correlation study of some methods used for insulin sensitivity assessment: the Egyptian experience

ABSTRACT

Various surrogate methods for the quantification of insulin sensitivity have been proposed. A comparative evaluation is lacking and is relevant for the standardization of investigative methods and comparability of results. A gold standard in measuring insulin sensitivity is the minimal model derived estimates of insulin sensitivity, but this method is difficult to apply in large studies. Therefore, indirect indices of insulin sensitivity were proposed, namely I, In [I], 10[4] [I x G], I/G, HOMA-IR, In [HOMA-IR], QUICKI, AIR, and HOMA-beta cell. The aim of the present study was to compare these simple indices with data from the MINMOD. Forty Egyptian subjects of whom 15 were obese and 25 were nonobese. All subjects underwent an FSIGT. Plasma glucose was determined and serum insulin was measured. The insulin sensitivity, SI was calculated with the MINMOD program. HOMA IR, beta cell function and QUICKI were calculated. In [HOMA-IR] correlated best with the MINMOD-derived SI [r = -0.401, P = 0.010]. This was followed by In [I] and QUICKI which were equally correlated with SI [r = -0.389, P = 0.013 and r = 0.388, P = 0.013, respectively]. This was followed by I/G and HOMA-IR which correlated better with the MINMOD-derived SI than 10[4] / [Insulin x G] and I, respectively [r -0.352, P = 0.026 and r = -0.350, P = 0.027 versus r = 0.343, P = 0.030 and r = -0.343, P = 0.030, respectively]. AIR and HOMA-beta cell did not show any statistically significant correlation with the MINMOD-derived SI. In both obese and nonobese subjects HOMA-IR showed a highly significant negative correlation with QUICKI [r = -0.875, P = 0.000 and r = -0.890, P = 0.000, respectively]. Also, In [HOMA-IR] showed a highly significant negative correlation with QUICKI in both obese and nonobese subjects [r = -0.968, P = 0.000 and r = -0.947, P = 0.000, respectively]. Fasting insulin-derived measures of insulin sensitivity proposed are relatively crude methods for the quantification of insulin sensitivity in comparison to the Minimal Model analysis of the FSIVGTT and are, therefore, of limited value for the assessment of the metabolic status of an individual patient. When used in epidemiologic studies, fasting insulin alone, or the widely used HOMA or QUICKI perform at least as surrogate measures of insulin sensitivity or means for the identification of individuals with the metabolic syndrome in the general population