Systematic review of Plasmodium falciparum and Plasmodium vivax polyclonal infections: Impact of prevalence, study population characteristics, and laboratory procedures

Multiple infections of genetically distinct clones of the same Plasmodium species are common in many malaria endemic settings. Mean multiplicity of infection (MOI) and the proportion of polyclonal infections are often reported as surrogate marker of transmission intensity, yet the relationship with traditional measures such as parasite prevalence is not well understood. We have searched Pubmed for articles on P. falciparum and P. vivax multiplicity, and compared the proportion of polyclonal infections and mean MOI to population prevalence. The impact of the genotyping method, number of genotyping markers, method for diagnosis (microscopy/RDT vs. PCR), presence of clinical symptoms, age, geographic region, and year of sample collection on multiplicity indices were assessed. For P. falciparum, 153 studies met inclusion criteria, yielding 275 individual data points and 33,526 genotyped individuals. The proportion of polyclonal infections ranged from 0–96%, and mean MOI from 1–6.1. For P. vivax, 54 studies met inclusion criteria, yielding 115 data points and 13,325 genotyped individuals. The proportion of polyclonal infections ranged from 0–100%, and mean MOI from 1–3.8. For both species, the proportion of polyclonal infections ranged from very low to close to 100% at low prevalence, while at high prevalence it was always high. Each percentage point increase in prevalence resulted in a 0.34% increase in the proportion of polyclonal P. falciparum infections (P<0.001), and a 0.78% increase in the proportion of polyclonal P. vivax infections (P<0.001). In multivariable analysis, higher prevalence, typing multiple markers, diagnosis of infections by PCR, and sampling in Africa were found to result in a higher proportion of P. falciparum polyclonal infections. For P. vivax, prevalence, year of study, typing multiple markers, and geographic region were significant predictors. In conclusion, polyclonal infections are frequently present in all settings, but the association between multiplicity and prevalence is weak.