Modulating effects of fitness and physical activity on Alzheimer's disease: Implications from a six-month randomized controlled sports intervention

Abstract

Background

Physical activity and fitness are major targets in Alzheimer's disease (AD) preventive research. However, current research is heterogeneous and often disregards the relationship between these parameters and disease outcomes.

Objective

To assess the effects of physical activity and fitness on AD within the context of a multicomponent sports intervention.

Methods

46 participants with early-stage AD (mean age 70 ± 7 years, 18 women, mean Montreal Cognitive Assessment (MoCA) score 19±5) were included in a six-month randomized controlled trial (Dementia-MOVE), participating in either a multicomponent sports intervention or a control condition with a psychoeducational program. The modulating effect of fitness and physical activity changes on AD outcome parameters such as cognition, function and cerebral brain structure from 3T-MRI were examined using multiple linear regression analyses.

Results

An increase in VO₂max was associated with assignment to the intervention group (p = 0.016), lower baseline fitness (p = 0.001), and an increased rate of physical activity (p = 0.046). Only in the intervention group, ΔVO₂max had a beneficial modulating effect on the MoCA score (p = 0.039), the executive functions (p = 0.017) and regional brain volumes of the temporal lobe, e.g., the hippocampus (p = 0.044). High daily step count was associated with preserved executive functions (p = 0.001), and caregivers’ quality of life (p ≤ 0.001) in the overall sample.

Conclusions

Our results confirm that multicomponent exercise improves cardiorespiratory fitness in AD, which is associated with advantageous developments in cognitive performance and preservation of brain structure. These findings suggest that especially patients with comparably worse cognition and fitness benefit and should be encouraged for activity engagement.

Keywords

Alzheimer's disease exercise fitness development intervention lifestyle modification magnetic resonance imaging mild cognitive impairment neurodegeneration neuroimaging

Introduction

Aging society and rising life expectancy are associated with an increased prevalence of chronic diseases like dementia, accompanied by higher rates of care dependency and mortality, as well as elevated treatment costs.¹ Looking at Alzheimer's disease (AD) as the most common type of dementia, the disease often progresses over years, from mild cognitive impairment (MCI) or early manifest AD, characterized by clinically assessed minor cognitive limitations, to increasing symptom burden and disability as the neurodegeneration advances.^2,3 Aside from pharmacological therapies, exercise that increases cardiorespiratory fitness (CRF) has previously been shown to be an effective, low-cost, and easily accessible treatment option for both elder individuals at risk, as well as patients diagnosed with amnestic MCI (aMCI) or AD.^4,5

While there is increasing evidence that general physical activity, defined in contrast to sedentary behavior, can protect against cognitive decline,^6,7 the underlying effects at the cerebral level with regard to neurodegeneration are still subject to intensive research. Current findings suggest that physical activity influences neuronal molecular pathways, optimizes brain metabolism, and increases regional cerebral blood flow.^8,9 Neuroimaging studies based on these considerations show that high physical activity levels and short-term interventions for up to six months could influence brain structure in AD-affected regions such as the hippocampus and further regions of the temporal lobe. Yet their findings are inconsistent due to the general heterogeneity of study designs and often not clearly characterized study populations.^10–13 Thus, the question of how exactly fitness and physical activity affect brain integrity during neurodegeneration still remains unanswered.

Furthermore, previous exercise trials often fail to consider the complex mutual relation of physical activity and fitness improvement in the context of a sports intervention; hence, their interacting effect on the different aspects of AD pathology is unclear.¹⁴ Establishing a suitable intervention and finding the optimal exercise dose for the elderly and cognitively impaired population is therefore challenging but at the same time highly clinically relevant. So far, multicomponent exercise interventions comprising moderate-to-vigorous physical activity (MVPA) at least twice a week for a minimum duration of three to six months appear to be the most promising for increasing CRF.^15,16 In line with these findings, the World Health Organization (WHO) has also published recommendations on the rate of physical activity in elder subjects, constituting at least 150–300 min of moderate-intensity or at least 75–150 min of vigorous-intensity aerobic physical activity per week, with the recommendation to perform multicomponent activity to prevent falls and increase muscle strength.¹⁷ Additional psychoeducational and lifestyle-modifying programs, for instance based on the FINGER study design, can also increase adherence to and acceptance of such recommendations and interventions,^18,19 pointing again to the relevance of multicomponent lifestyle interventions.

The Dementia-MOVE trial (Multi-Objective Validation of Exercise in Dementia) therefore examined the influence of a six-month multicomponent exercise intervention based on these previous recommendations, while adjusted to the needs of people with AD. Fitness development, the influence of CRF and activity changes on the different domains and facets of AD progression are analyzed for a holistic approach to the disease.

We expect that people with AD participating in the exercise intervention program will improve their fitness compared to a control group attending a psychoeducational program only. We hypothesize that increased fitness and a higher level of physical activity due to participation in the intervention program leads to a positive impact on cognitive performance, neuropsychiatric health and brain integrity in AD. As an underlying goal, we examine the feasibility and acceptance of this intensified program as a treatment option in a cohort of elderly people affected by AD.

Methods

Study design and participants

The study Dementia-MOVE is a randomized controlled clinical trial with two arms, including an intervention group receiving a six-month multicomponent exercise training and a control group undergoing a psychoeducational program.²⁰ In summary, a total of 46 people aged between 50 to 80 years with early AD, who were cognitively and physically eligible for a physical exercise program, were recruited at the memory clinic of the Department of Neurology of the RWTH Aachen University Hospital (UKA), and via advertising and search in newspapers (see Figure 1). All participants were characterized as having mild cognitive impairment (MCI) or mild AD with a cerebrospinal fluid (CSF) constellation of amyloid-β 1-42 or amyloid-β 42/40 ratio (A), phospho-tau (T), and tau (N) based on Jack et al. (2018), resulting in their characteristic ATN-status alongside the AD-continuum.²¹ According to this, all participants included showed a pathological amyloid CSF concentration (A+), with phospho-Tau CSF concentrations in either elevated (A⁺T⁺; n = 32) or normal range (A⁺T⁻, n = 9). In cases where CSF biomarkers were not available (n = 5), the diagnosis of AD was based on neuroimaging and clinical evaluation (i.e., neurological examination, and neuropsychological assessment) as part of an expert consensus at the memory clinic at UKA. All participants presented an early-stage disease with a Clinical Dementia Rating scale (CDR) ≤1. Exclusion criteria mainly included severe systemic, cardiopulmonary, orthopedic and/or psychiatric diseases (for a complete overview of the inclusion and exclusion criteria, see also Supplemental Table 1). Sample size calculations were based on power analyses which were derived from our primary outcome parameter presented by metabolic MR imaging and our previous results in this context.^20,22,23 After performing baseline assessments at time point T₁, participants were randomized into either the exercise intervention group (n = 26) or the control group (n = 20) by a blinded member of the study team using the WINPEPI software.²⁴ Stratification followed age and fitness level, as determined via VO₂max, or in case of non-availability, the result of the 6-min walk test (6MWT). The site of residency was further considered when allocating between Aachen and Simmerath. In case of unequal numbers of participants, allocation was performed in favor of the intervention group.

Figure 1.

Participant flow chart of the screening and intervention period and the subsequent analysis. A total of 41 patients (22 in the intervention group and 19 in the control group) were finally included for statistical data analysis. Note that even when nine participants dropped out of the program, four participants still agreed to participate in the outcome assessments.

The study and all protocol changes were approved by the Ethics Committee at the Medical Faculty of the Rheinisch-Westfälische Technische Hochschule Aachen (RWTH Aachen) (EK 306/18) and were performed according to the latest Declaration of Helsinki. All participants gave written informed consent for study participation and protocol changes. The trial was registered at ClinicalTrials.gov (NCT03939286). Further details on the study design of the Dementia-MOVE protocol have been previously published.²⁰

The multicomponent intervention program

A detailed overview of all intervention components is presented in Figure 2. The intervention program started after randomization and took place for a total duration of 26 weeks (six months plus two weeks for optional compensation on missed appointments). Participants assigned to the intervention group attended a supervised exercise session for 60 min per week, taking place at two study sites, the UKA or the hospital Eifelklinik St Brigida in Simmerath, located approximately 20 km south-east of Aachen, with its affiliated professional physiotherapy practice. The program was homogenized between both sites and included aerobic, resistance, and coordination exercises. It was supervised by physiotherapists and members of the study team. Sports groups usually consisted of about five participants per group, therefore assuring an optimized supervision and individualized guidance of the sports program. During the first supervised training session, the optimal exercise load for the program was determined for each individual: (1) ergometer wattage was estimated from 6-MWT distance according to Luxton et al.,²⁵ and (2) load for a moderate resistance training intensity was calculated according to the Epley formula for 40% of one-repetition-maximum.²⁶ Over the course of the intervention period, the exercise load was then gradually increased according to moderate-to-vigorous physical exertion regularly surveyed using the Borg RPE-scale during the training sessions.²⁷ An exemplary training plan including mean initial exercise loads is provided in the Supplemental Table 2.

Figure 2.

Outline and timeline of the multicomponent intervention design with its three main components: the supervised exercise, the psychoeducation program and the different methods for activity monitoring as well as the three acquisition time points. Borg RPE-Scale: Borg Rating of Perceived Exertion Scale; BZgA: German Federal Centre for Health Education.

In addition to the participation in the supervised training program, participants of the intervention group were further instructed to perform a moderate-to-vigorous intensity home-based training for at least 30 min per week, as well as stretching and toning exercises for at least 15 min twice per week, following current recommendations on repetitive physical exercise.²⁸ Supporting training cards from the German Federal Centre for Health Education (BZgA) for the stretching and toning exercises were distributed to the intervention participants.²⁹

Furthermore, both groups participated in monthly psychoeducational lectures at both study sites, respectively, with a focus on lifestyle modifying factors, disease management, and help offers, as well as therapy options based on the content of the FINGER study (see Figure 2),³⁰ thus, also controlling for the social aspect of the intervention as control condition.

During the first COVID-19 lockdown in Germany, the supervised exercise program was substituted by a home-based training for nine weeks from March to May 2020, for 18 participants. Exercises in the home-based training were chosen to optimally resemble the supervised training and to address the same muscle groups without resorting to training machines, yet participants were allowed to implement available home gym equipment at any time. To optimize coordination training at home, the participants received an MFT Fun Disc^® balance plate, which was also used during the training sessions at both study sites. Instructions for a homogenized home-based training were sent in written form and orally explained to all participants of the sports groups and their relatives (see also Supplemental Table 2). The psychoeducational topics were summarized in paper form and sent to all participants during the lockdown period. The discussion of these lectures and a check for training adherence was carried out by weekly telephone calls through the study team. Since the T₃ assessment of five participants was not feasible at that time, the intervention was extended for a further nine weeks for these individuals to maintain the intervention effect, until the lockdown was terminated, and assessments could be safely performed.

Measuring fitness and physical activity

Aerobic fitness was assessed via estimation of VO₂max, a well-established and validated parameter in exercise testing, which defines a person's maximum capacity to absorb and metabolize oxygen.³¹ We therefore performed a submaximal ergometer test on a Lode Corival CPET^® device following a modified Astrand protocol.³² Additionally, the 6-MWT was performed for an alternative estimation of VO₂max when the participants failed to increase heart frequency during the Astrand protocol, for instance, due to β-blocker intake or orthopedic impairment impeding ergometer cycling, which can be common in an elder population.³³

Furthermore, handgrip strength was measured using a Saehan^® hand dynamometer,³⁴ as well as postural stability via the MFT-S3-Check^® system, which records stability, sensomotorics, and symmetry data of two passes, with a balance plate tilting front/back and left/right, respectively.³⁵ Previous studies have associated weak performance in both parameters with cognitive decline.^36,37 All fitness assessments were performed at time point T₁ at baseline, after three months at time point T_2, and after six months at time point T₃.

A written diary was used to monitor physical activity and compliance throughout the entire duration of the intervention. It was completed by the participants with the help of their caregivers if needed and included space for the participants to record types and duration of daily activities as well as additional information such as subjective exertion during exercise, hours of sleep, medication taken if necessary, and medical issues occurring during study participation. Filled pages were regularly collected by the study team and checked for missing or incomplete data. In case of such occurrences, the participants and caregivers were re-advised to avoid missing data on previous activities.³⁸

Furthermore, the PAQ50+ questionnaire, which is adapted to physical activity assessment in elder people, was used at each assessment time point four weeks prior to the visits T₁, T₂, and T₃.³⁹ To demonstrate the reliability of these two subjective assessing methods, the Fitbit Charge 2^® (Fitbit), a wearable fitness tracker, was distributed to the study participants for an average duration of 14 days around the middle of the intervention period to objectively record activity during that time. The Fitbit registered step count, walking distance, activity minutes of varying intensity, calorie consumption, as well as sleep.

Processing of physical activity and fitness assessment data

A detailed methodological description and comparison of the three assessment methods of physical activity and fitness, i.e., the wearable fitness tracker, the written diary, and the PAQ50+, is presented in our recent publication.³⁸ In short, daily metabolic equivalent of task (MET) scores were calculated based on the different output data as follows:

D a i l y M E T = \sum_{n = 1}^{A} \frac{t_{n} \times M E T_{n}}{60},

(1)

with A being the number of distinct daily activities performed, t_n the time in minutes spent on activity n, weighted by its individual MET MET_n; low intensity activities with a MET score <2 were filtered by our algorithm.^38,40

A further MVPA score was calculated by summarizing daily moderate and vigorous activity minutes for all three assessing methods. For the diary and the PAQ50+, activity minutes were computed based on the activities’ individual MET scores, classifying moderate activity as MET 3–5.9 and vigorous activity as MET >6 according to previous literature.^41,42

VO₂max as a measure of CRF was estimated from ergometer testing data according to von Dôbeln et al.,⁴³ taking the advanced age and differing body compositions as well as maximum heart rates of an elder study population into account compared to the classic nomogram introduced by Astrand and Rhyming (1954)³²:

V O 2 m a x = 1.29 \sqrt{\frac{l o a d [k p m]}{h e a r t r a t e [a t 5 - 6 m i n, b p m] - 60}} \times e^{- 0.00884 \times a g e [y e a r s]}

(2)

For eight participants who were not able to cycle or did not reach a defined threshold heart rate >110/min during the test protocol on at least one acquisition time point, VO₂max was estimated from the 6-MWT distance according to Burr et al. to avoid inaccuracies in estimation resulting from stronger heart rate deviations of the original Astrand protocol.⁴⁴

For handgrip strength, we focused on the participants’ dominant hand (right = 38, left = 2, ambidextrous = 1 of the included subjects). Furthermore, a mean composite score of all postural stability measuring components was calculated, including Z-standardized stability, sensomotoric and symmetry values of the two different passes performed.

Assessment of AD outcome parameters

Neuropsychiatric and cognitive screening

Neuropsychiatric and behavioral symptoms were assessed via questionnaires sent to the participants and their relatives in advance of each assessment appointment T₁, T₂, and T₃. This assessment included the Quality of Life in AD (QoL-AD),⁴⁵ the German version of the Hospital Anxiety and Depression Scale (HADS-D),⁴⁶ the Neuropsychiatric Inventory Questionnaire (NPI-Q),⁴⁷ as well as the Pittsburgh Sleep Quality Index (PSQ-I).⁴⁸ Absolute questionnaire scores were used for the subsequent analyses. For HADS-D and PSQ-I, higher scores indicate more severe depressive symptoms and worse sleep quality, respectively.

Global cognitive screening was evaluated via the Mini-Mental State Examination (MMSE, at baseline T₁ only),⁴⁹ and the Montreal Cognitive Assessment (MoCA).⁵⁰ Furthermore, a neuropsychological test battery focusing on attention, verbal memory, and executive functions was performed. Alternate versions of the tests were used at the three assessment time points T₁, T₂, and T₃, if available. The attention score comprised cued and non-cued reaction time of the Test of Attentional Performance (TAP),⁵¹ and digit span forward.⁵² Verbal memory was assessed via the California Verbal Learning Test (CVLT).⁵³ Executive functions domain included the Trail Making Test's (TMT) B/A-ratio,⁵⁴ phonemic and semantic word fluency,⁵⁵ digit span backwards,⁵² and the Stroop interference task.⁵⁶ Results of the individual tests were Z-standardized with baseline means and standard deviations as references and then summarized as mean composite scores for each domain.⁵⁷

Structural MRI analysis

All MRI images were acquired on the same 3-Tesla Siemens PRISMA whole-body MR scanner at baseline (T₁) and after the intervention (T₃). Isotropic high-resolution T1-weighted images of the brain were obtained with an MP-RAGE sequence (TR = 2400 ms, TI = 1000 ms, TE = 2.4 ms, FA = 8°, 0.8 mm isotropic resolution). Images were segmented into gray matter, white matter, CSF, brainstem and cerebellum via the antsLongitudinalCorticalThickness.sh pipeline implemented in Advanced Normalization Tools (ANTs).⁵⁸ Parcellation for a region of interest (ROI) approach was performed via the Hammersmith Brain Atlas, containing 68 brain regions,⁵⁹ as well as VolBrain for the segmentation of subcortical regions.⁶⁰ For the current analysis, eight ROIs in the temporal lobe were preselected, of particularly high interest in AD: hippocampus (HPC), amygdala (AMG), medial anterior temporal lobe (M-ATL), lateral anterior temporal lobe (L-ATL), parahippocampal and ambient gyri (PHC), superior temporal gyrus (STG), middle and inferior temporal gyri (MTG/ITG) and fusiform gyrus (FFG). All brain volumetric data were normalized for the total intracranial volume (TIV) and averaged between both hemispheres.

Furthermore, T1-weighted and fluid-attenuated inversion recovery (FLAIR) MR images (TR = 9000 ms, TI = 2500 ms, TE = 79 ms, FA = 150°, 0.4 mm × 0.4 mm × 3 mm resolution) were used for the segmentation of white matter lesions in VolBrain, resulting in the total lesion volume of all intracerebral lesions at baseline and after intervention.^61,62

For whole-brain group comparison, a tensor-based-morphometry analysis was performed. The logarithm of the determinant of the Jacobians (Log( $\nabla Jac$ )), which represent local expansions or contractions in each voxel of each individual T1-weighted image when normalized to the subject template (created by the antsLongitudinalCorticalThickness-pipelines) were used to analyze structural changes. The difference of the logarithms of the determinants of the Jacobians corresponding to the follow-up and baseline images, respectively, were normalized to the OASIS template of miccai2012-multi-atlas-challenge-data,⁶³ and smoothed (σ = 2 mm) for consequent group analysis.

Statistical data analysis

Statistical analyses were performed using IBM SPSS Statistics 26 software and MATLAB R2019b. First, all data were visually inspected for normal distribution and possible outliers in boxplots and scatter plots. MR data were checked for segmentation failure, movement artifacts, and structural brain lesions as well. Pairwise deletion was applied for missing data. The participants’ characteristics and measurement data were compared for group differences using mean (M), standard deviation (SD), median (Mdn), median absolute deviation (MAD) as well as independent samples t-test, Mann-Whitney-U-test and Chi-square analysis, as indicated. Correlation analyses of fitness and activity data were performed using Pearson's correlation and Spearman's rank correlation, reporting the correlation coefficients r and ρ, respectively. Differences between follow-up minus baseline values (T₃-T₁) for the different assessment parameters were calculated and indicated with delta (Δ). Changes in fitness parameters were analyzed via multiple linear regression analysis, including group, fitness baseline levels, age, and sex as predictors, considering collinearity tests. To investigate the influence of the COVID-19 study protocol changes, a categorical variable (encoding 0 = not affected, 1 = affected), further referred to as Cov-19_adj, was entered into the model as well.

Prediction of neuropsychiatric, cognitive, and brain structural changes was then carried out via multiple linear regression analysis. Changes in VO₂max, i.e., ΔVO₂max, and activity via Fitbit step count were entered as predictors in separate models. At first instance, we tested for potential confounding factors, including age, sex, baseline cognition, CSF status and study site as covariables in the model. Based on the results of this analysis, the design was simplified by excluding non-significant covariables for the final analysis, which was performed in each group separately regarding ΔVO₂max as the dependent variable, and in the overall cohort regarding the Fitbit step count. The influence of ATN and APOE4 carrier status was further investigated by additionally entering those parameters as categorial variables into the regression models (encoding 0 = A⁺T⁻, 1 = A⁺T⁺, as well as 0 = heterozygous APOE4 allele carrier/non-APOE4 allele carrier, 1 = homozygous APOE4 allele carrier) on a probatory basis.^64–66 For neuropsychiatric and cognitive parameters, the respective baseline level was included in all models to adjust for the individual clinical disease manifestation. If the model included more than one possible predictor, the best fit was determined via changes in RMSE and R² as well as overall significance in all models. Significance is determined by an uncorrected p < 0.05 to meet the exploratory concept of the analysis. Effect size estimates were calculated using Cohen's d and Pearson's r as indicated. The 95% confidence intervals (95% CI) are reported. Furthermore, for the whole brain analysis, a voxel-wise permutation analysis (n = 2000 permutations) for comparison of the intervention and the control group was performed in BROCCOLI,⁶⁷ on voxels inside the brain tissue mask with a statistical threshold of p < 0.05, including MoCA at baseline, VO₂max at baseline, and education in years as covariates into the model.

Results

Participant flow

A total of 46 participants were initially included at baseline and randomized into either the intervention group (n = 26) or the control group (n = 20). At the UKA site, 19 people participated in the sports program, and seven in Simmerath. 43 participants completed the three-month and 43 completed the six-month follow-up (see Figure 1). Participants who missed all exercise sessions or psychoeducational lectures, yet still participated at least at the T₁ and T₃ assessments (n = 3) remained in the final analysis. Apart from that, one participant originally randomized to the intervention group, but unable to participate in any training session due to urinary incontinence and therefore sedentary, yet regularly attending the psychoeducational lectures, was allocated to the control group for final analysis. Furthermore, two participants of the intervention group were excluded from the final analysis for non-adherence to the home-based exercise program during the COVID-19 lockdown (German lockdown 03/19/2020 to 05/12/2020). Brain MRI was omitted in four out of the 41 remaining participants (n = 1 intervention group, n = 3 from control group) from at least one assessment time point. Reasons for not performing the MRI on these subjects were claustrophobia (n = 3), and the presence of a cardiac pacemaker unit (n = 1). For a detailed illustration of the participant flow, see also Figure 1.

Adherence to protocol

Mean adherence to the supervised exercise intervention was 72% (SD = 16%) and could be increased to 76% (SD = 16%) including two additional appointments that were offered if attendance was below the targeted 80% after 26 weeks of intervention. Eleven participants of the intervention group (50%) had an attendance of more than 80%. Reasons for reducing training participation for longer than a week were chronic musculosceletal pain (n = 5), rehabilitation after minor surgery (n = 3), respiratory infection (n = 2) and a toe abscess (n = 1).

Participants of the intervention group attended an average of four out of the seven psychoeducation lectures offered during the intervention period, while those in the control group attended an average of five events.

Regarding the activity diary, an average of 85% (SD = 22%) versus 77% (SD = 36%) of diary entries were completed, based on 182 days (26 weeks), in the intervention and the control group, respectively.

Occurrence of adverse events

There were no severe adverse events to report in both groups. Non-serious adverse events possibly associated with the intervention during the supervised exercise mainly included musculoskeletal pain (n = 4), vertigo (n = 2), nausea (n = 1), arterial hypertension (n = 1), and headache (n = 1). In case of discomfort, the exercise load was reduced, or the training session was terminated prematurely, which occurred once for three subjects and twice for one subject over the course of the intervention.

Baseline results and patient characteristics

An overview of the baseline characteristics and measurements of the analyzed study sample is given in Table 1A and 1B. At baseline, there was a group difference between the intervention group compared to the control group in terms of mean years in education (12.5 ±. 3.8 years versus 14.8 ±. 2.4; U = 107, z = -2.71, 95% CI (1.00, 4.00), p = 0.01, r = 0.42) and mean MoCA score (17.1 ± 4 versus 21.3 ± 4.8; t (39) = -3.02, 95% CI (−6.89, −1.36), p = < 0.01, |d| = 0.95). Performance in the executive functions domain also differed between the intervention and the control group (−0.21 ± 0.61 versus 0.24 ± 0.6; t (39) = -2.41, 95% CI (−0.84, −0.07), p = 0.02, |d| = 0.75). There were no group differences at baseline in other cognitive domains. Furthermore, the amyloid-β 42/40-ratio as well as tau and phospho-tau levels from CSF differed between both groups. However, it is important to note that CSF markers were acquired during the diagnostic process and not at the time point of study inclusion, with a maximum time distance of 1954 days before the intervention start. Therefore, interpretation of these differences in CSF pathology must be made with caution.

Table 1A.

Basic characteristics of the included study sample (n = 41/46) at baseline.

	Intervention Group(n = 22)	Control Group(n = 19)	p
Intervention Site (AC/SI)	15/7	15/4	0.438
Age (y)	72.1 (5.8)	68 (8.2)	0.067
Sex (m/f)	13/9	12/7	0.790
Education (y)	12.5 (3.8)	14.8 (2.4)	0.007*
Disease duration (y)	3.6 (2.2)	3.4 (1.7)	0.936
CDR	0.8 (0.3)	0.6 (0.4)	0.151
CSF Amyloid-β 1–42 (pg/ml)^a	402.3 (161.7)	430.4 (176.8)	0.693
CSF Amyloid-β 42/40-ratio^a	0.4 (0.1)	0.7 (0.9)	0.013*
CSF Tau (pg/ml)^a	735.9 (322.7)	446.4 (263.2)	0.005*
CSF Phospho-Tau (pg/ml)^a	98 (36.5)	70 (27.8)	0.011*
A⁺T⁺; 2018, Jack et al. (%)	77.3	63.2	0.322
APOE4 Status (hetero-/homozygous, %)	40.9/22.7	63.2/10.5	0.333
Dementia medication (%)^b	72.7	47.4	0.097
Psychiatric medication (%)^c	31.8	31.6	0.987
Changes in dementia medication (%)^d	31.8	36.8	0.735
Cardiovascular risk^e	1.6 (1)	1.3 (1)	0.155

Frequencies are indicated for categorical variables, mean values and standard deviations for metric variables, respectively. *p < 0.05. Missing values: CSF Amyloid-β 1–42 (n = 4), CSF Amyloid-β 42/40-ratio (n = 5); CSF Tau (n = 3); CSF Phospho-Tau (n = 3). AC, Aachen; SI, Simmerath; CDR, Clinical Dementia Rating; CSF, cerebrospinal fluid; APOE4, Apolipoprotein E4

Lab-specific AD-relevant cut-off values for cerebrospinal fluid parameters are as follows: amyloid-β 1–42 <450 pg/ml, amyloid-β 42/40-ratio <0.5, tau >450 pg/ml, phospho-tau >61 pg/ml, analysis was performed in the Neurochemical Laboratory at the University of Göttingen.^82,83 One participant had the liquor diagnostics performed in an external practice, with a deviating lab-specific cut-off for amyloid-β 1–42 of <500 pg/ml.

Including the substances Donepezil, Rivastigmine, Galantamine, and Memantine licensed for the treatment of AD in Germany.

Including all kinds of psychopharmaceuticals such as antipsychotic, antidepressant or anxiolytic medication.

Indicated the change in dementia medication during the course of the intervention.

1 point each was assigned for the presence of arterial hypertension (anamnestic or medication), diabetes mellitus (anamnestic or medication), hypercholesterinemia (>5.16 mmol/L as the UKA internal laboratory reference value or medication), elevated body mass index (>30 kg/m²), active smoking (anamnestic); max. score = 5.²⁰

Table 1B.

Measurement results at time points T₁ and T₃. Mean values and standard deviations, as well as group tests in direct time-point comparison, are indicated.

	Time point T₁ (Baseline)			Time point T₃
	Intervention Group(n = 22)	Control Group(n = 19)	P	Intervention Group(n = 22)	Control Group(n = 19)	p
Fitness assessment
VO₂max (ml/min/kg)	29.8 (4.72)	29.8 (6.17)	1.00	31.84 (4.12)	30.35 (5.55)	0.32
6-MWT (m)	532.32 (104.73)	532.92 (83.92)	0.98	533.83 (95.27)	547.01 (71.91)	0.63
Handgrip strength, dominant hand (kg)	30.68 (9.68)	30.84 (10.3)	0.96	31.02 (12.11)	32.82 (11.26)	0.63
Postural Stability (Z-score)	−0.03 (0.51)	0.03 (0.6)	0.78	0.04 (0.44)	−0.03 (0.58)	0.45
BMI (kg/m²)	23.78 (3.2)	24.31 (2.89)	0.58	23.57 (3.1)	23.97 (2.95)	0.68
Cognitive performance
MMSE	24.09 (3.16)	25.05 (3.6)	0.37
MoCA	17.14 (3.97)	21.26 (4.78)	<0.01*	16.23 (3.9)	21.32 (5.21)	<0.01*
Attention (Z-score)	−0.16 (0.76)	0.19 (0.65)	0.08	−0.22 (0.79)	0.05 (0.78)	0.33
Verbal memory (Z-score)	−0.12 (0.44)	0.14 (0.53)	0.09	0.06 (0.60)	0.12 (0.57)	0.78
Executive functions (Z-score)	−0.21 (0.61)	0.24 (0.6)	0.02*	−0.25 (0.53)	0.22 (0.53)	<0.01*
Neuropsychiatric symptoms
QoL (patient)	37.27 (4.59)	37 (5.31)	0.60	36.91 (4.77)	35 (4.49)	0.31
QoL (family)	35.5 (3.82)	34.16 (6.88)	0.33	33.86 (4.79)	34.22 (4.4)	0.79
HADS (total score)	6.59 (5.8)	9.16 (6.88)	0.20	9.27 (5.93)	11 (9.82)	0.51
NPI-10	7.72 (6.31)	7.19 (10.09)	0.34	7.86 (9.68)	7.56 (14.8)	0.31
PSQ-I	4.36 (3.47)	5.58 (3.79)	0.29	4.27 (3.53)	3.89 (2.33)	0.90
MRI
Hippocampus Volume (%TIV)	0.22 (0.03)	0.23 (0.04)	0.48	0.22 (0.03)	0.23 (0.04)	0.32
Gray Matter Volume (%TIV)	34.19 (1.42)	34.95 (2.44)	0.24	33.82 (1.47)	34.6 (2.75)	0.31
Cortical Thickness (mm)	2.61 (0.12)	2.7 (0.23)	0.16	2.55 (0.14)	2.66 (0.25)	0.12
White Matter Volume (%TIV)	30.92 (1.78)	30.75 (1.54)	0.76	30.89 (1.87)	31.02 (2)	0.84
White Matter Lesion Volume (%TIV)	0.46 (0.7)	0.42 (0.63)	0.19	0.54 (0.75)	0.43 (0.64)	0.10

p < 0.05. Missing values: T₁, postural stability (n = 1), Quality of Life (family) (n = 2), NPI-10 (n = 7), all MRI parameters (n = 4); T₃, 6-MWT (n = 1), Verbal Memory (Z-score) (n = 1), Quality of Life (patient) (n = 1), Quality of Life (family) (n = 1), NPI-10 (n = 2), all MRI parameters (n = 4). VO₂max: maximal oxygen uptake; 6-MWT: 6-Minute Walk Test; MMSE: Mini-Mental State Examination; MoCA: Montreal Cognitive Assessment; HADS: Hospital Anxiety and Depression Scale; TIV: total intracranial volume.

Monitoring of physical activity and fitness during the intervention

Assessment and monitoring of physical activity served mainly two purposes: (1) to check for protocol adherence, and (2) to assess general physical activity and fitness during the study. In a recent publication, we demonstrated the good validity of the diary and fitness trackers for the assessment of physical activity and fitness in the same patient group,³⁸ with comparable values in MET-scores and MVPA between the assessment via the fitness tracker and the diary (see Supplemental Table 3 and Supplemental Figure 1). Therefore, fitness tracker activity levels were considered as representative for general activity level for the whole intervention period, while physical activity reported by the PAQ50+ questionnaire was used as an assessment prior to intervention start.

Based on the physical activity reported in the diaries, both the intervention and the control group fulfilled, on average, the current WHO recommendations for general physical activity for the elderly during the study period, with a median weekly activity time of 171 min of moderate-intensity and 91 min of vigorous activity in the intervention group, and 281 min of moderate-intensity and 72 min of vigorous activity in the control group per week.¹⁷ At the subject level, the WHO recommendation was achieved by 19 participants in the intervention group and 16 participants in the control group.

Over the six-month intervention period, the global level of daily MET derived from the diary remained constant over time in both groups. The rate of vigorous activity reported was higher in the intervention group compared to the control group (intervention, Mdn = 15.18, MAD = 3.50; control, Mdn = 10.39, MAD = 6.87; U = 112, z = -2.13, 95% CI (−11.93, −0.57), p = 0.03, r = 0.34). However, the rate of moderate activity appeared to be higher in the control group compared to the intervention group (see also Figure 3), where we observed that a few outlying individuals in the control group with an exceptionally high activity level led to a generally higher level of reported moderately intense activity on average (intervention, Mdn = 27.39, MAD = 12.94; control, Mdn = 48.19, MAD = 20.25; U = 86, z = -2.86, 95% CI (6.99, 41.60), p = < 0.01, r = 0.46).

Figure 3.

Physical activity development during the intervention period derived from the paper-based physical activity diary. Daily median moderate (A) and vigorous (B) activity minutes for every intervention month in both groups are indicated, including the 95% confidence intervals. For the intervention group, the supervised exercise program and all home sports activities are included. *significant results in the Mann-Whitney U test with p < 0.05. Missing values: n = 2.

The development of fitness parameters over the intervention is illustrated in Figure 4: all fitness parameters increased in the intervention group. Compared to the control group, however, only VO₂max (intervention, M_T3–1 = 2.04, SD = 2.34; control, M_T3–1 = 0.55, SD = 2.57) and the postural stability score (intervention, M_T3–1 = 0.1, SD = 0.41; control, M_T3–1 = -0.06, SD = 0.45) improved in the intervention group.

Figure 4.

Development of fitness parameters assessed at all three visits (at baseline T₁, T₂, and T₃) in both groups. Mean values are indicated, including the 95% confidence intervals, significant group difference in ΔVO₂max in multiple linear regression analysis with p = 0.02, considering baseline VO₂max and ΔPAQ50+ as covariables (see also Figure 5). There were no significant group differences concerning the other fitness parameters.

We then analyzed which factors positively predicted ΔVO₂max in multiple linear regression analyses. As expected, ΔVO₂max was significantly associated with an affiliation to the intervention group (t = 2.52, p = 0.02), furthermore with an increase of activity compared to before the intervention start as reflected by ΔPAQ50+_T2/3–1 (t = 2.07, p = 0.05), and a lower baseline VO₂max (t = -3.59, p = < 0.01) (see Figure 5). Regarding the additional fitness parameters, there were no group differences revealed in the regression analyses. A summary of all statistical models is available in Supplemental Table 4.

Figure 5.

Multiple linear regression model of ΔVO₂max. The unstandardized predicted values of the model, including group, baseline VO₂max, and ΔPAQ50+ as significant predictors are plotted on the x-axis against ΔVO₂max on the y-axis. Colors indicate data points of the intervention group in red and data points of the control group in blue, respectively. The model summary is included below, indicating the regression coefficient B, its 95% confidence interval (95% CI), the standardized regression coefficient β, the respective t value, and the significance level p for each predictor, as well as the overall model accuracy, including the effect size $R_{a d j}^{2}$ , p, and root mean squared error (RMSE). The significance level is defined as p < 0.05.

No significant change was detected in the general activity levels, measured by diary MET, of participants affected by changes to the study protocol during the COVID-19 pandemic (z = -1.34, 95% CI (−3.08, 0.57), p = 0.26, r = 0.28), compared to those who were not affected. Furthermore, multiple regression analysis revealed no influence of the COVID-19 adjustment variable (Cov-19_adj) on development of VO₂max, 6-MWT distance, and postural stability, leading to its subsequent exclusion from the models. However, a certifiable beneficial effect on the development of handgrip strength could be reported for people being affected by the COVID-19 protocol changes (t = 2.35, p = 0.03). None of the participants in the exercise group presented clinical symptoms or were tested positive for COVID-19 during their six-month intervention period.

Influence of physical activity and fitness development on changes in neuropsychiatric symptoms, cognitive performance, and brain integrity

As mentioned before, baseline VO₂max showed a negative association with ΔVO₂max (r = -0.44, 95% CI [-0.66, −0.15], p = < 0.01) and was therefore initially regressed out from further analyses, applying the unstandardized residuals of ΔVO₂max ∼ VO₂max_T1 in subsequent regression analyses. Since both groups revealed differences in cognitive and educational status and potentially in the neurodegenerative markers from CSF at baseline, and to circumvent that baseline imbalances potentially hampered direct group comparisons (see Table 1B), we performed multiple linear regression analyses for disease outcome parameters separately in both groups. After screening the covariables, only the respective baseline levels of neuropsychiatric and cognitive parameters, showing a significant modifying effect, remained in the model; all other covariables were subsequently excluded to simplify the analysis. Interestingly, only in the intervention group CRF improvement did positively predict changes in MoCA (RMSE = 2.48, p = 0.03, $R_{a d j}^{2}$ = 0.25) and in executive functions (RMSE = 0.26, p = < 0.01, $R_{a d j}^{2}$ = 0.45), as well as in volumetric changes of regions of the temporal lobe (HPC, RMSE = 4×10⁻³, p = 0.04, $R_{a d j}^{2}$ = 0.16; AMG, RMSE = < 0.01, p = 0.03, $R_{a d j}^{2}$ = 0.20; M-ATL, RMSE = 0.01, p = 0.01, $R_{a d j}^{2}$ = 0.28; PHC, RMSE = 0.01, p = 0.01, $R_{a d j}^{2}$ = 0.25; STG, RMSE = 0.01, p = 0.05, $R_{a d j}^{2}$ = 0.15), which was not the case for the control group, indicating an isolated mediating positive role of ΔVO₂max on these outcome parameters in the intervention group (Figure 6A, B). In models that considered changes in neuropsychiatric and cognitive parameters, an inverse association was found between the baseline levels of these parameters and their changes from baseline to T3 (T_3–1 deltas). Furthermore, neither the A⁺T⁺ status nor the APOE4 carrier status played a moderating role in either group, which led to their exclusion from the analysis models.

Figure 6.

T-values of the regression analyses for predicting disease outcome parameters from changes in VO₂max (i.e., ΔVO₂max), separately in the intervention group (A) and the control group (B), and physical activity via Fitbit step count in the overall cohort (C). ΔVO₂max was preliminary corrected for baseline VO₂max in both groups by determining the residuals of the regression ΔVO₂max ∼ VO₂max_T1. For neuropsychiatric symptom and cognitive performance variables, baseline variables were additionally added to the model. The best models were chosen via comparison of R², RMSE, and significance level. QoL: quality of life; HADS: Hospital Anxiety and Depression Scale; NPI-10: Neuropsychiatric Inventory; PSQ-I: Pittsburgh Sleep Quality Index; MoCA: Montreal Cognitive Assessment; HPC: hippocampus; AMG: amygdala; M-ATL: medial anterior temporal lobe; L-ATL: lateral anterior temporal lobe; PHC: parahippocampal and ambient gyri; STG: superior temporal gyrus; MTG/ITG: middle and inferior temporal gyri; FFG: fusiform gyrus; WML: white matter lesions.

As a second step, the influence of the physical activity level derived from Fitbit step count was analyzed on the outcome parameters, as this was shown to be the generally most reliable measurement of the Fitbit tracker in previous studies.^68,69 Since mean activity levels via Fitbit recordings were similar between groups, we hypothesized that the influence of physical activity itself might differ from the immediate intervention effect reflected by the CRF improvement. Supporting this assumption, a small to moderate positive association was found between the Fitbit step count as the absolute physical activity level and baseline VO₂max (ρ = -0.37, 95% CI (0.05, 0.62), p = 0.02), but not ΔVO₂max (ρ = -0.15, 95% CI (−0.45, 0.18), p = 0.36). Again, multiple linear regression analyses were performed to analyze the influence of physical activity via Fitbit step count on disease outcome parameters in the overall cohort (n = 39) (Figure 6C). Higher physical activity levels were associated with an increased Quality of Life for caregivers (RMSE = 3.53, p = 0.01, $R_{a d j}^{2}$ = 0.23), as well as better performance in executive functions (RMSE = 0.31, p < 0.01, $R_{a d j}^{2}$ = 0.27). Again, there was a negative association with the baseline levels of these variables. There was also a positive effect of the Fitbit step count on the regional cortical thickness of the lateral anterior temporal gyrus (RMSE = 0.18, p = 0.03, $R_{a d j}^{2}$ = 0.11). No influence of ΔVO₂max was observed on vascular lesion change, sleep, and neuropsychiatric and depressive symptoms from NPI and HADS. See Supplemental Table 5 for further details on the analyses.

Whole-brain analysis revealed higher brain volume in the intervention group in the upper frontal brain lobe compared to the psychoeducation control group when taking MoCA and VO₂max baseline levels, as well as education, as covariates into account at a liberal cluster threshold of p < 0.1 with n = 2000 permutations. There were no significant group clusters at a cluster threshold of p < 0.05 (see Figure 7).

Figure 7.

Illustration of the TBM results from the voxel-wise permutation analysis at the whole- brain level, revealing higher brain volume in the upper frontal lobe in the contrast intervention group > control group with MoCA and VO₂max at baseline and education as covariates, correction at the cluster-level at p < 0.1, n = 2000 permutations.

Discussion

This study presents the results of activity and fitness development from the Dementia-MOVE trial, a six-month multicomponent exercise intervention in AD. We examined the influence of CRF and physical activity on different clinical aspects of AD pathology, such as neuropsychiatric symptoms, sleep, quality of life, cognitive performance, as well as structural cerebral changes assessed by 3T MRI.

The physical activity and fitness assessment methods used in our study demonstrated high compliance with our protocol instructions, adhering to the current WHO recommendations on weekly activity rates for the elderly.¹⁷ The physical activity assessment via wearable fitness trackers showed comparable physical activity levels between the intervention and the control group. However, the written diary revealed differences, particularly in the reporting of moderate activity levels. It, therefore, cannot be entirely ruled out that participants in the control group were more diligent and detailed in keeping the diary, especially when referring to moderate activities, compared to the intervention group, which might have focused more on the intensive training and its reporting. One might further argue that a difference in the cognitive state of the control group at baseline might lead to better activity reporting in the diaries in the control group. However, our findings did not show an association between cognitive state, as represented by the MoCA score, and compliance in keeping the diary in our sample.

We further analyzed the association between physical activity and CRF, represented by VO₂max. As expected, high activity levels derived from the Fitbit were positively associated with VO₂max from baseline, but not with ΔVO₂max. Interestingly, participants with lower baseline CRF levels who successfully increased their physical activity levels, as indicated by the ΔPAQ50+_T2/3−1 score, benefited the most from our multicomponent intervention program. They increased their VO₂max, which had a positive effect on various AD progress parameters, including a positive mediating effect on regional brain structure and cognitive performance. Therefore, it can be assumed that, although there is a detectable positive influence of increased physical activity on CRF development, the individual baseline activity level must be considered separately concerning its influence on AD progression.^14,70 CRF improvement itself, reflected by an increase in VO₂max, was positively associated with an affiliation to the intervention group validating the effectiveness of our structured training concept for improving fitness in a sample of elderly and cognitively impaired participants. This aligns with previous findings emphasizing that increased physical activity is a crucial element to improve CRF, yet conveying its greatest effect through organized and regular training sessions of especially vigorous intensity,^71,72 which, regardless of the average activity level, was only ensured in our intervention group. It could therefore be inferred that a specific level of activity at a specific intensity level is regularly required to produce benefits. This is underlined by the results of the intervention group, which showed a higher rate of activity on a vigorous level, while the control group engaged at a more moderate activity level, not revealing a significant impact. A positive trend was also reported regarding the development of postural stability, a domain that was actively trained in our program, which should be recommended even apart from its sheer effect on AD progression as an important preventive measure to reduce the risk of falls.⁷³ On the other hand, there was no remarkable response in other fitness parameters that were not specifically addressed in our intervention program, such as handgrip strength and the 6-MWT distance. A beneficial effect of the COVID-19 protocol changes on development of handgrip strength can probably be explained by training adjustments to home-based and space-restricted training forms with a focus on single muscle groups during the lockdown.

We further analyzed how VO₂max improvement and average activity levels influenced AD outcome parameters. A modifying beneficial effect of CRF improvement was only observed in the intervention group that successfully increased VO₂max over the course of the intervention period, supporting previous findings that demonstrated an effect of exercise-induced CRF increase on cognitive performance and AD-affected brain regions of the temporal lobe. This effect could be related to a combination of reduced brain inflammation, higher brain reserve, and reduced brain damage in terms of reduced vascular, neurotoxic, or oxidative stress,⁷⁴ leading to a reduced cellular loss and, therefore, maintained brain volumes and cortical thickness. Additionally, we were able to validate the previously described beneficial effects of VO₂max improvement on cognitive performance,⁷⁵ as represented in the MoCA and executive functions composite score in the intervention group. Further negative associations with included baseline levels as non-specific progress markers for neuropsychiatric and cognitive parameters suggest that clinically more severely affected individuals of our sample were able to benefit more from the intervention in combination with their thereby achieved fitness improvement.^64,76 However, there was no clear beneficial effect of CRF improvement on neuropsychiatric symptoms which has been reported in previous studies.^77,78 In contrast, we found higher activity levels associated with increased quality of life for caregivers, highlighting another quality sign of multicomponent intervention concepts, by effectively relieving caregivers and offering space for discourse of likewise affected people and their relatives, even if this effect was not significantly reflected by the participants themselves.⁷⁹ Not only CRF, but also high physical activity led to a beneficial effect on the executive functions in the overall sample, pointing out that not only CRF increase but also physical activity on a regular base itself contributes to a positive development of cognitive performance in AD. This beneficial effect on executive functions is also reflected on the brain level, where our brain voxel-wise permutation analysis points at positive volume effects in frontal brain regions in trend. Recent research has pointed to an assumed predominant effect of multicomponent exercise on executive functions which is also in line with our cerebral findings of effects in the frontal lobe.⁸⁰ There was no beneficial effect of physical activity and CRF on vascular lesions, which one would assume given the positive mediating influence of these factors on the cardiovascular risk profile.⁶² However, it might be the case that the intervention period of six months is too short to observe large effects in vascular changes on the cerebral level but would rather be observed on long-term outcomes.

In conclusion, we demonstrated that our multicomponent exercise intervention design can be an effective method to improve CRF in a sample of people affected with AD, especially for those with previously low fitness levels after six months of intervention. VO₂max improvement was associated with advantageous developments in cognitive performance and a positive mediating effect on brain structure in AD-relevant regions of the temporal lobe in the intervention group, while negative associations with cognitive baseline levels suggest a higher benefit for people with more cognitive impairment. Furthermore, we found high levels of physical activity to be associated with increased cognitive performance, sustained cortical thickness in the temporal lobe area, and higher quality of life in caregivers of AD-affected individuals, marking another crucial domain to reduce the disease burden through physical interventions.

Even though there was a randomization process before group allocation, group differences occurring at baseline in single assessment parameters are inevitable, and our intervention group may have been more strongly affected by AD pathology based on their initial CSF status at the time of diagnosis and cognitively more impaired and with a lower cognitive reserve compared to the control group, pointing to differences in the MoCA score, the executive function, and the education level at baseline. This was taken into account in our statistical analyses, and we excluded a significant influence of these potential confounding factors on the outcome parameters. Still, this aspect has to be considered when looking at absolute outcome parameters beyond the change from pre- to post-intervention state alone. As suggested by previous publications, it is presumably challenging to account for all potential confounding factors during a temporary intervention, which may influence the outcomes and hinder direct comparisons between groups. In addition, we encountered minor fluctuations in protocol adherence due to medical issues related not only to the underlying disease but also expected in interventions involving an older population. Furthermore, necessary adjustments to the protocol were required during the COVID-19 lockdown. Nevertheless, a strong point of the present sample is the very detailed clinical characterization, which is often lacking in previously reported studies.¹¹ Considering the current body of research and the mentioned limitations, which are common in interventional research, our study demonstrates the positive influence of both physical activity and fitness improvement on AD pathology. The underlying effects on the cerebral level should be considered multifactorial and will need to be the subject of preclinical and clinical intensive research in the future.⁸¹ Our study provides an important gain in knowledge on the rarely considered relationship between fitness and physical activity, both of which should be taken into account according to our study results. The present study has a limited sample size, which was based on initial power analyses related to our primary outcome parameter (see also⁶²). However, it can support the design of future intervention studies, which should also include well-characterized samples, as in the present study, and additionally incorporate larger sample sizes, longer intervention periods, and long-term follow-ups to enable subgroup analyses and compensate for group differences that might influence the outcomes, particularly concerning baseline cognition, fitness, activity levels, and differences in AD pathology, which might influence the distinct response to intervention effects.

According to our findings, which show a positively mediating influence of cardiovascular fitness and physical activity on several disease outcome parameters in AD, regular, structured exercise and activity increase in people affected by AD should be aimed at, both for primary and secondary prevention. Additionally, the intensity of physical activity must always be considered. Connection to professional, multicomponent training programs should be established early in the disease, but it is also highly effective for people with advanced disease manifestations, as shown in our study, and is therefore never too late.

Supplemental Material

sj-docx-1-alz-10.1177_13872877241303764 - Supplemental material for Modulating effects of fitness and physical activity on Alzheimer's disease: Implications from a six-month randomized controlled sports intervention

Supplemental material, sj-docx-1-alz-10.1177_13872877241303764 for Modulating effects of fitness and physical activity on Alzheimer's disease: Implications from a six-month randomized controlled sports intervention by Shari David, Ana S Costa, Christian Hohenfeld, Sandro Romanzetti, Shahram Mirzazade, Jennifer Pahl, Luisa Haberl, Kai M Schneider, Axel Kilders, Thomas Eggermann, Christian Trautwein, Frank Hildebrand, Jörg B Schulz, Kathrin Reetz and Alexa Haeger in Journal of Alzheimer's Disease

Footnotes

Acknowledgments

The authors thank the team from the physiotherapy department at RWTH Aachen University Hospital, especially Rudolph Schifflers and David Uebis, as well as Gökhan Giray and Rebecca Schulze from the MedAix physiotherapy practice in Simmerath for their assistance during the intervention, and Michael Beulertz for his assistance in data acquisition. The authors also thank Prof. Ingo Kurth from the Department of Human Genetics at RWTH Aachen University for his support in ApoE-analysis. This work was supported by the Brain Imaging Facility of the Interdisciplinary Center for Clinical Research (IZKF) Aachen within the Faculty of Medicine at RWTH Aachen University.

ORCID iDs

Shari David

Kathrin Reetz

Alexa Haeger

Author contributions

Shari David (Data curation; Formal analysis; Investigation; Supervision; Visualization; Writing – original draft; Writing – review & editing); Ana S Costa (Conceptualization; Investigation; Methodology; Writing – review & editing); Christian Hohenfeld (Formal analysis; Software; Writing – review & editing); Sandro Romanzetti (Data curation; Writing – review & editing); Shahram Mirzazade (Investigation; Writing – review & editing); Jennifer Pahl (Investigation; Writing – review & editing); Luisa Haberl (Investigation; Writing – review & editing); Kai M Schneider (Resources; Writing – review & editing); Axel Kilders (Investigation; Writing – review & editing); Thomas Eggermann (Resources; Writing – review & editing); Christian Trautwein (Resources; Writing – review & editing); Frank Hildebrand (Resources; Writing – review & editing); Jörg B Schulz (Conceptualization; Resources; Writing – review & editing); Kathrin Reetz (Conceptualization; Investigation; Project administration; Supervision; Writing – review & editing); Alexa Haeger (Conceptualization; Data curation; Formal analysis; Funding acquisition; Investigation; Methodology; Project administration; Supervision; Validation; Writing – review & editing).

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The research project was funded by RWTH-Startup granted to AH (StUpPD_329-18) and further supported by the Stiftung Universitätsmedizin Aachen. SD and LH were funded by the IRTG Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) – 269953372/GRK2150. This research project was supported by the Rotation Program and Short Term Program of the Faculty of Medicine RWTH Aachen University.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data availability

The data supporting the findings of this study are available on reasonable request from the corresponding author.

Supplemental material

Supplemental material for this article is available online.

References

Prince

Wimo

Guerchet

, et al. World Alzheimer Report 2015. The Global Impact of Dementia: An analysis of prevalence, incidence, cost and trends. London, UK: Alzheimer’s Disease International, 2015.

Gustavsson

Norton

Fast

, et al. Global estimates on the number of persons across the Alzheimer’s disease continuum. Alzheimers Dement 2023; 19: 658–670.

Crous-Bou

Minguillón

Gramunt

, et al. Alzheimer’s disease prevention: from risk factors to early intervention. Alzheimers Res Ther 2017; 9: 71.

Ngandu

Lehtisalo

Solomon

, et al. A 2 year multidomain intervention of diet, exercise, cognitive training, and vascular risk monitoring versus control to prevent cognitive decline in at-risk elderly people (FINGER): a randomised controlled trial. Lancet 2015; 385: 2255–2263.

Morris

Vidoni

Johnson

, et al. Aerobic exercise for Alzheimer’s disease: a randomized controlled pilot trial. PLoS One 2017; 12: e0170547.

Livingston

Huntley

Sommerlad

, et al. Dementia prevention, intervention, and care: 2020 report of the Lancet Commission. Lancet 2020; 396: 413–446.

Hersi

Irvine

Gupta

, et al. Risk factors associated with the onset and progression of Alzheimer’s disease: a systematic review of the evidence. Neurotoxicology 2017; 61: 143–187.

López-Ortiz

Pinto-Fraga

Valenzuela

, et al. Physical exercise and Alzheimer’s disease: effects on pathophysiological molecular pathways of the disease. Int J Mol Sci 2021; 22: 2897.

Thomas

Tarumi

Sheng

, et al. Brain perfusion change in patients with mild cognitive impairment after 12 months of aerobic exercise training. J Alzheimers Dis 2020; 75: 617–631.

10.

Sinha

Berg

Yassa

, et al. Increased dynamic flexibility in the medial temporal lobe network following an exercise intervention mediates generalization of prior learning. Neurobiol Learn Mem 2021; 177: 107340.

11.

Haeger

Costa

Schulz

, et al. Cerebral changes improved by physical activity during cognitive decline: a systematic review on MRI studies. Neuroimage Clin 2019; 23: 101933.

12.

Anderson-Hanley

Barcelos

Zimmerman

, et al. The Aerobic and Cognitive Exercise Study (ACES) for community-dwelling older adults with or at-risk for mild cognitive impairment (MCI): neuropsychological, neurobiological and neuroimaging outcomes of a randomized clinical trial. Front Aging Neurosci 2018; 10: 76.

13.

Suzuki

Shimada

Makizako

, et al. A randomized controlled trial of multicomponent exercise in older adults with mild cognitive impairment. PLoS One 2013; 8: e61483.

14.

Dougherty

Boots

Lindheimer

, et al. Fitness, independent of physical activity is associated with cerebral blood flow in adults at risk for Alzheimer’s disease. Brain Imaging Behav 2020; 14: 1154–1163.

15.

Banerjee

Peterson

Oliver

, et al. Validating a commercial device for continuous activity measurement in the older adult population for dementia management. Smart Heal 2018; 5–6: 51–62.

16.

Borges-Machado

Ribeiro

Sampaio

, et al. Feasibility and impact of a multicomponent exercise intervention in patients with Alzheimer’s disease: a pilot study. Am J Alzheimers Dis Other Demen 2019; 34: 95–103.

17.

Bull

Al-Ansari

Biddle

, et al. World Health Organization 2020 guidelines on physical activity and sedentary behaviour. Br J Sports Med 2020; 54: 1451–1462.

18.

Coupé

Tourdias

Linck

, et al. Lesionbrain: an online tool for white matter lesion segmentation. In: International Workshop on Patch-based Techniques in Medical Imaging (MICCAI), Grenade, Spain, September 2018, pp.95–103.

19.

Glenn

Madero

Gray

, et al. Engagement with a digital platform for multimodal cognitive assessment and multidomain intervention in a Japanese population: pilot, quasi-experimental, longitudinal study. JMIR mHealth UHealth 2019; 7: e15733.

20.

Haeger

Costa

Romanzetti

, et al. Effect of a multicomponent exercise intervention on brain metabolism: a randomized controlled trial on Alzheimer’s pathology (Dementia-MOVE). Alzheimers Dement (N Y) 2020; 6: e12032.

21.

Jack

Bennett

Blennow

, et al. NIA-AA research framework: toward a biological definition of Alzheimer’s disease. Alzheimers Dement 2018; 14: 535–562.

22.

Haeger

Bottlaender

Lagarde

, et al.

What can 7T sodium MRI tell us about cellular energy depletion and neurotransmission in Alzheimer’s disease?

Alzheimers Dement 2021; 17: 1843–1854.

23.

Haeger

Boumezbeur

Bottlaender

, et al. 3T Sodium MR imaging in Alzheimer’s disease shows stage-dependent sodium increase influenced by age and local brain volume. Neuroimage Clin 2022; 36: 103274.

24.

Abramson

. WINPEPI Updated: computer programs for epidemiologists, and their teaching potential. Epidemiol Perspect Innov 2011; 8: 1.

25.

Luxton

Alison

, et al. Relationship between field walking tests and incremental cycle ergometry in COPD. Respirology 2008; 13: 856–862.

26.

Epley

. Poundage chart. In: Boyd epley workout. Lincoln, NE: Body Enterprises, 2985, 1985, pp.86.

27.

Borg

. Psychophysical bases of perceived exertion. Med Sci Sports Exerc 1982; 14: 377–381.

28.

Wang

Liu

H-Y

Cheng

Y-C

, et al. Exercise dosage in reducing the risk of dementia development: mode, duration, and intensity-a narrative review. Int J Environ Res Public Health 2021; 18: 13331.

29.

Bundeszentrale für gesundheitliche Bildung. Bewegungspackung. https://shop.bzga.de/bewegungspackung-60582387 (accessed March 31, 2024).

30.

Kivipelto

Solomon

Ahtiluoto

, et al. The Finnish geriatric intervention study to prevent cognitive impairment and disability (FINGER): study design and progress. Alzheimers Dement 2013; 9: 657–665.

31.

Albouaini

Egred

Alahmar

, et al. Cardiopulmonary exercise testing and its application. Postgrad Med J 2007; 83: 675–682.

32.

Åstrand

P-O

Ryhming

. A nomogram for calculation of aerobic capacity (physical fitness) from pulse rate during submaximal work. J Appl Physiol 1954; 7: 218–221.

33.

Ries

Echternach

Nof

, et al. Test-retest reliability and minimal detectable change scores for the timed “up & go” test, the six-Minute walk test, and gait speed in people with Alzheimer disease. Phys Ther 2009; 89: 569–579.

34.

Rijk

Roos

Deckx

, et al. Prognostic value of handgrip strength in people aged 60 years and older: a systematic review and meta-analysis. Geriatr Gerontol Int 2016; 16: 5–20.

35.

Raschner

Lembert

Platzer

H-P

, et al. S3-Check - evaluation and generation of normal values of a test for balance ability and postural stability. Sportverletz Sportschaden 2008; 22: 100–105.

36.

Ide

Ota

Hada

, et al. Dynamic balance deficit and the neural network in Alzheimer’s disease and mild cognitive impairment. Gait Posture 2022; 93: 252–258.

37.

Shaughnessy

Hackney

Clark

, et al. A narrative review of handgrip strength and cognitive functioning: bringing a new characteristic to muscle memory. J Alzheimers Dis 2020; 73: 1265–1278.

38.

David

Hohenfeld

Haberl

, et al. Physical activity monitoring in Alzheimer’s disease during sport interventions: a multi-methodological perspective. Front Neurol 2023; 14: 1195694.

39.

Huy

Schneider

. [Instrument for the assessment of middle-aged and older adults’ physical activity: design, eliability and application of the German-PAQ-50+]. Z Gerontol Geriatr 2008; 41: 208–216.

40.

Ainsworth

Haskell

Whitt

, et al. Compendium of physical activities: an update of activity codes and MET intensities. Med Sci Sports Exerc 2000; 32: S498–S504.

41.

Mathiason

Han

, et al. Mechanistic effects of aerobic exercise in Alzheimer’s disease: imaging findings from the pilot FIT-AD trial. Front Aging Neuroscience 2021; 13: 703691.

42.

Van Blarigan

Kenfield

Tantum

, et al. The Fitbit One physical activity tracker in men with prostate cancer: validation study. JMIR Cancer 2017; 3: e5.

43.

von Dôbeln

Astrand

Bergström

. An analysis of age and other factors related to maximal oxygen uptake. J Appl Physiol 1967; 22: 934–938.

44.

Burr

Bredin

SSD

Faktor

, et al. The 6-minute walk test as a predictor of objectively measured aerobic fitness in healthy working-aged adults. Phys Sportsmed 2011; 39: 133–139.

45.

Logsdon

Gibbons

McCurry

, et al. Assessing quality of life in older adults with cognitive impairment. Psychosom Med 2002; 64: 510–519.

46.

Herrmann-Lingen

Buss

Snaith

. Hospital anxiety and depression scale: HADS-D; Deutsche Version; ein Fragebogen zur Erfassung von Angst und Depressivität in der somatischen Medizin; Testdokumentation und Handanweisung. Huber, 2005.

47.

Cummings

Mega

Gray

, et al. The neuropsychiatric inventory: comprehensive assessment of psychopathology in dementia. Neurology 1994; 44: 2308–2314.

48.

Buysse

Reynolds

Monk

, et al. The Pittsburgh sleep quality index: a new instrument for psychiatric practice and research. Psychiatry Res 1989; 28: 193–213.

49.

Folstein

McHugh

. Mini-mental state’. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975; 12: 189–198.

50.

Nasreddine

Phillips

Bédirian

, et al. The Montreal Cognitive Assessment, MoCA: a brief screening tool for mild cognitive impairment. J Am Geriatr Soc 2005; 53: 695–699.

51.

Zimmermann

Fimm

. Testbatterie zur Aufmerksamkeitsprüfung (TAP) Version 1.7. Herzogenrath: PSYTEST Verlag, 2002.

52.

Aste

Neubauer

Horn

. Wechsler Intelligenztest fuer Erwachsene (WIE). Deutschsprachige Bearbeitung und Adaptation des WAIS- III von David Wechsler. Frankfurt: Harcourt Test Services, 2006.

53.

Niemann

Sturm

Thöne-Otto

AIT

, et al. CVLT California Verbal Learning Test. German adaptation. Manual.

54.

Wagner

Helmreich

Dahmen

, et al. Reliability of three alternate forms of the Trail Making Tests A and B. Arch Clin Neuropsychol 2011; 26: 314–321.

55.

Aschenbrenner

Tucha

Lange

. Regensburg Word Fluency Test [Regensburger Wortflüssigkeits-Test (RWT)]. Göttingen: Hogrefe, 2000.

56.

Bäumler

. Farbe-Wort-Interferenztest (FWIT) nach J. R. Stroop.

57.

Schmitz-Hübsch

Fimmers

Rakowicz

, et al. Responsiveness of different rating instruments in spinocerebellar ataxia patients. Neurology 2010; 74: 678–684.

58.

Avants

Tustison

Song

, et al. A reproducible evaluation of ANTs similarity metric performance in brain image registration. Neuroimage 2011; 54: 2033–2044.

59.

Hammers

Allom

Koepp

, et al. Three-dimensional maximum probability atlas of the human brain, with particular reference to the temporal lobe. Hum Brain Mapp 2003; 19: 224–247.

60.

Manjón

Coupé

. Volbrain: an online MRI brain volumetry system. Front Neuroinform 2016; 10: 30.

61.

Romero

Coupé

Tourdias

, et al. LesionBrain: An online tool for white matter lesion segmentation.

62.

Haeger

Mangin

J-F

Vignaud

, et al. Imaging the aging brain: study design and baseline findings of the SENIOR cohort. Alzheimers Res Ther 2020; 12: 77.

63.

Open Access Series of Imaging Studies (OASIS). Retrieved from https://www.oasis-brains.org (accessed May 31, 2024).

64.

Tokgöz

Claassen

JAHR

. Exercise as potential therapeutic target to modulate Alzheimer’s disease pathology in APOE ε4 carriers: a systematic review. Cardiol Ther 2021; 10: 67–88.

65.

Aslanyan

Ortega

Fenton

, et al. Protective effects of sleep duration and physical activity on cognitive performance are influenced by β-amyloid and brain volume but not tau burden among cognitively unimpaired older adults. Neuroimage Clin 2023; 39: 103460.

66.

de Frutos-Lucas

Frost

Erickson

, et al. Does APOE genotype moderate the relationship between physical activity, brain health and dementia risk? A systematic review. Ageing Res Rev 2020; 64: 101173.

67.

Eklund

Dufort

Villani

, et al. BROCCOLI: software for fast fMRI analysis on many-core CPUs and GPUs. Front Neuroinform 2014; 8: 24.

68.

Germini

Noronha

Borg Debono

, et al. Accuracy and acceptability of wrist-wearable activity-tracking devices: systematic review of the literature. J Med Internet Res 2022; 24: e30791.

69.

Fuller

Colwell

Low

, et al. Reliability and validity of commercially available wearable devices for measuring steps, energy expenditure, and heart rate: systematic review. JMIR mHealth UHealth 2020; 8: e18694.

70.

Lee

D-C

Sui

Ortega

, et al. Comparisons of leisure-time physical activity and cardiorespiratory fitness as predictors of all-cause mortality in men and women. Br J Sports Med 2011; 45: 504–510.

71.

Lakoski

Barlow

Farrell

, et al. Impact of body mass index, physical activity, and other clinical factors on cardiorespiratory fitness (from the Cooper Center longitudinal study). Am J Cardiol 2011; 108: 34–39.

72.

Zhang

Cong

, et al. Effect of physical activity on risk of Alzheimer’s disease: a systematic review and meta-analysis of twenty-nine prospective cohort studies. Ageing Res Rev 2023; 92: 102127.

73.

Dolatabadi

Van Ooteghem

Taati

, et al. Quantitative mobility assessment for fall risk prediction in dementia: a systematic review. Dement Geriatr Cogn Disord 2018; 45: 353–367.

74.

Livingston

Sommerlad

Orgeta

, et al. Dementia prevention, intervention, and care. Lancet 2017; 390: 2673–2734.

75.

Huang

Zhao

, et al. Comparative efficacy of various exercise interventions on cognitive function in patients with mild cognitive impairment or dementia: a systematic review and network meta-analysis. J Sport Heal Sci 2022; 11: 212–223.

76.

Brown

Peiffer

Rainey-Smith

. Exploring the relationship between physical activity, beta-amyloid and tau: a narrative review. Ageing Res Rev 2019; 50: 9–18.

77.

Chen

Kuo

Chang

, et al. Resistance band exercises reduce depression and behavioral problems of wheelchair-bound older adults with dementia: a cluster-randomized controlled trial. J Am Geriatr Soc 2017; 65: 356–363.

78.

Liu

Zhang

, et al. Comparative efficacy of various exercise interventions on depression in older adults with mild cognitive impairment: a systematic review and network meta-analysis. Ageing Res Rev 2023; 91: 102071.

79.

Song

DSF

. Effects of a moderate-intensity aerobic exercise programme on the cognitive function and quality of life of community-dwelling elderly people with mild cognitive impairment: a randomised controlled trial. Int J Nurs Stud 2019; 93: 97–105.

80.

Yan

Guo

, et al. Effect of multicomponent exercise on cognition, physical function and activities of daily life in older adults with dementia or mild cognitive impairment: a systematic review and meta-analysis. Arch Phys Med Rehabil 2023; 104: 2092–2108.

81.

Abdullahi

Wong

. Understanding the mechanisms of disease modifying effects of aerobic exercise in people with Alzheimer’s disease. Ageing Res Rev 2024; 94: 102202.

82.

Storch-Hagenlocher

Reiber

Wildemann

, et al. Liquordiagnostik. In: Wildemann

Reiber

Oschmann

(eds) Neurologische Labordiagnostik. Stuttgart: Georg Thieme Verlag KG, 2006, pp.30–73.

83.

Zettl

Lehmitz

Mix

. Klinische Liquordiagnostik. New York: De Gruyter, 2005.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

2.88 MB