Predictors of Large and Small Artery Elasticity in Healthly Subjects 9 to 89 Years Old

Abstract

Background

We identified demographic variables, cardiovascular risk factors, and ambulatory activity measures that predict large and small artery elasticity in apparently healthy subjects between 9 and 89 years of age.

Methods

480 subjects were assessed on large artery elasticity index (LAEI), small artery elasticity index (SAEI), demographic measures, cardiovascular risk factors, and daily ambulation during seven consecutive days. All possible regression and Mallow's Cp were used to select multivariate models for prediction of LAEI and SAEI.

Results

In subjects 20 years of age and younger, LAEI model (R² = 0.25, p < 0.001) included age, average ambulatory cadence, and obesity. SAEI model (R² = 0.39, p < 0.001) contained body mass index (BMI), maximum daily ambulatory cadence for 30 continuous minutes, age, and total ambulatory strides. In subjects between 21 and 50 years, LAEI model (R² = 0.41, p < 0.001) included systolic blood pressure (SBP), gender, race, and diastolic blood pressure (DBP). SAEI model (R² = 0.42, p < 0.001) contained gender, BMI, DBP, race, dyslipidemia, and SBP. In subjects older than 50 years, LAEI model (R² = 0.54, p < 0.001) included SBP, gender, age, and BMI. SAEI model (R² = 0.45, p < 0.001) contained gender, age, BMI, DBP, current smoking, and SBP.

Conclusions

Daily ambulatory activity, particularly cadence of 30 continuous minutes of ambulation, is positively associated with arterial elasticity in children and adolescents. In contrast, the predominant factors related to the decline in arterial elasticity in adults are blood pressure and age.

Introduction

Early detection of vascular dysfunction is important to better identify those at higher risk for subsequent cardiovascular outcomes, and to modify their clinical course by utilizing appropriate medical and behavioral interventions. Reduced arterial elasticity, also expressed as reduced compliance or increased stiffness, is a non-invasive marker predictive of cardiovascular events.^{1, 2} Arterial elasticity is reduced with modifiable cardiovascular risk factors such as smoking,^3-5 hypertension,^{1, 4, 6-9} increased cholesterol,^{1, 6, 7, 9} elevated low-density lipoprotein cholesterol,^{5, 6, 9, 10} increased triglycerides,^{4, 5, 10} diabetes,^{1, 11} elevated insulin^{4, 10} and glucose^{7, 10} levels, increased high-sensitive C-reactive protein,⁷ and increased levels of endothelial biomarkers P-selectin and urinary albuminuria.⁴ In contrast, exercise increases arterial elasticity,^{9, 12-15} with self-reported vigorous physical activity, as opposed to light-to-moderate activity, having a protective effect in a longitudinal follow-up of young adults.¹⁶ Arterial elasticity may be an early marker for cardiovascular dysfunction and disease.^{1, 17}

In addition to modifiable cardiovascular risk factors and endothelial biomarkers, arterial elasticity is associated with age. We previously found that in subjects free of cardiovascular risk factors, arterial elasticity increased with age in children, adolescents, and young adults up to age 30, and declined thereafter, particularly after 50 years of age.¹⁸ The age-related decline in arterial elasticity supported previous investigations in adults,^{4, 6, 7, 19-21} and was the first to describe an increase in young subjects.¹⁸ Few previous studies have utilized multivariate techniques to evaluate the influence of demographic and risk factor measures on arterial elasticity in different age groups,^{4, 8, 20} and none have objectively measured daily physical activity levels as a potential covariate of arterial elasticity.

The purpose of this study is to identify demographic variables, cardiovascular risk factors, and ambulatory activity measures that predict large and small artery elasticity in apparently healthy subjects between 9 and 89 years of age.

Methods

Subjects

A total of 480 subjects between 9 and 89 years of age participated. The subjects were recruited by local newspaper advertisements, university email advertisements, and informational flyers distributed in Oklahoma City and surrounding areas. Subjects were excluded for having a history of coronary artery disease, stroke, congestive heart failure, peripheral arterial disease, chronic obstructive pulmonary disease, renal disease, liver disease, and active cancer due to their potential confounding influences on the outcome measures. Subjects agreed to participate by signing the informed consent form approved by the Institutional Review Boards at the University of Oklahoma, and the University of Oklahoma Health Sciences Center.

Measurements

Medical Screening

To begin the study, subjects were evaluated during a medical history and physical examination. Demographic information, height, weight, body mass index, cardiovascular risk factors, co-morbid conditions, and a list of current medications were obtained.

Diastolic Pulse Contour Analysis (PCA)

Arterial elasticity measurements were obtained in the morning following an overnight fast of at least eight hours, and prior to engaging in any strenuous physical activity. The large artery elasticity index (LAEI) and small artery elasticity index (SAEI) were obtained by an HDI/Pulsewave™ CR-2000 Cardiovascular Profiling System (Hypertension Diagnostic, Inc., Eagan, Minnesota, USA) following 5 to 10 minutes of rest in the supine position as previously described.^{19, 22-24} To convert values to whole numbers, the units for LAEI (ml × mmHg⁻¹) were multiplied by 10 and the units for SAEI (ml × mmHg⁻¹) were multiplied by 100. An appropriately sized blood pressure cuff was placed around the subject's left upper-arm, and a rigid plastic wrist stabilizer was placed on the subject's right wrist to minimize wrist movement and stabilize the radial artery during the measurement. An Arterial Pulsewave™ Sensor was placed on the skin directly over the radial artery at the point of the strongest pulse, while the arm rested in a supine position. The non-invasive acoustic sensor was adjusted to the highest relative signal strength, and arterial waveforms were recorded for 30 seconds and the diastolic portion was digitized at 200 samples per second to determine LAEI and SAEI values,⁷ which assess the elasticity of the large and small arteries throughout the arterial system. Measurements were averaged over three continuous 30-second trials. The test-retest intraclass reliability coefficient is R = 0.87 for LAEI and R = 0.83 for SAEI.²²

Ambulatory Activity Monitoring

Instrumentation and Procedures

Daily ambulatory activity was assessed using a step activity monitor (Step Watch 3, Cyma Inc., Mountlake Terrace, WA) as previously described.²⁵ Ambulatory activity was measured during seven consecutive days in which subjects were instructed to wear the monitor during waking hours and to remove it before retiring to bed. The step activity monitor was attached to the right ankle above the lateral malleolus using elastic Velcro straps, and continuously recorded the number of steps taken on a minute-to-minute basis. The accuracy of the step activity monitor exceeds 99 ± 1%.²⁵

Variables Obtained from the Step Activity Monitor

The step activity monitor records the number of ambulatory strides taken per minute for each minute throughout a 24-hour period. After downloading data from the step activity monitor to a computer, the software program displays the number of strides taken and the number of minutes spent ambulating each day, from which an average daily cadence is calculated. Any minute in which ambulation occurred (ie, at least one stride was recorded) was defined as an active minute, and the total number of strides were divided by the total number of active minutes to yield an average cadence of ambulation. The daily ambulatory strides and time are further analyzed by the software program, and are quantified into the maximum daily ambulatory cadences for 20 and 30 continuous minutes (i.e., the highest ambulatory cadence for 20 and for 30 continuous minutes each day). All of these outcome measures are recorded and averaged for each day, and then the daily averages are averaged over the seven-day monitoring period. In apparently healthy adults, the test-retest intraclass reliability coefficient for the measurement of total daily strides and total daily minutes of activity over the 7-day period are R = 0.94 and R = 0.91, respectively.²⁵ The intraclass reliability coefficients for the remaining variables pertaining to daily ambulatory cadences range from R = 0.83 to R = 0.94.²⁵

Statistical Analyses

In our recent publication both LAEI and SAEI were shown to have a strong association with age across the ages from 9 to 77, but the association was far from linear.¹⁸ However, three age intervals were found over which the association was primarily linear. A positive linear trend was demonstrated from early age into the twenties, followed by a slightly negative trend up to 50 years, and then a much larger in magnitude negative trend to 77 years. To obtain age intervals for which regression was primarily linear for both LAEI and SAEI, polynomial regression models were first fit on the interval of 90 to 80 years, then the interval was lengthened by reducing lower bound in steps of 10 years until a significant (p<0.10) curvilinear component was detected for either LAEI or SAEI. The first detection occurred at extension from 50 years to 40 years and, therefore, 50 years was selected as lower bound for the older group, and as a new start for continuing the procedure for the next age interval. The next detection occurred at extension to 10 years. Thus, 20 years was selected as a lower bound for the next age interval. No further curvilinear components were detected from 20 to 9 years. This was the rationale for partitioning subjects into three age groups, less than 21, 21 to less than 51, and 51 to less than 90 years. Data were summarized within each group and means of measurement variables compared across groups using one way ANOVA. Dichotomized variables were compared using Chi Square analysis. Within each age group, an All Possible Regression procedure was used to obtain models and associated Mallow's Cp statistics for multiple linear regression of LAEI and also for SAEI. The model with Cp nearest to ideal value of number of variables in model +1 and with all variables significant at alpha = 0.10 was selected. All analyses were performed using the NCSS statistical package.

Results

The iterative examination of regression of LAEI and SAEI on age resulted in the following three age intervals: 9 to 20 years, 21 to 50 years, and 51 to 89 years. No significant curvilinear component was detected within each age group. For LAEI, the corresponding slopes of regression line for the three respective age groups were, 0.8140 (p < 0.01), -0.0146 (p = 0.76), and -0.1853 (p < 0.01) (ml × mmHg⁻¹) × 10 per year. For SAEI, the slopes were .3780 (p < 0.01), -0.0709 (p < 0.01), and -0.922 (p < 0.01) (ml × mmHg⁻¹) × 100 per year.

The clinical characteristics of the subjects are displayed in Table 1. The three age groups consisted of similar percentages of males and females, whereas the oldest group had the highest percentage of Caucasians and the youngest group had the lowest percentage (p < 0.001). Significant group differences existed for all remaining variables, except current smoking, with the oldest group typically having the least favorable values. The youngest group had the most favorable values for systolic blood pressure, diastolic blood pressure, and percentage with hypertension, whereas they had similar or worse measures than the middle-aged group for obesity, body mass index, LAEI, SAEI, and ambulatory activity measures.

Table 1.

Clinical characteristics of younger, middle-aged, and older subjects. Values are means (SD) or percentages.

Variables	Subjects 9 to 20 Years of Age (N = 99)	Subjects 21 to 50 Years of Age (N = 127)	Subjects 51 to 89 Years of Age (N = 254)	P Value
Age (years)	14 (2)	34 (10)	66 (10)	---------
Weight (kg)	70.6 (30.1)	77.3 (19.4)	82.0 (18)	<0.001
Body Mass Index (kg/m2)	26.5 (9.6)	26.7 (6.5)	28.8 (5.6)	<0.001
Systolic Blood Pressure (mmHg)	113 (11)	121 (16)	136 (20)	<0.001
Diastolic Blood Pressure (mmHg)	59 (7)	69 (10)	76 (10)	<0.001
LAEI (ml × mmHg−1) × 10	14.5 (4.5)	18.1 (5.5)	14.0 (4.9)	<0.001
SAEI (ml × mmHg−1) × 100	8.2 (2.5)	8.1 (3.0)	4.5 (2.4)	<0.001
Total Strides (strides/day)	4923 (1669)	5433 (2035)	4369 (1881)	<0.001
Total Ambulatory Time (min/day)	359 (94)	350 (99)	319 (111)	0.001
Average Daily Ambulatory Cadence (strides/min)	14.1 (2.3)	15.5 (3.6)	14.0 (4.5)	<0.001
Maximum Daily Ambulatory Cadence for 20 Continuous minutes (strides/min)	27.7 (6.8)	30.7 (9.7)	24.6 (11.2)	0.001
Maximum Daily Ambulatory Cadence for 30 Continuous minutes (strides/min)	23.7 (6.2)	26.7 (9.9)	21.2 (10.8)	<0.001
Gender (% male)	52	42	47	0.50
Race (% Caucasian)	37	64	85	<0.001
Current Smoking (% yes)	-------	15	11	0.46*
Diabetes (% yes)	5	3	11	0.03
Hypertension (% yes)	3	12	46	<0.001
Dyslipidemia (% yes)	--------	21	45	<0.001*
Obesity (% yes)	33	21	36	0.01

^*Younger Group not included

In subjects 20 years of age and younger, LAEI was positively associated with age (p < 0.001), average daily cadence of ambulation (p = 0.01), and obesity (p = 0.07) (Table 2). After adjustment, average daily cadence explained 6% of the variance in LAEI. The predictors of SAEI were age (p = 0.001), BMI (p < 0.001), maximum daily ambulatory cadence for 30 continuous minutes (p < 0.001), and total ambulatory strides (p = 0.01). After adjustment, age accounted for average daily ambulatory cadence for 30 continuous minutes explained 15% of the variance in SAEI.

Table 2.

Regression coefficient summary for independent variables used in regression models for large artery elasticity index (LAEI) and small artery elasticity index (SAEI) of subjects 9 to 20 years of age.

Variables	Predictors	Regression Coefficient	95% CI	R2†	P Value
LAEI * (ml × mmHg−1) × 10	Age (years)	0.7185	0.3527 to 1.0843	0.1433	< 0.001
	Average_Cadence (strides/min)	0.4864	0.1429 to 0.8299	0.0800	0.01
	Obesity	1.6272	−0.1214 to 3.3758	0.0362	0.07
	Intercept	−2.5714	−9.3313 to 4.1885
SAEI ** (ml × mmHg−1) × 100	BMI (kg/m2)	0.1140	0.0704 to 0.1575	0.2310	< 0.001
	Maximum 30-minute cadence (strides/min)	0.2259	0.1149 to 0.3368	0.1538	< 0.001
	Age (years)	0.3329	0.1431 to 0.5228	0.1189	0.001
	Total Strides (strides/day)	−0.0006	−0.0010 to −0.0001	0.0733	0.01
	Intercept	−1.9692	−5.3925 to 1.4541

^*Overall model results for LAEI: R² = 0.2497, p < 0.001, Cp = 4.232944.

^**Overall model results for SAEI: R² = 0.3896, p < 0.001, Cp = 5.513766.

^†Partial R² adjusted for all other predictors.

In subjects between 21 and 50 years of age, the predictors of LAEI were systolic blood pressure (p < 0.001), race (p = 0.001), gender (p < 0.001), and diastolic blood pressure (p = 0.03) (Table 3). After adjustment, systolic blood pressure accounted for 22% of the variance in LAEI. The predictors of SAEI were diastolic blood pressure (p = 0.01), body mass index (p < 0.001), gender (p = 0.001), dyslipidemia (p = 0.02), race (p = 0.01), and systolic blood pressure (p = 0.09). After adjustment, diastolic blood pressure explained 40% of the variance in SAEI.

Table 3.

Regression coefficient summary for independent variables used in regression models for large artery elasticity index (LAEI) and small artery elasticity index (SAEI) of subjects 21 to 50 years of age.

Variables	Predictors	Regression Coefficient	95% CI	R2†	P Value
LAEI * (ml × mmHg−1) × 10	SBP (mmHg)	−0.2736	−0.3660 to −0.1812	0.2212	< 0.001
	Female Gender	−3.7353	−5.3785 to −2.0921	0.1434	< 0.001
	Non-Caucasian Race	−2.8548	−4.4550 to −1.2546	0.0935	0.001
	DBP (mmHg)	0.1511	0.0129 to 0.2894	0.0373	0.03
	Intercept	44.0849	37.8113 to 50.3586
SAEI ** (ml x× mmHg−1) × 100	DBP (mmHg)	−0.0976	−0.1727 to −0.0225	0.0527	0.01
	BMI (kg/m2)	0.1333	0.0541 to 0.2126	0.0853	0.001
	Female Gender	−1.6134	−2.5311 to −0.6957	0.0924	0.001
	Dyslipidemia	−1.8854	−3.3957 to −0.3751	0.0488	0.02
	Non-Caucasian Race	1.1429	0.2465 to 2.0393	0.0508	.01
	SBP (mmHg)	−0.0455	−0.0980 to 0.0070	0.0242	.09
	Intercept	17.2615	13.4523 to 21.0706

^*Overall model results for LAEI: R² = 0.4107, p < 0.001, Cp = 4.087225.

^**Overall model results for SAEI: R² = 0.4166, p < 0.001, Cp = 7.145884.

^†Partial R² adjusted for all other predictors.

In subjects older than 50 years of age, the predictors of LAEI were systolic blood pressure (p < 0.001), gender (p < 0.001), age (p < 0.001), and body mass index (p < 0.001) (Table 4). After adjustment, systolic blood pressure accounted for 26% of the variance in LAEI. The predictors of SAEI were age (p < 0.001), gender (p < 0.001), diastolic blood pressure (p < 0.01), systolic blood pressure (p = 0.07), body mass index (p < 0.001), and current smoking (p = 0.01). After adjustment, age explained 18% of the variance in SAEI.

Table 4.

Regression coefficient summary for independent variables used in regression models for large artery elasticity index (LAEI) and small artery elasticity index (SAEI) of subjects 50 to 89 years of age.

Variables	Predictors	Regression Coefficient	95% CI	R2†	P Value
LAEI * (ml × mmHg−1) × 10	SBP (mmHg)	−0.1337	−0.1588 to −0.1086	0.3055	< 0.001
	Female Gender	−4.0473	−4.8851 to −3.2096	0.2663	< 0.001
	Age (years)	−0.1286	−0.1744 to −0.0827	0.1090	< 0.001
	BMI (kg/m2)	0.1514	0.0735 to 0.2292	0.0555	< 0.001
	Intercept	38.4112	34.0064 to 42.8160
SAEI ** (ml × mmHg−1) × 100	Age (years)	−0.0848	−0.1135 -0.0561	0.1377	< 0.001
	Female Gender	−1.9806	−2.4612 to −1.4999	0.2370	< 0.001
	DBP (mmHg)	−0.0549	−0.0922 to −0.0175	0.0380	< 0.01
	SBP (mmHg	−0.0181	−0.0375 to 0.0012	0.0158	.07
	BMI (kg/m2)	0.0772	0.0343 to 0.1201	0.0559	< 0.001
	Smoking	−1.0931	−1.8786 to −0.3075	0.0342	0.01
	Intercept	15.4828	12.3821 to 18.5836

^*Overall model results for LAEI: R² = 0.5416, p < 0.001, Cp =0.526386.

^**Overall model results for SAEI: R² =0.4470, p < 0.001, Cp =7.040744

^†Partial R² adjusted for all other predictors.

Discussion

A novel finding in this study was that objectively measured daily ambulatory cadence was associated with LAEI and SAEI in children and adolescents. The beneficial influence of daily ambulatory cadence on arterial elasticity in young subjects may be evident because cardiovascular risk factors have had less time to elicit chronic impairments in the cardiovascular system typically seen in older adults. Our study supports a previous finding that aerobic fitness, estimated by a 20 meter incremental shuttle run, is associated with large and small artery compliance in children between 9 and 11 years of age,¹⁵ and a more recent longitudinal investigation showing that self-reported lifetime vigorous physical activity is a predictor of brachial and femoral arterial compliance in young adults.¹⁶ In the current study, average daily cadence is a predictor of LAEI, and the maximum daily ambulatory cadence for 30 continuous minutes is a predictor of SAEI. Both of these objective measures reflect an intensity component to ambulation, which may be a surrogate marker of aerobic fitness as measured previously.¹⁵ A public health implication is that higher ambulatory cadence (ie, exercise intensity) performed continuously for at least 30 minutes daily is a positive predictor of microvascular function in children, adolescents, and young adults, and that higher average ambulatory cadence maintained throughout the course of a day is predictive of favorable macrovascular function. Although these findings support the beneficial effects of continuous ambulation of at least 30 minutes duration on arterial elasticity, a dose-response pattern cannot be determined from the design of our study.

This study supports our previous report that LAEI and SAEI increase with age in children, adolescents, and young adults.¹⁸ These results suggest that the elasticity of the macrovasculature and microvasculature increase with physical maturation. This supports the findings that large artery compliance is associated with height in children, and that small artery compliance is related to Tanner stages.¹⁵ We also noted that body mass index was positively associated with SAEI, possibly providing further evidence of the influence of physical maturation. However, body mass index also was a predictor of SAEI in adults between 21 and 50 years of age, suggesting that additional factors besides physical maturation may explain this relationship in children. Previous work supports these findings, as body weight positively correlates with small artery compliance in children,¹⁵ and obese children and adolescents have slower pulse wave velocity (ie, higher arterial elasticity) than lean controls.²⁶

Blood pressure, gender, and race were the primary predictors of LAEI and SAEI in subjects between 21 and 50 years of age. Systolic blood pressure negatively correlated with LAEI, whereas diastolic blood pressure negatively correlated with SAEI. These findings agree with previous reports showing that elevated blood pressure reduces both LAEI and SAEI in subjects between 25 and 89 years of age,¹ and in subjects between 27 and 42 years of age in the CARDIA Study.⁴ However, not all studies consistently show these trends.^6-8 The strength of association between diastolic blood pressure and SAEI in the current study is noteworthy, as diastolic pressure explained 40% of the variance in SAEI after adjustment for other covariates, suggesting that our subjects between 21 and 50 years of age had high peripheral resistance. Additional predictors of both LAEI and SAEI were gender and race, as women and non-Caucasians had lower values. These findings agree with previous work showing that women had lower LAEI and SAEI,⁶ and a recent report that African-Americans had lower SAEI, but not LAEI, than Caucasians in subjects between 45 and 84 years of age in the MESA study.²⁷

Blood pressure and gender also were primary predictors of LAEI and SAEI in subjects older than 50 years of age. Systolic blood pressure was the primary predictor of LAEI, and diastolic blood pressure was the main predictor of SAEI. Furthermore, older women had lower values of LAEI and SAEI than older men. In contrast to the subjects who were 21 to 50 years old, age was a significant predictor for the decline in both LAEI and SAEI in the older group, and race was not. Age was the strongest predictor for SAEI, suggesting that age negatively impacts the microcirculation to an even greater extent than the macrocirculation. The age-related declines in LAEI and SAEI support previous findings from our lab and those of others.^{4, 6, 7, 18-21} Age-related decreases in LAEI occur even in the absence of coexisting diseases^{18, 19, 21} due to structural changes within the arterial wall. These changes include increased fragmentation and decreased density of elastin,²⁸ increased concentration of collagen,²⁹ hypertrophy of vascular smooth muscle,²⁹ and decreased nitric oxide production.³⁰ Less is known about the age-related changes in SAEI. We have found that in subjects free of cardiovascular risk factors, SAEI sharply drops in adults beyond 50 years of age.¹⁸

There are limitations to this study. The cross-sectional research design of this study does not allow causality to be established when examining the relationship between arterial elasticity and the multiple predictor variables. A self-selection bias may also exist regarding study participation. Another limitation is that diastolic PCA is a non-invasive technique to determine elasticity of the large and small arteries.^{7, 31, 32} However, this technique has been validated with invasive measures of arterial compliance³¹ and provides reliable measurement of arterial elasticity.²² A limitation is also associated with the assessment of physical activity. Although we directly measured ambulatory activity with a step activity monitor, non-ambulatory activity such as upper extremity movements and resistance training are not recorded by the monitor, and self-reported daily activities was not assessed. However, we believe non-ambulatory activity had minimal impact on the study results because ambulation is a large component of daily activity, and very few subjects indicated during the medical history that they were actively engaged in any type of resistance training exercise. An additional limitation is that the medication data obtained during the medical history was only used to better characterize subjects for hypertension, dyslipidemia, and diabetes. We believe that antihypertensive medications had minimal influence on the study results because none of the younger subjects and only a few of the middle-aged subjects were taking medications, and fewer than half of the hypertensive older subjects were taking medications. Further, hypertension was not correlated with LAEI or SAEI in any of the groups. A limitation associated with the statistical models is that other selection methods and criteria might produce slightly different models since some of the variables are highly correlated with others (e.g., BMI and obesity), and substitution of one for the other may not greatly change the model R-square. A final limitation is that the present findings only apply to apparently healthy subjects.

In conclusion, daily ambulatory activity, particularly the cadence of at least 30 continuous minutes of ambulation and the average daily cadence, are positively associated with arterial elasticity in children and adolescents. In contrast, the predominant factors related to the decline in arterial elasticity in adults are blood pressure, female gender, non-Caucasian race, and age.