Associations of long-term insulin resistance trajectories with incident cardiovascular disease by LDL-C status, with time-updated lipid control: a two-decade population-based study
F Esmaeili, M Tohidi, F Azizi, F HadaeghAbstract
Background/Introduction
Insulin resistance drives atherosclerotic CVD and often accompanies dyslipidaemia. While single measurements predict CVD, the risk conveyed by long-term insulin resistance trajectories, especially in people with controlled LDL-C, is less clear. Demonstrating residual risk from worsening insulin resistance despite favourable LDL-C could sharpen cardiometabolic risk stratification beyond LDL-focused prevention.
Purpose
To investigate the association between longitudinal HOMA-IR trajectory patterns and incident CVD, stratified by LDL-C status at index, and to assess whether the relationship persists after accounting for time-updated LDL-C category and statin use.
Methods
This trajectory analysis included 4,051 participants (1903 male) free of CVD at baseline with at least three repeated measurements of fasting insulin and fasting plasma glucose, enabling calculation of the HOMA-IR. HOMA-IR trajectories were derived over an exposure period ending at the index visit, and incident CVD events were subsequently ascertained during follow-up. Using group-based trajectory modelling, participants were classified into three HOMA-IR patterns: low stable, moderate stable, and rising. LDL-C status at index was categorised as controlled (<100 mg/dL) or not controlled. Associations between HOMA-IR trajectories and CVD incidence were assessed using multivariable regression, reported as ORs with 95% CIs, and presented separately by LDL-C stratum; effect modification was evaluated using a trajectory×LDL-C interaction term. In sensitivity analyses, LDL-C category and statin use were modelled as time-varying covariates. Analyses were adjusted for a large set of covariates at three models.
Results
During an average 10.1-year even accrual period, 280 participants developed incident CVD (145/2431 with LDL-C controlled; 135/1620 with LDL-C not controlled). In the fully adjusted model, the rising HOMA-IR trajectory was associated with higher CVD risk in both strata (LDL-C controlled: OR=1.35, 95% CI [1.07-2.34] p=0.03; LDL-C not controlled: 1.52[1.07-2.15] p=0.02), whereas the moderate stable trajectory was not (controlled: OR=1.14[0.89-1.82] p=0.06; not controlled: 1.18[0.82-2.37] p=0.07). There was no evidence of effect modification by LDL-C status (P for interaction in Model 3=0.42). Joint-phenotype analysis showed a graded risk increase, peaking in rising HOMA-IR with LDL-C not controlled (crude OR=5.05[3.10–8.20]). Findings were robust when LDL-C status and statin use were modelled as time-varying covariates (rising vs low-stable: OR=1.33[1.04-1.70] in LDL-C controlled; 1.48[1.09-2.01] in LDL-C not controlled).
Conclusion
A rising HOMA-IR trajectory identified higher CVD incidence risk even with LDL-C controlled at baseline, and remained robust after time-updated LDL-C and statin adjustment. Worsening insulin resistance signals substantial residual risk beyond LDL-focused prevention and warrants earlier targeted intervention.figure 1