OBJECTIVE: The objective of this study was to test the association of positive affect as measured by the Center for Epidemiologic Studies Depression Scale (CES-D) with risk of AIDS mortality, controlling for the other CES-D subscales and laboratory measures of disease progression.

METHODS: Data come from the San Francisco Men’s Health Study, a prospective study of a household probability sample of single men in San Francisco. The subjects were 407 men who were HIV+ at study baseline.

RESULTS: In time-dependent Cox proportional hazards models, the positive affect subscale of the CES-D was significantly associated with lower risk of AIDS mortality (RR = 0.89, CI = 0.84–0.95). When risk estimates were adjusted for time-dependent covariates of CD4, serum ß₂-microglobulin, P24 antigen, antiretroviral use, and the other subscales of the CES-D, positive affect remained significantly predictive of lower risk of AIDS mortality (RR = 0.90, CI = 0.85–0.97). When the CES-D subscale predictors were lagged by 12, 24, and 36 months in order to address the possibility that positive affect was simply a marker for better health, positive affect remained significantly predictive lagged by 12 months and marginally predictive lagged by 24 months.

CONCLUSIONS: Positive affect seems to be the "active ingredient" in the association of scores on the CES-D depressive mood scale and mortality in this sample of HIV+ men. Future work should expand the traditional negative-affect-only focus to encompass the significant role that positive affect plays in living with HIV.