The joint impact of temperature, humidity, and air pollution on COVID-19 incidence: a multi-country time-series study in 439 cities

Wagatsuma K.; Feurer D.; Yu W.; Xu R.; Riffe T.; Kniffka M.S.; Acosta E.; Armstrong B.; Mistry M.; Lowe R.; Royé D.; Hashizume M.; Tobias A.; Vicedo-Cabrera A.M.; Madaniyazi L.; Sheng Ng C.F.; Íñiguez C.; Ragettli M.S.; Lavigne E.; Correa P.M.; Ortega N.V.; Kyselý J.; Urban A.; Orru H.; Indermitte E.; Maasikmets M.; Breitner-Busch S.; Schneider A.; Honda Y.; Alahmad B.; Zanobetti A.; Schwartz J.; Carrasco G.; Holobâca I.H.; Kim H.; Lee W.; Bell M.L.; Scovronick N.; Acquaotta F.; de Sousa Zanotti Stagliorio Coelho M.; Diaz M.H.; Félix Arellano E.E.; Michelozzi P.; Stafoggia M.; de’Donato F.; Rao S.; Seposo X.; Tong S.; Klompmaker J.; Guo Y.; Masselot P.; Gasparrini A.; Sera F. · Environment International (2026) · DOI: 10.1016/j.envint.2026.110090

Resumen

Several studies have explored the short-term effects of environmental stressors on coronavirus disease 2019 (COVID-19) transmission and severity. However, evidence on the interactive effects of meteorological conditions and air pollution remains limited and geographically variable. We therefore aimed to quantify the independent and interactive effects of short-term exposure to humidex, a composite index of temperature and relative humidity, and fine particulate matter ≤ 2.5 μm (PM2.5) on daily COVID-19 incidence across multiple cities and in multiple countries. Daily time-series data on confirmed COVID-19 cases, meteorological factors, and PM2.5 concentrations were collected from 439 cities in 22 countries during January 2020–August 2022 as part of the Multi-Country Multi-City Collaborative Research Network. A two-stage design was applied: first, city-specific quasi-Poisson models with distributed lag non-linear models estimated exposure–response associations; second, multilevel random-effects meta -analyses pooled city-specific estimates. Effect modification by PM2.5 was assessed using a product term between non-linear humidex function and linear PM2.5 function. Approximately 95.1 million confirmed COVID-19 cases were analyzed. Lower humidex values (0.1 °C versus 15.1 °C) were associated with increased daily cases (relative risk [RR]: 1.1192, 95% confidence interval [CI]: 1.0214–1.2262). A 10 μg/m3 increase in PM2.5 over the current and preceding 2 days was associated with a modest increase in daily cases (RR: 1.0079, 95% CI: 1.0001–1.0161). No statistically significant interaction between humidex and PM2.5 was observed. Short-term exposure to cold–dry conditions and elevated PM2.5 independently increased COVID-19 incidence, highlighting the need to consider both thermal environment and air quality when designing climate-resilient public health responses. These findings enhance understanding of how climate-related environmental stressors influence COVID-19 transmission.

Instituciones: Universidad Peruana Cayetano Heredia