Heat Stress and Cardiovascular Hospitalizations in a Semi-Arid Megacity: A Multi-Method Epidemiological and Machine-Learning Analysis in Isfahan, Iran Publisher Pubmed

Summary: Can heat waves strain heart care? A study linked extreme heat to 42% more cardiovascular admissions and longer stays, supporting hospital heat alerts. #ClimateHealth #Cardiology

Javari P ; Javari M
Source: BMC Public Health Published:2026

Abstract

Climate change is intensifying heat exposure worldwide and is expected to increase the burden of cardiovascular diseases (CVD), particularly in semi-arid urban environments where temperature extremes are common; however, individual-level evidence from hot and dry settings remains limited. This study examined associations between multiple thermal stress indices and CVD hospitalizations (n = 2,760), identified predictors of prolonged length of stay (LOS > 7 days) and in-hospital mortality, and explored projected risks under ongoing heat trends in Isfahan, Iran, from 2019 to 2024. Daily meteorological data from ERA5-Land reanalysis and six local weather stations were linked with de-identified hospital admission records. Thermal exposure was assessed using the Heat Index (HI), Humidex, Wet-Bulb Globe Temperature (WBGT), Daily Temperature Range (DTR), and Warm Spell Duration Index (WSDI). Quasi-Poisson regression models were used to estimate associations between heat exposure and CVD admissions, adjusting for age, sex, socioeconomic status, fine particulate matter (PM₂.₅), day of the week, and long-term trends. Logistic regression, Extreme Gradient Boosting (XGBoost), and Random Forest models were applied to predict prolonged LOS, with model performance evaluated using five-fold cross-validation and SHAP-based feature interpretation. Each one–standard deviation increase in same-day HI (≈ 4.3 °C) was associated with a 12% increase in CVD admissions (IRR = 1.12; 95% CI: 1.05–1.19), while warm spells lasting five or more consecutive days increased admission risk by 42% (IRR = 1.42; 95% CI: 1.21–1.67). Extreme heat exposure was also associated with prolonged hospitalization (OR = 1.85; 95% CI: 1.21–2.83). Among the machine learning models, XGBoost showed strong discriminatory performance (AUC = 0.84; 95% CI: 0.81–0.87), with HI, WBGT, and age identified as the most influential predictors. Overall, multiple thermal stress indices independently predicted higher CVD admission rates and longer hospital stays in this semi-arid urban setting, supporting the integration of climate-based early warning systems into hospital preparedness strategies to reduce heat-related cardiovascular risks. © The Author(s) 2026.