Abstract
Background: Homeostasis model assessment (HOMA) is a surrogate index widely used to study the role of insulin sensitivity or resistance in associated disease states. However, reported values for the definition of insulin resistance (the top 25% of the distribution in non-diabetic subjects) vary widely. This study evaluates the effect of HOMA computational methods [original HOMA model formula (HOMA1) and online calculator computer model (HOMA2)] on the associations and cut-off limits for insulin resistance.
Methods: Anthropometric measurements, fasting adiponectin, leptin, leptin receptor, insulin, glucose, high-sensitivity C-reactive protein and a lipid profile were measured in type 2 diabetic patients (n=226) and their normoglycemic first degree relatives (n=319). HOMA1 and HOMA2 were used to estimate insulin resistance, β-cell function and insulin sensitivity. Subjects were classified as metabolic syndrome positive or negative (International Diabetes Federation criteria). Bland-Altmann analysis was used to evaluate agreement between the computational methods. Univariate and multivariate logistic regression analyses were used to relate the HOMA computational methods with metabolic variables and metabolic syndrome status.
Results: The two computational methods had different cut-offs for the definition of insulin resistance: HOMA1 formula ≥2.5; HOMA2 calculator ≥1.4. Correlations of the two HOMA computational methods with anthropometric and metabolic variables showed some degree of variation. The odds ratios of the associations with the metabolic syndrome for the HOMA1 formula and HOMA2 calculator computational methods were 2.04 and 1.43, respectively. Receiver operating characteristic analysis for diagnosis of the metabolic syndrome showed that areas under the receiver operating characteristic curves for the HOMA1 formula and HOMA2 calculator computational methods were 0.741 and 0.680, respectively.
Conclusions: The HOMA computational method is a significant determinant of the associations and classification of insulin resistance.
Clin Chem Lab Med 2010;48:1629–34.
©2010 by Walter de Gruyter Berlin New York
