To summarize the association between alcohol use and human immunodeficiency virus (HIV) infection based on studies conducted in Africa, EMBASE and PubMed were searched for African studies that related alcohol use to HIV infection. Meta-analyses were conducted to obtain pooled univariate and multivariate relative risk estimates. Subgroup analyses were performed for studies having different sample types: males or females and population-based or high-risk, and ones that differentiated between problem and asymptomatic drinkers. Alcohol drinkers were more apt to be HIV+ than nondrinkers. The pooled unadjusted odds ratio (OR) from 20 studies was 1.70 (95% confidence interval, CI = 1.45-1.99). Results from 11 studies that adjusted for other risk factors produced a pooled risk estimate of 1.57 (95% CI = 1.42-1.72). Males and females had similar risk estimates, while studies involving high-risk samples tended to report larger pooled odds ratios than studies of the general population. When compared with nondrinkers, the pooled estimates of HIV risk were 1.57 (95% CI = 1.33-1.86) for non-problem drinkers versus 2.04 (95% CI = 1.61-2.58) for problem drinkers, a statistically significant difference (z = 2.08, P <0.04). Alcohol use was associated with HIV infection in Africa and alcohol-related interventions might help reduce further expansion of the epidemic.