Effects of age and sex on middle cerebral artery blood velocity and flow pulsatility index across the adult lifespan

Abstract

Reduced middle cerebral artery blood velocity (MCAv) and flow pulsatility are contributors to age-related cerebrovascular disease pathogenesis. It is unknown whether the rate of changes in MCAv and flow pulsatility support the hypothesis of sex-specific trajectories with aging. Therefore, we sought to characterize the rate of changes in MCAv and flow pulsatility across the adult lifespan in females and males as well as within specified age ranges. Participant characteristics, mean arterial pressure, end-tidal carbon dioxide, unilateral MCAv, and flow pulsatility index (PI) were determined from study records compiled from three institutional sites. A total of 524 participants [18–90 yr; females 57 (17) yr, n = 319; males 50 (21) yr, n = 205] were included in the analysis. MCAv was significantly higher in females within the second (P < 0.001), fifth (P = 0.01), and sixth (P < 0.01) decades of life. Flow PI was significantly lower in females within the second decade of life (P < 0.01). Rate of MCAv decline was significantly greater in females than males (−0.39 vs. −0.26 cm s⁻¹·yr, P = 0.04). Rate of flow PI rise was significantly greater in females than males (0.006 vs. 0.003 flow PI, P = 0.01). Rate of MCAv change was significantly greater in females than males in the sixth decade of life (−1.44 vs. 0.13 cm s⁻¹·yr, P = 0.04). These findings indicate that sex significantly contributes to age-related differences in both MCAv and flow PI. Therefore, further investigation into cerebrovascular function within and between sexes is warranted to improve our understanding of the reported sex differences in cerebrovascular disease prevalence.

NEW & NOTEWORTHY We present the largest dataset (n = 524) pooled from three institutions to study how age and sex affect middle cerebral artery blood velocity (MCAv) and flow pulsatility index (PI) across the adult lifespan. We report the rate of MCAv decline and flow PI rise is significantly greater in females compared with in males. These data suggest that sex-specific trajectories with aging and therapeutic interventions to promote healthy brain aging should consider these findings.

INTRODUCTION

Changes in cerebral blood velocity over the course of the lifespan have important implications for brain health, as velocity decline has been tied to cognitive impairment and dementia pathogenesis (1–5). The literature is well established that velocity decreases with increasing age (6–8). However, the few large studies that have attempted to quantify the rate of decline by measuring middle cerebral artery mean blood velocity (MCAv, an estimate of cerebral blood flow) have made conflicting estimates (9–12). Ainslie et al. (9) reported that age-related MCAv decline was −0.8 cm·s⁻¹ per yr in men 18–79 yr of age, but Bailey et al. (10) reported a −0.3 cm·s⁻¹ per yr decline in sedentary men and −0.4 cm·s⁻¹ per year in trained men. In addition, Bakker et al. (11) reported −0.6 cm·s⁻¹ per yr decline in females and −0.4 cm·s⁻¹ per yr decline in males 55 yr of age and over. Lastly, Tegeler et al. (12) reported MCAv decreased by ∼4%–5% each decade from <30 to 80 yr of age in healthy females and males. Importantly, disparities may be due to differences in rate of MCAv change by sex and life stage, which complicate attempts to determine one single MCAv decline rate for the general population over an entire lifetime.

Sex differences in resting MCAv have been reported (8, 12, 13). However, conclusions have been pieced together with investigations only including smaller age ranges that did not include assessment across the entirety of the adult lifespan (11, 13), involving small sample sizes (8, 13), or which did not access the interaction effect of age and sex (8, 12). Young females have significantly higher resting MCAv than age-matched males. However, this disparity may disappear in later adulthood, with females and males having similar resting MCAv by the sixth to eighth decade of life (8, 11, 12). This disappearance of the MCAv disparity suggests that there may be sex-specific trajectories of MCAv with increasing age. However, there has been minimal investigation into the rate of MCAv decline by sex and previous studies have failed to account for this potential heterogeneity in the rate of MCAv decline.

Along with MCAv, increasing evidence indicates an important role of the large extracranial artery function in the development of cerebrovascular and cognitive diseases (14, 15). Specifically, the manner in which flow pulsatility is translated into the brain has emerged as a critical contributor to brain health that is dependent on large extracranial artery stiffness (16). The proximal aorta and carotid arteries dampen pressure pulsatility to ensure a steady cerebral blood flow in the brain microvasculature (17). However, the ability of the cerebrovasculature to dampen pressure has been reported to become less effective with advancing age (18–20). A commonly used hemodynamic index to assess cerebral flow pulsatility, measured with transcranial Doppler ultrasound, is the Gosling flow pulsatility index (PI) (21). Data suggest an age effect on flow PI with younger adults showing a significantly lower flow PI (0.73, n = 114) compared with older adults (0.84, n = 148), in generally healthy participants (18). However, in this cross-sectional study, an interaction effect between age and sex was not detected, possibly due to dichotomization of sex into premenopausal/young and postmenopausal/older adults.

Effects of sex on MCAv and flow PI could have implications for differences in brain health and disease pathogenesis. Although both sexes are impacted by cerebrovascular and cognitive diseases, females may have a higher prevalence of cerebrovascular disease (22) and greater risk and severity of cognitive decline than males (23). Therefore, it is necessary to systematically investigate whether MCAv and flow PI differ between sexes across the adult lifespan. It is unlikely that MCAv and flow PI would change at one single, steady rate from early to late adulthood, considering markers of other brain pathology follow heterogeneous patterns with age (24, 25). To address these gaps in knowledge, we aimed to characterize the rate of MCAv decline and flow PI rise for females and males across the adult lifespan using pooled cross-sectional data from three institutional sites. In addition, we aimed to determine if the rate of change in MCAv and flow PI differed within specified decades of life between sexes. We hypothesized 1) a significant age and sex interaction effect for MCAv and flow PI, with females having a significantly greater rate of MCAv decline and rate of flow PI rise compared with males across age and 2) disparities in the rate of MCAv decline and flow PI rise would be greater between sexes in the older age ranges.

METHODS

Study Population

Participants had previously been involved in research studies at one of three institutions (University of Kansas Medical Center, KS; University of Wisconsin-Madison, WI; Université Laval, Canada) from 2013 to 2019. This study included secondary analysis from previous institutional review board (IRB) approved studies and the use of deidentified data sharing, thus, the University of Kansas Medical Center IRB deemed this study to be nonhuman subjects research and did not require IRB review. Comprehensive inclusion and exclusion criteria from the previous independent institutional studies have been published elsewhere (26–37). Briefly, participants were enrolled if 1) between the ages of 18 to 90 years, 2) had varying levels of physical (in)activity and exercise engagement, 3) cognitively normal/nondemented, 4) without a history of major cardiovascular, pulmonary, metabolic, or neurological diseases, and 5) without major orthopedic disability. Premenopausal females were assessed during days 1–10 of their menstrual cycle or during the nonactive phase if taking oral contraceptives. Females were not pregnant, which was confirmed verbally or with a urine pregnancy test.

Laboratory Procedures

Data were retrieved from study records, which included demographics (age, sex, height, weight, and cardiovascular risk factors), mean arterial pressure (MAP), end-tidal carbon dioxide ($\text{PE}{\text{T}}_{\text{C}{\text{O}}_2}$), and unilateral MCAv. Body mass index (BMI) was calculated from height and weight measurements. Cardiovascular disease (CVD) risk classification was determined according to the American College of Sports Medicine (ACSM) criteria (38). Participants abstained from consuming caffeine for a minimum of 6 h and exercising for 12 h before experimental procedures.

Participants were in a seated or supine position and completed a rest period of at least 10 min in a quiet, temperature-controlled room (22°C–24°C) before data collection. $\text{PE}{\text{T}}_{\text{C}{\text{O}}_2}$ was measured via a nasal cannula or mouthpiece attached to a breath-by-breath gas analyzer (BCI Capnocheck 9004, Smiths Medical; Datex Ohmeda, GE Healthcare, Fairfield, CT; Breezesuite, MedGraphics Corp., Saint Paul, MN). MAP was measured using a finger plethysmograph (Finometer Pro/Finometer Nova, Finapres Medical Systems, Amsterdam, The Netherlands; Nexfin, Edwards Lifesciences, Irvine, CA; Nexfin, Edwards Lifesciences, ON, Canada). MCAv was measured using a 2-MHz transcranial Doppler ultrasound probe placed on the left or right trans-temporal window (TCD; RobotoC2MD System, Multigon Industries, Yonkers, NY; Neurovision System, Multigon Industries, Yonkers, NY; Spencer Technologies, Redmond, WA; Doppler Box, Compumedics DWL USA, Inc., San Juan Capistrano, CA). The TCD probe was secured with a headgear device to maintain the optimal position and angle.

All recordings lasted for at least 15 s with participants in a resting state. Data were sampled at 250–1,000 Hz (National Instruments, Austin, TX; WinDaq DATAQ Instruments, Akron, OH; AD Instruments, Colorado Springs, CO). Mean MCAv was calculated by averaging the cerebral blood velocity within each cardiac cycle. MCAv, MAP, and $\text{PE}{\text{T}}_{\text{C}{\text{O}}_2}$ were averaged over the duration of the recording for each participant. $\text{PE}{\text{T}}_{\text{C}{\text{O}}_2}$ data were not collected in 28 participants from one site. Flow PI was calculated as the difference between systolic and diastolic MCAv divided by the mean MCAv (MCAv_systolic − MCAv_diastolic)/MCAv_mean (21).

To examine whether the rate of MCAv and flow PI change varied within specified decades of life, participants were divided into five subgroups: 18–30 yr, 31–50 yr, 51–60 yr, 61–70 yr, and 71–90 yr.

Statistical Analysis

All statistical analyses were performed using MATLAB (R2019b v. 9.6.0, MathWorks, Natick, MA). Between-group (female versus male) differences of quantitative normally distributed BMI, MAP, $\text{PE}{\text{T}}_{\text{C}{\text{O}}_2}$, MCAv, and flow PI were assessed using an analysis of variance (ANOVA). Between-group (female versus male) differences in medications, postural position, and CVD risk classifications were evaluated using χ² test.

The association between age and 1) MCAv and 2) flow PI, as a function of sex was examined using an analysis of covariance (ANCOVA), with “sex” being the independent categorical variable, “MCAv and flow PI” being the continuous dependent variables, and “age” being a continuous covariate. With age as the variable of interest, an ANCOVA as an extension of multiple regression, was used to test whether the rate of MCAv decline and rate of flow PI rise (e.g., slope) was significantly different between females and males. The “age × sex” interaction expressed the difference in rate of change and the ANCOVA coefficient showed the significance of this term. In addition, within the five subgroups, an ANCOVA was used to compare the rate of change for MCAv and flow PI between females and males within the specified age ranges.

Statistical significance was established a priori at P < 0.05. Data are expressed as means (standard deviation) unless otherwise stated.

RESULTS

We included 524 unique, deidentified data sets (no repeated data) in the analysis (age: females 57 ± 17 yr, males 50 ± 21 yr; 61% female). Sixty-nine females were premenopausal and 250 postmenopausal. Twenty-four premenopausal females were taking oral contraceptives and 29 postmenopausal females were taking oral estrogen, eight postmenopausal females were using topical estrogen, and eight were on hormone replacement therapy. One postmenopausal female and three males were taking testosterone. Sixty-four females and 47 males were taking antihyperlipidemic medications (P = 0.43) and 64 females and 25 males were taking antihypertensive medications (P = 0.02).

Two hundred and fifty-seven participants (144 females, 113 males) had recordings in the seated position and 267 participants (175 females, 92 males) in the supine position (P = 0.03).

Average $\text{PE}{\text{T}}_{\text{C}{\text{O}}_2}$ was significantly higher in females compared with in males (Table 1). In addition, there were significantly fewer participants in the high CVD risk classification than the low CVD risk classification (P = 0.003; Bonferroni corrected).

Table 1.

Participant characteristics (n = 524)

	Total			18–30 yr		31–50 yr		51–60 yr		61–70 yr		71–90 yr
	Females n = 319	Males n = 205	P Value	Females n = 53	Males n = 66	Females n = 16	Males n = 26	Females n = 74	Males n = 20	Females n = 125	Males n = 57	Females n = 51	Males n = 36
Height, cm	164 ± 6	179 ± 6	<0.001	166 ± 6	180 ± 7	165 ± 6	180 ± 7	164 ± 5	181 ± 5	164 ± 6	177 ± 5	162 ± 8	179 ± 7
Body mass, kg	68 ± 13	82 ± 13	<0.001	63 ± 10	77 ± 11	63 ± 10	81 ± 10	70 ± 13	88 ± 12	70 ± 13	85 ± 12	69 ± 14	84 ± 15
BMI, kg m2	25 ± 5	26 ± 3	0.54	23 ± 4	24 ± 3	23 ± 3	25 ± 3	26 ± 5	27 ± 3	26 ± 5	27 ± 3	26 ± 5	26 ± 4
MAP, mmHg	84 ± 15	86 ± 13	0.31	83 ± 12	86 ± 10	80 ± 11	87 ± 15	92 ± 12	92 ± 11	83 ± 16	87 ± 14	79 ± 15	79 ± 12
$\text{PE}{\text{T}}_{\text{C}{\text{O}}_2}$, mmHg	38 ± 6	37 ± 6	0.01	38 ± 4	39 ± 6	37 ± 5	39 ± 4	42 ± 4	38 ± 4	39 ± 6	36 ± 6	32 ± 6	33 ± 5
CVD risk			0.01
Low	129 (40)	110 (54)		48 (91)	65 (98)	14 (88)	21 (81)	34 (46)	8 (40)	28 (22)	12 (21)	5 (10)	4 (11)
Moderate	95 (30)	54 (26)		1 (2)	0 (0)	1 (6)	2 (8)	22 (30)	9 (45)	51 (41)	29 (51)	20 (39)	14 (39)
High	95 (30)	41 (20)		4 (7)	1 (2)	1 (6)	3 (11)	18 (24)	3 (15)	46 (37)	16 (28)	26 (51)	18 (50)

Continuous variables are means ± standard deviation. CVR risk is count (percent). Statistical comparisons were only completed for the total column. One-way analysis of variance (ANOVA) was used for statistical comparisons of continuous variable and the chi-square test was used for statistical comparison of CVD risk. Significance, P < 0.05. Boldface values indicate significant differences. BMI, body mass index; MAP, mean arterial pressure (missing: females n = 1); $\text{PE}{\text{T}}_{\text{C}{\text{O}}_2}$, end-tidal CO₂ (missing: females n = 1, males n = 27). CVD risk, ACSM cardiovascular disease risk classification, significant difference between low and high CVD risk.

There was no interaction between age, sex, and postural position (supine or seated; P = 0.59). Despite $\text{PE}{\text{T}}_{\text{C}{\text{O}}_2}$ being significantly higher in females compared with in males, there was no interaction between age, sex, and $\text{PE}{\text{T}}_{\text{C}{\text{O}}_2}$ (P = 0.19). There was no interaction between age, sex, and MAP (P = 0.49). In addition, a lower proportion of participants were in the high ACSM CVD risk classification compared with the low-risk classification, however, there was no interaction between age, sex, and ACSM CVD risk classification (P = 0.61). These results indicated there was no influence of recorded position, $\text{PE}{\text{T}}_{\text{C}{\text{O}}_2}$, MAP, or ACSM CVD risk classification on sex-specific MCAv trajectories across the adult lifespan.

Middle Cerebral Artery Blood Velocity

Across the adult lifespan, the rate of MCAv decline was significantly greater in females than in males (Table 2, Fig. 1). The regression equations to predict MCAv (cm s⁻¹) within sex, were:

Table 2.

MCAv and MCAv rate of change (n = 524)

	MCAv (cm·s−1)				MCAv Rate of Change (cm·s−1·yr)
Age Group	Combined	Females	Males	P Value	Females	Males	P Value
18–30 yr	64.4 ± 12.7 n = 119	70.5 ± 12.7 n = 53	59.5 ± 10.3 n = 66	<0.001	−0.61 (−2.10, 0.87)	−0.38 (−1.32, 0.56)	0.78
31–50 yr	58.9 ± 13.3 n = 42	60.9 ± 12.5 n = 16	57.6 ± 13.9 n = 26	0.43	0.59 (−1.14, 2.32)	−0.31 (−1.21, 0.59)	0.35
51–60 yr	62.0 ± 14.2 n = 94	64.0 ± 14.7 n = 74	54.6 ± 8.6 n = 20	0.01	−0.33 (−1.80, 1.14)	0.34 (−2.01, 2.70)	0.72
61–70 yr	53.1 ± 12.8 n = 182	55.0 ± 13.0 n = 125	48.8 ± 11.1 n = 57	<0.01	−1.44 (−2.31, −0.57)	0.13 (−1.05, 1.32)	0.04
71–90 yr	47.6 ± 10.9 n = 87	48.8 ± 12.1 n = 51	45.8 ± 8.8 n = 36	0.22	−0.40 (−1.30, 0.49)	−0.69 (−1.82, 0.44)	0.72
Total	56.8 ± 14.1 n = 524	59.0 ± 15.0 n = 319	53.4 ± 11.9 n = 205	<0.001	−0.39 (−0.47, −0.30)	−0.26 (−0.33, −0.19)	0.04

MCAv data presented as means ± standard deviation. MCAv rate of change data presented as slope (95% confidence interval). Significance, P < 0.05. Boldface values indicate significant differences. MCAv P values, ANOVA between means (females vs. males). MCAv rate of change P values, pairwise comparison of slopes (females vs. males). MCAv, middle cerebral artery blood velocity.

Figure 1.

Relationship between age and middle cerebral artery velocity (MCAv) within females and males. The association between age and MCAv as a function of sex was examined using an analysis of covariance (ANCOVA). The hashed line represents regression fit for females (n = 319). The solid line represents regression fit for males (n = 205).

Female MCAv = [80.81 − (0.39 × Age)] + ɛ, Male MCAv = [65.41 − (0.26 × Age)] + ɛ.

MCAv was significantly higher in females than in males for all age groups, except within the 31–50 and 71–90 age groups (Table 2). The rate of MCAv change was significantly greater in females than in males only in the 61–70 age group.

Flow Pulsatility Index

Across the adult lifespan, the rate of flow PI rise was significantly higher in females than in males (Table 3, Fig. 2). The regression equations to predict flow PI within sex, were:

Table 3.

Flow pulsatility index (PI) and Flow PI rate of change (n = 524)

	Flow Pulsatility Index				Flow PI Rate of Change (Flow PI·yr)
Age Group	Combined	Females	Males	P Value	Females	Males	P Value
18–30 yr	0.79 ± 0.20 n = 119	0.73 ± 0.17 n = 53	0.83 ± 0.21 n = 66	<0.01	−0.004 (−0.023, 0.015)	0.001 (−0.019, 0.020)	0.74
31–50 yr	0.77 ± 0.19 n = 42	0.73 ± 0.13 n = 16	0.80 ± 0.22 n = 26	0.27	0.003 (−0.014, 0.021)	0.006 (−0.008, 0.020)	0.85
51–60 yr	0.80 ± 0.12 n = 94	0.79 ± 0.12 n = 74	0.81 ± 0.12 n = 20	0.63	0.023 (0.012, 0.033)	−0.006 (−0.038, 0.026)	0.05
61–70 yr	0.88 ± 0.18 n = 182	0.89 ± 0.18 n = 125	0.87 ± 0.18 n = 57	0.49	0.029 (0.017, 0.040)	0.027 (0.009, 0.045)	0.87
71–90 yr	1.07 ± 0.17 n = 87	1.07 ± 0.16 n = 51	1.08 ± 0.19 n = 36	0.71	−0.001 (−0.013, 0.011)	0.012 (−0.013, 0.037)	0.31
Total	0.87 ± 0.20 n = 524	0.86 ± 0.19 n = 319	0.88 ± 0.21 n = 205	0.30	0.006 (0.005, 0.007)	0.003 (0.002, 0.005)	0.01

Flow pulsatility index data presented as means ± standard deviation. Flow PI rate of change data presented as slope (95% confidence interval). Significance, P < 0.05. Boldface values indicate significant differences. Flow pulsatility index P value, ANOVA between means (females vs. males). Flow PI rate of change P value, pairwise comparison of slopes (females vs. males). Flow PI, flow pulsatility index.

Figure 2.

Relationship between age and flow pulsatility index (PI) within females and males. The association between age and flow PI as a function of sex was examined using an analysis of covariance (ANCOVA). The hashed line represents regression fit for females (n = 319). The solid line represents regression fit for males (n = 205).

Female flow PI = [0.54 + (0.006 × Age)] + ɛ, Male flow PI = [0.72 + (0.003 × Age)] + ɛ.

Flow PI was significantly lower in females than in males within the 18–30 age range (Table 3). The rate of flow PI change was not significantly different in any specified age ranges between females and males.

DISCUSSION

This study characterized the rate of MCAv decline and flow PI rise in females and males using pooled cross-sectional data from three institutional sites. To date, this is the first population-based study assessing cerebral blood velocity and flow pulsatility, measured by transcranial Doppler ultrasound, in both female and male participants across the adult lifespan. In support of our hypotheses, females had a greater rate of decline in MCAv and greater rate of rise in flow PI than males. The disparity in the rate of MCAv decline between females and males was significantly greater in the sixth decade of life. However, contrary to our hypothesis, the rate of flow PI rise was not significantly different within the specified decades of life between sexes.

In agreement with previous research, significant sex differences in MCAv were determined to be more pronounced in younger adults, but no longer manifested around the seventh decade of life. Martin et al. (8) reported significant sex differences no longer occurring in the sixth decade of life and Bakker et al. (11) reported significant sex differences no longer occurring in the eighth decade of life. Disparity in MCAv between sexes at different stages of life has been questioned by others, which have hypothesized that MCAv between sexes may not decline in a similar manner with aging. Evidence from our large sample of participants indicate that the rate of MCAv change in the sixth decade is significantly greater in females, compared with in males, resulting in a similar resting MCAv in both sexes within the seventh decade.

There are multiple hypothesized reasons for the disparity between cerebral blood velocity in females and males, with differences likely driven not only by hormonal differences between sexes but also hormonal changes within each sex that occur over the lifetime. For example, blood pressure starts to rise in the third decade of life in males, but not until the fourth or fifth decade in females, corresponding with the beginning of menopause (39). Although sympathetic nervous system activity and total peripheral resistance are strongly correlated in males, there is no association in females (40, 41). In contrast, a relationship arises between sympathetic nervous system activity and total peripheral resistance in older females (40, 41). Therefore, the loss of estrogen in older females may increase the effect of sympathetic nervous system activity on total peripheral resistance, blood pressure, and loss of endothelial vasodilation (39, 40, 42). Studies have shown that flow mediated dilation (a measure of endothelial function) decreases with advancing age in both sexes, up to 70 yr for males and 80 yr for females. However, age-related decline in endothelial function is greatest after age 45 yr in females, whereas a steady decline has been observed after age 30 yr in males (43). In relation to endothelial dysfunction, arterial stiffness progressively increases with aging and is associated with reduced cerebrovascular function (44–47).

In addition to estrogen, both progestins and androgens also change across the lifespan in both females and males, though their effects on the cerebrovasculature are not as well defined. In general, estrogen decreases cerebrovascular tone and androgens increase cerebrovascular tone (48, 49). Despite testosterone having vasoconstricting effects, paradoxically, higher testosterone levels have been positively associated with increased brain perfusion in older males (50, 51). This suggests that the gradual decline in testosterone levels over the lifespan in males may negatively impact cerebrovascular function (52). Most of this work on the impact of gonadal hormones on the cerebrovasculature has been done in animal models or in vitro. Future studies should investigate the interaction between gonadal hormones and aging on cerebral blood flow in humans. Although previous research may support our findings of an accelerated trajectory in female MCAv decline related to aging, the mechanisms underlying differences between females and males as it relates to changes in cerebrovascular function, or dysfunction, across the adult lifespan are underdeveloped.

In addition to the factors described in the previous paragraphs, it is possible that structural differences between sexes that occur with age contribute to the decline in MCAv. Findings from a previous study in individuals without neurologic deficits have suggested that there is a clear difference between females and males in the process of brain atrophy with increasing age. Within females, a significant reduction in the Brain Volume Index (a measure of atrophy) begins in the fifth decade of life and remains relatively constant into the sixth decade but decline significantly again in the seventh and eighth decades of life (53). However, in males, a significant reduction in the Brain Volume Index does not begin until the sixth decade of life, with a gradual and steady decline through the eighth decade (53). Varying hypotheses have been proposed as to whether cerebral atrophy precedes reduced cerebral blood flow or vice versa. Studies that have reported a relationship between brain volume and cerebral blood flow are mostly based on cross-sectional analyses (54, 55). To date, only one longitudinal study by Zonneveld et al. (56), reported a bidirectional relationship between reduced cerebral blood flow and brain atrophy in a sample of 3,263 nondemented participants over the age of 45 yr. Their results indicate that brain atrophy causes cerebral blood flow to decrease over time and only in individuals over the age of 65 yr was lower cerebral blood flow also related to brain atrophy over time. Therefore, it is plausible that cerebral atrophy contributed to the decline in MCAv determined from our findings. Further investigation into the contribution of brain atrophy on MCAv is necessary, as our study did not measure brain atrophy and the aforementioned longitudinal study (56) did not assess sex differences.

With aging, central (common carotid) pulsatile flow and pressure were only modestly correlated in a large population study, suggesting that changes in (common carotid and MCA) pulsatile flow are influenced by other mechanisms besides increases in central pulsatile pressure (57). The common carotid arteries are relatively stiffer (measured by magnetic resonance imaging) compared with the aortic arch in young adults (58) and stiffen linearly, with a greater rate in females, beginning earlier in the life span (59) compared with the aorta, which remains relatively stable until ∼55 yr of age (60). This may be initially protective for the brain by reflecting pulsatile energy away from the brain (57) until there is a loss of impedance mismatch in mid-life. At this point, downstream remodeling may increase or become more important. However, the increase in the transmission of central pulsatile pressure in older adults is unlikely to be the only determinant of cerebrovascular pulsatile flow and is complicated by changes in vascular anatomy with aging and between sexes. This excessive pulsatile flow can cause damage to brain structure (61) and contribute to the pathogenesis of cerebrovascular disease (62–65). In agreement with our findings, the increase in cerebral pulsatile flow is greater in females compared to males across aging (66, 67), however, the rate in the rise of flow PI within age groups does not appear to be different between sexes.

Cerebrovascular aging cannot be fully understood without consideration of females and investigation into major gaps in the extent of sex differences (9, 68). For example, in a large, population-based cohort (≥70 yr of age) it was determined, using FLAIR magnetic resonance imaging, that sex was independently associated with cerebrovascular pathologies (white matter hyperintensities and cortical infarcts) even after accounting for midlife risk factors (e.g., obesity, physical inactivity, systemic hypertension, etc.) (69). Increasing evidence highlights the importance of sex in vascular health and disease through differences in clinical manifestations, pathophysiology, and response to treatments (70–73). However, one neglected component is the heterogeneity of females included within samples. The phase of the menstrual cycle (in premenopausal or perimenopausal females) and years’ postmenopause in older females is often not accounted for in statistical analysis. Consideration of female hormonal factors in future research will allow for elucidation of other potential mechanisms that may be contributing to cerebrovascular regulation. Despite evidence that estrogen levels may play a significant role in cerebrovascular function, as stated within this discussion section, it is unlikely to be the only reason mechanism for sex differences exist, warranting further investigation.

It is important to note study limitations that should be taken into consideration. First, using transcranial Doppler ultrasound to quantify cerebral blood flow requires the assumption that vessel length, vessel radius, and blood viscosity remain constant (74) in order for velocity to be a surrogate for cerebral blood flow. Evidence suggests changes in MAP and $\text{PE}{\text{T}}_{\text{C}{\text{O}}_2}$ are associated with changes in MCA diameter (75). Transcranial Doppler ultrasound measurements for the present study were collected in a resting state and MCA diameter most likely was not affected by changes in MAP and/or $\text{PE}{\text{T}}_{\text{C}{\text{O}}_2}$, but consideration should be given to age-associated MCA remodeling that can occur (76) as well as greater variability in MCA cross-sectional area at rest in older adults (77), potentially confounding measurements. In addition, only unilateral MCAv was assessed, therefore, results may not reflect the trajectory of age-related change in posterior cerebral circulation and/or global cerebral circulation. Second, the sample size in the 31–50 age group is smaller than the other age groups and, therefore, a noticeable gap emerged. Research studies primarily recruit younger and/or older adults; however, given the evidence provided within the discussion section related to variability in physiological changes that occur over the adult lifespan, there is a growing interest in studying middle aged adults. For example, the presence of midlife CVD risk factors has been shown to accelerate age-related cognitive decline and increase the risk for Alzheimer’s disease later in life (78). Therefore, future work should assess middle-aged adults as an independent cohort to better understand preventative mechanisms as well as detrimental risk factors in age-related cerebrovascular diseases. Furthermore, our results suggest that the rate of MCAv change is not linear across the adult lifespan. However, considering the sample size within the 31–50 yr age group, conclusions should not be drawn and trajectory of MCAv should be further assessed within this age range. Third, there was a significant difference between sex and postural position of the participants during transcranial Doppler ultrasound recordings (seated or supine). Although there was not an interaction effect between age, sex, and postural position, previous research has reported both cerebral blood velocity and cerebral blood flow to be lower in the seated position compared to supine position (79, 80). However, previous research has been limited to younger populations and small sample sizes, therefore, future investigation into the effects of posture on MCAv and flow PI across the lifespan is necessary. Fourth, sex hormones were not directly measured but should be given consideration in future studies as the potential effects on cerebrovascular hemodynamics in older adults has been suggested by previous research, as described in the discussion section. Last, a strength of this study was pooling many participants from three institutions to address these important questions. We acknowledge that variability in methodology and analyses would be inherent between laboratories. However, there is significant benefit to leveraging datasets across institutions to be more representative of the population and address gaps in knowledge for MCAv.

Conclusions

Aging is associated with significant reductions in MCAv and increases in flow pulsatility within both females and males. Our findings provide evidence that the rate of MCAv decline and the rate of flow PI rise is significantly different across the lifespan between sexes and the rate of MCAv change in females varies within specific age ranges. Sex differences in vascular contributions to brain health may play a role in the previously reported observation of more severe cerebrovascular disease (66, 81) and greater lifetime risk of stroke (70) and Alzheimer’s disease (23) in females compared with in males.

GRANTS

S. A. Billinger’s contributing studies were supported in part by American Heart Association Grant 16GRNT30450008, the National Institutes of Health (NIH) K01-HD-067318 from the Eunice Kennedy Shriver National Institute of Child Health and Human Development, and the Wohlgemuth Faculty Scholar Award. Additional support was provided by the University of Kansas Alzheimer’s Disease Center (P30 AG-035982), the Institutional Clinical and Translational Science Award, NIH/NCATS Grant Number UL1TR000001, NIH National Institute on Aging (NIA) R01 AG-043962, and the Landon Center on Aging endowments. The Georgia Holland Research in Exercise and Cardiovascular Health (REACH) laboratory space was supported by the Georgia Holland Endowment Fund (KUMC). J. N. Barnes’s contributing studies were supported in part by the NIH grants HL118154 and Alzheimer’s Association AARG17-499398. P. Brassard’s contributing studies were supported by the Ministère de l’Éducation, du Loisir et du Sport du Québec and the foundation of the Institut Universitaire de Cardiologie et de Pneumologie de Québec. Additional support was provided by The Leo and Anne Albert Charitable Trust.

DISCLAIMERS

The contents are solely the responsibility of the authors and do not necessarily represent the official view of the NIH.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

M.R.A. J.N.B., P.B., and S.A.B. conceived and designed research; M.R.A., S.E.A., and S.A.B. analyzed data; M.R.A., S.E.A., C.S.K., J.N.B., P.B., J.L.W., K.B.M., A.J.H., L.L., and S.A.B. interpreted results of experiments; M.R.A. and S.E.A. prepared figures; M.R.A., S.E.A., C.S.K. P.B., and S.A.B. drafted manuscript; M.R.A., S.E.A., C.S.K., J.N.B., P.B., J.L.W., K.B.M., A.J.H., L.L., and S.A.B. edited and revised manuscript; M.R.A., S.E.A., C.S.K., J.N.B., P.B., J.L.W., K.B.M., A.J.H., L.L., and S.A.B. approved final version of manuscript.