Brachial-ankle pulse wave velocity trajectories in a middle-aged population

ObjectiveThe "trajectory" phenotype was observed in several cardiovascular risk factors with aging. We aim to identify multiple brachial-ankle Pulse Wave Velocity (baPWV) trajectory phenotypes and assess their determinants.MethodsAmong 5,182 participants with baPWV measurements (2010-2016) at no less than three time points in Kailuan Study, we derived baPWV trajectory pattern using SAS Proc Traj program. We applied the lowest Bayesian information criterion to identify the best typing model, related the identified trajectory pattern to baseline and changes in characteristics.ResultsAmong 5.3 +/- 1.7 years follow-up, four distinct baPWV trajectories were identified as low (1,961,37.8%), medium-low (1,846,35.6%), medium-high (1,024,19.8%), and high (351,6.8%) groups. In the stepwise models, mean arterial pressure and age were the main determinators of the trajectory patterns, with a Delta pseudo-R-2 of 0.335 and 0.164, respectively. With the low trajectory group as reference and multivariable adjustment, odd ratios of medium low, medium high and high associated with 1 mmHg increment of mean arterial pressure were 1.08(95%CI: 1.07-1.09), 1.13(1.12-1.14), and 1.16(1.15-1.18). The estimates for age were 1.08(1.07-1.10), 1.20(1.18-1.21) and 1.28(1.26-1.31). Additionally, baseline resting heart rate, low-density lipoprotein cholesterol, fasting blood glucose, hypersensitive C-reaction protein and uric acid, and changes in mean arterial pressure, resting heart rate, fasting blood glucose, and uric acid were positively associated with the trajectory, while BMI was negatively associated.ConclusionsThe changes in baPWV overtime followed a "trajectory" pattern, mainly determined by mean arterial pressure and age.