J Expo Sci Environ Epidemiol. 2026 Apr 8. doi: 10.1038/s41370-026-00890-7. Online ahead of print.
ABSTRACT
BACKGROUND: Extreme rainfall is expected to intensify with climate change, yet short-term health effects can plausibly be harmful (flooding, contamination, access disruption) or protective (pollution washout, cooling). Robust, comparable evidence at national scale is scarce for Brazil.
OBJECTIVE: We analyzed daily mortality for all-cause, cardiovascular, respiratory, and infectious outcomes across all Brazilian municipalities from 2003-2023.
METHODS: Rainfall extremes were defined hydrologically using regional frequency analysis. In the first stage, we fit state-level conditional quasi-Poisson models with distributed-lag nonlinear terms for rainfall (exposure) and temperature (confounder), cumulating effects over lags 0-14. In the second stage, we pooled reduced exposure-response coefficients using multivariate random-effects meta-analysis to obtain national curves.
RESULTS: We found substantial between-state heterogeneity. For all outcomes, some states showed positive associations on extreme-rain days, while others showed protective or null effects. Protective associations were common in parts of the South/Southeast (consistent with pollution washout/cooling), whereas in other states the results suggest positive associations, compatible with flooding, contamination, or access disruption. In contrast, the national pooled curves were not statistically significant across all the range of exposure for all-cause, cardiovascular, respiratory, or infectious mortality.
SIGNIFICANCE: Short-term mortality responses to extreme rainfall in Brazil are context-dependent, varying with climate regime, sanitation and drainage, emission profiles, vector ecology, housing, and health-system robustness. A single national effect obscures local vulnerability. Adaptation should be state-tailored-combining drainage/sanitation upgrades, vector control, continuity of emergency care, and maintenance of air-quality gains during wet periods.
IMPACT: Our findings suggest state-level heterogeneity. National pooled effects are null, but several states show elevated risk. Findings guide subnational adaptation in drainage, health, and air quality.
PMID:41951749 | DOI:10.1038/s41370-026-00890-7