This study was based on the China H-type Hypertension Registry Study (Registration number: ChiCTR1800017274); briefly, the China H-type Hypertension Registry Study is an ongoing, real-world, observational study conducted in Wuyuan, Jiangxi province of China. The full details with respect to the design and rationale of this study were extensively described elsewhere (21). Participants recruited in 2018 were consecutively enrolled if they: (1) were aged ≥18 years and (2) had hypertension, defined as seated resting blood pressure ≥140/90 mm Hg or self-reported use of antihypertensive medications. The participants were excluded when subjects with neurological abnormalities were unable to follow-up according to this study protocol under the Declaration of Helsinki conducted this study. Approval of this study protocol was provided by the Ethics Committees of the Biomedical Institute of Anhui Medical University (Approved No. of Ethic Committee: CH1059). Before each participant entered this study, they all signed the informed consent.

A total of 5,233 subjects had baPWV measurements taken at baseline. After excluding those with ankle-brachial indexes (ABI) <0.90 (n = 133) (22) or using lipid-lowering medications (n = 174), finally, 4,926 patients were selected for this study (Supplementary Figure 1).

Standardized questionnaires were used to obtain demographic factors, lifestyle behaviors (smoking and alcohol consumption), disease history, family history of diseases, and medication use at baseline. Current smoking was defined as smoking ≥1 cigarette per day for 1 year or more or a cumulative smoking amount ≥360 cigarettes per year. Alcohol consumption was defined as drinking an average of at least two or more times a week over a year. At the same time, we also collected the anthropometric indices included weight, height, WC, and hip circumference (HC) of all the participants. The BMI was calculated as the body weight in kilograms/square of the height in meters (kg/m²). Blood pressure (BP) was measured in the sitting position using an electronic sphygmomanometer (Omron; Dalian, China). After a 10-min rest period, BP was measured three times and the average of the three measurements was used for final analyses.

The baPWV (cm/s) and ABI were automatically measured simultaneously in the supine position after resting for more than 10 min, using an automatic waveform analyzer (BP-203RPE III device; Omron Health Care, Kyoto, Japan).

The specific measurement method is as follows: blood pressure cuff is wrapped on arms and ankles, the lower edge of the armband is located 2–3 cm above the cubital fossa transverse stripes, while the lower edge of the ankle band is located 1–2 cm above the medial malleolus. Furthermore, the baPWV value was calculated as the ratio of transmission distance from the brachium to ankle divided by the transit time:

Lb and La refer to the path length from suprasternal incision to the brachial muscle (Lb) and from suprasternal incision to the ankle (La), respectively. ΔTba is the ankle-brachial artery pressure wave foot time difference. baPWV was measured twice automatically and the higher value of the left and right sides was used in the final analysis.

After an 8–10 h fasting period, all the blood samples were taken from the cubital vein and delivered to the Biaojia Biotechnology Laboratory, Shenzhen, China. Automatic clinical analyzers (Beckman-Coulter Canada, Inc., Mississauga, Canada) were used to measure biochemicals included creatinine, uric acid, plasma homocysteine (Hcy), fasting total cholesterol (TC), TGs, high-density lipoprotein cholesterol (HDL-C), and fasting plasma glucose (FPG). LAP was determined from WC (cm) plus TG (mmol/l) for men [(WC-65) × TG] and women [(WC-58) × TG] (23); to prevent non-positive values for LAP, any male WC values of 65 cm or less were revised upward to 66.0 cm and any female WC values of 58 cm or less were revised upward to 59 cm (15). The estimated glomerular filtration rate (eGFR) was calculated by the Chronic Kidney Disease (CKD) Epidemiology Collaboration formula (24).

Based on quartiles of LAP levels, the study population was divided into four groups. The characteristics were presented as mean (SD) and categorical variables were expressed as count (percentage). We compared baseline characteristics on quartiles of participants of LAP levels by the one-way ANOVA for continuous variables and the chi-squared test or the Fisher's exact test for categorical variables, as appropriate. The distribution of LAP was strongly skewed toward the left. Thus, we performed the ln-transformed (lnLAP) on the LAP index before the analysis.

The predictive value of baPWV in cardiovascular events is unavailable. In this study, elevated baPWV is defined as a value >75% of baPWV value and >2,059 cm/s. The Pearson's correlation coefficient was used to assess the association of the LAP index with cardiovascular risk factors. Beta coefficients (β) and 95% CI used to investigate the association between LAP and baPWV were calculated using the multivariate linear regression analysis; odds ratios (ORs) and 95% CI used to investigate the association between LAP and elevated baPWV in hypertensive participants were calculated using the multivariate logistic regression analysis for five models. These five models are as follows: model 1 adjusted only for age and sex; model 2 adjusted for prior covariates and BMI, systolic blood pressure (SBP), diastolic blood pressure (DBP), pulse rate, pulse pressure, smoking status, and drinking status; model 3 adjusted for these prior covariates plus fasting blood glucose (FBG), TC, HDL, low-density lipoprotein cholesterol (LDL-C), Hcy, uric acid, eGFR, diabetes mellitus, antihypertensive drugs, and antiplatelet drugs; model 4 adjusted for these prior covariates plus the types of antihypertensive drugs and duration of hypertension; and model 5 adjusted for these prior covariates plus glucose-lowering drugs. Variables known as traditional risk factors for stroke and the potential confounders were selected, if the effect estimates individually changed by at least 10% (25). To characterize the shape of the dose-response relationship of the LAP index with baPWV and elevated baPWV, we performed a generalized additive model (GAM) and a fitted smoothing curve (penalized spline method). As additional exploratory analyses, possible modifications of the association between the LAP index and baPWV were also evaluated for variables including sex (males vs. females), age (<65 vs. ≥65 years), BMI (<25 vs. ≥25 kg/m²), current smoking (no vs. yes), current drinking (no vs. yes), SBP (<140, 140–159, ≥160 mm Hg), DBP (<90, 90–99, ≥100 mm Hg), diabetes mellitus (no vs. yes), and eGFR (<60 vs. ≥60 ml/min/1.73 m²). In order to ensure the robustness of data analysis, we conducted the following sensitivity analysis: (1) We have exclude 59 cancer patients and discussed the relationship between LAP and baPWV in non-cancer participants and (2) We evaluated the relationship between LAP and baPWV in men and women, respectively.

Data were analyzed using the Empower (R; www.empowerstats.com; X&Y Solutions Incorporation, Boston, Massachusetts, USA) and the statistical package (R) (http://www.R-project.org, The R Foundation). Effects that met the 5% significance level were considered as statistically significant, with all the tests being two-sided.

A representative sample of 4,926 participants of the China Hypertension Registry Study was eligible and enrolled in the analysis. From the study population, the mean age of those participants was 64.42 ± 9.44 years old and 2,440 (49.53%) people were men. The overall mean baseline LAP index was 40.65, ranging from 0.14 to 514.60 and mean baPWV was 1855.27 (414.58) cm/s. The baseline clinical characteristics of participants according to the quartile of the LAP index are given in Table 1. Compared with participants in the lowest quartile group, participants with the higher LAP index tended to be younger women, have diabetes mellitus and higher BMI, have a low prevalence of current smokers and drinkers, and have lower values of HDL and Hcy. Likewise, significant differences in biological parameters were observed among the groups. DBP, pulse rate, pulse pressure, FBG, TC, LDL, eGFR, and uric acid of the participants in the highest LAP index quartile were significantly higher than those in the first quartile. Moreover, the high LAP index groups had higher use of antihypertensive drugs, antiplatelet drugs, and glucose-lowering drugs (p < 0.05). In different genders, we describe the baseline characteristics according to the quartile of the LAP index (Supplementary Tables 3, 4). Supplementary Table 3 shows that there is no significant difference in Hcy and SBP among men in LAP four groups, but there are significant differences in other variables. Supplementary Table 4 shows that there is no significant difference in current drinking, SBP, Hcy, eGFR, and antiplatelet drugs among women in LAP four groups, but there are significant differences in other variables.

Table 2 shows the results of the Pearson's correlation analysis between the LAP index and cardiovascular risk factors. After age and sex adjustment, the LAP index was significantly correlated with BMI, SBP, DBP, FPG, TC, HDL-C, and LDL-C (p < 0.05).

Overall, the LAP index was found to be positively associated with baPWV in models 1–5. Values are the regression coefficients (β) of association, with model 1 adjusted only for age and sex; model 2 adjusted for prior covariates and BMI, SBP, DBP, pulse rate, pulse pressure, smoking status, and drinking status; model 3 adjusted for these prior covariates plus FBG, TC, HDL, LDL-C, Hcy, uric acid, eGFR, diabetes mellitus, antihypertensive drugs, and antiplatelet drugs; model 4 adjusted for these prior covariates plus the types of antihypertensive drugs and duration of hypertension; and model 5 adjusted for these prior covariates plus glucose-lowering drugs. Table 3 shows that in models 1–5, according to the β coefficients, for every 1-unit increase in ln LAP, the increases in baPWV were 19.81 (95% CI: 9.51, 30.11 cm/s), 42.39 (95% CI: 30.56, 54.22 cm/s), 25.06 (95% CI: 10.91, 39.22 cm/s), 24.12 (95% CI: 5.73, 42.51 cm/s), and 24.10 cm/s (95% CI: 5.72, 42.49 cm/s), respectively. As possible confounding factors were removed, the association became more significant. We also converted lnLAP from a continuous variable to a quartile variable. Compared with participants in the first quartile group, participants in quartile 2, quartile 3, and quartile 4 of the LAP index tended to have increased β coefficients of baPWV (p < 0.001) (Table 3). The results showed that LAP and baPWV had a positive linear correlation and were independent of cardiovascular risk factors.

As shown in Table 4, in a fully adjusted model 5, a positive association between the LAP and the risk of elevated baPWV was still existed. For every 1-unit increase in lnLAP, the adjusted ORs of elevated baPWV for participants in models 1–5 were 1.17 (95% CI: 1.09, 1.25), 1.42 (95% CI: 1.28, 1.58), 1.24 (95% CI: 1.09, 1.40), 1.19 (95% CI: 1.01, 1.41), and 1.19 (95% CI: 1.01, 1.41), respectively. When we converted lnLAP from a continuous variable to a quartile variable, the adjusted OR of elevated baPWV for participants in quartiles 2–4 was 1.41 (95% CI: 1.01, 1.96), 1.79 (95% CI: 1.20, 2.67), and 2.05 (95% CI: 1.28, 3.28) compared with participants in quartile 1 (p < 0.001). A generalized additive model and penalized spline method were used to further analyze the dose-response association between lnLAP and baPWV and elevated baPWV (Figures 1A,B), showing a significant positive linear association of the LAP index with baPWV and elevated baPWV. We conducted a sensitivity analysis to explore the association of the LAP index with baPWV and elevated baPWV in non-cancer participants and found that the results remained stable (Supplementary Tables 1, 2). Because there are gender differences in the calculation of the LAP index, we conducted a sensitivity analysis to explore the association of the LAP index with baPWV and elevated baPWV in different genders, the results are consistent with those in the general population (Supplementary Tables 5, 6).

To assess the relationship between lnLAP (per 1-unit increment) and baPWV in various groups, we performed exploratory subgroup analyses (Figure 2). There were no significant interactions in the following of the subgroup variables included sex (male vs. female; p-interaction = 0.808), age (<65 vs. ≥65 years; p-interaction = 0.587), BMI (<25 vs. ≥25 kg/m²; p-interaction = 0.568), current smoking (no vs. yes; p-interaction = 0.061), current drinking (no vs. yes; p-interaction = 0.861), SBP (<140, 140–159, ≥160 mm Hg; p-interaction = 0.283), diabetes mellitus (no vs. yes; p-interaction = 0.894), and eGFR (<60 vs. ≥60 ml/min/1.73 m²; p-interaction = 0.366). After adjustment for age, sex, BMI, SBP, DBP, pulse rate, pulse pressure, smoking status, drinking status, FBG, TC, HDL, LDL-C, Hcy, uric acid, eGFR, diabetes mellitus, antihypertensive drugs, antiplatelet drugs, the types of antihypertensive drugs, duration of hypertension, glucose-lowering drugs, except for the stratifying variables (p for interactions >0.05). Only in different subgroups of DBP, there is interaction between the LAP index and baPWV (DBP <90, 90–99, ≥100 mm Hg; p-interaction = 0.006). We found that the relationship between the LAP and baPWV was more significant among participants with DBP ≤ 99 mm Hg, DBP <90 mm Hg [27.98 (95% CI: 8.17, 47.79)], DBP = 90–99 mm Hg [33.89 (95% CI: 7.74, 60.05)], DBP ≥ 100 mm Hg [−28.47 (95% CI: −66.11, 9.16)].