Objective: The purpose of this clinical study was to determine the accuracy of infrared tympanic membrane thermometry compared to axillary temperature (tempAx) for detecting body temperature reliably in critically ill patients in the daily practice.
Materials and methods: Fifty adult patients admitted to a medical-surgical intensive care unit of an acute-care teaching hospital in Seville, Spain, during a 2-month period underwent prospective and simultaneous measurements of body temperature using a mercury-in-glass thermometer placed at the axilla of the dominant arm for at least 3 minutes and an infrared thermometer (ThermoScan 07, Braun Corporation, Kronberg, Germany) in both ears.
Results: A total of 429 simultaneous measurements of axillary temperature (tempAx) and tympanic temperature (tempTT) were made. The mean +/- SD tempAx was 36.90 degrees C +/- 1.06 degrees C and the mean tempTT was 36.94 degrees C +/- 0.97 degrees C ( P = NS; 95% CI, -0.18 to 0.10), with a difference between tempTT and tempAx means of 0.04 degrees C. There was a statistically significant correlation between tempAx and tempTT ( r = 0.813, P < .0005). When 20 extreme readings of both methods (<34.2 degrees C and >39.8 degrees C) were excluded, the mean tempAx was 36.91 degrees C +/- 0.86 degrees C and the mean tempTT was 36.9 degrees C +/- 0.89 degrees C ( P = NS; 95% CI, -0.05 to 0.06), with a difference of 0.01 degrees C and a statistically significant correlation between both measurements ( r = 0.80, P < .0005). The sensitivity and specificity of tempTT for different thresholds were 74% and 85% for 37 degrees C, 70% and 95% for 38 degrees C, and 25% and 99.8% for 39 degrees C, respectively. The negative predictive value for 39 degrees C was 99%.
Conclusions: In adult intensive care unit patients, the infrared tympanic thermometer (ThermoScan 07) produced highly reliable measurements when compared to tempAx measured using a conventional mercury-in-glass thermometer. Both methods correlated positively and significantly.