Determination of insulin resistance in surgery: the choice of method is crucial

Background & aims: In elective surgery, postoperative hyperglycaemia and insulin resistance are independent risk factors for complications. Since the simpler HOMA method has been used as an alternative to the hyperinsulinemic normoglycemic clamp in studies of surgery induced insulin resistance, we compared the two methods in patients undergoing elective surgery.

Methods: Data from 113 non-diabetic patients undergoing elective surgery were used. Insulin sensitivity, both before and after surgery, was quantified by the clamp and HOMA. Pre- and postoperatively, the results of the clamp were compared to HOMA using regression- and correlation analysis. Degree of agreement between the methods was studied using weighted linear kappa and the Bland-Altman test.

Results: Both the clamp and HOMA recorded a mean relative reduction in insulin sensitivity of 39 ± 24% and 39 ± 61% respectively after surgery; with significant correlations (p < 0.01) for pre- and post-operative measures as well as for relative changes. However r(2) values were low: 0.04, 0.07 and 0.03 respectively. The degree of agreement for the relative change in insulin sensitivity using the Bland-Altman test gave a mean of difference 0% but "limits of agreement" (± 2SD) was ± 125%. This poor inter-method agreement was consolidated by a weighted linear kappa value of 0.18.

Conclusion: While the hyperinsulinemic euglycemic clamp measures the postoperative changes in insulin sensitivity, HOMA measures something different. Data using the HOMA method must therefore be interpreted cautiously and is not interchangeable with data obtained from the clamp.