A pilot study of agreement between noninvasive thermometers and the core temperature of postoperative cardiothoracic surgical patients

Abstract

ObjectiveReliable and accurate temperature assessment is fundamental for clinical monitoring; noninvasive thermometers of various designs are widely used in intensive care units, sometimes without a specific assessment of their suitability and interchangeability. This study evaluated agreement of four noninvasive thermometers with a pulmonary artery catheter temperature. MethodsThis prospective method comparison study was conducted in an Australian adult intensive care unit. One hundred postoperative adult cardiothoracic surgery patients who had a pulmonary artery catheter (Edwards Lifescience) in situ were identified. The temperature reading from the pulmonary artery catheter was compared to contemporaneous measurements returned by four different thermometers—temporal Artery (TA, Technimed), Per Axilla (Axilla, Welch Allyn), Tympanic (Tymp, Covidien), and the NexTemp® (NEXT, Medical Indicators [used per axilla]). The time required to obtain each noninvasive temperature measurement was recorded. ResultsAgreements between each noninvasive temperature and the pulmonary artery catheter standard were assessed using summary statistics and the Bland–Altman method comparison approach. A clinically acceptable maximum difference from the standard was defined as ±0.5 °C. Temperature agreement with the pulmonary artery standard (mean difference °C [95% limits of agreement °C]) was greatest for Tymp (−0.20 [−0.92 to 0.52]), intermediate for AXILLA (−0.37 [−1.3 to 0.59]) and NEXT (−0.71 [−1.7 to 0.27]), and least for TA (−0.60 [−2.0 to 0.81]). The proportion of measurements within ±0.5 °C of the standard were TYMP (81%), AXILLA (63%), TA (45%), and NEXT (30%). The time to obtain measurements varied, with the Tymp and TA estimates immediate, the AXILLA a mean of 40 s (standard deviation = 11 s), while NEXT results were at the manufacturer-recommended 3-min point. ConclusionsTympanic thermometers showed closest agreement with the pulmonary artery standard. Deviations by more than 0.5 °C from that standard were relatively common with all noninvasive devices.