In this study, we examined the factors associated with the 5-year incidence of baPWV ≥ 1400 cm/s in Japan based on large-scale health examination data. The results showed that age, BP, FBS, TG, eGFR, and habitual drinking were factors significantly associated with the 5-year incidence of baPWV ≥ 1400 cm/s in men, whereas age, BMI, BP, and HR were significantly associated with the 5-year incidence of baPWV ≥ 1400 cm/s in women. Using these factors, we developed a risk-prediction score and risk prediction equation for each sex. The risk prediction score based on the abovementioned factors for the 5-year incidence of baPWV ≥ 1400 cm/s yielded an AUC = 0.68 for men and 0.71 for women. Furthermore, the risk prediction equation for the 5-year incidence of baPWV ≥ 1400 cm/s yielded an AUC = 0.71 for men and 0.77 for women, suggesting that the equation had better prediction power than that of the risk prediction score.

Hypertension, dyslipidemia, DM, smoking, obesity, CKD, age, and male sex are risk factors for atherosclerosis [12, 13, 25]. The Framingham risk score (FRS), which was developed for the American population, uses age, sex, SBP, LDL-C, HDL-C, and smoking as risk factors to predict the development of CHD [5]. In contrast, the Suita score, developed for Japanese subjects, includes age, sex, SBP, DBP, LDL-C, HDL-C, DM, smoking, and CKD as risk factors [26].

As an index of arterial stiffness, the simple and non-invasive baPWV has been used in clinical practice and large cohort studies [11]. Previous cross-sectional studies have reported that age, sex, SBP, DBP, HR, FBS, TG, and eGFR levels are significantly associated with elevated baPWV in men and women. Yamashina et al. reported that age, BMI, mean BP, FBS, HDL-C, TG, and smoking were significantly associated with baPWV ≥ 1400 cm/s in both men and women in a Japanese population aged ≥ 30 years.

Previous cross-sectional studies have reported that age is significantly associated with baPWV. In a cross-sectional study of 7881 Japanese health examination participants (mean age, 43 years, 4488 men and 3393 women), Tomiyama et al. reported that aging affected baPWV in both men and women, with the effect greater in women than in men [8]. We previously reported that baPWV was independently correlated with age in 1033 Japanese participants (567 men and 466 women) who underwent regular health examinations [27]. In the present longitudinal study, age was significantly associated with the 5-year incidence of baPWV ≥ 1400 cm/s in both men and women.

Cross-sectional and longitudinal studies have reported an association between BP and baPWV. We previously reported a strong correlation between the mean BP and baPWV in a cross-sectional study of 567 Japanese men and 466 Japanese women undergoing regular health examinations (men: r = 0.617, women: r = 0.753) [27]. In addition, in a longitudinal study of 1020 Chinese young adults aged 18–23 years (47.7% women) without cardiovascular disease who were followed up for up to 25 years, early stage 1 hypertension (130–139/80–89 mmHg) was compared with normotension in Chinese participants in middle adulthood. The risk of developing high-risk baPWV was reported to be 1.61 times higher in those with early stage 1 hypertension than in normotensive participants [15]. Our present 5-year longitudinal study demonstrated that normal-high BP (120–129/80-84 mmHg) and hypertension (≥ 130/ ≥ 85 mmHg) were significantly associated with a 5-year incidence of baPWV ≥ 1400 cm/s, as compared with normal BP ( < 120/ < 80 mmHg), in both men and women.

Cross-sectional and longitudinal studies have reported an association between HR and baPWV. In a cross-sectional study of 68 men and women (mean age 65.97 ± 9.90) living in China, the results demonstrated that changes in HR may affect the baPWV, and the baPWV values tended to be higher when HR accelerated [28]. In addition, in a longitudinal study of 1795 healthy Japanese individuals (mean age 39 ± 8 years) followed for 5–6 years, HR at the baseline examination and changes in HR during the follow-up period were significantly associated with the corresponding changes in baPWV during the study period [29]. Our 5-year longitudinal study revealed that HR is associated with the 5-year incidence of baPWV ≥ 1400 cm/s in only women.

Cross-sectional and longitudinal studies have reported an association between eGFR and baPWV. In a cross-sectional study of 1,368 men and women (mean age 58.1 ± 14.4 years) living in China, the baPWV values were significantly higher in patients at the CKD stage (eGFR < 60 mL/min/1.73 m2) and the early CKD stage (eGFR 60–80 mL/min/1.73 m2) [14]. In addition, in a longitudinal study of 8,045 Chinese (mean age 54 ± 12 years) followed for 5 years, participants with higher baPWV at baseline had greater declines in eGFR over time [30]. Our present 5-year longitudinal study revealed that CKD (eGFR < 60 mL/min/1.73 m2) is associated with the 5-year incidence of baPWV ≥ 1400 cm/s in men.

An association between the FBS and baPWV has also been reported in cross-sectional studies. A cross-sectional study of 601 drug-naïve healthy participants (men, 46.2%) in Korea reported that FBS was significantly associated with increased baPWV [31]. Another cross-sectional study of 232 men (mean age 65.2 ± 9.5 years) from a rural area in Japan reported that the normal FBS group had a baPWV of 1518 cm/s, whereas the group with diabetes (FBS ≥ 126 mg/dL) had a significantly higher baPWV of 1771 cm/s. Furthermore, multiple regression analyses, including age, SBP, total cholesterol (TC), and BMI, demonstrated that FBS was independently and significantly associated with baPWV in Japanese men [32]. However, the association between FBS and baPWV in longitudinal studies has not been reported. Here, we revealed that elevated FBS ( ≥ 126 mg/dL) is associated with the 5-year incidence of baPWV ≥ 1400 cm/s in men.

Cross-sectional and longitudinal studies have reported an association between lipid levels and baPWV. A cross-sectional study of 909 Chinese participants aged 24–84 years has reported that TG levels were significantly and positively associated with baPWV [33]. In addition, a longitudinal study of 659 Chinese men aged 18 years and older, followed for 4.1 years, examined the association of TC, TG, LDL-C, and HDL-C with elevated baPWV and found that baseline serum TG was independently associated with the incidence of elevated baPWV (baPWV ≥ 1400 m/s) in the normal baPWV (baPWV < 1400 m/s) at baseline [16]. However, their study used only lipid values and did not examine associations with other atherosclerotic risk factors. In our longitudinal study using multiple atherosclerotic risk factors, we found a significant association between TG and the 5-year incidence of baPWV ≥ 1400 m/s in men.

Cross-sectional and longitudinal studies have reported an association between alcohol consumption and baPWV. The baPWV was positively associated with alcohol consumption in Korean men aged 40 years and older, but no clear relationship was found in Korean women in a cross-sectional study [13]. In a longitudinal study of 4016 healthy male Japanese workers (mean age, 43 years), divided into 1306 non-drinkers, 1311 mild-moderate drinkers, and 1337 heavy drinkers, who were followed-up for 9 years, the mean baPWV of non-drinkers was 1306 cm/s, that of mild-moderate drinkers was 1311 cm/s, and that of heavy drinkers was 1337 cm/s, indicating that alcohol intake was significantly associated with elevated baPWV [17]. In our longitudinal study, habitual drinking was associated with the 5-year incidence of baPWV ≥ 1400 cm/s only in men but not in women.

An association between obesity and baPWV has also been reported in cross-sectional studies. In a Chinese study of 429 healthy volunteers (66% men) aged 18 years or older (mean age: 44 years), BMI levels were reported to be significantly associated with baPWV [34]. Furthermore, in a study of 3512 Japanese (1228 men and 2284 women), baPWV was significantly associated with BMI in women with obesity but not in non-obese women and men [35]. In our longitudinal study, BMI was significantly associated with the 5-year incidence of baPWV ≥ 1400 cm/s only in women, whereas this association was absent in men. As shown in Table 1, the percentage of habitual drinking differed significantly between men and women because eating and drinking habits may differ between both sexes.

The present study had some limitations. First, we recruited participants who underwent health examinations at a single facility in Japan and were interested in their health. Therefore, a participant selection bias could not be avoided. Second, our data were retrospectively collected during annual physical examinations, and the subjects were not continuously monitored. Therefore, we could not determine the onset of baPWV ≥ 1400 cm/s, which precluded us from performing a Cox regression analysis to account for time-to-event. Third, we did not assess the use of medications such as antihypertensive, antidiabetic, or antidyslipidemic agents after the initial examination. These medications must have affected subsequent changes in baPWV. Therefore, medication use should be evaluated during the follow-up period in the next study. Fourth, the data were not prospectively collected. In the future, multicenter prospective longitudinal studies should be conducted to examine the relationship between baPWV and risk factors for arterial stiffness. Fifth, prediction scores and equations were not validated. The predictive ability of these scores and equations for arterial stiffness should be validated in future prospective studies.