HIV-attributable causes of death in the medical ward at the Chris Hani Baragwanath Hospital, South Africa

Abstract

Introduction Data on the association between HIV infection and deaths from underlying medical conditions are needed to understand and assess the impact of HIV on mortality. We present data on mortality in the Chris Hani Baragwanath Hospital (CHBH) South Africa and analyse the relationship between each cause of death and HIV. Methods From 2006 to 2009 data were collected on 15,725 deaths including age, sex, day of admittance and of death, HIV status, ART initiation and CD4+ cell counts. Causes of death associated with HIV were cases, causes of death not associated with HIV were controls. We calculate the odds-ratios (ORs) for being HIV-positive and for each AIDS related condition the disease-attributable fraction (DAF) and the population-attributable fraction (PAF) due to HIV for cases relative to controls. Results Among those that died, the prevalence of HIV was 61% and of acquired immune deficiency syndrome (AIDS) related conditions was 69%. The HIV-attributable fraction was 36% in the whole sample and 60% in those that were HIV-positive. Cryptococcosis, Kaposi's sarcoma and Pneumocystis jirovecii, TB, gastroenteritis and anaemia were highly predictive of HIV with odds ratios for being HIV-positive ranging from 8 to 124, while genito-urinary conditions, meningitis, other respiratory conditions and sepsis, lymphoma and conditions of skin and bone were significantly associated with HIV with odds ratios for being HIV-positive ranging from 3 to 8. Most of the deaths attributable to HIV were among those dying of TB or of other respiratory conditions. Conclusions The high prevalence of HIV among those that died, peaking at 70% in those aged 30 years but still 7% in those aged 80 years, demonstrates the impact of the HIV epidemic on adult mortality and on hospital services and the extent to which early anti-retroviral treatment would have reduced the burden of both. These data make it possible to better assess mortality and morbidity due to HIV in this still high prevalence setting and, in particular, to identify those causes of death that are most strongly associated with HIV.