Sex-specific associations of reservoir-excess pressure parameters with age and subclinical vascular remodeling

Abstract

Objective:

Central artery reservoir pressure (RP) and excess pressure (XSP) are associated with cardiovascular disease (CVD) events and mortality. However, sex differences in the trajectory of central RP and XSP with advancing age and their relations with vascular markers of subclinical CVD risk are incompletely understood. Therefore, we tested the hypothesis that central RP and XSP would be positively associated with advancing age and vascular markers of subclinical CVD risk in men and women.

Method:

Healthy adults (n=398; aged 18 to 80 years, 60% female) had central (carotid) artery pressure waveforms acquired by applanation tonometry. RP and XSP peaks and integrals were derived retrospectively from carotid pressure waveforms using custom written software. Carotid artery intimal medial thickness (IMT) was measured by ultrasonography and aortic stiffness was determined from carotid-femoral pulse wave velocity (cfPWV).

Results:

RP peak, RP integral and XSP integral were higher with age in both men and women (P<0.05), while XSP peak was lower with age in men (P<0.05). In women, both RP peak (β=0.231, P<0.01) and RP integral (β=0.254, P<0.01) were associated with carotid artery IMT and RP peak was associated with cfPWV (β=0.120, P=0.02) after adjusting for CVD risk factors.

Conclusion:

Central artery RP and XSP were higher with advancing age in men and women, and RP peak was associated with both carotid artery wall thickness and aortic stiffness in women but not men. Central RP peak may provide some insight into sex differences in vascular remodeling and subclinical CVD risk with advancing age in healthy adults.

Introduction

Cardiovascular disease (CVD) remains the leading cause of mortality worldwide and elevated blood pressure (BP) is a primary risk factor for CVD [1,2]. Systolic and diastolic BP are conventionally determined from the peak and trough of the underlying brachial artery BP waveform, however this approach overlooks complex morphological features of the arterial waveform located between systolic and diastolic BP. This limitation of conventionally measured BP has spurred the development of pulse wave analysis, which seeks to extract more detailed hemodynamic indices from the BP waveform morphology that may provide pathophysiological insight into CVD [3]. Indeed, morphological features of the central artery waveform, such as central augmentation index, are associated with clinical CVD events independent of brachial artery cuff systolic and diastolic BP [4]. However, more comprehensive methods of BP waveform analysis exist, such as the reservoir excess pressure model [5], but the extent to which reservoir pressure (RP) and excess pressure (XSP) are associated with subclinical CVD has not been fully explored.

The reservoir excess pressure model describes ventriculo-vascular coupling by deconstructing the BP waveform into a RP and an XSP, the sum of which is equal to the total pressure [5]. RP represents the theoretical minimum hydraulic work required by the left ventricle to eject the stroke volume and XSP represents excess ventricular work that remains once RP has been subtracted from the total pressure [6]. Importantly, RP and XSP are associated with adverse CVD outcomes independent of traditional CVD risk factors including conventional brachial BP, in middle-aged and older adults [7–9]. However, these studies did not examine potential sex differences in the association of RP and XSP with CVD outcomes.

RP and XSP are associated with subclinical damage to the heart [9,10], kidneys [11–13], and brain [14,15] in middle-aged and older adults, indicating that RP and XSP may provide novel hemodynamic information related to ventriculo-vascular interactions and target organ damage with aging. However, whether central RP and XSP are associated with vascular markers of subclinical CVD risk, such as carotid artery intima media thickness (IMT) and aortic stiffness, and whether there are potential sex differences in the relation is less clear. Additionally, several studies have described differences in RP and XSP with advancing age [16–18], however these studies vary widely in participant age range and disease status and did not test age by sex differences. Therefore, the aim of our study was to 1) characterize age and sex-related differences in RP and XSP with advancing age in healthy adults free of overt CVD and 2) determine the degree to which age-related differences in RP and XSP are associated with subclinical vascular markers of CVD risk in men and women. We tested the hypothesis that higher central RP and XSP would be positively associated with advancing age and associated with carotid artery remodeling and aortic stiffness in men and women.

Methods

Participants

Data compiled from our laboratory database was used to perform a retrospective analysis in a cohort of 399 adults between the ages 18 and 80 years who had non-invasive carotid pressure waveforms collected with applanation tonometry between 2011 and 2021 at the University of Iowa. Participants were recruited from the Iowa City community using flyers and email advertisements to undergo baseline vascular testing for several studies. One participant was excluded because their carotid-femoral pulse wave velocity (cfPWV) was 6.5 standard deviations above the mean, leaving a total of 398 adults for the final analysis. Common exclusion criteria across the studies in which data were compiled included current smoker, current use or use of hormone therapy within the past 6 months, self-reported history of stroke, dementia, diabetes mellitus, pulmonary/renal/neurological/hepatic disease, previous CVD events (myocardial infarction, stent, bypass surgery, heart failure), or current CVD. Participants over 50 years of age were screened for subclinical CVD using a 12-lead electrocardiogram, symptom-limited exercise stress test. Body mass index (BMI) was computed after measurement of weight and height in kg/m². Normal weight was defined as a BMI<25 kg/m², overweight was defined as a BMI ≥25<30 kg/m²; and obese was defined as a BMI ≥30 kg/m². Postmenopausal status was documented in female participants as no menses for at least 12 months, and vascular testing in premenopausal women was limited to the early follicular phase of the menstrual cycle in most studies but not all. All vascular measurements were performed after 10 min of quiet, supine rest. On the day of vascular testing, participants were instructed to arrive following an overnight fast with minimum of 8-hours and abstain from caffeine intake the morning of the study. Additionally, participants were instructed to abstain from exercise and alcohol consumption at least 24-h prior to testing. Furthermore, participants taking vasoactive medications (e.g., anti-hypertensives) were included and asked to withhold their medication the morning of vascular testing. All procedures were approved by the University of Iowa Institutional Review Board and conducted in accordance with the Declaration of Helsinki. All participants provided written informed consent before participating in the research study.

Subclinical Markers of CVD

Carotid intimal medial thickness (IMT)

Carotid artery IMT was obtained in 308 participants. Carotid artery IMT and diameter were measured using B-mode ultrasound imaging with a 12-MHz linear array transducer (LOGIQ 7; GE Healthcare). Ultrasound images were recorded for 30 seconds at 15 frames per second and analysis was performed ~2 cm from the carotid bulb on a wall segment 5–10 mm in length using off-line edge detection software (Vascular Analysis Tools Analyzer 5.5; Medical Imaging Applications, Coralville, IA). Carotid artery IMT was calculated as the end-diastolic distance between the lumen/intima edge to the media/adventitia edge of the far wall [19]. Carotid artery IMT values obtained from 30 second video recordings were averaged over 20–30 cardiac cycles to a single value for each participant.

BP waveform acquisition

Brachial BP was obtained in triplicate using an inflatable cuff with a built-in microphone (Non-invasive Hemodynamics workstation; Cardiovascular Engineering, Inc., Norwood, MA). BP was determined via auscultation, where the 1^st and 5^th Korotkoff sounds denoted systolic and diastolic BP respectively during cuff deflation rate of 2 mmHg per second. Briefly, a tonometer was used to collect pressure waveforms from the brachial, radial, carotid and femoral arteries. The peak and trough of the ensembled brachial pressure waveform was calibrated to the systolic and diastolic brachial cuff BP. The ensembled carotid pressure waveform was then calibrated to the integrated mean and diastolic BP of the brachial pressure waveform.

Aortic stiffness

Aortic stiffness was obtained in 393 participants and assessed by the reference-standard cfPWV (NIHem workstation; Cardiovascular Engineering, Inc., Norwood, MA) as previously described [20,21]. Carotid and femoral artery waveforms were gated to the R-wave of the ECG to determine the foot-to-foot time delay. Carotid-femoral distance was quantified as the distance between the suprasternal notch to the femoral waveform pulse site minus the distance from the suprasternal notch to the carotid waveform pulse site. cfPWV was calculated as the carotid-femoral distance divided by the carotid-femoral foot-to-foot time delay.

Augmentation Index

Augmentation index (AIx) was derived from the ensembled carotid pressure waveform and calculated as AIx (%) = [(P2-P1)/PP] x 100, where P2 is the late systolic pressure inflection, P1 is the early systolic pressure inflection and PP is the central pulse pressure.

Reservoir-excess pressure parameters

A custom-written script in R (R Foundation for Statistical Computing, Vienna, Austria) was used to estimate RP from the ensembled carotid pressure waveform as follows:

$$\frac{dRP}{dt}=ks\left(P-\mathrm{R}\mathrm{P}\right)-kd\left(RP-P_{\mathrm{\infty }}\right)$$ (1)

where P is the total measured pressure, ks is the systolic rate constant, kd is the diastolic rate constant and P∞ is the arterial asymptotic pressure (Pinfinity). XSP was calculated as P – RP. In traditional applications of the reservoir-excess pressure model, Pinfinity (the pressure asymptote of the diastolic decay) is taken as a free parameter obtained by fitting a mono exponential function to the pressure waveform during diastole with the form: Rp-p∞ = (p0-p∞)∙e^-b∙t where p∞ is Pinfinity (pressure asymptote), p0 is the optimized pressure at end systole and b is the diastolic time constant. In the present study, pressure during diastole was fit using a mono exponential function with a constraint on Pinfinity to fall between ≥10 mmHg and ≤95% diastolic BP. Values for RP and XSP were presented as the peak less diastolic pressure (i.e. amplitudes) and integrals calculated as the area under the curve of RP and XSP.

Statistical analysis

Data are presented as mean ± standard deviation. Participant characteristics were categorized into age groups roughly by decade to demonstrate differences in participant characteristics with advancing age. Welch’s t-tests were used to test mean differences in participant characteristics, RP and XSP between men and women by age group. Linear regression was used to test associations of RP and XSP with age as a continuous variable and sex, carotid artery IMT and cfPWV. cfPWV was included as a confounding variable in age by sex models to assess the impact of aortic stiffness on RP and XSP with aging. Conventional CVD risk factors that were considered physiologically important or correlated with exposures and outcomes (P<0.05) were included as confounders in adjusted models assessing the relation between RP and XSP and vascular markers of subclinical CVD. Age, body mass index (BMI), resting heart rate (HR), mean arterial pressure (MAP), cfPWV and anti-hypertensive medication were included as confounding variables in models using carotid artery IMT as the outcome of interest. Age, BMI, HR, MAP and anti-hypertensive medication were included as confounding variables in models using cfPWV as the outcome of interest. Linear regression models were assessed for normality qualitatively using histograms and Q-Q plots of model residuals as well as quantitively using the Kolmogorov-Smirnov test. XSP peak and XSP integral were log transformed because of non-normally distributed model residuals. Sensitivity analyses were performed using Bland-Altman analysis to test the influence of constraining Pinfinity in our study on RP and XSP (Supplemental table 1). Lastly, we calculated vascular age as the predicted age in a multivariable regression model that included: sex, height, heart rate, cfPWV, SBP, DBP, total cholesterol and fasting blood glucose. Age difference (Δ-age) was calculated as chronological age minus predicted vascular age and early vascular aging (EVA) was defined as Δ-age < −5.7 years as previously described [22]. Logistic regression was used to assess the relation between RP peak and the probability of EVA. Data were analyzed using R 4.1.0 (R Foundation for Statistical Computing, Vienna, Austria).

Results

Participant characteristics

Participant characteristics are presented in Table 1. Participants were between 18 and 80 years of age (median age, 45 years) and included 159 men (40%) and 239 women (60%). BMI ranged from 18–50 kg/m², but on average was <30 kg/m² in all age groups and obesity categories were evenly represented in the study: 126 normal weight (32%), 135 overweight (34%) and 137 obese (34%). On average participants were not taking anti-hypertensive medications (89%), the remaining were receiving anti-hypertensive medication (11%). There were 140 premenopausal women (59%), the remaining 99 women were postmenopausal (41%). Postmenopausal women were between 50 and 77 years of age and were aged 62 ± 6 years on average.

Table 1.

Participant characteristics by age group (n=398)

	18–29 years		30–39 years		40–49 years		50–59 years		60–69 years		70+ years
Variable	Men n=48	Women n=55	Men n=17	Women n=56	Men n=16	Women n=17	Men n=29	Women n=44	Men n=35	Women n=50	Men n=14	Women n=17
Age, years	24.0 (3.5)	24.6 (3.7)	33.7 (3.7)	34.0 (2.6)	43.4 (2.7)	43.4 (3.1)	55.7 (2.9)	55.3 (2.4)	63.7 (2.6)	64.0 (2.6)	73.1 (2.8)	71.9 (2.0)
BMI, kg/m2	28.9 (4.7)	26.5 (6.9)*	26.6 (4.5)	28.2 (8.3)	30.2 (4.2)	30.5 (8.3)	28.0 (5.3)	29.2 (5.8)	28.2 (5.0)	27.0 (4.3)	29.4 (4.3)	29.6 (6.2)
Systolic BP, mmHg	123 (13)	114 (11)*	117 (12)	113 (12)	125 (10)	117 (13)	121 (15)	122 (14)	127 (15)	125 (16)	125 (14)	134 (14)
Diastolic BP, mmHg	63 (10)	65 (8)	67 (10)	64 (10)	73 (10)	73 (11)	71 (8)	72 (7)	73 (8)	69 (8)	71 (9)	71 (9)
MAP, mmHg	84 (12)	83 (10)	86 (11)	84 (12)	94 (10)	91 (13)	92 (10)	93 (10)	94 (10)	93 (11)	93 (9)	97 (12)
Brachial PP, mmHg	60 (11)	49 (8)*	49 (7)	48 (8)	53 (7)	45 (6)*	49 (10)	50 (11)	54 (12)	55 (13)	54 (12)	64 (10)*
Central PP, mmHg	50 (14)	41 (11)*	43 (9)	44 (12)	47 (10)	45 (12)	47 (13)	50 (14)	52 (16)	55 (15)	54 (12)	61 (20)
cfPWV, m/s#	5.8 (0.7)	5.6 (0.8)	6.4 (0.7)	5.9 (1.0)*	7.4 (1.4)	6.7 (1.0)	7.6 (1.4)	7.7 (1.6)	8.9 (2.0)	8.4 (1.8)	8.9 (1.8)	10.4 (2.5)
Carotid IMT, mm##	0.40 (0.1)	0.41 (0.1)	0.43 (0.2)	0.49 (0.1)	0.50 (0.1)	0.49 (0.1)	0.55 (0.1)	0.54 (0.1)	0.57 (0.1)	0.56 (0.1)	0.60 (0.1)	0.62 (0.1)
Anti-HTN, n (%)	1 (2)	2 (4)	0 (0)	4 (7)	1 (7)	4 (24)	2 (7)	2 (5)	10 (29)	10 (20)	2 (14)	7 (41)
Statins, n (%)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)	1 (3)	2 (5)	10 (29)	4 (8)	4 (29)	5 (29)
Contraceptives, n (%)	- -	27 (49)	- -	16 (29)	- -	3 (18)	- -	1 (2)	- -	0 (0)	- -	0 (0)
Post-menopausal, n (%)	- -	0 (0)	- -	0 (0)	- -	0 (0)	- -	37 (84)	- -	45 (90)	- -	17 (100)

Data presented as mean (standard deviation) for continuous variables and n (%) for categorical variables.

^#cfPWV was not collected in 5 participants (n=393).

^##Carotid artery IMT was not collected in 90 participants (n=308). BP, blood pressure; MAP, mean arterial pressure; PP, pulse pressure; cfPWV, carotid-femoral pulse wave velocity; IMT, intima-media thickness; HTN, hypertension. Welch’s t-tests were used to test mean differences between men and women in each age group.

^*P<0.05 compared to men within the same age group.

Reservoir-excess pressure parameters and age by sex

Mean values of RP and XSP for each age group by sex are presented in Supplemental Table 2. RP peak, RP integral and XSP integral were positively associated with age in both men and women (Figure 2). XSP peak was negatively associated with age in men (Figure 2). The association between RP peak, RP integral and XSP integral and age was stronger in women (Figure 2). cfPWV abolished the impact of age on RP peak, RP integral and XSP integral in men (Supplemental table 3). cfPWV attenuated the impact of age on RP peak, RP integral and XSP integral in women, however age remained significant in our models (Supplemental table 3). Higher central pulse pressure was positively associated with age in women, but not men (Supplemental figure 1). Age-corrected values of RP and XSP stratified by sex are presented in Supplemental Table 4. Logistic regression was used to assess the relation between RP peak and the odds of EVA. We observed for every 1 mmHg increase in RP peak, there was a 3% increase in the odds of EVA (P<0.05) (Supplemental Table 5).

FIGURE 2.

Differences in RP and log transformed XSP parameters with age. A) RP peak, B) RP integral, C) XSP peak and D) XSP integral mean values and standard error. Linear regression was used to assess the relation of RP and XSP parameters with age and sex. RP; reservoir pressure; int, integral; XSP, excess pressure; B, unstandardized beta coefficient. *P<0.05 (Men, n=159; Women, n=239)

Association of blood pressure waveform parameters and subclinical markers of CVD

RP peak and RP integral were positively associated with carotid artery IMT in women, but not men, after adjusting for age, BMI, HR, MAP, cfPWV and anti-hypertensive medication (Table 2). RP peak was associated with cfPWV in women but not men after adjusting for age, BMI, HR, MAP and anti-hypertensive medication (Table 3). Finally, cfPWV was associated with carotid artery IMT in men (P<0.05), but not women (P>0.05) in models adjusted for age, BMI, HR, MAP and anti-hypertensive medication. A sensitivity analysis was performed excluding individuals taking anti-hypertensive medication. Altogether, the principal findings were unchanged, however we did observe an attenuation in the association between RP peak and cfPWV in women (Supplemental Table 6). The ratio of XSP to RP was not associated with carotid artery IMT or cfPWV in women or men after adjusting for traditional CVD risk factors (P>0.05).

Table 2.

Association of reservoir pressure (RP) and excess pressure (XSP) parameters with carotid artery intima media thickness (CIMT) in men and women.

Carotid IMT	Men (n=117)			Women (n=188)
Variable	β (95% CI)	P	R2	β (95% CI)	P	R2
RP peak	0.020 (0.018, 0.023)	0.84	0.36	0.231 (0.229, 0.233)*	<0.01	0.332
RP integral	0.154 (0.148, 0.160)	0.14	0.373	0.254 (0.250, 0.258)*	<0.01	0.337
XSP peak	0.098 (0.025, 0.171)	0.27	0.367	0.093 (0.048, 0.138)	0.19	0.311
XSP integral	0.071 (0.057, 0.085)	0.42	0.364	0.113 (0.071, 0.154)	0.16	0.312

Linear regression was used to assess relations of RP and log transformed XSP parameters with carotid artery IMT in men and women. Models are adjusted for age, BMI, HR, MAP, cfPWV and anti-hypertensive medication. RP, reservoir pressure; XSP, excess pressure; β, standardized beta coefficient; IMT, intima-media thickness; BMI, body mass index; HR, heart rate; MAP, mean arterial pressure.

^*P <0.05 indicates significant association between variable and carotid IMT.

Table 3.

Associations of reservoir pressure (RP) and excess pressure (XSP) parameters with carotid femoral pulse wave velocity (cfPWV) in men and women.

cfPWV	Men (n=156)			Women (n=237)
Variable	β (95% CI)	P	R2	β (95% CI)	P	R2
RP peak	0.052 (0.026, 0.078)	0.45	0.592	0.120 (0.099, 0.140)*	0.02	0.63
RP integral	−0.019 (−0.074, 0.037)	0.81	0.591	0.094 (0.046, 0.142)	0.09	0.626
XSP peak	−0.009 (−0.667, 0.650)	0.88	0.591	−0.001 (−0.591, 0.589)	0.99	0.621
XSP integral	0.019 (−0.598, 0.636)	0.74	0.591	0.005 (−0.085, 0.095)	0.92	0.622

Linear regression was used to assess relations of RP and log transformed XSP parameters with cfPWV in men and women. Models are adjusted for age, BMI, HR, MAP and anti-hypertensive medication. RP, reservoir pressure; XSP, excess pressure; cfPWV, carotid-femoral pulse wave velocity; β, standardized beta coefficient; BMI, body mass index; HR, heart rate; MAP, mean arterial pressure.

^*P <0.05 indicates significant association between variable and cfPWV.

Central pulse pressure was not associated with carotid artery IMT in women (β=0.160, P=0.06) or men (β=0.065, P=0.48) after adjusting for age, BMI, HR, MAP, cfPWV and anti-hypertensive medication. Furthermore, central pulse pressure was not associated with cfPWV in women (β=0.089, P=0.10) or men (β=0.027, P=0.66) after adjusting for age, BMI, HR, MAP and anti-hypertensive medication. Carotid AIx was associated with carotid artery IMT in women (β=0.179, P=0.03), but not men (β=0.033, P=0.34) after adjusting for age, BMI, HR, MAP, cfPWV and anti-hypertensive medication. However, our observed association of RP peak and RP integral with carotid artery IMT was stronger than the association between carotid AIx and carotid artery IMT. Carotid AIx was not associated with cfPWV in women or men (P>0.05).

Discussion

The aim of our study was to characterize age and sex-related differences in RP and XSP with advancing age in healthy adults and determine the degree to which age-related differences in RP and XSP are associated with subclinical vascular markers of CVD risk, carotid artery IMT and aortic stiffness, in men and women. The primary finding was that RP peak, RP integral and XSP integral were higher with advancing age in both men and women, but XSP peak was lower with aging in men. Among women, higher RP peak and RP integral were associated with greater carotid artery IMT, and RP peak was associated with faster cfPWV. Taken together, these data demonstrate that age-related increases in central RP may provide hemodynamic insight into the pathophysiology of subclinical vascular remodeling and large artery stiffening with advancing age, particularly in women.

Our finding that RP peak is associated with advancing age is in agreement with previous findings [16–18]. For example, in patients ages 55–72 years (n=674) undergoing coronary angiography for suspected coronary artery disease, RP peak was positively related with age [16]. Additionally, another study examined patients ages 35–73 years (n=18) undergoing coronary angiography with low probability of coronary artery disease and demonstrated a positive association between RP peak and age [17]. Furthermore, among apparently healthy adults ages 35–55 years (n=1939), RP peak was higher with age, with a more significant elevation women [18]. Our observed age-by-sex interaction of RP and XSP occurs around 50 years of age and the age range of postmenopausal women was 50–77 years, thus menopause may contribute to the accelerated rise in RP and XSP in women with advancing age. Lastly, one study revealed a positive association between RP integral, XSP peak and XSP integral and age in high-risk older adults. Collectively, our results extend the results of these previous studies in that our data encompasses a wider age range across the lifespan and describes changes in RP and XSP among both men and women. Longitudinal studies are needed to confirm these findings regarding the effect of aging on RP integral, XSP peak and XSP integral.

Age-related increases in aortic stiffness may partially explain the elevation in RP with age observed in our study and others [16–18]. Following ventricular contraction, aortic inflow exceeds outflow and this volume differential is buffered by the radial expansion, and subsequent recoil, of the aortic wall [24]. In this regard, an increase in aortic stiffness would result in a loss of the buffering function of the aorta and an increase in RP. Indeed our data and others [17,25] reveal a positive association between RP peak and aortic stiffness, however the association in our study was independent of MAP and was only observed in women. The effect of anti-hypertensive medication on the association between RP peak and cfPWV in women warrants further investigation. In contrast to our hypothesis, RP peak was not associated with cfPWV in men in fully adjusted models, which is interesting considering that cfPWV increases similarly with aging in men and women [26]. However, RP peak is strongly associated with central and peripheral pulse pressure [27], and our data reveal central pulse pressure was lower in younger women than men and had a stronger association with age compared to men. This is in agreement with a longitudinal study that demonstrated peripheral pulse pressure is higher in young men compared with young women and plateaus after 40 years of age, whereas in women peripheral pulse pressure was lower in youth and increased linearly across all ages [28]. Thus, it is plausible that sex differences in the trajectory of central pulse pressure with aging may partially explain why RP peak was associated with cfPWV in women but not men in our cohort. Our observed sex difference in XSP peak with aging may be because of disparate aortic root remodeling between men and women. Aortic characteristic impedance (Zc) is inversely related to aortic root diameter and directly to aortic wall stiffness, however aortic Zc is 5 times more sensitive to diameter than stiffness [29]. One study revealed women have a lower aortic Zc in youth compared to men and an accelerated increase in aortic Zc with aging [30]. Additionally, a recent study found aortic Zc and aortic forward pressure amplitude to be the primary determinants of XSP [15], suggesting age-related sex differences in XSP peak may be a result of other factors such as aortic root diameter. Future studies should examine sex differences in the association of aortic Zc and XSP peak with aging.

Carotid artery IMT reflects subclinical arterial wall remodeling, mainly smooth muscle cell hyperplasia and fibrocellular hypertrophy in the medial layer [31], and is associated with aortic stiffness in healthy older adults [32]. Indeed, our data suggest higher RP peak and integral are associated with carotid artery IMT in women independent of several traditional CVD risk factors, including aortic stiffness. This is plausible because the carotid artery is likely exposed to excessive pulsatile energy as RP peak increases and aortic buffering capacity declines, thus causing compensatory carotid medial hypertrophy to overcome elevated wall stress. However, RP peak and integral were the second strongest correlate of carotid artery IMT in our models, with age being the strongest. Our observed association of RP peak and integral with carotid artery IMT is in contrast to previous observations, which found XSP to be associated with carotid artery IMT, not RP [9,25,33]. XSP increases from central to peripheral arteries and is highly related to systolic BP [27], thus the observed differences might be because these studies obtained pressure waveforms from peripheral sites using radial tonometry without the application of a transfer function [9] and BP from brachial cuff oscillometry [25,33] where pressures are higher because of PP amplification [34]. Therefore, pressure waveforms collected directly from the carotid artery in our study may explain the contrasting results of XSP and carotid artery IMT compared with other studies.

RP remains relatively constant throughout the arterial system [27,35], which would not explain the differences between our findings and the previous findings that did not see an association of RP with carotid artery IMT [9,25,33]. The use of the Sphygmocor^® XCEL (CardieX, Inc) brachial cuff to derive RP from brachial sub-diastolic volumetric displacement waveforms may partially explain differences between the studies by Peng et al. [25,33] and our study in the association of RP and XSP with carotid artery IMT. At higher pressures, the XCEL brachial cuff is predisposed to underestimate RP peak and integral and overestimate XSP peak and integral when compared to catheter measured RP and XSP respectively [36]. Lastly, Davies et al. [9] studied individuals with hypertension, who have elevated carotid artery IMT compared to age-matched normotensive controls [37]. Therefore, it is plausible that accelerated carotid artery IMT could have influenced the association with RP in their hypertensive cohort, compared to our cohort who the majority were normotensive (SBP <130 mmHg). Taken together, RP peak and integral are associated with carotid artery remodeling and elevated subclinical CVD risk with aging in healthy women but not men, however the mechanisms explaining sex differences in these relations are unclear and require further investigation.

A strength of the present study was the wide age range of participants, allowing us to characterize RP and XSP across the lifespan in both men and women. However, this was a cross-sectional analysis, therefore we do not have longitudinal data to track hemodynamic progression with aging. Additionally, we did not control for menstrual cycle phase in all studies when studying premenopausal women or account for perimenopause in middle-aged women in our study. Nevertheless, carotid artery IMT is unchanged over the menstrual cycle [38] and the need to control menstrual cycle during vascular testing is still widely debated but may be considered in future studies [39]. Lastly, our study constrained Pinfinity to be between 10 mmHg and 95% diastolic BP. Pinfinity is thought to approximate the microcirculatory zero flow pressure which has been shown to have a non-zero value. Indeed, Hughes and Parker have recently collated data showing that experimentally measured values of zero flow pressure in humans averages around 27 mmHg [40]. However, in individuals who exhibit diastolic decay with little concavity, estimates of Pinfinity may be underestimated and can even be negative. Importantly, the value of Pinfinity has a substantial influence on estimates of the diastole rate constant and consideration of our approach to constrain Pinfinity should be given when interpreting our results (especially as they relate to the diastolic rate constant). Nevertheless, it would seem that differences in Pinfinity have negligible influence on the RP and XSP peaks and integrals [7,41]

In conclusion, we found that central RP peak, RP integral and XSP integral were higher with aging in heathy men and women without CVD. Additionally, in women, RP peak and integral were associated with greater carotid artery IMT and RP peak was associated with higher aortic stiffness independent of conventional CVD risk factors. In this regard, RP peak derived from carotid tonometry may provide insight into central hemodynamic related vascular remodeling and subclinical CVD risk with aging in women. Future studies should explore whether sex differences in RP and XSP further explain disparities in vascular-mediated subclinical target organ damage such as vascular cognitive impairment.

FIGURE 1.

Example of a carotid pressure waveform (solid line) and reservoir-excess pressure parameters. The reservoir pressure and excess pressure can be expressed as peaks or integrals. RP, reservoir pressure; int, integral; XSP, excess pressure.

Abbreviations:

CVD

cardiovascular disease

BP

blood pressure

RP

reservoir pressure

XSP

excess pressure

IMT

intima-media thickness

cfPWV

carotid-femoral pulse wave velocity