Associations between Blood Pressure Indices and Brachial–ankle Pulse Wave Velocity in Treated Hypertensive Adults: results from the China Stroke Primary Prevention Trial (CSPPT)

Brachial–ankle pulse wave velocity (baPWV), as a marker of arterial stiffness, has been demonstrated to be associated with blood pressure (BP) and onset of hypertension. However, little information is available on the associations between baPWV and BP indices [systolic BP (SBP), diastolic BP (DBP), pulse pressure (PP), mean arterial pressure (MAP)] in treated hypertensive patients. We aimed to assess the associations between BP indices and baPWV. In this cross-sectional study, 14,598 hypertensive patients from China Stroke Primary Prevention Trial (CSPPT) at the exit visit of the trial were analyzed. Elevated baPWV was defined as ≥18.3 m/s. Multivariate linear and logistic regression analyses were performed to evaluate the associations of BP indices with baPWV and elevated baPWV. Moreover, the smooth curve fitting (penalized spline method) was conducted. Multivariate linear regression analyses showed that continuous SBP, DBP, PP and MAP were independently and positively associated with baPWV (β = 0.081, 0.084, 0.078 and 0.115, respectively, all P < 0.001). Compared with controlled SBP group (<140 mm Hg), uncontrolled SBP (≥140 mm Hg) was significantly associated with higher baPWV [β = 2.234, 95% confidence interval (CI): 2.137–2.332]. Similarly, compared with controlled DBP group (<90 mm Hg), uncontrolled DBP (≥90 mm Hg) was significantly associated with higher baPWV (β = 1.466, 95%CI: 1.341–1.590). Multiple logistic analyses also showed that SBP, DBP, PP and MAP were significantly and positively associated with elevated baPWV (OR = 1.056, 1.049, 1.052, and 1.075, respectively, all P < 0.001). The fully-adjusted smooth curve fitting presented a linear association between BP indices with baPWV. In conclusion, among treated hypertensive patients, SBP, DBP, PP and MAP levels were independently and positively associated with baPWV and elevated baPWV, suggesting that baPWV might be a way to predict uncontrolled BP.

is regarded as a public health challenge worldwide and is the leading modifiable risk factor for CVD 7,8 . Recent studies have showed that the control rate of hypertension is still low, although the awareness and treatment of hypertension have increased 9 . Poor blood pressure (BP) control further lead to target organ damage 10 . Therefore, it is important to identify the factors that influence individual response to anti-hypertensive treatment.
Indeed growing evidence has established the associations of baPWV with BP and onset of hypertension [11][12][13][14][15][16] . However, obvious conflicting results could be found among those reports. For instance, several epidemiological studies reported that BP indices was positively associated with baPWV [11][12][13][14] . In contrast, some studies showed that diastolic BP (DBP) was not associated with baPWV 15 . In addition, some studies indicated that DBP was negitvely associated with PWV 16 . These conflicting results might be attributed to the differences in cohort characteristics, sample size, and adjustment of confounders. Most of these previous studies have only discussed the association between BP and baPWV in general population. Uncertainty remains regarding the associations between BP indices [systolic BP (SBP), DBP, pulse pressure (PP), mean arterial pressure (MAP)] and baPWV because of limited data in treated hypertensive patients.
Therefore, our study aimed to assess the associations between BP indices and baPWV in a large sample of populations with treated hypertension.

Methods
Subject population and design. Our present study was a subset of the China Stroke Primary Prevention Trial (CSPPT) 17 . Briefly, the CSPPT is a large community-based, randomized, multicenter, double-blind, and actively controlled trial with a total of 20,702 hypertensive participants, conducted from May 2008 to August 2013 in Jiangsu and Anhui provinces of China. It was designed to confirm that enalapril maleate and folic acid tablets combined was more effective in preventing stroke among patients with hypertension when compared with enalapril maleate alone. Participants were scheduled for a visit (the first visit) after a 3-week run-in treatment period, and then were followed-up every 3 months until completion of the trial. Each visit involved recording BPs, heart rate, trial medication compliance, concomitant use of other medications, adverse events, and study outcome events. Details regarding inclusion/exclusion criteria, treatment assignment and outcome measures of the trial have been described in previous publications 17 and shared on a related website (http://clinicaltrials.gov/ct2/show/ NCT00794885). The study was approved by the ethics committee of the Nanfang Hospital, Guangzhou, China. The patients provided a written informed consent under the premises of the original CSPPT trial, including the possibility of post hoc analyses. This trial was registered with Clinicaltrials.gov (#NCT00794885). All methods were performed in accordance with the relevant guidelines and regulations.
At the exit visit of the trial, each participant was accpeted BaPWV measurement. In total, 20,702 eligible hypertensive participants were enrolled. Next, we selected participants according to the exclusion criteria listed below: missing baPWV values (n = 5504), missing BP values (n = 112), and ankle brachial index (ABI) < 0.90 (n = 488).

Data collection.
Brachial-ankle pulse wave velocity (baPWV). BaPWV was automatically measured by PWV/ABI instruments (form PWV/ABI, BP-203RPE; Omron-Colin, Japan) as previously described by trained volunteers from medical colleges 18 . Briefly, occlusion and monitoring cuffs matched with oscillometric sensors were wrapped around subjects' arms and the ankles, and pulse volume wave forms of the bilateral brachial and posterior tibial arteries were recorded simultaneously to determine the time interval between the initial increase in brachial and tibial waveforms (the transit time, Tba). The transmission distance from the brachium to ankle was calculated according to body height. The path length from the suprasternal notch to the brachium (Lb) was obtained using the following equation: Lb = 0.2159 × height of the patient (cm) − 2.0734. The path length from the suprasternal notch to the ankle (La) was obtained using the following equation: La = 0.8129 × height of the patient (cm) + 12.328. And the baPWV value was calculated as the ratio of transmission distance from the brachium to ankle divided by the transit time: baPWV = (La − Lb)/Tba. The maximum of the right and left-side baPWV values was used for analysis. The validation of this automatic device and its reproducibility have been previously published 2 .
BP and rest heart rate (RHR). BP and RHR was measured at the last follow up visit with the subject in the sitting position after having rested for more than 5 minutes, and using an electronic sphygmomanometer (Omron; Dalian, China). Three consecutive measurements were obtained on the right arm, with 1-minute intervals between each. Then SBP, DBP, and RHR were calculated as the mean of three independent measures. MAP was calculated as [(2 × DBP) + SBP] / 3. PP was calculated as SBP -DBP. Uncontrolled SBP was defined as a SBP of ≥140 mm Hg and uncontrolled DBP was defined as a DBP of ≥90 mm Hg among treated hypertensive individuals.
Covariables. We selected these covariates on the basis of their associations with baPWV. Continuous variables included age (years), body mass index (BMI, kg/m 2 ), serum homocysteine (hcy, μmol/L), fasting lipids [total cholesterol (TC, mmol/L), high-density lipoprotein-cholesterol (HDL-C, mmol/L), and triglycerides (TG, mmol/L)], fasting plasma glucose (FPG, mmol/L), creatinine (μmol/L), and uric acid (μmol/L). Categorical variables consisted of sex (male, female), Center (Lianyungang, Anqing), smoking (never smoking, former smoking or current smoking), alcohol consumption (never drinking, former drinking or current drinking), treatment group (enalapril, enalapril-folic acid), comorbidities (stroke, diabetes), medication use (combined with other antihypertensive drugs, lipid-lowering drugs and glucose-lowering drugs), methylenetetrahydrofolate reductase (MTHFR) C677T (rs1801133) polymorphism (CC, CT, TT). BMI was calculated as body weight in kilograms divided by the square of height in meters (kg/m 2 ). Laboratory data were measured using automatic clinical analyzers (Beckman Coulter) at the core laboratory of the National Clinical Research Center for Kidney Disease, Nanfang Hospital, www.nature.com/scientificreports www.nature.com/scientificreports/ Guangzhou, China. MTHFR C677T polymorphisms were detected on an ABI Prism 7900HT sequence detection system (Life Technologies) using the TaqMan assay. statistical analysis. Data are presented as mean ± SD for continuous variables and as frequency (%) for categorical variables. The population characteristics by baPWV quartiles were compared using ANOVA tests (continuous variables), or χ2 tests (categorical variables), accordingly. Taking BP indices (SBP, DBP, MAP, and PP) as independent variables and baPWV as an dependent variable, multivariate linear regression analyses were used to assess the β and 95% confidence interval (CI) of baPWV being associated with BP indices (SBP, DBP, MAP, and PP) adjustment for major covariables including age, sex, center, MTHFR C677T polymorphisms, treatment group, antihypertensive treatment, BMI, smoking status, alcohol consumption, RHR, TC, HDL-C, triglycerides, FPG, hcy, creatinine, uric acid. Meanwhile, Toshiaki Ohkuma et al. 19 proposed that the optimal cutoff value of baPWV for cardiovascular disease in patients with hypertension was 18.3 m/s. Therefore, the elevated baPWV was defined as ≥18.3 m/s. We investigated the associations of BP indices with elevated baPWV using multivariate binary logistic regression analysis. BP indices were evaluated using models for both continuous and categorical variables according to clinical normal values and quartiles. The lowest level of each BP indice was considered as a reference group. To ensure the robustness of data analysis, we also did the sensitivity analyses. We performed tests for linear trend by entering the median value of each quartiles of BP indices as continuous variables in the models. Moreover, to further characterize the shape of the associations between BP indices and baPWV, we used the smooth curve fitting (penalized spline method). Interaction and stratified analyses were performed to evaluate whether covariates influenced the associations between BP indices and baPWV.
All the analyses were performed using the statistical package R (http://www.R-project.org, The R Foundation) and Empower (R) (www.empowerstats.com; X&Y Solutions, Inc., Boston, MA). A 2-tailed P < 0.05 was considered to be statistically significant.

Results
Characteristics of the subjects. Based on the inclusion and exclusion criteria, a total of 14,598 participants (mean age: 64.4 ± 7.4 years; 40.2% males) at the exit visit were selected for final data analysis (  Table 1. Compared with participants with baPWV in the lower quartiles (Q1-Q3), participants with baPWV in the highest quartile (Q4) had higher values in age, RHR, SBP, DBP, MAP, PP, FPG, TC, TG, hcy, uric acid, creatinine and higher rates of stroke, diabetes, antihypertensive and glucose-lowering drugs (all P < 0.01). In contrast, populations in the highest baPWV group had significantly low BMI value. No significant differences were found  www.nature.com/scientificreports www.nature.com/scientificreports/ MAP were significantly and positively associated with baPWV (β = 0.081, 0.024, 0.093 and 0.079, respectively, all P < 0.001). In addition, after adjustment for all confounding factors including age, sex, center, MTHFR C677T polymorphisms, treatment group, antihypertensive treatment, BMI, smoking status, alcohol consumption, RHR, TC, HDL-C, TG, FPG, creatinine, hcy, and uric acid, all BP indices were independently and positively associated with baPWV (β = 0.081, 0.084, 0.078 and 0.115, respectively, all P < 0.001). BP indices were also evaluated for Patients with higher PP levels (≥40 mmHg) had higher baPWV (β = 1.888, 95%CI: 1.754-2.022, P < 0.001). Compared with normal MAP levels (70-105 mmHg), high MAP levels (≥105 mmHg) were positively associated with baPWV (β = 2.080, 95%CI: 1.973-2.187, P < 0.001), while low MAP levels (<70 mmHg) were negatively associated with baPWV (β = −3.224, 95%CI: −4.235-2.212, P < 0.001). When BP indices were classified into quartiles, we observed a significant and progressive increase in baPWV with BP quartiles (all P for trend < 0.001), suggesting a dose-dependent increase in baPWV with BP indices. Further analyses using smooth curve fitting (penalized spline method) confirmed that the associations of BP indices with baPWV were linear (Fig. 2). Table 3 shows the associations between BP indices and elevated baPWV. In adjusted model, SBP, DBP, PP and MAP were independently and positively associated with elevated baPWV in treated hypertensive patients [odds ratio (OR) = 1.056, 1.049, 1.052, and 1.075, respectively, all P < 0.001]. Compared with SBP control group, uncontrolled SBP was significantly associated with increased prevalence of elevated baPWV (OR = 4.514, 95%CI: 3.808-4.531, P < 0.001). Similarly, uncontrolled DBP was associated with increased prevalence of elevated baPWV (OR = 2.261, 95%CI: 2.038-2.508, P < 0.001) compared with DBP control group. Also, P for trend for all BP indices in the all models was significant, indicating that BP indices were positively associated with elevated baPWV. subgroup analysis. The role of other covariables on the association between BP indices and baPWV was further explored. Figure 3 shows the results of a subgroup analysis assessing the association of baPWV with SBP. Regardless of subgroup, SBP was positively associated with baPWV, and the effect was more significant in   Table 3. Associations between BP indices and elevated brachial-ankle pulse wave velocity Abbreviations: OR = odds ratio; CI = confidence interval; BP = blood pressure; SBP = systolic blood pressure; DBP = diastolic blood pressure; MAP = mean arterial pressure; PP = pulse pressure. * Adjusted for age, sex, center, MTHFR C677T polymorphisms, treatment group, antihypertensive treatment, BMI, smoking status, and alcohol consumption, RHR, TC, HDL-C, TG, FPG, creatinine, hcy, uric acid.

Discussion
In the present study, we found that SBP, DBP, PP and MAP levels were independently, significantly and positively associated with baPWV and elevated baPWV in treated hypertensive populations. The fully-adjusted smooth curve fitting showed a linear association between BP indices with baPWV. Moreover, uncontrolled SBP and DBP were significantly associated with higher baPWV. These findings suggest that baPWV might be a way to predict uncontrolled BP.  www.nature.com/scientificreports www.nature.com/scientificreports/ To our knowledge, few small sample population-based studies have addressed the relationship between BP indices and baPWV in treated hypertensive populations. Meili Zheng et al. 20 used data from 3056 treated hypertensive subjects and found that baPWV appeared to be an independent determinant of individual response to anti-hypertensive treatment, most notably for SBP (with estimated coefficients of −9.01 for the top quartile, as compared to the bottom quartile). Protogerou Athanase et al. 21 conducted a randomized, double-blind trial of 375 patients with hypertension and concluded that baseline PWV was a significant predictor of BP response to antihypertensive treatment, independent from age. Another study in patients with resistant hypertension indicated that a lower PWV reflected the predictors of higher BP reduction during spironolactone treatment 22 . Coutinho et al. 23 noted that PWV was associated with longitudinal increases in SBP, MAP and PP but not DBP in a community-based cohort of hypertensive individuals. Consistent with previous studies, our study included a large sample and further showed that BP indices (SBP, DBP, MAP and PP) were independently and positively associated with baPWV and elevated baPWV in hypertensive populations. We also found the linear relationship between BP indices and baPWV using the smooth curve fitting. Moreover, uncontrolled SBP and DBP were significantly associated with higher baPWV. The study raises the possibility that baPWV may serve as a simple and noninvasive measurement to identify hypertensive adults whose BP was not being well controlled and a novel therapeutic target to further reduce CVD risk. Further research is needed to clarify the issue.
According to STROBE statement 24 , subgroup analysis can make better use of data to reveal underlying truths. Our findings showed that BP indices were positively assoiated with baPWV in subgroups and some fators modified the associations. The results suggested that males showed the stronger association between BP indices and baPWV. Some potential mechanisms including sex hormones could help explain the sex differences in this association. Male was a common risk factor for hypertension and arterial stiffness 25,26 . In aging males, hypogonadism is associated with increased arterial stiffening in central and peripheral vessels, while testosterone replacement reduces PWV 27 . In addition, we found that age could modify the associations of baPWV with SBP and MAP. Some researchers have confirmed that arterial stiffness and BP are increasing with the aging process 28,29 . Shouling Wu et al. also found that arterial stiffness mediated the positive association between aging and BP 13 . However, although our results showed that the association between DBP and baPWV was weaker in elderly, there was no statistical difference in the interaction test between different age groups. Perhaps because DBP decreased with increasing age, which was a major determinant of increasing baPWV 30,31 . Interestingly, we found that the effect of BP indices on baPWV was more significant in the low BMI group, suggesting a negative relationship between BMI and baPWV 32,33 . This finding might be due to the characteristics of our study participants: the majority of whom were farmers. We suspected that BMI in this population might reflect muscle mass and physical activity rather than adiposity. This finding was also in accordance with prior research 34,35 . Compared to the findings of previous studies 25,32,36 , we found, moreover, that higher RHR still had a higher association between baPWV and BP indices, suggesting RHR was the common risk factor for BP and baPWV. This phenomenon might be related to sympathetic nerve activity. Furthermore, previous research reported that FPG and hcy were independent risk factors for baPWV 11,26 , the present study investigated a large community-based sample from China and further confirmed that those with the higher FPG and hcy had a stronger association between baPWV and SBP and MAP.
Limitations and strengths. Several limitations of the study are worth mentioning. This study was cross-sectional and failed to establish a cause-and-effect relationship between BP indices and baPWV. Additionally, the study participants were from a Chinese, rural hypertensive population aged 45 to 75 years old; thus, the generalizability of the results to other populations remained to be verified. Despite these limitations, this study was one of the largest to assess the associations between BP indices and baPWV in hypertensive populations. Our study also reported the linearity between BP indices and baPWV using smooth curve fitting (penalized spline method).

Conclusions
In summary, our study showed that BP indices (including SBP, DBP, PP and MAP) were independently and positively associated with baPWV in treated hypertensive patients. Moreover, linearity between BP indices with baPWV was found. These findings suggest that baPWV might be a way to predict uncontrolled BP in treated hypertensive patients.