# Insulin resistance indices for predicting circadian syndrome: estimated glucose disposal rate as a protective indicator through non-linear dose-response in Chinese adults

## TL;DR

This study shows that insulin resistance indices, especially eGDR, can predict circadian syndrome in Chinese adults, with eGDR acting as a protective factor.

## Contribution

The study introduces eGDR as a protective indicator for circadian syndrome through non-linear dose-response relationships.

## Key findings

- eGDR showed inverse associations with CircS, indicating a protective effect.
- TyG-BMI, METS-IR, and AIP were positively associated with CircS risk.
- Sex and age modified the associations, with stronger effects in women and adults under 60.

## Abstract

Circadian syndrome (CircS) extends metabolic syndrome by incorporating sleep and mood components, demonstrating superior predictive value for cardiovascular outcomes. Insulin resistance (IR) critically mediates the bidirectional metabolic-circadian disruption underlying CircS. While CircS affects nearly 40% of Chinese adults, the predictive utility of IR surrogate indices remains unexplored. This study evaluated the associations between six IR indices and both the prevalence and incidence of CircS among middle-aged and older Chinese adults.

We examined associations between six IR indices—including estimated glucose disposal rate (eGDR), triglyceride-glucose-body mass index (TyG-BMI), and metabolic score for insulin resistance (METS-IR)—and CircS among 8,392 participants aged ≥45 years from the China Health and Retirement Longitudinal Study (CHARLS, 2011–2015) using cross-sectional and longitudinal designs. Logistic and Cox regression models assessed associations with covariate adjustment. Subgroup analyses by age and sex included interaction testing. Restricted cubic splines (RCS) and receiver operating characteristic (ROC) curves evaluated dose-response relationships and discriminatory performance.

CircS prevalence was 37.0% (3,101/8,392), with marked sex differences: women accounted for 69.0% of CircS cases vs. 44.0% of non-CircS cases (P < 0.001). Diabetes prevalence was 3.1-fold higher in CircS patients (11.2 vs. 3.6%, P < 0.001). All indices demonstrated excellent discrimination (AUC: 0.893–0.907). After full adjustment, eGDR showed inverse associations with CircS (OR = 0.381; HR = 0.602), while TyG-BMI, METS-IR, and AIP were positively associated with CircS risk in both analyses. RCS revealed varied dose-response patterns, with most indices showing non-linearity and AIP demonstrating a linear relationship. Sex and age significantly modified several index-CircS associations, with stronger associations observed among women and adults aged < 60 years.

IR surrogate indices demonstrate strong associations with CircS. Notably, eGDR exhibited a protective effect with robust predictive value for incident CircS in longitudinal analysis. These readily calculable indices may be valuable for screening high-risk populations, particularly women and adults aged < 60 years.