Risk for exertional heat illness among US army enlistees: climate indexes, intrinsic factors and their interactions

Objectives To characterise intrinsic and extrinsic (climatic) risks for mild and severe exertional heat illness (EHI) among first-year army enlistees. Methods We examined 337 786 soldiers who enlisted between 2012 and 2019. Survival models were used to predict incident EHI from intrinsic factors (demographics, healthcare utilisation, chronic conditions, body mass index (BMI), Army Physical Fitness Test (APFT), upper/lower respiratory tract infections (URTI and LRTI), skin and soft-tissue infections (SSTI), extrinsic factors (geographical region, daily mean Universal Thermal Climate Index (UTCI), wet bulb globe temperature (WBGT)) and interactions. Results There were 1390 cases of mild and 359 cases of severe EHI. Females had a higher risk for mild (adjusted OR (aOR) 1.78; 95% CI 1.57 to 2.02) but a lower risk for severe (aOR 0.61; 95% CI 0.38 to 0.87) EHI. Obesity was associated with severe EHI (aOR: 1.76; 95% CI 1.09 to 2.84) but not mild EHI (aOR: 1.03; 95% CI 0.76 to 1.39). URTI was associated with severe (aOR: 2.44; 95% CI 1.12 to 5.30) and mild (aOR 3.72, 95% CI 2.84 to 4.87) EHI, as were LRTI (severe, aOR: 11.40; 95% CI 6.09 to 21.32; mild, aOR 2.06; 95% CI 1.22 to 3.46), but not SSTI. UTCI outperformed WBGT in predicting EHI. Outside the Southern USA, EHI risk was elevated at lower UTCI. Associations varied over climate conditions and generally did not increase with climatic heat stress. Conclusions Respiratory infections were associated with the highest risk for EHI in soldiers. Risk mitigation strategies may include monitoring prevention and recovery from respiratory infections. Female sex and obesity may have different associated risks over climate conditions. No data are available. Data are derived from official population military records and sources, and were provided to us under strict data use agreement guidelines. Therefore, we cannot reasonably share the data with outside parties. We are willing to share our code to other researchers with access to similar data.