The Effect of Aerobic Exercise on Cognitive Function in People with Alzheimer’s Disease: A Systematic Review and Meta-Analysis of Randomized Controlled Trials

A growing body of research has examined the effect of aerobic exercise on cognitive function in people with Alzheimer’s Disease (AD), but the findings of the available studies were conflicting. The aim of this study was to explore the effect of aerobic exercise on cognitive function in AD patients. Searches were performed in PubMed, Web of Science, and EBSCO databases from the inception of indexing until 12 November 2021. Cochrane risk assessment tool was used to evaluate the methodological quality of the included literature. From 1942 search records initially identified, 15 randomized controlled trials (RCTs) were considered eligible for systematic review and meta-analysis. Included studies involved 503 participants in 16 exercise groups (mean age: 69.2–84 years) and 406 participants (mean age: 68.9–84 years) in 15 control groups. There was a significant effect of aerobic exercise on increasing mini-mental state examination (MMSE) score in AD patients [weighted mean difference (WMD), 1.50 (95% CI, 0.55 to 2.45), p = 0.002]. Subgroup analyses showed that interventions conducted 30 min per session [WMD, 2.52 (95% CI, 0.84 to 4.20), p = 0.003], less than 150 min per week [WMD, 2.10 (95% CI, 0.84 to 3.37), p = 0.001], and up to three times per week [WMD, 1.68 (95% CI, 0.46 to 2.89), p = 0.007] increased MMSE score significantly. In addition, a worse basal cognitive status was associated with greater improvement in MMSE score. Our analysis indicated that aerobic exercise, especially conducted 30 min per session, less than 150 min per week, and up to three times per week, contributed to improving cognitive function in AD patients. Additionally, a worse basal cognitive status contributed to more significant improvements in cognitive function.


Introduction
Dementia is becoming a growing risk factor among the elderly population as the world's population grows and the problem of population aging intensifies. As of 2019, there were approximately 57.4 million people worldwide with dementia, and it is estimated that there will be more than 152 million in 2050 [1]. This trend will continue to intensify as COVID-19 impacts the public health and healthcare system around the world [2]. Alzheimer's disease (AD) is the most common form of dementia, accounting for 60%-80% of dementia patients [3,4].
AD is an insidious progressive neurodegenerative disease, typically characterized by memory impairment, cognitive impairment, and limitations in daily activities. Atypical symptoms are language, vision, and practical or executive problems that often begin at the onset of symptoms and cannot be effectively identified, leading to further disease progression [5]. This deterioration can lead to neuronal damage in the brain, severe

Eligibility Criteria
Inclusion criteria were as follows: eligible studies (1) should be RCTs; (2) should include both an intervention and control group; (3) should use AD patients as subjects; and (4) should use MMSE as the primary outcome measure.
Exclusion criteria were as follows: (1) non-English language publications; (2) reviews and conference articles; (3) animal model publications; (4) the control group received other treatments.

Data Extraction
Two authors independently extracted the data from each study included first author's last name, year of publication, sample size (n), intervention features (frequency, intensity, duration of intervention, session duration), participant characteristics (age and basal cognitive status), and treatment effects.

Methodological Quality Assessment
The quality of the included studies was evaluated using the Cochrane Collaboration tool according to the 7 domain biases as follows: (1) randomization sequence generation (selection bias), (2) allocation concealment (selection bias), (3) blinding of participants and personnel (performance bias), (4) blinding of outcome assessment (detection bias), (5) incomplete outcome data (attrition bias), (6) selective reporting (reporting bias), and (7) other bias [30]. Methodological quality assessment was performed independently by two authors. Discussions with the third author resolved the disagreements. The methodological quality of all the included studies was assessed by using a Physiotherapy Evidence Database (PEDro) Scale [32,33]. The PEDro Scale is an 11-item scale used to evaluate the quality of the RCTs of the physical therapy studies. The total Physiotherapy Evidence Database scores for RCTs were ranged from 0-10. Scores above 9, between 6 and 8, between 4 and 5, and below 4 are considered excellent, good, average, and poor quality, respectively [34].

Statistical Analysis
We calculated the changes in mean and standard deviation (SD) values of MMSE score. For studies reporting standard error (SE) or 95% confidence intervals (CIs), SD was calculated [22,35]. Data were pooled using random-effects models to obtain the weighted mean difference (WMD) and 95% CIs. If there was a high heterogeneity (I 2 > 50%), subgroup analysis or sensitivity analysis was used to interpret the results [36,37]. In the subgroup analyses, we tried to use minutes of intervention per session (30 min per session and more than 30 min per session), minutes of intervention per week (less than 150 min per week and 150 min or more per week), frequency of intervention per week (up to three times per week and more than three times per week), and basal cognitive status of participants (mild and moderate) to explore the impact on cognitive function. The analysis result, funnel plot, and forest chart were generated using the software RevMan.5. In terms of overall impact, p < 0.05 was considered statistically significant.

Statistical Analysis
We calculated the changes in mean and standard deviation (SD) values of MMSE score. For studies reporting standard error (SE) or 95% confidence intervals (CIs), SD was calculated [22,35]. Data were pooled using random-effects models to obtain the weighted mean difference (WMD) and 95% CIs. If there was a high heterogeneity (I 2 > 50%), subgroup analysis or sensitivity analysis was used to interpret the results [36,37]. In the subgroup analyses, we tried to use minutes of intervention per session (30 min per session and more than 30 min per session), minutes of intervention per week (less than 150 min per week and 150 min or more per week), frequency of intervention per week (up to three times per week and more than three times per week), and basal cognitive status of participants (mild and moderate) to explore the impact on cognitive function. The analysis result, funnel plot, and forest chart were generated using the software RevMan.5. In terms of overall impact, p < 0.05 was considered statistically significant.

Characteristics of the Included Studies
The main characteristics of the participants are shwon in Table S1. Of the fifteen studies, four were based in United States [45,[47][48][49], two each in Denmark [43,46] and Italy [20,41], and one each in South Korea [44], China [27], Brazil [42], Canada [38], Germany [40], and France [28]. All included studies contained a sample size of 909 subjects diagnosed with AD. Aerobic exercise and control groups were composed of 503 and 406 subjects, respectively. The 15 included studies were published between 2002 and 2020. Of the 15 included studies, 4 studies performed moderate-intensity to high-intensity aerobic exercise [39,41,43,46], 3 studies did not describe exercise intensity [20,38,40], and the remaining included studies performed moderate-intensity aerobic exercise [27,28,42,44,45,[47][48][49]. The overall duration varied from 9 weeks to 26 weeks, with an average intervention duration of 18.13 weeks. The frequency of interventions per week ranged from 2-5, with an average of 3.66 per week. The intervention time for each session varied from 30 to 60 min, with an average of 46.66 min per session. Interventions were administered at least 60 min per week, with an average weekly intervention time of 160 min.

Risk of Bias
The quality of the included studies was evaluated using the Cochrane Collaboration tool according to the following six aspects: selection bias, performance bias, detection bias, attrition bias, reporting bias, and other bias. The quality of included studies is shown in Figure 2. Possible publication bias was assessed by examining the funnel plot ( Figure 3), while the results of the Egger's test indicated that the small sample size studies of the included study were not enough to affect the final results (p = 0.381, Table 1).

Characteristics of the Included Studies
The main characteristics of the participants are shwon in Table S1. Of the fifteen studies, four were based in United States [45,[47][48][49], two each in Denmark [43,46] and Italy [20,41], and one each in South Korea [44], China [27], Brazil [42], Canada [38], Germany [40], and France [28]. All included studies contained a sample size of 909 subjects diagnosed with AD. Aerobic exercise and control groups were composed of 503 and 406 subjects, respectively. The 15 included studies were published between 2002 and 2020. Of the 15 included studies, 4 studies performed moderate-intensity to high-intensity aerobic exercise [39,41,43,46], 3 studies did not describe exercise intensity [20,38,40], and the remaining included studies performed moderate-intensity aerobic exercise [27,28,42,44,45,[47][48][49]. The overall duration varied from 9 weeks to 26 weeks, with an average intervention duration of 18.13 weeks. The frequency of interventions per week ranged from 2-5, with an average of 3.66 per week. The intervention time for each session varied from 30 to 60 min, with an average of 46.66 min per session. Interventions were administered at least 60 min per week, with an average weekly intervention time of 160 min.

Risk of Bias
The quality of the included studies was evaluated using the Cochrane Collaboration tool according to the following six aspects: selection bias, performance bias, detection bias, attrition bias, reporting bias, and other bias. The quality of included studies is shown in Figure 2. Possible publication bias was assessed by examining the funnel plot ( Figure 3), while the results of the Egger's test indicated that the small sample size studies of the included study were not enough to affect the final results (p = 0.381, Table 1).   The standard assessment of the 15 studies included in this systematic review using the PEDro scale is shown in Table 2. Of the fifteen included studies, twelve were of good quality [28] and three were of fair quality [28]. A limitation of the methodological quality of all studies was the lack of clarity on patients blinding. The mean PEDro score for all included studies was 6.33. Table 2. Methodological assessment of randomized controlled trials included in the systematic review using the PEDro scale.  The standard assessment of the 15 studies included in this systematic review using the PEDro scale is shown in Table 2. Of the fifteen included studies, twelve were of good quality and three were of fair quality [34]. A limitation of the methodological quality of all studies was the lack of clarity on patients blinding. The mean PEDro score for all included studies was 6.33.

Effects of Aerobic Exercise on Cognitive Function in AD Patients
After analyzing the data of all included studies, we found that aerobic exercise had a significant effect on increasing MMSE score in AD patients [WMD, 1.50 (95% CI, 0.55 to 2.45), p = 0.002, Figure 4], while there was significant heterogeneity (I 2 = 87%). To explain the heterogeneity between eleven studies and find modifiable factors of aerobic exercise, four further subgroup analyses were further performed.

Subgroup Analysis
Different results were observed when considering minutes of intervention per session (

Effects of Aerobic Exercise on Cognitive Function in AD Patients
After analyzing the data of all included studies, we found that aerobic exercise had a significant effect on increasing MMSE score in AD patients [WMD, 1.50 (95% CI, 0.55 to 2.45), p = 0.002, Figure 4], while there was significant heterogeneity (I 2 = 87%). To explain the heterogeneity between eleven studies and find modifiable factors of aerobic exercise, four further subgroup analyses were further performed.

Effects of Aerobic Exercise on Cognitive Function in AD Patients
After analyzing the data of all included studies, we found that aerobic exercise had a significant effect on increasing MMSE score in AD patients [WMD, 1.50 (95% CI, 0.55 to 2.45), p = 0.002, Figure 4], while there was significant heterogeneity (I 2 = 87%). To explain the heterogeneity between eleven studies and find modifiable factors of aerobic exercise, four further subgroup analyses were further performed.  In addition, different results were observed when considering minutes of intervention per week ( Figure 6). Specifically, interventions conducted less than 150 min per week increased MMSE score significantly [WMD, 2.10 (95% CI, 0.84 to 3.37), I 2 = 92%, p = 0.001]. However, interventions conducted 150 min or more per week exercise had no

Discussion
The aim of this study was to explore the effect of aerobic exercise on cognitive function in AD patients. From 1942 search records initially identified, 15 studies were considered eligible for systematic review and meta-analysis. Our results showed that there was a significant effect of aerobic exercise on increasing MMSE score in AD patients. Subgroup analyses showed that interventions conducted 30 min per session, less than 150 min per week, and up to three times per week increased MMSE score significantly. In addition, a worse basal cognitive status was associated with greater improvement in MMSE score.

Effects of Aerobic Exercise on Cognitive Function
This systematic review and meta-analysis indicated that aerobic exercise had the potential to improve the cognitive function in AD patients, as manifested by an improvement in MMSE score. Our study showed that aerobic exercise contributed to an overall improvement in MMSE score by 1.50 (WMD), which was of clinical importance for AD patients. We noticed that under the theme of exercise intervention in AD patients, most exercise interventions were aimed at increasing the patients' physical activity. The intervention methods included resistance, aerobic, balance, and combined exercise. Despite differences in exercise interventions, studies have demonstrated the link between aerobic exercise and improved cognitive function in AD patients [50][51][52].
Aerobic exercise can improve aerobic fitness, reduce the risk of chronic disease in older adults, and increase life expectancy. There are a growing number of studies suggesting that aerobic exercise can promote neuroplasticity, induce increased hippocampal neurogenesis [53,54], and protect against or reverse age-related hippocampal atrophy [55]. Regular exercise improves cerebrovascular and endothelial function, reduces oxidative stress and systemic inflammation, and helps regulate immune function, which may contribute to improved neuronal function [56][57][58]. Although the mechanism of aerobic exercise improving cognitive function had not been fully revealed, it was speculated that the beneficial effect of exercise on cognitive function might be strengthened through the following mechanisms.
Firstly, studies have shown that aerobic exercise can increase the level of brainderived neurotrophic factor (BDNF), which is involved in neuroprotection and promotes cell survival, neurite outgrowth, and synaptic plasticity [59][60][61] associated with cognitive function. In addition, aerobic exercise can also increase the brain' uptake of circulating insulin-like growth factor-1 (IGF-1), a factor that promotes neuronal differentiation of

Discussion
The aim of this study was to explore the effect of aerobic exercise on cognitive function in AD patients. From 1942 search records initially identified, 15 studies were considered eligible for systematic review and meta-analysis. Our results showed that there was a significant effect of aerobic exercise on increasing MMSE score in AD patients. Subgroup analyses showed that interventions conducted 30 min per session, less than 150 min per week, and up to three times per week increased MMSE score significantly. In addition, a worse basal cognitive status was associated with greater improvement in MMSE score.

Effects of Aerobic Exercise on Cognitive Function
This systematic review and meta-analysis indicated that aerobic exercise had the potential to improve the cognitive function in AD patients, as manifested by an improvement in MMSE score. Our study showed that aerobic exercise contributed to an overall improvement in MMSE score by 1.50 (WMD), which was of clinical importance for AD patients. We noticed that under the theme of exercise intervention in AD patients, most exercise interventions were aimed at increasing the patients' physical activity. The intervention methods included resistance, aerobic, balance, and combined exercise. Despite differences in exercise interventions, studies have demonstrated the link between aerobic exercise and improved cognitive function in AD patients [50][51][52].
Aerobic exercise can improve aerobic fitness, reduce the risk of chronic disease in older adults, and increase life expectancy. There are a growing number of studies suggesting that aerobic exercise can promote neuroplasticity, induce increased hippocampal neurogenesis [53,54], and protect against or reverse age-related hippocampal atrophy [55]. Regular exercise improves cerebrovascular and endothelial function, reduces oxidative stress and systemic inflammation, and helps regulate immune function, which may contribute to improved neuronal function [56][57][58]. Although the mechanism of aerobic exercise improving cognitive function had not been fully revealed, it was speculated that the beneficial effect of exercise on cognitive function might be strengthened through the following mechanisms.
Firstly, studies have shown that aerobic exercise can increase the level of brain-derived neurotrophic factor (BDNF), which is involved in neuroprotection and promotes cell survival, neurite outgrowth, and synaptic plasticity [59][60][61] associated with cognitive function. In addition, aerobic exercise can also increase the brain' uptake of circulating insulinlike growth factor-1 (IGF-1), a factor that promotes neuronal differentiation of progenitor cells [60,62]. Secondly, studies have shown that the pathogenesis of AD may be related to circadian rhythm arrhythmia [63]. AD patients often exhibit sleep fragmentation and abnormal circadian rhythm, and aerobic exercise can help improve circadian rhythm and sleep quality in AD patients, which in turn improves cognitive function [57]. Finally, aerobic exercise can improve vagal regulation by increasing heart rate variability (HRV), the vagal control of the cardiovascular system is associated with activity in the prefrontal cortex, a brain structure that supports executive function [64,65], which is associated with cognitive function [66].

Subgroup Analysis
In the studies we included, aerobic exercise significantly increased MMSE score in AD patients, but the heterogeneity between groups was high. Therefore, we used subgroup analysis to interpret the results. In the subgroup analyses, we sought to determine the effects of minutes of intervention per week, minutes of intervention per session, frequency of intervention per week, and the basal cognitive status of participants.
Aerobic exercise has a dose-response relationship for health effects, and appropriate loading is the key to the health benefits of exercise [67]. The cognitive benefits of aerobic exercise appear to apply here as well. The American College of Sports Medicine (ACSM) and the American Heart Association (AHA) recommend that older adults engage in at least 30 min of moderate-intensity aerobic exercise 5 days a week or at least 20 min of vigorous-intensity aerobic exercise 3 days a week [68]. A recently published expert consensus recommends that the frequency of aerobic exercise for older adults is 20-60 min, 3-7 days per week [67]. However, our results showed that interventions conducted 30 min per session had a significant effect on improving cognitive function, while interventions conducted more than 30 min per session had no significant associations with MMSE score in AD patients, which was consistent with the results of Jia et al. [10], showing that interventions conducted up to 30 min per session showed greater effectiveness for improving cognition of patients compared to those conducted more than 30 min per session. Many studies have shown that 20 minutes per session can significantly improve cognitive function in older adults [17,69]. Exercise is too short to induce changes in the level of physical arousal, brain structure, and brain function, while a long duration may lead to excessive fatigue and will not induce brain plasticity [70]. Therefore, it is important to determine the duration for which such changes are most effective [37]. Future studies should elucidate the effect of intervention-based exercise duration.
The World Health Organization (WHO) recommends that all adults perform 150-300 min of moderate-intensity aerobic exercise, 75-150 min of vigorous-intensity aerobic exercise, or an equivalent combination of moderate-intensity and vigorous-intensity aerobic exercise per week [71]. The WHO indicated that this recommendation also applies to older adults. Our subgroup analysis showed that interventions conducted less than 150 min per week increased MMSE score significantly, while interventions conducted 150 min or more per week had no significant associations with MMSE score in AD patients, which was in consistent with the results of Groot et al. [72], showing that although both interventions conducted up to 150 min per week and more than 150 min per week yielded a positive effect on cognitive function, interventions conducted up to 150 min per week appeared to be more effective. One study showed that the minimum threshold for cognitive improvement in healthy older adults was 90 min of physical activity per week [73]. In addition, Jia et al. [10] reported that interventions conducted up to 120 min per week had a tendency to show greater effectiveness for improving cognition of AD patients compared to those conducted more than 120 min per week. Considering the physical activity capacity of older AD patients, most RCTs set the weekly interventions at up to 150 min to reduce participant dropout and potential risk [74], which may lead to a conservative trend in the number of studies with exercise interventions more than 150 min per week.
As for frequency of intervention per week, our results showed that interventions conducted up to three times per week increased MMSE score significantly, while interventions conducted more than three times per week had no significant associations with MMSE scores in AD patients, which was in consistent with the results of Jia et al. [10], showing that interventions conducted up to three times per week showed greater effect on improving cognition of AD patients compared to those conducted more than three times per week, which indicated that high-frequency interventions did not have better cognitive effects than low-frequency interventions [10]. In addition, Cai et al. [70] also showed that low-frequency exercise (one or two times per week) and moderate-frequency exercise (three or four times per week) had a positive exercise effect on working memory in older adults, while high-frequency exercise (≥five times per week) had no such positive effect. Moreover, the effect size for older adults engaged in low-frequency exercise was larger than that for those engaged in moderate-frequency exercise or high-frequency exercise. We noted that the use of exercise intervention frequency alone cannot exclude the influence of other confounding factors, such as the duration of intervention per session and the duration of intervention per week, which may influence effects of aerobic exercise on cognitive function. Therefore, the best combination is yet to be discovered, and current evidence is insufficient to provide precise recommendations regarding the frequency of exercise interventions.
MMSE scores indicate the degree of cognitive impairment, including severe (0-10), moderate (10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20), mild (20)(21)(22)(23)(24)(25), and questionably significant (25)(26)(27)(28)(29)(30) [75]. In the above results, we suspected that the improvement of cognitive function by aerobic exercise was influenced by the basal cognitive status of participants, so this study divided the included studies into two subgroups, namely the mild group and moderate group. Our results showed that aerobic exercise increased MMSE score significantly in both mild and moderate AD patients, while a worse basal cognitive status was associated with greater improvement in MMSE score. The cognitive decline rate assessed by MMSE was 3 points per year, with the slowest rate of decline in mild patients, followed by moderate, and the fastest rate of decline in severe patients [76]. One previous study in patients with moderate cognitive impairment has reported a faster cognitive decline in the control group than in the intervention group [20], and thus the cognitive benefits of aerobic exercise can be determined more sensitively. However, since studies included in this systematic review and meta-analysis focused on mild and moderate AD patients, and studies focused on severe patients need to be cautious when referring to our findings.

Limitations of the Review
This study has some limitations that should be considered. Firstly, the included studies were all randomized controlled trials of aerobic exercise intervention, which could not be completely blinded. Therefore, in the process of quality evaluation, subjective factors would cause a certain degree of deviation. Secondly, some RCT research protocols did not describe the exercise intensity, so adequate statistics on the effect of aerobic exercise intensity on the cognitive function in AD patients were not available. Finally, the types of exercise interventions across the included studies were consistent, but the details of the interventions varied widely, which may be an important source of heterogeneity.

Conclusions
This systematic review and meta-analysis indicated that aerobic exercise, especially conducted 30 min per session, less than 150 min per week, and up to three times per week, contributed to improving cognitive function in AD patients. Additionally, a worse basal cognitive status contributed to more significant improvements in cognitive function.
Author Contributions: S.Z. and K.Z. wrote the manuscript. L.Y. and Y.L. contributed to the conception and design. S.Z. and K.Z. extracted data the data and evaluated the quality. Q.S. and Y.C. verified the data. S.Z., K.Z., Q.S., Y.C., Y.L. and L.Y. contributed to the analysis and interpretation of the data. S.Z., K.Z., Q.S., Y.C., Y.L. and L.Y. revised it critically for important intellectual content. All authors have read and agreed to the published version of the manuscript.