Variability of resting endoscopic grading for assessment of recurrent laryngeal neuropathy in horses

Reasons for performing study: The extent to which variability affects endoscopic grading of arytenoid cartilage movement is uncertain. Objective: To determine the observer and within horse variability of grading arytenoid cartilage movement in horses during resting endoscopic examination, using a 7-grade system. Methods: Endoscopic recordings of the upper respiratory tract made at rest in 270 draught horses were reviewed independently by 2 veterinarians to assess interobserver variability when scoring horses' laryngeal function with a 7-grade system. Grading was repeated by both examiners in 80 randomly selected recordings in order to assess intraobserver variability. In 120 horses, endoscopy was repeated after 24-48 h, with videos graded by both veterinarians to assess intrahorse variability. Results: The mean weighted K statistic for concordance within examiners was 0.867, with a mean intraobserver agreement of 76.3%. The weighted kappa statistic for concordance between the 2 examiners was 0.765, with an interobserver agreement of 63.1%. Of the horses receiving 2 endoscopic examinations, the same grade was assigned to 57.1% of horses at the second examination, when effects resulting from interobserver variability were removed. The mean weighted K statistic for concordance between the grade assigned at first vs. second examinations was 0.588, indicating only moderate agreement. Conclusions and potential relevance: Intra- and interobserver reliability of resting endoscopic grading of arytenoid cartilage movement using a 7-grade system was high when examinations were conducted by experienced veterinarians. However, there was moderate daily intrahorse variability, suggesting that results of resting endoscopic examinations performed on a single day should be interpreted with caution.