Fasting Insulin Guidelines

Eleanor Kellon, VMD

This is the abstract from the recent study on fasting insulin showing the cutoff of 20 uIU for normal was way too high:

Comparison of fasted basal insulin with the combined glucose-insulin test in horses and ponies with suspected insulin dysregulation


Basal (fasting) insulin is a convenient test of insulin dysregulation.

Previous cut-offs had poor sensitivity (15%) for diagnosis of insulin dysregulation.

Use of a receiver operating characteristic curve-derived cut-off improved the sensitivity (63%) of basal (fasting) insulin.

Specificity (87%) was maintained with this technique.


Fasting horses for measurement of basal serum insulin concentration (fasting insulin; FI) has been recommended to standardise testing for insulin dysregulation (ID), yet limited data exist comparing it to dynamic tests. This study aimed to compare FI with the combined glucose-insulin test (CGIT) in horses suspect for ID. We hypothesised that FI would have poor sensitivity for detecting ID compared to CGIT using conventional cut-offs. Records were retrieved from CGITs performed in horses fasted for approximately 8 h. Serum insulin and glucose concentrations were measured before and for 150 min following an IV bolus of glucose followed by insulin. Correlations between FI and CGIT values were assessed. Youden’s index analysis was used to determine the optimal cut-off for FI. Logistic regression and Mann–Whitney U tests were used to determine factors affecting the results.

CGITs (n = 130) from 62 horses were evaluated. Compared to CGIT, sensitivity and specificity of FI for diagnosis of ID were 14.6% and 100% at a cut-off of 20 μIU/mL and 63.4% and 87.2% at a cut-off of 5.2 μIU/mL, respectively. FI was significantly correlated with insulin at 45 min (rs = 0.66) and 75 min (rs = 0.72); area under the curve for insulin (AUCinsulin; rs = 0.67); glucose at 45 min (rs = 0.53); and AUCglucose (rs = 0.50). Obesity was significantly associated with increased odds of a positive CGIT and horses with a positive CGIT were significantly older (P < 0.05). In conclusion, FI correlated well with CGIT results and had adequate sensitivity and specificity at lower cut-offs, despite poor sensitivity at conventional cut-off values. Further research to derive cut-off values relevant to the fasting period is warranted.

Eleanor in PA 
EC Owner 2001

Join to automatically receive all group messages.