Clinical utility of serum 25-hydroxyvitamin D in the diagnosis of insulin resistance and estimation of optimal 25-hydroxyvitamin D in US adults

Aims: To assess the clinical utility of measuring serum 25-hydroxyvitamin-D [25(OH)D] along with traditional risk factors in the diagnosis of insulin resistance (IR) and to estimate the optimal 25(OH) D level associated with normal glucose and insulin homeostasis. Methods: A cross-sectional analysis of 6868 adults aged >= 20 years without diagnosed diabetes in the National Health and Nutrition Examination Survey, with available standardized 25(OH) D data (2001-2010). IR was defined by the homeostatic-model-assessment of insulin resistance (HOMA-IR; >= 75th percentile, sex-specific: 3.9 in men or 3.6 in women). Using logistic regression, two risk models were developed to estimate the risk of IR: Model 1 included established risk factors, and Model 2 additionally included serum 25(OH) D. Predictiveness curves and decision-curve analysis were used to assess differences in IR detection among models. Receiver-operating-characteristic curves were used to estimate the lower threshold for 25(OH) D. Results were validated in a testing sample. Results: Model 2 marginally improved detection of IR: at a risk threshold of 0.2, adding 25 (OH) D would identify an additional 2 to 4 cases per 1000 people. Overall, the lower 25 (OH) D threshold was estimated at 60 nmol/L, however, the threshold differed by ethnicity (Mexican-Americans: 54 nmo/L, non-Hispanic whites: 68 nmol/L, and non-Hispanic blacks: 41 nmol/L). Conclusion: Addition of serum 25(OH) D to traditional risk factors provided small incremental improvement in detection of IR in asymptomatic adults. The optimal 25(OH) D threshold was estimated to be at least 60 nmol/L, however, the threshold may differ by ethnic-background. Further research is needed to validate these results in other populations. (C) 2017 Elsevier B.V. All rights reserved.