Background: Body temperature is commonly used to screen patients for infectious diseases, establish diagnoses, monitor therapy, and guide management decisions.
Purpose: To determine the accuracy of peripheral thermometers for estimating core body temperature in adults and children.
Data sources: MEDLINE, EMBASE, Cochrane Central Register of Controlled Trials, and CINAHL Plus from inception to July 2015.
Study selection: Prospective studies comparing the accuracy of peripheral (tympanic membrane, temporal artery, axillary, or oral) thermometers with central (pulmonary artery catheter, urinary bladder, esophageal, or rectal) thermometers.
Data extraction: 2 reviewers extracted data on study characteristics, methods, and outcomes and assessed the quality of individual studies.
Data synthesis: 75 studies (8682 patients) were included. Most studies were at high or unclear risk of patient selection bias (74%) or index test bias (67%). Compared with central thermometers, peripheral thermometers had pooled 95% limits of agreement (random-effects meta-analysis) outside the predefined clinically acceptable range (± 0.5 °C), especially among patients with fever (-1.44 °C to 1.46 °C for adults; -1.49 °C to 0.43 °C for children) and hypothermia (-2.07 °C to 1.90 °C for adults; no data for children). For detection of fever (bivariate random-effects meta-analysis), sensitivity was low (64% [95% CI, 55% to 72%]; I2 = 95.7%; P < 0.001) but specificity was high (96% [CI, 93% to 97%]; I2 = 96.3%; P < 0.001). Only 1 study reported sensitivity and specificity for the detection of hypothermia.
Limitations: High-quality data for some temperature measurement techniques are limited. Pooled data are associated with interstudy heterogeneity that is not fully explained by stratified and metaregression analyses.
Conclusion: Peripheral thermometers do not have clinically acceptable accuracy and should not be used when accurate measurement of body temperature will influence clinical decisions.
Primary funding source: None.