Patients >10 kg in weight who had slide-confirmed acute P. falciparum malaria were included in the study, provided that they or their guardians gave fully informed written consent intheir own language, they were not pregnant, they had not received mefloquine in the previous 63 days and there were no other clinical or laboratory signs of severe illness and/or severe and complicated malaria 13. If they gave written informed consent, they were allocated randomly to receive either the six-dose regimen of artemether-lumefantrine (Coartem^® 20/120, Novartis Pharma AG, Basel, Switzerland) or mefloquine (Lariam^®, Hoffman-La Roche, Basel, Switzerland) plus artesunate (Guilin Pharmaceutical Factory No.1, Guilin, China). At enrolment (Day 0), a medical history was obtained, a full physical examination was performed and blood was taken for quantitative parasite counts and routine haematology (finger prick blood sample for malaria smear and haematocrit). All information was recorded on a standard case record form. All patients were examined and blood smears were taken daily until they became aparasitaemic, and then weekly for 6 weeks. At each visit a questionnaire on adverse events was completed. A blood smear was also taken from any patient complaining of fever or symptoms compatible with malaria during the follow-up period. Parasite counts were determined on Giemsa-stained thick and thin blood films. The person-gametocyte-weeks were calculated per 1,000 person-weeks after excluding the episodes on admission and during treatment.

Computerized randomization was in blocks of ten. Patients allocated to artemether-lumefantrine group (ALN) received the tablets at 0 and 8 hours and twice daily for the following 2 days. Artemether-lumefantrine was dispensed as a fixed dose combination tablet. Each tablet contained 20 mg of artemether and 120 mg of lumefantrine. The number of tablets was given according to the body weight. The minimum dosage for patients weighing less than 15 kg was one tablet per dose; patients between 15 and 24 kg received two tablets, those between 25 and 34 kg received three tablets and patients 35 kg and above were treated with four tablets per dose. Patients allocated to artesunate-mefloquine group (MAS3) received artesunate, 4 mg/kg oncedaily for 3 days (day 0 was the first day of treatment), plus mefloquine, 15 mg/kg on day 1 and 10 mg/kg on day 2.

Each patient was given antipyretics and cooled by tepid sponging if the tympanic temperature was equal or above 37.5°C before drug administration. Drug administration was observed in all patients and if vomiting occurred in less than 30 min, administration of the full dose was repeated, if vomiting occurred between 30 and 60 min, half the dose was repeated. Patients treated with artemether-lumefantrine were given a glass of chocolate milk (200 ml) with each dose to increase absorption 14.

The primary therapeutic outcome measure in this study was the incidence of microscopically and genotypically confirmedrecrudescent infections in both treatment groups by day 42. Parasite genotyping by the polymerase chain reaction (PCR) was used to distinguish recrudescent from newly acquired P. falciparum infections. P. falciparum infections were genotyped for allelic variation in three polymorphic antigen loci, merozoite surface proteins 1 and 2 (MSP-1 and MSP-2) and glutamate rich protein (GLURP), on admission and in case of parasite reappearance 1516. Secondary measures were the immediate treatment responses: parasite clearance, fever clearance, incidence of adverse events, and degree of anaemia. The sample size was calculated to detect a difference in failure rates of 7 % between the two regimens with 90% CI and 80 % power assuming a 20% drop out.

Adverse events were symptoms or signs that were not presenton admission and that developed after the start of treatment. All adverse events, including those probably related to malaria, were recording and compared among treatment groups. The rates of early vomiting (<1 h) after each dose and for each drug wererecorded and compared among the groups in the analysis.

Patients with uncomplicated recrudescent infections were re-treated with artesunate, 2 mg/kg/day for 7 days; patients >8 years oldalso received doxycycline, 4 mg/kg/day for 7 days.

Data were analysed using SPSS for Windows, version 11. Categorical data were compared using the Chi-square test with Yates' correction or by Fisher's exact test, as appropriate. Continuous variables conforming to a normal distribution were compared using Student's t test. Data not normally distributed were log-transformed or compared using the Mann-Whitney U test. The relative risks were calculated using cross-tabulations. The rates of adverse events at three different periods (days 1–2, days 3–7 and days 14–42) were compared among treatment groups. For each of the three periods, the events were counted only once (e.g., if a patient vomited on day one and day two, this was counted as one adverse event). The PCR-adjusted cure rates were evaluated by survival analysis and compared using the log-rank test. Patients for whom PCR genotyping was either inconclusive or missing were censored in the survival analysis on the day of parasite reappearance. For all statistical tests the significance level (p) was set at 0.05.

The initial responses to the two treatment groups were similar. None of the patients developed severe malaria. On admission, 55.0% (133/242) of the patients on ALN and 57.9% (140/242) of the patients in the MAS3 group had a tympanic temperature ≥ 37.5°C. All except three patients had a normal temperature on day 3 (2 in ALN and 1 in MAS3). There was no difference in fever clearance times between the two treatment groups (Figure 1). Parasite clearance times were short and most patients cleared their parasitaemia by day two. Figure 2 shows the percentage of patients with positive slide for asexual P. falciparum in both groups. By day three, four (1.8%) of 227 patients in the artemether-lumefantrine recipients and three (1.3%) of 238 artesunate-mefloquine recipients still had a positive blood film (P > 0.05). Overall, 12.3 % of patients were anaemic (haematocrit <30%) on admission, 10.8% in ALN group and 13.8% in MAS3 group (P = 0.33). The mean (SD) decrease in haematocrit value at day seven from baseline was greater in the group receiving MAS3 than in the ALN group: 9.3% (SD,11.5%; 95% CI, 7.7% to 10.9%) compared with 6.7% (SD, 11.4%; 95% CI, 5.1 to 8.3%) respectively (P = 0.023).

During the 42-day follow-up period, 27 new P. falciparum infections occurred among artemether-lumefantrine and 24 among artesunate-mefloquine recipients (P > 0.05) (Table 2). The PCR-adjusted cure rates by day 42 were 98.8% (95% CI, 96.4% to 99.6%) in the ALN group and 96.3% (95% CI, 93.1% to 98.0%) in the MAS3 group (P = 0.08). PCR confirmed treatment failures were more likely in children aged below 15 years than in adults (RR, 5.1; 95% CI, 1.4–18.7; P = 0.006). The mean age was 13.6 years (n = 12; SD = 8.5) in patients with treatment failure and 23.7 years (n = 438; SD = 15.2) in successfully treated patients. In this trial, only age group was independently associated with treatment failure, but not other factors e.g. higher parasitaemia (>10,000/μL); anaemia (haematocrit <30%); fever (tympanic temp ≥ 37.5°C) on admission; sites; treatment groups; early vomiting (within one hour following drug administration). The mediantime to recrudescence was comparable for MAS3 group (21 days; n = 9; range, 14–28 days) and ALN group (28 days; n = 3; range, 21–42 days; P > 0.05).

Of 452 patients, 119 (26.3%) had P. vivax parasitaemia detected during follow up. There were significantly fewer cases of vivax malaria in the MAS3 group (29 of 227) than in the ALN group (90 of 225) (P < 0.001). The median time to appearance of P. vivax parasitaemia was longer in the MAS3 group (40 days; range, 13–43 days) than in the ALN group (28 days; range, 14–43 days; P < 0.001).

Twenty patients (8.3%) in the artemether-lumefantrine group and 19 (7.9%) in the artesunate-mefloquine group had gametocytes detected during the first 3 days. All except one patient in ALN group cleared gametocytes within first week after start of treatment. After excluding these, gametocytes developed (between day 7 and 42) in 1.2% (3/241) of ALN group and 1.3 (3/240) of MAS3 group. The person-gametocyte weeks were low and similar: 2.7 (95% CI, 0.6–7.8) per 1000 person-weeks for both groups.

Both treatment regimens were well tolerated. No serious adverse events were reported. Overall, 5/242(2.1%) of the patients vomited one or more doses of medication in the ALN group and 2/242(0.8%) of the MAS3 treated patients (RR, 2.5; 95% CI, 0.5–12.7; P = 0.45). The rates of early vomiting (within one hour) of the drugs were very low (around 2%) and did not differ among groups (one in each group on day 2).

The most commonly reported and possibly drug-related adverse events to both combination therapies were effects on the gastrointestinal (abdominal pain, anorexia, nausea, diarrhoea and late vomiting e.g. >1 h after administration of treatment) and central nervous system (headache, dizziness). Figure 3 shows the proportions of possibly drug-related adverse events of those who did not have those symptoms at admission during follow-up in both groups. Overall, there were less adverse events in ALN group compared to MAS3, though the differences were not statistically significant.