Associations between weather and Plasmodium vivax malaria in an elimination setting in Peru: a distributed lag analysis
Gabriella R Barratt Heitmann,Xue Wu,Anna T Nguyen,Astrid Altamirano-Quiroz,Sydney Fine,Bryan Fernandez-Camacho,Antony Barja,Renato Cava,Veronica Soto-Calle,Hugo Rodriguez,Gabriel Carrasco Escobar,Adam Bennett,Alejandro Llanos-Cuentas,Erin A Mordecai,Michelle S Hsiang,JADE R BENJAMIN-CHUNG
DOI: https://doi.org/10.1101/2024.11.26.24318000
2024-11-28
Abstract:Background
Plasmodium vivax (Pv) is the predominant malaria species in countries approaching elimination. In the context of climate change, understanding environmental drivers of transmission can guide interventions, yet evidence is limited, particularly in Latin America.
Objectives
We estimated the association between temperature and precipitation and Pv malaria incidence in a malaria elimination setting in Peru.
Methods
We analyzed malaria incidence data from 2021-2023 from 30 communities in Loreto, Peru with hourly weather data from the ERA5 dataset and land cover data from MapBiomas. Predictors included average weekly minimum and maximum temperature, high heat (>90th percentile mean temperature), total weekly precipitation, and heavy rain (>90th percentile total precipitation). We fit non-linear distributed lag models for continuous weather predictors and generalized additive models for binary predictors and the lookback period was 2-16 weeks. Temperature models adjusted for total precipitation; precipitation models adjusted for maximum temperature. We performed subgroup analyses by season, community type, and distance to forest edge.
Results
The median vs. lowest values of weekly average minimum temperature was associated with 2.16 to 3.93-fold higher incidence 3-16 weeks later (5-week lag incidence ratio (IR) =3.93 [95% CI 2.18, 7.09]); for maximum temperature, the association was hump-shaped across lags, with protective associations at 1-2 and 15-16 week lags and 1.07-1.66-fold higher incidence at 6-13 week lags. High heat (>27.5C) was associated with 1.23 to 1.37-fold higher incidence at 5-9 week lags (9-week lag IR = 1.25 [1.02, 1.53]). Associations between total precipitation and malaria incidence were hump-shaped across lags, with the strongest positive association at 750 mm of precipitation at a 9-week lag (IR=1.56; [1.27, 1.65]). Heavy rain (>186mm) was associated with 1.22-1.60-fold higher incidence at 2-10 week lags (9-week lag IR=1.23 [1.02, 1.49]).
Discussion
Higher temperatures and precipitation were generally associated with higher malaria incidence over 1-4 months.
Epidemiology