A double-blind, placebo-controlled trial was set up to compare clindamycin and pyrimethamine as prophylaxis for toxoplasmic encephalitis (TE) in HIV-infected patients at risk of the disorder. Interim analysis showed that clindamycin-treated patients were 4.4 (95% confidence interval 1.3-15.2) times more likely to experience an adverse effect that necessitated withdrawal of the study drug than those who received placebo. Diarrhoea and rash were reported in 16 (31%) and 11 (21%), respectively, of 52 patients treated with clindamycin (300 mg twice daily) compared with 2 (6%; p = 0.06) and none (p = 0.01) of the 32 placebo-treated patients. The clindamycin arm of the trial was prematurely terminated, although recruitment to the pyrimethamine arm continues.