Antiretroviral therapy (ART) is efficacious in improving clinical outcomes among people living with HIV (PLWH) and reducing HIV transmission when taken regularly. Research examining modifiable factors associated with ART non-adherence is critical for informing novel intervention development in settings with high HIV prevalence. Alcohol use has been linked with ART non-adherence in studies in sub-Saharan Africa; however, no review has pooled estimates across studies. We reviewed studies of alcohol use and ART non-adherence conducted in sub-Saharan Africa. We searched PubMed, CINAHL, EMBASE, and PsycINFO through August 2019 with terms related to ART non-adherence, alcohol use, and sub-Saharan Africa. One author reviewed titles/abstracts (n = 754) and two authors reviewed full texts (n = 308) for inclusion. Discrepancies were resolved by group consensus. Studies were retained if they quantitatively measured associations between alcohol use and ART non-adherence or viral non-suppression. We defined ART non-adherence using the definitions from each parent study (e.g., patients with > 5% missed ART doses during the previous four, seven or 30 days were considered non-adherent). A random effects meta-analysis was conducted to pool associations and we conducted additional analyses to assess between-study heterogeneity and publication bias and sensitivity analyses to determine robustness of our results when considering only certain study designs, alcohol use or ART scales, or studies that used viral non-suppression as their primary outcome. Of 56 articles meeting our inclusion criteria, 32 articles were included in the meta-analysis. All studies measured alcohol use via self-report. ART non-adherence was assessed using self-report, pill counts, or pharmacy records and definition of non-adherence varied depending on the measure used. Individuals who used alcohol had twice the odds of ART non-adherence compared with those who did not use alcohol (34% non-adherence among alcohol users vs. 18% among non-users; pooled odds ratio: 2.25; 95% confidence interval: 1.87-2.69; p < 0.001). We found evidence of a high degree of heterogeneity between studies (Cochrane Q statistic: 382.84, p< 0.001; I2 proportion: 91.9%) and evidence of publication bias. However, the magnitude of our pooled odds ratio was consistent across a number of sensitivity analyses to account for heterogeneity and publication bias. In a secondary analysis with studies using viral non-suppression as their primary outcome, we also estimated a statistically significant pooled effect of alcohol use on viral non-suppression (pooled odds ratio: 2.47; 95% confidence interval: 1.58-3.87). Evidence suggests alcohol use is associated with ART non-adherence in Sub-Saharan Africa, potentially hindering achievement of the UNAIDS 90-90-90 HIV treatment targets.
Keywords: ART adherence; Africa; Alcohol; HIV; Systematic review.