Across 4,823 cardiovascular-disease-free adults aged 45 or older (3,258 from the US Health and Retirement Study, 1,565 from the English Longitudinal Study of Ageing), latent class mixed models sorted estimated glucose disposal rate (eGDR) into four trajectories over three waves. Persistently low and declining trajectories carried higher incident CVD risk than a persistently high one in both cohorts. In the HRS, trajectory class improved model fit beyond cumulative average eGDR, but not the reverse. A declining-then-rising pattern fared better than steady decline.
eGDR is calculated from waist circumference, hypertension status and HbA1c, so it can be tracked without clamp studies or fasting insulin assays. The real contribution here is the idea that direction of travel matters, not just lifetime exposure. That would support repeated measurement over a one-off reading, and the apparent benefit of recovery is encouraging for anyone who improves waist size, blood pressure or glycemic control, though this study cannot show that lifestyle changes caused the recovery. Limitations are substantial: three waves is a thin basis for trajectory modeling, the ELSA estimates were imprecise and the recovery-group comparisons rest on small numbers, and eGDR's components overlap with established risk factors, so the signal may partly reflect them. Observational data cannot establish causation. This is a preprint that has not been peer reviewed, and the findings may change. Editorially, it is an incremental refinement of existing eGDR work that needs external replication before it informs practice.