Infrared Spectroscopy of Synovial Fluid Shows Accuracy as an Early Biomarker in an Equine Model of Traumatic Osteoarthritis

Simple Summary Osteoarthritis is a leading cause of lameness and joint disease in horses. A simple, economical, and accurate diagnostic test is required to routinely screen horses for OA. The study assessed the accuracy of infrared (IR) spectroscopy in analyzing synovial fluid (SF) to identify horses with early inflammatory changes related to equine carpal osteoarthritis (OA). OA was surgically induced in one group of horses while the others were allocated as controls. SF samples were collected before OA induction and weekly until 63 days. IR spectroscopy was used to analyze the SF samples, and predictive models were created to classify the samples. Overall, the accuracy for distinguishing between joints with OA and any other joint was 80%. Results show that IR spectroscopy could classify samples based on the day they were collected with 87% accuracy. Distinguishing between OA vs. OA Control and OA vs. Sham joints had lower accuracies of 75% and 70%, respectively. The authors conclude that IR spectroscopy accurately discriminates between SF in joints with induced OA and controls.Abstract Osteoarthritis is a leading cause of lameness and joint disease in horses. A simple, economical, and accurate diagnostic test is required for routine screening for OA. This study aimed to evaluate infrared (IR)-based synovial fluid biomarker profiling to detect early changes associated with a traumatically induced model of equine carpal osteoarthritis (OA). Unilateral carpal OA was induced arthroscopically in 9 of 17 healthy thoroughbred fillies; the remainder served as Sham-operated controls. The median age of both groups was 2 years. Synovial fluid (SF) was obtained before surgical induction of OA (Day 0) and weekly until Day 63. IR absorbance spectra were acquired from dried SF films. Following spectral pre-processing, predictive models using random forests were used to differentiate OA, Sham, and Control samples. The accuracy for distinguishing between OA and any other joint group was 80%. The classification accuracy by sampling day was 87%. For paired classification tasks, the accuracies by joint were 75% for OA vs. OA Control and 70% for OA vs. Sham. The accuracy for separating horses by group (OA vs. Sham) was 68%. In conclusion, SF IR spectroscopy accurately discriminates traumatically induced OA joints from controls.