Agreement between mean weighted skin temperature formulas during exercise

There are many formulas to calculate mean weighted skin temperature (T_mws) with little evidence for their agreement with each other. This investigation assessed the agreement of five common T_mws formulas (≤7 sites) with a 12-site formula (Hardy-DuBois 12-site, HDB12). It was hypothesized that a formula with more measurement sites (Hardy-Dubois 7-site, HDB7) would have more agreement during steady-state exercise in moderate (20°C) and warm (30°C) conditions. Ten well-trained male participants (age: 37 ± 12 yr, height: 178 ± 7 cm, mass: 76 ± 9 kg, body fat: 8 ± 2%, V̇o_2peak: 66 ± 10 mL/kg/min) cycled at 55% of peak oxygen uptake for 60 min. The T_mws were calculated for HDB12, HDB7, Palmes/Park 6-site (PP6), Teichner 6-site (TEI6), Ramanathan 4-site (RAM4), and Burton 3-site (BTN3) formulas. Good agreement was determined when the mean bias, mean absolute error (MAE), or root mean squared error (RMSE) was ≤0.5°C or the limits of agreement (LOA) were ≤1.0°C. The T_mws formulas were in better agreement with HDB12 in warm compared with moderate conditions. The best performing formulas were RAM4, BRT3, and TEI6; mean bias was low in RAM4 (moderate: 0.2°C; warm: 0.1°C) and BRT3 (moderate: 0.1°C; warm: 0.0°C) and RMSE was low in TEI6 (moderate: 0.6°C; warm: 0.4°C) compared with the others. The large deviation in torso skin temperature for HDB7 (moderate: -2.0°C; warm: -1.6°C) resulted in consistent underestimation of HDB12 (mean bias, moderate: -0.8°C; warm: -0.5°C). Due to its popularity, the four-site RAM4 is acceptable for cycling in 20°C and 30°C conditions, when a 12-site formula is not feasible.NEW & NOTEWORTHY This investigation is the first to show T_mws formula agreement during exercise in moderate (20°C) and warm (30°C) environments. This study uniquely contributes to the present knowledge by showing better agreement between all formulas and the criterion in the 30°C environment than in the 20°C environment during exercise. This study supports the prevalent use of the Ramanathan formula currently in research.