RESULTS Compared with younger patients, patients aged 50 years or older were less likely to have acquired HIV via intravenous drug use or homosexual contact (p = .0001). Older patients were more likely to have comorbid diseases (12% vs 4%; p = .0001), had more frequent neurologic findings on admission (19% vs 9%; p = .0001), and scored higher on a PCP-specific severity-of-illness scale indicating more severe disease (p = .0001). Older patients had more intensive care unit admissions and intubations (p = .0001). Patients aged 50 years or older were less likely to have a diagnosis of HIV mentioned in their progress notes during the first 2 days of admission (75% vs 85%; p = .0001), less likely to receive PCP-specific therapy within the first 2 hospital days (58% vs 76%; p = .0001), and more likely to receive steroids (32% vs 22%; p = .0001). Older patients had a greater in-hospital mortality (32% vs 18%; p = .0001). However, in logistic regression analysis with mortality as the outcome, the effect of older age was diminished when adjustments were made for insurance status, severity of illness, comorbidity, timely PCP therapy, and inpatient use of steroids. CONCLUSIONS Age differences in mortality for AIDS-related PCP may be explained by increased severity of presenting illness, underrecognition of HIV, and delay in initiation of PCP-specific therapy. Physicians may need to consider HIV-related infections for persons aged 50 years or older at risk of HIV infection.