Objective: Inflammatory arthritis, especially rheumatoid arthritis (RA), imposes substantial morbidity that diminishes quality of life and escalates health care costs when left untreated. Timely diagnosis and treatment are pivotal. However, limited access to rheumatologists underscores the importance of triaging referrals based on symptom severity. We developed a protocol for using infrared thermography (IRT) to detect joint inflammation and implemented it in an academic primary care clinic as a screening tool for rheumatology referrals. Materials and Methods: We enrolled people with RA, osteoarthritis (OA), and controls to undergo joint power doppler ultrasound (PDUS) and IRT. IRT image analysis employed manual segmentation with specialized software to determine surface joint temperatures. IRT temperature cutoff points using PDUS as the gold standard for joint inflammation were established. Subsequently, we recruited people with hand or foot joint pain and fast-tracked to rheumatology those surpassing IRT cutoff points for joint inflammation. Results: Thirty-two people with RA, 10 with OA, and 9 controls were enrolled. Most participants (86.3%) were females, and 37.3% were black. Temperature measurements showed robust interrater reliability. Cutoff points at metacarpophalangeal (MCP) (T center ≥32.93 °C, T mean ≥32.67 °C) and wrist (T center ≥33.76 °C, T mean ≥33.79 °C) joints discriminated between inflamed and noninflamed joints. Of the 20 pilot participants, N = 10 (50%) were referred to rheumatology using IRT findings, with N = 8 (80%) seen within 4 weeks. Diagnoses included osteoarthritis, gout, and calcium pyrophosphate deposition disease. Conclusion: IRT demonstrates potential as a reliable tool for identifying inflamed joints and offers a feasible pathway for identifying patients with joint-level thermal abnormalities who may benefit from expedited rheumatologic evaluation. Further research is needed to refine IRT-based rheumatology referral protocols and optimize their utilization.
Chavez et al. (Mon,) studied this question.