Résumé : ■ Objective: To compare pyrimethamine plus clindamycin (PC) to pyrimethamine plus sulfadiazine (PS) as a treatment for toxoplasmic encephalitis (TE) in patients with the acquired immunodeficiency syndrome (AIDS). ■ Design: Randomized, unblinded phase II, multicenter trial with provision for crossover for failure or intolerance of the assigned regimen. ■ Setting: University hospitals. ■ Patients: Eighty-four patients with presumptive TE were entered. Thirteen were excluded when they were found to have another diagnosis, and 12 were excluded because they did not meet entry criteria. The baseline characteristics in the remaining 26 patients randomized to PC and 33 randomized to PS were comparable. ■ Interventions: Patients were treated for 6 weeks with pyrimethamine and folinic acid plus either sulfadiazine or clindamycin. Clindamycin was given intravenously during the first 3 weeks. ■ Measurements and Main Results: There was a trend toward greater survival in patients randomized to PS (hazard ratio, 3.25; 95% Cl, 0.63 to 16.8; P = 0.13), but most study deaths were not directly related to TE. In contrast, patients randomized to PC appeared more likely to achieve complete clinical (odds ratio, 0.67; Cl, 0.2 to 1.97; P > 0.2) and radiologic responses (odds ratio, 0.28; Cl, 0.08 to 0.96; P= 0.02). Multivariate analysis revealed drug effects to be largely independent of other variables. Similar efficacy of the treatments was also suggested by a hazard analysis of resolution of abnormal mental status, fever, and headache. Skin rash was the most common adverse event in both treatment arms. Because of toxicity, six patients randomized to PC and 11 patients randomized to PS had to switch to the alternate treatment, but only three were unable to complete therapy after crossover. ■ Conclusions: The results of several end points of efficacy, taken together, suggest that the relative efficacy of PC approximately equals that of PS. PC appears to be an acceptable alternative in patients unable to tolerate PS.