Lacrimal caruncula thermography overestimates rectal temperature in horses: Influence of camera type and measurement distance.

Abstract

BACKGROUND: Rectal temperature (RT) is the gold standard for assessing body temperature in horses, but handling and welfare concerns limit its use. Thermography of the lacrimal caruncula (LC) has been proposed as a noncontact alternative, although its accuracy may be influenced by device characteristics and measurement distance. OBJECTIVES: To evaluate agreement between LC thermography and RT in horses, compare two infrared thermal cameras and determine the effect of camera-to-subject distance on measurement bias and variability. METHODS: LC temperature was recorded at 0.5, 1 and 2 using two cameras (IR FlexCam S and UNI-T UTi165A). Eighty-four healthy Thoroughbreds contributed three measurements per camera. Agreement between LC and RT was evaluated using Bland-Altman analysis (bias and limits of agreement, LoA). Linear regression of the difference versus mean temperature assessed proportional bias. RESULTS: Both cameras produced higher LC values than RT at all distances, and agreement decreased as distance increased. For FlexCam, biases were 3.75°C at 0.5 m, 3.71°C at 1 m and 5.28°C at 2 m. LoA widened from 1.06-6.44°C at 0.5 m to 1.23-9.33°C at 2 m. For UTi165A, biases were 2.97°C at 0.5 m, 3.85°C at 1 m and 5.21°C at 2 m, with LoA from 0.86-5.08°C at 0.5 m to 2.04-8.39°C at 2 m. Negative regression slopes indicated proportional bias for both devices. MAIN LIMITATIONS: The indoor setting, evaluation of a single anatomical site and inclusion of only healthy Thoroughbreds limit generalisability. Only two camera systems were assessed. CONCLUSION: UTi165A showed narrower LoA and better repeatability at 0.5 m but did not demonstrate agreement sufficient to replace RT. Short acquisition distances, device selection and standardised protocols are essential if LC thermography is to be used as a screening adjunct in equine practice.