ObjectiveTo assess the incidence ofPneumocystis cariniipneumonia (PCP) after discontinuation of either primary or secondary prophylaxis.DesignThis was a prospective, non-randomized, non-blinded study.SettingTwenty-five University-based AIDS Clinical Trials Group units.ParticipantsParticipants either had a CD4 cell count ⩽ 100 × 106/l at any time in the past and no history of confirmed PCP (group I; n = 144), or had a confirmed episode of PCP ⩾ 6 months prior to study entry (group II; n = 129). All subjects had sustained CD4 cell counts > 200 × 106/l in response to antiretroviral therapy.InterventionsSubjects discontinued PCP prophylaxis within 3 months or at the time of study entry. Evaluations for symptoms of PCP and CD4 cell counts were performed every 8 weeks. Prophylaxis was resumed if two consecutive CD4 cell counts were < 200 × 106/l.Main outcome measure(s)The main outcome was development of PCP.ResultsNo cases of PCP occurred in 144 subjects (median follow-up, 82 weeks) in group I or in the 129 subjects (median follow-up, 63 weeks) in group II (95% upper confidence limits on the rates of 1.3 per 100 person-years and 1.96 per 100 person-years for groups I and II, respectively). Eight subjects (five in group I and three in group II) resumed PCP prophylaxis after two consecutive CD4 cell counts < 200 × 106/l.ConclusionsThe risk of developing initial or recurrent PCP after discontinuing prophylaxis is low in HIV-infected individuals who have sustained CD4 cell count increases in response to antiretroviral therapy. Neither lifelong primary nor secondary PCP prophylaxis is necessary.