We report here the results of a randomized double blind trial comparing coartemether (CGP56697), a combination of artemether and benflumetol, with pyrimethamine/sulfadoxine (P/S). Two hundred eighty-seven children 1-5 years of age with uncomplicated falciparum malaria were enrolled at two centers in The Gambia between July 1996 and December 1996. Following treatment, children were visited at home every 24 hr until a blood film free of asexual parasites was obtained. Genotyping of parasites was used to distinguish recrudescence from new infections. Three days after the start of treatment, 133 (100%) of the CGP56697-treated children compared with 128 (93.4%) of children treated with P/S had cleared their parasites (P = 0.003). The day 15 cure rate was 93.3% for CGP56697 and 97.7% for P/S (P = 0.13). Within the third and fourth week after initiation of therapy, 20 children treated with CGP56697 and one of the P/S-treated children returned with second malaria episodes (P < 0.0001). Genotyping suggested that the majority (19 of 23 [82.6%]) of these second episodes were due to new infections, supporting the World Health Organization recommendation that longer follow-up is not relevant for the assessment of drug efficacy. At the two-week follow-up, 28.9% of the P/S treated children but none of the CGP56697-treated children carried gametocytes (P < 0.0001). This study showed that CGP56697 is safe in African children with acute uncomplicated falciparum malaria, clears parasites more rapidly than P/S, and results in fewer gametocyte carriers. More frequent new infections within the third and fourth week following treatment with CGP56697 than treatment with P/S are likely to be due to the short prophylactic effect of CGP56697.