Accuracy of Zero-Heat-Flux Cutaneous Temperature in Intensive Care Adults

Objectives: To compare accuracy of a continuous noninvasive cutaneous temperature using zero-heat-flux method to esophageal temperature and arterial temperature. Design: Prospective study. Setting: ICU and NeuroICU, University Hospital. Patients: Fifty-two ICU patients over a 4-month period who required continuous temperature monitoring were included in the study, after informed consent. Interventions: All patients had esophageal temperature probe and a noninvasive cutaneous device to monitor their core temperature continuously. In seven patients who required cardiac output monitoring, continuous iliac arterial temperature was collected. Simultaneous core temperatures were recorded from 1 to 5 days. Comparison to the esophageal temperature, considered as the reference in this study, used the Bland and Altman method with adjustment for multiple measurements per patient. Measurements and Main Results: The esophageal temperature ranged from 33 degrees C to 39.7 degrees C, 61,298 pairs of temperature using zero-heat-flux and esophageal temperature were collected and 1,850 triple of temperature using zero-heat-flux, esophageal temperature, and arterial temperature. Bias and limits of agreement for temperature using zero-heat-flux were 0.19 degrees C +/- 0.53 degrees C compared with esophageal temperature with an absolute difference of temperature pairs equal to or lower than 0.5 degrees C of 92.6% (95% CI, 91.9-93.4%) of cases and equal to or lower than 1 degrees C for 99.9% (95% CI, 99.7-100.0%) of cases. Compared with arterial temperature, bias and limits of agreement were -0.00 degrees C +/- 0.36 degrees C with an absolute difference of temperature pairs equal to or lower than 0.5 degrees C of 99.8% (95% CI, 95.3-100%) of cases. All absolute difference of temperature pairs between temperature using zero-heat-flux and arterial temperature and between arterial temperature and esophageal temperature were equal to or lower than 1 degrees C. No local or systemic serious complication was observed. Conclusions: These results suggest a comparable reliability of the cutaneous sensor using the zero-heat-flux method compared with esophageal or iliac arterial temperatures measurements.