Population-level spatial patterns of body-surface temperature in older adults: a multicenter infrared thermography study across diverse environmental settings

Background Body-surface temperature reflects peripheral physiological regulation and may provide a useful phenotype for population studies of ageing. Infrared thermography (IRT) enables non-invasive assessment of temperature distribution, but large multicenter evidence remains limited. We aimed to characterise body-surface temperature patterns in older adults and examine their associations with individual and contextual factors. Methods In this multicenter observational study, 5,225 adults aged ≥60 years underwent standardized IRT across multiple regions in China. Temperature was measured at 12 predefined trunk regions. Both absolute regional temperature and plane-specific relative temperature (ΔT) were analyzed. Multivariable regression models adjusted for demographic, anthropometric, physiological, and center-related variables were used to assess constitution-by-region interactions, followed by false discovery rate-adjusted region-specific analysis. Sensitivity analysis included absolute-temperature models and leave-one-site-out validation. Results A total of 62,700 region-level observations were analyzed. Regional setting was the largest contributor to thermal variability, followed by body composition, whereas constitution explained a relatively small proportion of overall variance (2.0–10.9%). Significant constitution-by-region interactions indicated spatially heterogeneous temperature patterns (χ²=732.41, df=88, p<0.001). Constitution-related differences were localized mainly to abdominal, posterior lumbar, and selected thoracic regions, with modest effect sizes (0.08–0.27 °C) that remained directionally consistent across sensitivity analysis. Conclusion Body-surface temperature in older adults exhibits modest but spatially organized variation influenced predominantly by regional setting and body composition. Constitution-related differences are localized and reproducible but should be interpreted as exploratory surface-temperature phenotypes rather than clinically actionable findings.