One-year exercise improves cognition and fitness and decreases vascular stiffness and reactivity to CO2 in amnestic mild cognitive impairment

Abstract

Background

Amnestic mild cognitive impairment (aMCI) is often a precursor stage to Alzheimer's disease (AD). Aerobic exercise (AE) has received increasing attention in the prevention of AD. While there is some evidence that it improves neurocognitive function in older individuals, the effect of exercise in the long-term is not well understood.

Objective

To assess the effect of long-term exercise on cognition, fitness, vascular stiffness, and cerebrovascular reactivity (CVR).

Methods

In this prospective clinical trial, 27 aMCI participants were enrolled into two groups and underwent 12 months of intervention. One group (n = 11) underwent AE training (6M/5F, age = 66.2 years), and the control group (n = 16) performed stretch training (ST group, 9M/7F, age = 66.4 years). Both groups performed training three times per week with duration and intensity gradually increased over time. CVR was measured at pre- and post-training using blood-oxygenation-level-dependent MRI.

Results

In the AE group, aerobic fitness improved (p = 0.034) and carotid artery stiffness decreased (p = 0.005), which was not observed in the ST group. In all participants, decreases in carotid artery stiffness were associated with increases in aerobic fitness (p = 0.043). The AE group displayed decreases in CVR in the anterior cingulate cortex and middle frontal gyrus (p < 0.05, FWE corrected); the ST group did not show significant changes in CVR. Several measures of cognition (i.e., inhibition and delayed recall), neuropsychiatric symptoms, and functional status ratings improved only in the AE group.

Conclusions

These results suggest that AE may alter cerebral hemodynamics in patients with aMCI which may improve cognitive, psychological, and functional status.

Introduction

Aerobic exercise (AE) has gained attention in recent years as a potential non-pharmacological intervention to mitigate cognitive decline in populations at risk for Alzheimer's disease (AD). ¹ Modern advances in the treatment of AD have resulted in the development and approval of monoclonal antibodies against amyloid plaque. ² Despite the broad progress that has resulted in these achievements, the clinical significance of such therapies is still under question, with current research suggesting these therapies may at best help slow progression of AD by about 30%.^3,4 Given these clinical limitations, physical exercise has been explored as an adjuvant strategy to improve multiple domains of cognitive function such as speed, spatial, controlled, and executive tasks in some investigations. ⁵ However, the mechanism of physical exercise in slowing or preventing AD progression are not yet fully understood, and evidence on the impact of AE on progression of AD is still largely ambiguous. One opportunity is to utilize AE in those individuals at greater risk for AD but do not meet the criteria for dementia, such as in amnestic mild cognitive impairment (aMCI). aMCI is often a precursor stage to the development of AD, with declines in memory without the significant functional deficits present in dementia. ⁴

One proposed mechanism underlying AE and effects on cognition is improved cerebrovascular health. Decreased cerebral blood flow (CBF), and waste clearance due to microvascular damage may play a role in altered metabolism which may ultimately exacerbate or even lead to AD pathology. ⁶ In normal aging, CBF naturally declines; however, participants with AD and MCI have been shown to have decreased CBF when compared to healthy participants.^7–9 In contrast, AE may improve cardiorespiratory fitness and CBF in a way that may curb or reverse this dysfunction, resulting in cognitive improvements. Cardiorespiratory fitness characterizes the body's circulatory ability to deliver oxygen for utilization by tissue, and thus cardiovascular health may closely influence brain function. For example, AE has been shown to increase neuroplasticity of the brain by increasing processes such as gliogenesis and synaptogenesis, as well as being associated with increases in ¹⁰ white and gray matter volume. ¹⁰ Our previous work supports these findings, with an increase in CBF among older subjects undergoing AE, with corresponding improvements in cognitive function. ⁹ However, CBF data provides information about the static properties of blood vessels, i.e., its ability to supply blood, and does not provide any information on the dynamic properties of blood vessels, i.e., its ability to vasodilate in response to a challenge.

One common way to quantify the dynamic vasodilation capacity of vasculature is to measure cerebrovascular reactivity (CVR). CVR is the ability for blood vessels to dilate in response to a vasodilatory agent, e.g., CO2, and is measured using non-invasive blood oxygen level dependent (BOLD) functional MRI (fMRI). CVR has been shown to be a biomarker associated with cognitive decline in AD.^11,12 Prior studies, including our previous work in older athletes reported a negative relationship between CVR and measures of cardiorespiratory fitness assessed by peak oxygen consumption (VO2_peak).^13–15 Another parameter related to fitness and cerebrovascular health is arterial compliance, which has been shown to be higher in older individuals with greater fitness. However, the effects of fitness, cognition, arterial stiffness and CVR are unknown in aMCI.

Given the need for further study into the potential of AE, the purpose of the present study was to conduct a randomized prospective clinical trial of 1-year duration to primarily assess the AE-related changes in cerebral hemodynamics in aMCI, with any associated improvements in cognitive function. We aimed to assess long-term changes to CVR in aMCI patients as our primary outcome for our study. Also, we aimed to directly compare cognitive performance, measures of cardiorespiratory fitness, and MRI-derived measures of CVR across aMCI participants who completed AE-regimens for 1 year. To evaluate if the observed brain changes are specific to AE, we recruited participants with aMCI that only performed stretch training for 1 year as a group for comparison. We hypothesized that a long-term regimen of AE would yield meaningful changes in CVR and would be associated with preservation of neurocognitive function in aMCI participants at risk for further cognitive decline.

Methods

Participants characteristics

This is a prospective study registered as a clinical trial: NCT01146717, Aerobic Exercise Training in Mild Cognitive Impairment Study (AETMCI), URL: https://www.nia.nih.gov/alzheimers/clinicaltrials/aerobic-exercise-training-mild-cognitive-impairment. Recruitment was conducted between June 2010 and June 2014, and data collection was completed by September 2016. This study was approved by the Institutional Review Board of the University of Texas Southwestern Medical Center and Texas Health Presbyterian Hospital Dallas and was performed in accordance with the guidelines of the Declaration of Helsinki and Belmont Report. All participants gave informed written consent before being enrolled into the study. There are no conflicts of interest to report. Participants in this study are a subgroup of participants from the study performed by Tarumi et al. ¹⁶ and Thomas et al. ⁹

Recruitment of aMCI participants

Recruitment was conducted in the Dallas-Fort Worth metropolitan area using community-based advertisements and through the University of Texas Southwestern Medical Center Alzheimer's Disease Center. A telephone interview was first conducted to identify potential aMCI participants who (1) had memory concerns, (2) did not exercise regularly, and (3) were aged 55–80 years. These individuals were invited to visit the clinic for further screening.

Diagnosis of MCI

The diagnosis of aMCI was based on Petersen criteria, ⁴ as modified by the Alzheimer's Disease Neuroimaging Initiative project (http://adni-info.org). Specifically, aMCI participants met the following criteria: a global Clinical Dementia Rating scale of 0.5, with a score of 0.5 in the memory category, in addition to memory impairment as indicated by scores at or below 1.5 standard deviations beneath the normative reference mean on the Logical Memory (LM) subtest of the Wechsler Memory Scale-Revised, and a Mini-Mental State Exam (MMSE) score between 24 and 30.

Exclusion criteria

Participants were excluded if they had any of the following condition(s): diagnosis of AD or other type of dementia, major neurological, vascular, or psychiatric disorders including depression, history of a clinical diagnosis of B12 deficiency or hypothyroidism (stable treatment for at least 3 months is allowed), or chronic inflammatory diseases including lupus, rheumatoid arthritis, and polymyalgia rheumatica. Additionally, participation in regular exercise within the last 2 years, body mass index ≥35 kg/m², sleep disorders including clinically diagnosed or self-reported sleep apnea, uncontrolled hypertension, diabetes, and/or a history of smoking within the past 2 years were exclusion criteria. Prior to being enrolled into the study, each aMCI participant wore an accelerometer for 1 week (Actical, Philips Respironics, USA), and those who spent >90 min of moderate-to-vigorous physical activity (>4.0 metabolic equivalents, METs) per week were also excluded. For a list of all exclusion criteria, please refer to clinical trial: NCT01146717.

Study protocol

Figure 1 shows a flowchart of participants screening, attrition, and enrollment, including which patients were ultimately included into this subsample. Of the final group enrolled into the trial, each aMCI participant was assigned to one of the two study arms, AE versus stretching and toning group, using randomization and blinding procedures. Clinical, neuropsychological, cardiorespiratory, and imaging measures were collected before and after the 12-month training period. For this study, availability of CVR data for participants in either group was the main eligibility criteria to comprise this study.

Figure 1.

Flowchart of participant screening, exclusion, and enrollment numbers.

Randomization, blinding, group design

A randomized, single-blind, placebo-controlled trial design (AE versus stretching and toning groups) was used, and SAS V9.2 was used to generate stratified, randomization lists using a blocking factor of 4. Participants were stratified by age (55–70 and 71–80) and sex (men and women). Investigators conducting the analyses were blinded to treatment assignment. Participants were instructed to maintain their normal daily activities aside from the assigned interventions and were instructed not to disclose their group assignment or to discuss their interventions during outcome measurements or meetings with other participants. In both the AE and Stretch training programs, each participant was trained and supervised by a research assistant with a background in exercise physiology. AE and stretch training were performed at the convenience of the participants and not in groups. There was no social aspect to the training; however, this was not specifically controlled for. During the study period, they were asked to perform assigned intervention on top of their regular physical activities.

Aerobic exercise training regimen

The AE group was instructed to perform moderate to vigorous AE, which consisted of brisk walking, running, or cycling. The dose and intensity of the AE was based on each individual's fitness level, which was initially assessed by the peak oxygen uptake (VO2_peak) measured during a treadmill test. ¹⁷ Dose and intensity were progressively increased as participants adapted to previous workloads. Specifically, the program started with a frequency of 3 exercise sessions per week for 25–30 min per session at the intensity of 75–85% of maximal heart rate that was measured during the VO2_peak test at baseline. At week 11, participants started alternating between 3 (3 base pace) and 4 (3 base pace + 1 maximal steady state) exercise sessions per week for 30–35 min per session. Base pace is 75–85% HRmax (moderate intensity); maximal steady state is 85–90% HRmax (high intensity). For example, during the weeks in which they performed 3 exercise sessions per week, a high intensity exercise session was introduced which consisted of 30 min of walking at the intensity of 85–90% of maximal heart rate (e.g., brisk uphill walking). After week 26, participants performed 4–5 exercise sessions per week for 30–40 min, including two high intensity sessions. Each exercise session included a 5 min warm-up and a 5 min cool-down. After week 26, three 40-min base pace sessions were performed similarly with a 5 min warm-up and a 5 min cool-down. Any mode of AE was allowed if they maintained the prescribed training dose and intensity, as monitored by changes in heart rate during each of the exercise sessions. All participants were provided with a heart rate monitor (Polar RS400, Polar 201 Electro, USA). This AE program meets the national physical activity guidelines for older adults and has been used in our previous studies that showed significant improvement of cardiorespiratory fitness in sedentary individuals older than 65 years of age.^18,19

Stretching and toning regimen

A stretching program was used as active control to keep participants engaged with the same level of attention received from the investigators as those in the AE group. In the stretching group, we provided a program including light-intensity TheraBand^® (The Hygenic Company LLC., Akron, Ohio) exercises. The stretch group (ST) performed a stretch and balance routine that focused on the upper and lower body. Participants were trained to maintain their heart rate below 50% of maximal heart rate during each session. The frequency and duration of the stretch training program was the same as the AE program. At week 19, we introduced a second set of full body stretches that are more advanced than the previous set. At week 26, we introduced a set of low resistance TheraBand^® exercises that focused on strengthening the upper and lower body.

Assessment of adherence to intervention

To ensure adherence to each program, participants were required to make a training log in addition to heart rate monitoring. Each month, participants visited the clinic to download heart rate data and review their training log together with an exercise physiologist to ensure implementation of the prescribed training programs. When adherence to training programs was not met with the prescribed intensity, duration, and frequencies, in-person and/or telephone meetings were held to solve the issues and encourage participants to continue the program. The total amount of exercise performed by the AE group over 12 months was calculated by the training impulse (TRIMP) score, which is calculated by multiplying the duration of exercise session by the average heart rate achieved during each session weighted for exercise intensity. ²⁰ The compliance metric for each participant was calculated by the ratio of prescribed exercise sessions over the completed exercise sessions in which participants achieved the target heart rate.

Adverse events

Training for the first several weeks occurred at a local hospital where adverse events can be safely handled, until they could comfortably exercise by themselves at home. During the study, 10 adverse events occurred. During VO2_peak testing, 4 had arrhythmia, 1 reported foot pain, and 1 had pain in the mouth caused by wearing the mouthpiece. During AE training, 1 fell from the treadmill, 2 reported ankle pain, and 1 had knee pain. The number of participants who experienced adverse events was not different between the ST and AE groups.

Cardiorespiratory fitness assessment

Participants underwent cardiorespiratory assessments at baseline and after completion of the 12 months of AE or stretch training. Peak oxygen uptake (VO2_peak) test was performed to assess cardiorespiratory fitness. VO2_peak test measures the peak amount of oxygen a person can utilize during intense exercise. It is an index of an individual's fitness level. VO2_peak was assessed using a modified Astrand-Saltin protocol on a treadmill, ¹⁷ with all subjects reaching a plateau of VO2_peak values at the conclusion of the treadmill test.

MRI procedures

MRI experiments were performed on a 3Tesla MRI scanner using an 8-channel head coil (Philips Healthcare, Best, Netherlands). All participants were requested to refrain from consuming caffeine and alcohol for 8 h prior to the MRI scans. A body coil was used for RF transmission. Foam padding was placed around the head to minimize motion during MRI scan acquisition. The anatomic MRI protocol consisted, among other sequences, of a T1-weighted magnetization prepared rapid acquisition of gradient echo sequence (T1-MPRAGE) and a Fluid-Attenuated-Inversion-Recovery (FLAIR). ²¹ The scan parameters of the T1-MPRAGE sequence were as follows; TR/TE/TI = 8.1/3.7/1100 ms, shot interval 2100 ms, flip angle = 12 degrees, voxel size 1 × 1 × 1 mm³, number of slices 160, sagittal slice orientation and duration 3 min 57 s. The FLAIR image was acquired with TR/TE = 11,000/125 ms, TI = 2800 ms, FOV = 230 × 230 mm², 24 slices, 5 mm thick with 1 mm gap, reconstruction matrix = 512, scan duration = 3 min, 40 s.

Primary outcome: Cerebrovascular reactivity measurements

CVR response to CO2 was measured using a hypercapnia (inhalation of 5% CO2 mixed with 21% O2 and 74%N2) protocol described previously.^13,22,23 Alternating blocks of room air (1 min) and hypercapnia (1 min) was inhaled by the subject while BOLD images were acquired continuously for 7 min. Compared to long-duration CO2 paradigm, this 1-min paradigm has been shown to improve subject comfort while maintaining the quality of the data. ²³ Another advantage of the short block design is that potential effect of MRI signal drift could be reduced. The imaging parameters were: TR/TE = 1500 ms/30 ms, flip angle = 60°, voxel size = 3.0 × 3.0 × 5.0 mm³, FOV = 240 × 240 mm², 29 slices, and scan duration = 7 min. The air/gas mixture was delivered using a Douglas bag and switching between room air and CO2 was achieved via a valve on the bag. The air/gas mixture was delivered to the subject via a mouth piece and a nose clip was used to stop nasal breathing. End-tidal CO2 (EtCO2) is the partial pressure of CO2 at the end of an exhaled breath and was measured using a capnography monitor (Capnograd, model 1265, Novametrix Medical Systems, Wallingford, CT) during the experiment; EtCO2 is an indicator of the vasodilatory input to the brain. Other physiologic parameters including breath rate, heart rate and arterial oxygenation were also measured during the experiment by a physiology monitor (MEDRAD Inc., Pittsburgh, PA).

CVR data were analyzed using the Johns Hopkins University CVR-MRICloud. ²⁴ Briefly, voxel-wise whole-brain CVR maps were obtained using a general linear model (GLM) between whole-brain averaged BOLD signal and global-shifted EtCO2. Additionally, voxel-wise CVR maps were transformed into the individual's MPRAGE space and the Montreal Neurologic Institute (MNI) standard space. Regional CVR averages were also calculated. T1-segmentation on the MRICloud was used to parcellate the brain into regions of interest (ROI). Subsequently, using these voxel-wise maps, mean CVR was obtained from the following ROIs: cortical lobes (frontal, parietal, etc.), middle frontal gyrus, dorso-lateral pre-frontal cortex, anterior cingulate cortex (ACC), subcallosal ACC, subgenual ACC, rostral ACC, and posterior cingulate cortex. The choice of these ROIs were based on our previous findings highlighting changes in CBF across various ROIs in aMCI subjects after a regimen of AE, including anterior and posterior cingulate cortices. ⁸ Relative CVR maps were obtained by dividing each voxel by the average value from the Cerebellum ROI to normalize between subject differences. Use of relative values has been shown to be more sensitive in detecting regional differences, often superseding whole-brain absolute values differences. ²⁵

Table 2.

Changes in relative cerebrovascular reactivity (CVR) before and after intervention for stretch and exercise groups, normalized to mean cerebellar CVR, BOLD/BOLD ratio, (mean ± SD).

Regions of Interest	Stretch			Exercise			p Interaction effect
	Pre	Post	p	Pre	Post	p
Middle Frontal Gyrus (MFG)	0.89 ± 0.2	0.87 ± 0.1	0.624	0.87 ± 0.1	0.80 ± 0.2	0.021	0.304
Dorsolateral pre-frontal cortex	1.00 ± 0.1	0.99 ± 0.1	0.747	0.97 ± 0.1	0.87 ± 0.1	0.042	0.172
Posterior Cingulate Cortex (PCC)	0.93 ± 0.2	0.92 ± 0.1	0.773	0.96 ± 0.2	0.95 ± 0.2	0.854	0.980
Anterior Cingulate Cortex (ACC)	0.82 ± 0.1	0.83 ± 0.1	0.761	0.87 ± 0.1	0.78 ± 0.1	0.038	0.042
Subgenual ACC	0.81 ± 0.2	0.8 ± 0.2	0.816	0.98 ± 0.3	0.84 ± 0.3	0.047	0.089
Subcallosal ACC	0.28 ± 0.6	0.34 ± 0.9	0.748	0.61 ± 0.5	0.17 ± 0.8	0.040	0.093
Rostral ACC	0.84 ± 0.3	0.77 ± 0.3	0.297	0.89 ± 0.2	0.74 ± 0.2	0.059	0.365
Regional Averages
Frontal Lobe	0.88 ± 0.1	0.88 ± 0.1	0.949	0.94 ± 0.2	0.83 ± 0.2	0.024	0.043
Parietal Lobe	0.91 ± 0.1	0.94 ± 0.1	0.494	0.98 ± 0.2	0.90 ± 0.2	0.144	0.084
Temporal Lobe	0.99 ± 0.1	1.03 ± 0.1	0.195	1.01 ± 0.1	0.93 ± 0.2	0.135	0.034
Limbic System	0.84 ± 0.1	0.86 ± 0.1	0.556	0.87 ± 0.1	0.82 ± 0.1	0.252	0.166
Occipital Lobe	1.02 ± 0.2	1.05 ± 0.1	0.309	1.06 ± 0.2	0.97 ± 0.2	0.104	0.039
Hippocampus	0.7 ± 0.2	0.73 ± 0.2	0.238	0.71 ± 0.2	0.69 ± 0.1	0.658	0.330
Whole Brain	0.69 ± 0.1	0.71 ± 0.1	0.282	0.72 ± 0.1	0.68 ± 0.1	0.22	0.082

p-value: Two tailed paired T-test comparison of Pre and Post intervention data, significance: p < 0.05.

Interaction effect: Repeated measures ANOVA of intervention group between pre and post measurements, significance: p < 0.05.

p-values in bold font are significant at p < 0.05.

Other MRI measurements

The T1-weighted high-resolution image was used for the assessment of brain volumes. Volumetric measurements of the following ROIs before and after 1 year were obtained: cortical lobes (frontal, parietal, etc.), middle frontal gyrus, dorso-lateral pre-frontal cortex, anterior cingulate cortex (ACC), subcallosal ACC, subgenual ACC, rostral ACC, and posterior cingulate cortex. Quantification of white matter hyperintensity (WMH) volume used the FLAIR image, with a procedure described in detail elsewhere. ²⁶ To account for individual differences in head size, total volume of WMH was normalized to the intracranial volume. ²⁷

Carotid artery stiffness measured with ultrasound

Description of ultrasound methods used to measure carotid artery stiffness, are provided in detail in our group's prior report. ²⁸

Statistical analysis

Univariate analysis of pre- and post-intervention data was conducted using a paired t-test and between-group comparison was performed using a two-sample t-test. Jamovi 2.2.5, an open source statistical software, was used to conduct all statistical analysis. ²⁹ Repeated-measures ANOVA between time points and group classification was performed. A p-value of 0.05 or less was considered significant. Voxel-wise CVR data were analyzed in SPM12 and considered significant at a Family-Wise-Error (FWE) corrected threshold of p < 0.05, corrected for multiple comparisons.

Results

Characteristics of aMCI participants at baseline

From 1620 participants initially screened, 27 aMCI participants with complete CVR data were included in the analysis, with 16 in the ST group and 11 in the AE group, with participants excluded as detailed in Figure 1. Baseline demographic and clinical characteristics are displayed in Table 1.

Table 1.

Baseline (pre-training) characteristics of amnestic mild cognitive impairment participants in the exercise and stretch groups (mean ± SD or N%).

		Total (n = 27)	Stretch (n = 16)	Exercise (n = 11)	p
Demographics	Age (y)	66.4 ± 7.0	66.4 ± 6.9	66.2 ± 7.1	0.94
	Sex (Male/Female)	15 (55.6%) / 12 (44.4%)	9 (56.3%) / 7 (43.7%)	6 (54.5%) / 5(45.5%)	-
	Education (y)	16.3 ± 1.9	15.9 ± 2.0	16.8 ± 1.8	0.24
	Height (cm)	169.4 ± 10.4	170.7 ± 10.3	167.6 ± 10.2	0.47
	Weight (kg)	81.0 ± 15.3	82.9 ± 14.4	78.2 ± 16.2	0.45
	BMI (kg/m2)	28.1 ± 4.2	28.4 ± 4.2	27.7 ± 4.1	0.68
	Compliance with Aerobic Exercise Level (%)	-	-	79.0%	-
Clinical	MMSE	29.0 ± 1.4	28.9 ± 1.6	29.3 ± 1.0	0.49
	CDR	0.5 ± 0.05	0.5 ± 0.0	0.5 ± 0.0	>0.99
	Logical Memory Delayed Recall	8.3 ± 2.1	8.4 ± 1.5	8.3 ± 2.7	0.90
	Taking Anti-Hypertensive Medication	11 (40.7%)	7 (43.8%)	4 (36.4%)	-
	Taking Statin Medication	6 (22.2%)	4 (25%)	2 (18.2%)	-
Vascular	Systolic Blood Pressure (mm Hg)	122 ± 10.0	123 ± 10.9	121 ± 8.9	0.67
	Diastolic Blood Pressure (mm Hg)	71.7 ± 8.0	71.8 ± 8.2	72 ± 8.2	0.88
	White Matter Hyper-intensity Volume (mL)	2.9 ± 2.6	2.7 ± 2.1	3.1 ± 3.1	0.68

p: Two tailed paired T-test comparison of Pre and Post intervention data, significance: p < 0.05; BMI: body mass index; MMSE: Mini-Mental State Exam; CDR: Clinical Dementia Rating.

Baseline global cognitive function, as measured by the MMSE, had a mean score of 29.0 (SD: 1.4), and a mean Clinical Dementia Rating value of 0.5 (SD: 0.05). The brachial systolic blood pressure (SBP) and diastolic blood pressure (DBP) had mean values of 122 mmHg (SD: 8.2) and 71.7 mmHg (SD: 7.3), respectively, with 11 participants (40.7%) taking anti-hypertensive medication at baseline, and six participants (22.2%) taking an anti-cholesterol medication. The mean number of total WMH volume was 2.9 mL (SD: 2.6). Based on duration of exercise, mean compliance to the AE program was 79.0%. The TRIMP scores exhibited significant heterogeneity between participants and the average TRIMP score in the AE group was 11,985 (Supplemental Figure 1). Across measured demographic, clinical, and vascular parameters, baseline differences between stretch and AE groups were not statistically significant (p > 0.05).

Effect of intervention on cardiovascular function

Both VO2_peak and carotid stiffness were found to be similar (p = 0.400) in the AE and stretch groups before training. However, after one year, the AE group showed a significant increase in VO2_peak (p = 0.034), while the stretch group did not show a significant difference (p = 0.416) (Figure 2A), with repeated measures ANOVA showing a group-by-time interaction effect (p = 0.008, F = 31.32). After one year, the exercise group also had a significant decrease in carotid stiffness when comparing post-training to pre-training (p = 0.005), while the stretch group did not show a difference (p = 0.879) (Figure 2B). Repeated measures ANOVA revealed a group-by-time interaction effect (p = 0.033, F = 5.10) on carotid stiffness. When examining the magnitude of these associations, larger decreases in carotid stiffness were also associated with larger increases in VO2_peak (R²= 0.161, p = 0.043), as shown in Figure 2C.

Figure 2.

(A) Changes in VO2_peak in the exercise and stretch group. (B) Changes in carotid artery stiffness in the exercise and stretch group. (C) Increase in VO2_peak (post-pre VO2_peak) is associated with decrease in carotid stiffness (post-pre carotid stiffness).

Effects of intervention on primary outcome: Cerebrovascular reactivity

After one year AE intervention, the AE group had significant decreases in relative CVR in regions of the brain such as the frontal lobe and anterior cingulate cortex (ACC) (see Figure 3). Differences in CVR were not observed in the ST group after one-year ST intervention (FWE corrected p > 0.05). When isolating several regions of interest, interaction effects of CVR with respect to group and time point in the superior, middle, and frontal gyri, and anterior cingulate cortex and corresponding sub-regions, along with the frontal lobe, temporal lobe, and occipital lobe were significantly decreased after exercise (p = 0.047) while the stretch group did not change (see Table 2). Changes in CVR were not associated with changes in VO2_peak, cognition or carotid stiffness in either group (p > 0.05).

Figure 3.

Voxel-wise reductions in relative cerebrovascular reactivity (CVR) in aMCI participants after performing 1 year of aerobic exercise compared to baseline are shown in blue color, significant at a family wise error (FWE) corrected threshold of p < 0.05.

Effect of intervention on secondary outcome: Cognitive function

Participants had similar scores on the MMSE, Clinical Dementia Ratings, and overall cognitive test scores at baseline (p > 0.05). Broadly, participants in the exercise group showed slight but statistically significant improvements on some of their cognitive test scores over time (Table 3). For the Delis-Kaplan Executive Function System (DKEFS) test, AE participants’ scores on the inhibition sub-score increased (p = 0.005), as did scores on Logical Memory delayed recall (p = 0.028). In addition, slight but statistically significant decreases in reported problems on the Functional Assessment Questionnaire (p = 0.049) and psychiatric symptoms on the Neuropsychiatric Inventory Questionnaire (p = 0.011) were seen, even though these scores did not reflect clinically meaningful levels of impairment or distress. In contrast, participants in the stretch program did not show any statistically significant improvements on their cognitive tests (p > 0.05). In our analysis using repeated-measure ANOVA, the DKEFS category fluency test (F = 4.05, p = 0.046) had a significant time-interaction effect between exercise and control groups. Across both groups, participants who improved their VO2_peak also showed improvements in DKEFS Category Fluency (p = 0.021, R²= 0.241 and Category Switching (p = 0.018, R²= 0.382), as seen in Figures 4A and 4B, respectively.

Table 3.

Cognitive results before and after intervention for stretch and exercise groups (mean ± SD).

Cognitive Test	Stretch			Exercise			p: Interaction Effect
	Pre	Post	p	Pre	Post	p
Delis-Kaplan Executive Function System (DKEFS)
Word Interference test—Inhibition—SS	11.1 ± 2.1	12.0 ± 1.9	0.208	10.6 ± 1.8	11.5 ± 1.7	0.005	0.800
Trail Making test—Motor Speed—SS	11.7 ± 1.3	11.9 ± 1.0	0.653	10.5 ± 20.	11.8 ± 1.3	0.051	0.067
Verbal Fluency test—Category Fluency—SS	11.7 ± 3.7	11.6 ± 3.6	0.906	10.7 ± 3.1	12.8 ± 3.6	0.086	0.046
Verbal Fluency test—Category Switching—SS	12.7 ± 2.3	12.0 ± 2.8	0.308	10.8 ± 4.8	11.5 ± 4.0	0.624	0.327
Neuropsychiatric Inventory Questionnaire (NPI)—RS	3.2 ± 2.6	2.6 ± 4.7	0.545	4.6 ± 2.2	2.6 ± 2.0	0.011	0.339
Functional Assessment Questionnaire (FAQ)—RS	2.7 ± 2.6	1.7 ± 2.2	0.681	4.0 ± 4.1	1.8 ± 3.7	0.049	0.143
Logical Memory Immediate Recall—RS	11.1 ± 1.9	11.6 ± 3.3	0.629	11.4 ± 2.3	12.1 ± 3.0	0.351	0.749
Logical Memory Delayed Recall—RS	8.8 ± 2.2	11.1 ± 3.9	0.068	8.8 ± 3.3	12.9 ± 2.7	0.028	0.597
California Verbal Learning Test (CVLT)—TS	44.7 ± 13.9	45.4 ± 10.3	0.722	46.2 ± 9.7	51.7 ± 10.1	0.183	0.227
Boston Naming Test—RS	28.1 ± 1.9	27.9 ± 1.8	0.688	28.2 ± 0.9	27.9 ± 2.0	0.661	0.910
Clock Drawing Test—RS	12.6 ± 1.4	12.2 ± 2.1	0.615	12.4 ± 1.3	13.7 ± 0.7	0.111	0.275

p: Two tailed paired T-test comparison of Pre and Post intervention data, significance: p < 0.05.

Interaction effect: Repeated measures ANOVA of intervention group between pre and post measurements, significance: p < 0.05.

SS: Scaled Score; RS: Raw Score; TS: T-score.

p-values in bold font are significant at p < 0.05.

Figure 4.

(A) Increased VO2_peak (post-pre VO2_peak) is associated with improvements in Delis-Kaplan Executive Function System category fluency test (post-pre DKEFS category fluency). (B) Increased VO2_peak (post-pre VO2_peak) is associated with improvements in DKEFS category switching test (post-pre DKEFS category switching).

Effect of 12 months of intervention on brain volume

To investigate whether physiological changes in CVR were accompanied by anatomic changes, we compared brain volume of above-mentioned regions before and after interventions. No differences in brain volume in any ROIs studied were found between the two groups (p > 0.05 for all comparisons).

Summary

In summary, we provide a comprehensive analysis of the relationship of cardiorespiratory health, cognitive function, and neurovascular function in the setting of a 1-year long AE regimen in participants with aMCI. We observed that our exercise participants had slight improvements in several objective measures of cognition, mostly those related to executive function, compared to control subjects who were largely unchanged after 1 year. Furthermore, changes in cardiovascular parameters such as VO2_peak were correlated with decreases in carotid artery stiffness. Finally, we demonstrate significant decreases in MRI-based CVR in the exercise group, especially in the region of the anterior cingulate cortex and the frontal lobe.

Discussion

Despite the approval of new biologics for anti-amyloid therapy in the treatment of AD and related dementias, the clinical significance of these drugs is still limited. As a result, the role of AE in mitigating the progression of MCI to AD as an alternative or adjuvant therapy is a recent topic of interest. Aging adults are known to have changes in cerebrovasculature, which may play a role in amyloid burden and disease progression over time. ³⁰ To our knowledge, this is the first comprehensive prospective study that directly compares MRI- and ultrasound-based hemodynamic parameters, cognitive outcome findings, and cardiovascular health of participants with MCI undergoing AE over a duration of 1 year.

Here we report slight improvements in various cognitive tests related to executive function after AE in MCI participants over 1 year. Although these findings are statistically significant, improvements in scores are not clinically significant. This result is similar to the findings of the other 1-year-duration exercise study by Tomoto et al., which reported similar magnitude of cognitive differences between the exercise and stretch groups. ³¹ Tarumi et al. also found slight improvements in cognitive scores after 1 year, along with reduced hippocampal atrophy. ¹⁶ Despite these preliminary findings, overall outcomes are not well described on the benefit of exercise in MCI.^32,33 One notable finding in our study was that improvements in the DKEFS verbal fluency tests were associated with improved VO2_peak, and verbal fluency improvements have been reported after AE in previous MCI studies,^34,35 which have implications for the role of frontal brain regions in response to AE.

Regarding our finding that CO2-induced CVR decreases over 1 year of AE, broadly, there is no consensus about the influence of AE and CVR in older or MCI participants. For example, one study by Barnes et al. showed older participants at higher aerobic fitness had higher MCA velocity CVR. ³⁶ Another study by Guadagni et al. examining the effect of AE regimens on healthy older participants found a decrease in cerebrovascular resistance during acute exercise after 6 months. ³⁵ Other studies endorse a lack of significant change in CO2 reactivity after exercise. Zhu et al. found that athletic older adults did not have significant differences in CO2 reactivity compared to sedentary adults, and Nowak-Flück et al. had similar findings during acute exercise.^37,38 However, it is difficult to generalize these findings, as across these studies CVR was measured at different times (e.g., after acute exercise versus at rest), or did not compare changes before and after exercise intervention over a long-term period. Furthermore, none of these studies examined these changes across MCI participants with and without AE.

The only similar CVR study to our knowledge is by Tomoto et al. conducted over 1 year on 70 individuals with aMCI, which also found that AE decreased hypercapnic CVR as measured by transcranial Doppler. ³¹ Our observations were similar utilizing MRI- based CVR measures and extend the line of thinking in our previous work, which found a negative dose-response relationship with VO2_peak and CVR in high-fitness older athletes. ¹³ The relationship between AE, cardiorespiratory fitness, and cerebral hemodynamics are not yet well-established, as noted in the meta-analysis by Smith et al. ³⁹ One possible explanation for variable results across studies may be due to a non-linear relationship of CVR and exercise. Dubose et al. show a peaking of CVR at a VO2_peak of 28 mL/kg/min, after which CVR declined. ⁴⁰ In our exercise group, the mean VO2_peak was 24.6 mL/kg/min with a standard deviation of 6.0, ranging up to 35.6 mL/kg/min. Although our group did not significantly surpass Dubose et al.'s threshold, the much longer-term duration of our study (12 versus 3 months) may increase the influence of non-linear behavior in decreasing CVR.

The current study also observed that decreases in carotid artery stiffness were associated with increases in VO2_peak, with decreases in stiffness isolated to the AE group. This result falls in line with prior studies which also highlight that regular exercise can slow or reverse age-related increases in arterial stiffness.^41,42 Although our study did not investigate this, one possible explanation for both decreased arterial stiffness and decreased CVR after long-term exercise is that AE may change the levels of certain molecules in the CO2-mediator vasodilation pathway, such as adenosine receptor or nitric oxide synthase expression levels. A recent report examined the use of adenosine receptor antagonists (e.g., caffeine) to explain reductions in CVR, caused by a reduction of the adenosine receptor in cerebral vessels. ⁴³ These pathways have also been implicated in causing decreases in arterial stiffness due to exercise. ⁴⁴ These physiological changes may also support the idea of desensitization to hypercapnia through long-term exercise, a phenomenon present in other parts of the body, e.g., skeletal muscle. ⁴⁵ These changes may be independent of other cardiovascular adaptations to exercise. For example, in our study, both systolic and diastolic blood pressure did not change significantly before and after AE. Future study should examine how changes in expression of these molecules affect cerebral autoregulation in the setting of long-term AE for aMCI.

Our study found the main region of CVR decrease to be at the ACC, which is commonly implicated in many disease states and is important for various neurocognitive functions including episodic memory and has been shown to be influenced by exercise. For instance, some studies show a link between exercise and preserving cortical volume in the ACC.^46–49 However, findings on the hemodynamic effects in the ACC after AE have not reached a consensus. Our previous study examining the effect of 1 year duration of AE on MCI participants reported that exercise helps maintain CBF in the ACC, whereas the ACC CBF decreased in the MCI group that did not perform exercise. We also previously reported that voxel-wise positive correlations between CBF increases in the ACC and neighboring frontal regions were associated with improvement in long term delayed recall memory function. ⁹ In line with this previous study, our present study extends this finding of the effects of long-term exercise to a decrease in CVR in the ACC. The systematic review by Kleinloog et al. found 16 MRI studies that utilized exercise as an intervention with median duration of 14 weeks found uniform increases in CBF in the ACC. ⁵⁰ However, to our knowledge, there are no studies that directly compare CBF to CVR changes in the long-term. Our study adds to the literature highlighting the importance of the ACC in the long-term progression of cognitive decline. Future prospective work is needed to delineate how decreases in CVR within this region affect specific cognitive functions.

Strengths and limitations

Overall, one major strength of this study comes from the long-term duration. Effects of long-term exercise for a duration of 1 full year on aMCI patients are largely unknown, and a comparison of cognition to MRI measures adds further context to the host of changes that occur in the brains of participants with aMCI. The long-term effects of exercise on MRI-derived measures of CVR were also wholly unexplored and our analyses add greater breadth to the understanding of cerebrovascular function in cognitive impairment. Furthermore, this prospective clinical trial included an appropriately matched control group that was engaged in an active stretching regimen, which we believe improves the sensitivity of any differences observed between the groups.

Our study findings are also tempered by several limitations. Most notably, it comprised a single-center prospective trial with a limited sample size of 27 participants in total, and only 11 that underwent exercise. Along with reductions in generalizability, the limited sample size mutes the study's ability to find relevant subgroup differences, e.g., identifying differences by sex. Second, the changes in cognitive scores that were seen in the AE group were small and not clinically significant, even though some differences reached statistical significance. This is, however, in somewhat of a contrast to the findings in the ST group, which showed smaller or no changes over time and were universally non-significant. Third, we conducted several univariate comparisons for each ROI, which may introduce potential spurious findings due to multiple comparisons that are not corrected for, although the number of these comparisons were limited to a prior knowledge of ROIs associated with changes due to AE. Lastly, we did not find significant associations between changes in CVR and cognitive changes, and thus we are cautious extending conclusions regarding cognition and cerebrovascular dysfunction in aMCI beyond the scope of our study.

Conclusions

We conducted a prospective clinical trial assessing the effects of a 1-year long AE regimen in individuals with aMCI. Participants undergoing exercise showed some improvements in aspects of cognition, in addition to increased cardiorespiratory fitness, decreases in arterial stiffness and decreased CVR, while participants in the stretch group did not exhibit these changes. This study supports the idea that sustained AE alters cerebral hemodynamics and may help slow deterioration of cognitive function in aMCI.