Noninvasive Measurement of Body Temperature in Critically Ill Patients

Q What is the most accurate means of measuring body temperature noninvasively in an adult critical care patient, assuming the patient does not have a pulmonary artery catheter or a temperature-sensing bladder, rectal, or esophageal probe?A Elizabeth Bridges, RN, PhD, CCNS, and Karen Thomas, RN, PhD, reply:When evaluating the accuracy and precision of any temperature measurement method, the method should be compared against a "gold standard." In intensive care patients, the pulmonary artery (PA) temperature, which reflects the core temperature, is considered the standard for comparison. Table 1 summarizes studies evaluating the agreement of various non-invasive temperature measurement methods (oral, ear-based, temporal artery, and axillary) with the PA temperature. In previous research,7,9 a thermometer was considered accurate if the mean difference from the PA temperature was ±0.3°C and precise with a standard deviation ranging from 0.3°C to 0.5°C. As demonstrated in Table 1, the oral, ear-based, and temporal artery measurements are generally equivalent with regard to accuracy, whereas the axillary temperature is an underestimate of the PA temperature. However, the precision varies across methods (oral, SD=0.24–0.6°C; ear-based, SD=0.4–0.57°C; temporal artery, SD=0.5–1.1°C; and axillary, SD=0.16–0.6°C).The difficulty in comparing 2 alternative temperature measurement methods (eg, oral, temporal artery, ear-based) is that each measurement has its own error. For example, in a recent review on the evaluation of new fever in critically ill adults, O'Grady et al17 stated that after invasive methods (PA, esophageal, or bladder), the following methods should be used in this order: rectal, oral, and tympanic. Axillary, temporal artery, and chemical dot thermometers are not recommended. In a subsequent series of letters to the editor,18 the author stated that temporal artery measurements were not recommended because Lawson et al9 found that 20% of the temporal artery temperature measurements were greater than ±0.5°C different from the concurrent PA temperature. However, as summarized in Table 1, the bias and precision of the oral and temporal artery methods were similar, and 19% of the oral measurements were also greater than ±0.5°C different from the concurrent PA temperature, suggesting that the 2 methods are comparable.Similarly, Fetzer and Lawrence19 recently compared ear-based and temporal artery temperature measurements and reported that the bias between the 2 methods was -0.4±0.64°C (95% CI, −1.29 to 1.21), which is less accurate and precise than either method compared with PA temperature measurement (Table 1). Unlike the studies outlined in Table 1, the difference between the ear-based and temporal artery methods reflects the error in both measurements, and we cannot say that one method is more accurate than the other, only that they are not interchangeable.Limited research has addressed whether the thermometer correctly identifies patients with hyperthermia or hypothermia. In a study20 in 13 febrile patients, the oral thermometer was most accurately for detecting fever (>38.3°C), whereas the ear-based measurement had the lowest chance of a false-negative reading. In 2 additional studies,15,21 researchers evaluated the utility of ear-based temperature measurements for detecting fever (variably defined as a body temperature >38°C–38.5°C). In these studies, the ear-based thermometers also showed high specificity (0.92–1.00), indicating accuracy in identifying patients without a fever; but lower sensitivity (0.50–0.80) indicating accurate identification of only 50% to 80% of patients with a fever. Limited research has been done on temporal artery temperature measurement in febrile patients. In the study by Lawson et al,9 11 measurements were obtained from 3 patients where the PA temperature was 38°C or greater. The temporal artery measurement allowed correct characterization of the patient as febrile in 10 of 11 cases, and the oral and axillary temperatures allowed correct detection in 7 of 11 cases; whereas none of the fevers were detected with the ear-based measurement.In another study22 in which the temporal artery thermometer (Exergen Temporal Scanner TAT-5000A, Watertown, Massachusetts) was compared with a bladder probe, the temporal artery thermometer had a sensitivity of 0.71 and specificity of 0.97 (area under curve, 0.95) to detect a temperature greater than 37.8°C, and in patients with a bladder temperature less than 35°C, the temporal artery thermometer had a sensitivity of 0.3 and a specificity of 0.95. In the only study16 to evaluate the SensorTouch temporal artery thermometer, the sensitivity was 0 (ie, the device failed to detect any fevers) and the specificity was 1. In total, these studies indicate that noninvasive temperature measurements are accurate for ruling out hyperthermia and hypothermia but may fail to detect hyperthermia and hypothermia, depending on the thermometer used. Further research is needed in this area. Finally, it is important to recall that therapeutic decisions should not be made on the basis of a single vital sign.A critical issue to consider when using any thermometer is whether you are controlling the factors that affect the accuracy and precision of the measurement (Table 2). These factors must be addressed when educating staff on the use of the various temperature measurement methods and when critiquing the literature about a specific method and device.