Estimated pulse wave velocity and first-ever stroke risk in individuals with heart disease: findings from two prospective cohorts

Patients with heart disease are at high risk of first-ever stroke. Although estimated pulse wave velocity (ePWV) has been linked to stroke risk in general populations, its association with first-ever stroke in patients with heart disease remains unclear. This study evaluated the association between ePWV and first-ever stroke risk among patients with heart disease using two prospective cohorts. We included 9,486 UK Biobank participants with prevalent heart disease and no prior qualifying stroke. Cox proportional hazards and restricted cubic spline models were used to assess associations between baseline ePWV and first-ever stroke. Competing-risk and landmark sensitivity analyses assessed robustness. Prediction analyses compared a conventional age–blood-pressure model with an ePWV substitution model using repeated cross-validation and time-dependent area under the curve (AUC). The association was independently replicated in 521 participants with heart disease from the China Health and Retirement Longitudinal Study (CHARLS). In UK Biobank, higher ePWV was associated with increased risk of first-ever stroke in the primary multivariable-adjusted model (hazard ratio [HR] per 1 m/s, 1.20; 95% confidence interval [CI], 1.11–1.31). The association was approximately linear and remained robust in competing-risk, landmark, and age-excluded sensitivity analyses. In CHARLS, the direction of association was replicated in Model 4 (HR per 1 m/s, 1.30; 95% CI, 1.05–1.60), with an HR of 3.28 (95% CI, 1.31–8.23) for the highest versus lowest quartile (Q4 vs. Q1). Compared with the conventional model, the ePWV substitution model showed small improvements in discrimination in UK Biobank (difference in Harrell’s C statistic, 0.0053; difference in 5-year AUC, 0.0086; both Benjamini–Hochberg-adjusted P = 0.037). The 15-year AUC difference and the CHARLS comparisons were not statistically significant. Higher ePWV was associated with first-ever stroke risk among patients with established heart disease, with directional replication across two prospective cohorts. Its modest and inconsistent discrimination gains suggest that ePWV may provide complementary vascular ageing–related information but should not replace established risk factors.