We addressed the hypothesis that measuring the fasting insulin in healthy subjects would enable predicting the subsequent development of MS. This study has shown that in a general cohort of healthy Asian subjects, 8.5% developed MS over a 5 year period and that elevated fasting insulin levels predicted the development of subsequent MS even in the subgroup of patients without any MS component at baseline. This was particularly seen for the subjects in the highest quartile of the insulin levels when compared to those subjects with lower fasting insulin levels. Therefore, high baseline fasting insulin levels independently predicted the development of MS over time.
Numerous studies have been performed that showed that insulin resistance is associated with MS per se or the factors that comprise MS (e.g., NAFLD [21, 22], T2DM [23, 24], CVD ). In the study by Mykkänen et al ., insulin sensitivity assessed by frequently sampled intravenous glucose tolerance test and the minimal model showed significant association with the number of metabolic disorders. In addition, fasting hyperinsulinemia has been used as a surrogate estimate of insulin resistance according to various combinations of fasting insulin and the glucose concentration such as HOMA (homeostatic model assessment) [27, 28], but not the fasting insulin level per se. Assuming that insulin resistance and the obesity being the central pathogenic factor in the development of MS, a recent study reported the role of leptin, an adipokine associated with obesity, on the prediction of MS and CVD risk . However, our study is the only study that has measured the baseline insulin concentration as a surrogate marker in order to evaluate the risk of developing MS among individuals with and without any MS component and as stratified by the fasting insulin concentration.
Morrison et al. have recently shown that an interaction between the BMI and the HOMA score is associated with the development of MS in children  and De Boer et al. has shown in children that the features of MS are correlated with the fasting insulin levels . Not surprisingly, higher fasting glucose within the normal range in children is associated with increased insulin levels, yet that study was correlative rather than prognostic . Our study adds to the growing evidence by showing that elevated insulin occurs early and it predicts the subsequent development of MS. We have also found that an elevated fasting insulin level itself is potentially a key early feature in the pathophysiology of the clinical risk factor cluster of MS.
There have been many studies that have shown that patients with MS are at an increased risk of cardiovascular atherosclerosis [32–35]. Mehta et al. have recently shown that measures of insulin resistance, such as HOMA-IR, add predictive value over and above the diagnosis of MS itself in terms of the correlation with coronary artery calcification, which is a marker of the atherosclerotic risk . In the study by Tenenbaum et al ., insulin resistance assessed by HOMA-IR was an independent predictor for new major cardiovascular events among patients with preexisting coronary artery disease. Therefore, strategies that screen patients' insulin levels and therapies that specifically target hyperinsulinemia may have value to prevent this common risk factor for CVD.
There are several limitations in our study. There is ongoing controversy about whether the concept of metabolic syndrome is useful to predict CVD and T2DM [2, 5, 9]. The diagnosis of MS has recently been shown to be useful for predicting subsequent CVD events, including in Asian cohorts  and in a recent meta-analysis . Insulin sensitivity does not account for 100% of the variation in insulin response . Kim et al. showed that the fasting insulin concentration reflects less than 40% of the variability in insulin resistance as measured by a direct technique, although the fasting insulin concentration is significantly associated with insulin resistance . This study population was extremely well characterized, yet it was a homogenous population and it is representative of an adult Korean working population. This limits the generalizability of the data to other patient populations to some extent. Although the euglycemic insulin clamp test is currently the best and most accurate technique for assessing insulin sensitivity, the fasting insulin level was used as the marker for insulin resistance in this current study . The reason we didn't use the euglycemic insulin clamp technique is that it is almost impossible to perform this test in all the participants in this kind of large study population. Lastly, we did not analyze the risk for MS with Cox proportional hazard model or Kaplan-Meir survival estimate. This study was done in subjects who performed health examination twice in 5 years of interval, 2003 and 2008. We cannot specify the exact point when the subjects developed the disease of interest during those 5 years. Therefore, Cox proportional hazard model or Kaplan-Meir survival estimate could not be used, and this could have lowered the significance of the results. However, our study result has its own meaning study in that high baseline insulin concentration predicts the risk of developing MS among individuals with and without any MS component and as stratified by the fasting insulin concentration.
In conclusion, this study is the first to evaluate elevated fasting insulin as an independent predictive factor for the development of subsequent MS over a 5 year period in a well characterized cohort of apparently healthy adults. This study supports the concept that hyperinsulinemia is an early feature, if not the central feature, of the cardiovascular clusters of MS. We have to keep in mind the importance of high fasting insulin levels not just as a surrogate marker to predict future development of MS, but also the increased risk for future CVD itself as well. The usefulness of strategies to reverse hyperinsulinemia, such as lifestyle modifications, on the prevention of MS and CVD, warrants further investigation.