Axillary and rectal temperature measurements poorly agree in newborn infants

Aim: Evaluation of the agreement between axillary temperature measurements and rectal temperature measurements in neonates.

Methods: Rectal and axillary body temperatures were simultaneously measured for 3 min in 33 neonates (gestational age 25-42 weeks, weight 840-4,005 g). Two investigators performed paired measurements, one in each neonate. A single type of thermometer was used in this study: one thermometer for each rectal and another thermometer for each axillary measurement. The Bland-Altman method was used (95% 'limits of agreement': mean +/- 2 SD) to determine the level of agreement between axillary and rectal measurements.

Results: The axillary temperature was significantly lower than the rectal temperature (mean +/- SD 0.27 +/- 0.20 degrees C, p < 0.05). The '95% limits of agreement' ranged from -0.13 to +0.67 degrees C. Increasing postnatal age (days) showed a significant increase in temperature difference (rectal minus axillary; r = 0.54; p < 0.05).

Conclusions: The mean difference between axillary and rectal temperature shows a wide variation. Axillary temperature measurements cannot be used interchangeably with rectal measurements in neonates.