Causes of Death among People Living with AIDS in the Pre- and Post-HAART Eras in the City of São Paulo, Brazil

OBJECTIVE We examine the trend in causes of death among people living with AIDS in the city of São Paulo, Brazil, in the periods before and after the introduction of highly active antiretroviral therapy (HAART), and we investigate potential disparities across districts of residence. METHODS Descriptive study of three periods: pre-HAART (1991-1996); early post-HAART (1997-1999); and late post-HAART (2000-2006). The data source was the São Paulo State STD/AIDS Program and São Paulo State Data Analysis Foundation. Causes of death were classified by the ICD-9 (1991-1995) and ICD-10 (1996-2006). We estimated age-adjusted mortality rates for leading underlying causes of death and described underlying and associated causes of death according to sociodemographic characteristics and area of residence. We used Pearson's chi-square test or Fisher's exact test to compare categorical variables. Areas of residence were categorized using a socioeconomic index. To analyze trends we apply generalized linear model with Poisson regression. RESULTS We evaluated 32,808 AIDS-related deaths. Between the pre- and late post-HAART periods, the proportion of deaths whose underlying causes were non-AIDS-related diseases increased from 0.2% to 9.6% (p<0.001): from 0.01% to 1.67% (p<0.001) for cardiovascular diseases; 0.01% to 1.62% (p<0.001) for bacterial/unspecified pneumonia; and 0.03% to 1.46% (p<0.001) for non-AIDS-defining cancers. In the late post-HAART period, the most common associated causes of death were bacterial/unspecified pneumonia (35.94%), septicemia (33.46%), cardiovascular diseases (10.11%) and liver diseases (8.0%); and common underlying causes, besides AIDS disease, included non-AIDS-defining cancers in high-income areas, cardiovascular diseases in middle-income areas and assault in low-income areas. CONCLUSIONS The introduction of HAART has shifted the mortality profile away from AIDS-related conditions, suggesting changes in the pattern of morbidity, but heterogeneously according to area of residence. There is a need for public policies aimed at adapting health care services to address the new scenario.