In a comparative study of infection prophylaxis, patients with acute leukemia receiving remission induction therapy were assigned either no prophylaxis, sulfamethoxazole and trimethoprim, ketoconazole, or the combination of sulfamethoxazole and trimethoprim and ketoconazole. Both sulfamethoxazole and trimethoprim and the combination of sulfamethoxazole and trimethoprim and ketoconazole substantially reduced the overall incidence of infection consequent to a marked decrease in bacterial infection. However, sulfamethoxazole and trimethoprim were associated with an increased rate of fungal infection, while ketoconazole decreased this complication. No form of prophylaxis reduced infectious mortality or increased the complete remission rate. However, because of its effect in reducing infectious morbidity, we conclude that patients with acute leukemia receiving remission induction treatment should be given antibacterial and antifungal prophylaxis.