Assessing rectal temperature with a novel non-invasive sensor

Abstract Athletes, soldiers, and workers who perform intense physical activities under extreme hot conditions might encounter increased physiological thermal strain. Consequently, the increase in body core temperature (Tc) might result in heat exhaustion and heatstroke. Thus, continuously following changes in Tc is of utmost importance. Recently, the Tcore sensor (Drager, Germany), which employs a unique dual-sensor heat flux technology, became commercially available to measure Tc, in a hospital-controlled environment. This study aimed to evaluate the possibility of using the Tcore sensor to accurately monitor rectal temperature (Tre), reflecting Tc, under exercise-heat stress. Thirteen healthy young males completed the study protocol, consisting of 90 min of moderate exercise (walking on a treadmill - 5 km/h, 4% elevation) under controlled hot/dry and hot/wet climatic conditions (30 °C/60% rh, 34 °C/40% rh, and 40 °C/40% rh). Tcore sensors were placed on the forehead and the left wrist. Temperatures from both Tcore sensors were recorded continuously together with Tre using a rectal thermistor. The original algorithm used by the company to estimate Tre from the Tcore sensor was found to be inadequate under the study's conditions and new models for the forehead and the wrist measurements were developed. Nearly 150,000 measurement sets (after filtering) were used to build independent MATLAB software algorithms and test their reliability according to the cross-validation algorithm. Bland-Altman analysis was used to compare between the results obtained by the new models to Tre. The database consisted of a large Tre range (36.5–38.9 °C). The mean errors of the models were close to zero, and the mean absolute errors were 0.20 ± 0.16 °C and 0.27 ± 0.20 °C for the forehead and wrist, respectively. 95% of the measurements from the forehead model and 86% from the wrist model were within ±0.5 °C of T re , and 78% (forehead) and 64% (wrist) were within ±0.3 °C. Root Mean Square Deviation (RMSD) values were 0.29 °C and 0.40 °C for the forehead and wrist models, respectively. The developed models show the feasibility to use the Tcore sensor for assessing Tre under exercise-heat conditions. Furthermore, the sensor was found to be adequate for use on the wrist as well, which might be more practical for use in field conditions.