Epistatic Interactions Among Co‐Inherited Malaria‐Protective Red Blood Cell Polymorphisms and Their Effects Across the Infection‐Mortality Spectrum in Endemic Populations: A Systematic Review

Background Malaria remains a major global health burden, with red blood cell (RBC) polymorphisms conferring partial protection. However, the epistatic interactions among co‐inherited RBC polymorphisms are poorly understood. We aimed to characterise how epistatic interactions among malaria‐protective RBC polymorphisms modify risk across the infection‐to‐mortality spectrum in endemic populations. Methods We systematically searched seven databases, trial registries, preprint servers and grey literature sources for studies reporting multilocus genotype‐phenotype associations for P. falciparum malaria, with duplicates removed and records screened by title and abstract. Full‐text reports were assessed and data were extracted using a structured form, independently verified by two reviewers. Risk of bias was evaluated with a modified Newcastle–Ottawa Scale. Joint genotypes were compared against wild‐type references within each study, and single‐locus effects were collated to establish a multiplicative “no‐interaction” expectation. Observed joint effects were classified as synergistic, antagonistic, or indeterminate based on study‐specific thresholds. Results Ten studies met inclusion, yielding 18 joint‐genotype contrasts across eight polymorphism pairs. Nine contrasts (50%) were antagonistic, six (33%) indeterminate, and three (17%) synergistic. The most reproducible signal was the antagonism between HbAS and α + ‐thalassaemia. Similar negation was observed for HbAS + G6PD A − , CR1 Sl 2 + α + ‐thalassaemia, and G6PD A − + CYB5R3 T117S, although the findings were based on fewer and smaller strata. Synergy was rare and confined to haptoglobin 2‐1 + α + ‐thalassaemia and HbS allele carriers + G6PD A − . Certainty of evidence was low or very low for 14/18 contrasts owing to small dual carrier strata, hospital case‐control designs and heterogeneity in endpoints. Margin‐shift sensitivity analyses indicated that the predominance of antagonism and the direction of the main HbAS and α + ‐thalassaemia and CR1 Sl 2 + α + ‐thalassaemia were robust to reasonable changes in the departure thresholds. Conclusions In the limited and heterogenous literature on multilocus malaria resistance, statistical epistasis between malaria‐protective RBC polymorphisms is common and more often antagonistic than synergistic, although the certainty of evidence was generally low. These findings suggest polygenic risk models, vaccine efficacy analyses and transmission or gene‐drive models should account for prevalent multilocus genotypes and their interaction terms rather than assuming independent single‐locus effects. Larger harmonised cohorts coupled with CRISPR‐based mechanistic studies are needed to refine effect sizes and clarify the biological pathways underlying these statistical interactions.