A study of the VaSera arterial stiffness device in US patients

Abstract

The cardio‐ankle vascular index (CAVI) represents a promising index of arterial stiffness. However, neither the CAVI measure nor its measurement device, the VaSera, have undergone general testing in a North American clinical setting. To begin the process of collecting normal values in the United States, we studied 20 male and 28 female volunteers without reported cardiovascular or renal disease and no history of smoking. Their CAVIs, ankle‐brachial indices (ABIs), and four‐limb blood pressures were measured in three positions: supine, 7° Trendelenburg, and 7° reverse Trendelenburg. In addition, the ABI function was validated against an established ABI measurement technique. Position was found to affect CAVI and other hemodynamic parameters, indicating that CAVI is not robust to slight positional variations. No differences were found in the blood pressure between arms or legs (interbrachial or interankle), supporting recent findings from meta‐analyses and studies but contradicting other work. This study represents an early step in bringing the VaSera device and its CAVI measurement into clinical practice.

1. Introduction

The cardio‐ankle vascular index (CAVI) is an easy‐to‐measure proxy of arterial stiffness derived from arterial pulse wave velocity and measured by the VaSera device (Fukuda Denshi, Tokyo, Japan).1 Its main advantage over the current standard arterial stiffness measure, pulse wave velocity, is its theoretical independence of blood pressure (BP), allowing for accurate measure in normotensive and hypertensive individuals.2 Although in practice the CAVI is somewhat correlated with both diastolic and systolic BP, the measure has been independently correlated with the progression of renal disease, cardiovascular pathology, and stroke events.3, 4, 5, 6, 7

The CAVI measure has also shown utility as a diagnostic tool. For instance, CAVIs have been documented to fall (ie, arterial stiffness falls) as obese patients lose weight.8 Indeed, CAVI in conjunction with other measures of arterial stiffness can be used to differentiate between different pathologies of the arterial system. For example, in diabetes, recent data indicate that peripheral arteries stiffen while the carotid and other larger arteries do not.9

However, standardized reference values for CAVI in an adult US population are lacking. The device has been tested on Russian, Mongolian, Chinese, Taiwanese, Japanese, and Czech populations, but among these populations, only the Japanese have been extensively studied with great detail.10, 11, 12, 13, 14 Moreover, differences have been found among national average values, thus it is difficult to apply these results in a US population without first collecting population‐specific normal values.

1.1. Interlimb BP difference

Interarm differences in patients with healthy BP have been documented for nearly 100 years through various measurement methods, most recently through 24‐hour ambulatory BP monitoring.15, 16, 17 Indeed, measurement technique is critical, as sequential measurements can overestimate the prevalence and magnitude of interarm differences, and direct, invasive measurement reduces differences.15, 18 However, accurate measurement is critical as consistent interarm BP differences can indicate subclavian stenosis, and clinicians can miss hypertension if screening BPs are performed in only one arm.18, 19, 20, 21 Indeed, for systolic BP (SBP) differences >10 mm Hg, there is mounting evidence that interarm differences can help predict mortality, even without vascular imaging studies.22, 23, 24

However, the significance of smaller interarm differences in the absence of identified vascular findings remains unclear. These smaller differences have been documented in healthy populations,16 but other studies have not found a statistical difference.25 Some attribute a large part of the interarm difference to white‐coat hypertension,26 although those findings have been disputed.27 Others contend that differences are only consistent in vascular obstructive disease, even in populations with other chronic conditions.20 Nonetheless, the prevalence, cause, and implications of these differences remain unclear.

A related measure that is beginning to find clinical utility is the difference between ankle BPs. Although ankle‐brachial index (ABI) is a well‐established predictor of mortality and cardiovascular morbidity,28 the differences in leg BPs are beginning to be explored. Interankle BPs (taken from a four‐cuff simultaneous BP apparatus) have been shown to predict mortality in elderly Chinese populations,29 and many other risk factors for cardiovascular disease and mortality have been found to correlate with interankle difference.30 Furthermore, recent work has begun to associate interankle BP differences with progression of diabetic nephropathy31 as well as cardiovascular and all‐cause mortality in patients.32

1.2. Positional changes and measurement

Although clinically BPs are taken with patients in a sitting position, measurements of arterial stiffness are typically taken with patients in the supine position. It is unclear whether position makes a difference in arterial stiffness parameters, such as pulse wave velocity, as studies and theoretical work are inconsistent.33, 34 There are similarly conflicting results among studies of supine and sitting brachial BPs.35 Interestingly, changes in arterial stiffness have been implicated in some positional BP pathologies, such as orthostatic hypotension.36 Nonetheless, positional changes have been used to assess general cardiovascular function,37 and differences in positional responses of BP have been noted when comparing type I diabetic patients with nondiabetic, matched controls.38 Position‐dependent responses to hemodynamic drugs, such as salbutamol and nitroglycerin have also been noted.39

In this study, the limb BPs, CAVIs, ABIs, heart rates, and brachial and malleolar arterial wave forms of a group of healthy individuals were measured using a four‐cuff simultaneous BP VaSera device. Parameters were measured in patients in the supine, 7° Trendelenburg, and 7° reverse Trendelenburg positions to assess for changes in hemodynamic parameters. Supine ABIs were assessed using Doppler to further verify readings from the VaSera.

2. Methods

2.1. Institutional review and screening

In this single‐center study, adult participants, aged 18 to 65 years, were recruited from the staff, faculty, and students of the University of Pennsylvania. This study was approved by the institutional review board of the University of Pennsylvania, and all participants gave informed consent. Exclusion criteria included a history of smoking, hypertension, cardiovascular disease, and renal disease. Eligibility was determined by self‐report using a standard medical history checklist.

2.2. VaSera measurement

Before measurements were taken, participants comfortably rested supine for 10 minutes in a Stryker bed (Stryker Instruments, Kalamazoo, MI, USA) that showed a digital display of head‐to‐foot position in degrees (horizontal=0°). Oscillometric BPs, ABIs, CAVIs, and heart rates were measured using the VaSera VS‐1500AU (Fukuda Denshi) according to the manufacturer's instructions. Two measurements were taken in each position (a third measurement was taken if CAVIs differed by more than 10%). The supine measurement was taken first and the order of the Trendelenburg and reverse Trendelenburg measurements were alternated. When shifting a participant into a new position, they were asked to lie quietly for 1 minute before measurements commenced. CAVI was calculated with the machine's internal algorithm. Both the simultaneous (all four limbs assayed at the same time) and sequential (right limbs followed by left limbs) settings for BP readings were used. Cuffs were placed approximately 2 cm above the intercubital fossa on the arms and approximately 2 cm above the medial malleolus on the legs. Parallel wrapping of cuffs was preferred, but if not possible, spiral wrapping was used. Cuff size was adjusted according to arm size as per the manufacturer's instructions. Data were transferred for analysis using the VaSera Data Management Software VSS‐10U (Fakuda Denshi), and any data not captured by the software was manually entered into data files by one user.

2.3. ABI Doppler measurement

We incorporated ABI measurements into the protocol by an amendment at about the halfway point in the study. ABI was measured in accordance with the American Heart Association's most current scientific statement on the parameter, and both left and right brachial, posterior tibial, and dorsalis pedis pressures were obtained.40 ABI measurements were taken after the VaSera measurement. A Baumanometer sphygmomanometer (Hawksley & Sons, Lancing, West Sussex, UK) and Tycos adult size cuff (Welch Allyn, Skaneateles Falls, NY, USA) were used to assess pressures. The return of arterial sounds (and hence an estimate of systolic pressure) was determined using the LifeDop Doppler (Wallach Surgical Devices, Trumbull, CT, USA). Doppler signals were enhanced using Aquasonic 100 Ultrasound Gel (Parker Laboratories, Fairfield, NJ, USA).

2.4. Data analysis

Initial data processing was performed in Excel 2013 (Microsoft, Redmond, WA, USA). Statistical analysis was performed in R 3.1.0 (R Foundation for Statistical Computing, Vienna, Austria), and the MethComp package (Bendix Carstensen, Gentofte, Denmark) was utilized for the Bland‐Altman plot.

3. Results

Self‐reported demographic characteristics of the data are presented in Table 1 and are notable for the under‐representation of African Americans compared with other groups. In addition, alcohol consumption varied greatly in this population as standard deviations exceeded averages for that metric in both sexes. Analyzing the data by BP, sex, or race did not result in significant differences, although brachial SBP was statistically lower in women (data not shown).

Table 1.

Characteristics of Patients^a

Demographics	No. (% or SD)
	Men (n=20)	Women (n=28)
Raceb
Caucasian	13 (65%)	20 (71%)
Asian	3 (15%)	4 (14%)
South Asian	3 (15%)	1 (4%)
African American	0 (0%)d	3 (11%)
Mixed	1 (5%)d	0 (0%)
Age, y	33 (13)	33 (12%)
18–25b	10 (50%)	11 (39%)
25–40b	5 (25%)	12 (43%)
40–65b	5 (25%)	5 (18%)
Height, cmc	180 (7.5)	160 (7.3)
Weightc	76 (14)	61 (8.6)
Body mass index	24 (3.0)	22 (2.9)
Alcohol consumption (drinks per wk)	2.3 (2.9)	1.6 (1.7)
Handednessb
Right	16 (80%)	25 (89%)
Left	4 (20%)	3 (11%)

^aData are rounded to two significant digits. Percentages are rounded to the nearest ones place.

^bData for race, handedness, and the breakdown of ages presented as number (percentage of sex) rather than number (standard deviation [SD]). Percentages may not add up to 100% due to rounding.

^cSignificant difference as defined by Welch's t test P<.05 when applicable. In the case of height, P=2.2×10⁻⁷. In the case of weight, P=8.1×10⁻⁵.

^dThe individual who identified as mixed was of African and Caucasian descent.

3.1. Positional differences in BPs and arterial stiffness measures

Changes in position did significantly alter ankle BPs and arterial stiffness measures of CAVI and ABI (Figure 1). Ankle pressures were highest in reverse Trendelenburg, when the head was elevated relative to the horizontal and the ankles were lower. Conversely, pressures were lowest in Trendelenburg, when the head was lower relative to the horizontal. On average, between the two extremes of position, Trendelenburg and reverse Trendelenburg, diastolic ankle pressures rose nearly 25 mm Hg (95% confidence interval [CI], 21–28), mean pressures rose approximately 25 mm Hg (95% CI, 20–30), and systolic pressures rose by 26 mm Hg (95% CI, 20–33). The data were similar in the left ankle (data not shown). Brachial pressures did not significantly change throughout the range of positions (Table 1).

Figure 1.

Changes in ankle blood pressure and arterial stiffness with position. All measures increase from Trendelenburg to supine to reverse Trendelenburg. (A) Diastolic, mean, and systolic pressure changes with positional change. Error bars are drawn for systolic pressures. (B) Left and right ankle‐brachial index (ABI) changes with position. Error bars omitted for clarity. (C) Left and right cardio‐ankle vascular index (CAVI) change with position. Error bars omitted for clarity. ***P<.001

3.2. Lack of interlimb differences in BP and hemodynamic parameters

A summary of all BP measurements is presented in Table 2 and remaining measured hemodynamic parameters are presented in Table 3 (arterial stiffness and heart rate). These parameters did not exhibit statistically significant differences in any position. As with positional differences, the measurement setting of the VaSera did not impact the results.

Table 2.

Summary of Limb Blood Pressures in Various Positions^a

Position	Blood Pressure, mean (SD), mm Hg
	Right Arm			Left Arm			Right Ankle			Left Ankle
	Sys	Dia	Mean	Sys	Dia	Mean	Sys	Dia	Mean	Sys	Dia	Mean
Supine	116 (12)	71 (8)	87 (9)	116 (12)	70 (9)	87 (8)	132 (17)	71 (10)	94 (12)	133 (17)	70 (8)	93 (12)
Trendelenburg	113 (10)	68 (7)	85 (8)	113 (11)	68 (8)	85 (8)	117 (15)	57 (8)	79 (10)	117 (16)	56 (7)	78 (11)
Reverse Trendelenburg	114 (12)	70 (8)	86 (8)	113 (11)	70 (8)	86 (8)	143 (16)	81 (10)	104 (12)	143 (16)	81 (10)	103 (12)

Abbreviations: Dia, diastolic pressure; Mean, mean arterial pressure calculated from the VaSera internal algorithm; SD, standard deviation; Sys, systolic pressure.

^aData are rounded to the nearest ones place.

Table 3.

Summary of CAVI Parameters and HR^a

Position	Men, Mean (SD)					Women, Mean (SD)
	R CAVI	L CAVI	R ABI	L ABI	HR	R CAVI	L CAVI	R ABI	L ABI	HR
Supine	6.74 (1.23)	6.70 (1.20)	1.12 (0.10)	1.12 (0.10)	63 (8)	6.24 (0.97)	6.30 (0.97)	1.12 (0.08)	1.12 (0.09)	63 (11)
Trendelenburg	6.10 (1.10)	6.09 (1.07)	1.00 (0.10)	1.01 (0.10)	65 (11)	5.61 (1.00)	5.67 (0.96)	1.02 (0.08)	1.02 (0.09)	63 (12)
Reverse Trendelenburg	7.11 (1.04)	7.05 (0.98)	1.22 (0.09)	1.22 (0.10)	63 (8)	6.73 (0.86)	6.79 (0.85)	1.25 (0.07)	1.24 (0.07)	61 (11)

Abbreviations: L, left; R, right; SD, standard deviation.

Heart rate (HR) is reported to the nearest ones place.

^aCardio‐ankle vascular index (CAVI) and ankle‐brachial index (ABI) are reported to the nearest two decimal places.

3.3. CAVI correlates with age and sex

In multivariate regressions of right and left CAVI with available parameters, only age, sex, and right and left ankle systolic pressures showed any significant correlations (Table 4). CAVI does not appear to correlate with alcohol consumption or other demographic factors when age and sex are factored. CAVI, as an index of arterial stiffness, tends to rise with age and is lower in women.

Table 4.

Cardio‐Ankle Vascular Index Correlates With Age and Sex^a

Variable	Left Cardio‐Ankle Vascular Index			Right Cardio‐Ankle Vascular Index
	Coefficient Estimate	Standard Error	P Value	Coefficient Estimate	Standard Error	P Value
Intercept	220	310	.47	200	310	.51
Age, y	0.039	0.0076	1.0×10 −6	0.039	0.0076	8.5×10 −7
Sex (1=male, 2=female)	−0.66	0.21	.023	−0.75	0.21	.00065
Height, cm	0.0075	0.011	.49	0.0074	0.011	.50
Weight, kg	−0.015	0.0082	.065	−0.016	0.0082	.050
Alcohol consumption, drinks per wk	−0.025	0.033	.45	−0.030	0.034	.36
Position (0=supine, 1=Trendelenburg, 2=reverse Trendelenburg)	−0.020	0.080	.80	−0.014	0.080	.86
Trendelenburg first? (1=yes, 2=0)	0.13	0.13	.33	0.21	0.14	.12
Time and date	−0.0052	0.0074	.48	−0.0048	0.0074	.52
Right brachial systolic	−0.014	0.019	.46	−0.10	0.019	.59
Right brachial mean	0.013	0.031	.68	0.0087	0.031	.78
Right brachial diastolic	−0.022	0.030	.46	−0.018	0.030	.55
Left brachial systolic	−0.0097	0.017	.58	−0.0062	0.017	.72
Left brachial mean	−0.034	0.032	.29	−0.041	0.032	.21
Left brachial diastolic	0.033	0.030	.26	0.040	0.030	.18
Right ankle systolic	0.032	0.015	.036	0.034	0.015	.026
Right ankle mean	0.0066	0.021	.76	0.0058	0.021	.78
Right ankle diastolic	0.032	0.024	.20	0.047	0.025	.059
Left ankle systolic	−0.024	0.013	.077	−0.026	0.013	.048
Left ankle mean	0.012	0.019	.54	0.0020	0.019	.91
Left ankle diastolic	−0.015	0.023	.53	−0.019	0.024	.42

All variables with P values <.05 are presented in bold. Each patient is represented with three data points in this regression since there are three positions possible. Regressions are not particularly altered when each position is considered separately. In fact, only age remains a consistent statistically significant coefficient when each position is considered separately.

^aAll data are rounded to two significant figures.

3.4. Preliminary validation of VaSera ABI function

Although the main purpose of this study was not to validate the VaSera oscillometric BP monitor (and more subjects would be required to conduct a formal validation), a preliminary comparison of machine‐generated ABI data to Doppler ABI was conducted. Because of the late addition into the protocol, only 23 of the 48 participants underwent ABI testing. The ABIs calculated by the VaSera were relatively accurate, as the Bland‐Altman plot for the measurements indicates (Figure 2). The number of measurements two standard deviations away from the mean difference did not exceed the number that would be expected in a normal distribution.

Figure 2.

A Bland‐Altman plot demonstrates no significant difference in measuring ankle‐brachial index (ABI) through Doppler or VaSera. The difference of the measurement between the left and right ABI does not differ significantly from 0. No measurements greatly exceed the second standard deviation line (at least no more than would be expected in a normal distribution). Parameters of the plot: left—average: −0.0083, standard deviation: 0.12%, 2.5% limit: −0.24%, 97.5% limit: 0.27; right—average: 0.010, standard deviation: 0.10%, 2.5% limit: −0.18%, 97.5% limit: 0.20. All values rounded to two significant figures

4. Discussion

To our knowledge, this is the first systemic study of the VaSera device in healthy controls from a US adult population, providing important baseline normative data for subsequent studies of cardiovascular disease and other disorders. Our study observed that CAVI correlates with sex and age and agrees with previous studies of cardiovascular parameters.41, 42 Also to our knowledge, the ABI validation is the first published validation of this feature of the VaSera. We acknowledge that it is underpowered, but it provides support that the VaSera device performed well in this diagnostic maneuver. Considering that the device can obtain BP, CAVI, and ABI data simultaneously, its ability to aid in hemodynamic studies is promising.

The positional studies we undertook were designed to determine the sensitivity of the VaSera device to changes in position. To our knowledge, these types of studies have not been investigated extensively in the past; however, work by Zwain and colleagues43 using much greater tilt angles of 30° and 60° showed changes in CAVI thought to be due to changes in cardiac index with position. The VaSera manufacturer's instructions indicate that patients should lay supine. However, they do not specify whether pillows can be used to maintain patient comfort. As the work on white‐coat hypertension has shown that discomfort can alter BP and other cardiovascular parameters,44 it is important to understand that even a 7° change in position can cause a significant alteration in VaSera output.

However, while the data do show a change in CAVI by position, these changes are not uniform between individuals. While these interindividual changes might represent normal variation, they might also hint at future pathological differences.

When we planned this investigation, we expected to see a small interarm difference in BP values. Although a recent meta‐analysis found no significant differences between brachial BPs in healthy volunteers,16 these studies have largely consisted of measurements at a single time (ie, a clinical visit). Some recent work with ambulatory BP monitoring has found interarm differences over 24 hours, while other ambulatory BP monitoring studies have shown no such difference.16 As the VaSera can take BP measurements automatically and simultaneously in both arms on the same cardiac cycle, this provides evidence supporting a lack of significant interarm differences in BP in individuals without cardiovascular pathology.

The lack of interankle BP differences is to be expected considering the lack of differences in the arms; however, considering that studies have been limited but promising so far, the full implication of our results remains to be seen. Considering that peripheral artery disease (especially in the vasculature of the leg) is a tracked clinical parameter, ankle BPs might be more useful in the future in certain populations, such as those with claudications or other vascular abnormalities.

Because cardiovascular disease is an important cause of mortality in African Americans, we plan to recruit more patients from this group in order to obtain a better picture of CAVI and arterial stiffness in this population. In the meantime, this initial study provides the groundwork for a standard to be developed for the use and interpretation of results from the VaSera device.

Study Limitations

We acknowledge several limitations to our study. Firstly, we were able to perform routine laboratory testing (complete blood cell count, chemistry panel, and lipid profile) and electrocardiography on 31 of the participants we enrolled. The results showed no abnormalities in laboratory data, supporting our screening process that ensured healthy adults were studied (see Table S1). Each participant was studied only once, and we are not able to report on the reproducibility of our results. Next, it is worth noting that CAVI computation uses brachial SBP under the assumption that the BP in the brachial artery might be representative of the whole BP in the artery of the origin of the aorta to the ankle. This is not necessarily the case due to pulse pressure amplification as the pulse wave travels to the periphery. The discrepancy between the brachial artery SBP and the SBP in other arterial segments (such as the lower extremity arteries) was likely exacerbated with changes in tilt, which may underlie the changes in CAVI observed in our study. Finally, African American men are underrepresented in our study, and our overall study group was small.

5. Conclusions

This study presents normative data using the VaSera device and emphasizes the importance of patient position. The ability to simultaneously measure BP and arterial stiffness in all limbs with a relatively simple device provides both valuable research and clinical data. Our study contributes preliminary data needed for establishing normal measurement values, and tentatively confirms the ability to conveniently and dependably track cardiovascular risk in a healthy US adult population. Further studies in US populations with comorbidities are warranted.

Conflicts of Interest

George Maliha: none. Raymond Townsend: grant support from Fukuda Denshi.

Supporting information

 

Maliha G, Townsend RR. A study of the VaSera arterial stiffness device in US patients. J Clin Hypertens. 2017;19:661‐668. 10.1111/jch.12967