Epidemiology of disappearing Plasmodium vivax malaria: a case study in rural Amazonia

Abstract

Background: New frontier settlements across the Amazon Basin pose a major challenge for malaria elimination in Brazil. Here we describe the epidemiology of malaria during the early phases of occupation of farming settlements in Remansinho area, Brazilian Amazonia. We examine the relative contribution of low-density and asymptomatic parasitemias to the overall Plasmodium vivax burden over a period of declining transmission and discuss potential hurdles for malaria elimination in Remansinho and similar settings.

Methods: Eight community-wide cross-sectional surveys, involving 584 subjects, were carried out in Remansinho over 3 years and complemented by active and passive surveillance of febrile illnesses between the surveys. We used quantitative PCR to detect low-density asexual parasitemias and gametocytemias missed by conventional microscopy. Mixed-effects multiple logistic regression models were used to characterize independent risk factors for P. vivax infection and disease.

Principal findings/conclusions: P. vivax prevalence decreased from 23.8% (March-April 2010) to 3.0% (April-May 2013), with no P. falciparum infections diagnosed after March-April 2011. Although migrants from malaria-free areas were at increased risk of malaria, their odds of having P. vivax infection and disease decreased by 2-3% with each year of residence in Amazonia. Several findings indicate that low-density and asymptomatic P. vivax parasitemias may complicate residual malaria elimination in Remansinho: (a) the proportion of subpatent infections (i.e. missed by microscopy) increased from 43.8% to 73.1% as P. vivax transmission declined; (b) most (56.6%) P. vivax infections were asymptomatic and 32.8% of them were both subpatent and asymptomatic; (c) asymptomatic parasite carriers accounted for 54.4% of the total P. vivax biomass in the host population; (d) over 90% subpatent and asymptomatic P. vivax had PCR-detectable gametocytemias; and (e) few (17.0%) asymptomatic and subpatent P. vivax infections that were left untreated progressed to clinical disease over 6 weeks of follow-up and became detectable by routine malaria surveillance.

Figure 1. Location of the field site, Remansinho, southern Amazonas State, Brazilian Amazonia.

The map also shows the village of Nova Califórnia (western Rondônia State), the nearest town, Acrelândia (eastern Acre State), where our field laboratory is situated, and the BR 364 interstate highway, which connects Acre, Rondônia and southern Amazonas to the rest of the country.

Figure 2. Proportion of P.vivax infections diagnosed during 8 consecutive cross-sectional surveys in Remansinho, Brazil, that were symptomatic (black bar segments) and asymptomatic (white bar segments) according to parasite density estimated by quantitative PCR.

The bar widths are proportional to the number of cases within each parasite density class. A total of 129 P. vivax infections were classified according to the presence of symptoms and parasite density.

Figure 3. Venn diagram showing the proportion of P. vivax infections diagnosed by quantitative PCR during 8 consecutive cross-sectional surveys in Remansinho, Brazil, that were asymptomatic, subpatent (i.e., missed by conventional microscopy) and agametocytemic.

The latter group comprises infections with no pvs25 gene transcripts detected by quantitative reverse transcriptase PCR; note that all agametocytemic infections were both asymptomatic and subpatent.

Figure 4. Correlation between length of residence in Amazonia (in years), a proxy of cumulative exposure to malaria, and the probability of having a P. vivax infection (continuous red line) and a clinical vivax malaria episode (continuous black line).

Lines represent median individual probabilities derived from the final (fully adjusted) mixed-effects logistic regression models; the shaded area surrounding the lines represent interquartile ranges.

Figure 5. Kaplan-Meier estimates of the proportion of P. vivax-infected (continuous black line) and uninfected (continuous red line) asymptomatic subjects who remained free of slide-confirmed clinical vivax malaria over the follow-up period.

Dashed lines represent the respective 95% confidence intervals. The small vertical tick-marks indicate the occurrence of a slide positive case of P.vivax, corresponding to the right censoring of the individual survival time. A Cox proportional hazards model revealed no significant difference, between the two groups, in overall risk of vivax malaria episodes, after controlling for potential confounders (hazard ratio = 1.07; 95% CI, 0.52–2.22, P = 0.840).