Infrared cameras overestimate skin temperature during rewarming from cold exposure

The primary aim of this study was to assess the accuracy of an infrared camera and that of a skin thermistor, both commercially available. The study aimed to assess the agreement over a wide range of skin temperatures following cold exposure.

Fifty-two males placed their right hand in a thin plastic bag and immersed it in 8 °C water for 30 min whilst seated in an air temperature of 30 °C. Following hand immersion, participants removed the bag and rested their hand at heart level for 10 min. Index finger skin temperature (T_sk) was measured with a thermistor, affixed to the finger pad, and an infrared camera measured 1 cm distally to the thermistor. Agreement between the infrared camera and thermistor was assessed by mean difference (infrared camera minus thermistor) and 95% limits of agreement analysis, accounting for the repeated measures over time. The clinically significant threshold for T_sk differences was set at ±0.5 °C and limits of agreement ±1 °C.

As an average across all time points, the infrared camera recorded T_sk 1.80 (SD 1.16) °C warmer than the thermistor, with 95% limits of agreement ranging from -0.46 °C to 4.07 °C.

Collectively, the results show the infrared camera overestimated T_sk at every time point following local cooling. Further, measurement of finger T_sk from the infrared camera consistently fell outside the acceptable level of agreement (i.e. mean difference exceeding ±0.5 °C). Considering these results, infrared cameras may overestimate peripheral T_sk following cold exposure and clinicians and practitioners should, therefore, adjust their risk/withdrawal criteria accordingly.