Background: Prophylaxis against Pneumocystis carinii pneumonia is indicated in patients with human immunodeficiency virus (HIV) infection who have less than 200 CD4 cells per cubic millimeter and in those with a history of P. carinii pneumonia. However, it is not clear whether prophylaxis can be safely discontinued after CD4 cell counts increase in response to highly active antiretroviral therapy. Methods: We conducted a randomized trial of the discontinuation of primary or secondary prophylaxis against P. carinii pneumonia in HIV-infected patients with a sustained response to antiretroviral therapy, defined by a CD4 cell count of 200 or more per cubic millimeter and a plasma HIV type 1 (HIV-1) RNA level of less than 5000 copies per milliliter for at least three months. Prophylactic treatment was restarted if the CD4 cell count declined to less than 200 per cubic millimeter. Results: The 474 patients receiving primary prophylaxis had a median CD4 cell count at entry of 342 per cubic millimeter, and 38 percent had detectable HIV-1 RNA. After a median follow-up period of 20 months (388 person-years), there had been no episodes of P. carinii pneumonia in the 240 patients who discontinued prophylaxis (95 percent confidence interval, 0 to 0.85 episode per 100 person-years). For the 113 patients receiving secondary prophylaxis, the median CD4 cell count at entry was 355 per cubic millimeter, and 24 percent had detectable HIV-1 RNA. After a median follow-up period of 12 months (65 person-years), there had been no episodes of P. carinii pneumonia in the 60 patients who discontinued prophylaxis (95 percent confidence interval, 0 to 4.57 episodes per 100 person-years). Conclusions: In HIV-infected patients receiving highly active antiretroviral therapy, primary and secondary prophylaxis against P. carinii pneumonia can be safely discontinued after the CD4 cell count has increased to 200 or more per cubic millimeter for more than three months.