Abstract Background and aims Stroke subtypes exhibit pathophysiological mechanisms that may confer differential vulnerability to environmental stressors. While temperature associations with overall stroke incidence are well-established in temperate regions, evidence from tropical-subtropical settings remains scarce. We investigated whether ischemic stroke (IS), intracerebral hemorrhage (ICH), and subarachnoid hemorrhage (SAH) demonstrate heterogeneous temperature-mortality relationships in Brazil. Methods We conducted a time-series analysis linking monthly nationwide stroke mortality data (ICD-10: I60, I61-I62, I63) from Brazil's Unified Health System with meteorological data from the National Institute of Meteorology (2020-2024). Pearson correlations between minimum temperature (TMIN) and subtype-specific mortality were stratified by season (Southern Hemisphere). Mann-Whitney U tests compared winter versus summer mortality rates. Results 1,720 stroke deaths (194 IS, 975 ICH, 551 SAH), overall temperature-mortality correlations differed by subtype: SAH (r=-0.255, p=0.049), IS (r=-0.164, p=0.209), and ICH (r=+0.101, p=0.444). Seasonal stratification revealed heterogeneity: IS demonstrated strong inverse correlation during winter (r=-0.663, p=0.007); ICH showed inverse correlation during spring (r=-0.584, p=0.022); SAH exhibited consistent inverse trends across autumn (r=-0.437) and winter (r=-0.395), though not reaching statistical significance. Winter-to-summer mortality comparisons showed opposing patterns: SAH increased 12.5% (8.5→9.6 deaths/month), IS increased 5.7%, while ICH decreased 5.0% during colder months. Conclusions Stroke subtypes demonstrate divergent temperature-mortality associations in tropical-subtropical Brazil, with SAH and IS showing cold-season vulnerability while ICH exhibits paradoxical warm-season peaks. These findings suggest subtype-specific physiological responses to thermal stress and the importance of disaggregated analyses in climate-stroke research for tailored prevention strategies in climatically diverse regions. Conflict of interest None of the authors: nothing to disclose Figure 1 - belongs to Results
Rodrigues et al. (Fri,) studied this question.