Prospective assessment of malaria infection in a semi-isolated Amazonian indigenous Yanomami community: Transmission heterogeneity and predominance of submicroscopic infection

Abstract

In the Amazon basin, indigenous forest-dwelling communities typically suffer from a high burden of infectious diseases, including malaria. Difficulties in accessing these isolated ethnic groups, such as the semi-nomadic Yanomami, make official malaria data largely underestimated. In the current study, we longitudinally surveyed microscopic and submicroscopic malaria infection in four Yanomami villages of the Marari community in the northern-most region of the Brazilian Amazon. Malaria parasite species-specific PCR-based detection of ribosomal and non-ribosomal targets showed that approximately 75% to 80% of all malaria infections were submicroscopic, with the ratio of submicroscopic to microscopic infection remaining stable over the 4-month follow-up period. Although the prevalence of malaria infection fluctuated over time, microscopically-detectable parasitemia was only found in children and adolescents, presumably reflecting their higher susceptibility to malaria infection. As well as temporal variation, the prevalence of malaria infection differed significantly between villages (from 1% to 19%), demonstrating a marked heterogeneity at micro-scales. Over the study period, Plasmodium vivax was the most commonly detected malaria parasite species, followed by P. malariae, and much less frequently P. falciparum. Consecutive blood samples from 859 out of the 981 studied Yanomami showed that malaria parasites were detected in only 8% of the previously malaria-positive individuals, with most of them young children (median age 3 yrs). Overall, our results show that molecular tools are more sensitive for the identification of malaria infection among the Yanomami, which is characterized by heterogeneous transmission, a predominance of low-density infections, circulation of multiple malaria parasite species, and a higher susceptibility in young children. Our findings are important for the design and implementation of the new strategic interventions that will be required for the elimination of malaria from isolated indigenous populations in Latin America.

Fig 1. Map of the Brazilian-Venezuelan border showing the study area.

The Marari community (yellow in the top inset map) is located in the Yanomami indigenous area (green in the inset maps) in the northern-most region of the state of Amazonas (red in the bottom inset map), roughly 454 miles from the state capital, Manaus, in Brazil. The studied villages are marked with yellow dots, while the main rivers are shown in blue. Studied villages shown in the main map are located in the municipality of Barcelos (light green) or Santa Isabel do Rio Negro (dark green). Map sources: the Brazilian Water National Agency, Multi-scale Ottocoded Hydrographic Base, and the Brazilian Institute of Geography and Statistics (IBGE). Amazonas municipal cartographic base (SRC selected- EPSG: 4674, SIRGAS 2000). Scale 1:128.000.

Fig 2. Monthly-time series of microscopically-confirmed malaria cases in the Marari community of the Yanomami indigenous area (Amazonas, Brazil) during the study period, 2014–2015.

The study design included three cross-sectional surveys: the first survey (baseline) was carried out in September 2014, the second in November 2014 (2 months later) and the third in January 2015 (4 months after the first baseline survey). Data are presented as microscopically-confirmed cases per 1,000 inhabitants, and were provided by the National Malaria Surveillance System Registry (SIVEP-Malaria) from the Brazilian Ministry of Health. Malaria infection caused by P. vivax (light blue) and P. falciparum (dark blue) are plotted by month.

Fig 3. Odds ratios (ORs) for the relative risk of malaria infection in each village compared with Castanha/Ahima (the village with the lowest number of malaria infections, used as the reference).

At first baseline survey, the inhabitants of Taibrapa were temporarily located in their alternative residence (Taibrapa-II). Crude OR were obtained by using 2x2 contingency tables with a 95% confidence interval (CI). P < 0.05 was considered significant. Malaria infection was confirmed by either conventional light microscopy (microscopically-confirmed cases) or species-specific PCR-based assays (submicroscopic or low-density infections).

Fig 4. Prevalence of microscopic and submicroscopic malaria infection stratified by age at baseline survey.

Submicroscopic infections were determined by PCR-based protocols (PCR) and microscopic infections by conventional Giemsa-stained thick blood smears (Microscopy).

Fig 5. Radar charts showing the malaria prevalence by village over the 4-month study period.

Three cross-sectional surveys were carried as described in the legend of Fig 2 (baseline, and then 2 and 4 months later). On each radar chart the data are shown as the percentage of individuals positive by either conventional microscopy or PCR-based methods. The 4 Yanomami villages are represented as Taibrapa (T-I or T-II, according to the alternative residences used by the Yanomami during the study), Gasolina (G), Alapusi (Al) and Castanha/Ahima (C/Ah). The overall percentage of positive individuals and their 95% confidence intervals for each survey are shown underneath their respective radar charts, with different letters (a-c) indicating statistically significant differences between the surveys (Fisher’s exact test, P < 0.05).

Fig 6. Comparison of the frequency of species-specific positivity (P. vivax, P. malariae, P. falciparum and mixed infections) during the study period (baseline, 2 and 4 months later).

Fig 7. Species-specific malaria positivity stratified by age during the 4 month study period (baseline, 2 and 4 months later).

Infections caused by P. vivax, P. falciparum and P. malariae were determined by microscopy and/or PCR-based protocols, as described in the methods.