Abstract / Summary
Factors associated with longitudinal progression of brachial-ankle pulse wave velocity (baPWV) are less well established than correlates of single-time-point measurements. We examined associations of apolipoprotein B (ApoB) relative to LDL cholesterol and diabetes with baPWV progression. Adults with at least three baPWV measurements were classified using individual slopes from a linear mixed-effects model. Risers had slopes > 10 cm/s/year, and maintainers had slopes from − 10 to 10 cm/s/year; participants with slopes < − 10 cm/s/year were excluded from the primary comparison. ApoB residual was defined as measured ApoB minus the value predicted from LDL cholesterol by linear regression. Multivariable logistic regression examined associations with riser status. Among 6,378 participants in the two-group cohort, 1,389 (21.8%) were classified as risers. In the main multivariable model, adjusted odds ratios were 1.13 (95% CI 1.03–1.23) per 1-SD higher ApoB residual and 1.41 (1.06–1.85) for diabetes. The positive eGFR association in the main model was not evident under an alternative slope definition. The ApoB residual model had the lowest Akaike information criterion among the lipid models examined. Hypertension and systolic blood pressure were associated with baseline baPWV but not independently with riser status. Higher ApoB relative to LDL cholesterol and diabetes were associated with baPWV progression in this selected screening cohort. The ApoB association was modest and varied across model specifications. These observational findings require replication and do not establish causality or clinical treatment thresholds.