Mid-infrared spectroscopy of equine synovial fluid from normal and diseased joints

Infrared (IR) absorption patterns of equine synovial fluid (SF) were studied. The objectives were to: (1) develop and optimize laboratory protocols suitable for mid-infrared (MIR) spectroscopic analysis of equine SF, (2) identify significant differences among MIR spectra of SF from anatomically different types of equine joints, and (3) determine the feasibility of using MIR spectroscopy and classification algorithms for the differentiation of SF samples from diseased and control joints.;Inter-articular variability among 3 clinically normal high motion joints (antebrachiocarpal (AC), midcarpal (MC) and tarsocrural (TC)) was investigated. Statistical comparisons of MIR absorption patterns of SF from study joints were made. Samples from the contralateral pairs were likely to yield a similar MIR absorption pattern. Differences in spectral features between ipsilateral AC and MC were detected, and comparisons between the spectra of the carpal and TC joints revealed more widespread discriminatory absorption bands. The results suggested that inter-articular variation should be considered when using this technique.;The feasibility of the use of IR spectroscopy combined with statistical classification algorithms was assessed by using SF samples from joints with traumatic arthritis (TA) and control joints. The MIR absorption patterns of SF from joints with TA differed significantly from corresponding patterns for controls. A classification model was developed based on characteristics of 3 optimal MIR regions, and yielded an overall accuracy of 97% (sensitivity 93%; specificity 100%) in the calibration dataset. The same model with cost-adjusted prior probability of 0.60:0.40 produced an overall accuracy of 89% (sensitivity 83%; specificity 100%) for a validation dataset, and 100% correct classification for a second validation set of normal control SF.;The feasibility of this technique was further confirmed by comparing SF samples from tarsocrural joints with osteochondrosis (OC) and control joints. Disease-associated characteristics within MIR spectra of SF were identified by the use of statistical modeling. The classification model developed was based on the characteristics of 6 optimal MIR regions, and yielded an overall accuracy of 77% (73% sensitivity; 81% specificity).;The technique of MIR spectroscopy of dried films was optimized for equine SF. Suitable MIR spectra of equine SF were obtained from the 8 μL aliquots of 3:1 SF to aqueous potassium thiocyanate solution deposited onto a silicon microplate by using an optimal spectral acquisition protocol. The overall MIR absorption pattern of equine SF is similar to the MIR absorption pattern of human SF reported in literature. The laboratory methods used in the current research for collecting the MIR spectra are technically straightforward and economical.;The feasibility of IR and statistical classification algorithms for the differentiation of spectra derived from samples of diseased and control joints were demonstrated in this current research project. These findings favor the further development of this method for diagnosis of equine joint disease. Further recruitment of samples from both diseased and normal equine populations is required to evaluate the clinical usefulness of IR spectroscopy in diagnosis of equine joint diseases.