Abstract / Summary
Background: Sub-Saharan Africa (SSA) bears the world’s greatest HIV burden, prompting sustained interest in host genetic modifiers of susceptibility, including the ABO blood group system. Yet reported associations between ABO phenotype and HIV infection remain inconsistent, and whether they reflect true biological effects or uncontrolled population structure is unresolved. Methods We conducted a PRISMA-informed structured narrative review of studies examining ABO phenotype and HIV infection in SSA, searched across PubMed/MEDLINE, Google Scholar, Scopus, Web of Science, Embase, and African Journals Online. Given marked heterogeneity in design, denominators, and reporting across eligible studies, a formal meta-analysis was not undertaken. Instead, for studies reporting extractable population-level ABO frequencies, we explicitly tested whether HIV-associated phenotype patterns were distinguishable from background distribution. Results Fifteen studies from nine SSA countries met inclusion criteria. Blood group O was the most prevalent phenotype both in the general population and among HIV-positive individuals in every study reporting background data. Among six studies with quantifiable denominators, only one (Korir et al., Kenya) used an independent background comparator; its HIV-positive proportion (45.6%, 95% CI 39.8–51.4%) fell squarely within the background range (45–49%), showing no detectable excess. The remaining five studies derived HIV and background estimates from the same sample, precluding independent confounding assessment. A notable outlier — 46.2% blood group AB among an HIV-exposed infant cohort (Okonkwo et al.) versus an expected 3–5% — represented the largest deviation observed, though based on a small subgroup requiring replication. Only two of fifteen studies reported adjusted estimates. Conclusion Existing SSA evidence does not support a reproducible, confounding-independent association between ABO phenotype and HIV infection; apparent group-O associations largely reflect background prevalence. Standardized reporting of adjusted estimates and population phenotype distributions is needed going forward.