correspondence E. Gorissen, Burgemeester Bickerstraat 5c, 1111 BZ Diemen, The Netherlands
Summary We conducted two randomized clinical trials to determine the in vivo efficacy of amodiaquine and sulfadoxine/pyrimethamine in treating Plasmodium falciparum malaria. Seventy-five patients under the age of 10 years in Kibwezi, Kenya, and 171 patients in Kigoma, Tanzania, were enrolled for treatment. Due to loss of eight patients in Kibwezi and 37 in Kigoma to follow-up, we used best and worst case scenarios for the parasitological response. The in vivo sensitivity of Plasmodium falciparum to amodiaquine was 75% (no loss to follow-up) in Kibwezi and ranged from 85% in the best to 65% in the worst case scenario in Kigoma. The sensitivity to sulfadoxine/pyrimethamine was 70% to 88% in Kibwezi and 65% to 89% in Kigoma. R1 resistance to amodiaquine was 22% in Kibwezi and varied from 6% in the best to 26% for the worst case scenario in Kigoma. The R1 resistance to sulfadoxine/pyrimethamine was 5% to 23% in Kibwezi and 2% to 26% in Kigoma. R2 resistance was 3% for amodiaquine and 7% for sulfadoxine/pyrimethamine in Kibwezi and 9% in Kigoma for each treatment group. There was no statistically significant difference between treatment groups at either study site, except for a slight difference in R1 resistance in the best case scenario, Kibwezi, in favour of S/P. Although both amodiaquine and sulfadoxine/pyrimethamine resistance seems to be increasing, these antimalarials are still effective in parasite clearance.
Plasmodium falciparum malaria remains a major cause of childhood morbidity and mortality in Kenya and Tanzania. The first-line treatment for malaria is still chloroquine, but increasing resistance requires a change in this policy. Amodiaquine (AQ) is beneficial for a certain period after occurrence of significant resistance against chloroquine, thus delaying the switch to sulfadoxine/pyrimethamine (S/P) or other more expensive drugs (Ollario et al. 1996). To determine the in vivo efficacy of both AQ and S/P, two studies were conducted in Kibwezi, Kenya and Kigoma, Tanzania, as part of the continuous monitoring of drug resistance in Kenya and Tanzania coordinated by the African Medical and Research Foundation (AMREF).
The studies were undertaken at the end of the rainy season, May to August 1997, in Kibwezi Health Center, Makueni District of Eastern Province, Kenya, a mesoendemic malaria area, and in Maweni Hospital, Kigoma, Tanzania, which is situated near Lake Tanganyika and known for holoendemic malaria.
Children up to 10 years of age with Plasmodium falciparum monoinfections presenting at the outpatient clinic of the study sites were eligible for the study if they fulfilled the inclusion criteria ( Table 1) and informed verbal consent was obtained from parent or guardian.
Table 1. Inclusion criteria for an in vivo efficacy study of amodiaquine and sulfadoxine/pyrimethamine for treatment of non-severe malaria in Kibwezi, Kenya and Kigoma, Tanzania
Age, sex, drug allergy, symptoms and therapy during current illness were recorded. The physical examination included rectal (age < 3 years) or oral (age ≥ 3 years) measurement of temperature, weight and checking for hepatosplenomegaly and jaundice. Fingerprick blood samples were collected for a thin and a thick bloodsmear and haemoglobin measurement.
Treatment and follow-up
Patients were assigned to the treatment groups as follows: the first patient was allocated AQ (30 mg/kg over 3 days) or a single dose of S/P (1/4 tablet/5 kg body weight; one tablet contains 25 mg pyrimethamine and 500 mg sulfadoxine) by tossing a coin. The next patient received the other drug and subsequent patients were treated with one of these drugs in turn. If a child vomited within 30 min after medication, the full dose was given again. Patients were reviewed on days 1, 2, 7 and 14 after start of treatment (day 0). AQ was administered again under supervision on day 1 and 2. On each occasion, the symptoms were recorded, the temperature was taken, a clinical examination and a finger prick for parasite count performed. To measure the clinical response to therapy, a score was used in which each malaria symptom contributed 1 point ( Table 1).
On days 0 and 14 the haemoglobin concentration was measured. Patients with haemoglobin levels of 5 and 7 g/dl were given ferrous sulphate tablets (10 mg/kg/day) from day 2 until day 30. All patients with a temperature > 37.9 °C received paracetamol syrup in accordance with local policy, although it was recently shown that paracetamol may prolong parasite clearance ( Brandts et al. 1997 ).
Patients with a parasite density on day 2 of > 50% of the baseline value were evaluated again on day 3. If the parasite count had not decreased by 75% then, this was considered a clinical failure and the patient was treated with single-dose S/P or intravenous quinine. Patients who showed resistance to AQ or S/P on day 14 were treated with S/P.
Clinical failures were considered to be patients with life-threatening clinical deterioration and febrile (> 37.9 °C) individuals with parasitaemia on days 4–7 without any other clinical cause. They were treated with quinine (15 mg/kg loading dose, then 10 mg/kg every 12 h for 3 days) and/or S/P (1/4 tablet/5 kg bodyweight). Cure was defined as a good clinical and parasitological response, no signs and symptoms of malaria and no parasitaemia (after counting 100 white blood cells in the thick film) on day 14. A negative blood smear was a parasite-free thick film.
All blood smears were stained using Field stain (Nevill et al. 1995) and the number of asexual parasites was counted per 100 white blood cells using the thick smear. The fixed thin smear was used to verify if the patient had pure Plasmodium falciparum malaria. In Kibwezi the haemoglobin concentration was measured by a Compur #3 (Minilab, Bayer Diagnostics, Leverkusen, Germany) haemoglobin meter, after mixing 20 µl finger prick blood with 5 ml Drabkin solution ( Cook 1985). In Kigoma the haemoglobin concentration was determined by measuring the packed cell volume.
Response to therapy
The response of Plasmodium falciparum to a drug has been standardized by WHO (Nevill et al. 1995) into four parasitological categories:
Sensitive: sustained parasite clearance on day 7 with no reoccurrences until day 28;
Resistant R1: parasite clearance on day 7, with recrudescence between day 7 and day 28;
Resistant R2: ≥ 75% reduction in parasitaemia within 48 h but no clearance of parasites until day 7;
Resistant R3: < 75% reduction in parasitaemia within 48 h and clinical deterioration.
We presumed that no de novo infection would develop within 14 days and that recurrent parasitaemia was true recrudescence.
A minimal sample size of 30 subjects per drug is recommended by WHO (Nevill et al. 1995). Normally distributed data means were compared using Student's t-test for independent samples while drug efficacy was compared by using the chi-square test. The 95%-confidence interval for the differences in percentages was calculated using SPSS.
AMREF has been given ethical clearance by the Kenya and Tanzania Ministry of Health to monitor drug resistance patterns in both countries. Verbal informed consent was obtained from parent or guardian after information was given in the local language.
Findings in Kibwezi, Kenya
Of 130 eligible patients randomized to the different treatment groups, 55 had to be withdrawn eventually because they did not fulfil the inclusion criteria. Forty-five of these patients (82%) had inadequate parasitaemia on day 0, six (11%) were older than 10 years, three (5%) had severe anaemia requiring blood transfusion (one of those died) and one (2%) had a history of S/P use.
Of the remaining 75 patients, 32 (42.7%) had been randomized to AQ treatment and 43 (57.3%) to S/P. Both groups were comparable for age (P = 0.275), sex (P = 0.460), weight (P = 0.822), parasitological (P = 0.070) and haematological indices (P = 0.682) on day 0 ( Table 2) and history of chloroquine use.
Table 2. Baseline charcteristics of an in vivo efficacy study of amodiaquine (AQ) and sulfadoxine/pyrimethamine (S/P) in Kibwezi, Kenya and Kigoma, Tanzania
The parasitological responses to therapy are summarized in Table 3. Because eight patients in the S/P group were lost to follow-up, two scenarios for the parasitological response are possible: a best case scenario, in which these eight patients are considered cured; and a worst case scenario, in which the response is assumed to be R1 resistant. A difference in R1 resistance was found in the best case scenario in Kibwezi, showing significantly less R1 resistance in the S/P treatment group. The other differences in sensitivity and resistance between the two treatment groups in both worst and best case scenario are statistically not significant.
Table 3. Parasitological response of an in vivo efficacy study of amodiaquine (AQ) and sulfadoxine/pyrimethamine (S/P) in Kibwezi. Those lost to follow-up are assumed to be cured (sensitive) in the best case scenario and resistant (R1) in the worst case scenario. Table contains absolute numbers, percentages (in parentheses), differences between AQ and S/P in percentages and the 95% confidence interval of the differences (in brackets)
Patients with parasitaemia on day 14 received fresh treatment with either a repeated dose of S/P or intravenous quinine. The seven patients of the AQ group who showed R1 resistance received S/P; six responded well. One developed severe malaria but was successfully treated with intravenous quinine. All patients of the S/P group with R1 and R2 resistance were treated with a repeated dose of S/P and cured. There were no significant differences between the treatment groups in clinical improvement.
Overall, the mean haemoglobin concentration tended to increase regardless of therapy received. The number of patients with moderate anaemia (haemoglobin < 7 g/dl) was 24 on day 0; 13 were treated with AQ and 11 with S/P. On day 14, 5 patients still had moderate anaemia, three in the AQ and two in the S/P group. These differences are not statistically significant.
Twenty-eight patients were afebrile (temperature < 37.9 °C) at the time of inclusion and remained so throughout the illness. The mean temperatures during follow-up of the remaining 47 patients with fever decreased significantly faster in the AQ than in the S/P group (fever was cleared on day 1 in the AQ group and on day 2 in the S/P group; P = 0.049, Fig. 1).
Findings in Kigoma, Tanzania
Of the 171 patients included, 85 received AQ and 86 S/P. There were no significant differences in age (P = 0.278), sex (P = 0.492), weight (P = 0.758), parasitological (P = 0.213) and haematological indices (P = 0.738) ( Table 2) or history of chloroquine use between the two treatment groups on day 0.
Thirty-seven patients (including Rwandan refugees) did not complete the follow-up, 17 of whom received treatment with AQ and 20 with S/P. Results of best and worst case scenarios are shown in Table 4. All patients with R1 and R2 resistance, both in the AQ and the S/P group, were cured with a repeated dose of S/P.
Table 4. Parasitological response of an in vivo efficacy study of amodiaquine (AQ) and sulfadoxine/pyrimethamine (S/P) in Kigoma. Those lost to follow-up are assumed to be cured (sensitive) in the best case scenario and resistant (R1) in the worst case scenario. Table contains absolute numbers, percentages (in parentheses), differences between AQ and S/P in percentages and the 95% confidence interval of the differences (in brackets)
All patients improved clinically, but those treated with S/P showed significantly fewer symptoms on days 7 and 14 (mean score day 7: 1.116, day 14: 0.656) than patients in the AQ group (mean score day 7: 1.634, P = 0.037 and day 14: 1.443, P = 0.000).
The number of patients with moderate anaemia was 16 on day 0, equally divided over the two treatment groups. Two patients in the AQ group and none in the S/P group still had anaemia on day 14. These differences are not statistically significant. 103 Patients remained afebrile during their illness. The mean temperature of the 68 febrile patients decreased significantly faster in the AQ group than in the S/P group (day 2: P = 0.020, Fig. 1).
Resistance of Plasmodium falciparum to antimalarials is a major problem. We studied the in vivo effectivity of AQ and S/P in children with uncomplicated falciparum malaria in East-Africa, where chloroquine resistance is widespread (Mutabingwa et al. 1992). We used the 14 day in vivo-test. In endemic countries recurrence of parasites may be due to recrudescence or reinfection. Since the incubation period of Plasmodium falciparum is about 7–10 days and therapeutic blood levels of both AQ and S/P persist for 5–9 days, most recurrences of parasites within 14 days are due to recrudescence rather than reinfection.
Both AQ and S/P are still effective in the treatment of malaria and S/P may be the best first-line treatment for malaria at present (Nevill et al. 1995; Henry et al. 1996 ). However, our results ( Tables 3 and 4) show that the in vivo sensitivity of Plasmodium falciparum to AQ is 75% in Kibwezi and 65–85% in Kigoma, while the sensitivity to S/P is 70–88% in Kibwezi and 65–89% in Kigoma. These differences are statistically not significant. The only difference is found in Kibwezi, where R1 resistance is significantly lower in the S/P group. However, in the Kigoma study and according to Dillen et al. (1999) the efficacies of AQ and S/P are equal, thus we conclude that, because of the small number of patients and the wide 95% confidence interval, the Kibwezi results are not all consistent with reality.
Which antimalarial could act as first-line treatment remains an important question. Chloroquine resistance is unacceptably high in Kenya ( Bloland et al. 1993 ). Both AQ and S/P, with rare true clinical failures, are more effective antimalarials, although the resistance levels are high. Halofantrine, an alternative antimalarial, was shown to be superior to chloroquine, AQ and S/P treatment in Eldoret District, Kenya ( Anabwani et al. 1996 ), but is much more expensive and complicated to use, requiring assessment of QT time on the ECG.
Therefore, in our opinion, AQ and/or S/P should replace chloroquine as first-line treatment. In choosing between AQ and S/P, which are equally effective in parasite clearance, other aspects are important. AQ is significantly faster in fever clearance than S/P, as demonstrated in this study and by others (Ollario et al. 1996), and the efficacy of AQ could be improved by a higher dosage regimen of 35 mg/kg, as used in West Africa ( Dillen et al. 1999 ). S/P, on the other hand, is superior to AQ both in reducing the symptoms on day 14 according to the Kigoma study ( Muller et al. 1996 ; Ollario et al. 1996), and because of its single-dose administration. AQ, by contrast, has to be given over 3 days, which does not improve compliance.
Several steps are necessary for the control of malaria in East Africa. Effective and affordable first-line treatment should be available and used properly everywhere. Combination treatment will reduce the rate at which resistance will develop. Education about malaria treatment and information for health workers and the population are important.
We thank those who supported this malaria research project, especially AMREF East-Africa, the Kibwezi Primary Health Center and Maweni Hospital in Kigoma. We are grateful to Drs A. H. A. ten Asbroek and Dr J. F. Wendte, Department of Social Medicine, Academic Medical Center, University of Amsterdam. Funding was provided by AMREF, the Spinoza Fund (Amsterdam University Union) and the Department of Infectious Diseases, Tropical Medicine and AIDS, Academic Medical Center, Amsterdam.