Cardiorespiratory Fitness Suppresses Age‐Related Arterial Stiffening in Healthy Adults: A 2‐Year Longitudinal Observational Study

Abstract

Cardiorespiratory fitness is negatively associated with arterial stiffness, although it is unclear whether it is associated with prospective arterial stiffness changes. The authors examined cardiorespiratory fitness and arterial stiffness progression in a 2‐year follow‐up study of 470 healthy men and women aged 26 to 69 years. Peak oxygen uptake ($\dot{\mathrm{V}}{\mathrm{O}}_{2\mathrm{peak}}$) was measured at baseline using a graded cycle exercise test. Arterial stiffness was assessed using brachial‐ankle pulse wave velocity (baPWV) at baseline and after 2 years. Two‐year changes in baPWV were significantly higher in patients in the lowest $\dot{\mathrm{V}}{\mathrm{O}}_{2\mathrm{peak}}$ tertile (28.8±7.6 cm/s) compared with those in the highest $\dot{\mathrm{V}}{\mathrm{O}}_{2\mathrm{peak}}$ tertile (−1.4±7.5 cm/s) (P=.024) and were inversely correlated with $\dot{\mathrm{V}}{\mathrm{O}}_{2\mathrm{peak}}$ (r=−.112, P=.015). Stepwise multiple regression analysis revealed that age, glucose, baPWV, $\dot{\mathrm{V}}{\mathrm{O}}_{2\mathrm{peak}}$, and sex were independent correlates of 2‐year changes in baPWV, suggesting that higher cardiorespiratory fitness is associated with age‐related arterial stiffening suppression.

Arterial stiffness increases progressively with advancing age, even in healthy individuals.1 This arterial stiffening is associated with future hypertension2 and cardiovascular (CV) events3 and is recognized as a surrogate marker for CV disease. Therefore, the prevention of arterial stiffening is of great clinical importance.

Carotid‐femoral pulse wave velocity (cfPWV) is a standard method for assessing aortic stiffness.4 cfPWV is used in clinical practice mainly in Europe, and in the United States to a lesser extent. Recently, brachial‐ankle pulse wave velocity (baPWV) was proposed as an alternative method for assessing arterial stiffness in Asian populations.5 A major advantage of baPWV is its measurement method, which simply involves wrapping the four extremities in blood pressure (BP) cuffs.6, 7 Moreover, the use of either cfPWV or baPWV is accepted by the Japanese guidelines for the management of hypertension as a tool for assessing subclinical target organ damage.8 In addition, baPWV has been shown to be associated with an increased risk of total CV events and all‐cause mortality,9, 10 as is cfPWV.11

Cardiorespiratory fitness (CRF) is independently associated with a lower risk of all‐cause mortality and CV events.12 Thus, previous studies have investigated the relationship between CRF and arterial stiffness as a surrogate marker for CV disease in cross‐sectional research and have suggested that higher CRF was associated with lower arterial stiffness.13, 14, 15, 16 To our knowledge, only Ferreira and colleagues17, 18 have reported longitudinal research from adolescence to young adulthood. There is no information regarding middle‐aged and elderly populations. In addition, there has been no previous study on whether CRF is associated with the progression of arterial stiffness as assessed by baPWV in longitudinal research.

Previous studies have demonstrated that regular aerobic exercise is effective in preventing and reversing arterial stiffening in healthy adults.19 Regular aerobic exercise results in higher CRF,20 and, consequently, CRF may be associated with a lower CV disease risk. Therefore, we hypothesized that higher CRF would be associated with less progression of age‐related arterial stiffening in healthy adults. To test our hypothesis, we examined the relationship between CRF and the progression of arterial stiffening with a 2‐year follow‐up study.

Methods

Participants

The participants were recruited from the community around the National Institute of Health and Nutrition, Tokyo, Japan, or the Okayama Southern Institute of Health, Okayama Health Foundation, Okayama, Japan. We used data from 470 Japanese adults (128 men and 342 women; mean age, 48.8±9.5 years) selected from among 1125 participants who met the following criteria: (1) they received anthropometric, CRF, exercise habit, physical activity (n=459), and arterial stiffness assessments and underwent blood examinations (baseline measurement); (2) they underwent arterial stiffness measurement at 2‐year follow‐up; and (3) they had no history of stroke, cardiac disease, or chronic renal failure; they were receiving no medical treatment for hypertension, dyslipidemia, or diabetes; they were currently nonsmokers; and they had an ankle‐brachial pressure index between 0.9 and 1.3 at both baseline and the follow‐up visit (during the observation period). Ethical approval for the study was obtained from the Human Research Committees of the National Institute of Health and Nutrition and Okayama Health Foundation, and it was performed in accordance with the guidelines of the Declaration of Helsinki.

To assess the effects of CRF level on progression of arterial stiffening, the participants were categorized into tertiles based on their peak oxygen uptake ($\dot{\mathrm{V}}{\mathrm{O}}_{2\mathrm{peak}}$) levels during an incremental cycle exercise test using a cycle ergometer (described below) at baseline for each sex and the following age‐specific distributions (20–29, 30–39, 40–49, 50–59, and 60–69 years) (Table 1): lowest tertile, unfit; middle tertile, mid‐range fitness; and highest tertile, fit.

Table 1.

Age‐Specific Distributions for Cardiorespiratory Fitness Based on Peak Oxygen Uptake Tertile

		Peak Oxygen Uptake, mL/kg/min
	Age, y	Low	Middle	High
Men	20–29 (n=3)	39.5	40.4	44.0
	30–39 (n=38)	28.2–36.6	37.6–44.5	44.6–55.7
	40–49 (n=47)	18.8–30.4	30.5–39.4	39.6–54.3
	50–59 (n=25)	26.8–31.2	31.9–35.4	35.9–47.1
	60–69 (n=15)	22.4–25.6	26.8–34.8	36.1–47.7
Women	20–29 (n=2)	–	37.0	39.6
	30–39 (n=48)	20.6–29.9	30.3–33.9	34.0–48.9
	40–49 (n=100)	17.8–28.2	28.3–32.7	32.8–44.6
	50–59 (n=135)	13.9–25.5	25.6–29.5	29.7–39.8
	60–69 (n=57)	16.9–24.9	25.1–28.8	29.0–37.4

Arterial stiffness

Participants were observed under quiet resting conditions in the supine position. baPWV and BP were measured with a vascular testing device (form PWV/ABI device; Omron Colin, Kyoto, Japan), according to the method previously described7 Bilateral brachial and ankle arterial pressure waveforms were stored for 10 seconds by the extremity cuffs connected to a plethysmographic sensor and an oscillometric pressure sensor wrapped around the participant's arms and ankles. The baPWV was calculated from the distance between the two arterial recording sites divided by the transit time.21 The coefficient of variation for interobserver reproducibility of baPWV was 4% in our laboratory. The intraclass correlation coefficient was 0.84 (95% confidence interval, 0.78–0.90), and the minimal detectable change with 95% confidence was 371 cm/s for baPWV in a previous study.22

Recordings were made in triplicate, and they strictly conformed to American Heart Association guidelines.23 The mean of the right and left baPWV values were used for analysis.

Cardiorespiratory fitness

CRF, which was assessed from the peak oxygen uptake ($\dot{\mathrm{V}}{\mathrm{O}}_{2\mathrm{peak}}$), was measured by an incremental cycle exercise test using a cycle ergometer (Monark Ergomedic 828E Test Cycle [National Institute of Health and Nutrition], Varberg, Sweden, or Excalibur V2.0 [Okayama Southern Institute of Health], Lode BV, Groningen, The Netherlands), as described previously.14, 24, 25 The incremental cycle exercise began at a work load of 30 W to 60 W for women and 60 W to 120 W for men, which was then increased by 15 W/min until the participants could no longer maintain the fixed pedaling frequency (60 rpm). The participants were encouraged during the ergometer test to exercise at the level of maximum intensity. The expired air of participants who were tested at the National Institute of Health and Nutrition was collected over 30‐second intervals in Douglas bags. Expired oxygen (O₂) and carbon dioxide (CO₂) gas concentrations were measured by mass spectrometry (Arco‐1000; Arco System, Ogaki, Japan), and gas volume was determined using a dry gas meter (DC‐5; Shinagawa Seisakusho, Tokyo, Japan). The expired air of participants who were tested at Okayama Southern Institute of Health was collected and the rates of O₂ consumption and CO₂ production were measured breath by breath using a cardiopulmonary gas exchange system (Oxycon Alpha; Mijnhrdt B.V., The Netherlands). The heart rate and rating of perceived exertion26 were monitored on a minute‐by‐minute basis during exercise. The highest value of $\dot{\mathrm{V}}{\mathrm{O}}_2$ during the exercise test was designated as $\dot{\mathrm{V}}{\mathrm{O}}_{2\mathrm{peak}}$. The CRF assessment was performed after all other tests.

Anthropometric measures and body composition

Height, weight, and waist circumference were measured and body mass index (BMI) was calculated. Body composition was determined by dual‐energy radiography absorptiometry (Hologic QDR‐4500; Hologic, Waltham, MA) with participants in the supine position.

Exercise habits

Data on exercise habits were obtained at interviews conducted by well‐trained staff using the structured method of the National Nutrition Survey in Japan. The patients were asked if they currently exercised (over 30 minutes per session, two times per week for a duration of 3 months). Participants who answered “yes” were classified as participants with exercise habits (exercised regularly). Participants who answered “no” were classified as participants without exercise habits.

Physical activity

The duration and intensity of physical activity were evaluated by triaxial accelerometry (Actimarker EW4800; Panasonic Electric Works, Osaka, Japan), as described previously.27, 28 Participants were asked to wear a triaxial accelerometer for 28 days; we used data for 14 days, during which the accelerometer was worn continuously from the time the participant awoke until he or she went to bed. We obtained the duration of daily physical activity corresponding with 1.1 METs to 2.9 METs (light), 3.0 METs to 5.9 METs (moderate), and ≥6.0 METs (vigorous).29 Time spent in inactivity was defined as the sum of time spent sedentary (<1.1 METs) and the time at which the accelerometer was not worn, which was calculated as 1440 – (daily time spent in light physical activity+moderate+vigorous).

Blood samples

Blood samples were taken from participants following an overnight fast of at least 10 hours. Venous blood withdrawn from the antecubital vein was collected into tubes without additives or EDTA and was immediately centrifuged at 3000 rpm for 20 minutes to obtain serum or plasma. The levels of glucose and glycated hemoglobin (HbA_1c) in plasma and total cholesterol, high‐density lipoprotein (HDL) cholesterol, and triglycerides in serum were determined.

Statistical Analyses

Data are expressed as mean±standard deviation unless otherwise indicated. The differences across the CRF levels were assessed by one‐way analysis of variance or Kruskal‐Wallis test. A general linear model was used to analyze the association of the 2‐year change in baPWV (Δ: follow‐up baPWV – baseline baPWV) across the CRF levels. The 2‐year change in baPWV was entered as a dependent variable; the tertile $\dot{\mathrm{V}}{\mathrm{O}}_{2\mathrm{peak}}$ categories were entered as fixed factors; and baseline age, BMI, body fat, glucose, triglycerides, HDL cholesterol, baPWV, and sex were entered as covariates for adjustment. Bonferroni's test was applied for post hoc pair‐wise comparisons. In the general linear model analyses, data were expressed as estimated marginal mean±standard error.

Pearson's correlation coefficients were used to analyze the relationships between the 2‐year change in baPWV and the baseline variables (age, height, weight, BMI, waist circumference, body fat, glucose, HbA_1c, triglycerides, total cholesterol, HDL cholesterol, systolic BP (SBP) and diastolic BP (DBP), baPWV, and $\dot{\mathrm{V}}{\mathrm{O}}_{2\mathrm{peak}}$). A stepwise multiple regression analysis was used to determine the influences of baseline variables on the 2‐year change in baPWV. P<.05 was considered statistically significant. Statistical analyses were performed using SPSS software, version 20.0 (IBM Japan, Tokyo, Japan).

Results

Table 2 shows the baseline characteristics of the participants divided by CRF level. There were significant differences in $\dot{\mathrm{V}}{\mathrm{O}}_{2\mathrm{peak}}$, BMI, body fat, glucose, HbA_1c, triglycerides, HDL cholesterol, heart rate, baPWV, exercise habits, and physical activity among the three groups. The Figure highlights the significant association of the 2‐year change in baPWV across CRF levels. General linear model analysis revealed that the changes in baPWV during the study period were significantly higher in patients in the low CRF group (unfit) than they were in patients in the high CRF group (fit). An inverse relationship was observed between CRF level and the 2‐year changes in baPWV (28.8±7.6, 16.0±6.8, and −1.4±7.5, respectively; P for trend=.029). The 2‐year changes in baPWV were significantly higher in the low CRF group than they were in the high CRF group (P=.024). These data indicate that a higher CRF is associated with slower progression of age‐related arterial stiffening.

Table 2.

Physical Characteristics of Participants by CRF Level

		CRF Level
Variable	Overall	Low	Middle	High	P Value
No.	470	153	161	156
Peak oxygen uptake, mL/kg/min	31.3±7.0	25.4±4.2	30.9±4.4	37.6±6.2	<.01a
Men/women, No.	128/342	41/112	45/116	42/114	NSb
Premenopausal women, No. (%)	170 (36)	52 (34)	58 (36)	60 (39)	NSb
Age, y	48.8±9.5	49.5±9.4	49.0±9.8	48.1±9.2	NSa
BMI, kg/m2	22.2±2.8	23.1±3.3	22.0±2.5	21.5±2.2	<.01a
Body fat, %	25.7±6.7	29.0±6.5	25.6±6.1	22.7±5.9	<.01a
Glucose, mg/dL	89.4±9.8	91.4±12.0	87.9±8.3	89.0±8.4	<.01a
Glycated hemoglobin, %	5.3±0.5	5.4±0.6	5.3±0.3	5.3±0.3	.03a
Triglycerides, mg/dL	85±55	99±66	82±57	76±36	<.01a
Total cholesterol, mg/dL	208±34	212±32	206±35	208±36	NSa
HDL cholesterol, mg/dL	67±17	63±16	66±15	71±19	<.01a
Heart rate, beats per min	61±11	65±10	62±11	59±10	<.01a
SBP, mm Hg	116±14	118±15	114±13	117±12	NSa
DBP, mm Hg	71±10	72±11	69±10	71±10	NSa
baPWV, cm/s	1229±166	1272±197	1214±164	1204±122	<.01a
Patients with exercise habits, No. (%)	228 (48.5)	41 (26.8)	77 (47.8)	110 (70.5)	<.01b
Jogging	39 (8.3)	4 (2.6)	8 (5.0)	27 (17.3)	–
Swimming	88 (18.7)	16 (10.5)	27 (16.8)	45 (28.8)	–
Dancing	60 (12.8)	14 (9.2)	18 (11.2)	28 (17.9)	–
Stretching	43 (9.1)	10 (6.5)	13 (8.1)	20 (12.8)	–
Resistance	58 (12.3)	15 (9.8)	17 (10.6)	26 (16.7)	–
No.	459	151	160	148
Daily time spent in physical activity
Light, min/d	579±113	578±124	577±109	581±105	NSa
Moderate, min/d	61±25	52±20	63±26	67±26	<.01a
Vigorous, min/d	2.7±7.9	1.6±5.0	1.7±4.4	5.0±11.8	<.01a
Inactivity, min/d	798±115	808±124	798±112	787±109	NSa

Abbreviations: baPWV, brachial‐ankle pulse wave velocity; BMI, body mass index; DBP, diastolic blood pressure; HDL, high‐density lipoprotein; SBP, systolic blood pressure; NS; not significant. Data are presented as mean±standard deviation unless otherwise noted.

^aAnalysis of variance.

^bKruskal‐Wallis test.

Figure 1.

Estimated marginal mean±standard error of the 2‐year changes in brachial‐ankle pulse wave velocity (baPWV) grouped by peak oxygen uptake ($\dot{\mathrm{V}}{\mathrm{O}}_{2\mathrm{peak}}$) tertiles in healthy adults. Data are adjusted for baseline age, body mass index, body fat, glucose, triglycerides, high‐density lipoprotein cholesterol, baPWV, and sex.

Pearson's correlation coefficients between the 2‐year change in baPWV and baseline parameters are as follows: the 2‐year change in baPWV correlated with age (r=.133, P<.01), glucose level (r=.105, P=.023), HbA_1c concentration (r=.095, P=.039), baPWV (r=−.121, P<.01), and $\dot{\mathrm{V}}{\mathrm{O}}_{2\mathrm{peak}}$ (r=−.112, P=.015). These data indicate that CRF is correlated with age‐related arterial stiffening.

The stepwise multiple regression analysis revealed that age, glucose level, baPWV, $\dot{\mathrm{V}}{\mathrm{O}}_{2\mathrm{peak}}$, and sex were independent correlates of 2‐year changes in baPWV (Table 3). These data indicate that CRF is associated with arterial stiffness, independent of other factors already known to be related to arterial stiffness.

Table 3.

Stepwise Multiple Regression Analyses Showing Association Between 2‐Year Changes in baPWV and Baseline Variables

	Nonstandardized Coefficient	95% CI	Standardized Coefficient	P Value
2‐year changes in baPWV
Age, y	1.712	0.73–2.69	0.181	<.01
Glucose, mg/dL	1.053	0.21–1.90	0.115	.015
baPWV, cm/s	−0.151	−0.21 to −0.10	−0.281	<.01
Peak oxygen uptake, mL/kg/min	−2.187	−3.59 to −0.78	−0.172	<.01
Sex, male=0; female=1	−37.783	−59.2 to −16.4	−0.188	<.01

Abbreviations: baPWV, brachial‐ankle pulse wave velocity; CI, confidence interval.

Table 4 shows the results of the 2‐year change in baPWV according to subgroup. There were no significant differences in the rate of progression of baPWV between the younger (<50 years) and older (>50 years) age groups or between the BMI groups (<25 kg/m² and ≥25 kg/m²). The annual rate of change in baPWV was lower in women than it was in men.

Table 4.

Two‐Year Changes in baPWV According to Subgroup

	Two‐Year Changes in baPWV, cm/s
Subgroup	Crude	P Value	Adjusted	P Value
Age,a y
<50 (n=238)	12.1±5.4	NS	9.5±6.2	NS
≥50 (n=232)	16.8±6.3		19.5±6.3
Sexb
Men (n=128)	20.6±8.7	NS	34.3±9.7	.03
Women (n=342)	12.1±4.6		6.9±5.1
BMIc, kg/m2
<25 (n=404)	13.3±4.3	NS	13.3±4.3	NS
≥25 (n=66)	21.1±12.5		21.2±10.9

Abbreviations: BMI, body mass index; NS, not significant. ^aAdjusted for sex, body fat, glucose, total cholesterol, systolic blood pressure (SBP), and brachial‐ankle pulse wave velocity (baPWV). ^bAdjusted for age, body fat, glucose, total cholesterol, SBP, and baPWV. ^cAdjusted for age, glucose, total cholesterol, SBP, baPWV, and sex. Bold value indicates significance.

Discussion

In this longitudinal observational study, the 2‐year changes in baPWV were higher in patients in the low CRF groups compared with those in the high CRF groups. Moreover, CRF was inversely associated with 2 years of baPWV progression in healthy adults. This association was independent of other confounders. To our knowledge, this is the first study to examine the relationship between CRF and the progression of baPWV in healthy adults in a longitudinal study. These findings suggest that higher CRF is associated with slower progression of age‐related arterial stiffening in healthy adults.

Previous cross‐sectional studies indicated that a high level of CRF was associated with arterial stiffness.13, 14, 15, 16 However, little is known about the longitudinal relationships between CRF and the progression of baPWV. Here, we determined the relationship between CRF and the 2‐year changes in baPWV. The strengths of the present study were that the CRF levels of all participants were evaluated by maximal exercise testing, and the sample size was relatively large. Similar to previous cross‐sectional findings, the present study also showed that the progression of baPWV was significantly related to CRF. Ferreira and colleagues17, 18 showed that longitudinal changes in CRF were inversely and significantly associated with arterial stiffness (distensibility, compliance coefficients, and Young's elastic modulus) in adolescence to young adulthood. In comparison to the studies by Ferreira and coworkers,17, 18 our study not only used a larger cohort of participants but also analyzed middle‐aged and older members of the population as well as young adults. In addition, we used baPWV for the arterial stiffness assessment. Therefore, this is the first study to examine the relationship between CRF and the progression of arterial stiffness as assessed by baPWV. We were unable to directly compare Ferreira's results with ours because of the different methods of arterial stiffness assessment. However, our present study indicated a similar tendency and confirmed that CRF is inversely associated with baPWV progression in middle‐aged and older populations. Our results suggest that maintaining greater CRF would generate a protective effect on the progression of age‐related arterial stiffening in middle‐aged and elderly populations.

BP is the most important determinant of arterial stiffening. Determination of baPWV has been the subject of many cross‐sectional studies, but few studies have examined the rate of progression of baPWV, especially in healthy adults. Thus, it is unclear whether BP changes or baseline BP correlates with changes in baPWV in healthy adults. Although the baseline SBP was highly correlated with the baseline baPWV (r=.72) in our study, we did not find a significant correlation between the changes in baPWV and the baseline SBP, which is consistent with a previous report by Wildman and colleagues.30 These findings may indicate a strong and independent association between BP levels and arterial stiffening. Several previous studies have indicated that SBP changes or baseline BPs correlate with changes in PWV. However, the SBP of participants in these studies was higher than that in our study participants. SBP is the main longitudinal determinant of PWV. We think that a normal SBP value (<120 mm Hg) may not affect PWV trajectories, while an SBP value >120 mm Hg may affect PWV.31 Since our study participants were very healthy (no history of stroke, cardiac disease, or chronic renal failure; received no medical treatment for hypertension, dyslipidemia, or diabetes; were currently nonsmokers; and had an ankle‐brachial pressure index of 0.9–1.3), the SBP was lower than that in previous studies. On the other hand, the $\dot{\mathrm{V}}{\mathrm{O}}_{2\mathrm{peak}}$ was correlated with both the baseline baPWV and the changes in baPWV. Therefore, we think that CRF is one of the relevant factors affecting baPWV.

It is well established that the annual rate of change in PWV accelerates with advancing age31 and BMI.30 In the present study, however, there were no significant differences in the rate of arterial stiffness progression as assessed by baPWV in the younger (<50 year) and older (>50 year) groups or in the BMI <25 kg/m² and BMI ≥25 kg/m² groups. Because we extracted a population of apparently healthy participants who had no atherogenic or metabolic disorders, there might have been no differences between these groups. On the other hand, the annual rate of change in baPWV was lower in women than in men, which is consistent with a previous report by Tomiyama and colleagues.32

Our study showed that CRF was inversely associated with increased baPWV. One possible reason for this finding could be that glucose, HbA_1c, and body fat were lower in people with high CRF (Table 2). Hyperglycemia and obesity have been shown to be associated with arteriopathy.33 A previous study suggested that glucose or HbA_1c and body fat had a negative correlation with PWV.34, 35 However, in multiple linear regression and general linear model analyses that included these factors, the $\dot{\mathrm{V}}{\mathrm{O}}_{2\mathrm{peak}}$ was independently related to the progression of arterial stiffness as assessed by baPWV. Therefore, maintaining a higher CRF may directly influence the progression of baPWV. The endurance‐trained state is associated with an elevated overall content of elastin, reduced calcium content,36 reduced formation of advanced glycation end products and collagen cross‐linking in the arterial wall,37 and improved endothelial function.38, 39 Thus, habitual exercise and a higher fitness level are thought to decrease arterial stiffness by preventing arterial remodeling and dysfunction. Moreover, CRF heritability in sedentary people ranged from 25% to 65%40 and some individuals can perform very little physical activity and still possess a high $\dot{\mathrm{V}}{\mathrm{O}}_{2\mathrm{peak}}$. It is possible that this study's findings are also affected by genetic factors. Therefore, additional studies are required to further investigate this topic in longer observational or interventional studies.

Study Strengths and Limitations

Our findings have several important implications. First, the present study showed that a higher level of CRF was associated with lower levels of arterial stiffness progression in healthy adults. The baPWV is a risk factor for total CV events and all‐cause mortality.9, 10 The maintenance of a higher level of CRF may have a protective effect against CV events by attenuating age‐related arterial stiffening. Therefore, the maintenance of a higher level of CRF may be important for primary prevention strategies. Second, the annual rate of change in baPWV in the intermediate CRF level group was 8.1 cm/s per year (16.1 cm/s per 2 years). Compared with previous cross‐sectional findings of arterial stiffening with aging,32, 41 our data are similar, which adds to its validity.

There are several potential limitations to the present study. First, CRF was measured only once at baseline. It is possible that the participants' CRF levels changed during the observation period. During our 2‐year observational study, we observed that a higher level of CRF was associated with lower levels of arterial stiffness progression in healthy adults. However, the field would benefit from more longitudinal research to determine the cause‐effect relationships between CRF and arterial stiffness progression. Second, depending on which clinic the participants attended, two different methods to analyze O₂ consumption and CO₂ production were used. However, a previous study reported that O₂ consumption and CO₂ production attained from the incremental cycle test were similar for the two different methods.42 Third, we used baPWV to assess arterial stiffness; this method has been popularized primarily in Asian countries for over 15 years. However, the cfPWV is considered the gold‐standard method for assessing aortic stiffness. Numerous studies have showed that baPWV performs comparatively well to cfPWV in identifying vascular damage.5, 7, 43 Therefore, we believe that baPWV provides qualitatively similar information to that derived from the aortic pulse wave velocity. Measuring baPWV is a simple but powerful technique that can be easily applied in clinical practice. Fourth, arterial stiffness undergoes major changes during the menstrual cycle. Since the present study did not control for the menstrual cycle, it is possible that our findings were affected by this. However, the number of premenopausal women included in our study was relatively low (36%), and there were no significant differences in the number of premenopausal women among the three groups. Therefore, we think that there was a relatively small effect on the progression of baPWV.

Conclusions

This was the first prospective study to indicate that CRF is inversely associated with age‐related arterial stiffening. This association was independent of some confounders, specifically age and glucose. Therefore, CRF may be an effective measure for preventing age‐related arterial stiffening.

Statement of Financial Disclosure

This study was supported by a Grant‐in‐Aid for Scientific Research (#23240089, M. Miyachi; #25750363, Y. Gando) and a Grant‐in‐Aid for Scientific Research from the Ministry of Health, Labor and Welfare of Japan (M. Miyachi). The authors declare no conflicts of interest.

J Clin Hypertens (Greenwich). 2016;18:292–298. DOI: 10.1111/jch.12753. ©2015 Wiley Periodicals, Inc.