Effects of home-based exercise with or without cognitive training on cognition and mobility in cardiac patients: A randomized clinical trial

Abstract

This randomized controlled trial compared the effects of home-based exercise, with or without cognitive training, on cognition and physical function in individuals aged 50 years and older with stable CVD during the COVID-19 pandemic. 122 patients (67.3 ± 7.9 years, 71% men) with stable CVD (77% coronary heart disease) were randomly assigned (1:1) to (1) Home-based physical exercise alone, or (2) Home-based physical exercise combined with cognitive training. Cognition (executive functions (primary outcome), processing speed, episodic memory, and working memory) and physical functions were assessed remotely at baseline, 3 months, and 6 months. Adjusted mean changes from baseline to 3 months and 6 months for executive functions, episodic memory, working memory, sit-to-stand test, gait speed, and timed up-and-go test were significant in the overall sample (p < 0.05). Furthermore, executive functions, episodic memory, sit-to-stand test, and timed up-and-go performances were significantly improved at 6 months in both groups when analyzed separately although no group differences were observed. Mean exercise dose differed significantly between the 2 groups: 1413 vs 953 METs.min⁻¹ week⁻¹ respectively for the exercise and combined group (p < 0.01). Mean cognitive training duration was 25.6 ± 16.6 min.week⁻¹ for the combined intervention group. Results remained unchanged after accounting for the exercise dose. In adults affected by CVD, a remote combined intervention integrating sequential cognitive and exercise training yields comparable enhancements in executive function, episodic memory, and physical performances compared to exercise training alone. ClinicalTrials.gov: NCT04661189.

Graphical abstract

Supplementary information

The online version contains supplementary material available at 10.1007/s11357-025-01530-y.

Introduction

The COVID-19 pandemic resulted in repeated confinements of the general population to their homes, drastically reducing social interactions and modifying physical activity habits and dietary behaviors [1]. Collateral effects of this unexpected lockdown on cognition, psychosocial well-being, and physical health [2–4], especially in individuals at high risk for cognitive decline, such as older adults [2, 5, 6] and patients with cardiovascular disease (CVD), have been documented [2, 7, 8]. In a 5-year cohort study involving 9304 cognitively healthy older participants aged ≥ 50 years, Li et al. showed that participants with multimorbidity (n = 4375) suffered accelerated declines in global cognition and elevated dementia risk, compared with individuals without morbidity (n = 1818) during the pandemic period [2]. These results were confirmed by another cohort study [9], which observed significant worsening of executive function and working memory in the first year of the pandemic. Cognitive decline in executive functions was significantly associated with reduced physical activity while working memory impairment was significantly associated with increased alcohol use, depression, history of COVID-19 infection, and loneliness [9]. Reduced physical activity continued to affect executive function in the second year of the pandemic [9].

Aerobic and resistance training, as non-pharmacological treatments, play a crucial role in comprehensive cardiac rehabilitation (CR) programs, which are highly recommended (class I, level A) for their established clinical benefits, including cognition, specifically executive function [10–13]. In this context, during the pandemic, remote interventions such as telerehabilitation with remote home-based exercise were introduced, in an effort to maintain cardiovascular, physical and cognitive health in cardiac patients [10, 14]. Exercise under remote supervision has been shown to be safe for people with stable CVD and has the same positive effects on health outcomes compared to center-based CR [10, 15, 16]. Additionally, cognitive training is known to enhance cognitive abilities in both heart surgery patients and older individuals [17, 18]. Exercise and cognitive training separately have demonstrated enhancements in cognition and physical functions and a potential synergistic effect on cognition in healthy older adults has already been documented [19–22]. Studies have already shown that combining exercise and cognitive training in a laboratory setting may further boost cognition in healthy older adults and patients with mild cognitive impairment compared to exercise training alone [19, 22, 23]. Nevertheless, it remains to be explored whether a synergistic impact exists in older individuals with CVD, particularly in the context of home-based training without direct supervision and in the context of remote evaluations.

The COVEPICARDIO study examined the effects of home-based exercise, with or without cognitive training, on cognition and physical function in people aged 50 years and over with stable CVD during the COVID-19 pandemic. As the first hypothesis, we assumed that cognition would be different after a six-month period between individuals with stable CVD randomly assigned to a program combining home-based physical exercise and cognitive training and those assigned to home-based physical exercise only, expecting larger improvements in the former group. For the second hypothesis, we investigated the impact of the amount of physical exercise performed over 6 months on cognition, anticipating greater cognitive improvements for participants who engaged in higher dose of physical exercise.

Methods

Study design

The COVEPICARDIO study was a prospective, evaluator-blinded, randomized trial with two parallel intervention arms (1:1): 1/ Physical exercise (EXE): remote monitoring and coaching of home-based physical exercise, 2/ Physical exercise and cognitive training (COMB): remote monitoring and coaching of home-based physical exercise and cognitive training. The intervention lasted six months. Three testing periods, each lasting 7 to 10 days, were conducted remotely at baseline (T0), then after three months (T3), and six months (T6) of intervention, utilizing videoconference supervision and online questionnaires. The protocol of the COVEPICARDIO study was prospectively registered in ClinicalTrials.gov (NCT04661189) and was published in 2021 [24].

Participants

A total of 122 participants aged 50 years and older with stable CVD were recruited and randomized between May 2020 and February 2021. Participants were recruited by the Montreal Heart Institute’s (MHI) physicians (N.B., J.I.G., A.N., M.J.).

Eligibility criteria

Individuals were enrolled in the study if they had access to a tablet (i.e., iPad or Android) or a computer with an internet connection. In addition, participants had to manifest an absence of significant cognitive impairment (i.e., a score of 19/23 or higher on the telephone version of the Mini-Mental State Examination, MMSE [24]), absence of non-cardiopulmonary limitation (e.g. severe arthritis) to exercise or severe exercise intolerance, and absence of respiratory disease (e.g., severe asthma, severe coronavirus disease 2019 related symptoms, COPD) as documented with a medical questionnaire. Participants were invited to take part in the study if they presented CVD such as stable coronary artery disease, stable chronic heart failure, corrected valvular heart disease, or atrial fibrillation, with a low-risk profile, as determined by the physician, and without contraindications to exercise training.

Participants provided written informed consent by email before starting the study. The COVEPICARDIO was conducted in compliance with the International Conference on Harmonization Good Clinical Practice (ICH-GCP) and in accordance with the Helsinki Declaration. The study protocol was approved by the Montreal Heart Institute’s (MHI) Research Ethics Board (FWA00003235; research project: ICM 2020–2785).

Methods

The participants completed 4 pre-testing sessions (baseline), 4 mid-testing sessions (at 3 months) and 4 post-testing sessions (at 6 months). All assessments, as well as the weekly training follow-ups, were done remotely during the COVID-19 pandemic, via videoconference or using self-reported online questionnaires. Before starting the assessments, a videoconference session was scheduled with each participant, which consisted of a technology tutorial to ensure that internet connectivity and the necessary tools for the tests were prepared and worked correctly. Demographics, medical information (i.e., chronic diseases, comorbidities, and medications) and the Physical Activity Readiness Questionnaire (PARQ +) were also completed. Then, participants performed a neuropsychological assessment to assess general cognitive functioning, executive functions, processing speed, and episodic memory. Additionally, different physical functions were assessed (balance, lower limb muscle strength, mobility and predicted cardiorespiratory fitness).

Interventions

Participants received a group-specific training guide according to their randomization. Only the participants in the combined training group received the cognitive training guide in addition to the physical exercise guide, which everyone received. This training guide contained only recommendations about training sessions' type, intensity, and frequency. Participants then attended an introductory session over the phone with their kinesiologist and another one with the cognitive training supervisor (if applicable).

Physical exercise

Participants were encouraged to complete home-based physical exercise at least 5 times a week using video capsules available via Facebook or YouTube, created by the MHI EPIC Center’s team of kinesiologists. Aerobic, resistance, flexibility, and/or balance exercises were proposed in these videos and did not require any equipment. Each video lasted approximately 15 min with a 3-min warm-up, followed by a 10-min training, and a 2-min cool-down period. Kinesiologists provided guidelines to allow participants to adapt the exercise to their physical capacities. Participants could also follow other video or web-based training programs or could engage in outdoor activities (e.g., walking or cycling). To track their adherence, the duration, intensity (Borg Rating of Perceived Exertion (RPE) graduated from 0 to 10), and the type of activity performed during each session were reported by the participant in a journal throughout the intervention. Participants transmitted this information to their kinesiologists during their weekly follow-up calls. Participants’ training volume was calculated based on the type of each activity, its intensity and its frequency. At the end of the study, to precisely establish the dose of physical exercise (weekly amount of physical exercise measured in METs.min⁻¹.week⁻¹), a panel of five experts (FB, LB, MG, EGD and TV) convened to categorize each activity according to the Compendium of Physical Activities (Supplementary material Table S1) [24].

Cognitive training

The combined intervention included home-based cognitive training in addition to physical exercise. Participants were asked to complete their home-based cognitive training prior to exercise training or at least 3 h post-exercise. Validated computerized cognitive training tasks [25], as well as video capsules were provided. Participants were encouraged to complete a minimum of 3 sessions per week (15–20 min/session, maximum 1 session/day) of the following type: a) dual-task training, during which participants had to share attention between two concurrent tasks, b) inhibition training, during which participants had to refrain from giving an automatic response (i.e., modified Stroop task), and c) working memory training, in which participants had to maintain and update information in working memory to recall items presented earlier (i.e., N-Back task). Task difficulty increased gradually over the participant's cognitive training to avoid task automatization and to maintain stimulation [26]. Live visual feedback was provided during training, as well as a histogram of daily performances to encourage improvement. Once per week, participants were asked to complete a strategy-based training, via video capsules. This training is an adapted version of the MEMO + program, which has been validated previously [27]. Participants learned different mnemonics (e.g., face-name association, visual imaging), and learned about age-related changes in memory. In addition, some video capsules were targeted to help participants develop strategies to cope with multiple aspects of their daily life, such as sleep, anxiety, and nutrition. To track cognitive training adherence, participants were asked to complete a follow-up agenda and to indicate days and times when they completed cognitive training sessions. In addition, objective cognitive training data for the computerized training tasks was collected on the application’s server (e.g., duration, number of training sessions).

Remote monitoring and coaching

Once a week, kinesiologists called participants to 1) motivate them to stay engaged in their training, 2) ensure the safety of the exercise training sessions, and 3) to collect exercise and cognitive training data reported by participants in their training journal during the previous week.

Measurements and outcomes

Cognitive assessment

Cognitive functioning was assessed by a neuropsychologist or a trained student in neuropsychology via videoconference (i.e. Zoom platform). Paper–pencil tests and computerized tasks were administered in a fixed order. First, during the Rey auditory verbal learning test (RAVLT) participants were asked to remember a list of 15 nouns, which was read 5 times. Participants were asked to give back as many words as they remembered after each reading (immediate recall) and after a delay of 20 min (delayed recall). Then, during the Digit Span Forward and the Digit Span Backwards, participants were required to repeat in the same order and then in reverse order a list of numbers just after the assessor had read it out loud (variables of interest: total score). The Mental Alternation test (MAT) included three parts: (A) counting from 1 to 20 while being timed; (B) reciting the alphabet while being timed; and (C) naming as many numbers and letters alternating in order during 30 s (i.e. “1-A-2-B-3-C”…) (variable of interest: total number of correct alternations). Lastly, during the Phonological and Semantic Fluency test from the D-KEFS battery, participants were asked to say as many nouns as possible, starting with the letter “P” in 60 s.

The composite scores were computed as follows: all cognitive scores were first transformed into Z-scores by using the mean and the standard deviation value of all subjects across all time points (combining baseline, T3, and T6 of the two groups) (Z-score = (Xi—mean)/SD). The change in Z score is interpreted as a change in distance from the distribution. Z-score change provides a reliable measure of the intervention effect size, which allows comparison of experimental groups using a standardized method. Four composite scores were calculated by averaging the following cognitive subcomponents from the tests above: 1) Executive functions = total number of correct switched (MAT-C) and total number of words letter “P” (Phonological Fluency); 2) Processing speed = Counting time (MAT-A) and Alphabet time (MAT-B); 3) working memory = Digit span forward and digit span backward scores; and 4) Episodic memory = immediate recall (RAVLT), delayed recall (RAVLT), and total words recalled during the 5 learning trials (RAVLT). The Z-score components of the processing speed composite score were multiplied by “−1” so that higher values reflect higher performance, as is the case for all other scores.

Physical functions assessment

Balance, lower limb muscle strength and mobility were assessed remotely via videoconference under the supervision of a research assistant or a kinesiologist. Participants received test instructions and demonstration videos by email beforehand. At the start of the evaluation, the assessor verified the setup to ensure proper testing conditions. Balance was evaluated with the one-leg balance test, where participants stood on one leg for up to 60 s. Four trials were performed, two on each leg. Lower limb strength was evaluated with the Five-Time-Sit-to-Stand (STS) test, performed twice. Mobility was evaluated using the 4-m walking test (4-MWT) and the Timed Up and Go (TUG) test. The assessor ensured the walking track length was measured and checked the standard sitting height chair used for the TUG test. For the TUG, participants started seated, stood up, walked 3 m, turned, and returned to a seated position, with time recorded. For the 4-MWT, participants walked a 4-m track at both normal and fast speeds, with three trials at each speed. The details of the procedures were previously published in the protocol [24]. These tests were previously performed at home with older adults in a remote setting and were found to be reliable with good agreement [28].

The primary outcome was the change in executive functions from baseline to 6 months. The secondary outcomes included the change from baseline to 6 months in other cognitive performance measures (processing speed, working and episodic memory) as well as functional tests.

Randomization and Blinding

Following their inclusion in the study, participants were randomized into one of the two arms: (1) home-based physical training alone or (2) combined physical and cognitive home-based training. A biostatistician from the Montreal Health Innovations Coordinating Centre (MHICC) generated the randomization sequence using blocks of length 4 with a 1:1 allocation ratio.

Assessors performing the evaluations and investigators were blinded to group allocation. Only the kinesiologists executing the weekly follow-up of physical exercise and cognitive training were aware of the assigned intervention. Kinesiologists did not take part in any assessment.

Data analysis

Sample size calculation

A detailed sample size calculation is presented in the published protocol [24]. An independent team of biostatisticians from the MHICC performed the sample size calculation based on the comparison between the intervention groups on the primary outcome, i.e., the change from baseline to 6 months in executive functions. A sample size of 49 subjects in each group would have 80% power to detect a difference in means of −0.186 (corresponding to, for example, a difference between physical training alone mean of 0.239 and combined training mean of 0.425, for a standardized effect size of 0.572), assuming that the common standard deviation is 0.325 (slightly larger than in a traditional context because of the home-based administration of neuropsychological tests) using a two-group t-test (0.050 two-sided significance level). As we expected a 20% attrition rate based on our previous exercise studies, we recruited a total of 122 patients with CVD.

Statistical analysis

Baseline characteristics and baseline values for neuropsychological and physical functions assessments were summarized using mean ± standard deviation (SD) for continuous variables and using frequencies and proportions for categorical variables. The mean of the 2 trials for the STS test and the mean of the 3 trials for the TUG and gait speed assessments, performed at spontaneous speed and fast speed, were analyzed. Prior to the analyses, basic assumptions underlying the models were checked. Homogeneity of group variances was assessed by conducting the Brown-Forsythe test on the residuals, normality of residuals by looking at Skewness and Kurtosis values and visual inspection of the distribution with an histogram and parallel slopes assumptions for the treatment regression lines by testing the mean centered baseline value of the test x group interaction at each timepoint. No major violation was observed. Mean changes in cognitive and physical performance from baseline to three and six months were analyzed using a restricted maximum likelihood (REML)-based repeated measures approach including effects of time, intervention group (EXE and COMB), and time x intervention group interaction, with mean centered-baseline value of the test, mean centered-baseline value of the test x time interaction, sex and age as covariates. For cognitive performance, years of education were also included as a covariate. An unstructured covariance structure was used to model the within-patient errors and the Kenward-Roger approximation was used to estimate denominator degrees of freedom. Adjusted means for the changes from baseline, by intervention group and overall, along with their 95% confidence intervals are reported. The between group comparison (EXE vs COMB) of the changes from pre- to post-intervention addressed the added value of cognitive training compared to physical exercise alone (i.e., primary hypothesis). Following the intention-to-treat principle, analyses were conducted on all randomized individuals. The proportion of missing data for baseline executive function was 9.8% and for age and years of education 2.5%. The proportion of missing follow-up data for executive function was 24.6% at 3-months and 32.8% at 6-months. Multiple imputation by fully conditional specification (FCS) with the MI procedure in SAS was used to account for missing data in the executive function analysis. When executive function values were missing, imputation was performed at the test level and executive function was calculated afterwards from the imputed data. The variables included in the imputation model were the z-scores for switching and phonological fluency at baseline, 3 months and 6 months, as well as sex, years of education and age at baseline and 40 imputations were generated. The imputed datasets were analysed jointly by using the MIANALYZE procedure in SAS and the combined results for the adjusted means, along with their 95% confidence interval, are presented. Results for the executive function score are reported both for the observed data and the imputed data. Otherwise, analyses were performed on observed data only. Physical exercise dose between T0 and T6 was summarized using the mean along with a 95% confidence interval and was compared between groups with a Student t-test. Cognitive training dose from T0 to T3 and from T3 to T6 were compared using a paired t-test.

As a secondary analysis, physical exercise dose (in METs.min⁻¹.week⁻¹) was added first as covariate to the initial set of explanatory variables in a linear mixed model accounting for repeated measures to evaluate the effect of the intervention for each group (EXE and COMB), on cognitive functions while adjusting for participants’ mean weekly dose of exercise training during the 6-month intervention. Then, the physical exercise dose was separated at a cut off of 750 METs.min-1.week-1 corresponding to the recommended level [29]. Participants performing a lower or higher dose of exercise during the intervention in both the COMB and EXE groups were compared to assess changes in cognitive performance. All statistical tests used a two-tailed significance level of 0.05. Statistical analyses were performed with the use of SAS software, version 9.4 (SAS Institute) and STATA SE 15 (StataCorp LP, College Station, Texas, USA).

Results

Baseline clinical characteristics of the 122 randomized participants (67.3 ± 7.9 years, 71% men, BMI was 28.9 ± 5.6 kg.m⁻², 94 (77%) coronary heart disease participants) are summarized in Table 1 and Table 2. Ninety-three participants completed the study at 6 months (EXE n = 44 and COMB n = 49) (Supplementary Material Figure S1: CONSORT flowchart).

Table 1.

Baseline characteristics of participants

	Overall n = 122			Exercise n = 60			Combined n = 62
	Mean	±	SD	Mean	±	SD	Mean	±	SD
Age (years)	67.30	±	7.93	66.62	±	7.80	67.95	±	8.07
Sex (n men, %)	87 (71.3%)			42 (70.0%)			45 (72.6%)
Education (years)	16.20	±	3.63	16.61	±	3.83	15.80	±	3.40
Mini-Mental State Examination	21.91	±	1.07	21.77	±	1.04	22.05	±	1.08
Height (m)	1.70	±	0.09	1.71	±	0.10	1.70	±	0.08
Weight (kg)	83.80	±	16.41	87.00	±	18.23	80.82	±	14.02
Body mass index (kg.m−2)	28.89	±	5.56	29.63	±	6.22	28.21	±	4.83
Systolic blood pressure (mmHg)	123.94	±	15.27	123.67	±	15.41	124.19	±	15.32
Diastolic blood pressure (mmHg)	75.04	±	9.57	75.56	±	10.54	74.56	±	8.70
VO2max (mL.kg−1.min−1) (Matthew's quest.)	27.16	±	7.93	27.15	±	8.44	27.18	±	7.49
PASE score	126.71	±	63.44	126.32	±	69.46	127.07	±	58.06
Cardiovascular risk factors
Tobacco (n, %)	8 (6.5%)			2 (3.3%)			6 (9.7%)
OH drinks ≥ 2 per week (n, %)	63 (51.6%)			29 (48.3%)			34 (54.8%)
High blood pressure (n, %)	61 (50%)			35 (58.3%)			26 (41.9%)
Dyslipidemia (n, %)	60 (49.2%)			28 (46.7%)			32 (51.6%)
Hyperglycemia (n, %)	26 (21.3%)			14 (23.3%)			12 (19.3%)
Overweight: BMI ≥ 25 and < 30 kg.m−2	47 (38.5%)			18 (30.0%)			29 (46.8%)
Obesity I: BMI ≥ 30 and < 35 kg.m−2	20 (16.4%)			10 (16.7%)			10 (16.1%)
Obesity II: BMI: ≥ 35 and < 40 kg.m−2	12 (9.8%)			7 (11.7%)			5 (8.1%)
Obesity III: BMI ≥ 40 kg.m−2	6 (4.9%)			5 (8.3%)			1 (1.6%)
Family history of CVD (n, %)	15 (12.3%)			9 (15%)			6 (9.7%)
Cardiovascular disease history
CHD (n, %)	94 (77.0%)			44 (73.3%)			50 (80.6%)
Arrhythmia (n, %)	26 (21.3%)			12 (20.0%)			14 (22.6%)
Valvular heart disease (n, %)	14 (11.5%)			9 (15.0%)			5 (8.0%)
HFrEF (n, %)	9 (7.4%)			5 (8.3%)			4 (6.5%)
Stroke (n, %)	11 (9.0%)			5 (8.3%)			6 (9.7%)
Congenital heart diseases (n, %)	1 (0.8%)			0			1 (1.6%)
Pharmacological treatment
Number of drugs/day: from 1 to 2	9 (7.4%)			4 (6.7%)			5 (8%)
Number of drugs/day: from 3 to 4	15 (12.3%)			6 (10%)			9 (14.5%)
Number of drugs/day: 5 or more	16 (13.1%)			9 (15%)			7 (11.3%)
Attention deficit hyperactivity disorder	1 (0.8%)			0 (0%)			1 (1.6%)
Anxiolytique	8 (6.5%)			3 (5.0%)			5 (8%)
Sleep disorders treatments	3 (2.4%)			2 (3.3%)			1 (1.6%)
Anti inflammatory	4 (3.3%)			2 (3.3%)			2 (3.2%)
Antidiabetic	4 (3.3%)			1 (1.7%)			3 (4.8%)
Antithrombotic	25 (20.5%)			13 (21.7%)			12 (19.3%)
Beta blockers	19 (15.6%)			11 (18.3%)			8 (12.9%)
Blood pressure control	25 (20.5%)			16 (26.7%)			9 (14.5%)
Cholesterol	26 (21.3%)			13 (21.7%)			13 (20.9%)
Diuretics	7 (5.7%)			4 (6.7%)			3 (4.8%)
Hormonotherapy	7 (5.7%)			3 (5.0%)			4 (6.5%)
Pain killer	5 (4%)			1 (1.7%)			4 (6.5%)
Proton pump inhibitors	11 (9%)			4 (6.7%)			7 (11.3%)

BMI body mass index; CHD coronary heart disease; CVD cardiovascular disease; HFrEF heart failure with reduced ejection fraction; PASE physical activity scale for the elderly; OH alcohol

Table 2.

Baseline value for neuropsychological and functional tests

	Overall (n = 122)			Exercise (n = 60)			Combined (n = 62)
Neuropsychological tests	Mean	±	SD	Mean	±	SD	Mean	±	SD
Executive function	−0.124	±	0.825	−0.014	±	0.813	−0.233	±	0.830
Switching (total switches MAT)	25.83	±	6.66	26.21	±	6.81	25.44	±	6.55
Phonological fluency (P letter)	14.21	±	4.12	15.05	±	3.63	13.40	±	4.42
Processing speed	−0.084	±	0.971	−0.133	±	1.040	−0.035	±	0.905
TMT alphabet (sec)	11.89	±	5.85	12.13	±	5.94	11.66	±	5.81
TMT counting (sec)	8.81	±	3.94	9.03	±	4.35	8.60	±	3.51
Episodic memory	−0.304	±	0.954	−0.202	±	0.912	−0.404	±	0.991
Immediate recall (RAVLT)	9.93	±	3.30	10.12	±	3.24	9.75	±	3.38
Delayed recall (RAVLT)	9.50	±	3.67	10.00	±	3.49	9.00	±	3.80
Total 5 trials (RAVLT)	45.62	±	10.38	46.64	±	10.14	44.61	±	10.61
Working memory	−0.140	±	0.831	−0.001	±	0.891	−0.277	±	0.751
Digit Span Forward Score	9.56	±	2.23	9.81	±	2.33	9.32	±	2.12
Digit Span Backward score	8.62	±	2.12	8.98	±	2.37	8.27	±	1.78
Functional tests	Mean	±	SD	Mean	±	SD	Mean	±	SD
TUG usual speed (sec)	8.60	±	1.66	8.65	±	1.87	8.55	±	1.44
TUG fast speed (sec)	6.65	±	1.27	6.65	±	1.33	6.65	±	1.21
One leg balance (sec)	44.4	±	20.1	46.5	±	19.6	42.4	±	20.5
STS test (sec)	11.59	±	3.21	11.93	±	3.81	11.24	±	2.46
Walking test (4 m) usual speed (m/s)	0.987	±	0.194	0.989	±	0.206	0.984	±	0.183
Walking test (4 m) fast speed (m/s)	1.398	±	0.300	1.416	±	0.325	1.379	±	0.273

MAT mental alternation test; RAVLT rey auditory verbal learning test; STS sit-to-stand test; TUG timed-up-and-go test; TM trail making test

The adjusted mean changes from baseline to 3 and 6 months for neuropsychological tests are presented in Table 3, and Supplementary Material Table S2 while the raw changes are illustrated in the Fig. 1. Based on the imputed data, in the overall sample, the adjusted means for the change in executive function performance were significantly improved from baseline to 3 months (+ 0.164; 95%CI 0.041 to 0.286; p = 0.009) and baseline to 6 months (+ 0.267; 95%CI 0.141 to 0.392; p < 0.001). Furthermore, executive functions improved significantly in both groups at T6 (+ 0.224; 95%CI 0.047 to 0.401 p = 0.013 and + 0.310; 95%CI 0.149 to 0.471 p < 0.001, respectively for the EXE and COMB groups). However, no between-group differences was observed. Similar results were obtained from the analysis conducted on observed data for executive function.

Table 3.

Adjusted changes from baseline to 3 and 6 months for neuropsychological tests

Neuropsychological tests	Overall				Exercise				Combined				Between group
Multiple Imputation analysis	Estimate	(95% CI)		p	Estimate	(95% CI)		p	Estimate	(95% CI)		p	P-value
Executive function from T0 to T3	0.164	0.041	0.286	0.009	0.119	−0.051	0.290	0.169	0.208	0.041	0.375	0.015	0.453
Executive function from T0 to T6	0.267	0.141	0.392	< 0.001	0.224	0.047	0.401	0.013	0.310	0.149	0.471	< 0.001	0.455
Observed data
Executive function from T0 to T3	0.154	0.026	0.281	0.019	0.105	−0.074	0.284	0.248	0.202	0.027	0.378	0.025	0.434
Executive function from T0 to T6	0.277	0.156	0.398	< 0.001	0.215	0.041	0.389	0.016	0.339	0.175	0.502	< 0.001	0.297
Processing speed from T0 to T3	0.034	−0.078	0.145	0.552	−0.007	−0.161	0.148	0.930	0.074	−0.080	0.228	0.341	0.450
Processing speed from T0 to T6	−0.025	−0.173	0.123	0.740	−0.056	−0.265	0.154	0.600	0.006	−0.198	0.210	0.954	0.673
Episodic memory from T0 to T3	0.338	0.213	0.462	< 0.001	0.253	0.081	0.426	0.004	0.422	0.252	0.592	< 0.001	0.156
Episodic memory from T0 to T6	0.608	0.470	0.746	< 0.001	0.599	0.406	0.793	< 0.001	0.617	0.430	0.805	< 0.001	0.893
Working memory from T0 to T3	0.161	0.031	0.292	0.016	0.117	−0.065	0.299	0.207	0.206	0.027	0.385	0.025	0.477
Working memory from T0 to T6	0.171	0.027	0.315	0.020	0.132	−0.072	0.335	0.203	0.211	0.014	0.408	0.036	0.574

For observed data: 87 subjects were included in the analysis of Executive function. 100 subjects were included in the analysis of Processing speed. 98 subjects were included in the analysis of Episodic memory and 100 subjects were included in the analysis of working memory

Fig. 1.

Changes from baseline to 3 and 6 months for Executive function and Processing speed Z scores. The violin box were created from raw data. The continuous red line is the median, the black dotted line is the interquartile range. Statistical tests are based on the linear mixed model. Stars indicate that the change from baseline is significant. *p < 0.05; **p < 0.01; ***p < 0.001

In the overall sample, the scores in episodic memory and working memory both significantly improved from baseline to 3 and 6 months (p < 0.05). Both groups showed significant improvement in episodic memory at 3 and 6 months (p < 0.01) with no difference between the two groups (between groups at 6 months, p = 0.893). However, only the COMB group showed an enhancement in working memory performances at both time points (p < 0.05), unlike the exercise-only group, but the group difference was not statistically significant at 6 months (p = 0.574) (Table 3, Fig. 2, Supplementary Material Table S2).

Fig. 2.

Changes from baseline to 3 and 6 months for Episodic memory and Working memory Z scores. The violin box were created from raw data. The continuous red line is the median, the black dotted line is the interquartile range. Statistical tests are based on the linear mixed model. Stars indicate that the change from baseline is significant. *p < 0.05; **p < 0.01; ***p < 0.001

The adjusted mean changes from baseline to 3 and 6 months for functional tests are presented in Table 4. Briefly, the performances for the TUG at spontaneous pace (but not maximal speed), and STS test increased significantly for the two groups at 6 months (p < 0.05). However, the improvement in physical performances at 6 months was not significantly different between groups (Table 4 and Supplementary Material Figure S2).

Table 4.

Adjusted changes from baseline to 3 and 6 months for functional tests

	Overall				Exercise				Combined
Functionnal tests (observed data)	Estimate	(95% CI)		p	Estimate	(95% CI)		p	Estimate	(95% CI)		p	Between group P-value
TUG usual speed (sec) from T0 to T3	−0.247	−0.469	−0.025	0.030	−0.208	−0.517	0.102	0.186	−0.286	−0.584	0.012	0.060	0.708
TUG usual speed (sec) from T0 to T6	−0.448	−0.691	−0.205	< 0.001	−0.377	−0.718	−0.037	0.030	−0.518	−0.852	−0.184	0.003	0.551
TUG fast speed (sec) from T0 to T3	−0.070	−0.247	0.108	0.437	−0.048	−0.294	0.199	0.702	−0.092	−0.330	0.146	0.445	0.790
TUG fast speed (sec) from T0 to T6	−0.099	−0.298	0.100	0.325	−0.022	−0.300	0.256	0.876	−0.177	−0.451	0.098	0.204	0.424
One leg balance (sec) from T0 to T3	−0.649	−2.732	1.435	0.542	0.354	−2.723	3.432	0.821	−0.529	−3.582	2.523	0.734	0.679
One leg balance (sec) from T0 to T6	1.965	−0.228	4.157	0.079	1.585	−1.615	4.786	0.332	3.437	0.241	6.633	0.035	0.410
STS test (sec) from T0 to T3	−0.664	−1.058	−0.269	0.001	−0.746	−1.293	−0.199	0.008	−0.581	−1.119	−0.043	0.035	0.661
STS test (sec) from T0 to T6	−0.760	−1.130	−0.389	< 0.001	−0.781	−1.294	−0.268	0.003	−0.738	−1.249	−0.228	0.005	0.905
Walking, usual speed from T0 to T3	−0.311	−0.417	−0.205	< 0.001	−0.259	−0.405	−0.114	0.001	−0.363	−0.506	−0.219	< 0.001	0.298
Walking, usual speed from T0 to T6	−0.257	−0.390	−0.124	< 0.001	−0.184	−0.368	0.001	0.051	−0.331	−0.516	−0.146	0.001	0.259
Walking, fast speed from T0 to T3	−0.166	−0.261	−0.071	0.001	−0.131	−0.261	−0.001	0.048	−0.201	−0.329	−0.072	0.003	0.430
Walking, fast speed from T0 to T6	−0.094	−0.213	0.025	0.120	−0.065	−0.230	0.100	0.434	−0.123	−0.288	0.042	0.142	0.615

STS sit-to-stand test; TUG timed-up-and-go test. For observed data: 99 subjects were included in the analysis of the TUG (usual and fast speed) and the STS test. 98 subjects were included in the analysis of the walking test (usual and fast speed)

Physical exercise dose statistically differed between groups: 1413 METs.min⁻¹ week⁻¹ (95% CI: 1144; 1681) and 953 METs.min⁻¹ week⁻¹ (95% CI: 743; 1164) respectively for the EXE and COMB group (p = 0.008). The details regarding physical activity dose and type of activities are presented in Supplementary Material Figure S3-S4 and Supplementary Material Table S1. Additionally, the average training duration for cognitive training in the COMB group was 25.6 ± 16.6 min.week⁻¹. Even after accounting for physical exercise dose, the results are similar to those reported previously (Supplementary Tables S3 and Supplementary Material Figure S5).

Discussion

The main objective of this study was to compare the effects of home-based exercise, with or without cognitive training, on cognition and physical function in individuals aged 50 years and over with stable CVD during the COVID-19 pandemic. With the COVID-19 crisis and the confinement measures in place, current recommendations for physical activity and sedentary behavior (see WHO Guidelines [30]) were difficult to meet, specifically in cardiac patients, who already tend to have a sedentary/inactive profile. To our knowledge, this trial is the first to propose fully remote assessments and interventions with exercise and cognitive training in individuals with CVD.

After six months of home-based physical exercise, individuals with stable CVD exhibited improvements in executive function, episodic memory, working memory, and physical functioning (TUG, STS, and walking speed) in both the exercise-alone and combined-training groups. However, we did not confirm our primary hypothesis: participants randomly assigned to the combined intervention did not show larger improvements in cognitive performance or physical functioning compared to those in the exercise-only group.

Similar outcomes observed in both intervention groups highlight the importance of direct supervision in both exercise and cognitive training programs in CVD patients. Home-based interventions lack the personalized coaching and immediate feedback that can be provided in center-based programs. Supervision by professionals ensures correct exercise technique, adherence, and the necessary intensity to achieve benefits. Moreover, supervised programs enhance motivation through social interaction and structured group activities, which are vital for the success of these interventions. The social aspect fosters a sense of community and mutual encouragement among participants, which is often missing in home-based settings. Furthermore, the direct supervision allows for the immediate adjustment of exercises and cognitive tasks to match the individual’s progress and capabilities, optimizing the training benefits.

The cognitive training dose in this study was likely insufficient, with sessions being less frequent and shorter than in other successful interventions. Therefore, the lack of direct supervision and the low dose of cognitive training might have contributed to the similar outcomes observed in both intervention groups. Future research should aim to compare the efficacy of supervised center-based programs with remote home-based interventions to fully understand the impact of supervision and adequate dosing on exercise and cognitive training outcomes.

To our knowledge, only one pilot study in cardiac disease (chronic heart failure, analyzed n = 39, age: 61 ± 10 years, LVEF: 35 ± 15%) evaluated a 3-month home-based program with two interventional groups (a combined group with aerobic exercise and cognitive training and an exercise only training group) and an attentional control usual care group with stretching and flexibility program [31]. Participants in both intervention groups (combined and exercise alone) significantly improved verbal memory and 6-min walking test performances at 3 months compared to the control group [31]. Unfortunately, there is no available data on the comparison between the two interventional groups.

Secondly, although there are no specific studies in individuals with coronary artery disease, a meta-analysis of 28 studies, including 2711 older adults with mild cognitive impairment, showed greater effect sizes in the multidomain intervention combining exercise and cognitive interventions group compared to the single intervention group for global cognition (SMD, 0.41; 95% CI, 0.23–0.59; p < 0.001), executive function (SMD, 0.20; 95% CI, 0.04–0.36; p = 0.01) as well as for memory and verbal fluency [32]. Consistent with the design of the present study, among the 28 studies included in the meta-analysis [32], 7 studies (comprising 928 participants) combined aerobic and resistance exercise with cognitive training. Only three of these studies [33–35] demonstrated that the combined intervention had a greater effect on cognition than exercise alone. In line with the results of the present study, the four other studies have failed to demonstrate the superiority of combined training compared to an exercise-only program. It is important to note that programs vary widely from one study to another in terms of duration, session frequency, and their duration and intensity, making comparisons difficult. Altogether, results are conflicting and further clinical trials are needed with extensive details about the optimal dose of physical exercise and cognitive training [36]. Future research should also aim to compare the efficacy of supervised center-based programs with remote home-based interventions to fully understand the impact of supervision and adequate dosing on exercise and cognitive training outcomes.

The recent SYNERGIC trial confirmed that an intervention combining aerobic, resistance exercise, and cognitive training in older adults with mild cognitive impairment led to greater improvements in the Alzheimer Disease Assessment Scale Cognitive 13 (mean difference: − 1.45 points; 95%CI, − 2.70 to − 0.21 points; p = 0.02; d = 0.39) compared to exercise training alone [22]. The cognitive training program in the SYNERGIC trial consisted of 30 min sessions, three times per week, which is threefold higher compared to the present study. Therefore, an insufficient cognitive training dose could explain the lack of a superior effect of combined training compared to the exercise-only group in the present study.

In a systematic review with meta-analysis (44 studies with 4793 participants, 58% women, 74 ± 6 years old) examining the dose–response relationship between overall and specific types of exercise with cognitive performances in older adults [37], the authors found a non-linear dose–response association between overall exercise and cognition. The estimated minimal exercise dose associated with clinically relevant changes in cognition was 724 METs.min⁻¹.week⁻¹, and doses beyond 1200 METs.min⁻¹.week⁻¹ provided less clear benefits. Interestingly, the authors found that the dose response association was exercise-type dependent with superior effects of resistance exercises over other modalities, and results showed that clinically important effects may occur at lower doses for many types of exercises [37]. The meta-analysis only used studies that evaluated general cognitive function through the Mini-Mental State Examination but not specific cognitive domains such as executive functions. Surprisingly, significant effects were found for walking (from 557 to 851 METs.min⁻¹.week⁻¹) but not for cycling [37]. The authors also found that BMI status significantly influenced the dose–response association between exercise dose and cognition and identified an ascending, non-linear dose–response association for participants with a healthy BMI, whereas an inverted U-shaped dose–response relationship between exercise dose and cognition was observed for overweight/obese individuals [37].

In the present study, both groups had an average dose of physical activity that exceeded the minimum dose found by Gallardo-Gomez et al. [37] (724 METs.min⁻¹.week⁻¹) to be associated with benefits in cognition, but 53% of the participants randomized to the combined group did not achieve this minimal amount of physical activity, while only 26% exceeded this dose in the exercise group. This difference may explain the similar results on cognitive function in both groups. Furthermore, the combined group had a higher proportion of overweight participants than the exercise alone group, which could also mitigate the findings regarding the effects of combined training versus exercise alone.

In a recent systematic review [38] investigating the effects of exercise training on cognition in CVD patients (22 studies, 1125 patients who trained 2–18 weeks at various intensities (60%−80% of maximum heart rate), exercise training improved global cognition, executive function, memory, attention, and processing speed in patients with CVD even with short duration programs but with high frequency (2 weeks, 5 times per week), also suggesting that training dose is important [38].

Regarding cognitive training alone, data suggests that it may be an efficient way to improve specifically targeted cognitive function in aging individuals (i.e., working memory, episodic memory, speed of processing), although few studies exist on the effect of cognitive training interventions in patients with cardiovascular diseases [13, 18, 39]. Results are encouraging in patients receiving coronary artery bypass grafting [18, 39] and patients with chronic heart failure [40]. In the ACTIVE trial [41], a randomized controlled trial on cognitive training in older adults, obese individuals exhibited a lower response to memory training compared to their non-obese counterparts. However, the impact of processing speed and reasoning trainings mirrored that of healthy control participants. Cognitive training dose in healthy adults was discussed in the systematic review and meta-analysis by Lampit et al. [42]. Significant effects on cognition were found in studies that administered 1 to 3 sessions per week but not in studies that trained their participants more than 3 times per week. Moreover, the authors found that home-based cognitive training was ineffective compared to group-based training, as well as when session duration was less than 30 min [42]. In the present study, the mean duration of home-based cognitive training session was less than 30 min/week, which could have minimized the effects of our combined intervention.

Study limitations

Some limitations of our trial should be acknowledged. Due to ethical considerations during the pandemic context, our research team was not able to implement a control group receiving no intervention. Furthermore, a control group with cognitive training but without physical training was also not included for ethical reasons. Our region was heavily affected by restrictions with strict home confinement, leading to increased sedentary time, reduced physical activity, and adverse health effects. Given that exercise rehabilitation is standard of care (grade 1A) for individuals with cardiovascular disease physical exercise training was offered to all participants to mitigate these effects. A design including cognitive training only would have allowed to control for the mere effect of cognitive training. Secondly, in the combined intervention group, participant adherence to the cognitive training at home was suboptimal, suggesting the importance of supervision. Additionally, the data collected were solely based on self-reporting. Furthermore, the recruitment and randomization processes did not account for sex differences despite documented variations in cognition and exercise-training effects on cognition among older adults [43]. To date, no study has documented potential sex-related differences in the effects of exercise, cognitive training, or their combination on cognition and physical outcomes in older adults with cardiovascular disease.

Conclusion

The COVEPICARDIO study demonstrated that a home-based 6-month combined intervention, integrating sequential cognitive and exercise training, yields comparable enhancements in executive function and episodic memory performances as well as in functional test performances compared to a 6-month exercise training alone intervention. These results remain similar even after accounting for training doses. The synergistic effect of home-based combined intervention with exercise and cognitive training on cognition remains to be shown in older adults with CVD, with a specific emphasis on the dose of each type of intervention. Future research should also focus on comparing the efficacy of supervised center-based programs with remote home-based interventions to fully understand the impact of supervision and optimal dosing on exercise and cognitive training outcomes.

Supplementary information

Below is the link to the electronic supplementary material.

Author contributions

Conception and design of the research: L.B., F.B, E.G.D., C.G., M.G., T.Vi., C.A.B., C.A.G, K.S., P.V. Principal investigator: L.B. Co-investigators: A.N. Collaborators: F.B., E.G.D, C.G., M.G., T.Vi., T.Vr., C.A.B., J.I.G., S.B, K.S., P.V. Data analysis: M.C.B and F.B. Drafting and revision of the manuscript: F.B wrote the first version of the manuscript, all authors revised it and contributed significantly to write the final version that was accepted.

Funding

This study is supported by the Montreal Heart Institute Foundation and the Mirella and Lino Saputo Research Chair in Cardiovascular Health and the Prevention of Cognitive Deficits from University of Montreal at the Montreal Heart Institute. FB is financially supported by a fellowship from the Fonds de Recherche du Québec – Santé (FRQ-S).

Data Availability

The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

Participants provided written informed consent by email before starting the study. The COVEPICARDIO was conducted in compliance with the International Conference on Harmonization Good Clinical Practice (ICH-GCP) and in accordance with the Helsinki Declaration. The study protocol was approved by the Montreal Heart Institute’s (MHI) Research Ethics Board (FWA00003235; research project: ICM 2020–2785).

Consent for publication

All authors carefully revised the manuscript, and they all approved the final version for publication.

Conflict of interest

The authors declare that the research is conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.