Abstract
BACKGROUND: Determining the risk factors for brachial-ankle pulse wave velocity (baPWV) may help to identify people susceptible to diabetic atherosclerosis and could prevent diabetic macrovascular complications in the early stages. We aim to comprehensively investigate risk factors contributing to arterial stiffness in patients with and without diabetes. METHODS: BaPWV was measured in 5651 individuals who attended health check-ups at baseline and follow-up. Lasso regression was used to screen for risk factors. Mixed models and multiple linear regressions were subsequently established to evaluate the effect size of the potential risk factors on baPWV and PWV change rates. All analyses were stratified by diabetes. Mediation analysis was also conducted to demonstrate the mechanisms of arterial stiffness in patients with diabetes. RESULTS: In lasso regression, postprandial 2-h glucose (P2hG), systolic blood pressure (SBP) and age were associated with baPWV regardless of diabetes. Platelet counts (PLT), mean corpuscular volume (MCV) and coronary heart disease (CHD) were associated with baPWV in patients with diabetes. In the mixed models, PLT were positively associated with baPWV in patients with diabetes (β(platelet, perSD) = 25.80; 18.26-33.33). Elevated PLTs could also significantly increase the PWV change rate in patients with diabetes (β(platelet, perSD) = 54.05; 10.00-107.10). In mediation analysis, diabetes had a significant average direct effect on baPWV. The average causal mediation effect (ACME) of PLTs was 1.76, with a range of 0.17 to 3.70. CONCLUSIONS: Elevated PLT counts can increase baPWV in diabetes and are a potential mediator between diabetes and atherosclerosis.