2005, 19:343–358. Liu G. In the neighborhood: connecting plant biomass and 25. childhood obesity. Contemporary Pediatrics, 2005, (May):14. Al-Nozha MM et al. Obesity in Saudi Arabia. Saudi Medical 26. Journal, 2005, 26:824–829. adolescent  ages. The  prevalence  of  obesity was higher among males  than  females,  who  had  a  tendency  to  be  overweight  rather  than  obese.  Envi- ronmental factors may have more of an  influence on  the prevalence of obesity  than do genetic  factors.  Interventions  to  encourage  healthier  lifestyles  for  children  are  needed  at  the  national  level.  Acknowledgements The authors thank Hassan Amay (MBS  Paediatrics) for her valued contribution  to this article. طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 1009 Review Childhood obesity in the Middle East: a review P. Mirmiran,1 R. Sherafat-Kazemzadeh,1 S. Jalali-Farahani 1 and F. Azizi 1 ABSTRACT Accurate and comprehensive data on the extent of the problem of childhood obesity is lacking in countries of the Middle East. This review, based on a Medline search, summarizes the prevalence of obesity among children and adolescents in the region during 1990–2007. The highest rates of obesity and overweight were reported from Bahrain and the lowest from the Islamic Republic of Iran. Studies from Saudi Arabia, Islamic Republic of Iran and Kuwait showed an upwards trend in childhood obesity compared with a decade ago. Lack of uniformity in reference standards and reporting systems renders comparisons difficult. Nevertheless, the high prevalence of childhood obesity in the Middle East should stimulate policy-makers in the region to set up effective national and regional surveillance systems. 1Nutrition Unit, Obesity Research Centre, Research Institute for Endocrine Sciences, Shaheed Beheshti University of Medical Sciences, Tehran, Islamic Republic of Iran (Correspondence to F. Azizi: azizi@endocrine.ac.ir). Received: 24/01/08; accepted: 13/10/08 ةعجارم :طسولأا قشرلا لافطأ في ةنم ِّسلا يزيزع نوديرف ،نياهارف – ليلاج اراس ،هداز مظاك - تفا َرش ايور ،نايرميرم نيورب ثحبلا لىع ةزكترلما ،ةعجارلما هذه ص ِّخلتو .طسولأا قشرلا نادلب لافطأ في ةنم ِّسلا ةلكشم راشتنا لوح ةلماشو ةقيقد تايطعم رفاوتت لا :ةصلالخا تلادعم لىعأ تناكو .2007-1990 ماوعلأا للاخ ميلقلإا في ينقهارلماو لافطلأا ينب ةنم ِّسلا راشتنا ،Medline ةيبطلا تايشرنلا عاجرـتسا طخ في ةيبرعلا ةكلملما في تيرجأ يتلا تاساردلا ترهظأو .ةيملاسلإا ناريإ ةيروهجم في اه ُّلقأ ناك ينح في ،نيرحبلا ةلود نم اهنع غلبلما نزولا طرفو ةنم ِّسلا غيلبتلا مظنو ةيعجرلما يرياعلما صقن نأ ولو .ضيالما دْق َرعلاب ةنراقم لافطلأا في ةنم ِّسلا دايدزا لىإ ًلايم تيوكلاو ةيملاسلإا ناريإ ةيروهجمو ةيدوعسلا باحصأ زفيح نأ بيج طسولأا قشرلا في لافطلأا في ةنم ِّسلا راشتنا لّدعم عافترا نإف رمأ نم نكي ماهمو .تانراقلما ءارجإ بعصلا نم لعيج ةدحولما .ةلاعف ةيميلقإو ةينطو د ُّصرـت مظن دادعإ لىع ميلقلإا في رارقلا Revue sur l’obésité de l’enfant au Moyen-Orient RÉSUMÉ Les pays du Moyen-Orient manquent de données précises et exhaustives sur l’étendue du problème de l’obésité de l’enfant. Cette revue, qui repose sur des recherches effectuées dans Medline, offre une synthèse sur la prévalence de l’obésité chez les enfants et les adolescents dans cette région, entre 1990 et 2007. Les taux d’obésité et de surpoids les plus élevés ont été enregistrés à Bahreïn, et les plus faibles en République islamique d’Iran. Des études réalisées en Arabie saoudite, en République islamique d’Iran et au Koweït ont montré une tendance à la hausse de l’obésité infantile par rapport à la décennie précédente. Le manque d’uniformité dans les normes de référence et les systèmes de signalement rend les comparaisons difficiles. Néanmoins, la forte prévalence de l’obésité de l’enfant au Moyen-Orient doit inciter les responsables politiques de la région à mettre en place des systèmes de surveillance nationaux et régionaux efficaces. EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 1010 Introduction Obesity has  long been considered as a  predisposing factor that affects an indi- vidual’s health. Yet,  the  importance of  obesity and overweight among children  has  only  been  highlighted  relatively  recently [1–3]. Worldwide,  there are a  total of 155 million (1  in 10) children  overweight, and around 30–45 million  classified as obese [4–6]. Investigators now  suggest  that  in- dices of  adiposity are useful  screening  tools  for multiple  risk  factors  such as  hyperlipidaemia,  insulin  resistance and  high blood pressure  among  children  [1]. The  incidence of  obesity-related  conditions  is  increasing steeply among  adolescents;  these  include  type 2 dia- betes, a disease that was previously not  normally  seen  until middle  or  older  age  [5]. Childhood body mass  index  (BMI) has been shown to be associated  with adult adiposity [1] and  its  related  conditions [7]. The World Health Or- ganization  (WHO)  is  working  with  its Member States  to  implement  the  Global Strategy on Diet, Physical Activ- ity  and Health  to  combat  childhood  obesity [8]. WHO estimates that three- quarters of all deaths in the developing  world by  the year 2020 will be due  to  noncommunicable diseases [9].  In developed societies, several stud- ies  have  shown  increasing  numbers  of overweight  children  [1,10–12]. Al- though many nutritional  surveys were  conducted  in  the 1980s and 1990s  in  developing  countries,  they were  not  generally analysed for the prevalence of  overweight, focusing mainly on the un- derweight and malnutrition [13]. There  are no consistent data on overweight  available for more than 70% of children  aged  less  than 5 years  [13]. Nonethe- less,  reports  indicate  that worldwide  more  than 22 million  children under  5  years  old  are  obese  or  overweight,  and more  than 17 million of  them are  presumed  to  live  in developing coun- tries  [9,12]. A  study of 94 developing  countries showed that the nations with  the highest  prevalence of overweight  were located mainly in the Middle East,  North Africa and Latin America [13].  The Middle East region has the high- est dietary energy surplus among devel- oping countries,  and  there  is evidence  of a rapid rise in noncommunicable dis- ease risk factors, especially obesity [12].  If there is to be a regional commitment  to control the growing problem of child- hood obesity, access to reliable national  data  and  accurate  information  about  the rates and time trends are indispensa- ble. Still, there are few studies that have  reviewed  the extent of  this problem  in  this region, and those that have did not  take  account  of  the methodology  or  tabulate  the data comprehensively.  In  this review we explore studies about the  extent of overweight and obesity among  children and adolescents in countries of  the Middle East indexed in Medline, the  most widely used medical database. We  endeavour to provide a comprehensible  systematic  review of  the  studies  in  the  region with  the  emphasis  on  cut-off  values  and  definitions  of  overweight  and obesity. Literature review methods We  retrieved  surveys  concerning  children’s  anthropometric data  from  different countries. Due  to  the  limited  availability of national  surveys,  articles  with  local  datasets  (e.g.  limited  to  1  province) were also  included  to deter- mine  a  raw  estimate of  the  extent  of  childhood obesity in the region.  Definition of terms There are multiple reference values cur- rently used as  the  standards  to  assess  children’s obesity:  The WHO  reference  defines  BMI • percentiles as cut-offs  for overweight  among 10–19-year-olds and weight- for-height Z-scores for obesity in chil- dren < 10  years old. The  reference  values  include age-  and  sex-specific  data  from  the first National Health  and Nutrition Examination Survey  (NHANES  I)  collected  in  1971– 74  in  the United States of America  [14–16].  The Centers for Disease Control and • Prevention  (CDC)  reference  uses  BMI percentiles  for ages 2–20 years.  The  year  2000  reference  included  BMI-for-age-growth charts and was a  revised version of the 1977 National  Health Care Survey (NCHS) growth  charts [14,17].  The Childhood Obesity Working • Group of  the  International Obesity  Task Force  (IOTF)  reference uses  sex- and age-specific BMI cutoffs that  correspond  to  BMI  25  kg/m2  for  overweight and 30 kg/m2 for obesity  at age 18 years [14,18]. The reference  population was made up of 6 nation- ally representative datasets.  The NCHS reference data use  sub-• scapular  and  triceps  skinfold  thick- ness [19,20].  For this review we included all suit- able reports despite differences between  them  in  the definitions of  childhood  obesity. Method of selection An extensive  literature  review was per- formed using the Medline search engine.  We first  retrieved  all medical  subject  headings  (MeSH)  related  to  the  con- cepts of nutrition, obesity, overweight,  body weight, BMI, statistics, epidemiol- ogy,  children and adolescents,  as well  as  keywords  for  categorizing  articles  according  to  countries of  the Middle  East. Finally, we developed a set of ma- jor  topic headings  for  these concepts:  “Obesity/epidemiology”,  “Children”,  “Growth and development” and “Mid- dle East”. We employed  these MeSH words,  limiting  the  search  to  the  time  period 1990–2007.  Performing  the  search  in PubMed,  we obtained 70 results. These included  studies  pertaining  to  the  prevalence  of obesity among migrants  from Mid- dle East origin  in other countries;  the  طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 1011 prevalence of  related conditions  such  as diabetes or  cardiovascular diseases  and  their  genetic  and  environmental  risk  factors;  body perceptions;  social  and physical habits in a country; and the  nutritional  status of wasting and mal- nourished children, but not obese ones.  We excluded these studies. The results  included countries such as Turkey that  is  included  in MeSH  as  a Middle East  country but is not in the Eastern Medi- terranean Region  (EMR)  of WHO,  plus  there were African countries such  as Somalia and Sudan  that are part of  EMR but not listed by MeSH as a Mid- dle Eastern country. We excluded these  countries  for convenience of  searching  and drawing conclusions.  Surveys The  final  review  covered  48  articles  retrieved from Medline-indexed journals  that were directly  related  to  the preva- lence of overweight and obesity among  children and adolescents in countries of  the Middle East.  The studies  focused mainly on  the  prevalence of childhood obesity either  in a local or national setting. Multistage,  random sampling methods were used in  all the papers, and all were cross-section- al except for 2 studies, from Saudi Arabia  and Oman, which were cohort studies.  Because of  the  limited  availability  of  nationally representative anthropomet- ric data and the lack of uniformity in the  sampled populations  in  the  available  reports, we  selected only  reports with  larger  sample sizes or greater coverage  of different  age groups.  In  this  review  we have summarized and tabulated the  methodology and prevalence  rates  in  each paper  for comparison. Finally, we  compared  the  reviewed  studies with  reports from both other developing and  developed countries. Regional coverage of studies The  largest number of  studies on  the  epidemiology of obesity  among chil- dren and adolescents were  from Saudi  Arabia  and  the  Islamic  Republic  of  Iran (11 each),  followed by Kuwait (7  articles) and Bahrain (4 articles). Other  countries  included were: Jordan, Leba- non, Libyan Arab  Jamahiriya, Oman,  Pakistan, Qatar and United Arab Emir- ates. We were not able  to  retrieve any  articles originating  from Afghanistan,  Iraq or Syrian Arab Republic during this  period (Table 1).  Of  the  48  selected  articles  13  in- volved data  from a sample size greater  than  5000  individuals  (Table  1); the  largest  sample  sizes were  from the  Islamic  Republic  of  Iran  (21 111),  United Arab Emirates (16 391), Kuwait  (14 659) and Saudi Arabia  (12 701).  Studies  from Saudi Arabia and Oman  presented cohort  studies  for  the  trend  of obesity in different ages; others were  cross-sectional studies using multistage  stratified random sampling for data col- lection. The  studies  covered  different  age  groups and  in order  to compare data- sets  in a general way we  tabulated  the  information  covering  the widest  age  range  and  the  greatest  sample  size  in  each country in Table 2. Bahrain, Jordan,  Libyan Arab Jamahiriya, Oman and Qa- tar did not have reports  from every age  group, while  there were  single  reports  from the United Arab Emirates, Saudi  Arabia  and  Lebanon  covering  large  samples  from all  groups. As  illustrated  in Table 3, different  surveys employed  different standards for defining the cut- off point  for overweight/obesity. After  analysing  the  studies  shown on Table  2, we  then summarized  the prevalence  rates  from surveys covering  the  largest  sample size  in each country  in Table 3.  Different rates in both sexes were shown  if the data were available in the article.  Prevalence of obesity and overweight in the Middle East The highest prevalence of overweight  was  reported  from  Bahrain  (38.5%)  [41],  followed by Kuwait, with  a  rate  of 31.8% among girls  [27]. The  lowest  prevalence of obesity was reported from  the  Islamic Republic of  Iran  (2% –3%  by different cut-offs) [22],  followed by  Lebanon (3.2% among 3–19 year-old  girls)  [31]. The  study  from  Bahrain  included only 584 girls and since  there  were larger reports from this country, we  did not show these results in the tables. Among  the Middle East  countries  our review demonstrates that  there are  few nationally representative studies on  Table 1 Studies providing information for this review about the prevalence of obesity among children and adolescents in Middle East countries Country No. of articles retrieved Assessments in > 1 province Studies with > 5000 cases Bahrain 4 + 0 Islamic Republic of Iran 11 + 2 Jordan 1 + 1 Kuwait 7 + 3 Lebanon 4 + 1 Libyan Arab Jamahiriya 1 + 0 Oman 1 – 0 Pakistan 2 – 0 Qatar 2 + 1 Saudi Arabia 11 + 4 United Arab Emirates 4 + 1 EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 1012 the pattern of obesity among children  that  can be used  for  future  reference.  A year 2000 review of the available na- tional datasets  showed  that  the  rates  of overweight exceeded 5%  in Kuwait  and Qatar,  suggesting  that  a popula- tion-wide  shift  in which  overweight  was  replacing wasting,  but  they  also  reported rates lower than 5% in Afghan- istan, Bahrain, Islamic Republic of Iran,  Oman, Pakistan and Yemen,  indicat- ing  that underweight was  still  a major  problem [13]. Our review supports the  notion of nutrition  transition  in  this  region; the highest levels of overweight  and obesity were  among adolescents  in Bahrain and Kuwait (over 30%). By  comparison, obesity was  reported  to  be 29% among Canadian, 26% among  Italian and 19.7% among Dutch ado- lescents [20].  Table 2 Summary of the studies conducted in different countries on the childhood obesity in different age groups Country Ref. no. Year of publication Sample size Region Age groupa Preschool Childhood Adolescence Bahrain [21] 2003 506 Different provinces – + + Islamic Republic of Iran [22] 2007 21 111 Different provinces – + + Islamic Republic of Iran [23] 2002 4 315 Sistan, Guilan + – – Islamic Republic of Iran [24] 2004 2 321 Tehran – + + Islamic Republic of Iran [25] 2006 6 565 Tehran + + + Jordan [26] 2003 5 223 Different provinces – + + Kuwait [27] 2004 14 659 Different provinces – + + Kuwait [28] 1998 7 419 Different provinces + – – Kuwait [29] 2000 8 957 Different provinces – + – Lebanon [30] 2006 12 299 Different provinces – + + Lebanon [31] 2003 2 104 Different provinces + + + Libyan Arab Jamahiriya [32] 2002 1 614 Al Jabel Al Garby, Tripoli + – – Oman [33] 2004 550 Muscat – + + Pakistan [34] 2001 1 050 Karachi + + + Qatar [35] 2006 3 923 Different provinces – – + Qatar [36] 2004 8 231 Different provinces + – – Saudi Arabia [37] 2002 12 701 Different provinces + + + Saudi Arabia [38] 1996 9 061 Different provinces – + + United Arab Emirates [39] 2006 4 381 Different provinces + + + United Arab Emirates [40] 2005 16 391 Different provinces + + + aStudies did not use a uniform scheme for selecting the age groups under study, current categories are used as a basis for comparison, however, numbers indicate the exact the study sample age in each category. طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 1013 Ta bl e 3 Re po rt ed p re va le nc e of o ve rw ei gh t a nd o be si ty a m on g ch ild re n in d iff er en t M id dl e Ea st c ou nt ri es a nd s ta nd ar ds u se d to d efi ne c ut -o ffs C ou nt ry Re f. no . A ge ra ng e (y ea rs ) G ir ls Bo ys St an da rd N ot es O ve rw ei gh t (% ) O be si ty (% ) O ve rw ei gh t (% ) O be si ty (% ) Ba hr ai n [2 1] 12 –1 7 –a 35 .0 – 21 .0 W H O /N C H S C om pa re d W H O /N H A N ES v er su s I O TF cr ite ria . – 18 .0 – 15 .0 IO TF Is la m ic R ep ub lic o f I ra n [2 2] 6– 18 4. 6 2. 0 4. 3 2. 5 C D C D iff er en t r ef er en ce v al ue s w er e us ed a nd co m pa re d. F ig ur es h er e re fe r t o IO TF a nd C D C c rit er ia . 5. 9 1.3 5. 4 1.6 IO TF Fi gu re s h er e re fe r t o IO TF a nd C D C c rit er ia . 4. 9 2. 4 5. 2 2. 4 N at io na l Is la m ic R ep ub lic o f I ra n [2 5] 10 –1 9 12 .1, 15 .5 , 5 .7 5. 0 , 6 .1, 5 .7 1.1 , 1 2. 8, 19 .4 2. 2, 6 .0 , 9 .1 IO TF Fi gu re s h er e re fe r t o IO TF a nd C D C c rit er ia . 10 .1, 14 .0 , 1 3. 6 9. 0 , 7 .9 , 6 .9 5. 5, 15 .1, 10 .1 3. 3, 12 .8 , 1 1.6 C D C Is la m ic R ep ub lic o f I ra n [2 3] 2– 5 22 .3 10 .0 20 .4 9. 0 IO TF C om pa re d to IO TF a nd C D C c rit er ia . Jo rd an [2 6] 6. 5– 17 .5 – – – 3. 9, 10 .9 , 5 .7 C D C Ku w ai t [2 7] 10 –1 4 31 .8 13 .1 30 .0 14 .7 N C H S/ C D C Le ba no n [3 0 ] 10 –1 8 19 .0 4. 2 28 .8 10 .1 IO TF Le ba no n [3 1] 3– 19 16 .1 3. 2 22 .5 7.5 W H O /N C H S Li by an A ra b Ja m ah iri ya [3 2] < 5 – 7.5 – 3. 3 W H O /N C H S O m an [3 3] 15 –1 6 10 .7, 14 .7, 13 .3 4. 7, 4. 0 , 6 .0 3. 3, 10 .5 , 1 5. 5 1.0 , 4 .5 , 9 .3 IO TF C oh or t p re va le nc e. Pa ki st an [3 4] 2– 18 – 7.8 – 8. 8 N H A N ES Q at ar [3 6] < 5 – 5. 0 – 4. 7 W H O /N C H S D efi ne d as w ei gh t- fo r- ag e > 2 SD o f t he st an da rd . Sa ud i A ra bi a [3 7] 1– 18 12 .7 6. 7 10 .7 6. 0 IO TF U ni te d A ra b Em ira te s [4 0 ] 4– 1 8 20 .1 7.1 17 .1 7.7 IO TF a A d as h in di ca te s t ha t d at a w er e no t a ss es se d in th at ca te go ry . SD = st an da rd d ev ia tio n. W H O = W or ld H ea lth O rg an iz at io n. N C H S = N at io na l H ea lth C ar e Su rv ey . C D C = C en te rs fo r D is ea se C on tr ol ;. O TF = In te rn at io na l O be si ty T as k Fo rc e. N H AN ES I = 1s t N at io na l H ea lth a nd N ut rit io n Ex am in at io n Su rv ey . EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 1014 The prevalence of 24% in the Euro- pean Union in 2002 was already higher  than the predicted peak for 2010 based  on original  speculations [3]. Some re- ports place Asia  in  the  third position  for overweight among all nations with  a prevalence of 2.9% and an absolute  number of 10.6 million overweight chil- dren [13]. Reported rates of overweight  in Asia vary greatly,  from over 30%  in  Taipei, Taiwan [2]  to 1.7%  in  rural ar- eas  of China  [42]. While  there  have  been  several nationally  representative  nutrition  surveys  conducted  in Asian  countries,  the  focus was on  the  lower  end of the distribution and, thus, rates of  overweight were rarely reported.  Internationally,  obesity  in  adoles- cence  has  increased  greatly  over  the  past 30–40 years,  especially  since  the  mid-1990s [3]. The rapid pace of socio- economic development  in  the Middle  East  has  led  to  rapid modernization,  accompanied  by  decreased  levels  of  physical  activity  and  increased  calo- rie  consumption. Although very  little  information  is  currently  available  on  obesity trends in childhood, the current  review seems to confirm a rising preva- lence over time. Studies in Saudi Arabia,  Islamic Republic of  Iran,  and Kuwait  compared their dataset with similar data  from a decade previously, and reported  a  rising prevalence of obesity  among  children  and  adolescents  [24,28,43].  A  study  comparing 2  cross-sectional  samples of children  in Tehran,  Islamic  Republic of Iran, reported an increasing  trend over a 3-year interval [25]. Recent  data  on  adult  obesity  are  alarming and  it could be assumed that  fat  families will  raise  fat  children who  will  further contribute  to  the epidemic  of  obesity-related  diseases. This  no- tion  is  supported by findings  from the  Islamic Republic  of  Iran  and Untied  Arab Emirates  [44,45],  showing  that  obese children generally  live  in obese  families. Dietary  choices  seem  to  be  responsible  for  familial  clustering  of  obesity [44,46,47].  Prevalence of obesity and overweight by age Studies  in  the Middle East providing  a  comparison  between  different  age  groups indicated a steady increase in the  prevalence of overweight/obesity with  increasing growth and stage of puberty  [27,33,48]. The  age  group  at  highest  risk  for overweight/obesity  could not  be established from this review because  there were  a  limited number of  stud- ies  in different countries and a  lack of  uniformity  in  selecting  age  groups. A  comparison of  the  rates of overweight  among different age groups  in 4 Asian  and 5 European countries reported that  the  raw number of  cases was highest  among 10-year-olds but could not give  an accurate estimate of the age group at  highest risk for overweight/obesity [2].  Prevalence of obesity and overweight by sex The level of  fatness among adolescent  girls was found to be higher than boys in  the NHANES I study in the USA [20].  Our  review, however,  found a higher  prevalence  of  either  overweight  or  obesity reported among boys than girls  from Saudi Arabia,  Islamic Republic  of  Iran, Lebanon, Kuwait and Pakistan  [23,25,27,31,34,49]. International com- parisons also report that the prevalence  of obesity  is greater  among boys  than  girls  in developing countries. Such dif- ferences may  reflect different  cultural  habits  and attitudes  towards nutrition  and physical activity [2].  Prevalence of obesity and overweight by socioeconomic status Studies from adults find correlations be- tween obesity and socioeconomic status  (SES), but  the  results are  inconsistent  for  children  [11].  In our  review,  stud- ies examining the associated social and  economic  factors  revealed  that  seden- tary  lifestyles,  less physical activity and  a  tendency  towards high-calorie diets  was a  factor contributing  to obesity, as  in  some countries of  the Middle East  childhood obesity was most dominant  in urban areas  and among  the higher  socioeconomic classes [26,38,50,51]. A  similar picture was seen in Egypt, where  the prevalence of obesity among high  SES adolescents was more than double  that  among  low  SES  groups  [52].  A  cross-national comparison study came  to  the  conclusion  that  higher  SES  children were more  likely  to be obese  in China  and Russia,  but  in  the USA  low-SES groups were  at  a higher  risk  of obesity [11]. A rising  trend  towards  obesity  among  lower SES groups has  been reported from the USA [3], where  there is rising concern about the urban  and  lower-income  classes, who  seem  to be particularly vulnerable because of  poor diet and limited opportunities for  physical activity. However, in a study in  Pakistan, children from middle-income  families  were more  overweight  than  their  counterparts  from  higher  and  lower  socioeconomic  classes  [34].  Therefore,  the  relationship of  SES  to  body  fat accumulation  is complex and  varies across different cultures.  Food  consumption  patterns  and  dietary  quality  are  highly  income- dependent,  but  dietary  choices,  par- ticularly  in higher-income groups,  are  also  driven  by  non-economic  forces  [16]. Among other  factors  influencing  the obesity epidemic, mothers’  level of  education and cultural beliefs regarding  body  image are  frequently  reported as  related factors [26,38,50,51, 53]. The re- markable variation in the prevalence of  obesity across populations suggests that  social,  economic  and  environmental  factors are important influences on the  epidemic, although  it may also be  true  that genetic differences across popula- tions also play a role. طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 1015 Standards for defining obesity and overweight A major problem  in  interpreting and  comparing studies  is  the use of differ- ent standard values for the definition of  overweight, as demonstrated  in Table  2. WHO is currently raising awareness  of  the need  to  establish  a  consistent  and pragmatic definition for childhood  obesity  to  address  the  issue globally  [54]. The definition of obesity among  children  is  not  as  clear  as  in  adults  where BMI (weight/height2) has been  accepted  internationally as a  standard  for  assessing  obesity  [2]. There  are  several  factors such as growth and pu- berty  that  influence  fat  accumulation  among  adolescents  and  these make  delineating a global standard for assess- ing overweight and obesity a matter for  debate [55].  Traditionally,  sex-  and  age-specific  BMI percentiles based on the NHANES-I  survey in the USA are used to define child- hood obesity. Nonetheless, the accuracy  of these data as a baseline for estimating  fatness  among other nations has been  disputable [4,11]. The evidence supports  the notion that lower BMI cut-off points  may need to be set for Asian populations,  due to their predisposition to deposit fat  around  the waist  [1,56].  Investigators  point  to other problems. For  example,  the NCHS cut-offs  are  skewed  toward  the higher  end,  reflecting a  substantial  rate of childhood obesity [13,19].  A  WHO  expert  committee  rec- ommended using both  the BMI 85th  percentile  and  the  triceps  skinfold  thickness  90th  percentile  to  define  adolescent obesity [19]. This has been  used infrequently due to the difficulty of  measuring  triceps skinfold  thickness  in  large population-based studies. Instead,  many have chosen to use the BMI 95th  percentile  to define child  and adoles- cent obesity [11,20]. An alternative are  the proposed values  set by  the  IOTF,  as described earlier. For validating  the  standards,  analysis of  the  relationship  of overweight (according to the defined  cut-offs)  to  levels of  adverse  risk  fac- tors  should be  conducted, which  is  a  time-consuming and difficult  task but  has been done  for previous  standards  [1,2,42,57]. Several  international  efforts  are  underway  to compare different  stand- ards with  each  other.  Some  indicate  that  these  references produce  similar  estimates of  the overall prevalence of  overweight [14]. In a comparison study  Wang et  al.  applied  IOTF and WHO  criteria in China, Russia and the United  States and came to the conclusion that  the  IOTF was better  for  international  use [58], although Reilly et al. believed  that  this standard showed a  lower sen- sitivity  for defining obesity  in English  children [59]. Al-Sendi et al. in Bahrain  concluded  that  IOTF cut-offs would  be more practical  to use  [21], but Al- Almaie  from Saudi Arabia  found  no  significant  difference  between  IOTF  and NHANES  criteria  [50].  Esmail- lzadeh et al. from the Islamic Republic of  Iran compared the prevalence of obesity  among Tehran children aged 3–5, 6–11  and 12–19 years old using IOTF refer- ence values and standardized percentile  BMI  curves of  Iranian  children  [25].  The values  for 2001–02  showed  that  the prevalence of obesity among  these  age groups was 5.0%, 17.6% and 10.4%  for girls and 2.2%, 19.5% and 22.6% for  boys  respectively. The corresponding  figures  for overweight  in  the  same age  groups were: 14.1%, 12.2% and 13.7%  for girls and 6.6%, 15.2% and 15.4% for  boys.  IOTF  and CDC  cut-offs were  also applied and the results covering the  largest  sample size  in each country are  presented  in Table 3  for  comparison.  The authors concluded  that  there was  a  lower  sensitivity  for  IOTF  cut-offs  among 6–19-year-old  children, while  the  sensitivity  did  not  differ  for  the  lower age group [25]. Dorosty et al. also  showed that obesity prevalence among  Iranian children  from 2 different prov- inces using  IOTF  reference data was  significantly  higher  in  2–3-year-olds  than when using  the  Iranian  reference  data [23]. Again,  there was a compari- son between CDC, IOTF and national  cut-offs by Kelishadi et al., who  found  an excellent correlation between IOTF  and CDC cut-offs with national cut-off  points,  yet  they  reported  that  IOTF  criteria overestimated  the prevalence  of  overweight  and  underestimated  the  prevalence  of  obesity  [22]. One  explanation  for  this difference may be  that IOTF cut-offs did not include any  data from this region. Therefore, current  cut-offs used for childhood overweight  might not be generalizable  to different  ethnic groups. Discussion There  is a  rapid  rate of  social develop- ment  and modernization  in Middle  East countries [22,60], which is causing  inequalities  in SES within populations  [12]. This may  to some extent explain  the high  rates of both overweight and  wasting  in  this  region [13]. We should  also remember the importance of social  norms about diet and activity (among  girls  and boys)  and  the culturally de- termined  positive  image  that  being  overweight  carries  in  these  societies  [61]. Although  factors  such as  family  history, sedentary lifestyle, urbanization,  income and family diet patterns are as- sociated with a high prevalence of over- weight and obesity  in  the Middle East  [26,34,37,38,44,50,51,53], it seems that  the main underlying causes may be poor  knowledge about food choices and lack  of physical  activity  [44].  Investigators  in  the  Islamic Republic of  Iran  found  that nutrient density  is not entirely an  income-driven  issue and,  even  for  the  better-off  families,  satisfaction of needs  for nutrients in the diet was only attain- able  at  energy  levels of 3000 kcal per  day [12]. Hence, to tackle the problem  at  the country  level,  it  is useful  to have  as much information as possible about  the levels of overweight and obesity (e.g.  ethnicity, gender, social class, education,  EMHJ  •  Vol. 16  No. 9  •  2010 Eastern Mediterranean Health Journal La Revue de Santé de la Méditerranée orientale 1016 habits)  in order  to  target  interventions  as precisely as possible.  There are  several methodological  constraints that made the summation  and  summarization of  the  different  datasets difficult in this review. Differ- ent  definitions  and  standard  values  for assessing body weight  status was  the principal difficulty,  as discussed  earlier  in  this  section. Another  con- straint was  the  variation  among  the  surveys in the age groups studied and  we could not locate a uniform scheme  for comparison. In addition, it is likely  that the equipment and measurement  techniques  used  were  not  system- atically  standardized. Cross-national  comparisons  should  be  conducted  with caution because a country may  Freedman DS et al. The relation of overweight to cardiovascu-1. lar risk factors among children and adolescents: the Bogalusa Heart Study. Pediatrics, 1999, 103:1175–82. Bellizzi MC et al. Prevalence of childhood and adolescent over-2. weight and obesity in Asian and European countries. In: Chen C, Dietz WH, eds. Obesity in childhood and adolescence. (Nestle Nutrition Workshop Series. Pediatric Program. Volume 49.) Phila- delphia, Lippincott Williams and Wilkins, 2002:23–35. EU childhood obesity “out of control”. 3. International Obes- ity Task Force [press release] (http://www.iotf.org/popout. asp?linkto=http://www.iotf.org/media/IOTFmay28.pdf, ac- cessed 7 March 2010). Obesity: preventing and managing the global epidemic. Report 4. of a WHO consultation. Geneva, World Health Organization, 1999. Fight childhood obesity to help prevent diabetes, say WHO & IDF 5. [press release]. Geneva, World Health Organization, 2004 (http://www.who.int/mediacentre/news/releases/2004/ pr81/en/print.html, accessed 5 March 2010). Chinn S, Rona RJ. Prevalence and trends in overweight and 6. obesity in three cross-sectional studies of British children, 1974–94. British Medical Journal, 2001, 322:24–26. Freedman DS et al. The relation of childhood BMI to adult adi-7. posity: the Bogalusa Heart Study. Pediatrics, 2005, 115(1):22–27. The Work of WHO in the Eastern Mediterranean Region: an-8. nual report of the Regional Director. Alexandria, World Health Organization Regional Office for the Eastern Mediterranean, 2003. Global strategy for non-communicable disease prevention and 9. control (draft). Geneva, World Health Organization, 1997 (WHO/NCD/GS/97.1). Ogden CL et al. Prevalence of overweight among preschool 10. children in the United States, 1971 through 1994. Pediatrics, 1997, 99(4):e1 (doi:10.1542/peds.99.4.e1). Wang Y. Cross-national comparison of childhood obesity: the 11. epidemic and the relationship between obesity and socio- References economic status. International Journal of Epidemiology, 2001, 30:1129–1136. Kelishadi R. Childhood overweight, obesity and the metabolic 12. syndrome in developing countries. Epidemiologic Reviews, 2007, 29:62–76. De Onis M, Blossner M. prevalence and trends of overweight 13. among preschool children in developing countries. American Journal of Clinical Nutrition, 2000, 72:1032–1039. Wang Y, Wang JQ. A comparison of international references 14. for the assessment of child and adolescent overweight and obesity in different populations. European Journal of Clinical Nutrition, 2002, 56(10):973–982. Physical status: the use and interpretation of anthropometry. 15. Report of a WHO Expert Committee. Geneva, World Health Or- ganization, 1995 (WHO Technical Report Series No. 854). Bellizzi MC, Dietz WH. Workshop on childhood obesity: sum-16. mary of the discussion. American Journal of Clinical Nutrition, 1999, 70:173–175S. Defining childhood overweight and obesity. 17. Centers for Disease Control and Prevention [website] (http://www.cdc.gov/obes- ity/childhood/defining.html, accessed 16 June 2010). Cole TJ et al. Establishing a standard definition for child over-18. weight and obesity worldwide: international survey. British Medical Journal 2000, 320:1240–1243. De Onis M, Habicht JP. 19. Anthropometric reference data for inter- national use: recommendations from a WHO Expert Committee (http://www.unu.edu/Unupress/food/V182e/ch12.htm, ac- cessed 7 March 2010). Must A, Dallal GE, Dietz WH. Reference data for obesity: 85th 20. and 95th percentiles of body mass index (wt/ht2) and triceps skinfold thickness. American Journal of Clinical Nutrition, 1991, 53:839–846. Al-Sendi AM, Shetty P, Musaiger MO. Prevalence of over-21. weight and obesity among Bahraini adolescents: a comparison between three different sets of criteria. European Journal of Clinical Nutrition, 2003, 57(3):471–474. have  great  variation  by  region,  as  highlighted  in  the  Islamic Republic  of Iran and Saudi Arabia, where differ- ent provinces and ethnic groups were  studied  [22–24,30,31,35,37–40,43].  This may make  it difficult  to  reach a  general view of the extent of the prob- lem of overweight and obesity, at least  unless nationally  representative data  are available, but many of the surveys  reproduced here do not meet such re- quirements. Not all of the studies had  multiple  data  points  available,  thus  impeding  the estimation of  trends  in  overweight over  time. Finally,  there  may be  reports published  in  journals  not indexed in Medline (including lo- cal journals), which were not accessed  and included in this review. Conclusions This article quantifies  the magnitude  of  childhood  overweight  and  obes- ity  in  the Middle East  region which  can  serve  as  a  baseline  for  assessing  future  patterns.  Although  the  shift  towards  a more obese population  is  perceived  from  the current  literature,  few  countries  in  the  region  enjoy  access  to  nationwide  representative  data about  the extent of  this problem.  The present  estimates  can  also help  identify countries and regions  in need  of  population-wide  interventions  to  prevent childhood overweight, and  to  encourage cross-national  initiatives  to  establish monitoring  and preventing  programmes. طسوتلما قشرل ةيحصلا ةلجلماشرع سداسلا دلجلما عساتلا ددعلا 1017 Kelishadi R et al. Thinness, overweight and obesity in a national 22. sample of Iranian children and adolescents: CASPIAN Study. Child: Care, Health and Development, 2007, 34:44–54. Dorosty AR, Siassi F, Reilly RR. Obesity in Iranian children. 23. Ar- chives of Diseases in Childhood, 2002, 87(5):388–391. Mohammadpour-Ahranjani B et al. Prevalence of overweight 24. and obesity in adolescent Tehrani students, 2000–2001: an epidemic health problem. Public Health Nutrition, 2004, 7(5):645–648. Esmaillzadeh A et al. High prevalence of metabolic syndrome 25. among Iranian adolescents. Obesity, 2006, 14(3):377–382. Majali H et al. Socio-economic discrepancies in growth status 26. of Jordanian children in military-run schools at the turn of the twentieth century. Saudi Medical Journal, 2003, 24(5):548– 549. Al-Isa AN. Body mass index, overweight and obesity among 27. Kuwaiti intermediate school adolescents aged 10–14 years. Eu- ropean Journal of Clinical Nutrition, 2004, 58(9):1273–1277. Al-Isa AN, Moussa MA. Obesity among Kuwaiti pre-school 28. children aged 0–5 years: prevalence and comparison with the NCHS/CDC reference population. Nutrition and Health, 1998, 12(4):235–246. Al-Isa AN, Moussa MA. Nutritional status of Kuwaiti elementary 29. school children aged 6–10 years: comparison with the NCHS/ CDC reference population. International Journal of Food Sci- ences and Nutrition, 2000, 51(4):221–228. Chakar H, Salameh PR. Adolescent obesity in Lebanese private 30. schools. European Journal of Public Health, 2006, 16(6):648–651. Sibai AM et al. Prevalence and correlates of obesity in Leba-31. non: findings from the first epidemiological study. Obesity Research, 2003, 11(11):1353–1361. Hameida J, Billot L, Deschamps JP. Growth of preschool 32. children in the Libyan Arab Jamahiriya: regional and sociode- mographic differences. Eastern Mediterranean Health Journal, 2002, 8(4–5):458–469. Osman YF et al. Progression of obesity among Seeb school 33. children in Oman. A preliminary study. Saudi Medical Journal, 2004, 25(12):2038–2040. Hakeem R. Socioeconomic differences in height and body 34. mass index of children and adults living in urban areas of Karachi, Pakistan. European Journal of Clinical Nutrition, 2001, 55(4):400–406. Bener A. Prevalence of obesity, overweight and underweight in 35. Qatari adolescents. Food and Nutrition Bulletin, 2006, 27(1):39– 45. Kamal AA, Bener A, Al-Mulla AK. Growth pattern of Qatari pre-36. school children. Croatian Medical Journal, 2004, 45(4):461–465. El-Hazmi MA, Warsy AS. A comparative study of prevalence 37. of overweight and obesity in children in different provinces of Saudi Arabia. Journal of Tropical Pediatrics, 2002, 48(3):172– 177. Al-Nuaim AR, Bamgboye EA, Al-Herbish A. The pattern of 38. growth and obesity in Saudi Arabian male school children. International Journal of Obesity and Related Metabolic Disorders, 1996, 20(11):1000–1005. Malik M, Bakir A. Prevalence of overweight and obesity among 39. children in the United Arab Emirates. Obesity Reviews, 2006, 8:15–20. Al-Haddad FH, Bertis L, Ghafar AG. Childhood obesity in 40. United Arab Emirates schoolchildren: a national study. Annals of Human Biology, 2005, 32(1):72–79. Musaiger AO, Gregory WB. Anthropometry of adolescent girls 41. in Bahrain, including body fat distribution. Annals of Human Biology, 2000, 27(5):507–515. Hesketh T, Ding QJ. Standard definition of child overweight 42. and obesity worldwide. British Medical Journal, 2000, 321:1158– 1159. Abalkhail B. Overweight and obesity among Saudi Arabian 43. children and adolescents between 1994 and 2000. Eastern Mediterranean Health Journal, 2002, 8(4–5):470–479. Mirmiran P, Mirbolooki M, Azizi F. Familial clustering of obes-44. ity and the role of nutrition: Tehran Lipid and Glucose Study. International Journal of Obesity and Related Metabolic Disorders, 2002, 26(12):1617–1622. Moussa MA et al. Factors associated with obesity in Kuwaiti 45. children. European Journal of Epidemiology, 1999, 15(1):41–49. Al-Mahroos F, Al-Roomi K. Overweight and obesity in the 46. Arabian Peninsula: an overview. Journal of the Royal Society of Health, 1999, 119(4):251–253. Moussa MA et al. Factors associated with obesity in school 47. children. International Journal of Obesity and Related Metabolic Disorders, 1994, 18(7):513–515. Al-Hourani HM, Henry CJ, Lightoweler HJ. Prevalence of over-48. weight among adolescent females in the United Arab Emirates. American Journal of Human Biology, 2003, 15(6):758–764. Al-Almaie SM. Prevalence of obesity and overweight among 49. Saudi adolescents in Eastern Saudi Arabia. Saudi Medical Jour- nal, 2005, 26(4):607–611. Al-Isa AN, Moussa MA. Factors associated with overweight and 50. obesity among Kuwaiti kindergarten children aged 3–5 years. Nutrition and Health, 1999, 13(3):125–139. Kelishadi R et al. Obesity and associated modifiable environ-51. mental factors in Iranian adolescents: Isfahan Healthy Heart Program—heart health promotion from childhood. Pediatrics International, 2003, 45(4):435–342. El-Tawila S et al. 52. Transitions to adulthood: a national survey of Egyptian adolescents. Cairo, Population Council, 1999:25– 30. Al-Rukban MO. Obesity among Saudi male adolescents in 53. Riyadh, Saudi Arabia. Saudi Medical Journal, 2003, 24(1):27– 33. De Onis M, Blossner M. The World Health Organization global 54. database on child growth and malnutrition: methodology and application. International Journal of Epidemiology, 2003, 32:518–526. Guillaume M. Defining obesity in childhood: current practice. 55. American Journal of Clinical Nutrition, 1999, 70:126–130. The Asia–Pacific perspective: redefining obesity and its treatment56. . Manila, World Health Organization Regional Office for the Western Pacific/International Association for the Study of Obesity/International Obesity Task Force, 2000. Al-Hazza HM. Physical activity, fitness and fatness among Saudi 57. children and adolescents. Implications for cardiovascular health. Saudi Medical Journal, 2002, 23(2):144–150. Wang Y et al. Standard definition of child overweight and obes-58. ity worldwide. British Medical Journal, 200, 321:1158. Reilly JJ, Dorosty AR, Emmett PM, et al. Identification of the 59. obese child: adequacy of the body mass index for clinical practice and epidemiology. International Journal of Obesity and Related Metabolic Disorders, 2000, 24(12):1623–1627. UNICEF at a glance: Bahrain60. . United Nations Children’s Fund [website]. (www.unicef.org/infobycountry/bahrain.html, ac- cessed 7 March 2010). O’Dea JA. Gender, ethnicity, culture and social class influ-61. ences on childhood obesity among Australian schoolchil- dren: implications for treatment, prevention and community education. Health and Social Care in the Community, 2008, 16(3):282–290. طسوتلما قشرل ةيلماعلا ةحصلا ةمظنلم ةيميلقلإا ةنجللا ءاضعأ نادلبلا ةيملاسلإا ناريإ ةيروهجم . ةيبيللا ةيبرعلا ةييرهمالجا . سنوت . نيرحبلا . ناتسكاب . ةدحتلما ةيبرعلا تاراملإا . ناتسناغفأ . ندرلأا صرم . نانبل . تيوكلا . رطق . ينطسلف . نماُع . قارعلا . لاموصلا . نادوسلا . تيوبيج . ةينميلا ةيروهملجا . ةيروسلا ةيبرعلا ةيروهملجا ةيدوعسلا ةيبرعلا ةكلملما . برغلما Subscriptions and Distribution Enquiries regarding subscriptions and distribution of the print edition of EMHJ should be addressed to: Printing and Marketing of Publications at: email: pam@emro.who.int; tel: (+202) 2276 5000; fax: (+202) 2670 2492 or 2670 2494 Permissions Requests for permission to reproduce or translate articles, whether for sale or non-commercial distribution should be addressed to EMHJ at: emhj@emro.who.int Members of the WHO Regional Committee for the Eastern Mediterranean Afghanistan . Bahrain . Djibouti . Egypt . Islamic Republic of Iran . Iraq . Jordan . Kuwait . Lebanon Libyan Arab Jamahiriya . Morocco . Oman . Pakistan . Palestine . Qatar . Saudi Arabia . Somalia Sudan . Syrian Arab Republic . Tunisia . United Arab Emirates . Republic of Yemen Membres du Comité régional de l’OMS pour la Méditerranée orientale Afghanistan . Arabie saoudite . Bahreïn . Djibouti . Égypte . Émirats arabes unis . République islamique d’Iran Iraq . Jamahiriya arabe libyenne . Jordanie . Koweït . Liban . Maroc . Oman . Pakistan . Palestine . Qatar République arabe syrienne . Somalie . Soudan . Tunisie . République du Yémen Correspondence Editor-in-chief EMHJ WHO Regional Office for the Eastern Mediterranean P.O. Box 7608 Nasr City, Cairo 11371 Egypt Tel: (+202) 2276 5000 Fax: (+202) 2670 2492/(+202) 2670 2494 Email: sabrib@emro.who.int/emhj@emro.who.int EASTERN 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, con- cepts, 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 Col- laborating Centres and individuals within and outside the Region. LA REVUE DE SANTÉ DE LA MÉDITERRANÉE ORIENTALE EST une revue de santé officielle publiée par le Bureau régional de l’Organisation mondiale de la Santé pour la Méditerranée orientale. Elle offre une tribune pour la présentation et la promotion de nouvelles politiques et initiatives dans le domaine des ser-vices de santé ainsi qu’à l’échange d’idées, de concepts, de données épidémiologiques, de résultats de recherches et d’autres informations, se rapportant plus particulièrement à la Région de la Méditerranée orientale. Elle s’adresse à tous les professionnels de la santé, aux membres des instituts médicaux et autres instituts de formation médico-sanitaire, aux ONG, Centres collabora- teurs de l’OMS et personnes concernés au sein et hors de la Région. EMHJ is a trilingual, peer reviewed, open access journal and the full contents are freely available at its website: http://www/emro.who.int/emhj.htm EMHJ is abstracted/indexed in the Index Medicus and MEDLINE (Medical Literature Analysis and Retrieval Systems on Line) and the ExtraMed-Full text on CD-ROM, the Cumulative Index to Nursing and Allied Health Literature (CINAHL), CAB International, Lexis Nexis, Scopus and the Index Medicus for the WHO Eastern Mediterranean Region (IMEMR). ©World Health Organization 2010 All rights reserved Disclaimer 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 World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. Dotted lines on maps represent approximate border lines for which there may not yet be full agreement. 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. All reasonable precautions have been taken by the World Health Organization to verify the information contained in this publication. However, the published material is being distributed without warranty of any kind, either express or implied. The responsibility for the interpretation and use of the material lies with the reader. In no event shall the World Health Organization be liable for damages arising from its use. The named authors alone are responsible for the views expressed in this publication. ISSN 1020-3397 Cover designed by Diana Tawadros Internal layout designed by Emad Marji and Diana Tawadros Printed by WHO Regional Office for the Eastern Mediterranean ميدقتل برنم ىهو .ةيلماعلا ةحصلا ةمظنمب طسوتلما قشرل ىميلقلإا بتكلما نع ردصت ىتلا ةيمسرلا ةلجلما ىه ةيئابولا تايطعلماو ميهافلماو ءارلآا لدابتلو ،اله جيوترلاو ةيحصلا تامدلخا فى ةديدلجا تاردابلماو تاسايسلا لك لىإ ةهجوم ىهو .طسوتلما قشر ميلقإب اهنم قلعتي ام ةصاخو ،تامولعلما نم كلذ يرغو ثاحبلأا جئاتنو زكارلماو ،ةينعلما ةيموكلحا يرغ تماظنلما اذكو ،ةيميلعتلا دهاعلما رئاسو ةيبطلا تايلكلاو ،ةيحصلا نهلما ءاضعأ .هجراخو ميلقلإا فى ةحصلاب ينمتهلما دارفلأاو ةيلماعلا ةحصلا ةمظنم عم ةنواعتلما طسوتلما قشرل ةيحصلا ةلجلما Cover 6.indd 2 5/24/2010 11:46:47 AM