Positive association between baseline brachial–ankle pulse wave velocity and the risk of new-onset diabetes in hypertensive patients

All participants were part of the CSPPT (clinicaltrials.gov identifier: NCT00794885). Detailed methods and major findings of the CSPPT trial have been described previously [19‐22]. Briefly, the CSPPT was a large, community-based, multi-site, randomized, double-blind, and actively controlled trial with a total of 20,702 participants in China. Eligible participants were men and women aged 45 to 75 years old who had hypertension, defined as seated, resting systolic BP (SBP) ≥ 140 mmHg or diastolic BP (DBP) ≥ 90 mmHg at both the screening and the recruitment visit, or who were on antihypertensive medications. The major exclusion criteria included history of physician-diagnosed stroke, myocardial infarction, heart failure, post-coronary revascularization, and/or congenital heart disease.

The present study is a post hoc analysis of the CSPPT on 3532 subjects with baPWV measurements at baseline. Of those, 2429 participants without peripheral artery occlusive disease (PAD) assessed by ankle–brachial index (ABI < 0.9) [6, 16], with complete data on fasting glucose at baseline, and with physician-diagnosed diabetes or use of glucose-lowering drugs during the follow-up or fasting glucose data at the exit visit, as well as who were free of diabetes (physician-diagnosed diabetes or using glucose-lowering drugs) and whose fasting glucose (FG) was < 126.0 mg/dL at baseline, were included in the final analysis (Additional file 1: Figure S1).

The parent study (the CSPPT) and the current study were approved by the Ethics Committee of the Institute of Biomedicine, Anhui Medical University, Hefei, China (Federalwide Assurance Number 00001263). All participants provided written informed consent.

Eligible participants were randomly assigned, in a 1:1 ratio, to one of two treatment groups: a daily oral dose of one tablet containing 10 mg enalapril and 0.8 mg folic acid (the enalapril–folic acid group), or a daily oral dose of one tablet containing 10 mg enalapril only (the enalapril group). Participants were followed up every 3 months. During each follow-up visit, blood pressure was measured; study drug compliance, concomitant medication use, adverse events and possible endpoint events were documented by trained research staff and physicians. The study drug compliance was calculated as the percentage of days taking the study drugs during the trial.

Baseline data collection was conducted by trained research staff according to a standard operating procedure. Each participant was interviewed using a standardized questionnaire designed specifically for this study. The question about socioeconomic status was phrased as follows, “How does your standard of living compare to others?” and a choice of three responses: bad, medium, and good was provided. The question about physical activity was phrased as follows, “How do you describe your daily physical activity level?” and a choice of three responses: low, moderate, and high was provided [23‐25].

Serum fasting glucose (FG), lipids, and creatinine levels were measured using automatic clinical analyzers (Beckman Coulter) at the core laboratory of the National Clinical Research Center for Kidney Disease, Nanfang Hospital, Guangzhou, China. Serum folate was measured at baseline by a commercial laboratory using a chemiluminescent immunoassay (New Industrial).

baPWV, calculated as the ratio of transmission distance from the brachium to the ankle divided by the transit time, was used in this study. Participants were asked to remain in the supine position for at least 5 min after which baseline baPWV was measured using an automatic waveform analyzer (form PWV/ABI, BP-203RPE; Omron-Colin, Japan) according to published guidelines. The details describing the method of obtaining baPWV measurements are published elsewhere [8, 16]. In brief, two, bilateral readings of baPWV measurements were simultaneously taken and the maximum reading from each side was used for the analysis.

The primary study outcome was new-onset diabetes, defined as physician-diagnosed diabetes, or use of glucose-lowering drugs during follow-up, or new onset FG ≥ 126.0 mg/dL at the exit visit.

The secondary study outcomes include: (1) physician-diagnosed diabetes, or use of glucose-lowering drugs during follow-up; (2) the change in FG, calculated as FG at the exit visit minus that at baseline. The analysis of change in FG included subjects without physician-diagnosed diabetes, or use of glucose-lowering drugs during the follow-up.

Baseline characteristics are presented as mean ± standard deviation (SD) for continuous variables and proportions for categorical variables. Differences in baseline characteristics by baPWV quartiles were compared using ANOVA tests, or Chi-square tests, accordingly. The relationship of baPWV quartiles (< 15.9, 15.9–< 17.9, 17.9–< 20.7, and ≥ 20.7 m/s) with new-onset diabetes (primary outcome), physician-diagnosed diabetes or use of glucose-lowering drugs during follow-up, and change in FG (secondary outcomes) were evaluated using multivariable logistic regression models, Cox proportional hazard regression models and generalized linear regression models, respectively, without and with adjustment for age, sex, study center, study treatment group, body mass index (BMI), heart rate, smoking, systolic blood pressure (SBP), fasting glucose, total cholesterol (TC), creatinine, and folate at baseline, as well as time-averaged SBP during the treatment period. As additional exploratory analyses, possible modifications on the association between baPWV and new-onset diabetes were also evaluated by stratified analyses and interaction testing.

A two-tailed P < 0.05 was considered to be statistically significant in all analyses. R software (version 3.4.3, http://​www.​R-project.​org) were used for all statistical analyses.

As illustrated in the flow chart (Additional file 1: Figure S1), a total of 2429 hypertensive participants of the CSPPT without diabetes at the baseline were included in the final analysis.

Baseline characteristics of the study participants by baPWV quartiles are shown in Table 1. The mean age of the participants was 59.7 (SD, 7.4) years; 1049 were men (43.2%). Mean baseline baPWV was 18.6 (SD, 3.7) m/s. baPWV levels were inversely associated with BMI and time averaged diastolic blood pressure (DBP) during the treatment period, and positively associated with age, heart rate, SBP, DBP, TC, fasting glucose, high-density-lipid cholesterol (HDL-C) at baseline, as well as time averaged SBP during the treatment period. Moreover, participants with higher baPWV seemed to have lower physical activity levels (Additional file 1: Table S1).

During a median follow-up duration of 4.5 years (IQR, 4.2–4.7 years), new-onset diabetes occurred in 287 (11.8%) participants.

Overall, there was a significant positive association between baPWV and the risk of new-onset diabetes (Fig. 1). Per SD increment (3.7 m/s), higher baPWV was associated with a 33% increase in the adjusted risk of new-onset diabetes (OR, 1.33; 95% CI 1.13, 1.56) (Additional file 1: Table S2). When baPWV was assessed as quartiles, the adjusted odds ratios for participants in the second, third and fourth quartiles were 1.63 (95% CI 1.06, 2.49), 1.87 (95% CI 1.20, 2.91) and 2.48 (95% CI 1.53, 4.03), respectively, compared with those in quartile 1 (P for trend < 0.001). A significantly higher risk of new-onset diabetes was found in participants in quartiles 2–4 (adjusted OR, 1.80; 95% CI 1.22, 2.65) compared with those in quartile 1 (Fig. 2, Additional file 1: Table S2).

Consistently, per each SD increment (3.7 m/s), higher baPWV was associated with a 61% increase in the adjusted risk of physician-diagnosed diabetes or use of glucose-lowering drugs during follow-up (HR, 1.61; 95% CI 1.05, 2.47) (Additional file 1: Figure S2, Table S3). Accordingly, a significantly higher increase of FG levels (FG at the exit visit minus that at baseline) was also found in participants in quartiles 4 (adjusted β, 2.89 mg/dL; 95% CI 0.12, 5.66) compared with those in quartile 1 of the baPWV levels among participants without physician-diagnosed diabetes, or use of glucose-lowering drugs during follow-up (Additional file 1: Figure S3, Table S4).

During the treatment period, participants with higher baPWV had a higher frequency in use of calcium channel blockers (CCB) or diuretics (Additional file 1: Table S5). However, further adjustment for two variables: the use of calcium channel blockers (CCB), and the use of diuretics during the treatment period, did not substantially change the association between baPWV and new-onset diabetes (Additional file 1: Table S6). Moreover, the similar results were also found in participants without the concomitant use of diuretics during the treatment period (Additional file 1: Table S7).

In the CSPPT, all participants used enalapril or enalapril- folic acid during the follow up. However, further adjustment for the study drug (enalapril or enalapril-folic acid) compliance during the trial did not substantially change the results (Additional file 1: Table S8). Furthermore, we have further adjusted for family history of diabetes, physical activity, alcohol consumption and socioeconomic status, the association between baPWV and new-onset diabetes also did not been changed materially (Additional file 1: Table S9). More importantly, the similar results were also observed with further adjustment for the change in BMI (calculated as BMI at the exit visit minus that at baseline) (Additional file 1: Table S10).

In addition, we further explored the relationship of pulse pressure (PP) (both baseline PP and time-averaged PP during the treatment period) with new-onset diabetes (Additional file 1: Table S11). Overall, there was no significant association between baseline PP and new-onset diabetes. However, there was a significant positive relationship of time-averaged PP with new-onset diabetes (quartile 3–4 vs. quartile 1; adjusted OR, 1.55; 95% CI 1.03, 2.33). Nevertheless, further adjustment for baseline PP and time-averaged PP during the treatment period did not substantially alter the positive association between baseline baPWV and the risk of new-onset diabetes (per SD increment; adjusted OR, 1.30; 95% CI 1.11, 1.61) (Additional file 1: Table S12).

None of other variables, including sex (P-interaction = 0.275), age (< 60 vs. ≥ 60 years; P-interaction = 0.836), BMI (< 24, 24–28, ≥ 28 kg/m²; P-interaction = 0.350), treatment group (enalapril vs. enalapril + folic acid; P-interaction = 0.135), heart rate [< 72 (median) vs. ≥ 72 bmp; P-interaction = 0.420], SBP (< 160 vs. ≥ 160 mmHg; P-interaction = 0.747), TC [< 5.2 vs. ≥ 5.2 mmol/L; P-interaction = 0.702], FG (< 100.8 vs. 100.8–< 126 mg/mL; P-interaction = 0.486), folate [< 7.7 (median) vs. ≥ 7.7 ng/mL; P-interaction = 0.552] at baseline, and time-averaged SBP (< 140 vs. ≥ 140 mmHg; P-interaction = 0.292) during the treatment period, significantly modified the association between baPWV and new-onset diabetes (Fig. 3).

Our results indicate that the traditional assumption that diabetes precedes arterial stiffness might need to be reconsidered. While some previous cross-sectional studies [13‐15] have reported a positive association between baPWV and the prevalence of diabetes, the prospective association between diabetes and arterial stiffness is however, still inconclusive. de Oliveira Alvim et al. [26] found no significant difference in PWV progression after a 5-year follow-up in a subset of diabetics compared to non-diabetics. In contrast, Ferreira et al. [27] reported that better glycemic control, together with reductions in blood pressure and heart rate, was inversely associated with PWV changes. These results show that the association between arterial stiffness and diabetes is likely bidirectional. Diabetes might be in part a consequence of vascular impairment [28].

Our study findings are supported by previous studies. First, in a study on patients with untreated essential hypertension [29], those with higher pulse pressure (PP) exhibited impaired insulin secretion, increased post-challenge glucose concentrations and greater glucose spikes (PGS) during 75 g oral glucose tolerance testing. Our current study also found that there was a significant positive relationship of time-averaged PP with new-onset diabetes. Moreover, the positive association between baPWV and new-onset diabetes was independent of baseline PP and time-averaged PP during the treatment period. These results suggested that baPWV may be a more accurate marker of arterial stiffness. Second, baPWV has been shown to be associated with endothelial dysfunction [30], hypertension [10], inflammation [31], triglyceride glucose (TyG) index [32], left ventricular hypertrophy (LVH) [12] and lower muscle tissue [33]. Previous studies have found that endothelial dysfunction [34], hypertension [35], inflammation [36], TyG index [37], and LVH [28, 38, 39] precede incident diabetes. Moreover, patients with type 2 diabetes and visceral fat accumulation usually had lower muscle quality [40]. Third, previous clinical trials in diabetic patients indicated that the control of hypertension and hypercholesterolemia is more effective than glucose-lowering therapy in reducing cardiovascular events [41]. Fourth, angiotensin II receptor blocker (ARB) was found to inhibit the progression of arterial stiffness independent of blood pressure reduction [42]. Accordingly, ARB treatment significantly reduced the incidence of diabetes in previous studies [43, 44]. Our results provide further evidence for the assumption that diabetes may be partly a disease of vascular origin. However, further studies are needed to verify this hypothesis.

The exact mechanisms underlying the relationship of baPWV with new-onset diabetes remain to be elucidated. Some previously proposed potential mechanisms are outlined here. The propagation of increased pressure and flow pulsations to the pancreatic bed may lead to pancreatic dysfunction [18]. At the same time, arterial stiffness leads to increased arterial pulse pressure and pulsatile shear, resulting in endothelial dysfunction and metabolic dysregulation [30]. Endothelial dysfunction and impaired endothelium-dependent vasodilation may exacerbate insulin resistance by limiting the delivery of glucose to key target tissues [34]. More studies are needed to confirm our findings, and to further investigating the underlying mechanisms involved in this association.