We analyzed finger-prick blood samples collected in 9 villages in Yambio at the end of the malaria peak (January–February 2020) from persons >6 months of age with symptomatic malaria infection positively diagnosed by pan-pLDH–based RDT (CareStart Malaria PAN [pLDH] Ag RDT; Access Bio, https://accessbio.net). We performed malaria confirmation and speciation of 594 dried blood spot samples by multiplex PCR (5). Confirmed P. falciparum samples with high DNA quality (n = 518) underwent genotyping and molecular analysis for deletions in exon 2 of pfhrp2 and pfhrp3 (5) (Appendix). Demographic information was collected for all samples (Appendix). We defined multiplicity of infection (MOI) as the number of parasitic genotypes per infection and analyzed for a random subsample (n = 419) by amplifying P. falciparum merozoite surface protein 1 and 2 genes (pfmsp1, pfmsp2) (6). We defined monoclonal infection as the detection of a single PCR fragment for each locus and polyclonal infection as the detection of >1 PCR fragment for >1 locus.

Figure 1

Figure 1. Frequencies of Plasmodium falciparum single and double pfhrp2 and pfhrp3deletions in malaria-hyperendemic region, South Sudan. Color represents the type of deletion and proportion of...

Overall deletion frequency (including samples with both single and double deletions) among 518 genotyped PCR-positive samples was 15.6% for pfhrp2 and 20.0% for pfhrp3 (Table 1). Double deletions were found in 7.5% of isolates; patients at Yambio State Hospital had nearly twice the rate of double deletions (15.9%) as patients at all other sites (Figure 1). In 7/9 study sites, >10% of samples did not amplify pfhrp2; >5% of isolates were double-deleted in nearly half (4/10) of sites. Pfhrp2 deletion rates in South Sudan were as high as or higher than the country’s immediate neighbors, where reported deletion rates from specific sites have varied from 26% in Ethiopia to 19% in Kenya, 6% in the Democratic Republic of the Congo, 3% in Uganda, and <1% in Sudan (7–11).

Monoclonality was the only factor significantly associated with both pfhrp2 and pfhrp3 deletions and double deletions (Table 2). Even so, the frequency of deletion among polyclonal infections was higher than expected. Severe malaria cases exhibited significantly more pfhrp3 and double deletions than uncomplicated infections. Patients >14 years of age were more likely to harbor deletions than were patients <5 and 5–14 years of age, although the difference was significant only for pfhrp3 (Table 2).

Unequal deletion distribution between proximate geographic zones has been previously reported (9,10), and variability in double deletion rates by study site in Yambio was notable, suggesting possible hot spots. However, the higher rates seen at Yambio State Hospital could also be explained by its significantly higher proportion of monoclonal infections (Appendix). The small sample sizes at individual sites leave conclusions about hot spots unsettled, but they underscore the importance of realistic malaria control strategies targeted to the local molecular marker landscape.

Figure 2

Figure 2. Multivariable regression model of Plasmodium falciparum pfhrp2 deletions (A), pfhrp3deletions (B), and double deletions (C) in malaria-hyperendemic region, South Sudan. Panels A, C, and...

We used a statistically significant multivariable logistic regression model for total MOI, infection severity, and patient age factors for pfhrp2, pfhrp3 and double deletions (Figure 2). Age was the only significant predictor of pfhrp2 deletions (after adjusting for MOI and severity of infection), whereas age, MOI, and clinical severity were all predictors for pfhrp3 deletions. Only MOI was a significant predictor of double deletions (Appendix).

Most research surrounding the interaction between MOI and pfhrp2 and pfhrp3 deletions concludes that polyclonal infections mask deletions and lead to underestimates in deletion prevalence (12). Our results support this conclusion, finding lower MOI in Yambio associated with most deletions. In high-transmission settings, younger persons tend to have higher MOIs (13). We also found lower MOI and older patient age associated with deletions, contrasting with studies that have linked deletions to lower age but failed to consider MOI as a confounder (9).

In addition, the fitness-cost of deletions (the effects on the parasite after losing 1 of its proteins) could be another way that age, disease severity, and deletion risk interact, because milder disease has previously been associated with pfhrp2 deletion (14) and persons acquire immunity against P. falciparum as they age (13). In this area, our cohort breaks with consensus, finding deletions more commonly in complicated malaria patients. We believe this difference might reflect the difficulties of diagnosing febrile disease in South Sudan, where the signs of severe malaria might be caused by other undetected infections.

Sample collection in this study occurred at the end of the high-intensity malaria transmission season, when potentially high parasitic diversity but low prevalence could favor spread of gene-deleted organisms, making deletions easier to detect (14). When interpreting pfhrp2 deletion surveillance, the transmission period should be considered (9,15).

This study was limited because it was a secondary analysis of a study of molecular markers of antimalarial drug resistance and did not follow WHO protocols for pfhrp2 and pfhrp3 deletion surveillance. Consequently, the precise prevalence of pfhrp2 and pfhrp3 gene deletions causing false-negative results on RDTs in South Sudan was not generated to assess whether it is within the 5% threshold established by WHO (2). We also exclusively used pan p-LDH RDT–positive samples, preventing us from evaluating the effects of deletions on malaria diagnoses.

Characterizing pfhrp2 and pfhrp3 deletions is critical to designing effective public health strategies for malaria control. This study describes these deletions in a clinical cohort in a country with little previous endemic evidence of pfhrp2 and pfhrp3 deletion. Monoclonal infections were a principal predictor of deletions. We identified high levels of single and double deletions of pfhrp2 and pfhrp3, which if more widely present in this or other regions of South Sudan, could s eriously jeopardize HRP2-based RDT effectiveness moving forward. Future studies should be designed according to WHO protocol to produce precise estimates to measure the risk those deletions pose in South Sudan. Local variation in prevalence suggests the potential for deletion hotspots within the country and should be considered when designing malaria control strategies.