Surrogate markers of insulin resistance in subjects with metabolic syndrome – data of the Berlin Aging Study II

Background: The metabolic syndrome (MetS) is a cluster of metabolic changes which is associated with insulin resistance (IR). Cutoff values for the homeostasis model of insulin resistance (HOMA-IR) – a surrogate marker of IR-to identify subjects with MetS are not established. Methods: Cross-sectional data of 446 younger (53% women, 28±3 years old) and 1271 elderly study participants (52% women, 68±4 years old) without diabetes were available for the current analysis. MetS was defined according to the IDF/AHA/NHLBI (International Diabetes Federation/American Heart Association/National Heart, Lung, and Blood Institute) criteria of 2009. Using receiving operating characteristics (ROC) analysis cutoff values for HOMA-IR were calculated above which participants with MetS could be identified with highest sensitivity and specificity. Finally, binary logistic regression models were calculated. Results: The prevalence of MetS was 6.7% in young and 28.3% in elderly subjects. HOMA-IR cutoff values for the detection of MetS were HOMA-IR >1.88 (young subjects; sensitivity 80%, specificity 85.3%, positive predictive value 80%, negative predictive value 15%) and HOMAIR >1.98 (elderly subjects; sensitivity 73.6%, specificity 72.9%, positive predictive value 74%, negative predictive value 27%). In adjusted regression models [age, body mass index (BMI), sex, physical activity and age groups] subjects above these cutoff-values had odds of 5.7 [95% confidence interval (CI): 4.1–7.9] in elderly and 22.2 (95% CI: 7.0–70.5) in younger study participants to have MetS. Conclusions: Cutoff values for HOMA-IR are not established in clinical practice; however, they could be used to identify subjects with MetS, even if a diagnosis of MetS cannot made based solely on HOMA-IR considering the negative predictive values.