The data used in this study were obtained from the China H-type Hypertension Registry Study (registration number: ChiCTR1800017274;20/07/2018), an ongoing real-world prospective cohort study in Wuyuan, China, which began in March 2018. This registration study has been previously described in detail (25). Briefly, patients >18 years of age are eligible for study participation study if they (1) have been diagnosed with hypertension (systolic blood pressure [SBP] ≥ 140 mmHg or diastolic blood pressure [DBP] ≥ 90 mmHg) or are taking antihypertensive drugs during the screening phase, and (2) provide informed written consent. The exclusion criteria are as follows: (1) mental or nervous system dysfunction or inability to express their will, (2) inability to complete the follow-up as required or planning to migrate to the field soon, and (3) patients considered by the investigator to be unsuitable for inclusion in the group or for long-term follow-up. The study protocol was approved by the Ethics Committees of the Institute of Biology at Anhui Medical University (No. CH1059) and the Second Affiliated Hospital of Nanchang University (No. 2018019).

At baseline, in March 2018, we examined the study population of 14,234 patients with hypertension. All patients were followed up until August 2022. After excluding two participants lost to follow-up, the final analysis included 14,232 patients with hypertension (Figure 1).

All participants underwent baseline and follow-up health assessments using a structured modified screening questionnaire to determine their demographic characteristics, including age, sex, physical activity, smoking history, drinking history, medical history, and medications. The anthropometric indices included height, weight, and waist circumference. Body mass index (BMI) was defined as body weight/height² (kg/m²). Trained medical staff assessed blood pressure (BP) to limit interobserver variability. After the participants rested for 5 min, seated BP and heart rate were measured using an electronic sphygmomanometer (Omron; Dalian, China) following the standard method and using appropriately sized cuffs. Three measurements on the right arm were performed at 1-min intervals between successive readings, and the mean value was calculated. Alcohol consumption was defined as drinking an average of two or more times per week over a year. Drinking habits were categorized as occasional (drinking alcohol monthly or less) or regular drinking (drinking alcohol at least twice per month). Current smoking was defined as smoking ≥1 cigarette per day for ≥1 year or a cumulative smoking amount of ≥360 cigarettes per year.

Blood samples were collected after the participants had fasted for 8–12 h. The samples were quickly processed to obtain serum and stored at −80°C until they were sent for analysis (Biaojia Biotechnology, Shenzhen, Guangdong Province, China). Automatic clinical analyzers (Beckman Coulter, USA) were used to measure all biochemical parameters including fasting plasma glucose (FPG), homocysteine (Hcy), triglyceride (TG), total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), serum albumin, and serum uric acid levels. The glomerular filtration rate (eGFR) was estimated using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation (26). Diabetes mellitus was defined as a self-reported physician diagnosis of diabetes, FPG concentration ≥7.0 mmol/L, or use of hypoglycemic drugs. Medical history of coronary heart disease (CHD) and stroke were self-reported by the participants primarily through the administration of questionnaires. Each participant was queried regarding the presence of corresponding symptoms, type of treatment received, and availability of pertinent medical records (including discharge summaries, biochemical examination data, and imaging data) during episodes of CHD or stroke. Atrial fibrillation (AF) was defined as the presence of AF on a standard 12-lead electrocardiogram or a documented history of AF.

The primary and secondary outcome measures of this longitudinal cohort study were all-cause and cardiovascular mortalities, respectively. Data on deaths were collected by three means from August 31, 2018–August 31, 2022. First, a professionally trained Clinical Research Coordinator (CRC) conducted telephone follow-ups and household surveys. Second, community hospitals in the affiliated jurisdictions of Wuyuan County provided death certificates. Third, we extracted the death information of the subjects from the Wuyuan County Medical Security Bureau, reviewed the endpoint events collected as described above by the endpoint committee from each participant's inpatient and outpatient medical records, and determined the type of endpoint event. The all-cause deaths included those attributed to (1) cardiovascular diseases, (2) malignant tumors, and (3) other causes (1). Cardiovascular disease was defined as a patient's medical history indicating a diagnosis by a class II hospital or above, including cardiovascular disease (myocardial infarction and heart failure) and stroke (cerebral infarction, cerebral hemorrhage, and subarachnoid hemorrhage) (27). Cardiovascular death was defined as death due to sudden cardiac death, myocardial infarction, heart failure, stroke, invasive cardiovascular surgery, or cardiovascular hemorrhage during the follow-up period (28).

The exposure variable was ePWV. We used a formula derived from the Reference Values for Arterial Stiffness' Collaboration (29), as described by Greve et al. (16) ePWV was calculated using age (years) and mean blood pressure (MBP, mmHg) as follows: ePWV = 9.587−0.402 × age + 4.560 × 10⁻³ × age²−2.621 × 10⁻⁵ × age² × MBP + 3.176 × 10⁻³ × age × MBP−1.832 × 10⁻² × MBP. MBP was calculated as DBP + 0.4 (SBP–DBP). In this study, 0.4 was used to calculate MBP instead of 0.33 as a coefficient of 0.4 is more consistent with the change in heart rate contour with age and its relationship with target organ damage (30, 31). In addition, baPWV was detected in selected populations in this study. baPWV (cm/s) was automatically measured simultaneously using an automatic waveform analyzer (BP-203RPE III device; Omron Health Care, Kyoto, Japan) with the participants in the supine position after resting for >10 min.

All data in this study were analyzed using Empower (R; www.empowerstats.com; X&Y Solutions, Inc, Boston, MA, USA) and the R language (http://www.R-project.org, The R Foundation). For all analyses, two-tailed P < 0.05 was considered statistically significant.

We used means with standard deviations and medians with interquartile ranges to analyze continuous variables with normal and non-normal distributions, respectively. Categorical variables are presented as frequencies and percentages. The study population was divided according to ePWV quartiles. Analysis of variance (ANOVA) and the chi-square test were used to explore differences in quartile characteristics at the ePWV level for continuous and categorical variables, respectively. The proportionality of the hazards assumption was assessed using the Schoenfeld residuals technique. If the proportional hazard assumption was satisfied (p = 0.29 for testing departures from proportionality), Cox proportional hazard models were used to estimate the associations of ePWV and baPWV with mortality after adjusting for study covariates. We constructed three models: Model 1 was unadjusted. Model 2 was adjusted for sex, BMI, and heart rate. Model 3 was additionally adjusted for diabetes; stroke; CHD; AF; smoking status; drinking status; Hcy, FPG, TC, TG, uric acid, HDL-c, LDL-c, serum albumin, and eGFR levels; and antihypertensive, hypoglycemic, lipid-lowering, and antiplatelet drug use. The dose-response relationship between ePWV and cardiovascular and all-cause mortality was plotted using a generalized additive model and smooth-fitting curves. Since we tested baPWV levels in 5,232 patients at baseline to ensure that the data were unbiased, we compared the predictive power of ePWV and baPWV in these populations. The predictive values of ePWV and baPWV for cardiovascular mortality and all-cause mortality were compared using the area under the curve (AUC) of the receiver operating characteristic (ROC), net reclassification improvement (NRI), and integrated discrimination improvement (IDI). In the Cox proportional risk regression model, the risk duration was defined as the time from study initiation to cardiovascular or all-cause deaths. The Kaplan–Meier curve (log-rank test) was used to evaluate cumulative cardiovascular and all-cause mortalities according to ePWV quartile. In addition, we performed a stratified analysis interaction test and evaluated the presence of modifying factors that altered the effect of the ePWV on all-cause and cardiovascular mortalities. The robustness of the results was evaluated using several sensitivity analyses.

Our analysis of data from a longitudinal cohort of patients with hypertension in China revealed a significant association between ePWV and cardiovascular and all-cause mortalities. As ePWV increased, the risks of cardiovascular and all-cause mortalities gradually increased. This relationship was independent of traditional cardiovascular risk factors. The results of the subgroup analyses demonstrated heterogeneous results attributed to varying outcomes, in which BMI significantly modified the positive correlation between ePWV and all-cause mortality. Additionally, high ePWV was associated with a significant increase in the risk of cardiovascular mortality among patients who did not receive hypoglycemic drugs. Finally, our findings indicated that ePWV outperformed baPWV in predicting both cardiovascular and all-cause mortalities.

Since its introduction by Greve et al. in 2016, the ePWV has attracted significant attention among researchers who have investigated its correlations with cardiovascular disease and all-cause mortality and compared it to conventional risk factors. Ji et al. conducted a prospective cohort study involving 98,348 individuals undergoing physical examinations in the Kailuan cohort to assess the association between ePWV and cardiovascular events as well as all-cause mortality (mean EPWV: 9.11 ± 2.01 m/s). These findings demonstrated the significant relationship between ePWV and cardiovascular disease and all-cause mortality risks even after adjusting for traditional cardiovascular risk factors (22). Moreover, Liu et al. utilized data from the China Longitudinal Study on Health and Retirement (CHARLS), comprising 13,116 older adult individuals in China, to evaluate the impact of baseline and dynamic ePWV levels during follow-up. Following a median follow-up period of seven years, the findings revealed that elevated baseline ePWV and increasing ePWV levels during follow-up were associated with an increased risk of all-cause mortality in this population (21). Hsu et al. (23) investigated the relationship among ePWV, baPWV, and cardiovascular mortality in a cohort of 881 individuals undergoing physical examinations in Taiwan. Assessment of the comparative predictive abilities of these measures for mortality revealed that both ePWV and baPWV demonstrated significant prognostic value for cardiovascular mortality; however, ePWV exhibited superior predictive ability. Moreover, this previous study assessed these predictive capacities for cardiovascular mortality only in individuals without pre-existing conditions. Thus, the outcomes may not be universally applicable because of the limited sample size, with potential variations in sensitivity across different disease states. Hsu et al. conducted a prospective cohort study to assess the impact of ePWV on cardiovascular and all-cause mortalities in 187 patients with myocardial infarction. The findings revealed that, following a median follow-up period of 73 months, ePWV demonstrated an independent predictive ability for cardiovascular mortality after accounting for other risk factors. However, its predictive ability for all-cause mortality was observed only in univariate regression analysis (24).

In their prospective cohort study of 33,930 participants from the National Nutrition Cohort Study, Cheng et al. observed that ePWV was an independent risk factor for all-cause and cardiovascular mortalities among American adults and remained unaffected by traditional cardiovascular risk factors following a median follow-up period of 133 months. Moreover, this index exhibited a robust predictive value for both all-cause and cardiovascular mortalities, surpassing the performance of the Framingham Risk Score (Framingham) and Combined Cohort Equation (PCE) (32). Greve et al. demonstrated that the ePWV exhibited a higher predictive value in untreated than in treated patients with hypertension (16).The secondary analysis of the SPRINT study further demonstrated that ePWV exhibited a robust predictive value in patients with hypertension and effectively reflected the impact of intensive blood pressure management on arterial stiffness. Among 8,450 patients with hypertension with a median follow-up duration of 3.2 years in the SPRINT study, Charalambos et al. (17) demonstrated that the ePWV surpassed the Framingham Risk Score in accurately predicting cardiovascular events, thereby suggesting its supplementation to enhance the accuracy of cardiovascular disease risk prediction. Moreover, the group receiving intensive blood pressure therapy showed a notable decrease in the ePWV progression rate, leading to a relative reduction in cardiovascular event occurrence. Prospective cohort studies conducted across diverse racial populations have consistently demonstrated that the ePWV can serve as an independent prognostic indicator for overall mortality (18–20). Combined, these findings underscore the significance of ePWV as an autonomous predictor of cardiovascular events and mortality risk.

We validated our research hypotheses through a prospective study of the Wuyuan hypertension cohort, thereby expanding the research population and broadening the applicability of the ePWV index. Additionally, we provided two novel insights. First, we demonstrated that the ePWV outperformed the baPWV in predicting cardiovascular and all-cause mortalities among patients with hypertension. Specifically, the ePWV exhibited superior prognostic capability for assessing the risk of all-cause and cardiovascular deaths in individuals with hypertension. The comparison of these measures contributes to understanding the feasibility of utilizing PWV as a substitute index without requiring specialized equipment. Furthermore, we observed heterogeneity in mortality outcomes among various subgroups. In particular, patients with ePWV and BMI ≥ 25 kg/m² exhibited a more pronounced risk of all-cause mortality. This could be attributed to their increased susceptibility to metabolic disorders and the confounding effects of cardiovascular disease risk factors. Epidemiological studies have consistently demonstrated a significant association between overweight and obesity and an elevated all-cause mortality risk (33, 34). Therefore, it is imperative to emphasize weight loss and BMI reduction among patients with hypertension who are also overweight or obese to mitigate the risk of mortality associated with increased ePWV. Furthermore, the association between ePWV and cardiovascular mortality was more pronounced among individuals not taking hypoglycemic medications. This suggests that ePWV holds substantial significance as an independent risk indicator of cardiovascular mortality in this population. Furthermore, hypoglycemic agents can mitigate cardiovascular mortality risk by ameliorating glucose metabolism disorders. Metabolic disturbances or impaired vascular insulin sensitivity may alter arterial wall function and structure, leading to increased arterial stiffness (35). Therefore, the adverse effects of ePWV on cardiovascular mortality in patients taking hypoglycemic drugs were partially mitigated. However, previous studies have not reported such interactions; thus, the impact of hypoglycemic drugs on arterial stiffness progression and patient prognosis requires further study. These findings underscore the significance of the ePWV in predicting overall mortality, as well as disease-specific and cardiovascular mortalities, particularly among individuals who are overweight or obese and those not taking hypoglycemic medications. The assessment of ePWV can provide clinicians with crucial insights into the risk of cardiovascular and all-cause mortalities, aiding in the early identification and intervention of high-risk individuals to enhance patient prognosis and health outcomes.

While the biological mechanisms underlying the association between ePWV and cardiovascular and all-cause mortality remain unclear, we propose several possible explanations. First, the ePWV is a calculated index using age and average blood pressure, which provides an accurate reflection of the severity of vascular aging. Vascular aging involves multiple factors, including oxidative stress, chronic inflammation, and cellular senescence (36). Cardiovascular and non-cardiovascular diseases that lead to related deaths follow the same pathological process (37). An elevated ePWV may serve as an indicator of arterial stiffness and vascular dysfunction, which reflect abnormalities in vascular function. As a hallmark of aging, arterial stiffness is a crucial manifestation of vascular senescence (38). The development of arterial stiffness is characterized by the deposition of collagen and calcium, or the degradation of elastin induced by hemodynamics, leading to alterations in arterial wall thickness and function, as well as changes in elastin content. These modifications impact vascular tension and compliance, thereby emerging as important determinants of multi-organ injury (39, 40). These pathological mechanisms are intricately linked to the overall health status of patients and their susceptibility to cardiovascular diseases. Furthermore, an increase in arterial stiffness results in a heightened exposure of target organs to pulsatile hemodynamics, potentially leading to barotrauma-induced injury. This phenomenon is a potential intermediary factor contributing to increased mortality risk (41).

The strengths of the present cohort study include the large-scale population of patients with hypertension in China, prospective design, extended duration, high follow-up rate, reliable and comprehensive questionnaire survey, and standardized physical examination. These factors significantly enhanced the accuracy of the ePWV calculations. This investigation holds substantial public health significance for the management of patients with primary hypertension, particularly in rural regions of China, where the prevalence of hypertension is generally high and awareness and treatment rates remain relatively low. Moreover, given the unfavorable prognosis among patients in these areas, with limited resources available for the comprehensive management of hypertension, our findings demonstrate that a precise estimation of ePWV can be achieved by considering age and blood pressure levels alone. The effective monitoring of ePWV in patients with hypertension and timely interventions can help mitigate cardiovascular events and enhance the prognosis of hypertension, thereby demonstrating its robust clinical applicability. Recent research also highlights that enhancing comprehensive nursing care for patients with hypertension in low- and middle-income countries and regions can significantly alleviate the burden of cardiovascular diseases and ameliorate the impact of socioeconomic inequality (42). The findings of this study have significant implications for the development of region-specific health strategies under resource constraints, thereby contributing to improving patients' overall well-being.

When interpreting these results, the following limitations must be acknowledged. First, only baseline ePWV values were obtained, whereas multiple measurements of ePWV can provide more precise data on arterial stiffness. Secondly, the effects of heart rate on blood pressure and pulse wave velocity should also be considered. We accounted for this by adjusting for heart rate, thereby mitigating any associated effects. Third, owing to the progressive nature of arteriosclerosis, PWV cannot predict short-term cardiovascular risk factors; however, it holds significant potential for estimating arterial stiffness in the medium and long term. Therefore, ePWV-based estimation of arterial stiffness demonstrates considerable long-term value. As this study included only patients with hypertension, further validation of these findings across a diverse population is required.

In conclusion, ePWV was superior to baPWV as an independent predictor of cardiovascular and all-cause mortalities. Higher ePWV significantly increases the risk of all-cause mortality in overweight and obese patients, and the relationship between higher ePWV levels and increased risk of cardiovascular death is more significant in patients who do not take hypoglycemic drugs. In fact, considering the management and prevention of hypertension in rural areas of China, it is more necessary to find indicators as easily available as ePWV to predict cardiovascular events and deaths in high-risk patients with hypertension. This has important public health significance for chronic disease management in primary medical care.