Open access Original research

Effects of different aerobic exercises on

BMJ Open: first published as 10.1136/bmjopen-2022-067293 on 30 June 2023. Downloaded from http://bmjopen.bmj.com/ on August 1, 2023 by guest. Protected by copyright.
the global cognitive function of the
elderly with mild cognitive impairment:
a meta-­analysis
Conglin Han ‍ ‍,1,2 Weishuang Sun,1,2 Dan Zhang,1,2 Xiaoshuang Xi,3 Rong Zhang,4
Weijun Gong ‍ ‍5

To cite: Han C, Sun W, Zhang D, ABSTRACT

et al. Effects of different Objectives To summarise the effects of various types of
STRENGTHS AND LIMITATIONS OF THIS STUDY
aerobic exercises on the global aerobic exercise on the global cognitive function of the
cognitive function of the elderly ⇒ A rigorous methodology was used for this study,
elderly with mild cognitive impairment (MCI).
with mild cognitive impairment: including clear eligibility criteria, an extensive data-
a meta-­analysis. BMJ Open
Design A meta-­analysis of randomised controlled trials
(RCTs). base search, study selection by two reviewers work-
2023;13:e067293. doi:10.1136/
bmjopen-2022-067293 Data sources PubMed, EMBASE and the Cochrane ing independently and risk-­of-­bias assessment.
Library were searched for clinical RCTs from the earliest ⇒ This protocol is based on the Preferred Reporting
► Prepublication history and available records to March 2022. Items for Systematic Reviews and Meta-­Analyses
additional supplemental material project and the Cochrane Handbook for Systematic
Eligibility criteria for selecting studies We included
for this paper are available Reviews of Interventions.
online. To view these files, RCTs of subjects older than 60 years with MCI. The
outcome indicators of cognitive function of interest were ⇒ This study is the first to classify aerobic exercise
please visit the journal online
(http://dx.doi.org/10.1136/​ the Mini-­Mental State Examination (MMSE) and the and conduct a meta-­analysis to explore the effects
bmjopen-2022-067293). Montreal Cognitive Assessment (MoCA). of different aerobic exercises on cognitive function
Data extraction and synthesis Two researchers in older adults with mild cognitive impairment (MCI).
Received 11 August 2022 independently screened the literature, extracted data ⇒ This meta-­analysis will be used to inform future re-
Accepted 31 May 2023 and evaluated the quality of the included studies, search to guide and improve aerobic exercise train-
with disagreements resolved by a third researcher. ing practices in older adults with MCI.
The Cochrane Handbook for Systematic Reviews of ⇒ This study only analysed changes in global cognitive
Interventions was used to assess the risk of bias. functioning and did not include changes in specific
Meta-­analysis was performed by Review Manager V.5.3 cognitive domains.
software. Random-­effect models were used for meta-­
© Author(s) (or their
analysis.
employer(s)) 2023. Re-­use Results A total of 1680 patients who participated in 20 MCI. Nevertheless, the improvement effect of mind–body
permitted under CC BY. RCTs were included in this study. Based on outcomes of exercise is more reliable compared with multicomponent
Published by BMJ. MMSE analysis, the aerobic exercise, which was beneficial aerobic exercise and conventional aerobic exercise.
1
Beijing Rehabilitation Hospital, for global cognitive function in MCI patients, included PROSPERO registration number CRD42022327386.
Capital Medical University, multicomponent aerobic exercise (MD=1.79, 95% CI (1.41
Beijing, China to 2.17), p<0.01) and mind–body exercise (MD=1.28,
2
Rehabilitation Medicine 95% CI (0.83 to 1.74), p<0.01). The results of the meta-­ INTRODUCTION
Academy, Weifang Medical analysis of conventional aerobic exercise (MD=0.51, Along with the development and improve-
University, Weifang, Shandong, 95% CI (0.09 to 0.93), p=0.02) turned out to be statistically ment in modern science, medical technology
China
3
Beijing Rehabilitation Medicine
insignificant after sensitivity analysis (MD=0.14, 95% CI and treatment, the average life expectancy of
Academy, Capital Medical
(-­0.47 to 0.75), p=0.65). With the evaluation of MoCA, people around the world has been continu-
University, Beijing, China multicomponent aerobic exercise (MD=5.74, 95% CI (5.02 ously extended. Age-­related health problems
4
Second Clinical Medical to 6.46), p<0.01), mind–body exercise (MD=1.29, 95% CI are becoming more and more prominent.
Academy, Yunnan University of (0.67 to 1.90), p<0.01) and conventional aerobic exercise Dementia is one of the common diseases
Chinese Medicine, Kunming, (MD=2.06, 95% CI (1.46 to 2.65), p<0.01) were showed
Yunnan, China
of ageing and a major cause of disability in
significant beneficial effects for the patient. However,
5
Department of Neurological the elderly.1 The WHO estimates that about
there was a high degree of heterogeneity between the
Rehabilitation, Beijing results of multicomponent aerobic exercise (MMSE) and 50 million people worldwide suffer from
Rehabilitation Hospital, Beijing,
conventional aerobic exercise (MoCA), which was analysed dementia, with a new case every 3 s, and
China the number of people with dementia will be
and explored.
Correspondence to Conclusions In general, multicomponent aerobic three times by 2050.2 Severe dementia can
Dr Weijun Gong; exercise and mind–body exercise were beneficial in place a huge emotional and financial burden
​gwj197104@​ccmu.​edu.​cn improving global cognitive function in the elderly with on families and society, so it is important to

Han C, et al. BMJ Open 2023;13:e067293. doi:10.1136/bmjopen-2022-067293 1

Open access

take measures for early detection, early intervention and including breathing and physical exercise, medita-

BMJ Open: first published as 10.1136/bmjopen-2022-067293 on 30 June 2023. Downloaded from http://bmjopen.bmj.com/ on August 1, 2023 by guest. Protected by copyright.
early treatment of dementia. Mild cognitive impairment tion and so on.10 It is a type of aerobic exercise of low
(MCI) is an intermediate state between normal cognition to moderate intensity, with common forms such as tai
and dementia. MCI usually refers to cognitive decline chi, yoga and dance.11 12) and multicomponent aerobic
beyond that caused by normal ageing, but without exercise (aerobic exercise combined with other forms of
significant impairment in activity of daily living (ADL).3 exercise, such as resistance training, balance training).
Currently, medications have a limited role in preventing Aerobic exercise was defined as the involvement of major
and treating MCI,4 5 so there is a large body of research muscle groups throughout the body in addition to the
focusing on non-­pharmacological interventions. primary involvement of oxygen for energy supply.
It has been suggested that some non-­pharmacological
interventions, such as food, exercise, and cognitive stim- Types of control groups
ulation, may be beneficial for the improvement of cogni- Controls could be exercises of stretching, activities of
tive function in MCI patients.3 4 6 It has also shown that health education, routine care, daily lifestyle, and social
physical activity has a greater influence on the cogni- recreation.
tive performance of MCI patients than drugs. For the
elderly population with a decline in cognitive function, Types of outcomes
several studies have also confirmed that aerobic exercise The results are presented using measurable cognitive
appears to be effective in improving cognitive function.7–9 screening instruments: the Mini-­Mental State Examina-
However, there are various types of aerobic exercise, and tion (MMSE) and the Montreal Cognitive Assessment
the aerobic exercise that can be performed may vary (MoCA). and are available as mean (M) and SD.
in different populations (for example, it is difficult for Exclusion criteria
people with lower limb mobility restrictions to run). As Intervention methods could only be a variety of exer-
a result, an analysis is needed to compare the effects of cise training modalities; those with specific instructions
different types of aerobic exercise on the improvement of to have other training components, such as cognitive
global cognitive function in the elderly with MCI. training, were excluded. Non-­intervention studies, arti-
cles with only research protocols, review articles, unpub-
lished studies, abstracts or papers, and articles for which
METHODS full text or data were not available were excluded. Non-­
This study has been registered in the International English articles were excluded.
Prospective Systematic Evaluation Registry database
(PROSPERO, http://www.cdr.york.ac.uk/PROSPERO/) Search strategy
under the registration number: CRD42022327386. This We systematically searched electronic databases in
study was conducted following the Preferred Reporting PubMed, EMBASE and the Cochrane Library from the
Items for Systematic Reviews and Meta-­Analyses statement. earliest available records to March 2022. We searched
using subject terms and free words. The keywords used
Inclusion criteria included search terms related to sports exercise (eg,
Types of studies exercise or physical activity or aerobic exercise) and
Human subject studies designed as randomised controlled cognitive impairment (eg, or cognitive dysfunction or
trials (RCTs) published in English were included without amnestic MCI) and search terms related to older adults
year restriction. (eg, older adults or older patients). The search strate-
gies for EMBASE, PubMed and the Cochrane Library are
Types of participants described in online supplemental document 1. Endnote
Subjects defined as MCI by the original authors and X9.1 was used to store and sort the retrieved RCTs and
aged>60 years were included. We excluded healthy ageing to remove duplicates. Initial screening was performed by
elderly adults, patients with any form of dementia or those the first author based on the title and abstract. This was
diagnosed with cognitive impairment due to definite aeti- followed by an independent screening of the full text by
ologies, such as trauma or vascular or psychiatric diseases. two reviewers according to the eligibility criteria specified
Any clinical subtype of MCI was considered eligible. in the study protocol. Disagreements in included studies
were resolved by negotiation.
Types of interventions
We included supervised structured exercises of any Data extraction and analysis
frequency, intensity, duration, or time directed at A standardized data extraction form was applied to extract
improving physical fitness. The interventions were data based on the following parameters: (1) basic infor-
various forms of exercise containing aerobic exercise, mation about the included studies (author, year, country),
including conventional aerobic exercise (simple and (2) basic characteristics of the study population (sample
common aerobic exercises such as running, brisk walking size, age and gender of participants), (3) intervention
and cycling), mind–body exercise (mind–body exer- characteristics (type of intervention, the intervention
cise focuses on mind, body, psychology and behaviour, specifics, frequency and intensity of the intervention)

2 Han C, et al. BMJ Open 2023;13:e067293. doi:10.1136/bmjopen-2022-067293

 Open access

and (4) outcome measures: we chose the two most Patient and public involvement

BMJ Open: first published as 10.1136/bmjopen-2022-067293 on 30 June 2023. Downloaded from http://bmjopen.bmj.com/ on August 1, 2023 by guest. Protected by copyright.
commonly used scales for assessing global cognitive func- There was no direct patient or public involvement in this
tioning, the MMSE and the MoCA. We performed a meta-­ review.
analysis using the endpoint measures. First, we conducted
a meta-­analysis of all aerobic exercise studies to obtain
the results of the meta-­ analysis of MMSE and MoCA. RESULTS
After that, according to the type of aerobic exercise, they Search results
were divided into three groups (multicomponent aerobic By searching electronic databases and literature, we
exercise, mind–body exercise and conventional aerobic obtained a total of 24 255 relevant references. After dupli-
exercise) and meta-­ analysis was conducted separately. cates were removed, and titles and abstracts were checked
For each type of aerobic exercise, meta-­analysis results of against predefined inclusion criteria, 102 complete refer-
MMSE and MoCA were also obtained. ences were retained, and we reviewed each paper in detail.
For each outcome, the effect sizes are reported as the A total of 84 trials did not meet the predefined inclusion
mean difference (MD) with 95% CI for each study. We criteria: non-­RCTs (n=4), no control group meeting
pooled these values and conducted the meta-­ analysis the criteria (n=5), full text not available (n=6), assess-
using the Review Manager V.5.3. Heterogeneity between ment results not meeting the criteria (n=23), the study
trial design options was unclear; therefore, a random without results (n=9), data not available (n=13), interven-
effects model was chosen. Our random-­effect model was tion modality not meeting the predefined criteria (n=7)
based on the postulation that effect estimates were not the and study population not meeting the inclusion criteria
same but followed a normal distribution. While the area (n=17). Therefore, a total of 18 articles were included in
of the black square in forest plots denotes the weighted this study. The search process is shown in figure 1.
contribution of each study, p<0.05 (two tailed) was consid-
ered statistically significant. The I² statistic was computed Research characteristics
as a measure of heterogeneity.13 Cochrane’s handbook A total of 18 RCTs were included in this study, published
recommends heterogeneity as not important (0%–40%), from 2011 to 2022. The characteristics of all included
moderate heterogeneity (30%–60%), substantial hetero- studies are summarised in online supplemental docu-
geneity (50%–90%) and considerable heterogeneity ment 2 . The studies were conducted in different coun-
(75%–100%). Sensitivity analysis evaluated the stability of tries: China (n=8), the USA (n=2), Brazil (n=2), Japan
the results by leave-­one-­out method. Subgroup analyses (n=1), Korea (n=1), Turkey (n=1), Iran (n=1), Greece
were used to address significant clinical heterogeneity, or (n=1) and Spain (n=1). All included participants were
only descriptive analyses were performed. 1700 and they were assigned to either the experimental
or control group, and the sample size of the study ranged
from 19 to 389. All participants were diagnosed with MCI.
Risk of bias and quality assessment
One study15 was composed of the mind–body exercise
The Cochrane risk of-­bias assessment tool was used to
intervention group and the conventional aerobic inter-
independently assess the risk of bias in included trials. The
vention group with their corresponding control group.
assessment covered the following parameters: random
Another study16 was designed as two groups with different
sequence generation, allocation concealment, blinding,
exercise intensities (40% and 60% reserve heart rate) for
incomplete data, selective reporting and others (Cochrane
the conventional aerobic intervention. For all included
Handbook for Systematic Reviews of Interventions 8.5). Each studies, five experimental groups underwent multicom-
parameter was divided into three categories: low risk, ponent exercise training with aerobic exercise,17–21 seven
unclear and high risk. Each study was categorised using experimental groups underwent mind–body exercise
the following criteria: low risk of bias (all parameters were training15 22–27; and eight experimental groups under-
rated as low risk), moderate risk of bias (with parameters went conventional aerobic exercise training.15 16 28–32 The
rated as unclear and no parameters rated as high risk) total intervention duration for all studies ranged from 6
and high risk of bias (as long as one or more parameters weeks to 12 months. The frequency was 21–80 min per
were rated as high risk) (Cochrane Handbook for Systematic session, 1–5 times per week. Three experimental groups
Reviews of Interventions table 8.7.a). Two reviewers (CH had an exercise intensity of roughly low intensity,15–17
and WS) assessed each trial independently. Any disagree- that is, maximum heart rate<60% or Borg scale (level
ments were discussed with a third reviewer (WG) until a 10)<5 or Borg scale (level 20)<12 or metabolic equivalent
final decision was achieved. during exercise (MET)<3 or reserve heart rate<59%; 15
experimental groups had an exercise intensity of roughly
Assessment of evidence quality moderate to high intensity,15 16 18–20 22–29 31 32 that is,
The overall quality of the evidence was assessed using the maximum heart rate≥60% or Borg scale (10th grade)≥5 or
Recommendations for Assessment, Development and Borg scale (20th grade))≥12 or MET≥3 or reserve heart
Evaluation (GRADE) criteria.14 In the system, the quality rate≥59%; the intensity of exercise in two other exper-
of evidence is classified into four levels: high quality, imental groups was unclear.21 30 Ten studies assessed
moderate quality, low quality and very low quality.14 global cognitive function with the MMSE,16 18–21 26–28 31 32

Han C, et al. BMJ Open 2023;13:e067293. doi:10.1136/bmjopen-2022-067293 3

Open access

five studies assessed global cognitive function with the

BMJ Open: first published as 10.1136/bmjopen-2022-067293 on 30 June 2023. Downloaded from http://bmjopen.bmj.com/ on August 1, 2023 by guest. Protected by copyright.
MoCA15 22 24 29 30 and three additional studies used both
the MMSE and the MoCA.17 23 25

Literature quality evaluation

As shown in figure 2, the quality of studies included
in the trials varied. Two trials had a low overall risk
of bias,24 29 five trials had a moderate overall risk of
bias15 18 22 27 28 and eleven trials had a high overall risk of
bias.16 17 19–21 23 25 26 30–32 Regarding the method of rando-
misation, 13 trials15 17 18 20–25 29–32 had simple randomiza-
tion and 5 trials16 19 26–28 had stratified randomisation.
In terms of intervention allocation concealment, nine
trials17 19–21 23 25 27 28 30 used open randomisation tables,
three trials22 24 29 used sealed envelopes and the remaining
six trials15 16 18 26 31 32 were unclear.
We conducted a subgroup analysis based on the quality
of the studies in order to test the strength of the evidence.
The relevant results are shown in online supplemental
document 3. Generally, only the moderate risk group of
conventional aerobic exercise (MoCA) had statistically
insignificant results and there was only one study in this
group.

Results of meta-analysis of all included studies

The results of MMSE included a total of 14 trials, as
shown in figure 3. Meta-­analysis of the random effects
model showed that aerobic exercise was effective in
improving global cognitive function compared with the
control group (MD=1.23, 95% CI (0.99 to 1.47), p<0.01).
Significant heterogeneity was found in the study (I²=88%,
p<0.01).
The results of MoCA included a total of eight trials,
as shown in figure 3. Randomised effects model meta-­
analysis showed that aerobic exercise was effective in
improving global cognitive function compared with
control group (MD=2.94, 95% CI (2.58 to 3.30), p<0.01).
Significant heterogeneity was found in the study (I²=94%,
p<0.01).

Results of meta-analysis of different aerobic exercises

All trials were divided into multicomponent aerobic exer-
cise, mind–body exercise and regular aerobic exercise
by exercise modality and the results of meta-­analysis are
shown in figures 4 and 5. As can be seen from the figures,
after grouping by exercise modality, the between-­group
differences in the two outcome indicators of MMSE
(p<0.01) and MoCA (p<0.01) were statistically significant,
indicating that the grouping was meaningful.

Results of meta-analysis of multicomponent aerobic exercise

As shown in figure 4, multicomponent aerobic exercise
had a significant beneficial effect in the MMSE outcome
index (MD=1.79, 95% CI (1.41 to 2.17), p<0.01) but the
heterogeneity was very high (I²=95%, p<0.01), so further
Figure 1 Study flow diagram. analysis is needed. We performed a subgroup analysis
of the multicomponent intervention group, and the
results are shown in online supplemental document 4.
We performed a subgroup analysis based on the sample

4 Han C, et al. BMJ Open 2023;13:e067293. doi:10.1136/bmjopen-2022-067293

 Open access

size, exercise intensity, region, deviation risk rating and

BMJ Open: first published as 10.1136/bmjopen-2022-067293 on 30 June 2023. Downloaded from http://bmjopen.bmj.com/ on August 1, 2023 by guest. Protected by copyright.
duration, frequency and total time in the trial. From the
results of these analyses, it is clear that no meaningful
results were obtained after subgroup analysis. As shown
in figure 5, there was only one multicomponent exercise
trial with a MoCA outcome and the result was (MD=5.74,
95% CI (5.02 to 6.46), p<0.01).

Results of meta-analysis of mind–body exercise

As shown in figure 4, in the MMSE outcome, the mind–
body exercise intervention had a significant improvement
effect (MD=1.28, 95% CI (0.83 to 1.74), p<0.01) and was
not highly heterogeneous (I²=50%, p=0.11). As shown
in figure 5, there was also a clinically significant effect of
mind–body exercise in the results of MoCA (MD=1.29,
95% CI (0.67 to 1.90), p<0.01) and its heterogeneity was
also low (I²=6%, p=0.36).

Results of meta-analysis of conventional aerobic exercise

As shown in figure 4, in the outcome of MMSE, the
improvement in global cognitive function was signifi-
cant in the conventional aerobic exercise intervention
group (MD=0.51, 95% CI (0.09 to 0.93), p=0.02) and
without heterogeneity (I²=0%, p=0.46). Figure 5 demon-
strates the MoCA results for conventional aerobic
exercise. The results are also significant (MD=2.06,
95% CI (1.46 to 2.65), p<0.01), but the heterogeneity
(I²=84%, p=0.002) is high requiring further analysis and
interpretation.

Rating the body of evidence

Online supplemental document 5 presents the GRADE
evidence profile for the RCT outcomes. The quality of
evidence varies from moderate to very low. For multicom-
ponent exercise, there is very low evidence (MMSE and
MoCA) that it has a significant clinical effect on global
cognitive function. For mind–body exercise, moderate
(MMSE) and low (MoCA) evidence suggests a clini-
cally significant effect on global cognitive function. For
conventional aerobic exercise, there is low and very low
evidence of a clinically significant effect on global cogni-
tive function.

Sensitivity analysis
Using a leave-­one-­out method, we tested the effect of
excluding individual studies on the stability of the effect
estimates. We found no significant effect on effect esti-
mates after excluding each study, except for the results
for conventional aerobic exercise (MMSE). This suggests
that the vast majority of the final results were stable.
Regarding the MMSE results for conventional aerobic
exercise, after sensitivity analysis, we found that the results
were no longer significant after excluding the Wei and Ji’s
study32 (MD=0.14, 95% CI (-­0.47 to 0.75), p=0. 65). The
Figure 2 Risk-­of-­bias assessment of included trials.
details of the sensitivity analysis are presented in online
supplemental document 6.

Han C, et al. BMJ Open 2023;13:e067293. doi:10.1136/bmjopen-2022-067293 5

Open access

BMJ Open: first published as 10.1136/bmjopen-2022-067293 on 30 June 2023. Downloaded from http://bmjopen.bmj.com/ on August 1, 2023 by guest. Protected by copyright.
Figure 3 (A) Effect sizes of forest plots for aerobic exercise versus controls—MMSE outcomes. MMSE, Mini-­Mental State
Examination. (B) Effect sizes of forest plots for aerobic exercise versus controls—MoCA outcome. MoCA, Montreal Cognitive
Assessment.

DISCUSSION on cognitive function in the elderly with MCI, the inter-

To the best of our knowledge, our meta-­analysis explored vention criteria for the studies included in these meta-­
for the first time the effect of different aerobic exercise analyses were not only exercise but also cognitive training
types on global cognitive function in elderly patients and mental stimulation, which were not the same as in
with MCI. Our study analysed improvements in cognitive our study. The Meta-­analysis by Borges -Machado et al35
function through endpoint indicators of global cognitive discussed multicomponent exercise, which was defined
function and found that both multicomponent aerobic as exercise that combined aerobic, strength, posture and
exercise and mind–body exercise produced positive balance training. Although the study found improve-
effects. In addition, the results of conventional aerobic ments in ADL in the subjects, the study was conducted in
exercise were inconsistent and may not have a clinically patients with dementia and the results did not determine
significant effect on global cognitive function. The effects
the effects on cognitive function.
of conventional aerobic exercise need to be further inves-
Previous studies36 37 confirmed the beneficial effects
tigated in a large number of high-­quality studies.
of mind–body exercise on global cognitive function in
Summary of the literature people with MCI. Unlike the present study, they had slight
There were few previous meta-­analyses33 34 that explored differences in the type of trial studies, age and outcome
the effects of multicomponent exercise on cognitive func- evaluation indicators in the inclusion criteria. Our
tion in MCI. Although some of their results suggested a inclusion criteria were relatively more stringent. These
beneficial effect of multicomponent aerobic exercise previous studies combined different cognitive evaluation

6 Han C, et al. BMJ Open 2023;13:e067293. doi:10.1136/bmjopen-2022-067293

 Open access

BMJ Open: first published as 10.1136/bmjopen-2022-067293 on 30 June 2023. Downloaded from http://bmjopen.bmj.com/ on August 1, 2023 by guest. Protected by copyright.
Figure 4 Effect sizes of forest plots for the different aerobic exercise—MMSE outcomes. MMSE, Mini-­Mental State
Examination.

metrics expressed as standardised MD. We selected the cognitive function improvement than training alone.40
two most common indicators of global cognitive function Another possible explanation was that depression was an
evaluation (MMSE and MoCA) to be analysed separately important risk factor for cognitive decline in people with
and expressed the results as MD. MCI and that mind–body exercise improved mood in
We did not find any meta-­analysis related to conven- older adults better than other physical exercises.36
tional aerobic exercise in patients with MCI. Although
there are a few reviews9 38 39 reporting aerobic exercise Heterogeneity analysis of multicomponent aerobic exercise
in MCI patients, they did not analyse the types of aerobic From a total of five studies analysed by MMSE for multi-
exercise separately. Therefore, we believed that the anal- component aerobic exercise, two studies19 21 were no
ysis of the present study was still of practical relevance. meaningful change in outcomes. We analysed these
studies and speculated that the possible reasons for such
Analysis of the results of mind–body exercise high heterogeneity. In the trial studied by de Oliveira et
All studies of mind–body exercise reported positive al,19 the authors indicated that participants in both the
results. The results of both the MMSE and MoCA trial and control groups were receiving medication, but
meta-­analyses were significant effects and low hetero- no further medication-­related descriptions. We are not
geneity, which would seem to indicate the effectiveness sure if the medication affected the cognitive function of
of mind–body exercise in improving cognitive function the subjects and thus the effect of the intervention. In
in older adults with MCI. One possible explanation for another trial21 that reported negative results, each exer-
the beneficial results of the mind–body exercise was cise type of its multicomponent aerobic exercise was
that mind–body exercise was a form of physical exercise performed separately, whereas common multicompo-
supplemented with different degrees of cognitive func- nent exercises combine different components in a single
tion training. For example, Tai Chi, a Chinese system of exercise session.35 This was therefore the most different
slow meditative physical exercise designed for mind relax- aspect of this study from other multicomponent exercise
ation and balance, required attention, memory and other training, and one possible explanation was that different
aspects in addition to physical activity. Adding cognitive components were more effective when combined in
training to physical activity seems to have better effects on a single exercise session. One study showed41 that

Han C, et al. BMJ Open 2023;13:e067293. doi:10.1136/bmjopen-2022-067293 7

Open access

BMJ Open: first published as 10.1136/bmjopen-2022-067293 on 30 June 2023. Downloaded from http://bmjopen.bmj.com/ on August 1, 2023 by guest. Protected by copyright.
Figure 5 Effect sizes of forest plots for the different aerobic exercise—MoCA outcomes. MoCA, Montreal Cognitive
Assessment.

short-­term tai chi training did not significantly improve study47 demonstrated a sex bias in older adults regarding
cognitive function in adults when comparing long-­term the effects of aerobic exercise on cognitive function, with
and short-­term tai chi training. In another trial,17 the older women experiencing more significant cognitive
duration of aerobic exercise was too short at only 5 min. benefits from aerobic exercise interventions than men.
Thus, we speculated that there may be some difference in Second, one of the included studies15 had a small sample
the results produced by different time of aerobic exercise. size, which may have made the results of the trial less
reliable.
Analysis of the results of conventional aerobic exercise In summary, given the above two results, we are cautious
The preliminary MMSE results for conventional aerobic about the effect of conventional aerobic exercise. We
exercise were significant, which would seem to demon- believe that the improvement in cognitive function with
strate the effectiveness of conventional aerobic exer- conventional aerobic exercise may not always be signifi-
cise. However, after sensitivity analysis was performed cant and that future studies could improve the quality of
to exclude this Wei and Ji’s study,32 this result became research and provide more in-­depth studies and analyses
meaningless. This suggests that the MMSE results for of specific interventions.
conventional aerobic exercise are unstable and we cannot
conclude the effectiveness of conventional aerobic exer- Limitations
cise on this basis. In fact, despite the large number of There were several limitations of this review. The first
studies and recommendations9 42 43 that conclude that is the number of included studies. Although the total
aerobic exercise in life is beneficial for cognitive function, number of included studies is not small, the interven-
there are still many evidence-­based studies44 45 that have tions primarily explored in this study were a variety
yielded non-­positive results. A 4-­year trial46 showed that of different aerobic exercises, so the small number of
aerobic exercise alone was not sufficient to change cogni- included studies for each type of aerobic exercise makes
tive function, and that aerobic exercise up to a certain the interpretation of the results obtained rather difficult.
intensity and in combination with other forms of exercise Second, the quality of the studies we included varied,
might have better results. This may be the reason why the and low-­quality studies may make the results inaccurate.
results are not positive. The third is that there was considerable heterogeneity
The results of the meta-­analysis on MoCA produced between some of the included studies, representing
a high level of heterogeneity, which we tried to explain. both clinical and methodological heterogeneity, due to
First of all, in the three studies15 29 30 with this outcome, differences between participants (eg, older adults may
there was a larger proportion of female subjects, which have unknown comorbidities and the possibility of taking
may have made the studies unrepresentative. A previous medications as a result, which may affect the effectiveness

8 Han C, et al. BMJ Open 2023;13:e067293. doi:10.1136/bmjopen-2022-067293

 Open access

of the intervention), interventions, and study designs. Patient consent for publication Not applicable.

BMJ Open: first published as 10.1136/bmjopen-2022-067293 on 30 June 2023. Downloaded from http://bmjopen.bmj.com/ on August 1, 2023 by guest. Protected by copyright.
Fourth, it cannot be excluded that certain control groups Ethics approval Not applicable.
would also have different levels of benefit on cognitive Provenance and peer review Not commissioned; externally peer reviewed.
function such as social activities.48 Fifth, only some of Data availability statement All data relevant to the study are included in the
the trials that assessed global cognitive outcomes were article or uploaded as online supplemental information. This study summarised and
included, and not those that assessed separately cogni- analysed the data from the original experiment.
tive domains. Finally, only studies of trials with MMSE Supplemental material This content has been supplied by the author(s). It has
and MoCA outcomes were selected. Although these two not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been
peer-­reviewed. Any opinions or recommendations discussed are solely those
cognitive evaluation metrics are the most commonly
of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and
used, they also undoubtedly excluded other studies that responsibility arising from any reliance placed on the content. Where the content
might be of value. includes any translated material, BMJ does not warrant the accuracy and reliability
of the translations (including but not limited to local regulations, clinical guidelines,
Implications for future research terminology, drug names and drug dosages), and is not responsible for any error
Future studies should be designed with rigorous methods and/or omissions arising from translation and adaptation or otherwise.
as well as controlled for various biases to improve the Open access This is an open access article distributed in accordance with the
quality of the study. Different types of aerobic exercise Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits
others to copy, redistribute, remix, transform and build upon this work for any
have different experimental interventions. The duration, purpose, provided the original work is properly cited, a link to the licence is given,
frequency and intensity of the exercise intervention may and indication of whether changes were made. See: https://creativecommons.org/​
have different degrees of influence on the results of the licenses/by/4.0/.
trial. Follow-­up research protocols could focus on this ORCID iDs
to explore the design of better interventions. The exer- Conglin Han http://orcid.org/0000-0003-0621-4029
cise components of multicomponent aerobic training Weijun Gong http://orcid.org/0000-0002-9134-8218
are complex and future studies could investigate better
combinations of exercise components to achieve better
improvements.
In addition, previous studies have indicated that the REFERENCES
MoCA is superior to the MMSE for screening diagnoses in 1 Reinvigorating the public health response to dementia. Lancet Public
MCI patients, making it more recommended to replace Health 2021;6.
2 World Health Organization. WHO Guidelines Approved by the
the MMSE with the MoCA for future scales measuring Guidelines Review Committee, Risk Reduction of Cognitive Decline
global cognitive function.49 In addition to the MMSE and and Dementia: WHO Guidelines. Geneva, 2019.
3 Sanford AM. Mild cognitive Impairment[J]. Clinics in Geriatric
the MoCA, the Clinical Dementia Rating is a worthwhile Medicine 2017;33:325–37.
scale to refer to and use. In addition to the global cogni- 4 Jongsiriyanyong S, Limpawattana P. Mild cognitive impairment in
tive function, it would be interesting to assess specific clinical practice: a review Article. Am J Alzheimers Dis Other Demen
2018;33:500–7.
domains of cognitive function (eg, memory, attention), 5 Schneider LS, Mangialasche F, Andreasen N, et al. Clinical trials and
which may allow the inclusion of additional studies and late-­stage drug development for Alzheimer’s disease: an appraisal
from 1984 to 2014. J Intern Med 2014;275:251–83.
better delineate the sources of heterogeneity involved. 6 Ataollahi Eshkoor S, Mun CY, Ng CK, et al. Mild cognitive impairment
and its management in older People. CIA 2015;10:687.
7 Law C-­K, Lam FM, Chung RC, et al. Physical exercise attenuates
CONCLUSION cognitive decline and reduces behavioural problems in people with
mild cognitive impairment and dementia: a systematic Review.
The results of this meta-­analysis suggest that two types Journal of Physiotherapy 2020;66:9–18.
of aerobic exercise (including multicomponent aerobic 8 Ahlskog JE, Geda YE, Graff-­Radford NR, et al. Physical exercise as
a preventive or disease-­modifying treatment of dementia and brain
exercise and mind–body exercise) have a positive effect Aging. Mayo Clin Proc 2011;86:876–84.
on improving global cognitive function in older patients 9 Yong L, Liu L, Ding T, et al. Evidence of effect of aerobic exercise on
with MCI. However, the conclusion should be treated cognitive intervention in older adults with mild cognitive Impairment.
Front Psychiatry 2021;12:713671.
cautiously due to the heterogeneity and limited research. 10 Li Z, Liu S, Wang L, et al. Mind-­body exercise for anxiety and
Additional well-­designed RCTs are needed to provide depression in COPD patients: A systematic review and meta-­
Analysis. IJERPH 2019;17:22.
more evidence for this conclusion. 11 Zhang T, Liu W, Gao S. Effects of mind-­body exercises on cognitive
impairment in people with Parkinson’s disease: a mini-­review. Front
Contributors All authors made a substantial contribution to this work. CH, WS, Neurol 2022;13:931460.
DZ, XX, RZ and WG all contributed to the conception and design of the review. 12 Wu C, Yi Q, Zheng X, et al. Effects of mind-­body exercises on
cognitive function in older adults: A meta-­Analysis. J Am Geriatr Soc
CH drafted the paper with all the authors critically reviewing it and suggesting
2019;67:749–58.
amendments prior to submission. CH and WG are guarantors. The corresponding 13 Higgins JPT, Altman DG, Gøtzsche PC, et al. The Cochrane
author attests that all listed authors meet authorship criteria and that no others collaboration’s tool for assessing risk of bias in randomised trials.
meeting the criteria have been omitted. BMJ 2011;343:d5928.
14 Brozek JL, Akl EA, Alonso-­Coello P, et al. Grading quality of evidence
Funding This work was supported by Capital’s Funds for Health Improvement and
and strength of recommendations in clinical practice guidelines. part
Research (2022-­1-­2251), Natural Science Foundation of China (81972148) and 1 of 3. An overview of the GRADE approach and grading quality of
Beijing Natural Science Foundation of China (7222101). evidence about Interventions. Allergy 2009;64:669–77.
Competing interests None declared. 15 Tao J, Liu J, Chen X, et al. Mind-­body exercise improves cognitive
function and modulates the function and structure of the
Patient and public involvement Patients and/or the public were not involved in hippocampus and anterior cingulate cortex in patients with mild
the design, or conduct, or reporting, or dissemination plans of this research. cognitive Impairment[J]. Neuroimage Clin 2019;23:101834.

Han C, et al. BMJ Open 2023;13:e067293. doi:10.1136/bmjopen-2022-067293 9

Open access

16 Varela S, Ayán C, Cancela JM, et al. Effects of two different 32 Wei X, Ji L. Effect of Handball training on cognitive ability in elderly

BMJ Open: first published as 10.1136/bmjopen-2022-067293 on 30 June 2023. Downloaded from http://bmjopen.bmj.com/ on August 1, 2023 by guest. Protected by copyright.
intensities of aerobic exercise on elderly people with mild with mild cognitive Impairment[J]. Neurosci Lett 2014;566:98–101.
cognitive impairment: a randomized pilot Study[J]. Clin Rehabil 33 Özbe D, Graessel E, Donath C, et al. Immediate intervention effects
2012;26:442–50. of standardized Multicomponent group interventions on people
17 Li L, Liu M, Zeng H, et al. Multi-­component exercise training with cognitive impairment: A systematic Review. J Alzheimers Dis
improves the physical and cognitive function of the elderly with mild 2019;67:653–70.
cognitive impairment: a six-­month randomized controlled Trial[J]. Ann 34 Bruderer-­Hofstetter M, Rausch-­Osthoff A-­K, Meichtry A, et al.
Palliat Med 2021;10:8919–29. Effective Multicomponent interventions in comparison to active
18 Langoni C da S, Resende T de L, Barcellos AB, et al. Effect control and no interventions on physical capacity, cognitive function
of exercise on cognition, conditioning, muscle endurance, and instrumental activities of daily living in elderly people with and
and balance in older adults with mild cognitive impairment: a without mild impaired cognition - A systematic review and network
randomized controlled Trial[J]. Journal of Geriatric Physical Therapy meta-­analysis. Ageing Res Rev 2018;45:1–14.
2019;42:E15–22. 35 Borges-­Machado F, Silva N, Farinatti P, et al. Effectiveness of
19 de Oliveira Silva F, Ferreira JV, Plácido J, et al. Three months of Multicomponent exercise interventions in older adults with dementia:
Multimodal training contributes to mobility and executive function in A meta-­Analysis. Gerontologist 2021;61:e449–62.
elderly individuals with mild cognitive impairment, but not in those 36 Wang S, Yin H, Jia Y, et al. Effects of mind-­body exercise on
with Alzheimer's disease: A randomized controlled trial. Maturitas cognitive function in older adults with cognitive impairment:
2019;126:28–33. A systematic review and meta-­Analysis. J Nerv Ment Dis
20 Bademli K, Lok N, Canbaz M, et al. Effects of physical activity 2018;206:913–24.
program on cognitive function and sleep quality in elderly with 37 Zou L, Loprinzi PD, Yeung AS, et al. The beneficial effects of
mild cognitive impairment: A randomized controlled trial. Perspect mind-­body exercises for people with mild cognitive impairment: a
Psychiatr Care 2019;55:401–8. systematic review with meta-­Analysis. Archives of Physical Medicine
21 Lam LC, Chan WC, Leung T, et al. Would older adults with mild and Rehabilitation 2019;100:1556–73.
cognitive impairment adhere to and benefit from a structured 38 Cammisuli DM, Innocenti A, Franzoni F, et al. Aerobic exercise
Lifestyle activity intervention to enhance cognition?: a cluster effects upon cognition in mild cognitive impairment: A
randomized controlled Trial[J]. PLoS ONE 2015;10:e0118173. systematic review of randomized controlled trials. Arch Ital Biol
22 Li F, Harmer P, Fitzgerald K, et al. A cognitively enhanced Online Tai 2017;155:54–62.
Ji Quan training intervention for community-­dwelling older adults 39 Zheng G, Xia R, Zhou W, et al. Aerobic exercise ameliorates cognitive
with mild cognitive impairment: a feasibility Trial[J]. BMC Geriatr function in older adults with mild cognitive impairment: a systematic
2022;22:76. review and meta-­analysis of randomised controlled trials. Br J Sports
23 Qi M, Zhu Y, Zhang L, et al. The effect of aerobic dance intervention Med 2016;50:1443–50.
on brain spontaneous activity in older adults with mild cognitive 40 Gheysen F, Poppe L, DeSmet A, et al. Physical activity to improve
impairment: A resting-­state functional MRI Study[J]. Exp Ther Med cognition in older adults: can physical activity programs enriched
2019;17:715–22. with cognitive challenges enhance the effects? A systematic review
24 Zhu Y, Wu H, Qi M, et al. Effects of a specially designed aerobic and meta-­Analysis[J]. Int J Behav Nutr Phys Act 2018;15:63.
dance routine on mild cognitive impairment. Clin Interv Aging 41 Walsh JN, Manor B, Hausdorff J, et al. Impact of Short- and long-­
2018;13:1691–700. term Tai Chi mind-­body exercise training on cognitive function
25 Lazarou I, Parastatidis T, Tsolaki A, et al. International ballroom in healthy adults: results from a hybrid observational study and
dancing against neurodegeneration: a randomized controlled trial in randomized Trial[J]. Glob Adv Health Med 2015;4:38–48.
Greek community-­dwelling Elders with mild cognitive Impairment[J]. 42 World Health Organization. WHO Guidelines Approved by the
Am J Alzheimers Dis Other Demen 2017;32:489–99. Guidelines Review Committee, Global Recommendations on Physical
26 Doi T, Verghese J, Makizako H, et al. Effects of cognitive leisure Activity for Health. Geneva, 2010.
activity on cognition in mild cognitive impairment: results of a 43 Northey JM, Cherbuin N, Pumpa KL, et al. Exercise interventions for
randomized controlled Trial[J]. Journal of the American Medical cognitive function in adults older than 50: a systematic review with
Directors Association 2017;18:686–91. meta-­analysis. Br J Sports Med 2018;52:154–60.
27 Lam LC, Chan WM, Kwok TC. Effectiveness of Tai Chi in 44 Gates N, Fiatarone Singh MA, Sachdev PS, et al. The effect
maintenance of cognitive and functional abilities in mild cognitive of exercise training on cognitive function in older adults with
impairment: a randomised controlled trial. Hong Kong Med J mild cognitive impairment: a meta-­analysis of randomized
2014;20:20–3. controlled Trials[J]. The American Journal of Geriatric Psychiatry
28 Tomoto T, Liu J, Tseng BY, et al. One-­year aerobic exercise 2013;21:1086–97.
reduced carotid arterial stiffness and increased cerebral blood 45 Snowden M, Steinman L, Mochan K, et al. Effect of exercise on
flow in Amnestic mild cognitive impairment. J Alzheimers Dis cognitive performance in community-­dwelling older adults: review
2021;80:841–53. of intervention trials and recommendations for public health practice
29 Song D, Yu DSF. Effects of a moderate-­intensity aerobic exercise and research. J Am Geriatr Soc 2011;59:704–16.
programme on the cognitive function and quality of life of 46 Komulainen P, Tuomilehto J, Savonen K, et al. Exercise, diet, and
community-­dwelling elderly people with mild cognitive impairment: A cognition in a 4-­year randomized controlled trial: dose-­responses to
randomised controlled trial. Int J Nurs Stud 2019;93:97–105. exercise training (DR’s EXTRA). Am J Clin Nutr 2021;113:1428–39.
30 Choi W, Lee S. Ground Kayak Paddling exercise improves postural 47 Baker LD, Frank LL, Foster-­Schubert K, et al. Effects of aerobic
balance, muscle performance, and cognitive function in older adults exercise on mild cognitive impairment: a controlled trial. Arch Neurol
with mild cognitive impairment: a randomized controlled Trial[J]. Med 2010;67:71–9.
Sci Monit 2018;24:3909–15. 48 James BD, Wilson RS, Barnes LL, et al. Late-­life social activity
31 Kohanpour M-­A, Mohammad-­Ali K, Maghsoud P, et al. The effects and cognitive decline in old age. J Int Neuropsychol Soc
of aerobic exercise with lavender essence use on cognitive state 2011;17:998–1005.
and serum brain-­derived Neurotrophic factor levels in elderly with 49 Tiago CP, Leonardo M, Tatiana MB, et al. Is the Montreal cognitive
mild cognitive Impairment. Journal of Herbmed Pharmacology assessment (Moca) screening superior to the mini-­mental state
2017;6:80–4. examination (MMSE). Int Psychogeriatr 2019;31:491–504.

10 Han C, et al. BMJ Open 2023;13:e067293. doi:10.1136/bmjopen-2022-067293