Association between eGDR dynamic trajectories and depression risk: findings from a nationwide, population based,...

Depression and insulin resistance (IR) are highly comorbid. Previous studies, however, have relied on static, single-point IR assessments, failing to capture the prognostic significance of its dynamic changes. This study aims to investigate the association between longitudinal trajectories of the estimated glucose disposal rate (eGDR) and new-onset depression risk.

We analyzed data from adults aged ≥45 years without baseline depression from China Health and Retirement Longitudinal Study (CHARLS). eGDR was calculated from waist circumference, hypertension, and HbA1c. Trajectories were identified using K-means clustering. Depression was defined as a CES-D score ≥ 10. Cox models assessed depression risk. Restricted cubic splines evaluated nonlinear relationships of baseline eGDR and ΔeGDR with depression risk.

Among 4380 participants, three eGDR trajectories were identified: rapid decline (11.5 %), consistently low (25.9 %), and consistently high (62.6 %). Compared to the persistent low trajectory, the rapid decline and persistent high trajectories were associated with a 23.2 % increased (HR = 1.232, 95 % CI: 1.020-1.489) and a 16.2 % decreased (HR = 0.838, 95 % CI: 0.723-0.971) risk of depression, respectively. Both the rapid decline and consistently low trajectories represent higher-risk phenotypes compared to the consistently high trajectory. RCS revealed an inverted U-shaped association between baseline eGDR and depression risk (P < 0.001), with the peak risk located at eGDR = 8.634 (HR = 1.44). Similarly, ΔeGDR showed a U-shaped association with risk; a sharp decline (ΔeGDR ≤ -4.81) was associated with an HR > 1.5.

Longitudinal eGDR trajectories are independent predictors of depression risk. Maintaining high insulin sensitivity is protective, while a rapid decline confers significant risk. Monitoring dynamic eGDR changes is crucial for identifying high-risk individuals and implementing preemptive interventions.