Comparison between HOMA-IR and ISI-gly in detecting subjects with the metabolic syndrome

BACKGROUND: To verify whether, as index of insulin resistance, ISI-gly (insulin sensitivity index) is more efficient than HOMA-IR (homeostatic model assessment) or QUICKI (quantitative insulin sensitivity check index) in detecting patients with the metabolic syndrome. METHODS: Excluding patients with known diabetes, endocrine, liver and kidney diseases, we enrolled 553 subjects who were screened for metabolic abnormalities. After 5 days of a balanced weight maintenance diet, we performed an OGTT (oral glucose tolerance test) and measured fasting and 2-h postload blood sugar and insulin, from which we calculated ISI-gly, HOMA-IR and QUICKI, stratifying patients in tertiles. Statistical comparisons were performed for a number of metabolic variables between tertiles of the same index, as well as between tertiles of different indexes presumably expressing identical insulin resistance. RESULTS: All variables reflecting the metabolic syndrome were significantly more altered in the top as compared to the intermediate and the lowest tertile for HOMA-IR, the opposite for ISI-gly. Comparing homologous measurements of the top tertile of HOMA-IR with the lowest tertile of ISI-gly (presumably expressing identical insulin resistance), fasting glucose and insulin were significantly higher, while 2-h OGTT values were significantly lower. The opposite occurred comparing the lowest HOMA-IR to the highest ISI-gly tertile, the diagnostic predictive values of the latter in detecting metabolic derangements being also higher. Data from QUICKI 1st to 3rd tertiles exactly matched those obtained from HOMA-IR 3rd to 1st tertile. CONCLUSIONS: ISI-gly, which includes postload glucose and insulin concentrations, provides a more accurate estimate of whole-body insulin sensitivity than HOMA-IR or QUICKI, derived from fasting measurements only, thus constituting a more sensitive tool for screening and preventing metabolic abnormalities.