Objective: To quantify cause-specific mortality risk attributable to non-AIDS-related and AIDS-related causes before and after the advent of highly active antiretroviral therapy (HAART).
Methods: Competing-risk methods were used to determine the cumulative AIDS-related and non-AIDS-related risk of mortality between 1990 and the end of 2003 in the Johns Hopkins HIV Clinical Cohort, a prospective cohort study.
Results: Beginning in 1997 with the introduction of HAART, all-cause mortality declined and has remained stable at approximately 39 deaths per 1000 person-years. AIDS-related mortality continued to decline in this period (P = 0.008), whereas non-AIDS-related mortality increased (P < 0.001). Using competing-risk methods, the risk of dying attributable to AIDS-related causes remains significantly higher than the risk of dying attributable to non-AIDS-related causes for patients with a CD4 count <or=200 cells/mm in the HAART era. For those with a CD4 count >200 cells/mm, however, non-AIDS-related mortality was greater than AIDS-related mortality, particularly among injection drug users. Other transmission categories had similar AIDS-related and non-AIDS-related cumulative mortalities.
Conclusions: HAART has reduced mortality rates among HIV-infected individuals, but further efforts to reduce mortality in this population require increased attention to conditions that have not traditionally been considered to be HIV related.