We assessed the value of clinical and laboratory parameters for predicting the occurrence of skin reactions induced by pyrimethamine/sulfadiazine and pyrimethamine/clindamycin and the effects of continued therapy for patients with these reactions. We retrospectively studied all episodes of toxoplasmic encephalitis in patients with AIDS who were treated with pyrimethamine/sulfadiazine or pyrimethamine/clindamycin. Eighteen (75%) of 24 patients treated with pyrimethamine/sulfadiazine had cutaneous reactions after a mean of 11 days, whereas 15 (58%) of 26 patients treated with pyrimethamine/clindamycin had cutaneous reactions after a mean of 13 days (P = .56). Nine (50%) of the 18 patients continued to be treated with pyrimethamine/sulfadiazine throughout the duration of hypersensitivity, compared with all 15 patients who were treated with pyrimethamine/clindamycin (P = .002). Nine patients had to stop therapy with pyrimethamine/sulfadiazine (two had Stevens-Johnson syndrome and one had Lyell's syndrome). Thus, treatment throughout the duration of hypersensitivity is more likely to succeed for patients receiving pyrimethamine/clindamycin, whereas therapy with pyrimethamine/sulfadiazine is associated with a high risk of Lyell's syndrome and Stevens-Johnson syndrome.