Prevalence and risk of Plasmodium vivax infection among Duffy-negative individuals: a systematic review and meta-analysis

Abstract

A better understanding of the occurrence and risk of Plasmodium vivax infection among Duffy-negative individuals is required to guide further research on these infections across Africa. To address this, we used a meta-analysis approach to investigate the prevalence of P. vivax infection among Duffy-negative individuals and assessed the risk of infection in these individuals when compared with Duffy-positive individuals. This study was registered with The International Prospective Register of Systematic Reviews website (ID: CRD42021240202) and followed Preferred Reporting Items for Systematic review and Meta-Analyses guidelines. Literature searches were conducted using medical subject headings to retrieve relevant studies in Medline, Web of Science, and Scopus, from February 22, 2021 to January 31, 2022. Selected studies were methodologically evaluated using the Joanna Briggs Institute (JBI) Critical Appraisal Tools to assess the quality of cross-sectional, case-control, and cohort studies. The pooled prevalence of P. vivax infection among Duffy-negative individuals and the odds ratio (OR) of infection among these individuals when compared with Duffy-positive individuals was estimated using a random-effects model. Results from individual studies were represented in forest plots. Heterogeneity among studies was assessed using Cochrane Q and I² statistics. We also performed subgroup analysis of patient demographics and other relevant variables. Publication bias among studies was assessed using funnel plot asymmetry and the Egger's test. Of 1593 retrieved articles, 27 met eligibility criteria and were included for analysis. Of these, 24 (88.9%) reported P. vivax infection among Duffy-negative individuals in Africa, including Cameroon, Ethiopia, Sudan, Botswana, Nigeria, Madagascar, Angola, Benin, Kenya, Mali, Mauritania, Democratic Republic of the Congo, and Senegal; while three reported occurrences in South America (Brazil) and Asia (Iran). Among studies, 11 reported that all P. vivax infection cases occurred in Duffy-negative individuals (100%). Also, a meta-analysis on 14 studies showed that the pooled prevalence of P. vivax infection among Duffy-negative individuals was 25% (95% confidence interval (CI) - 3%-53%, I² = 99.96%). A meta-analysis of 11 studies demonstrated a decreased odds of P. vivax infection among Duffy-negative individuals (p = 0.009, pooled OR 0.46, 95% CI 0.26-0.82, I² = 80.8%). We confirmed that P. vivax infected Duffy-negative individuals over a wide prevalence range from 0 to 100% depending on geographical area. Future investigations on P. vivax infection in these individuals must determine if Duffy-negativity remains a protective factor for P. vivax infection.

Figure 1

Study flow diagram demonstrating study selection process.

Figure 2

Distribution of included studies on P. vivax infection among Duffy-negative individuals. Map was sourced and modified from https://mapchart.net/world.html by authors. Authors were allowed to use, edit and modify any map created with mapchart.net for publication freely by adding the reference to mapchart.net in publication.

Figure 3

Forrest plot demonstrated the pooled prevalence of P. vivax infection among Duffy negative individuals. ES prevalence estimate, CI confidence interval.

Figure 4

Forrest plot demonstrated the pooled prevalence of P. vivax infection among Duffy negative individuals stratified by continents. ES prevalence estimate, CI confidence interval.

Figure 5

Forrest plot demonstrated the odd of P. vivax infection among Duffy negative individuals. OR odds ratio, CI confidence interval.

Figure 6

Forrest plot demonstrated the odd of P. vivax infection among Duffy negative individuals stratified by continents. OR odds ratio, CI confidence interval l.

Figure 7

The funnel plot between odds ratio (OR) and standard error (se) of the logOR of the 11 studies demonstrated that the funnel plot was asymmetry. OR odds ratio, se standard error.

Figure 8

Contour-enhanced funnel plot demonstrated that the effect estimates were distributed in both significance and non-significance areas indicating that the funnel plot asymmetry was due to other causes.