Migration Typologies Predict Malaria Incidence in the Peruvian Amazon: A Prospective Cohort Study

Background: Plasmodium vivax and Plasmodium falciparum contribute a high burden of disease in the Peruvian Amazon. Migration is an important risk factor for malaria, as travel along the river systems accounts for the vast majority of movement within this region. This travel creates networks of transmission that allows individuals who travel to contract malaria as well as generating transmission across communities by repeatedly reintroducing the parasite into different areas, which can complicate control efforts.

Methods: Surveyors visited homes of participants three times a week to carry out active surveillance between July 2006 and October 2009. For participants returning from travel, data were collected on destination and duration away, which was used to define a typology for migrant populations. Incidence rates of malaria and migration typology characteristics were assessed through Negative Binomial regression models and migration rates with hydrological factors by logistic regression models.

Results: A cohort of 2,202 participants from 402 families were followed for three years and categorized by the average number of annual trips taken. High-frequency and low-frequency migrant populations reported 9.7 (IRR 7.59 (95% CI:.381, 13.160)) and 4.1 (IRR 2.89 (95% CI: 1.636, 5.099)) times more P. vivax cases than those considered non-migrants and 30.7 (IRR 32.42, (95% CI: 7.977, 131.765)) and 7.4 (IRR 7.44 (95% CI: 1.783, 31.066)) times more P. falciparum cases, respectively. Among high frequency migrants, males were at a 1.20 (95% CI: 0.872, 1.641) increased risk compared to female high frequency migrants. Working in agriculture, logging, and fishing were at 1.02, 1.36, 1.93 increased risk for P. vivax and 2.46, 2.35, 2.03 times increased risk of P. falciparum compared to those in non-migrant positions. However, the greatest occupation related risk was attributed to high frequency migrants employed in manual labor within their community. This group was at 2.45 (95% CI: 1.113, 5.416) times higher risk than non-employed low frequency migrants. Travel to lakes, rivers, and streams were 2.36, 1.81, and 1.95 times as many P. vivax and 1.36, 1.34, and 2.08 P. falciparum cases, respectively, compared to travel to communities. More migration episodes were recorded in the dry season, however, the seven day rolling mean streamflow was positively correlated with migration events (OR 1.25 (95% CI: 1.138, 1.368)).

Conclusions: Migration is a key risk factor for malaria; however, it alone does not characterize the distribution on risk within the migrant community. This study gives information about particular population groups to target control efforts.