Objective, noninvasive biomarkers of joint inflammation or pain are needed for dogs with osteoarthritis (OA). This prospective, observational study investigated whether body surface temperature quantified by infrared thermography was associated with joint pain and/or radiographic osteoarthritis (rOA) in dogs. Using thermal images, 14 regions of interest (ROI) were created for bilateral limb joints and the vertebral column for each dog and the mean temperature was calculated for each ROI. Orthopedic examination and radiographs classified each joint/region as affected or unaffected by rOA, pain, or both, called symptomatic OA (sxOA). Mean temperatures were compared between affected and unaffected joints. Eighteen dogs were enrolled, encompassing analysis of 250 ROI (rOA = 83/250, 33%; pain = 83/250, 33%; sxOA = 61/250, 24%). No temperature differences were detected between affected and unaffected joints for rOA (0.405 °C, 95% CI [−0.204, 1.025 °C]), pain (0.208 °C, 95% CI [−0.390, 0.806 °C]), or sxOA (0.276 °C, 95% CI [−0.344, 0.915 °C]). Elbow and stifle regions were warmer, and the lumbosacral region was cooler than the carpus. Long-haired dogs were cooler than short-coated dogs. Regional temperature variability and haircoat type might hinder detection of OA-related thermal signals. In dogs with OA, findings do not support infrared thermography as a standalone tool to identify affected joints.
Infrared thermography quantifies body surface temperature with growing applications in dogs, particularly among rehabilitation practitioners; however, incomplete understanding of normal temperature variation in healthy dogs limits interpretation. This study characterized body surface temperature distribution and evaluated factors influencing thermographic measurements in a heterogeneous group of healthy dogs. Client-owned, adult dogs were prospectively enrolled and underwent standardized thermographic imaging, with regions of interest defined for trunk, neck, and spine. Mean surface temperatures were assessed for symmetry, compared across regions and explored for relationships with individual dog characteristics. p < 0.05 was significant with adjustment for multiple comparisons. Thirty-three dogs were included with median age of 3.3 (1.0–10.9) years and median body weight of 18.6 (3.7–59.4) kgs. No significant left–right asymmetry was identified in paired regions. Mean (± SD) surface temperature for the dorsal neck (23.9° ± 1.6°) was significantly cooler than the lumbar spine (25.4° ± 1.9°, p = 0.009) and lateral trunk (25.2° ± 2.0°, p = 0.03) regions. Short-haired dogs exhibited significantly higher values than medium/long-haired dogs for the lateral trunk (26.8° ± 1.9° vs. 24.2° ± 1.4°, p = 0.001), dorsal neck (25.2° ± 1.8° vs. 23.2° ± 0.9°, p = 0.01) and sacral (26.0° ± 2.2 °C vs. 23.5° ± 1.1 °C, p = 0.006) regions. No other significant relationships were identified. Regional variation and coat length are important factors influencing thermography in healthy dogs, while bilateral symmetry is generally preserved. Establishing variability in healthy dogs is critical to clinical interpretation of thermography in abnormal dogs.