University of California, San Francisco Logo

University of California, San Francisco | About UCSF | Search UCSF | UCSF Medical Center

Temporal changes in causes of death among HIV-infected patients in the HAART era in Rio de Janeiro, Brazil
Global Health Sciences Literature Digest
Published January 14, 2010
Journal Article

Pacheco AG, Tuboi SH, May SB, et al. Temporal changes in causes of death among HIV-infected patients in the HAART era in Rio de Janeiro, Brazi. J Acquir Immune Defic Syndr. 2009 Aug;51(5):624-30.

In Context

With the expansion of highly active antiretroviral therapy (HAART), HIV/AIDS-related conditions and death rates in both higher- and lower-income countries have been reduced significantly. In developed countries, causes of death among HIV-infected individuals increasingly are related to other non-HIV-related diseases or underlying conditions.(1) Information from Brazil, the only low/middle-income country to make HAART universally available as early as 1996, demonstrated an initial rapid decline in death rates, which then remained stable since 1999.(2) This study examined causes of death among HIV-infected persons in Brazil initiating treatment over a 10-year period. Analysis of national death certificates revealed an increase in non-HIV-related conditions as causes of death from 1999-2004, particularly cardiovascular disease and diabetes. Brazil is in a unique position to examine these data because of its comprehensive data collection systems and national registries.


To evaluate mortality, causes of death and temporal trends among HIV-infected persons on treatment in Brazil


A university hospital (HUCFF) in Rio de Janeiro; this site participates in the International Epidemiologic Databases to Evaluate AIDS (IeDEA) network

Study Design

Retrospective analysis of a rolling cohort of HIV-infected persons on treatment


Patients who entered the treatment cohort between 1997 and 2006, who were aged 16 years or older, and who had at least one follow-up visit were eligible. Women who were followed only during pregnancy were excluded.


Mortality rates and causes of death over time, and their relation to age at start and end of observation (or death); gender; transmission group; baseline CD4 cell count; and HAART regimen.


Most data were obtained from hospital databases into which clinical information had been routinely entered following patient visits. Vital status was determined by evaluation of medical charts or linkage with the Rio de Janeiro Mortality Database. Causes of death were determined by examining death certificates, autopsy reports, medical records, and information from family members and/or health care providers. Classification of causes of death was based on the Coding Causes of Death in HIV (CoDe) protocol and evaluated by independent reviewers. Causes of death were classified as AIDS -related on non-AIDS-related depending on the presence of an AIDS-defining condition according to the United States CDC classification. Underlying or contributing causes of death using the World Health Organization (WHO) definition, were also classified as AIDS-related or non-AIDS-related. Those with unknown causes but who had a CD4 count within six months of death that was ≤200 cells/mL were considered to have died of an AIDS-related condition. Persons were considered lost to follow-up if no vital status information was available for more than one year.

Mortality was evaluated as two-year rates of AIDS-related and non-AIDS-related causes of death per 100 person-years. Results were compared to detect temporal changes. Survival analysis regression models were used, with reporting of adjusted hazard ratios, and these were performed separately for AIDS-related and non-AIDS-related causes.


Of 1538 eligible patients, 50% were women; median age at enrollment was 34 years. The primary transmission category was heterosexual transmission (50%); 16% were men who have sex with men, 2% were injection drug users. Median follow-up time was 4.6 years, and 170 were lost to follow-up (2.4 per 100 person-years). A total of 226 persons died (3.2 deaths per 100 person-years), of whom 111(49%) died due to AIDS-related causes, 98 (43%) died due to non-AIDS-related causes, and 29 (7.5%) died due to unknown causes. The primary cause of AIDS-related deaths were opportunistic infections (85/111 or 77%), which contributed to 38% of all mortality. Among non-AIDS-related causes of death, infectious diseases were the most common (19/98, or 19%); cardiovascular diseases contributed to 9/98 deaths or 9%.

Overall, death rates among those with AIDS-related compared with non-AIDS-related causes were not significantly different (1.6 vs. 1.4 deaths per 100 person-years); however, those who died from non-AIDS-related diseases were followed longer (median 3.4 yrs vs. 1.5 yrs). Time-trend analysis indicated that AIDS-associated deaths decreased significantly over time, from 2.4 per 100 person-years during 1997-1998 compared with 1.2 per 100 person-years during 2005-2006 (P<0.01), In contrast, non-AIDS-related death rates did not change significantly during the same periods (1.9 per 100 person-years vs. 1.6 per 100 person-years (P=ns).

The average age of all patients on treatment increased over time; the median age was 33 years in 1997 compared to 39 years at the end of the study. Having a low CD4 count (50 cells/mL) at baseline was associated with greater risk of an AIDS-related death; having a low baseline CD4 count and being an injection drug user were associated with greater risk of non-AIDS-related death. Although having a higher baseline age was associated with risk of AIDS-related causes of death, this risk significantly decreased over time; age had no significant effect on the risk for non-AIDS-related causes of death.


This study demonstrates that there has been a significant increase in non-AIDS-related causes of death among HIV-infected patients in Rio de Janeiro. The primary reason appears to be longer survival times due to successful treatment with HAART. These results are similar to studies from high-income countries in which there have been documented increases in cancer, cardiovascular disease, and liver disease among those on treatment. A study from New York, for example, showed an increase in the proportion of non-AIDS-related causes of death from 20% in 1999 to 26% in 2004,(3) and in another study, a more dramatic rise from 13% to 42% from 1996 to 2004.(1)

Quality Rating

This is a well-reported and high-quality study that made use of complementary data sources, validated causes of death, and employed appropriate statistical analyses. Loss to follow-up was minimized by linking to a vital statistics database. The study sample was drawn from a university hospital setting, and the extent to which the clients are representative of people in Rio de Janeiro, or other parts of Brazil, is not known. The small number of deaths, while speaking to a high quality of care, limited the ability to evaluate associations in multivariable models.

Programmatic Implications

As the life span of those successfully treated on HAART increases, attention must be paid to other chronic, long-term conditions, such as diabetes and cardiovascular diseases; avoiding "external" causes, such as violence and accidents; and appropriately treating non-HIV-related infections.


  1. Palella FJ Jr, Baker RK, Moorman AC, et al. Mortality in the highly active antiretroviral therapy era: changing causes of death and disease in the HIV outpatient study. J Acquir Immune Defic Syndr. 2006;43:27-34.
  2. Marins JR, Jamal LF, Chen SY, et al. Dramatic improvement in survival among adult Brazilian AIDS patients. AIDS. 2003;17:167582.
  3. Sackoff JE, Hanna DB, Pfeiffer MR, et al. Causes of death among persons with AIDS in the era of highly active antiretroviral therapy: New York City. Ann Intern Med. 2006;145:397-406.