Reliability of infrared tympanic thermometry in the detection of rectal fever in children

Abstract

Study objective: Recently published clinical guidelines for the management of febrile children are based on studies that used rectal temperature data to stratify the risk of bacteremia and septic complications. Appropriate management decisions rely on accurate detection and categorization of fever. Accordingly, this study compared the newer infrared tympanic thermometry (ITT) to rectal thermometry in this regard.

Design: Prospective observational study.

Setting: Urban teaching hospital ED with annual census of 60,000.

Participants: Consecutive children 6 months to 6 years old who had rectal temperatures measured.

Interventions: Triage nurses recorded rectal temperatures and bilateral ITT temperatures. Temperatures were correlated by Pearson correlation coefficients and compared using paired t tests with significance set at P < .01. Children were categorized by degree of fever using rectal temperature (afebrile, less than 100.4 degrees F; low fever, 100.4 to 102.9 degrees F; and high fever, more than 102.9 degrees F), and the accuracy of ITT in detecting fever and high fever was determined.

Results: Three hundred seventy patients were enrolled in the study. The mean age was 18.4 +/- 11.3 months; boys comprised 56% of patients. The mean temperatures were rectal, 101.0 +/- 2.0 degrees F; right tympanic membrane, 100.4 +/- 1.9 degrees F; and left tympanic membrane, 100.3 +/- 1.9 degrees F. The tympanic membrane temperatures were significantly lower than rectal readings (P << .001 for both right and left versus rectal). Rectal temperatures showed good correlation with both right and left tympanic membrane temperatures (r = .83 and .85, respectively). ITT was 76% sensitive and 92% specific in detecting fever of 100.4 degrees F or more (positive predictive value, 0.92; negative predictive value, 0.76). In the detection of high fever, ITT was only 57% sensitive but 98% specific (positive predictive value, 0.90; negative predictive value, 0.90). Rectal and TM temperatures differed by at least 0.5 degree F in 70% of the patients, 1.0 degree F in 41%, 2.0 degrees F in 12%, and 3.0 degrees F in 3%.

Conclusion: Despite the statistical correlation between ITT and rectal temperatures, the modalities may yield significantly different temperatures. The poor sensitivity of ITT in detecting fever and high fever may result in clinically important miscategorizations of individual patients. Current clinical management that is based on the presence and height of fever may be adversely affected if ITT is used.