Midlife aerobic exercise and dynamic cerebral autoregulation: associations with baroreflex sensitivity and central arterial stiffness

Abstract

Midlife aerobic exercise may significantly impact age-related changes in the cerebro- and cardiovascular regulations. This study investigated the associations of midlife aerobic exercise with dynamic cerebral autoregulation (dCA), cardiovagal baroreflex sensitivity (BRS), and central arterial stiffness. Twenty middle-aged athletes (MA) who had aerobic training for >10 yr were compared with 20 young (YS) and 20 middle-aged sedentary (MS) adults. Beat-to-beat cerebral blood flow velocity, blood pressure (BP), and heart rate were measured at rest and during forced BP oscillations induced by repeated sit-stand maneuvers at 0.05 Hz. Transfer function analysis was used to calculate dCA and BRS parameters. Carotid distensibility was measured by ultrasonography. MA had the highest peak oxygen uptake (V̇o_2peak) among all groups. During forced BP oscillations, MS showed lower BRS gain than YS, but this age-related reduction was absent in MA. Conversely, dCA was similar among all groups. At rest, BRS and dCA gains at low frequency (∼0.1 Hz) were higher in the MA than in MS and YS groups. Carotid distensibility was similar between MA and YS groups, but it was lower in the MS. Across all subjects, V̇o_2peak was positively associated with BRS gains at rest and during forced BP oscillations (r = 0.257∼0.382, P = 0.003∼0.050) and carotid distensibility (r = 0.428∼0.490, P = 0.001). Furthermore, dCA gain at rest and carotid distensibility were positively correlated with BRS gain at rest in YS and MA groups (all P < 0.05). These findings suggest that midlife aerobic exercise improves central arterial elasticity and BRS, which may contribute to cerebral blood flow (CBF) regulation through dCA.

NEW & NOTEWORTHY Middle-aged athletes (MA) showed intact dynamic cerebral autoregulation (dCA) during sit-stand maneuvers when compared with young (YS) and middle-aged sedentary (MS) adults. Conversely, MA showed the significant attenuation of age-related carotid distensibility and baroreflex sensitivity (BRS) impairments. In MA and YS groups, BRS was positively associated with dCA gain at rest and carotid distensibility. Our findings suggest that midlife aerobic exercise improves BRS by reducing central arterial stiffness, which contributes to CBF regulation through dCA.

INTRODUCTION

Midlife vascular dysfunction may significantly impact brain health and increase the risk of cognitive impairment and dementia in later life (1–3). Particularly, impaired blood pressure (BP) and cerebral blood flow (CBF) regulations and increased central arterial stiffness may elevate the risk of cognitive decline (4, 5). In the lack of effective treatment for preventing or slowing age-related cognitive decline and impairment (6), improving midlife vascular function may reduce the future risk of cerebrovascular and neurological diseases (7).

Arterial BP is partly controlled by baroreceptor reflex, whereas CBF is regulated by dynamic cerebral autoregulation (dCA) in time scales of several seconds to minutes (8). The baroreflex controls short-term BP changes through mechanical distortion of the arterial baroreceptors located in the aortic arch and carotid sinus and provides feedback control of heart rate (HR) and systemic vascular resistance (9). On the other hand, dCA maintains CBF in response to BP fluctuations through adjustment of cerebrovascular resistance (CVR) (4), which may be modulated by autonomic neural and myogenic functions (10).

Mounting evidence suggests that vascular function decreases after middle age (11–14). The aortic and carotid stiffness starts increasing after midlife and progressively worsens in later life (11, 13). The age-related arterial stiffening can decrease mechanical distortion of the vessel wall in response to change in transmural pressure (i.e., arterial BP) and thereby baroreflex sensitivity (BRS) (9). Age may also alter dCA. Our recent study showed that low-frequency dCA gain at rest was higher in old adults than in younger adults (14). Moreover, a significant correlation between dCA and BRS has been observed in young but not middle-aged or older adults (14, 15). Collectively, these observations suggest that age may significantly alter CBF regulation due to both the local and systemic changes in cerebro- and cardiovascular functions, including dCA, BRS, and central arterial stiffness.

Midlife aerobic exercise may delay the onset of vascular aging and improve CBF regulation and brain health (16). The benefits of regular aerobic exercise for reducing central arterial stiffness have been well documented (13, 17). Besides, regular aerobic exercise is associated with increased cardiovagal BRS (12, 18). Conversely, the impact of aerobic exercise on dCA has been less clear, with some studies showing no change but the others showing impaired dCA in endurance athletes (18–20). However, there is currently no study that comprehensively investigated the impact of age and aerobic exercise training on dCA, BRS, and central arterial stiffness in middle-aged adults. Understanding the interactions of CBF regulation (e.g., dCA) with BRS and central arterial stiffness in aerobically trained middle-aged adults may contribute to elucidating the physiological mechanism by which midlife physical activity is associated with the reduced risk of late-life dementia (21).

The purpose of this study was to determine the associations among dCA, cardiovagal BRS, and central arterial stiffness in aerobically trained middle-aged athletes (MA) and sedentary middle-aged (MS) and young (YS) adults. Based on the literature, we hypothesized that MA would exhibit intact dCA that is associated with higher cardiovagal BRS and lower central arterial stiffness when compared with the MS and/or YS groups.

MATERIALS AND METHODS

Participants

YS and MS participants were recruited for our previous studies that investigated the effect of normal aging on brain and vascular functions (14, 22). MA participants were newly recruited for the present study, and their neurocognitive and hemodynamic data that do not pertain to the purpose of the present study have been published from our laboratory (16, 23). The proportions of men and women in the YS and MS groups were matched to the MA group because dCA and BRS may be influenced by sex (14). In this study, middle age was defined as 45–64 yr according to the widely accepted retirement age (24). Young age was defined as <45 yr consistently with our previous study (14, 16).

To ensure recruitment of aerobically trained adults, MA participants had to meet all of the following criteria: 1) at least 10 yr of aerobic exercise training (e.g., running, cycling, swimming, or multimodal training with moderate-to-vigorous intensity), 2) participation in at least two competing events per year, and 3) currently training for an event. Recruitment was conducted through a community-based advertisement using local newspapers and word-of-mouth at running clubs in the Dallas-Fort Worth metropolitan area. The criteria for YS and MS groups included no participation in structured exercise or physical activity program for the past >2 yr and <90 min of moderate-to-vigorous physical activity (>4.0 metabolic equivalents) per week that was individually confirmed by an Actical accelerometer (Actical, Philips Respironics, USA) (14).

Exclusion criteria for all groups included uncontrolled hypertension or diabetes; body mass index (BMI) >35 kg/m²; currently smoking or a history of smoking in the past 2 yr; pregnancy; and the presence or a history of cerebrovascular, metabolic, neurodegenerative, psychiatric, or inflammatory disease, brain trauma, hypothyroidism, active alcoholism, or drug abuse. Screening procedures included detailed medical history and medication questionnaires, a comprehensive physical examination, 12-lead electrocardiogram (ECG), and carotid atherosclerotic plaque or stenosis with >50% occlusion evaluated by ultrasound image at the common carotid and/or internal carotid artery (25).

This study was approved by the Institutional Review Board of the University of Texas Southwestern Medical Center and the Texas Health Presbyterian Hospital Dallas and performed in accordance with the guidelines of the Declaration of Helsinki and Belmont Report. All participants provided an informed written consent before participation. All data were de-identified before analysis.

Study Protocol

All participants underwent hemodynamic and cardiorespiratory fitness assessments on separate days. All assessments were completed within 2 mo from consented date. Participants were instructed to abstain from caffeinated beverages, alcohol, and vigorous exercise for at least 24 h before hemodynamic assessment. On the day of cardiorespiratory fitness evaluation, participants were allowed to have a light meal until a couple of hours before testing. All data were collected in an environmentally controlled laboratory with an ambient temperature of ∼22°C.

DCA and BRS Assessment

Five minutes of baseline data were collected during spontaneous breathing after resting in the quiet, seated position for >15 min. Next, repeated sit-stand maneuvers were performed for 5 min with duty cycles of 10-s sit and 10-s stand. During the maneuver, an investigator verbally coached each participant. This protocol was designed to induce hemodynamic oscillations that are clinically relevant for postural BP and CBF changes (26) and to improve coherence between hemodynamic variables at a point frequency of 0.05 Hz (26). The 0.05 Hz was chosen because both dCA and BRS are effective at this frequency (26, 27).

Instrumentation.

HR was measured by 3-lead ECG (Hewlett-Packard, Andover, MA). Arterial BP was recorded from the left middle finger at the participant’s heart level using photoplethysmography (Finapres Medical Systems, Amsterdam, The Netherlands). Intermittent arterial BP was measured from the right arm using ECG-gated electrosphygmomanometer (Tango+, Suntech Medical, Morrisville, NC). The brachial BP recording was used to corroborate the finger measurement so that finger mean arterial pressure (MAP) and DBP values were close to brachial BP values. End-tidal carbon dioxide (EtCO₂), a surrogate measure of the arterial partial pressure of CO₂ (28), was monitored with a nasal cannula using capnograph (Capnogard, Novamatrix). Participants were instructed to exhale through their noses during data collection. CBF velocity (CBFV) was measured from the right middle cerebral artery (MCA) using a 2-MHz transcranial Doppler probe (TCD: Multi-Dop X2, Compumedics/DWL, Singen, Germany) over the temporal window. The Doppler sampling depth ranged from 42 to 55 mm and the insonation angle was adjusted to optimize the signal quality for each individual subject according to the standard procedure (29). A headgear (Spencer Technologies, Northborough, MA) was used to fix the TCD probe to maintain the optimal signal. TCD has a high-temporal resolution (>100 ms) and allows noninvasive and repeatable measurement of changes in CBFV on a beat-to-beat basis under the assumption that the insonated vessel diameter remains constant. All data were simultaneously recorded at a sampling frequency ≥250 Hz and stored for offline analysis.

Preprocessing.

Hemodynamic data were analyzed by AcqKnowledge (BIOPAC Systems, Goleta, CA) and DADiSP (Newton, MA) software according to the procedures described in detail elsewhere (30). Briefly, artifacts (e.g., spikes in TCD signal) were visually identified and cleaned before spectral and transfer function analyses. First, all hemodynamic data were averaged for each cardiac cycle to obtain beat-to-beat mean values (14). Normalized CBFV (CBFV%) was calculated relative to the mean value of analyzed data segment (14, 18). Next, the time series of beat-to-beat HR, R-R interval, BP, CBFV, and CBFV% were linearly interpolated, resampled at 2 Hz to obtain equidistant time series, and detrended by the third-order polynomial curve fitting. Finally, these data were subdivided into 256-data point segments (128 s) with 50% overlap, and a Hanning window was applied for spectral and transfer function analyses (30). The processing above resulted in the spectral resolution of 0.0078125 Hz. CVR index (CVRi) was calculated by dividing MAP by mean CBFV. Pulsatility index (PI) was calculated as [(systolic CBFV − diastolic CBFV)/mean CBFV].

Cardiovascular variability and BRS calculations.

Consistent with our previous publications (14, 18, 31), spectral powers of R-R interval, systolic BP (SBP), and diastolic BP (DBP) were calculated from the low-frequency range (LF: 0.05–0.15 Hz) at baseline and a point frequency of 0.05 Hz during repeated sit-stand maneuvers. BRS was determined by transfer function gain between SBP and R-R interval in LF at baseline and a point frequency of 0.05 Hz during sit-stand maneuvers. The frequency band of 0.05–0.15 Hz was selected a priori based on the literature that coherence between SBP and R-R interval is low below 0.05 Hz, which may compromise the validity of estimating cardiovagal baroreflex function (32). Also, spectral power and transfer function estimate at frequency >0.15 Hz may be influenced by respiration (31).

Cerebrovascular variability and dCA calculations.

Spectral powers of MAP, CBFV, and CBFV% variability, as well as their transfer function estimates, were calculated from very low (VLF: 0.02–0.07 Hz), low (LF: 0.07–0.20 Hz), and high (HF: 0.20–0.35 Hz) frequency ranges at baseline and at a point frequency of 0.05 Hz during repeated sit-stand maneuvers as reported in our previous studies (14, 18, 26, 30). The explanation of transfer function parameters has been described previously (30). In brief, transfer function gain quantifies the magnitude relationship between the input and output signals (i.e., SBP → R-R interval for cardiovagal BRS and MAP → CBFV or CBFV% for dCA), whereas the phase represents their temporal displacement. Coherence represents a strength of the linear relation between the input and output signals, which ranges from 0 to 1, with 1 being the perfect linearity. The physiological interpretations of dCA and BRS transfer function parameters have been described by previous studies (15, 31). Briefly, effective dCA would show the smaller gain and larger phase, whereas the higher BRS gain and smaller phase would suggest the better baroreflex function.

Central Arterial Stiffness

Central arterial stiffness was assessed with carotid-femoral pulse wave velocity (cfPWV) and carotid distensibility that was measured from the right common carotid artery (CCA). After resting in the supine position for >15 min, cfPWV was measured between the right carotid and left femoral arteries at least twice using applanation tonometry (SphygmoCor 8.0, AtCor Medical) (33). Carotid distensibility was assessed by ultrasonography and applanation tonometry from >15 cardiac cycles. Briefly, the right CCA images were obtained by a 3-to-12 MHz linear array transducer in duplex ultrasonography mode (CX50, Philips Ultrasound, Bothell, WA). The systolic and diastolic CCA diameters were measured using an edge-detection software (Vascular Tool 5; Medical Imaging Applications, Coralville, IA). Applanation tonometry (SphygmoCor 80, AtCor Medical) was used to obtain a continuous arterial pressure waveform from the right carotid artery. The carotid pressure waveform was calibrated to the brachial MAP and DBP using a procedure described previously (34). Carotid distensibility was calculated by the following equation (35).

$$\text{Carotid distensibility = }\{\text{2 × }\left[\right(D_{\mathrm{s}} -D_{\mathrm{d}})\text{/}{D}_{\mathrm{d}}\left]\text{/}\left({\text{P}}_{\text{s }}-{\text{ P}}_{\text{d}}\right)\text{ × 1,000}\right\}$$

where D_s and D_d represent the systolic and diastolic diameters and P_s and P_d are the systolic and diastolic carotid arterial pressures.

Carotid intima-media thickness (CIMT) was assessed as a surrogate measure of carotid atherosclerotic severity. Briefly, a >10-mm segment of the far wall IMT at the CA was measured proximal to the carotid bifurcation at the end-diastolic phase by using automatic edge-detection software (Vascular Tool 5; Medical Imaging Applications, Coralville, IA) (36).

Cardiorespiratory Fitness

The peak oxygen uptake (V̇o_2peak) was assessed by a modified Astrand–Saltin protocol using a treadmill (37). The treadmill grade was increased by 2% every 2 min until exhaustion while participants walked, jogged, or ran at a fixed speed. The speed was selected based on the individual cardiorespiratory fitness level, which was determined by a submaximal test conducted before V̇o_2peak testing. Between submaximal and V̇o_2peak testing, a ∼5-min recovery time was given to ensure that participants’ vital signs return to the resting values. During V̇o_2peak testing, oxygen uptake (V̇o₂), carbon dioxide production, and respiratory exchange ratio (RER) were continuously monitored. V̇o₂ was measured during the second minute of each stage using the Douglas bag method (38). Gas fractions were analyzed by mass spectrometry (Marquette MGA 1100), and ventilation volume was measured by a Tissot spirometer. Mass spectrometry and gas sampling system were calibrated before each test to ensure measurement accuracy and reliability. Exercise BP, 12-lead ECG, and HR were monitored continuously by a registered nurse or a board-certified cardiologist. The criteria to confirm that V̇o_2peak was achieved included an increase in V̇o₂ <150 mL despite increasing work rate of 2% grade, RER ≥1.1, and HR within 5 beats/min of the age-predicted maximal value (220 – age). In all cases, at least two of these criteria were achieved, confirming the identification of V̇o_2peak based on the American College of Sports Medicine guidelines (39). V̇o_2peak was defined as the highest V̇o₂ measured from a >30-s Douglas bag during the last stage of testing.

Statistical Analysis

One-way analysis of variance (ANOVA) was used for comparing continuous variables among the YS, MS, and MA groups. In addition, analysis of covariance was performed to adjust for potential covariates. In case of a significant main effect, post hoc pairwise comparisons were performed with the Bonferroni correction. Simple linear correlations were assessed by the Pearson product-moment correlation analysis. Partial correlation analysis was also used to adjust for age and sex which could significantly impact outcome measures (i.e., dCA, BRS, carotid distensibility, cfPWV, and V̇o_2peak). The results were also confirmed by nonparametric test (e.g., Kruskal–Wallis one-way ANOVA and Spearman’s correlation analysis). Statistical significance was set a priori at P < 0.05 for two-tailed tests. All data are reported as means ± standard deviation. Data were analyzed with SPSS 20.0 (SPSS, Inc, Chicago, IL).

RESULTS

Participant Characteristics and Baseline Hemodynamics

All groups were similar in height, body mass, BMI, and brachial BP (Table 1). EtCO₂, respiratory rate, CBFV, PI, and CVRi were also similar among groups. MA had significantly greater V̇o_2peak and lower HR than the YS and MS groups, and YS had greater V̇o_2peak than MS. MA and YS had significantly higher carotid distensibility and lower cfPWV than MS. However, these differences in the arterial stiffness measures disappeared after adjusting for HR. Carotid IMT in MA and MS was thicker than YS.

Table 1.

Subject characteristics and baseline hemodynamics

Variable	Young Sedentary	Middle-Aged Sedentary	Middle-Aged Athlete	P Value
Men/Women, n	9/11	9/11	9/11
Age, yr	32 ± 7	53 ± 5*	53 ± 4*	<0.001
Height, cm	168 ± 9	170 ± 9	173 ± 8	0.268
Body mass, kg	69 ± 14	76 ± 15	72 ± 12	0.219
BMI, kg/m2	24 ± 4	26 ± 4	24 ± 3	0.148
V̇o2max, mL/kg/min	35.2 ± 6.6	27.8 ± 4.4*	42.6 ± 5.1*†	<0.001
Steady-state hemodynamics
HR, beats/min	75 ± 6	72 ± 10	54 ± 10*†	<0.001
Brachial BP, mmHg
Systolic	117 ± 11	116 ± 10	121 ± 12	0.309
Diastolic	73 ± 11	77 ± 7	77 ± 9	0.221
Mean	88 ± 9	90 ± 7	92 ± 8	0.303
Pulse	44 ± 11	38 ± 8	44 ± 11	0.146
CBFV, cm/s
Systolic	92 ± 15	82 ± 15	84 ± 21	0.141
Diastolic	36 ± 5	33 ± 6	32 ± 5	0.075
Mean	56 ± 7	52 ± 9	53 ± 11	0.314
PI	0.99 ± 0.18	0.93 ± 0.11	0.98 ± 0.14	0.301
cfPWV, m/s	6.2 ± 1.1	7.6 ± 1.6*	6.3 ± 0.6†	<0.001
Carotid distensibility, 10−3·mmHg−1	4.1 ± 0.9	3.2 ± 0.8*	4.1 ± 0.8†	<0.001
Carotid IMT, mm	0.50 ± 0.06	0.60 ± 0.14*	0.62 ± 0.11*	<0.001
CVRi, mmHg/cm/s	1.69 ± 0.35	1.90 ± 0.39	1.86 ± 0.49	0.244
EtCO2, mmHg	38 ± 3	38 ± 4	37 ± 3	0.811
Respiratory rate, breaths/min	16 ± 3	15 ± 3	16 ± 3	0.746

Data are means ± standard deviation. Bold values represent P < 0.05. BMI, body mass index; BP, blood pressure; CBFV, cerebral blood flow velocity; cfPWV, carotid-femoral pulse wave velocity; CVRi, cerebrovascular resistance index; EtCO₂, end-tidal CO₂; HR, heart rate; IMT, intima-media thickness; PI, pulsatility index; V̇o_2max, maximal oxygen uptake.

*P < 0.05 vs. sedentary young; †P < 0.05 vs. middle age.

Cardiovascular Variability and BRS

Table 2 and Fig. 1 summarize the results of spectral analysis of R-R interval, SBP, and DBP variabilities in the LF range (0.05–0.15 Hz). R-R interval variabilities at rest and during sit-stand maneuvers were significantly higher in MA than those in MS. SBP variability during sit-stand maneuvers and DBP variabilities at rest and during sit-stand maneuvers were significantly higher in MA than those in YS.

Table 2.

Spectral powers of blood pressure, R-R interval, and cerebral blood flow velocity measured at rest and repeated sit-stand maneuvers at 0.05 Hz

Variable	Young Sedentary	Middle-Aged Sedentary	Middle-Aged Athlete	P Value
Cardiovascular hemodynamics
Low frequency (0.05–0.15 Hz)
R-R interval, ms2	445 ± 343	282 ± 159	631 ± 558†	0.023
SBP, mmHg2	7.28 ± 4.67	6.43 ± 3.21	5.52 ± 6.66	0.550
DBP, mmHg2	3.59 ± 2.60	2.50 ± 1.72	1.42 ± 1.17*	0.003
Sit-stand (0.05 Hz)
R-R interval, ms2	1,907.82 ± 1,671.91	1,048.88 ± 887.06	2,467.26 ± 1,635.40†	0.011
SBP, mmHg2	35.64 ± 37.14	51.11 ± 40.74	74.59 ± 57.48*	0.033
DBP, mmHg2	12.67 ± 9.87	13.56 ± 7.91	22.14 ± 16.55*	0.029
Cerebral hemodynamics
Very low frequency (0.02–0.07 Hz)
MAP, mmHg2	4.12 ± 2.58	5.40 ± 3.57	3.81 ± 2.75	0.217
CBFV%, %2	10.71 ± 6.17	17.22 ± 15.47	15.49 ± 9.74	0.169
Low frequency (0.07–0.20 Hz)
MAP, mmHg2	3.16 ± 1.58	2.57 ± 1.58	1.65 ± 1.76*	0.018
CBFV%, %2	9.75 ± 5.37	8.28 ± 5.07	7.20 ± 4.95	0.294
High frequency (0.20–0.35 Hz)
MAP, mmHg2	0.36 ± 0.32	0.60 ± 0.55	0.42 ± 0.39	0.208
CBFV%, %2	2.00 ± 1.47	2.27 ± 1.75	1.89 ± 1.80	0.767
Sit-stand (0.05 Hz)
MAP, mmHg2	19.00 ± 15.80	23.41 ± 14.76	36.27 ± 28.16*	0.028
CBFV%, %2	47.61 ± 33.87	58.95 ± 41.10	94.84 ± 67.39*	0.010

Data are means ± standard deviation. Bold values represent P < 0.05. CBFV, cerebral blood flow velocity; DBP, diastolic blood pressure; MAP, mean arterial pressure; SBP, systolic blood pressure.

*P < 0.05 vs. sedentary young; †P < 0.05 vs. middle age.

Figure 1.

Power spectral densities of R-R interval (A and B), systolic blood pressure (SBP; C and D), and diastolic blood pressure (DBP; E and F) measured at rest and during repeated sit-stand maneuvers performed at 0.05 Hz. Each line represents group-averaged data. Low frequency (LF: 0.05–0.15 Hz).

At rest, BRS gain was significantly higher in MA than in YS and MS. During sit-stand maneuvers, BRS gain was significantly higher in YS than in MS, but it was similar between YS and MA (Table 3 and Fig. 2).

Table 3.

Transfer function data of cardiovagal baroreflex function and dynamic cerebral autoregulation

Variable	Young Sedentary	Middle-Aged Sedentary	Middle-Aged Athlete	P Value
Cardiovagal baroreflex function
Low frequency (0.05–0.15 Hz)
Gain, ms/mmHg	7.23 ± 2.57	6.37 ± 1.77	10.88 ± 4.20*†	<0.001
Phase, rad	−0.91 ± 0.29	−1.06 ± 0.42	−1.16 ± 0.42	0.108
Coherence	0.67 ± 0.11	0.68 ± 0.09	0.61 ± 0.16	0.148
Sit-stand (0.05 Hz)
Gain, ms/mmHg	8.17 ± 3.57	4.96 ± 2.55*	6.09 ± 1.86	0.002
Phase, rad	−0.66 ± 0.50	−0.67 ± 0.22	−0.84 ± 0.32	0.196
Coherence	0.93 ± 0.09	0.94 ± 0.06	0.96 ± 0.05	0.522
Dynamic cerebral autoregulation
Very low frequency (0.02–0.07 Hz)
Gain, %/mmHg	1.25 ± 0.36	1.36 ± 0.42	1.69 ± 0.51*	0.005
Phase, rad	0.92 ± 0.54	0.81 ± 0.61	0.72 ± 0.39	0.484
Coherence	0.56 ± 0.11	0.62 ± 0.19	0.60 ± 0.13	0.380
Low frequency (0.07–0.20 Hz)
Gain, %/mmHg	1.65 ± 0.45	1.73 ± 0.39	2.09 ± 0.54*†	0.009
Phase, rad	0.52 ± 0.26	0.62 ± 0.21	0.44 ± 0.17†	0.033
Coherence	0.71 ± 0.16	0.70 ± 0.12	0.64 ± 0.12	0.175
High frequency (0.20–0.35 Hz)
Gain, %/mmHg	2.07 ± 0.47	1.90 ± 0.43	1.92 ± 0.42	0.399
Phase, rad	0.16 ± 0.21	0.19 ± 0.25	0.17 ± 0.29	0.915
Coherence	0.69 ± 0.19	0.72 ± 0.12	0.67 ± 0.16	0.579
Sit-stand (0.05 Hz)
Gain, %/mmHg	1.69 ± 0.41	1.58 ± 0.42	1.64 ± 0.25	0.667
Phase, rad	0.79 ± 0.30	0.85 ± 0.19	0.74 ± 0.20	0.309
Coherence	0.96 ± 0.03	0.98 ± 0.01	0.98 ± 0.01*	0.023

Data are means ± standard deviation. Bold values represent P < 0.05.

*P < 0.05 vs. sedentary young; †P < 0.05 vs. middle age.

Figure 2.

Transfer function gain (A and B), phase (C and D), and coherence (E and F) of the cardiovagal baroreflex sensitivity measured at rest and during repeated sit-stand maneuvers performed at 0.05 Hz. Each line represents group-averaged data. Low frequency (LF: 0.05–0.15 Hz).

Cerebrovascular Variability and dCA

Spectral powers of MAP, CBFV, and CBFV% are presented in Table 2 and Fig. 3. At rest, MAP variability in the LF range was significantly lower in MA than in YS. During sit-stand maneuvers, MAP, CBFV, and CBFV% variabilities were significantly higher in MA than in YS.

Figure 3.

Power spectral densities of mean arterial pressure (MAP; A and B) and cerebral blood flow velocity (CBFV%; C and D) measured at rest and during repeated sit-stand maneuvers performed at 0.05 Hz. Each line represents group-averaged data. Very low frequency (VLF: 0.02–0.07 Hz), low frequency (LF: 0.07–0.20 Hz), and high frequency (HF: 0.20–0.35 Hz).

At rest, MA had higher dCA gains at the VLF and LF ranges than YS and MS. The dCA phase at the LF was significantly smaller in MA than in YS and MS (Table 3 and Fig. 4). These findings suggest less effective dCA at rest in MA compared with YS and MS. During sit-stand maneuvers, these differences in dCA gain and phase were not observed among the three groups.

Figure 4.

Transfer function gain (A and B), phase (C and D), and coherence (E and F) of dynamic cerebral autoregulation measured at rest and during repeated sit-stand maneuvers performed at 0.05 Hz. Each line represents group-averaged data. Very low frequency (VLF: 0.02–0.07 Hz), low frequency (LF: 0.07–0.20 Hz), and high frequency (HF: 0.20–0.35 Hz).

Correlation Analysis

Table 4 shows simple correlations among V̇o_2peak, carotid distensibility, cfPWV, dCA, and BRS across all subjects. V̇o_2peak was correlated with lower cfPWV, higher carotid distensibility, and higher BRS at rest and during sit-stand maneuvers and higher LF dCA gain at rest across all subjects. Moreover, lower cfPWV was correlated with higher BRS gain at rest and during sit-stand maneuvers. Higher carotid distensibility was correlated with higher BRS at rest and during sit-stand maneuvers. With adjustment for age and sex, partial correlation analysis further showed that greater V̇o_2peak was correlated with lower cfPWV (r = −0.597, P < 0.001), higher carotid distensibility (r = 0.341, P = 0.010), and higher BRS at rest (r = 0.404, P = 0.002) and during sit-stand maneuvers (r = 0.320, P = 0.016) and higher LF dCA gain at rest (r = 0.266, P = 0.048) across all subjects. Moreover, lower cfPWV was correlated with higher BRS gain at rest (partial r = −0.355, P = 0.007) and during sit-stand maneuvers (partial r = −0.365, P = 0.006). Higher carotid distensibility was correlated with higher BRS at rest (partial r = 0.548, P < 0.001) and during sit-stand maneuvers (partial r = 0.306, P = 0.013).

Table 4.

Pearson’s product-moment correlation coefficients and (P values) illustrating the associations among V̇o_2peak, central arterial stiffness, dCA, and BRS

	(B)	(C)	(D)	(E)	(F)	(G)	(H)	(I)	(J)	(K)
(A) V̇o2peak	0.277 (0.034)	−0.477 (<0.001)	0.235 (0.074)	−0.273 (0.036)	−0.054 (0.684)	−0.239 (0.068)	0.382 (0.003)	0.130 (0.326)	0.257 (0.050)	−0.162 (0.220)
(B) Carotid distensibility		−0.468 (<0.001)	0.177 (0.175)	−0.334 (0.009)	0.263 (0.043)	−0.143 (0.275)	0.490 (<0.001)	−0.056 (0.670)	0.428 (<0.001)	−0.001 (0.992)
(C) cfPWV			−0.160 (0.227)	0.324 (0.012)	−0.148 (0.264)	0.170 (0.199)	−0.279 (0.033)	−0.058 (0.664)	−0.491 (<0.001)	0.148 (0.263)
(D) dCA gain (rest)				−0.352 (0.006)	0.163 (0.213)	−0.058 (0.661)	0.584 (<0.001)	−0.179 (0.172)	0.040 (0.763)	0.075 (0.570)
(E) dCA phase (rest)					−0.267 (0.039)	0.399 (0.002)	−0.263 (0.042)	0.267 (0.039)	−0.046 (0.726)	0.240 (0.065)
(F) dCA gain (sit-stand)						−0.227 (0.080)	0.018 (0.889)	−0.132 (0.313)	0.207 (0.113)	−0.241 (0.064)
(G) dCA phase (sit-stand)							−0.039 (0.767)	−0.087 (0.511)	−0.042 (0.749)	0.437 (<0.001)
(H) BRS gain (rest)								−0.013 (0.924)	0.156 (0.234)	0.071 (0.592)
(I) BRS phase (rest)									0.380 (0.003)	0.301 (0.020)
(J) BRS gain (sit-stand)										−0.061 (0.642)
(K) BRS phase (sit-stand)

Bold values represent P < 0.05. BRS, cardiovagal baroreflex sensitivity; cfPWV, carotid-femoral pulse wave velocity; dCA, dynamic cerebral autoregulation; V̇o_2peak, peak oxygen uptake. dCA and BRS gain and phase in the low-frequency ranges at rest and during sit-stand maneuvers.

Figure 5 shows that higher BRS at rest and during sit-stand maneuvers were correlated with higher carotid distensibility in YS and MA groups but not the MS. The higher BRS at rest in YS and MA groups was also correlated with higher LF dCA gain at rest (Fig. 5).

Figure 5.

Simple correlations between carotid distensibility and cardiovagal baroreflex sensitivity (BRS) gain in the low-frequency range at rest (A) and during repeated sit-stand maneuvers (B), and between BRS and dynamic cerebral autoregulation (dCA) in the low-frequency range at rest (C).

DISCUSSION

This is one of the first studies that comprehensively investigated the associations of age and aerobic exercise with dCA, BRS, and central arterial stiffness in middle-aged and younger adults. The main findings from this study are as follows. First, compared with YS and MS groups, dCA in MA was intact during forced BP oscillations, but it was impaired during the spontaneous resting condition. Conversely, YS and MS groups showed similar dCA during the sit-stand maneuver and resting conditions. Second, during the resting condition, dCA gain was positively associated with BRS in MA and YS groups but not in MS group. Third, MA showed significantly higher BRS and lower aortic and carotid stiffness than YS and MS groups. Moreover, BRS and vascular stiffness measures were correlated in YS and MA groups but not in the MS group. Therefore, these findings collectively suggest that aerobic exercise training during middle age reduces central arterial stiffness and improves BRS, which may contribute to CBF regulation via dCA.

Effect of Endurance Training on dCA

There has been a growing interest to understand the influence of aerobic exercise training on CBF regulation in normal aging adults and patients with cerebrovascular and neurological disease (21, 40). In normal adults, CBF decreases ∼6% per decade across the adult lifespan, as measured by MRI (41), although recent studies suggested that habitual physical activity may attenuate the age-related reduction (23, 29, 42). For example, we previously reported that regional cerebral perfusion measured by arterial spin labeling MRI is increased in aerobically trained old adults (42). In this present study, we did not observe such difference using TCD that measured mean CBFV at the MCA. This could be explained by the methodology of CBF measurement. TCD cannot quantify volumetric CBF without diameter measurement of the insonated artery (43). Also, the effect of aerobic exercise training on CBF may be regional (44, 45), although this is still controversial and needs further investigations using different modalities of CBF measurements.

Age may have small impact on dCA during midlife (46). Using autoregulatory index, Carey et al. (46) compared young (≤40 yr) and older (≥55 yr) groups matched with sex and BP and found no group difference in dCA measured during lower body negative pressure, Valsalva maneuver, and spontaneous increase in BP during the resting condition. Using the transfer function analysis, we studied 136 healthy adults aged 21–80 yr and found that dCA during the spontaneous resting condition was similar between young (<45 yr) and middle-aged (45–64 yr) adults, but it was impaired in old adults (≥65 yr) compared with young adults. However, the age-related impairment in old adults disappeared during sit-stand maneuvers (14). Therefore, these findings collectively suggest that dCA remains intact during middle age but may be impaired in older adults, particularly during the spontaneous resting condition.

The impact of aerobic exercise training on dCA lacks consistent evidence and may be related to the assessment technique (18–20). Using transfer function analysis, Aengevaeren et al. (18) showed that old endurance athletes had lower VLF gain at rest than age-matched sedentary control subjects, which suggests better dCA in athletes, but sit-stand maneuvers abolished this group difference. Ichikawa et al. (20) showed that young athletes and age-matched sedentary adults have similar cerebral autoregulatory function measured by thigh cuff release technique. In contrast, Lind-Holst et al. (19) showed that dCA assessed by both thigh cuff release technique and transfer function analysis was impaired at the resting condition in aerobically trained young athletes. In the current study, MA group showed higher VLF and LF dCA gains and the lower LF phase than MS and YS groups during the resting condition, but these group differences disappeared during sit-stand maneuvers.

The discrepant findings of dCA assessed during the spontaneous resting and sit-stand maneuver conditions and the lack of their correlations (Table 4) may be explained by several reasons. First, cerebral autoregulation in essence represents a nonlinear relation between change in BP and CBF; therefore, dCA may be influenced by the magnitude of BP fluctuation (47). Second, dCA may be influenced by aspects other than the autoregulatory function. Particularly, the cerebrovascular Windkessel function may have a significant impact on dCA, independently from the active control of vascular tone (e.g., myogenic and autonomic neural controls) (48). In this regard, our data showed that carotid (extracranial) distensibility measured at rest was greater in MA group than in MS group. Thus, if MA participants also had greater “intracranial” cerebrovascular distensibility, dCA at rest may be different between groups. Third, dCA may be influenced by static cerebrovascular tone (i.e., CVR). Zhang et al. (49) reported that patients with untreated hypertension have greater CVR and lower dCA gain at rest. Moreover, antihypertensive treatment decreased CVR while increasing dCA gain (49). Therefore, dCA at rest and during sit-stand maneuvers may reveal different regulatory mechanisms (50, 51), such as active versus passive vascular responses to changes in arterial pressure, and may not necessarily relate to each other linearly when BP fluctuation (i.e., input to the autoregulatory system) is significantly augmented by sit-stand maneuvers.

Effect of Endurance Training on BRS

Cardiovagal BRS decreases with aging (12, 14) and improves with aerobic exercise training (12, 18). Consistently, we observed that MS had lower BRS gain during sit-stand maneuvers than YS group. On the other hand, MA showed the highest BRS gain at rest among all groups, whereas it was similar between YS and MA groups during repeated sit-stand maneuvers. The different results of BRS and dCA during the resting and sit-stand maneuver conditions may be related to lower coherence value or signal-to-noise ratio in the estimation of transfer function parameters during the resting condition, as discussed in detail above (Effect of Endurance Training on dCA). Although the resting and sit-stand assessments may provide physiologically similar information, dCA and BRS measured during forced BP oscillations may be more relevant to clinical conditions, such as orthostatic hypotension and syncope, than the resting measurement (50).

There are several potential mechanisms by which aerobic exercise training can improve baroreflex function. First, aerobic exercise may reduce age-related central arterial stiffening, which in turn may improve mechanical transduction of BP changes and BRS gain (52). Consistent with this notion, we observed that carotid distensibility was greater in MA than in the MS group, and the greater carotid distensibility was associated with higher BRS gain. Second, aerobic exercise training may elevate vagal activities that have been shown to decrease with advancing age (9, 53). In the current study, MA had lower resting HR and greater variability of R-R interval than MS and YS groups, and MS had the lower R-R interval variability than YS. These results suggest that age decreased but aerobic exercise training increased vagal activities in our middle-aged participants. Third, aerobic exercise training may increase cardiac cholinergic responsiveness that may lead to increased BRS (12).

Higher SBP variability observed during sit-stand maneuvers in the MA group may be explained by the training-related changes in cardiovascular mechanics (54). The greater arterial and/or left ventricular distensibility, together with a steeper slope of the Starling relationship between the left ventricular filling pressure and stroke volume, may have resulted in large oscillations of stroke volume and increased SBP variability during sit-stand maneuvers in the MA group. Of note, these results are consistent with our previous findings from older endurance athletes (18).

In this study, aerobic exercise was associated with higher BRS in middle-aged adults, but previous studies showed that aerobic exercise attenuates BRS in young adults (55–58). Although the current study did not include young fit subjects and cannot determine the reason why the association between aerobic exercise and BRS differs by age groups, such difference may be related to age-related changes in autonomic neural function and/or elasticity of the barosensory arteries, as well as the methodology used to assess BRS (e.g., pharmacological vs. nonpharmacological).

Interaction between dCA and BRS

The arterial baroreflex controls short-term BP change, which subsequently may have a significant impact on CBF regulation via dCA (8). Previously, Tzeng et al. (15) reported an inverse correlation between cardiovagal BRS assessed by pharmacological intervention (i.e., sodium nitroprusside and phenylephrine hydrochloride bolus injections) and dCA assessed by rapid thigh cuff release technique and transfer function methods in young individuals. Based on these findings, they suggested that attenuated dCA function may be compensated by better control of BP through BRS to maintain stable CBF. Consistent with their findings, our previous study also reported the same correlation in young adults but not in middle-aged and older sedentary adults (14). These findings have been interpreted in a way that the compensatory interaction between BRS and dCA gains may be altered by aging. In the current study, we consistently observed the positive correlation between BRS and dCA gains at rest in YS participants, but this correlation was also present in the MA group (Fig. 5). Moreover, we found that BRS was positively associated with carotid distensibility in the MA and YS but not in MS groups (Fig. 5). Therefore, these associations may suggest that increased elasticity of the barosensory (i.e., carotid) artery through aerobic exercise training improves BRS gain which, in turn, may compensate for less effective dCA in middle-aged adults. Nevertheless, the correlations in Fig. 5 should be interpreted in the context of following cautions. First, we did not observe the correlation of carotid distensibility with BRS and dCA gains at rest and during sit-stand maneuvers in MS group. This may be explained by a possibility that reduced carotid distensibility in MS group led to smaller BRS gain and its individual variability, which subsequently may have resulted in a lack of the BRS-dCA gain correlation in MS group. Second, carotid distensibility only explained 21%–27% of the variances in BRS gains measured during sit-stand maneuvers in YS and MA groups. Therefore, BRS gain is not only determined by carotid distensibility but also other factors such as HR, arterial BP, and systemic vascular resistance.

Our data supported a hypothesis that the arterial baroreflex function may compensate for the cerebral autoregulatory function using a nonpharmacological method in young sedentary and middle-aged fit participants. In contrast to our hypothesis, Ogoh et al. (59) demonstrated that acute BRS dysfunction induced pharmacologically by autonomic blockade attenuated dCA function in young healthy individuals. Although we cannot determine a reason for the discrepant findings, autonomic neural function is not only associated with baroreflex function but also with cerebral autoregulatory functions (60). Thus, blocking the autonomic neural system may impair both BRS and dCA. These previous studies using pharmacological intervention should be interpreted with caution, as these drugs affect BP and BP variability that contributes to both dCA and BRS in addition to their effects on the autonomic loop of the baroreflex. Therefore, the relationship between BRS and dCA needs further investigation.

Clinical Perspectives

Mounting evidence suggests that impaired cardio- and cerebrovascular functions contribute to the progression of age-related cognitive decline (4, 5) and increase the risks of cognitive impairment and dementia, including Alzheimer’s disease (61). Conversely, regular physical activity and exercise during midlife have been shown to associate with the reduced risk of late-life cognitive impairment and dementia, which may potentially be mediated by improved cardio- and cerebrovascular health (3–5). Recently, we reported that middle-aged endurance athletes have lower CVR, better white matter fiber integrity, and greater regional cortical thickness than the age-matched sedentary participants (16, 23). Therefore, our current results are in line with these previous observations that support the benefits of midlife aerobic exercise and further add to the literature that aerobic exercise training improves BP and CBF regulations via improvements in central arterial stiffness, cardiovagal BRS, and dCA in middle-aged adults.

Strengths and Limitations

Several study limitations need to be acknowledged. First, this study is cross-sectional design that limits our understanding of causal relations between midlife endurance exercise and changes in BRS and dCA. However, conducting a longitudinal study with long-term exercise interventions for >10 yr would be daunting to address these questions, if not impossible. Second, the use of linear transfer function methods may limit its capability to quantify the nonlinear characteristics of cardio- and cerebrovascular hemodynamics. However, the average values of coherence function were >0.5 for both cardiovagal BRS and dCA assessment at rest. Also, we used repeated sit-stand maneuvers to examine the effect of forced BP oscillations on CBF. Third, cardiovagal BRS estimated in this study represents closed-loop properties of the baroreflex function which may not be extrapolated to the open-loop characteristics (62). However, baroreflex normally operates under closed-loop conditions, which may be more relevant to BP control in daily life. Fourth, changes in CBFV measured by TCD reflect changes in volumetric CBF, only if the insonated arterial diameter remains relatively constant. Recent studies using high-resolution MRI demonstrated change in the MCA diameter under moderate hypo- and hypercapnia (63), which is unlikely to have occurred in the present study with normal spontaneous breathing. The effects of respiration and sympathetic neural activity on the MCA diameter during sit-stand maneuvers are unknown. Fifth, we did not collect menopausal status or control menstrual cycle in female participants. Therefore, we cannot rule out the potential effect of sex hormones on our outcome measures. However, based on previous studies, the impact of menstrual cycle on cerebral hemodynamic regulation (i.e., mean CBFV, CVRi, and dCA) remains to be controversial (64). Sixth, the absence of young exercise-trained group would have been necessary to assess the interaction effects of age and exercise training. However, based on previous studies, the influence of exercise training status on central arterial stiffness and cardiovagal BRS is inconsistent in young individuals, with some studies showing less effects (12, 13) but the others showing attenuated BRS (55, 57, 58).

Despite these limitations, this study has several important strengths. First, we focused on aerobic exercise training in middle-aged adults who received relatively little attention in the past, but it is an important phase of the lifespan for the early detection and prevention of age-related vascular and neurological disorders (7). Therefore, the results from this study showing that midlife aerobic exercise has a significant impact on the BP and CBF regulations are likely to have important clinical implications. Second, our MA participants were trained well, as they have been exercising for ∼25 yr and had ∼90% of V̇o_2peak according to the American College of Sports Medicine guideline (39). Thus, our results should reflect at least some impact of aerobic exercise training on their cardio- and cerebrovascular functions when compared with the sedentary groups. Third, this study is strengthened by the comprehensive assessment of vascular functions, including central arterial stiffness, BRS, dCA, and V̇o_2peak within a single study. Moreover, repeated sit-stand maneuvers were used to induce clinically relevant BP oscillations and improve coherence function for estimating BRS and dCA. Finally, V̇o_2peak provided the objective, gold-standard index of cardiorespiratory fitness and the complementary results of the group-level analysis.

Conclusions

Middle-aged endurance athletes exhibited intact dCA during forced BP oscillations, higher cardiovagal BRS, and lower central arterial stiffness than sedentary middle-aged and younger adults. On the other hand, dCA during the spontaneous resting condition was impaired in these athletes compared with the sedentary groups. During rest, dCA gain was positively correlated with BRS in middle-aged athletes and young sedentary adults but not in the middle-aged sedentary group. Therefore, these findings collectively suggest that aerobic exercise training in middle-aged adults improves central arterial stiffness and BRS, which may contribute to CBF regulation via dCA, particularly during the spontaneous resting condition.

GRANTS

This work was supported by National Heart, Lung, and Blood Institute Grants K99HL133449 and R01HL102457 and Japanese Society for the Promotion of Science Grant 19K19970.

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

R.Z. and T. Tarumi conceived and designed research; T. Tomoto, J.R., and T. Tarumi performed experiments; T. Tomoto and T. Tarumi analyzed data; T. Tomoto, R.Z., and T. Tarumi interpreted results of experiments; T. Tomoto and T. Tarumi prepared figures; T. Tomoto drafted manuscript; T. Tomoto, J.R., R.Z., and T. Tarumi edited and revised manuscript; T. Tomoto, J.R., R.Z., and T. Tarumi approved final version of manuscript.