Change in Ankle-Brachial Index Over Time in a Screened Japanese Cohort　– The Okinawa Peripheral Arterial Disease Study –

Abstract

Background: The temporal change in ankle-brachial index (ABI) in the general population, especially in those aged <40 years, remains unclear.

Methods and Results: ABIs of 23,673 individuals were measured in 1-day health checkups between 2003 and 2010. Among them, 1,117 participants aged 28–76 years (mean 53±9 years) whose ABI was measured at least twice within an interval of ≥4 years (mean: 4.9 years) were selected for this study. Baseline ABI was the lowest at age <40 years and increased with age. ABI significantly increased in participants aged <40 and 40–49 years, but not in participants aged 50–59 and ≥60 years. ABI increased in participants with borderline-low baseline ABI (0.9<ABI<1.0, 0.09; P<0.001) and normal baseline ABI (1.0≤ABI<1.2, 0.006; P=0.017). ABI decreased in participants with high-normal baseline ABI (1.2≤ABI<1.4, −0.04; P<0.001). Stepwise multivariate analysis revealed that ABI change was independently associated with baseline ABI (β=−0.566), height (β=0.162), body mass index (β=0.093), and sex (women, β=−0.08).

Conclusions: ABI was lowest at age <40 years and increased with age. In participants aged <50 years, ABI significantly increased over the mean observation period of 4.9 years. (Circ J 2016; 80: 2004–2009)

The ankle-brachial index (ABI) is the ratio of the systolic blood pressure (BP) measured at the ankle to that in the arm. A cutoff ABI value ≤0.90 has good sensitivity and excellent specificity for lower-extremity arterial stenosis >50% on digital subtraction angiography.¹ The ABI is also an indicator of future cardiovascular events and death.^2,3 Both a low ABI (≤0.9) and a borderline-low ABI (0.9<ABI<1.0) are associated with endothelial dysfunction,⁴ an increased risk of myocardial infarction, stroke, and both overall death and cardiovascular-related death.^5–7 Several studies have reported differences in the normal ABI values according to sex and ethnicity.⁸ However, the same cutoff value has been used for the diagnosis of peripheral arterial disease (PAD) regardless of sex, ethnic group, and age. Lower-extremity PAD is mostly an atherosclerotic disease and predominantly affects patients at high risk of cardiovascular diseases. Therefore, ABI screening is recommended for patients with suspected lower-extremity PAD, who are defined as individuals with ≥1 of the following: exertional leg symptoms, non-healing wounds, aged ≥65 years, or ≥50 years with a history of smoking or diabetes.⁹

Previous studies have reported that the ABI decreases with age, most likely because of increased prevalence and progression to lower-extremity PAD.¹⁰ However, we recently reported that the ABI was lowest at age <40 years, and increased with age until 60–69 years in both sexes among subjects of a screened cohort.¹¹ In participants aged 21–39 years, 18% of women and 8% of men had a borderline-low ABI.¹² For both men and women aged <60 years the prevalence of hypertension was significantly lower in participants with borderline-low ABI compared with those with 1.0≤ABI<1.4. We assumed that young adults did not always have arterial stenosis despite borderline-low ABI. We reported that the ABI positively correlated with systolic BP, pulse pressure, and the brachial-ankle pulse wave velocity (baPWV), indices of arterial stiffness. A high-normal ABI (1.2≤ABI<1.4) was also associated with target organ damage, including proteinuria, only in participants aged <60 years.¹² In contrast, a low ABI was associated with proteinuria only in participants aged ≥60 years. Therefore, we hypothesized that ABI increases with age and arterial stiffening and decreases when flow-limiting atherosclerotic arterial stenosis occurs in a lower limb. The aim of this study was to evaluate the ABI change (∆ABI) over time in a younger population.

Methods

Subjects

The present study was an observational, longitudinal, single-center study. Participants were recruited from the 1-day health evaluation of the general population held by the Okinawa General Health Maintenance Association. A total of 91,962 individuals participated and for 23,673 participants, the ABI was voluntarily measured as an optional checkup between July 2003 and March 2010. Some individuals attended checkup multiple times during this period, so we enrolled subjects who participated at least twice over a 4-year interval (ie, visit 1 in 2003–2005 and visit 3 in 2008–2010). Participants with an ABI ≤0.9 (n=4) at baseline were excluded. No one had an ABI ≥1.4. After applying the exclusion criteria, 1,117 participants (509 men, 608 women) remained and of these, 825 subjects participated during the 2007–2008 period (visit 2). All participants provided written consent to participate in the study, which was approved by the Research and Ethics Committee of the University of the Ryukyus.

ABI Measurement

ABI and baPWV were measured using a validated automatic device (BP-203RPE, II form PWV/ABI; Omron-Colin, Japan), which simultaneously measures pulse volume in the brachial and ankle arteries using an oscillometric method together with bilateral arm and ankle BP. Both the ABI and baPWV were measured after the participant had rested supine for at least 5 min. ABI was calculated bilaterally as the ratio of systolic BP in the arms. The lower ABI value of both legs was used. ∆ABI between visit 1 and visit 3 was calculated as ABI from visit 1 minus ABI of the same side measured on visit 3. ABI was categorized into 4 ranges: (1) low (≤0.9), (2) borderline-low (0.9<ABI<1.0), (3) normal (1.0≤ABI<1.2), and (4) high-normal (1.2≤ABI<1.4). The reliability and reproducibility of this instrument in measuring ABI are similar to that of the Doppler probe.¹³

Data Collection

Individual medical histories were determined using self-administered questionnaires and were confirmed by physician interview. Trained nurses measured BP using a standard mercury sphygmomanometer after the participants sat quietly for 15 min. Blood was drawn after an overnight fast. Hypertension was defined as the systolic BP ≥140 mmHg and/or the diastolic BP ≥90 mmHg, or the current use of any antihypertensive medication. Body mass index (BMI, kg/m²) was calculated as the body weight divided by the square of the height. Diabetes was defined as fasting blood sugar level ≥126 mg/dl, hemoglobin A1c level (HbA1c) ≥6.5%, and/or taking of any medication for diabetes. Dyslipidemia was defined as serum total cholesterol level ≥220 mg/dl, serum triglyceride level ≥150 mg/dl, high-density-lipoprotein-cholesterol level <40 mg/dl, and/or taking of any medication for dyslipidemia. Smoking was defined as current, past, or never as determined by interview.

Statistical Analysis

All parametric values are expressed as the mean±standard deviation (SD). Comparisons between quantitative variables were performed with Student’s t-test. The difference in the measured values between visit 1 and visit 3 were assessed using a paired t-test. Two-sided tests with P<0.05 were considered to be significant. If a clinically worthwhile effect for ABI is 0.02¹⁴ and the between subjects’ standard deviation is 0.05,¹⁵ we estimated the sample size is 100 with an 80% power and 5% significance. All statistical analyses were performed using JMP (version 10.0.0). The difference in the mean values between groups was tested using the Tukey-Kramer test. Stepwise forward multiple linear regression analyses were performed to determine independent predictors of ∆ABI. Independent variables considered in the models were those that were found to be clinically relevant (eg, age; sex; BMI; height; baseline ABI; baseline baPWV; current smoking status; presence of hypertension, diabetes mellitus, and dyslipidemia).

Results

Baseline Characteristics

The mean age of participants was 53±10 years (range, 28–76 years). For women, participants with a borderline-low ABI were found to be younger and have higher HDL-cholesterol, and lower baPWV values than those with normal ABI. Participants with high-normal ABI were found be older and to have a higher prevalence of hypertension than those with normal ABI (Table 1). For men, participants with a borderline-low ABI were found to be younger than those with a normal ABI. Participants with a high-normal ABI were found to have a higher prevalence of hypertension and diabetes as well as higher systolic BP, and diastolic BP than those with normal ABI.

Table 1. Baseline Characteristics of Study Subjects by Sex and ABI

Variables	Women (n=608)				Men (n=509)
	0.9<ABI<1.0	1.0≤ABI<1.2	1.2≤ABI<1.4	P value	0.9<ABI<1.0	1.0≤ABI<1.2	1.2≤ABI<1.4	P value
n	35	531	42		13	415	81
Age (years)	50±7*	53±8	57±9**	<0.001	45±9*	52±11	54±10	0.009
Height (cm)	151±5	152±5	153±5	0.464	167±6	166±6	167±6	0.163
BMI (kg/m2)	23±3	24±3	25±4	0.034	25±3	25±3	26±3	0.086
Hypertension (%)	4 (11)	139 (26)	21 (50)**	<0.001	3 (23)	136 (33)	40 (49)**	0.011
Diabetes (%)	0 (0)	22 (4)	1 (2)	0.408	2 (15)	28 (7)	14 (17)**	0.006
Dyslipidemia (%)	11 (31)	232 (44)	15 (36)	0.240	6 (46)	244 (59)	46 (57)	0.637
Current smoking (%)	2 (6)	29 (5)	1 (2)	0.629	6 (46)	111 (27)	17 (21)	0.162
baPWV (m/s)	13.46±2.21*	14.21±2.65	14.87±2.51	0.024	14.70±1.23	14.70±2.60	15.26±2.65	0.504
SBP (mmHg)	115±13	121±15	127±15	0.004	121±8	125±14	132±16**	<0.001
DBP (mmHg)	71±9	74±10	77±8	0.034	74±10	77±10	81±10*	0.009
HbA1c (%)	5.4±0.1	5.6±0.0	5.6±0.1	0.477	5.7±0.2	5.6±0.0	5.7±0.1	0.339
Total cholesterol (mg/dl)	202±28	209±33	210±29	0.463	204±31	201±33	198±29	0.684
HDL-cholesterol (mg/dl)	68±14*	63±14	60±11	0.020	52±11	51±12	50±10	0.756
Triglycerides (log10)	4.4±0.5	4.5±0.5	4.5±0.4	0.282	5.0±0.6	4.9±0.6	4.9±0.6	0.816

Values are mean±SD or %. *P<0.05, **P<0.01 vs. 1.0≤ABI<1.2 group. ABI, ankle-brachial index; baPWV, brachial-ankle pulse wave velocity; BMI, body mass index; DBP, diastolic blood pressure; HbA1c, hemoglobin A1c; HDL, high-density lipoprotein; SBP, systolic blood pressure.

Change in ABI Over Time

The mean duration between visit 1 and visit 3 was 4.9±0.3 years (range, 4.0–5.9 years). The frequency of baseline borderline-low ABI was highest in participants aged <40 years and decreased with age in both sexes (Figure 1A). Figure 1B shows the distribution of ∆ABI by sex. The mean ∆ABI was quite small and exhibited no significant difference over 4.9 years (0.003±0.06, P=0.096). Baseline ABI was the lowest at age <40 years and increased with age until 60–69 years (Figure 2A) as we previously reported.¹¹ ABI significantly increased at visit 3 only in participants aged <40 years (0.02±0.07, P<0.01) and 40–49 years (0.01±0.07, P<0.01). ∆ABI decreased in participants aged 50–59 years (−0.001±0.06, P=0.660) and ≥60 years (−0.004±0.06, P=0.852), but the difference was not statistically significant. Only 1 participant had an ABI ≤0.9 (0.93–0.84) and no individuals had an ABI ≥1.4 at visit 3. We also examined 3 serial ABI measurements by age category in 825 participants (Figure 2B). In participants aged <40 years and 40–49 years, the ABI was significantly higher at visits 2 and 3 than at visit 1. In participants aged 50–59 years, the ABI decreased at visit 3. In participants aged ≥60 years, the ABI did not change over time.

Figure 1.

(A) Prevalence of baseline ABI by age group. (B) Frequency of change in ABI distribution by sex. ABI, ankle-brachial index.

Figure 2.

(A) Change in ABI between visits 1 and 3 by age group. (B) Three serial ABI by age group. *P<0.05, **P<0.01 vs. visit 1. ^†P<0.05 vs. visit 2. ABI, ankle-brachial index.

Determinants of ∆ABI

Stepwise forward multiple regression analyses revealed that baseline ABI, height, sex, and BMI were independent determinants of ∆ABI, accounting for 31% of the variability of ∆ABI. Baseline height and BMI were positively associated with ∆ABI, and baseline ABI and sex (female) were negatively associated with ∆ABI (Table 2).

Table 2. Determinants of Change in Ankle-Brachial Index (ΔABI) in Study Subjects

Variables	Univariate correlations		Stepwise multiple regression model (r2=0.31, P<0.001)
	R2	P value	β	P value
Age	0.013	<0.001	–
Women	0.008	0.003	−0.08	0.047
Height	0.013	<0.001	0.16	<0.001
BMI	0.002	0.186	0.08	0.001
Hypertension	0.008	0.003	–
Diabetes	0.001	0.239	–
Dyslipidemia	<0.001	0.671	−0.04	0.150
Current smoking	0.001	0.257	−0.04	0.167
Baseline ABI	0.253	<0.001	−0.57	<0.001
baPWV	0.002	0.125	0.04	0.136

Abbreviations as in Table 1.

Discussion

We observed that ∆ABI over the mean observation period of 4.9 years was quite small as in previous reports,¹⁶ and increased only in individuals who were aged <40 years and 40–49 years. In participants aged ≥60 years, ∆ABI decreased, but was not statistically significant. Therefore, the results in this study are in agreement with those of our previous cross-sectional study that showed that ABI was lowest in individuals aged <40 years, increased with age until 60–69 years, and decreased thereafter.¹¹

Several previous studies have reported the prevalence of PAD in Japan. The prevalence of an ABI <0.9 was 1.5% in men and 0.7% in women <60 years, and 3.6% in men and 3.3% in women ≥60 years in subjects aged ≥40 years.¹⁷ In the Hisayama study, the prevalence of ABI ≤0.9 was 1.47% in the 3,061 subjects aged ≥40 years.¹⁸ We previously reported that the prevalence of ABI ≤0.9 was 0.5% (n=67) in a screened cohort of Japanese (n=13,211).¹¹ In this study, the prevalence of an ABI ≤0.9 was 0.4% (n=4) at baseline, which is much lower than the prevalence in previous reports, and only 1 subject progressed to an ABI ≤0.9 from a borderline-low ABI. One reason for this lower prevalence of ABI ≤0.9 may be that the study subjects were much younger, as 34% of the participants were aged <50 years.

Several studies have investigated ∆ABI over time. In the cross-sectional Edinburgh Artery Study, the mean ABI decreased with age in both sexes in subjects aged 55–74 years.¹⁹ In a longitudinal study, a decrease of 0.03 was observed over 5 years, and only 0.01 over 12 years was observed in 695 subjects aged 55–74 years in a population-based study.²⁰ In a group of 750 Spanish volunteers aged 60–79 years, ABI decreased 0.02 over 4 years.¹⁴ In the Cardiovascular Health Study, a reduction of only 0.02±0.13 in participants was observed by a decline in ABI from <0.15 to >0.9.¹⁶ Even in a series of patients assessed in a vascular laboratory in which 45% had a baseline ABI <0.9, ABI decreased by a mean of 0.06 over 4.6 years.²¹ Overall, most of the previous studies demonstrated that ABI decreased with age in subjects over 55 years old, but the change was relatively small. ∆ABI over time even in subject over 65 years old in our study was quite low (−0.002) compared with American and European populations. These results support the reports that the Japanese have a lower prevalence of atherosclerotic diseases, including PAD, than in any other country²² and also may have a slower rate of PAD progression.

A few longitudinal studies reported that a certain percentage of subjects had an increase in ABI over time. In a multi-ethnic study of atherosclerosis,²³ 1.3% participants with 0.9<ABI<1.4 developed an ABI ≥1.4 at the 3-year follow-up. The risk factors for progression to ABI ≥1.4 were the male sex, as well as a higher baseline BMI and ABI. Lahoz et al reported that a decrease in ABI >10% was observed in 22% and an increase of ABI >10% was observed in 16% of participants.¹⁴ Compared with the participants with ∆ABI <10%, participants with an increase of ∆ABI >10% had higher rates of low HDL-cholesterol levels and Framingham risk score, but a lower baseline ABI and rate of β-blocker medication. However, the multivariate analysis showed no significant variables associated with an increase of ∆ABI >10%. The mechanism for the increase in ABI has yet to be elucidated, but an abnormally high ABI (≥1.4) has been thought to be related to calcification of the arterial media. This is more frequent in individuals with diabetes, end-stage renal disease, and advanced age.¹⁰ In the present study, no individuals exhibited a baseline or follow-up ABI ≥1.4. In addition, ∆ABI increased only in the subjects aged <40 and 40–49 years, who did not have higher prevalence of diabetes or renal failure.

The BP waveform amplifies as it travels distally from the heart, resulting in a progressive increase in systolic BP and may be the case in the ABI.¹⁰ Both arterial stiffness and reflected waves contribute to systolic BP amplification. The amplification of pressure in the upper limbs is different from that below the aorta in the femoral and more distal arteries.²⁴ The reflected wave returning from the lower body provides a substantial contribution to the contour of the brachial pressure wave, and is responsible for the secondary late systolic surge. In contrast, the reflection from the upper limb has little effect on the lower body pulse, indicating that the age-related increase in systolic BP amplification might be more enhanced in the ankle than in the elbow, and could result in age-related increases of the ABI. Previous studies have demonstrated a positive correlation between height and ABI as a consequence of the progressive systolic BP increase with greater distance from the heart.^25,26 Sex differences in the ABI have been reported in previous studies,⁸ including ours,¹¹ and height is sometimes considered the reason. However, we and others have reported that the influence of height on ABI is small and not significant after adjusting for sex.^11,27 The relationship between height and ∆ABI over time has not been reported to date. However, high stature may be correlated with an age-dependent increase of ABI because of a greater number of bifurcations in the lower limbs than in the upper limbs, leading to an amplification of systolic BP via increased reflected waves.²⁸ Previous studies have reported that PWV is the highest in the arms during adolescence compared with the aorta and lower extremities, but age-related changes in PWV are the most gradual in the arms.²⁹ These results indicate that age-related changes in systolic BP might be slowest in the arms compared with the aorta and lower extremities, which could lead to age-related increases in the ABI.

In this study, BMI was positively associated with ∆ABI over time via multiple linear regression analyses. In the ARIC study, BMI was higher in the subjects with an ABI >1.3 than in those with 0.9≤ABI<1.3.²⁷ The MESA study also reported that BMI was a risk factor for progression to an ABI ≥1.4. In that study, neither diabetes nor renal failure was associated with an increase of ABI above 1.4. Tabara et al reported a positive correlation between ABI and the femoral muscle cross-sectional area, which is also associated with BMI.³⁰ They concluded that a large muscle mass might act as a resistance to cull compression of the arteries, and thus caused an artifactual elevation of the ankle systolic BP. We analyzed the relationship between changes in body weight and ABI. There was no significant relationship between them (P=0.512), even adjusted for sex and age (P=0.367). This result indicated that a change in muscle mass may not be a potent determinant of DABI.

The ∆ABI over time was significantly associated with baseline ABI: the higher the value, the higher the probability of decrease, and vice versa. These results may indicate statistical convergence to an average value that would lead to an overestimate of the real variation. In this study, however, ∆ABI significantly increased in participants aged <40 years and 40–49 years, but ∆ABI did not decrease in the 50–59 years and ≥60 years age groups. We analyzed 3 serial ABI measurements and revealed that the ABI significantly increased at both visit 2 and visit 3 compared with baseline, in participants aged <40 years and 40–49 years. In participants aged 50–59 years and ≥60 years, the ABI did not change significantly at visit 2 compared with baseline. Therefore, we believe that the ∆ABI over time in participants aged <50 years is different from that of regression to a mean value.

Study Limitations

Participants in this study were self-selected and might have been more concerned about their health than the general population. In addition, the value of ∆ABI was quite small compared with previous studies. Therefore, we need a much larger sample size and a longer follow-up to detect the apparent ∆ABI over time. The numbers of participants with borderline-low ABI and high-normal ABI were not enough to evaluate the effects of age and cardiovascular risk factors on ∆ABI over time by ABI categories. We did not perform other examinations, such as computed tomography angiography, magnetic resonance angiography, and ultrasound duplex scan; therefore, we did not evaluate the prevalence of false-normal results. We did not evaluate the effects of drugs, such as antihypertensive and statins, which may have an influence on arterial stiffness and wave reflections. Because ethnicity affects the ABI value, further studies are needed to establish whether these results apply to any other ethnicities.

Conclusions

ABI was lowest at age <40 years and increased with age. In participants aged less than 50 years, ABI significantly increased over the mean observation period of 4.9 years, indicating that borderline-low ABI in low-risk younger subjects does not always indicate lower limb arterial stenosis.

Acknowledgments

We thank the staff of the Okinawa Health Promotion Foundation for collecting the data. We also acknowledge Ms Chiho Iseki for her assistance in the preparation of the manuscript.

Conflict of Interests / Disclosures / Funding Sources

None.

References

• 1.    Lijmer  JG,  Hunink  MG,  van den Dungen  JJ,  Loonstra  J,  Smit  AJ. ROC analysis of noninvasive tests for peripheral arterial disease. Ultrasound Med Biol 1996; 22: 391–398.
• 2.    Norgren  L,  Hiatt  WR,  Dormandy  JA,  Nehler  MR,  Harris  KA,  Fowkes  FG, et al. Inter-society consensus for the management of peripheral arterial disease (TASC II). Eur J Vasc Endovasc Surg 2007; 33(Suppl 1): S1–S75.
• 3.    Kusunose  K,  Sato  M,  Yamada  H,  Saijo  Y,  Bando  M,  Hirata  Y, et al. Prognostic implications of non-invasive vascular function tests in high-risk atherosclerosis patients. Circ J 2016; 80: 1034–1040.
• 4.    Kajikawa  M,  Maruhashi  T,  Iwamoto  Y,  Iwamoto  A,  Matsumoto  T,  Hidaka  T, et al. Borderline ankle-brachial index value of 0.91–0.99 is associated with endothelial dysfunction. Circ J 2014; 78: 1740–1745.
• 5.    Fowkes  FG,  Murray  GD,  Butcher  I,  Heald  CL,  Lee  RJ,  Chambless  LE, et al. Ankle brachial index combined with Framingham Risk Score to predict cardiovascular events and mortality: A meta-analysis. JAMA 2008; 300: 197–208.
• 6.    Nakamura  Y,  Kunii  H,  Yoshihisa  A,  Takiguchi  M,  Shimizu  T,  Yamauchi  H, et al. Impact of peripheral artery disease on prognosis in hospitalized heart failure patients. Circ J 2015; 79: 785–793.
• 7.    Shirasu  T,  Hoshina  K,  Yamamoto  S,  Shigematsu  K,  Miyata  T,  Watanabe  T. Poor prognosis in critical limb ischemia without pre-onset intermittent claudication. Circ J 2015; 79: 1618–1623.
• 8.    Aboyans  V,  Criqui  MH,  McClelland  RL,  Allison  MA,  McDermott  MM,  Goff  DC, et al. Intrinsic contribution of gender and ethnicity to normal ankle-brachial index values: The Multi-Ethnic Study of Atherosclerosis (MESA). J Vasc Surg 2007; 45: 319–327.
• 9.    Rooke  TW,  Hirsch  AT,  Misra  S,  Sidawy  AN,  Beckman  JA,  Findeiss  LK, et al. 2011 ACCF/AHA Focused Update of the Guideline for the Management of Patients With Peripheral Artery Disease (updating the 2005 guideline): A report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol 2011; 58: 2020–2045.
• 10.    Aboyans  V,  Criqui  MH,  Abraham  P,  Allison  MA,  Creager  MA,  Diehm  C, et al. Measurement and interpretation of the ankle-brachial index: A scientific statement from the American Heart Association. Circulation 2012; 126: 2890–2909.
• 11.    Ishida  A,  Miyagi  M,  Kinjo  K,  Ohya  Y. Age- and sex-related effects on ankle-brachial index in a screened cohort of Japanese: The Okinawa Peripheral Arterial Disease Study (OPADS). Eur J Prev Cardiol 2012; 21: 712–718.
• 12.    Ishida  A,  Nakachi-Miyagi  M,  Kinjo  K,  Iseki  K,  Ohya  Y. A high normal ankle-brachial index is associated with proteinuria in a screened cohort of Japanese: The Okinawa Peripheral Arterial Disease Study. J Hypertens 2014; 32: 1435–1443.
• 13.    Cortez-Cooper  MY,  Supak  JA,  Tanaka  H. A new device for automatic measurements of arterial stiffness and ankle-brachial index. Am J Cardiol 2003; 91: 1519–1522, A1519.
• 14.    Lahoz  C,  Garcia-Fernandez  T,  Barrionuevo  M,  Vicente  I,  Gonzalez-Alegre  T,  Mostaza  JM. Differences in the ankle-brachial index in the general population after 4 years of follow-up. Vasa 2013; 42: 112–119.
• 15.    Richart  T,  Kuznetsova  T,  Wizner  B,  Struijker-Boudier  HA,  Staessen  JA. Validation of automated oscillometric versus manual measurement of the ankle-brachial index. Hypertens Res 2009; 32: 884–888.
• 16.    Kennedy  M,  Solomon  C,  Manolio  TA,  Criqui  MH,  Newman  AB,  Polak  JF, et al. Risk factors for declining ankle-brachial index in men and women 65 years or older: The Cardiovascular Health Study. Arch Intern Med 2005; 165: 1896–1902.
• 17.    Fujiwara  T,  Saitoh  S,  Takagi  S,  Ohnishi  H,  Ohata  J,  Takeuchi  H, et al. Prevalence of asymptomatic arteriosclerosis obliterans and its relationship with risk factors in inhabitants of rural communities in Japan: Tanno-Sobetsu study. Atherosclerosis 2004; 177: 83–88.
• 18.    Usui  T,  Ninomiya  T,  Nagata  M,  Doi  Y,  Hata  J,  Fukuhara  M, et al. Albuminuria as a risk factor for peripheral arterial disease in a general population: The Hisayama study. J Atheroscler Thromb 2011; 18: 705–712.
• 19.    Fowkes  FG,  Housley  E,  Cawood  EH,  Macintyre  CC,  Ruckley  CV,  Prescott  RJ. Edinburgh Artery Study: Prevalence of asymptomatic and symptomatic peripheral arterial disease in the general population. Int J Epidemiol 1991; 20: 384–392.
• 20.    Smith  FB,  Lee  AJ,  Price  JF,  van Wijk  MC,  Fowkes  FG. Changes in ankle brachial index in symptomatic and asymptomatic subjects in the general population. J Vasc Surg 2003; 38: 1323–1330.
• 21.    Aboyans  V,  Criqui  MH,  Denenberg  JO,  Knoke  JD,  Ridker  PM,  Fronek  A. Risk factors for progression of peripheral arterial disease in large and small vessels. Circulation 2006; 113: 2623–2629.
• 22.    Ueshima  H,  Sekikawa  A,  Miura  K,  Turin  TC,  Takashima  N,  Kita  Y, et al. Cardiovascular disease and risk factors in Asia: A selected review. Circulation 2008; 118: 2702–2709.
• 23.    Allison  MA,  Cushman  M,  Solomon  C,  Aboyans  V,  McDermott  MM,  Goff  DC, et al. Ethnicity and risk factors for change in the ankle-brachial index: The Multi-Ethnic Study of Atherosclerosis. J Vasc Surg 2009; 50: 1049–1056.
• 24.    Nichols  WW,  O’Rourke  M,  Vlachopoulos  C. McDonald’s blood flow in arteries: Theoretical, experimental and clinical principles, 6th edn. London, UK: Arnold, 2005.
• 25.    Hiatt  WR,  Hoag  S,  Hamman  RF. Effect of diagnostic criteria on the prevalence of peripheral arterial disease: The San Luis Valley Diabetes Study. Circulation 1995; 91: 1472–1479.
• 26.    London  GM,  Guerin  AP,  Pannier  B,  Marchais  SJ,  Stimpel  M. Influence of sex on arterial hemodynamics and blood pressure: Role of body height. Hypertension 1995; 26: 514–519.
• 27.    Wattanakit  K,  Folsom  AR,  Duprez  DA,  Weatherley  BD,  Hirsch  AT. Clinical significance of a high ankle-brachial index: Insights from the Atherosclerosis Risk in Communities (ARIC) Study. Atherosclerosis 2007; 190: 459–464.
• 28.    Avolio  AP,  Chen  SG,  Wang  RP,  Zhang  CL,  Li  MF,  O’Rourke  MF. Effects of aging on changing arterial compliance and left ventricular load in a northern Chinese urban community. Circulation 1983; 68: 50–58.
• 29.    Wilkins  JT,  McDermott  MM,  Liu  K,  Chan  C,  Criqui  MH,  Lloyd-Jones  DM. Associations of noninvasive measures of arterial compliance and ankle-brachial index: The Multi-Ethnic Study of Atherosclerosis (MESA). Am J Hypertens 2012; 25: 535–541.
• 30.    Tabara  Y,  Igase  M,  Kido  T,  Ochi  N,  Miki  T,  Kohara  K. Composition of lower extremity in relation to a high ankle-brachial index. J Hypertens 2009; 27: 167–173.