Characterization of Glucose Response Curves after Insulin Injection in Sensitive versus Insensitive Mares

The glucose responses to intravenous injections of a range of doses of recombinant human insulin were determined for six mares known to be insulin sensitive and six mares known to be insulin insensitive, with the goal of better characterizing the regression lines resulting from the two categories of mares. Insulin doses between 8 and 198 mU of insulin per kg of body weight (mU/kg BW) were administered intravenously between September 13 and 26, 2010, starting with 50 mU/kg BW on the first day. Higher and lower doses were administered on alternate days to obtain percentages of decreases in blood glucose concentrations between 10% and 70%. Linear regression analysis revealed that insulin-insensitive mares have glucose response curves with higher y intercepts (P = .066), less steep slopes (P = .0003), and less goodness of fit (P = .053) in addition to the expected greater dose required to produce a 50% reduction in blood glucose concentrations (ED50; P = .006), despite the similarities between their body weights and those of insulin-sensitive mares. Linear and nonlinear regression of responses to the 32, 50, and 79 mU/kg BW insulin doses with the overall estimates of ED50 and the natural log of ED50 indicated that the 50 mU/kg BW dose had the greatest coefficient of determination (>0.95). Generally, it appears that estimates of insulin sensitivity based on a single injection of insulin or on multiple injections of insulin are least variable for insulin-sensitive mares.