Association Between Estimated Pulse Wave Velocity and Moderate-to-Severe Cerebral White Matter Hyperintensities

Background: Estimated pulse-wave velocity (ePWV) is a surrogate marker associated with arterial stiffness which is mathematically derived from age and mean blood pressure (MBP) to estimate PWV (a direct measure of arterial stiffness). We aimed to study whether ePWV was associated with moderate-to-severe cerebral white matter hyperintensities (MS-cWMH) and whether it provides information beyond age and blood pressure themselves. Methods: This was a cross-sectional study including 1011 healthy participants in South Korea for routine health examinations between March 2008 and December 2014, and the participants were divided into three groups based on ePWV tertiles. Logistic regression analysis and restricted cubic spline (RCS) analysis were used to investigate the relationship between ePWV and MS-cWMH. Because age and MBP are the deterministic components of ePWV, we performed additional incremental analyses comparing the discriminatory performance of age alone, MBP alone, age plus blood pressure, ePWV alone, and a conventional clinical model with and without ePWV (DeLong tests). Exploratory subgroup analyses were performed, and apparent discriminatory performance was evaluated by receiver operating characteristic (ROC) analysis. Results: The prevalence of MS-cWMH was 29.08%. After adjustment for potential confounders, each 1 m/s increase in ePWV was associated with a 55% increase in the odds of MS-cWMH (OR, 1.55; 95% CI, 1.40– 1.71). When ePWV was divided into tertiles, participants in the third tertile had 6.49-fold higher odds of MS-cWMH than those in the first tertile (OR, 6.49; 95% CI, 4.15– 10.16). Restricted cubic spline analysis suggested a predominantly linear association (overall p < 0.0001; p for non-linearity = 0.262). However, the association was almost entirely attenuated after adjustment for age alone (OR, 1.05; 95% CI, 0.90– 1.23), and a likelihood-ratio test indicated no incremental information of ePWV beyond age and MBP (p = 0.46). The ROC curve analysis showed that ePWV had moderate apparent discrimination for MS-cWMH (AUC, 0.742; 95% CI, 0.709– 0.775), which was lower than that of age alone (AUC, 0.770; DeLong p = 0.005) and of age plus MBP (AUC, 0.770; p = 0.002). An exploratory subgroup analysis revealed a significant interaction by diabetes mellitus status in the association between ePWV and MS-cWMH (p for interaction = 0.027), which warrants confirmation in independent cohorts. Conclusion: Higher ePWV was associated with prevalent MS-cWMH in stroke-free adults undergoing routine health examinations; however, this association was largely explained by age, the dominant component of ePWV, and ePWV provided no discriminatory value beyond age and blood pressure. ePWV should be interpreted as a convenient composite of age and blood pressure burden rather than as an independent marker of arterial stiffness. Prospective, externally validated studies incorporating directly measured PWV are needed to determine whether ePWV adds predictive information beyond its constituent variables.