Antiretroviral therapy and declining AIDS mortality in New York City.

The objective was to evaluate the association between antiretroviral therapy and AIDS mortality in New York City (NYC). Design was a population-based case-control study. We randomly selected 150 case patients and 150 control patients whose AIDS diagnosis was made during 1994 to 1996 (male:female, 2:1) from among 19,238 persons reported to the NYC Health Department HIV/AIDS Reporting System (HARS). Case patients had died of AIDS-related causes in 1996. Control patients, category matched with case patients on gender, were not known to have died by the end of 1996. Analysis was performed on 279 patients (142 cases and 137 controls). Cases and controls were similar in age, gender, race, HIV transmission category, and health insurance coverage. The median baseline CD4 count was 30 cells/microL for those who died and 103 cells/microL for survivors (p < 0.001). The prescription of HAART (antiretroviral combination that includes at least one protease inhibitor) in 1996 was strongly associated with survival in univariate analysis (OR = 5.1, 95%CI = 2.5-10.2). This association remained in a logistic regression analysis after adjusting for sex, age, race, health insurance status, HIV transmission categories, year of AIDS diagnosis, baseline CD4 count, and other antiretroviral therapy (AOR = 8.6, 95%CI = 3.5-20.7). Prescription of combination therapy other than HAART in 1996 and baseline CD4 count were also associated with survival, but less strongly so. The survival benefit of HAART extends beyond the confines of a few highly selected patients into the "real world," reducing AIDS deaths at the population level. This population-based study supports the likelihood that the introduction of HAART in 1996 played a primary role in the decline in NYC AIDS mortality.