Comparative performance of invasive, semi-invasive, and non-invasive insulin resistance-related indices across prediabetes and normoglycemia

Abstract

Insulin resistance, a defining abnormality in the development of type 2 diabetes, often emerges silently during prediabetes. This study evaluated the comparative performance of multiple insulin resistance-related indices across adults with prediabetes and normoglycemia. A cross-sectional analysis was conducted among 100 adults aged 25–55 years, equally divided between the two glycemic groups. Twenty-five candidate indices were examined, spanning blood-based indices, lipid–anthropometry combinations, and non-invasive anthropometric and body-composition measures, including several newly derived indices. Correlation analysis was performed to assess the relationship between candidate indices and the Homeostasis Model Assessment for Insulin Resistance, followed by logistic regression and receiver operating characteristic curve evaluation for indices showing significant correlations. Individuals with prediabetes demonstrated significantly higher adiposity and differences in metabolic parameters compared to individuals with normoglycemia (p < 0.05 for most variables). Seventeen of the twenty-five candidate indices showed significant correlations with the Homeostasis Model Assessment for Insulin Resistance. Among laboratory-based indices, the triglyceride–glucose index and the homeostasis model assessment showed good discriminative performance. Within lipid–anthropometry combinations, the triglyceride–glucose waist-to-height ratio and triglyceride–glucose waist circumference provided the best performance, each yielding an area under the curve of approximately 0.79. Several non-invasive indices also demonstrated good comparative performance. The bioimpedance-based somatotype adiposity index showed the highest overall performance among individual indices, with an area under the curve of 0.82, sensitivity of 78 percent, specificity of 80 percent, and an adjusted odds ratio of 14.9. These findings suggest that selected non-invasive anthropometric and body-composition indices may provide a practical and low-cost approach for assessing metabolic alterations associated with insulin resistance across glycemic groups, although further validation is required.