Activity-Dependent Differences Between Wearable Heat-Flux-Based Core Temperature Estimates and Gastrointestinal Temperature Measured by an Ingestible Capsule Under Free-Living Conditions

Abstract

Wearable systems that estimate core body temperature may facilitate long-term physiological monitoring under free-living conditions; however, the extent to which differences from gastrointestinal temperature vary across daily-life contexts remains unclear. This exploratory study examined activity-dependent differences between wearable heat-flux based core temperature estimates and gastrointestinal temperature measured using an ingestible capsule as a practical comparator. One-minute physiological data were synchronized with activity logs recorded at 30-min intervals. The primary outcome was the absolute temperature difference between the two measurement systems, and the signed temperature difference was evaluated descriptively. Nine participants were enrolled, and eight healthy young adults were included in the final analysis. A total of 12,528 valid paired observations were analyzed. Mean absolute difference was lowest during sleep (<inline-formula> <tex-math notation="LaTeX">$0.262~^{\circ }$ </tex-math></inline-formula>C) and highest during outdoor activity away from home (<inline-formula> <tex-math notation="LaTeX">$0.637~^{\circ }$ </tex-math></inline-formula>C). Absolute differences <inline-formula> <tex-math notation="LaTeX">$\ge 0.5~^{\circ }$ </tex-math></inline-formula>C occurred in 9.1% of sleep observations and 56.5% of outdoor-activity observations. Absolute difference increased with physical activity level (<inline-formula> <tex-math notation="LaTeX">$\rho =0.142$ </tex-math></inline-formula>, p < 0.001), whereas no statistically significant association was detected with melanopic equivalent daylight illuminance (melanopic EDI; <inline-formula> <tex-math notation="LaTeX">$\rho =0.016$ </tex-math></inline-formula>, p = 0.068). In the mixed-effects model, outdoor activity away from home (<inline-formula> <tex-math notation="LaTeX">$\beta =0.365$ </tex-math></inline-formula>, p < 0.001), indoor activity away from home (<inline-formula> <tex-math notation="LaTeX">$\beta =0.116$ </tex-math></inline-formula>, p < 0.001), and physical activity (<inline-formula> <tex-math notation="LaTeX">$\beta =0.022$ </tex-math></inline-formula>, p < 0.001) were associated with larger absolute differences. Exercise also yielded a large coefficient, but this estimate was based on only 50 observations and should be interpreted as exploratory. These findings suggest that discrepancies between the two measurement systems vary across daily-life contexts. The results do not establish absolute accuracy or interchangeability with a gold-standard measure of true core body temperature.