October 28, 1996

Depressive Affect and Survival Among Gay and Bisexual Men Infected With HIV

Author Affiliations

From the Center for AIDS Prevention Studies (Drs Mayne, Vittinghoff, Chesney, Barrett, and Coates), Division of General Internal Medicine, Department of Medicine (Drs Mayne, Chesney, Barrett, and Coates), and Division of Clinical Epidemiology, Department of Epidemiology (Drs Chesney and Barrett), University of California, San Francisco.

Arch Intern Med. 1996;156(19):2233-2238. doi:10.1001/archinte.1996.00440180095012

Background:  Although depression has been related to chronic disease processes and outcomes, studies examining the relationship between depression and disease progression in persons with human immunodeficiency virus (HIV) infection have produced inconsistent results.

Objective:  To investigate whether depressive affect is associated with HIV mortality.

Methods:  This was a prospective cohort study (San Francisco Men's Health Study) using a population-based probability sample of single men living in areas of San Francisco, Calif, with high case rates of acquired immunodeficiency syndrome. Data collection was at 6-month intervals with up to 15 waves of follow-up. Subjects included all 402 homosexual or bisexual men who in July 1984 had serologic evidence of HIV infection and who survived a minimum of 1 wave after baseline. Depressive affect was measured at each wave with the affective subscale of the Center for Epidemiologic Studies— Depression Scale. Laboratory markers, World Health Organization staging, antiretroviral use, and hospitalizations were also used as predictor variables. The primary outcome measure was survival.

Results:  In a Cox proportional hazards model, a time-dependent measure of depressive affect was associated with greater mortality (adjusted risk ratio, 1.67; 95% confidence interval, 1.01-2.78).

Conclusions:  Depressive affect was associated with mortality risk, highlighting the importance of diagnosis and treatment of depression among HIV-infected gay and bisexual men. Replication of this effect in incident cohorts with well-characterized dates of HIV infection is needed and, if replicated, an exploration of mediating pathways suggested.Arch Intern Med. 1996;156:2233-2238