Surrogate measures of insulin sensitivity vs the hyperinsulinaemic–euglycaemic clamp: a meta-analysis. Are there not some surrogate indexes lost in this story?

Excerpt

To the Editor: We read with interest the meta-analysis recently published by Otten and colleagues in Diabetologia on surrogate measures of insulin sensitivity vs the hyperinsulinaemic–euglycaemic clamp [1]. We agree with the comment from J. R. Petrie in the same issue of Diabetologia [2], and think that this study is both timely and useful. However, we have concerns about its conclusions. Indeed, we are afraid that the conclusion of the article may direct readers towards several surrogates, studied in the meta-analysis, and discourage the use of other reliable surrogates that could be more useful in some situations. For example, we previously described an OGTT-derived surrogate called the ‘simple index assessing insulin sensitivity derived from OGTT’ (SIisOGTT) [3], in comparison with the hyperinsulinaemic–euglycaemic clamp and with other fasting and OGTT-derived surrogates [3, 4]. While the original study [3] was cited by the authors, it was not studied in the meta-analysis [1]. However, other independent groups have described it as a particularly relevant index in some cases and as good as the well-known Matsuda index [5, 6]. Indeed, three OGTT-derived indexes (the Matsuda index, SIisOGTT and Avignon’s insulin sensitivity index) have been described as being highly correlated with the direct measurement of insulin sensitivity using the hyperinsulinaemic–euglycaemic clamp [5]. Two of these indexes, the Matsuda index and SIisOGTT, were found to be more reliable than others in a study that investigated differences in insulin resistance between race and sex [6]. Therefore, we think that when the revised-QUICKI is not measurable due to the unavailability of NEFA, SIisOGTT is probably as good as the other OGTT-derived indexes proposed by Otten et al [1]. Other fasting indexes such as the Disse index, which we developed few years ago and which takes into account NEFA, HDL-cholesterol:total-cholesterol ratio and insulin, have not been tested in the meta-analysis, while the Disse index has been previously described as being as good as the revised-QUICKI in comparison with the hyperinsulinaemic–euglycaemic clamp [7]. Another point that was not discussed is the use of these surrogate indexes of insulin resistance in considering cardiovascular risk, for which simple fasting indexes that include triacylglycerols in their calculation have been shown to perform well [8]. Finally, it should also be noted that indexes initially validated against the insulin-modified frequently sampled IVGTT and thereafter compared with the hyperinsulinaemic–euglycaemic clamp have been shown by different studies to be relevant [5, 8]. …