Validity of Telemetry Measurement of Core Body Temperature in Persons with Spinal Cord Injury

Study Objective Core body temperature (Tcore) in persons with spinal cord injury (SCI) has greater than normal variability and drifts with exposure to low or high ambient temperatures. This inability to thermoregulate causes persons with SCI to be vulnerable to hypo and hyperthermia. Accurate Tcore measurement can assess efficacy of interventions and guide behavioral responses. Our aim was to determine if telemetry measurement of Tcore, using an ingestible capsule (CorTemp), is valid and agrees with rectal thermocouple (Trec) measurement in persons with higher levels of spinal cord injury ( Hi‐SCI ; ≥thoracic level 1), before and after a meal. Methods Nine persons with Hi‐SCI (C4‐T1, AIS A‐C) and 9 matched able‐bodied ( AB ) controls consumed either a protein or carbohydrate meal (in random order on two separate days) while seated in a 25°C room. Participants ingested the CorTemp capsule one‐hour prior to baseline ( BL ). CorTemp and Trec were used to assess Tcore simultaneously every minute during the 30 minutes of BL. Simultaneous assessment of Tcore continued during post‐meal ( PM ) every five minutes for two hours, starting one hour post‐prandial. Pearson correlation coefficient ( r ), one‐sample t‐tests, and Bland‐Altman ( BA ) analysis were used to determine the association of change, mean bias , and limits of agreement ( LoA ) for the values of Tcore between CorTemp and Trec. Results During BL : r = 0.905 and 0.538 for Hi‐SCI and AB, respectively. Mean bias for Hi‐SCI (Trec‐CorTemp) = −0.067°C (95% CI, −0.0849 to −0.0489°C; p<0.0001) with 95% LoA within ±0.36°C (ranged from −0.43 to 0.30°C). Mean bias for AB = −0.0042°C (95% CI, −0.0214 to 0.0131°C; p=0.635) with 95% LoA within ±0.40°C (ranged from −0.40 to 0.40°C). During PM : r = 0.869 and 0.677 for Hi‐SCI and AB, respectively. Mean bias for Hi‐SCI = 0.014°C (95% CI, −0.0047 to 0.0327°C; p=0.141) with 95% LoA within ±0.33°C (ranged from −0.32 to 0.34°C). Mean bias for AB = 0.045°C (95% CI, 0.0241 to 0.0659°C; p<0.0001) with 95% LoA within ±0.43°C (ranged from −0.39 to 0.48°C). Conclusions The r values for Hi‐SCI and AB during both BL and PM demonstrate a strong association of change between CorTemp and Trec measurements of Tcore. Acceptable clinical thresholds for agreement were set as bias <0.1°C and 95% LoA within ±0.4°C as per Byrne et al., 2007. Our mean bias and 95% LoA are within the acceptable clinical thresholds. Therefore, these findings demonstrate that CorTemp is valid and agreeable compared to Trec and is an acceptable method of Tcore measurement in persons with Hi‐SCI, starting as soon as one‐hour post capsule ingestion and one‐hour post prandial. Support or Funding Information Department of Veterans Affairs Rehabilitation Research and Development Service (Award # B2020‐C). Bland‐Altman Plot Comparing CorTemp and Trec Measurements of Tcore in Subjects with SCI during Baseline SD, standard deviation. Bland‐Altman plot showing comparisons of CorTemp and Trec data points (N=364) every minute during the 30 minutes of BL measurement in subjects with SCI. Lines depicting bias and upper and lower limits of agreement are included. Figure 1 Bland‐Altman Plot Comparing CorTemp and Trec Measurements of Tcore in AB Subjects during Baseline SD, standard deviation. Bland‐Altman plot showing comparisons of CorTemp and Trec data points (N=525) every minute during the 30 minutes of BL measurement in AB subjects. Lines depicting bias and upper and lower limits of agreement are included. Figure 2