Background Increasing resistance to sulfadoxine-pyrimethamine is leading to a decline in its effectiveness. We aimed to assess the safety profile of chlorproguanil-dapsone (CD), and to compare the safety and efficacy of this drug with that of sulfadoxine-pyrimethamine (SP) as treatment for uncomplicated falciparum malaria. Methods We undertook a double-blind, randomised trial in 1850 consecutively recruited children with uncomplicated falciparum malaria, pooling data from five African countries. Analyses were based on all randomised patients with available data. Findings CD was significantly more efficacious than SP (odds ratio 3·1 [95% CI 2·0-4·8]); 1313 patients (96%) given CD and 306 (89%) given SP achieved acceptable clinical and parasitological response by day 14. Adverse events were reported in 46% and 50% of patients randomised to CD and SP, respectively (treatment difference -4·4%, [95% CI -10·1 to 1·3]). Haemoglobin in the CD group was significantly lower than in the SP group at day 7, a difference of -4 g/L (95% CI -6 to -2). Mean day 14 haemoglobin (measured only for the small number of patients whose day 7 data caused concern) was 94 g/L (92-96) and 97 g/L (92-102) after CD and SP, respectively. Glucose-6-phosphate dehydrogenase deficient patients on CD had greater odds than those on SP of having a fall of 20 g/dL or more in haemoglobin when baseline temperature was high. Methaemoglobinaemia was seen in the CD group (n=320, mean 0·4% [95% CI 0·4-0·4]) before treatment, 4·2% (95% CI 3·8-4·6) (n=301) at day 3, and 0·6% (0·6-0·7) (n=300) at day 7). Interpretation CD had greater efficacy than SP in Africa and was well tolerated. Haematological adverse effects were more common with CD than with SP and were reversible. CD is a useful alternative where SP is failing due to resistance.