Association between healthy vascular aging and the risk of the first stroke in a community-based Chinese cohort

In this study we tested whether vascular aging is associated with the risk of first stroke in the Kailuan cohort, a community-based Chinese cohort. For participants aged ≥ 50 years, healthy vascular aging (HVA) was defined as an absence of hypertension and a brachial-ankle pulse wave velocity < the mean + 2 standard deviations, which was determined from a reference sample of healthy participants aged < 30 years. The primary outcome was first stroke (ischemic or hemorrhagic). In total, 11,474 participants were enrolled. The prevalence of HVA decreased from 36.0% in participants aged 50-59 years to 4.7% in those aged ≥ 70 years. During a median follow-up of 3.3 years, the incidence of first stroke was 0.5% in the HVA group but was 2.6% in the Non-HVA group. After adjusting for confounding variables, HVA was associated with a 0.32-fold lower risk of first stroke compared to the Non-HVA group (95% confidence interval, 0.18-0.56; p < 0.001). It thus appears that HVA reduced the risk of first stroke in a community-based Chinese population. This suggests that evaluation of vascular aging as part of public health screening may be useful for stroke risk assessment.

AGING arterial stiffness, has been shown to be associated with a reduced risk of cardiovascular disease in Western populations [12,13]. We therefore speculated that HVA may be related to a reduced risk of first stroke. To test that idea, we investigated the association between HVA and the risk of first stroke in community-based Chinese cohort.

Study participants
From January 1, 2010 to December 31, 2016, brachialankle pulse wave velocity (baPWV) measurements were made in 30,772 participants in the Kailuan cohort. Among them, 13,280 were older than 50 years. After excluding participants for whom blood pressure data was lacking (N=1112) and those with a history of cerebrovascular diseases before the baPWV measurement (N=694), a total of 11,474 participants were included in the study (Figure 1). Among that group, 3003 (26.2%) exhibited HVA at baseline, whereas the remaining 8471 (73.8%) did not. A comparison of demographic and clinical information between the HVA and Non-HVA groups is shown in Table 1.

The prevalence and determinants of HVA
The prevalence of HVA was 36.0% among participants aged 50 to 59 years, 18.2% among participants aged 60 to 69 years, and 4.7% among those aged ≥70 years ( Figure  2). Using a multivariable analysis, we found that younger age, female sex, lower body mass index (BMI), lower high-sensitive C-reactive protein (hs-CRP), lower fasting blood glucose (FBG), higher education level, active physical activity, absence of current alcohol drinking, and the absence of diabetes mellitus and hyperlipidemia were significantly associated with HVA (p<0.05) ( Table 2).

DISCUSSION
The main finding of this study was that HVA was significantly associated with a lower risk of first stroke in a community-based Chinese cohort. Although the relationship between arterial stiffness and blood pressure is complex [14][15][16][17][18], the combination of normal pulse wave velocity and an absence of hypertension may be meaningful for prediction of cardiovascular disease risk. HVA was previously evaluated using two easily obtained metrics of vascular health: blood pressure and pulse wave velocity [12]. Considering the racial and geographic differences between Western and Chinese populations, our study builds a Chinese reference sample and establishes normal baPWV values for defining HVA, instead of adopting the reported references from a Western community [19]. Comparing the results from the Framingham heart study with our findings, we find that the prevalence of HVA was higher in this Chinese population than in the Western population (26.2% vs. 17.7%), particularly for participants aged 60 to 69 years (18.2% vs. 7.4%). Data from both the Framingham and Kailuan studies show that the prevalence of HVA decreases sharply with age. HVA appears achievable but is rare among participants aged ≥ 70 years. This may be attributable to greater arterial stiffness and a higher prevalence of hypertension among elderly participants.
Our study also suggests that HVA status may be influenced by both non-modifiable (age and sex) and modifiable (BMI, hs-CRP, FBG, physical activity, alcohol consumption, education level, history of diabetes mellitus and hyperlipidemia) factors [20,21]. Notably, in  AGING a univariate analysis, the fraction of participants currently smoking was higher in the HVA group than in the Non-HVA group. However, we did not observe an independent association between smoking status and HVA after multivariable logistic regression, which is consistent with an earlier study [12]. In addition, some epidemiologic studies have observed that smokers might have lower rates of hypertension and decreased arterial stiffness [22,23]. Therefore, the impact of smoking on vascular health remains unclear and seems not straightforward.
The large sample size in our study increases its statistical reliability. In addition, all of the participants had access to local medical insurance, which guaranteed that the outcome data could be tracked through the medical insurance system. Nevertheless, several limitations should be noted. First, to evaluate the arterial stiffness, baPWV was adopted instead of carotid-femoral pulse wave velocity (cfPWV) [24][25][26]. However, because of the simplicity of its measurement and good correlation with cfPWV, we suggest baPWV could be potentially applied in clinical practice [27][28][29]. Second, participants in the Kailuan cohort volunteered to undergo baPWV measurement, which could potentially introduce selection bias. Third, since the Kailuan Group is a coal energy and chemical enterprise, males make up the majority of employees. All employees (≥ 18 years) were enrolled in the Kailuan cohort, which led to a gender imbalance among the participants in the analysis (males, 75%; females, 25%). Finally, because the numbers of incident strokes in this community-based population were limited, we did not conduct a stratification analysis based on age. In the future, the follow-up will go on to determine whether the effect of HVA is similar in different age groups.
In conclusion, it appears that HVA may have reduced the risk of first stroke in a community-based Chinese cohort. This suggests that evaluation of vascular aging as part of public health screening may be useful for stroke risk assessment and for stroke prevention.

MATERIALS AND METHODS
The data supporting the findings of this study are available from the corresponding author upon reasonable request.

Study participants
The Kailuan cohort is an ongoing prospective cohort based in the Kailuan community in Tangshan City in the north of China [30,31]. All employees (≥ 18 years) of the Kailuan Group, a coal energy and chemical enterprise, were enrolled with the aim of investigating the risk factors of various diseases (e.g., stroke and cancers). As of December 31, 2016, 171,414 participants were included in the Kailuan cohort. The study's protocol was approved by the participating hospitals' ethics committees. All participants provided written informed consent. All participants underwent questionnaire assessments, clinical examinations and laboratory tests at the 11 hospitals responsible for healthcare in the Kailuan community.
To capture a distinct sample of individuals with HVA, in the present study we considered only participants in late middle age or older (aged ≥ 50 years), and we only enrolled participants who underwent baPWV measurement between 2010 and 2016. We excluded participants for whom blood pressure data was lacking as well as those with a history of cerebrovascular diseases before the baPWV measurement.

Assessment of baPWV
BaPWV was measured using a BP-203 RPE III networked arteriosclerosis detection device produced by Omron Health Medical Co., LTD (China) [30,31]. Participants received baPWV assessment after 5 min of rest in a supine position. Cuffs were applied to both arms and ankles. The lower edge of the arm cuff was positioned 2-3 cm above the cubital fossa transverse striation, and the lower edge of the ankle cuff was positioned 1-2 cm above the medial malleolus. The heartbeat monitor was placed at the left edge of the sternum, and electrocardiogram electrodes were placed on both wrists. BaPWV was calculated as the distance between the two sites divided by the pulse transit time (defined as the time interval between the wave fronts of the brachial and ankle waveforms). The distance between the sampling points was calculated automatically according to the subject's height. The maximum value of the bilateral baPWV was used in our study. Using the detection device, we were able to read the baPWV value directly. The methodology for baPWV measurement was the same for all participants.
Baseline data on demographics and cardiovascular risk factors were collected through face-face interviews when baPWV was measured. The collected data included age, sex, BMI, smoking, alcohol consumption, physical activity, history of diabetes, history of hyperlipidemia, monthly income, education level, FBG, TC, LDL, hs-CRP, use of antidiabetic or lipid-lowering drugs.
Physical activity was evaluated from responses to questions about the type and frequency of physical activity at work and during leisure time. Inactive physical activity was defined as no physical activity or less than 30 minutes every week.

Assessment of HVA
We defined HVA for participants aged ≥ 50 years as follows [12]: 1) absence of hypertension, which was defined as a systolic blood pressure < 140 mmHg, a diastolic blood pressure < 90 mmHg (blood pressure was measured at least twice at different occasions), and an absence of antihypertensive drugs; 2) baPWV < mean + 2 standard deviations (SD), which was determined from a reference sample of healthy participants aged 18-30 years with no cardiovascular risk factors (e.g., hypertension, diabetes, smoking, hyperlipidemia) and without a history of myocardial infarction, stroke, and other cardiovascular disease. To obtain the reference value for baPWV, we enrolled 595 healthy participants aged < 30 years from the Kailuan study ( Figure 4). The results showed that the reference value for baPWV was 11.90 ± 1.65 m/s. Thus, HVA was defined as a normotensive blood pressure and baPWV < 15.20 m/s. Based on these data, participants were classified into HVA and Non-HVA groups.

Outcome assessment and follow-up
The participants were followed up by face-to-face interviews at every two-year routine medical examination until December 31, 2017 or to death. The outcome information for the participants without faceto-face follow-up was obtained by referring to death certificates from provincial vital statistics offices, discharge summaries from the 11 hospitals, and medical records from medical insurance. The followups were performed by hospital physicians, research physicians, and research nurses who were blinded to the baseline data. All stroke records were reviewed by two independent stroke specialists. If there were instances of disagreement in a single case, the final evaluation was made by the event adjudication committee.
The primary outcome was the first occurrence of stroke (both ischemic and hemorrhagic). These included the first nonfatal stroke event or stroke death without a preceding nonfatal event. Stroke was diagnosed according to the World Health Organization criteria [32] combined with brain computed tomography (CT) or magnetic resonance (MR) imaging for confirmation. The secondary outcome was the first occurrence of ischemic stroke. All study outcomes were validated by the Data Safety Monitoring Board and the Arbitration Committee for Clinical Outcomes.

Statistical analysis
Continuous variables are presented as the mean ± SD or median with interquartile range, while categorical variables are presented as frequencies with percentages. The nonparametric Wilcoxon or t tests were used to compare group differences for continuous variables and χ 2 tests were used for categorical variables.
The association between baseline characteristics and HVA was investigated using multivariate backward logistic regression, and variables with p values < 0.05 were selected as confounding variables. Odds ratios (ORs) with the 95% CI were reported. We also examined associations between HVA and first stroke or ischemic stroke using Cox proportional hazards regression models. For each outcome, 4 multivariable models were performed. In model 1, we only adjusted for age and sex. In model 2, we additionally adjusted for other cardiovascular risk factors (including BMI, smoking, alcohol consumption, physical activity, history of diabetes, history of hyperlipidemia). In model 3, we further adjusted for monthly income, education level, FBG, TC, LDL, hs-CRP, and use of antidiabetic or lipidlowering medications. In model 4, we further adjusted for systolic blood pressure. Life table methods were used to calculate the cumulative rate of stroke-free survival, and a curve was generated.
Two-sided p values < 0.05 were considered significant. All analyses were performed using SAS 9.4 (SAS Institute, Cary, North Carolina).

ACKNOWLEDGMENTS
We gratefully acknowledge all the participating hospitals in the study and thank all enrolled participants and their relatives.

CONFLICTS OF INTEREST
The authors declare that they have no conflicts of interest.

FUNDING
The study was supported by grants from the National Natural Science Foundation of China (81825007)