|Home | About | Journals | Submit | Contact Us | Français|
Examine the correlation of cardiorespiratory fitness with brain atrophy and cognition in early-stage Alzheimer’s disease (AD).
In normal aging physical fitness appears to mitigate functional and structural age-related brain changes. Whether this is observed in AD is not known.
Nondemented (n=64) and early-stage AD subjects (n=57) had MRI and standard clinical and psychometric evaluations. Peak oxygen consumption (VO2peak), the standard measure of cardiorespiratory fitness, was assessed during a graded treadmill test. Normalized whole brain volume, a brain atrophy estimate, was determined by MRI. Pearson’s correlation and linear regression were used to assess fitness in relation to brain volume and cognitive performance.
Cardiorespiratory fitness (VO2peak) was modestly reduced in AD (34.7(5.0)ml/kg/min) vs. nondemented subjects (38.1(6.3)ml/kg/min, p=0.002). In early AD, VO2peak was associated with whole brain volume (beta=0.35, p=0.02) and white matter volume (beta=0.35, p=0.04) after controlling for age. Controlling for additional covariates of sex, dementia severity, physical activity, and physical frailty did not attenuate the relationships. VO2peak was associated with performance on delayed memory and digit symbol in early AD but not after controlling for age. In nondemented participants, there was no relationship between fitness and brain atrophy. Fitness in nondemented participants was associated with better global cognitive performance (r=0.30, p=0.02) and performance on Trailmaking A and B, Stroop, and delayed logical memory but not after controlling for age.
Increased cardiorespiratory fitness is associated with reduced brain atrophy in AD. Cardiorespiratory fitness may moderate AD-related brain atrophy or a common underlying AD-related process may impact both brain atrophy and cardiorespiratory fitness.
Studies in cognitively normal older adults suggest that maintaining physical fitness may oppose functional and structural changes of the brain that occur with normal aging.1, 2 Less is known about the role of exercise and fitness in pathological brain states, such as Alzheimer’s disease (AD). A recent meta-analysis of studies utilizing various exercise techniques found evidence that exercise provides some benefits to individuals with dementia3 although studies assessing exercise in AD have produced mixed results.4-8 Animal study data suggest that physical activity can decrease the neuropathological burden in a mouse model of AD.9 Whether exercise and fitness modify the disease process underlying AD in humans remains unknown.
Studies on physical activity and exercise often rely on self-reported activity measures and are thus limited by a lack of standard and objective definitions of physical activity.10 Maximal treadmill testing to determine peak oxygen consumption (VO2peak) is the standard measure of cardiorespiratory fitness. Cardiorespiratory fitness is closely related to physical activity level11 and is stronger than self-reported physical activity as a predictor of many health outcomes.12 At least one study has demonstrated that healthy older adults with higher cardiorespiratory fitness have lower rates of cognitive decline over 6 years than individuals with lower cardiorespiratory fitness.13 There is little data on cardiorespiratory fitness in AD, particularly in the early stages of the disease. Thus, we performed a cross-sectional study examining cardiorespiratory fitness in relation to brain structure and cognition in individuals in the earliest clinical stages of AD and nondemented controls. We hypothesized that increased cardiorespiratory fitness would be associated with less brain atrophy and higher cognitive performance in early AD.
Nondemented (Clinical Dementia Rating (CDR) 0, n = 64) and early-stage AD subjects (CDR 0.5 and 1, n = 57) aged 60 and over were enrolled in the University of Kansas Brain Aging Project. The CDR 0.5 (n = 45) and CDR 1 (n = 12) participants together represent early-stage AD. Study exclusions include neurologic disease other than AD, diabetes mellitus (defined as a clinical diagnosis and use of an anti-diabetic agent), history of ischemic heart disease (acute coronary artery event, angina), schizophrenia, clinically significant depressive symptoms, abnormalities in B12, RPR, or thyroid function, use of antipsychotic and investigational medications, and significant visual or auditory impairment, systemic illness, or orthopedic issues that could impair completion of the study.
The clinical assessment included a semi-structured interview with the participant and with a collateral source knowledgeable about the participant. Diagnostic criteria for AD require the gradual onset and progression of impairment in memory and in at least one other cognitive and functional domain.14 The presence or absence of dementia, and its severity if present, was determined using the CDR.15 These methods have a diagnostic accuracy for AD of 93%,16 are sensitive to detecting the earliest stages of AD by focusing on intraindividual change rather than comparison with group norms17 and are accurate in identifying the subset of individuals meeting criteria for MCI who have early stage AD.18 Only CDR 0, 0.5, and 1 subjects were enrolled into this study. The CDR sum of boxes was used as a measure of dementia severity.
Medications, past medical history, education, and demographic information were collected from the collateral source by the nurse clinician. A standard physical and neurological examination was performed to assess abnormalities in visual fields, cranial nerves, motor strength, sensation, reflexes, plantar responses, coordination, praxis, and gait.
A trained psychometrician administered a psychometric battery including standard measures of memory (WMS-R Logical Memory I and II19, Free and Cued Selective Reminding Task20), language (Boston Naming Test - 15 item21), working memory (WMS III Digit Span Forwards and Backwards19, WAIS Letter-number sequencing22), executive function (Trailmaking A and B,23 Verbal Fluency (animals, fruits and vegetables), and Stroop Color-Word Test24), and visuospatial ability (WAIS Block Design22). The mini-mental status examinantion (MMSE)25 was administered as a measure of global cognition. Additionally, all cognitive performance scores were converted to z-scores (with higher scores representing better performance) based on the mean and standard deviation of all nondemented and early AD subjects. The mean of each participant’s performance z-scores was determined to create an index of global cognitive performance.
VO2peak, the standard measure of cardiorespiratory fitness, was measured during a symptom-limited graded treadmill test, as previously described.26 Subjects were attached to a 12-lead electrocardiograph to monitor cardiac stability and wore a non-rebreathing facemask. Subjects began walking at a pace of 1.7 miles per hour at 0% grade. Every two minutes the speed or grade was increased according to a protocol designed for a geriatric population.27 Blood pressure and the participant’s self-assessment of exertion using the 15-point Borg Rating of Perceived Exertion scale28 were acquired during the last 30 seconds of each 2-minute stage. Heart rates were recorded every 15 seconds. Expired air was measured for oxygen and carbon dioxide at 15-second intervals using a Parvomedics system. Gas calibrations were performed on the metabolic cart before each test according to the specifications of the manufacturer.
VO2peak was considered the highest observed value during the test.29 Participant effort during the treadmill test was assessed with measures of the respiratory exchange ratio (value ≥1.0 considered near-maximal effort30) and 17 individuals were excluded for not attaining an RER≥1.0 (six nondemented vs. 11 early AD subjects, p=0.17). Subjects excluded for an RER<1.0 were predominantly women (14 women and three men) and tended to be older than those achieving an RER≥1.0 (77.5 (7.1) vs. 73.5 (6.5) years, p=0.04). There was no difference in global cognition, MMSE, and dementia severity in those excluded vs. those achieving an RER≥1.0.
The primary fitness measure was VO2peak normalized by lean mass determined by DEXA scanning (Lunar Corporation). This measure was chosen as our primary fitness measure for several reasons. Normalizing VO2peak for lean mass is more physiologic than indexing for body weight29 given that 95% of oxygen consumption during exercise occurs in exercising muscles (i.e., lean mass). Additionally, VO2peak normalized by lean mass minimizes gender differences in VO2peak. For instance, gender differences were not evident in our overall cohort (n=121) for VO2peak normalized by lean mass (37.1 (6.8) in men vs. 36.1 (5.1) ml/kg/min in women, p=0.37) while robust gender differences are present when normalizing VO2peak for total body weight (24.4 (5.2) in men vs. 19.9 (3.6) ml/kg/min in women, p<0.0001).
The participant’s level of habitual physical activity was estimated using the Physical Activity Scale in the Elderly (PASE). The PASE is a reliable and valid measure of physical activity and physical function developed specifically for older individuals.31 The PASE assesses an individual’s level of physical activity within the last seven days as an estimate of habitual physical activity. We modified the PASE by administering it to the subject’s study partner for both nondemented and AD subjects. The Physical Performance Test is a short battery of timed physical tasks serving as a composite measure of physical performance and used to assess physical frailty. The modified Physical Performance Test32 includes 1) writing a sentence, 2) simulated eating, 3) lifting a book and placing it on a shelf above shoulder height, 4) putting on and removing a jacket, 5) picking up a penny from the floor, 5) turning 360 degrees and 7) walking 50 feet, 8) chair rises and 9) a progressive Romberg test.
Structural MRI was obtained using a Siemens 3.0 Tesla Allegra MRI scanner at the Hoglund Brain Imaging Center. High-resolution T1 weighted anatomic images were acquired to provide detailed gross anatomy with high gray-white matter contrast (MP-RAGE; 1x1x1mm voxels; TR=2500, TE=4.38, TI=1100, FOV 256X256 with 18% oversample, 1mm slice thickness, flip angle=8 degrees). Normalized Whole Brain Volume was computed for each image session using a validated set of imaging tools from FMRIB Software Library (FSL; www.fmrib.ox.ac.uk/fsl). The images were pre-processed and skull-stripped using Brain Extraction Tool (BET). The skull stripped images were then segmented into white matter, gray matter, and cerebrospinal fluid using FMRIB’s Automated Segmentation Tool (FAST) by registering them to a Montreal Neuroimaging Institute avg152 template. Image processing was conducted utilizing the Laboratory of Neuroimaging Pipeline (University of California Los Angeles, www.pipeline.loni.ucla.edu). Normalized volumes for white matter, gray matter, and whole brain (sum of white and gray matter) were calculated by dividing each by the total intracranial volume (the sum of white, gray, and cerebrospinal fluid volumes) and expressed as the percent of total intracranial volume. Normalized brain volumes minimize gender differences and produce an estimate of brain atrophy. As expected, normalized whole brain volumes were not related to gender (77.1% (3.3) in women vs. 76.9% (3.4) in men, p=0.76), lending validity to the normalization procedure.
Analyses were conducted using SPSS 15.0. Continuous variables were summarized by means and standard deviations while categorical variables were summarized by frequency and percent. Student’s T-test was used to compare continuous demographic and imaging variables in early AD and nondemented groups. A chi square test was used to compare categorical variables. Pearson’s correlation coefficients were calculated to assess simple relationships between variables. Linear regression was used to assess the relationships between variables while controlling for covariates. A univariate general linear model was used to examine for the presence of group × VO2peak in predicting brain volume. To account for the apparent differential relationship between fitness and brain volume across groups (group × VO2peak interaction, see results), we conducted all analyses within diagnostic groups (early AD and nondemented subjects separately) as opposed to the entire cohort combined. To minimize the confounding influence of age-related changes in brain volume we controlled all analyses for age.
Table 1 summarizes the characteristics of the nondemented and AD groups. The mean age of the cohort (n = 121) was 73.5 (6.5) years with no difference between nondemented and early-stage AD groups. Individuals in the early AD group had mild global cognitive dysfunction with a mean MMSE score of 26.2. Whole brain atrophy was apparent in the early AD group, with whole brain differences primarily related to group differences in gray matter volume, consistent with studies suggesting gray matter is more vulnerable to AD than white matter.33, 34 Whole brain volume was associated with global cognitive performance (r=0.53, p<0.001) and MMSE performance (r=0.33, p<0.001).
Modest group differences in cardiorespiratory fitness were observed between AD (34.7 (5.0) ml/kg/min) and nondemented groups (38.1 (6.3) ml/kg/min, p=0.002). Group means of peak heart rate and RER were consistent with maximal effort (i.e., 90% of predicted maximum heart rate (220 - age) and RER≥1.1 for both groups). The duration of exercise testing to achieve VO2peak was lower in the early AD group than nondemented participants.
The role of dementia severity on treadmill testing was examined by assessing the correlation between dementia severity (CDR box score) and measures of effort and VO2peak in the early AD group. Dementia severity was not related to VO2peak (r=-0.09, p=0.49), RER (r=-0.01, p=0.94), peak heart rate (r=0.14, p=0.30), exercise duration (r=-0.16, p=0.25), and rating of perceived exertion (r=-0.12, p=0.37).
Simple correlations were first examined between VO2peak, brain atrophy, and cognition (table 2). Given the presence of a group × VO2peak interaction in predicting whole brain volume (p=0.004), all analyses were performed within diagnostic groups (i.e., nondemented and early AD) rather than the overall group combined. VO2peak was associated with whole brain volume in early AD (r=0.54, p<0.001; figure 1) but not nondemented controls (r=0.18, p=0.17, figure 2). In early AD, fitness was associated with both white matter volume (r=0.39, p = 0.003) and gray matter volume (r=0.36, p=0.007) with higher fitness levels associated with increased brain volume.
Next, these relationships were assessed when controlling for the effect of age (table 1) and other covariates. In early AD, controlling for age did not attenuate the relationship of fitness with whole brain volume (beta=0.35, p=0.02) and white matter volume (beta=0.35, p=0.04). Additionally, controlling for age and individual covariates of gender, dementia severity (CDR sum of boxes), physical activity level, and a measure of physical frailty (physical performance test) did not alter the relationship of fitness with whole brain volume and white matter volume. An overall model including all covariates did not alter the relationship between fitness and whole brain volume (beta=0.30, p=0.03) and partially attenuated the relationship of fitness and white matter volume (beta=0.27, p=0.096). After controlling for age, there was no relationship between fitness and gray matter volume in the early AD participants. In nondemented participants, there was no relationship between fitness and whole brain, white matter and gray matter volumes after controlling for age.
Simple correlations between VO2peak and overall cognitive performance were assessed first. In early AD, VO2peak was not associated with global cognitive performance or performance on the MMSE. On specific cognitive tasks, VO2peak was associated with performance on delayed memory (logical memory II, r=0.30, p=0.02) and digit symbol (r=0.37, p=0.05) in early AD although these associations were not present when controlling for age. In nondemented controls, VO2peak was associated with global cognitive performance (r=0.30, p=0.02) with increasing fitness associated with higher global cognitive performance. On specific cognitive tasks, VO2peak in nondemented participants was most strongly associated with timed tasks of Trailmaking A (r=-0.32, p=0.01) and Trailmaking B (r=-0.25, p=0.005). Additionally, fitness was associated with performance on the Stroop interference task (r=0.30, p=0.02), a measure reflecting executive function, and performance on delayed memory (logical memory II; r=0.26, p=0.04). These associations, however, were primarily driven by age and were not present when controlling for the effect of age.
These results demonstrate a relationship between cardiorespiratory fitness and brain atrophy in the earliest clinical stages of AD. Higher fitness levels in early AD participants were associated with preserved brain volume (less brain atrophy) independent of age and dementia severity. To our knowledge this is the first study to demonstrate a direct relationship between cardiorespiratory fitness and measures of whole brain atrophy in AD. Previous structural neuroimaging studies assessed nondemented controls, rather than AD, using less direct methods (i.e., voxel-based morphometry) to demonstrate regional brain structure (as opposed to whole brain) relationships with fitness. Importantly, the causal nature of these associations cannot be determined due to the cross-sectional design. As a result at least three explanations for these findings exist: 1) cardiorespiratory fitness moderates AD-related brain atrophy, 2) the AD process itself modifies cardiorespiratory fitness, or 3) a common underlying factor modifies both AD-related brain atrophy and cardiorespiratory fitness.
First, the possibility that enhanced cardiorespiratory fitness may moderate AD-related brain changes is biologically-plausible and supported by animal studies. Higher fitness through increased physical activity has been associated with enhanced neuronal survival to brain insults,35, 36 increased vascularization,37 and elevations of growth factors in areas important for memory.38, 39 Physical activity increases neurogenesis in the hippocampus and enhances learning in mice.40 Additionally, evidence suggests that increased physical activity may modify the AD process. Increased physical activity in an AD mouse model was associated with decreased neuropathological burden in both cortical and hippocampal regions, suggesting that exercise may mediate the amyloid cascade in favor of reduced production of beta-amyloid.9 Our data is consistent with, but does not establish, cardiorespiratory fitness as a moderating factor in neurodegeneration.
A second explanation of the current results is that the AD process itself may modify an individual’s fitness level resulting in fitness declines co-occuring with AD-related brain atrophy. Behavioral changes associated with the disease include reductions in physical activity26 which in turn may result in reduced fitness levels. Individuals with AD may also potentially have reduced effort or increased frailty limiting their ability to perform a maximal treadmill test although we found little influence of these factors on the association between fitness and brain atrophy in AD. For instance, controlling for physical activity level (PASE) and physical frailty (physical performance test) did not alter the relationship between fitness and brain atrophy. Additionally, severity of dementia (CDR box score) was not associated with cardiorespiratory fitness or physiologic measures of effort (RER and peak heart rate). All participants met criteria for peak effort (RER≥1.0) while peak heart rate and RER were consistent with maximal effort for both groups. Lastly, the 11 AD subjects excluded for not meeting peak effort criteria were not more severely demented or cognitively impaired than those meeting performance criteria suggesting that reduced effort was not related to dementia severity or cognitive impairment. These observations suggest the relationship between fitness and brain atrophy was not driven by the most severely demented of these early-stage AD subjects.
Third, it is possible that a common underlying AD-related process may impact both brain atrophy and cardiorespiratory fitness. Additionally, accumulating evidence suggests that AD may not only impact the brain but may have widespread systemic manifestations. For instance, mitochondrial dysfunction is evident in the CNS and periphery in AD43 and abnormalities in metabolic and biochemical processes described in AD brains are reported in peripheral cells such as skin fibroblasts derived from AD patients.44, 45 VO2peak represents the peak rate of metabolic activity (oxygen utilization) in exercising muscle (ml of oxygen consumed per minute per kg of muscle mass) and as such our findings of 1) reduced VO2peak in AD and 2) the correlation of VO2peak with brain atrophy in AD suggests the possibility of metabolic dysfunction co-occurring in brain and muscle. Another possibility is that fitness may modify vascular-related damage (such as white matter lesions and subclinical infarcts) known to influence both brain atrophy41 and cognitive symptoms of AD.42 Further study of the relationship between fitness and vascular-related damage is needed to assess this possibility.
The cross-sectional study design limits the interpretation of the results. Further longitudinal and interventional studies will be necessary to define the role of fitness in modifying the brain aging and AD process. Validated clinical methods were used for diagnosing AD although these methods are imperfect in predicting AD pathology. Additionally, although cardiorespiratory fitness is influenced by physical activity and exercise, this study only assessed the standard measure of cardiorespiratory fitness VO2peak at one point in time and therefore the data should be cautiously interpreted with respect to the role of exercise in early AD.
This study was supported by grants R03AG026374 and R21 AG029615 from the National Institutes of Aging and K23NS058252 from the National Institute on Neurological Disorders and Stroke and by generous support from the University of Kansas Endowment Association and the Fraternal Order of Eagles. The University of Kansas General Clinical Research Center (M01RR023940) provided essential space, expertise, and nursing support. The Hoglund Brain Imaging Center is supported by grant C76 HF00201 and Dr. Brooks is supported by NS039123 and AG026482. This study utilized the LONI Pipeline environment (http://pipeline.loni.ucla.edu), which was developed by the Laboratory of Neuro Imaging and partially funded by NIH grants P41 RR013642, R01 MH71940 and U54 RR021813.
We would also like to acknowledge and thank Hal Oppenheimer and the Oppenheimer Foundation for financial support and inspiration.
Disclosure: Dr. Burns has received honoraria for speaking from Pfizer and Novartis and as a consultant from Pfizer. Dr. Swerdlow has received honoraria from Forest, Eisai, Pfizer, and Ortho-McNeil for speaking. There are no disclosures from the other investigators.