Comparison of equivalence and determination of diagnostic utility of min-mod and clamp methods for insulin resistance in diabetes free subjects: a meta-analysis.

The gold standard for quantifying insulin sensitivity (IS) is the hyperinsulinemic-euglycemic clamp (Clamp) with a cut-off point of 5x10-2 (dL/min)/(microU/mL) or less to indicate insulin resistance. Bergman's minimal model (Min-Mod) is also being used to estimate IS, but there are doubts as to its equivalence with Clamp. The objective of the present study is to determine if Clamp and the tolbutamide and insulin techniques of Min-Mod are equivalent. Meta-analysis based on a bibliographic search from 1970 until the present was made for the MeSH terms: insulin resistance, hyperglycemic-clamp, euglycemic-clamp, Min-Mod, minimal model approach. Concordance was determined with both simple and intraclass correlation and Bland and Altman's concordance limits using R. Three of the 109 articles found were included. The concordance limits indicate that Clamp and Min-Mod are not equivalent, which could result in diagnostic errors if the accepted cut-off point is used for both methods. Given this lack of equivalence, a ROC analysis was performed and new diagnostic cut-off points of 2.4 and 4.6x0-2 (dL/min)/(microU/mL) for insulin and tolbutamide techniques of Min-Mod, respectively, are proposed, with adequate sensitivity, specificity, and predictive value. These values should be prospectively validated.