Review of evidence from risk factor analyses associating HIV infection in African adults with medical injections and multiple sexual partners

Abstract

Studies of risk factors for human immunodeficiency virus (HIV) infections in sub-Saharan Africa present a wealth of evidence relevant to ongoing debates about the contributions of unsterile health care and sexual transmission to Africa's HIV epidemics. From studies which meet search criteria (n = 39) we calculate population attributable fractions (PAFs) for incident and prevalent HIV infections associated with exposure to medical injections and with having more than one sexual partner. Median and mean crude PAFs for injections for both incident and prevalent HIV exceed those for multiple partners. Evidence suggests that adjustments for reverse causation (people with HIV-related symptoms seeking injections) and confounding do not explain away the large PAFs for injections. Misreporting of sexual behaviour has an unknown impact on PAFs for multiple partners. However, most PAFs for incident HIV infections are from communities with low-growth epidemics; hence data about the most important risk factors fuelling high-growth epidemics are sparse. Empirical resolution of ongoing debates may be possible with the aid of future research—especially in high-growth epidemic settings—that investigates risks for HIV transmission through a full range of sexual and parenteral exposures with attention to reverse causation, confound, and quality of sexual behaviour data.