A clinical trial with halofantrine on patients with falciparum malaria in Colombia.

A clinical trial with halofantrine on patients with falciparum malaria in Colombia M. Restrepo,1 D. Botero,2 R.E. Marquez,3 E.F. Boudreau,4 & V. Navaratnam5 A total of 120 semi-immune adult male malaria patients from an area of multidrug-resistant Plasmodium falciparum malaria were hospitalized for 42 days in Medellfn, Colombia (an area of no malaria transmission), and treated with halofantrine in a double-blind, randomized, prospective clinical trial according to five different treatment schedules. Each patient was assigned to one of the following halofantrine schedules: /, one dose of 1000mg; 11, three doses of 500mg; 111, two doses of 500mg; /V, three doses of250mg; and V, one dose of 750mg. Best results (75% cure rate) were obtained with schedule II, although there was no statistically significant difference compared with the other schedules. A total of 46 patients experienced recrudescent malaria. Drug levels in plasma 72 hours after beginning treatment showed no statistically significant difference between relapsing and cured patients. Side-effects (mainly gastrointestinal) were uncommon and mild. Cardiotoxicity was studied by electrocardiogram. A mean prolongation of 28.5ms (6.6 ± 6.3% increase from baseline) was observed in the Q-T, interval on day 1 of the trial. Introduction Halofantrine is a phenanthrenecarbinol antima- larial that has proved effective in patients with chloroquine-resistant malaria caused by Plasmo- dium falciparum. Recently, cardiotoxicity has been reported with therapeutic doses of halofantrine (1). Phase I studies of orally administered halofantrine hydrochloride in healthy male volunteers revealed no significant clinical or laboratory abnormalities (2), and the drug was tolerated in single oral doses of up to 2000mg. Minimal clinical symptoms such as nausea and mild abdominal pain were recorded. A few patients had slightly elevated levels of aspartate aminotransferase (SGOT) and/or alanine amino- transferase (SGPT), which returned to normal within 2 weeks after treatment (3). In pharmaco- kinetic studies, the lag time averaged 1 hour, peak blood levels were reached 3.4-6.0 hours after admin- istration, and the mean elimination half-life was 2.6 days (4). Phase II and III studies have been carried out in patients with symptomatic P. falciparum infection, I Instituto Colombiano de Medecina Tropical and Corporaci6n para Investigaciones Biol6gicas, Apartado A6reo 52162, Medellin, Colombia. Requests for reprints should be sent to Dr Restrepo at this address. 2 Instituto Colombiano de Medecina Tropical and Universidad Pontificia Boliviariana, Medellin, Colombia. 3 El Bagre, Antioquia, Colombia. 4 Pharmaceutical Systems, Columbia, MD, USA. 5 Centre for Drug Research, University Sains, Pulau Pinang, Malaysia. Reprint No. 5742 including some in areas of known multidrug- resistant malaria. Results suggest that halofantrine administered as a single oral dose, or divided into several doses, is effective in semi-immune individu- als (3, 5). The present study is a phase II/III double-blind, randomized, comparative dose-finding clinical trial of halofantrine hydrochloride in patients with un- complicated P. falciparum malaria. Parasitaemia clearance and fever clearance time, tolerance, and response to different dosage schedules were evaluated. Subjects and methods Patient selection The 120 patients selected were semi-immune adult males with symptomatic P. falciparum parasitaemia >200 asexual forms per pl, who were recruited mainly from the coastal regions of Colombia where multidrug-resistant falciparum malaria is endemic (6, 7). All the patients were flown to Medellin, an area of no malaria transmission, and hospitalized in a tropical medicine ward. The patients' ages ranged from 11 years to 72 years, with weights in the range 27-80kg (Table 1). Each patient gave his informed consent to par- ticipate in the trial and to remain hospitalized for 42 days. Patients with severe signs of malaria or accom- panying diseases were not included in the study. Other exclusion criteria were P. falciparum parasit- aemia >50000 asexual forms per tl and central nervous system, renal, or hepatic involvement. Bulletin of the World Health Organization, 1996, 74 (6): 591-597 © World Health Organization 1996 591 M. Restrepo et al. Table 1: Comparison of age and weight in the treatment groups Patients Group Group II Group IlIl Group IV Group V Differencea No. treated 27 27 27 27 12 Age (years) Mean 26.3 31.3 29.2 27.9 21.8 No Range 14-56 11-67 14-62 15-72 14-30 SDb 10.0 15.4 10.8 13.7 4.6 - Weight (kg) Mean 56.7 58.2 61.0 57.5 55.3 No Range 32-71 27-80 43-75 44-77 39-73 SD 8.4 9.7 8.2 8.1 10.2 - a Statistically significant at the 5% level. b SD = standard deviation. Patients who could not receive oral therapy or who had received other antimalarial medications dur- ing the preceding 8 days were also excluded; this latter exclusion was made on the basis of the pa- tient's self-reported history and not on urinary drug analysis. Patient examination and follow-up Initial malaria symptoms were recorded before treatment began. Physical examinations included measurement of body weight and height; inspection of the skin and extremities; documentation of vital signs; pulmonary and cardiac auscultation; liver, spleen, and lymph node palpation; cranial nerve examination; and neurological assessment of deep tendon reflexes and motor and sensory function. Follow-up physical examinations were performed daily during the first post-treatment week, and then weekly until day 42 of the trial. Chest X-rays and electrocardiograms (ECGs) were performed on admission, and follow-up ECGs were recorded on days 1, 7, 14, 28, and 42 of the trial. Detailed analysis of the Q-T and R-R intervals was per- formed 18-30 hours after the beginning of treat- ment and compared with the baseline values of day 0 (admission). The following laboratory examinations were carried out before treatment began: haematology and clinical biochemistry tests, parasitaemia from thick and thin blood films, measurement of glucose- 6-phosphate dehydrogenase, stool examination, urine analysis, screening for haemoglobinopathies, and an in-vitro microtest for P. falciparum sensitivity to quinine, mefloquine, chloroquine, and amodia- quine. The haematological and biochemical exami- nations were repeated on days 1, 4, 7, 14, 28, and 42 of the trial. Plasma levels of halofantrine 72 hours after beginning treatment were measured using high- performance liquid chromatography (HPLC) (8). Drug administration Halofantrine tablets, each containing 250mg of the hydrochloride, were used. Each patient was assigned to a treatment schedule and received three doses of tablets in individual packets. The drug was adminis- tered by the responsible nurse assigned to the inves- tigation and ingestion of tablets was supervised. Placebo tablets were used as necessary to make the doses identical in appearance. Treatment schedules are specified in Table 2. Pre-packaged drug doses were supplied by SmithKline Beecham Pharmaceuti- cals, Brentford, England. Table 2: Treatment schedules used in the study No. of tablets Schedule First dose Second dose Third dose Total dose No. treated 4 activea 2 active, 2 placebo 2 active, 2 placebo 1 active, 3 placebo 3 active, 1 placebo 2 placebob 2 active 2 active 1 active, 1 placebo 2 placebo 2 placebo 2 active 2 placebo 1 active, 1 placebo 2 placebo a Active tablet = halofantrine (hydrochloride salt), 250mg. b Placebo tablet = inert substance identical in appearance to active tablet. WHO Bulletin OMS. Vol 74 1996 I1 11l IV V 1 000 mg 1 500 mg 1 000 mg 750mg 750 mg 27 27 27 27 12 592 A clinical trial with halofantrine in Colombia Assessment of halofantrine efficacy Antimalarial efficacy was evaluated by parasitaemia and fever clearance time and clinical cure rate. Peripheral thick blood films for asexual forms of P. falciparum were taken at 8-hour intervals until they became negative, and then daily until day 42 of the trial. Patients' temperatures were measured twice daily during hospitalization. Fever clear- ance time was defined as the period elapsed from the start of drug administration until a tempera- ture of .37.5°C was obtained for two consecutive readings. Treatment was considered curative for P. falciparum if parasites disappeared from blood films within 7 days after beginning treatment and did not recur during the 42-day follow-up period. Patients who experienced a recrudescence of parasitaemia or presented a P. vivax relapse were removed from the study. These patients were treated with amodiaquine and sulfadoxine + pyrimethamine (recrudescent P. falciparum) and chloroquine and primaquine (recrudescent P. vivax). Adverse reactions All adverse reactions emerging during the trial were recorded, as well as the date they occurred and when they disappeared. Statistical evaluation For each treatment group, the mean, range, and standard deviation were calculated for age, body weight, and for the efficacy parameters (parasi- taemia and fever clearance time). The five treatment groups were assessed for comparability before re- ceiving medication and subsequently to assess treat- ment efficacy. For continuous variables such as weight, age, and laboratory parameters, one-way analysis of variance was used to compare groups on admission to the study. Either the X2 or Fisher's exact test was used to compare variables expressed as pro- portions, such as parasite clearance or fever clear- ance time and cure rate. One-way analysis of variance was also used to compare the effect of treat- ment on the results of laboratory studies over time. Geometric means and ranges were calculated for initial parasite counts and minimal inhibitory con- centrations of antimalarial drugs as obtained from in-vitro sensitivity tests. Results Pretreatment comparability The five treatment groups were compared prior to treatment with respect to weight and age and no significant differences were found (Table 1). Initial malaria parameters (Table 3) were simi- lar, with the exception of prior antimalarial treat- ment, as more individuals in the two low-dose groups (IV and V) had been treated in the month preceding the study (P < 0.05). There were no significant dif- ferences between groups with respect to haematol- ogy, urine analysis, stool examination, or clinical chemistry tests at the time of admission to the study. Halofantrine efficacy Antimalarial effectiveness of the five halofantrine treatment schedules is summarized in Table 4. There were no significant differences between groups in cure rates, parasite clearance time, or fever clear- Table 3: Initial malarial parameters of the study participants, by treatment group Group Group II Group liI Group IV Group V Parasitaemiaa 7459 7330 7566 8875 7363 Median number of days ill prior to admission 4.5 7.0 5.5 5.0 5.5 Previous history of malaria (%) 56 44 56 54 75 Prior antimalarial treatment (%) 4 4 7 26 25 Sensitivity to:b Mefloquine 2.6 2.8 2.6 2.1 2.9 Chloroquine 17.9 18.4 24.0 17.8 11.8 Amodiaquine 1.4 1.4 1.3 1.6 1.0 Quinine 45.5 39.5 40.0 54.7 64.0 % patients with in-vitro results 77 74 63 63 75 a Asexual forms per gl (geometric mean). b Geometric mean minimal inhibitory concentration (pmol/50,ud). WHO Bulletin OMS. Vol 74 1996 593 M. Restrepo et al. Table 4: Therapeutic effectiveness of halofantrine at day 42 of follow-up, by treatment schedule Schedule Schedule II Schedule IlIl Schedule IV Schedule V No. evaluated 22 24 25 24 10 Parasite clearance time (hours) Mean 47.6 55 56.9 55.9 50.8 Range 8-96 16-112 16-118 24-180 32-82 SDa 20.9 18.9 24.7 32.8 16.2 Fever clearance time (hours) Mean 41.9 41.0 40.3 44.8 44.0 Range 24-84 24-72 12-120 12-96 12-96 SD 19.2 18.2 21.9 25.4 27.1 Cure rate (%) 55 75 48 56 40 a SD = standard deviation. ance time. Parasitaemia recurred in 46 patients; in 6 patients (from groups I, III-V), recrudescence appeared between day 28 and day 42 of the trial. Schedule II produced the highest cure rate (75%). Recrudescent patients Of the 46 recrudescent patients, in-vitro drug sensi- tivity tests on pretreatment blood samples were performed for only 7. None exhibited mefloquine resistance, but all showed chloroquine resistance. Sulfadoxine/pyrimethamine and halofantrine in- vitro sensitivities were not examined, since WHO standardized test kits were not available. In-vitro data obtained at the time of recrudescence did not demonstrate increased resistance to related drugs, such as mefloquine or quinine, in 5 out of 6 patients tested, compared with samples obtained prior to halofantrine treatment. The mefloquine minimum inhibitory concentration increased from 2 to 8 pmol/ 50,l in one patient. A total of 18 patients were withdrawn from the trial for radical treatment of P. vivax malaria. Of these, 4 had mixed infections and were counted as recrudescent falciparum patients. Recrudescent pa- tients were followed for 3-10 days following second- ary treatment. Only one patient exhibited RII resistance to the amodiaquine and sulfadoxine/ pyrimethamine regimen and was treated with quinine. Levels of halofantrine and its N-desbutyl metabolite in plasma were determined (8). Values for halofantrine plus metabolite plasma concentra- tion in recrudescent versus cured patients showed no statistically significant differences between the five groups. Plasma metabolite concentration in recru- descent versus cured patients showed a significant difference in group I (P < 0.05). There was no such difference in other groups. Adverse effects Signs and symptoms that were not present on admis- sion but appeared during the first 7 days of the trial are listed in Table 5. There were no significant differences between groups with respect to these clinical findings. Nor did laboratory evaluation demonstrate any dose-related toxicity during the first week of the trial. Electrocardiogram findings The ECGs of 85 patients could be'evaluated both on admission (day 0) and on day 1 of the trial. The day- 1 ECG was recorded 18-30 hours after the first dose of halofantrine. The Q-T and R-R intervals were measured directly with an ECG ruler on lead II, and Q-Tc intervals were calculated according to Bazett's formula (9). Direct measurement of the Q-T and R- R intervals allowed for a 20ms discrimination (i.e. +lOms for all readings). The results of day-1 ECGs are compared with day-0 baseline readings because day 1 of the trial is the period with the highest plasma concentration of halofantrine. The Q-Tc interval was prolonged in all treat- ment groups (Table 6). Aggregate results for all treatment groups show a mean increase from base- line Q-Tc interval of 25.8ms (6.6 ± 6.3%). Excluding five patients with a baseline Q-Tc interval >440ms, 23 patients (29%) had their Q-Tc interval rise to >440ms. The maximum change from baseline Q-Tc value was 21.4%, and the maximum Q-Tc interval was 495 ms. No clinically significant arrhythmias or conduction disturbances emerged during the trial. WHO Bulletin OMS. Vol 74 1996594 A clinical trial with halofantrine in Colombia Table 5: Post-treatment adverse reactions to halofantrine, by treatment group Group Group II Group IlIl Group IV Group V No. evaluated Symptom (%): Abdominal pain Heartburn Nausea Vomiting Diarrhoea Anorexia Dizziness Insomnia Tinnitus Confusion Weakness Headache Fever Chills Itching Rash Palpitation 27 27 Discussion Clinical studies of halofantrine conducted in differ- ent endemic countries and France (infected patients returning from abroad), using the currently recom- mended three-dose regimen of 500mg every 6 hours, with 28-day follow-up, have shown cure rates rang- ing from 85% to 100% (5,10-13). These studies have all achieved higher cure rates than those obtained with the treatment schedules in the current investi- gation (40-75o%). A possible explanation is that Colombian strains of P. falciparium are less sensitive to halofantrine. The lack of an effective in-vitro sensi- tivity test kit for halofantrine made it impossible to verify this hypothesis. Although it has been reported that drug resistance can occur with widespread halofantrine usage (14), such resistance cannot have affected our study since halofantrine was not avail- able in Colombia at the time of the study. In Thai- land, where multidrug resistance is widespread, the clinical efficacy of halofantrine has decreased sub- stantially over a period of years despite its very lim- ited use (5, 15). However, it has been shown that higher than recommended doses over a 3-day period are efficacious (14). The loss of efficacy with halofantrine has been assumed to be caused by the widespread use of mefloquine and quinine, resulting in cross-resistance to halofantrine. Although differences in drug absorption could have resulted in the lower efficacy observed in this study, at 72 hours treatment-group means for plasma drug concentration were similar and were not found to differ from published data. We did not administer halofantrine with food; however, according to one study the drug is better absorbed when taken with a fatty meal (16). Our results were not substantially affected by drug tolerance. Vomiting after administration oc- curred in only 3 out of 46 patients who relapsed. Two of these patients vomited 21 h and 24h after receiv- ing the drug. One patient in group III vomited 90 Table 6: Electrocardiogram findings, by treatment group No. evaluated Baseline Q-T, interval (ms) Day-1 Q-T, interval (ms) Absolute increase (ms) % increase Group 21 409.6 (18.3)a 429.8 (23.3) 20.2 5.1 (6.3) Group II 19 402.0 (24.5) 430.0 (30.0) 28.9 7.4 (7.3) Group IlIl 20 402.4 (25.0) 433.0 (23.4) 30.6 7.6 (6.7) Group IV 17 401.5 (20.1) 424.6 (24.6) 23.1 5.9 (6.2) Group V 8 403.1 (30.6) 430.6 (23.5) 27.5 7.0 (3.9) a Figures in parentheses are the standard deviation. WHO Bulletin OMS. Vol 74 1996 7 7 0 0 0 0 0 0 7 0 4 4 0 4 0 4 0 4 4 7 4 4 4 4 4 0 0 0 4 0 0 4 4 4 27 4 4 11 4 0 0 0 0 4 0 0 4 7 0 4 0 0 27 7 0 4 0 0 4 7 0 7 4 0 7 4 7 11 0 4 12 25 0 8 8 8 8 8 8 0 0 0 0 0 0 0 0 0 595 M. Restrepo et al. minutes after receiving halofantrine. This patient exhibited a prolonged parasite clearance time (118 h) and relapsed early on day 8 of the trial; thus, limited drug absorption could have resulted in treatment failure in this case. Adverse effects were minimal. Abdominal pain (8% of patients) was the most frequent side-effect. No evidence of haematological, hepatic, or renal tox- icity was noted following treatment; blood urea ni- trogen, creatinine, SGOT, SGPT, and total bilirubin levels decreased 7 days following treatment, as did white-blood-cell and platelet counts. Although a de- crease in haemoglobin was observed following treat- ment in all groups, this was probably related more to the disease (increased red-cell destruction) than to the treatment. Other studies have shown adverse effects such as diarrhoea (14%) (17), and more rarely, coughing, itching, and headache (13), which were unusual in our study. In our investigation, a careful follow-up ECG was recorded on five separate occasions after treat- ment. Comparison of ECGs on admission and on day 1 of the trial revealed a mean prolongation of Q-Tc values less than that seen in studies of halofantrine in Thailand. None of the 85 patients had Q-Tc intervals >550ms or that increased by >25% over baseline values, although in one study from Thailand 18 of 51 patients exhibited such in- creases (1). The ECGs showed no clinically signifi- cant arrhythmias and the relationship between halofantrine-induced Q-Tc prolongation and proar- rhythmia has yet to be determined. Acknowledgements We thank I.H. Vasquez and J. Lopera, who assisted in patient follow-up; R. Nieman and J. Arango, who analysed the electrocardiograms; M.H. Mejia and E. Zapata, who performed laboratory studies; and H. Roulet, for technical assistance. E.B. Doberstyn and D.E. Davidson offered useful advice. The investigation received financial support from the UNDP/World Bank/WHO Special Programme for Research and Training in Tropical Diseases. Resume Essai clinique de I'halofantrine chez des malades atteints de paludisme a falciparum en Colombie Un essai clinique de I'halofantrine a 6te realis6 sur 120 malades adultes semi-immuns, de sexe masculin, provenant d'une r6gion ou le paludisme a Plasmodium falciparum est polypharmacoresis- tant, hospitalises pendant les 42 jours de l'6tude a Medellin (Colombie), secteur indemne de transmis- sion du paludisme. Les sujets recrutes dans l'etude etaient porteurs d'une parasitemie asexuee a P. falciparum sans sympt6mes ni complications graves. Un examen clinique complet et des analy- ses de laboratoire ont ete effectues avant le traitement et les jours 1, 4, 7, 14, 28 et 42 de l'essai. Les formes asexuees de P. falciparum ont ete recherchees dans des etalements epais de sang peripherique toutes les 8 heures jusqu'A disparition puis une fois par jour jusqu'au 42e jour de l'essai. L'etude a consiste en un essai prospectif ran- domis6 en double aveugle avec des comprimes a 250mg de chlorhydrate d'halofantrine. Chaque malade recevait un schema th6rapeutique con- sistant en trois doses de comprimes distribu6es en sachets individuels. Cinq schemas therapeutiques ont ete essay6s: 1, une dose de 1000mg d'halo- fantrine; I1, trois doses de 500 mg; II1, deux doses de 500mg; IV, trois doses de 250mg; V, une dose de 750 mg. Les doses etaient administrees a inter- valles de 6 heures et avaient toutes le m6me aspect grace a I'addition d'un nombre adequat de com- prim6s placebo. Le groupe V comprenait 12 malades et les groupes I a IV 27 malades chacun. Les malades des cinq groupes ont e com- par6s avant le traitement; aucune difference significative n'a 6te observee au niveau des parametres etudies. Au 42e jour de l'essai, les taux de guerison obtenus avec les differents traitements etaient les suivants: 1, 55%; 11, 75%; III, 48%; IV, 56%; V, 40%. Le taux eleve trouv6 pour le groupe 11 n'a pas 6te juge significativement different des taux des autres groupes. Une recrudescence a e observee chez 46 malades. Les taux plasmatiques d'halofantrine 72 heures apres le debut du traite- ment n'6taient pas significativement differents entre les malades gu6ris et les cas de rechute. Les effets secondaires, essentiellement diges- tifs, ont 6te rares et b6nins. La cardiotoxicit6 du traitement a et6 etudiee par 6lectrocardiographie. On a observ6 le premier jour de l'essai un allonge- ment moyen de 28,5 ms (soit une augmentation de 6,6 ± 6,3% par rapport a la valeur de base) de l'intervalle QT. Aucun cas cliniquement significatif d'arythmies ou de troubles de la conduction n'a 6te observ6 a la suite du traitement. References 1. Nosten F et al. Cardiac effects of antimalarial treat- ment with halofantrine. Lancet, 1993, 341: 1054- 1056. 2. Rinehart J et al. Evaluation of two phenanthrene- methanols for antimalarial activity in man: WR 122,455 and WR 171,669. American journal of tropi- cal medicine and hygiene, 1976, 25: 769-774. 596 WHO Bulletin OMS. Vol 74 1996 A clinical trial with halofantrine in Colombia 3. Cosgriff TM et al. Evaluation of the antimalarial activ- ity of the phenanthrenemethanol halofantrine (WR 171,669) American journal of tropical medicine and hygiene, 1982, 31: 1075-1079. 4. Fleckenstein L et al. Pharmacokinetics and bio- availability of orally administered halofantrine (WR 171,669) in healthy volunteers. Investigational new drug report, 1983, 9847 (suppl. 10): 99-133. 5. Boudreau EF et al. Malaria: treatment efficacy of halofantrine (WR 171,669) in initial field trials in Thai- land. Bulletin of the World Health Organization, 1988, 66: 227-235. 6. Restrepo A, Alvarez L, Restrepo M. [In- vivo study of P. falciparum resistance to chloroquine in Colombia. Description of resistance ll.] Acta m6dica Colom- biana, 1980, 5: 367-379. 7. Blair-Trujillo S. [Drug resistance of P. falciparum in Colombia: 1961-1986.] Biom6dica, 1986, 6: 95-100. 8. Bouchaud LK et al. Clinical efficacy and pharma- cokinetics of micronized halofantrine for the treatment of acute uncomplicated falciparum malaria in nonimmune patients. American journal of tropical medicine and hygiene, 1994, 51: 204-213. 9. Bazett HC. An analysis of the time relations of electro- cardiograms. Heart, 1920, 7: 353. 10. Ambroise-Thomas P et al. Halofantrine in the treat- ment of P. falciparum malaria in Mali (152 cases). Proceedings of the Ninth International Congress of Infectious and Parasitic Diseases, Munich, 20-26 July 1986. Abstract 1286. 11. Coulaud JP et al. Treatment of imported cases of falciparum malaria in France with halofantrine. Trans- actions of the Royal Society of Tropical Medicine and Hygiene, 1986, 80: 615-617. 12. Richard-Lenoble D et al. Efficacy, safety and ac- ceptability of halofantrine in the treatment of acute Plasmodium falciparum malaria in African children (Gabon). Parasitology today, 1989, 5 (suppl.): 59-63. 13. Nothdurft HD et al. Halofantrine: a new substance for treatment of multidrug-resistant malaria. Clinical in- vestigation, 1993, 71: 69-73. 14. Brasseur P et al. Fast emergence of Plasmodium falciparum resistance to halofantrine. Lancet, 1993, 341: 901-902. 15. Shanks GD et al. Halofantrine for the treatment of mefloquine chemoprophylaxis failures in Plasmodium falciparum infections. American joumal of tropical medicine and hygiene, 1991, 45: 488-491. 16. Shanks GD et al. Halofantrine given with food for falciparum malaria. Transactions of the Royal Society of Tropical Medicine and Hygiene, 1992, 86: 233-234. 17. Ter Kuile FO et al. Halofantrine versus mefloquine in treatment of multidrug-resistant falciparum malaria. Lancet, 1993, 341: 1044-1049. WHO Bulletin OMS. Vol 74 1996 597