Objective: To evaluate risk factors formortality among patients with AIDS in the era of highly active antiretroviral therapy (HAART), particularly the effect of cytomegalovi-rus (CMV). Design: Prospective cohort study of patients with AIDS, conducted from 1998 through 2003. Participants: One thousand five hundred eighty-three patients with AIDS, of whom 374 had CMV retinitis. Methods: Patients were contacted every 3 months, with examinations at least every 6 months, in which standardized data were collected on AIDS history and treatment, eye examinations, and hematologic, virologic, and immunologic laboratory data. Main Outcome Measure: Mortality. Results: The overall mortality rate was 0.07 deaths/ person-year. In a multivariate analysis, the following baseline risk factors were associated with an increased mortality: higher human immunodeficiency virus (HIV) viral load (relative risk [RR]= 4.6 for HIV viral load > 100 000 copies/ml vs. < 400 copies/ml; P < 0.0001), lower CD4+T-cell count at enrollment (RR=3.8 forCD4+T cell count 0-49 cells/μl vs. ≥200 cells/μl; P < 0.0001), CMV viral load ≥400 copies/ml (RR=1.9; P= 0.002), lower hemoglobin (RR=1.7 for hemoglobin < 10 g/dl; P=0.009), a history of cryptococcal meningitis (RR=1.7; P= 0.02), CMV retinitis (RR=1.6; P=0.0002), and Karnofsky score ≤80 (RR=1.4; P=0.008). Conclusions: In the era of HAART, CMV disease as manifested by CMV retinitis and a detectable CMV viral load were associated with an increased risk for mortality, even after adjusting for demographic, treatment, immunologic, and HIV virologic factors. Objective: To evaluate risk factors formortality among patients with AIDS in the era of highly active antiretroviral therapy (HAART), particularly the effect of cytomegalovi-rus (CMV). Design: Prospective cohort study of patients with AIDS, conducted from 1998 through 2003. Participants: One thousand five hundred eighty-three patients with AIDS, of whom 374 had CMV retinitis. Methods: Patients were contacted every 3 months, with examinations at least every 6 months, in standardized data were collected on AIDS history and treatment, eye Examinations, and hematologic, virologic, and immunologic laboratory data. Main Outcome Measure: Mortality. Results: The overall mortality rate was 0.07 deaths / person-year. In a multivariate analysis, the following baseline risk factors were associated with an increased mortality: higher human immunodeficiency virus (HIV) viral load (relative risk [RR] = 4.6 for HIV viral load> 100 000 copies / ml vs. <400 copies / ml; P <0.0001) ount at enrollment (RR = 3.8 for CD4 + T cell count 0-49 cells / μl vs. ≥ 200 cells / μl; P <0.0001), CMV viral load ≥ 400 copies / ml a history of cryptococcal meningitis (RR = 1.7; P = 0.02), CMV retinitis (RR = 1.6; P = 0.0002), and Karnofsky score ≤ 80 (RR = 1.7 for hemoglobin <10 g / dl; (RR = 1.4; P = 0.008). Conclusions: In the era of HAART, CMV disease as manifested by CMV retinitis and a detectable CMV viral load were associated with an increased risk for mortality, even after adjusting for demographic, treatment, immunologic, and HIV virologic factors.