Ambient temperature and non-accidental mortality: A nationwide space-time-stratified case-crossover study within the...

Research on ambient temperature effects on mortality leveraging cohort data remains scarce. This study estimates the association between ambient temperature and non-accidental mortality and investigates potential effect modification of the temperature-mortality relationship by a wide range of individual and household-level characteristics in a nationwide cohort. We designed a space-time-stratified case-crossover study within the 100 Million Brazilian Cohort, comprising low-income families enrolled in Brazil's Unified Registry for Social Programs (CadÚnico), including all deaths from 2000 to 2018. Daily mean temperatures were obtained from Brazilian Daily Weather Gridded Data based on 1,252 meteorological stations. The outcome was defined as daily counts of non-accidental deaths aggregated at the municipality level. The effect of ambient temperature on non-accidental mortality was estimated using conditional quasi-Poisson regression combined with distributed lag non-linear models. The study population included 3,554,422 deaths. Cold temperatures were associated with a 38 % increase in non-accidental mortality risk (RR = 1.38, 95 % CI: 1.34-1.43) nationally, with high regional variability. In contrast, heat was associated with a 4 % increase in mortality risk nationally, primarily driven by increased risk in the Southeast and Central-West regions. Across population subgroups, cold-related mortality risk was higher among older (≥60 years) adults (RR = 1.47, 95 % CI: 1.42-1.53) and Black individuals (RR = 1.56, 95 % CI: 1.49-1.64). Higher vulnerability to heat-related mortality was found among urban residents (RR = 1.07, 95 % CI: 1.05-1.10). However, we found no consistent evidence of effect modification of cold- and heat-related mortality by sex, income or access to key utilities (water, waste, and electricity). Our findings indicate significant regional differences, highlighting population strata particularly vulnerable to temperature-related mortality. This evidence informs the targeting of interventions for protecting vulnerable populations from extreme temperatures.