Prevalence of insulin dysregulation in a sport horse population as determined by traditional and stall‐side testing

Abstract Background The prevalence of insulin dysregulation (ID) has not been investigated in sport horses. Stall‐side insulin assays offer convenience but need further validation. Objectives To determine the prevalence of and risk factors for ID in a sport horse population using an oral sugar test (OST) and assess agreement between stall‐side and validated laboratory assays for insulin quantification. Study Design Prospective cross‐sectional study. Methods One hundred eighty sport horses presenting for orthopaedic evaluation at a university hospital ( N = 140) or local farms ( N = 40) were tested for ID using OST (0.15 mL/kg light corn syrup). Morphometric data were collected, and baseline serum leptin and adiponectin concentrations were measured. Insulin concentrations were measured using the validated fluorescence enzyme immunoassay (FEIA) and a stall‐side lateral flow assay (LFA). ID was diagnosed if FEIA insulin measured >20 μIU/mL at T0 or >30 μIU/mL at T60/T90. Factors associated with ID were assessed using a mixed effects binary logistic regression model. Assay agreement was assessed using Pearson's correlation coefficient and Bland–Altman plot. Receiver operator characteristic curve analysis was performed to determine diagnostic cut‐offs for LFA. Results ID was diagnosed in 22.8% (41/180) of horses. Only 11/41 ID horses had increased fasted insulin. Increasing leptin/adiponectin ratio in males was the only significant predictor of ID (OR = 1.102, 95% CI 1.036–1.171). Correlation between FEIA and LFA insulin measures was good ( r = 0.81, 95% CI 0.78–0.84), but the LFA read on average 13.62 μIU/mL higher than the FEIA and had greater variability at higher insulin concentrations. Area under the ROC curve showed the LFA was excellent at correctly classifying ID status (AUC = 0.95, 95% CI 0.93–0.97). Main Limitations Single‐centre study, voluntary enrolment may overestimate ID prevalence. Conclusions ID is common in sport horses and often requires dynamic insulin testing to diagnose. Physical appearance does not predict ID risk. Agreement between stall‐side and laboratory assays is good.