Abstract
Background: Although substantial evidence has shown that precipitation is associated with diarrhea in low- and middle-income countries, few multicountry studies have focused on middle- and high-income countries to comprehensively explore the exposure-response curves between precipitation and diarrhea across different countries/regions and climate zones (tropical, arid, temperate, and cold).
Methods: We collected weekly diarrhea morbidity data (primarily from hospital visits, hospital admissions, and surveillance systems) and corresponding exposure data (total precipitation and mean temperature) from 904 locations in 39 middle- and high-income countries/regions. To assess location-specific precipitation-diarrhea associations, we fitted time-stratified case-crossover analyses in distributed lag nonlinear models with 8 weeks of lag. Then, we pooled the estimates at the country/region level and by climate zone with multivariate meta-regression analysis.
Results: A total of 160,266,691 diarrhea cases were included in the study. The exposure-response association curves between total precipitation and diarrhea varied substantially at the country/region level. For example, U-shaped, J-shaped, or V-shaped curves were observed in Taiwan, Hungary, Germany, Denmark, Mexico, Brazil, and Ecuador. We identified a generally increased trend of diarrhea morbidity risk in the Philippines, Czech Republic, Norway, and USA when total precipitation exceeded the minimum risk level, defined at the first percentile of total precipitation. In China, Japan, Bangladesh, and Vietnam, diarrhea morbidity risk tended to decrease with increasing total precipitation. The country-specific relative risks ranged from 1.02 (95% confidence interval [CI]: 0.88, 1.19) in Finland to 2.67 (95% CI: 2.28, 3.13) in Japan for low levels of total precipitation and from 1.02 (95% CI: 0.93, 1.11) in Ecuador to 2.03 (95% CI: 1.49, 2.76) in Austria for high levels of total precipitation. Overall, associations between precipitation and diarrhea morbidity were stronger in arid regions and weaker in cold regions.
Conclusions: Associations between precipitation and diarrhea morbidity differed by country/region and climate zone. This study provides valuable evidence showing that location-specific prevention measures should be considered to mitigate the diarrhea burden associated with precipitation.