PhysAgeNet Deliverable D1.2 Systematic reviews using the database D1.1 submitted for publication related to PA interventions and EBM

1 CA20104 – Network on evidence-based physical activity in old age (PhysAgeNet) Deliverable D1.2 Systematic reviews using the database D1.1 submitted for publication related to PA interventions and EBM Contributors Working Group 1: Implementation and development of standards from evidence-based medicine (EBM) 1.3.1. Chronic exercise and depressive symptoms Leaders of the subgroup: Melanie Mack (CHE) and Michel Audiffren (FRA) Members of the subgroup: Claudia Voelcker-Rehage (DEU), Pinelopi Stavrinou (CYP), Julia Pavlova (UKR), Yaël Netz (ISR), Burcu Kömürcü (TUR), Antonia Kaltsatou (GRC), Sandra Haider (AUT), Christoforos Giannaki (CYP), Arzu Erden (TUR), Andreea Badache (SWE). Description of the activities of the subgroup: The subgroup “Chronic exercise and depressive symptoms” has met 10 times since the summer of 2022. The slides presented during the meetings were posted on the Riga Stradins University Teams space. The subgroup already completed the following operations: (1) search of randomized controlled trials examining the effect of chronic exercise on depressive symptoms in older adults in several databases (30,765 articles found), (2) uploading the references of these 30,765 articles in Rayyan, (3) removal of duplicates (10,347 duplicates removed), (4) selection of studies based on abstracts by pairs of reviewers (19,845 articles excluded), (5) selection of studies based on full-text by pairs of reviewers (411 articles excluded), (6) preparation of the questionnaire to extract the data on REDCap, (7) extraction of a sample of selected data and storage of these data in REDCap by pairs of reviewers, (8) conducting the meta-analysis on Comprehensive meta-analysis v4. The first results are presented in section 3. The review protocol has been registered on PROSPERO [CRD42022361418] and accepted for publication in PLOS One. The next stages to finish the systematic review and meta-analysis are described in the two Gantt charts hereafter. The submission of the final article is planned in January 2025. Timeline of the activities for 2024 and 2025: 2 …. 2. Results of the meta-analysis on the chronic effects of exercise on overall depression level in older adults As mentioned in section 1.3.1, the protocol of the present meta-analysis has been registered on PROSPERO [CRD42022361418] and accepted for publication in PLOS One (doi: 10.1371/journal.pone.0297348). We present hereafter the main results of the meta-analysis that will be published in a indexed international journal. 2.1. General results The analysis is based on 162 effect sizes (129 studies). A table with the references of the 129 studies is provided in appendix 1. The effect size index is the standardized difference in means (Cohen’s d). Computations were carried out using Comprehensive Meta-Analysis Version 4 (Borenstein et. al., 2022). 2.1.1. Statistical model The random-effects model was employed for the analysis. The studies in the analysis are assumed to be a random sample from a universe of potential studies, and this analysis will be used to make an inference to that universe (Borenstein, 2019; Borenstein et al., 2010; Borenstein et al., 2021; Hedges & Vevea, 1998; Higgins & Thomas, 2019). 2.1.2. What is the mean effect size? The mean effect size of the effect of chronic exercise on overall depression level is -0,675 with a 95% confidence interval of -0,779 to -0,572 (see Figure 1). The mean effect size in the universe of comparable studies could fall anywhere in this interval. This moderate effect size indicates that chronic exercise reduces the overall depression level by 0.675 standard deviation. 3 Figure 1: Caterpillar plot of the 162 effect sizes included in the meta-analysis The Z-value tests the null hypothesis that the mean effect size is zero. The Z-value is -12,776 with p < 0,001. Using a criterion alpha of 0,050, we reject the null hypothesis and conclude that in the universe of populations comparable to those in the analysis, the mean effect size is not precisely zero. 2.1.3. The Q-test for heterogeneity The Q-statistic provides a test of the null hypothesis that all studies in the analysis share a common effect size. If all studies shared the same true effect size, the expected value of Q would be equal to the degrees of freedom (the number of studies minus 1). The Q-value is 1143,668 with 161 degrees of freedom and p < 0,001. Using a criterion alpha of 0,100, we can reject the null hypothesis that the true effect size is the same in all these studies. 2.1.4. The I-squared statistic The I-squared statistic is 86%, which tells us that some 86% of the variance in observed effects reflects variance in true effects rather than sampling error. 2.1.5. Tau-squared and tau Tau-squared, the variance of true effect sizes, is 0,348 in d units. Tau, the standard deviation of true effect sizes, is 0,590 in d units. 2.1.6. The prediction interval If we assume that the true effects are normally distributed (in d units), we can estimate that the prediction interval is -1,845 to 0,495 (see Figure 2). The true effect size in 95% of all comparable populations falls in this interval (Borenstein, 2019, 2020; Borenstein et al., 2021; Borenstein et al., 2017; DerSimonian & Laird, 1986, 2015; Higgins, 2008; Higgins & Thompson, 2002; Higgins et al., 2003; Higgins & Thomas, 2019; IntHout et al., 2016). 4 Figure 2: Prediction interval of true effects and confidence interval of the mean effect size. !"#$$ !%#&$ !%#$$ !$#&$ $#$$ $#&$ %#$$ %#&$ "#$$ DE)*+E,-*E./0.102+-3041135*) S*T/8T+8E93808E113+3/530E/0:3T/)0;8< 2=30:3T/031135*0)E930E)0!$#>?0@E*=0T0A&B05./1E83/530E/*3+CTa0.10!$#b?0*.0!$#&b 2=30*+-3031135*0)E930E/0A&B0.10Taa05.:cT+T,a30c.c-aT*E./)01Taa)0E/0*=30E/*3+CTa0!%#?&0*.0$#dA 5 2.2. Moderator analyses 2.2.1. Exercise characteristics 2.2.1.1. Moderator: Type of exercise Rationale: Numerous recent meta-analyses have been conducted to investigate the differential effects of various types of exercise on overall depression levels in healthy older adults (Miller et al., 2020), and in older adults with Mild Cognitive Impairment (MCI) (Liu et al., 2023). For older adults, both with and without MCI, it was found that mind-body exercises were most beneficial in reducing depression levels (Liu et al., 2023; Miller et al., 2020). However, resistance, aerobic, and multicomponent exercises also proved effective, albeit slightly less so, still demonstrating moderate effect sizes. Therefore, we hypothesize that all the types of exercise included in our study are effective in reducing depressive symptoms, though through different mechanisms. Mind-body exercises, which combine lowintensity muscular activities (such as flexibility or balance training) with an internally guided focus fostering a self-contemplative state of mind, are hypothesized to be particularly effective. Results: A subgroup analysis, including 162 effect sizes, revealed that the type of exercise significantly explained the heterogeneity of the effect sizes, Q (6) = 19.830, p = .003. The efficacy of different types of exercise in reducing overall depression levels can be ranked in the following order: (1) exergames (n = 6; SMD = -1.705; 95 % CI = [-2.789, -0.622]), (2) resistance exercises (n = 27; SMD = -0.973; 95 % CI = [-1.244, -0.702]), (3) mind-body exercises (n = 33; SMD = -0.733; 95 % CI = [-0.979, -0.568]), (4) aerobic exercises (n = 30; SMD = -0.625; 95 % CI = [-0.947, -0.304]), (6) Coordinative exercises (n = 3; SMD = -0.529; 95 % CI = [-0.965, -0.093]), (7) Multi-component exercises (n = 55; SMD = -0.460; 95 % CI = [- 0.602, -0.318]), (8) Dance exercises (n = 8; SMD = -0.378; 95 % CI = [-0.087, -0.069]). Discussion: Contrary to our hypothesis and the results of the meta-analyses by Miller et al. (2020) and Liu et al. (2023), we did not find mind-body exercise to be the most effective type of exercise. However, similar to those studies, the effect sizes between mind-body exercises, aerobic exercise, and strength exercises were comparable (ranging from -0.97 to -0.63). Interestingly, for exergames, a type not included in the meta-analyses by Miller et al. (2020) and Liu et al. (2023), we found them to be the most effective, with an effect size of -1.71. This beneficial effect aligns with the findings of Yen & Chui (2021), who reported large effects of exergames on depressive outcomes in older adults. It might be suggested that exergames benefit the mental health of older adults by not only including cognitive and physical exercises but also promoting social interactions with other actors such as players, therapists, avatars, and animals. However, the within-study heterogeneity was quite high. Therefore, further research is needed to more specifically identify which aspects of exergames moderate the beneficial effect on depression in older adults. 2.2.1.2. Moderator: Intensity of exercise Rationale: The level of exercise intensity can significantly affect the type and extent of psychological and physiological adaptation to exercise training and can impact the human body's internal metabolic and cardiovascular stress (Mann, Lamberts, & Lambert, 2013; Xie et al., 2021). Regarding the anti-depressant properties of exercise, recent research has 6 shown that the intensity of exercise plays a significant role in the anti-depressant effect of various types of exercise, such as strength training, resistance training, yoga, and tai chi (Nebiker et al., 2018; Noetel et al., 2024). For instance, increasing the intensity of exercise by 10% has been found to have a notable positive effect on depression levels (Nebiker et al., 2018). Therefore, we hypothesized that the higher the exercise intensity, the larger the effect of exercise training intervention on overall depression levels. Results: There was a significant effect of the intensity of exercise on the effect size: Q (2) = 7.180; p = .028. The exercises with a vigorous intensity or higher (n = 25; SMD = -1.138; 95 % CI = [-1.535, -0.740]) led to a larger effect size than moderate (n = 80; SMD = -0.570; 95 % CI = [-0.707, -0.432]) and light intensity exercises (n = 57; SMD = -0.676; 95 % CI = [-0.842, - 0.509])). The difference between the mean effect size of light and moderate intensity exercises: Q (1) = 0.922; p = .337. Discussion: The results of the current analysis revealed that exercise training prescriptors, such as exercise intensity, can significantly impact the level of amelioration of overall depression symptoms in older people. The results of this study support recent meta-analyses on the relationship between exercise training and depression in younger populations that reported that high-intensity exercise interventions are more effective in reducing depression compared to moderate or light-intensity ones (Nebiker et al., 2018; Noetel et al., 2024). In conclusion, exercise intensity has a moderating effect on overall depression levels in older people, with high intensities being more effective than moderate and light intensities. 2.2.1.3. Moderator: Duration of exercise sessions Rationale: Based on a previous meta-analysis examining the effects of chronic exercise on cognitive aging (Colcombe & Kramer, 2003), we can hypothesize that RCTs that use moderate duration sessions lead to larger effect sizes because it could be considered that the optimal duration to maintain a high level of engagement of older participants is between 30 and 60 minutes. RCTs using shorter or longer intervention sessions could be considered as less optimal and leading to smaller effect sizes. Short session RCTs would not be sufficiently long to induce large and enduring changes, whereas long session RCTs would introduce acute fatigue effects. To perform this analysis, we distinguished four categories of studies: (1) short duration sessions (< 30 min), (2) moderate duration sessions (30 < duration < 60 min), (3) long duration session (> 60 min), (4) session duration not specified. Results: A subgroup analysis, including 162 effect sizes, revealed that the duration of exercise sessions did not significantly explain the heterogeneity of the effect sizes, Q (3) = 2.562, p = .464. There was no significant difference between the four categories of RCTs: RCTs using short duration sessions (n = 16; mean duration = 18.79 min; SMD = -0.434; 95 % CI = [-0.767, -0.101]), RCTs using moderate duration sessions (n = 90; mean duration = 48.95 min; SMD = -0.660; 95 % CI = [-0.794, -0.525]), RCTs using long duration sessions (n = 13; mean duration = 95.36 min; SMD = -0.687; 95 % CI = [-0.985, -0.389]), RCTs that did not specify the duration of sessions (n = 18; SMD = -0.482; 95 % CI = [-0.778, -0.186]). Discussion: Contrary to our hypothesis, the duration of the exercise sessions did not influence the size of the effect of chronic exercise on overall depression level. However, it can be noticed that moderate durations are the most used by researchers (75%). 7 2.2.1.4. Moderator: Frequency of exercise sessions Rationale: The meta-analysis of Sanders et al. (2019) suggested that the more frequent the exercise sessions were, the greater the benefit in cognitive function in older adults. Using the same logic, we can hypothesize that the more frequent the exercise sessions, the larger the reduction in overall depression level. Results: The meta-regression, including 145 effect sizes, indicated that the frequency of exercise sessions did not explain the heterogeneity of the effect sizes of chronic exercises on overall depression level (β = -0.008, R² = .00, p = .795). Discussion: Contrary to our expectations, we did not find a relationship between the frequency of exercise sessions and the effect size of the interventions. This moderator should not be considered alone, but certainly in combination with session duration and duration of the intervention. 2.2.1.5. Moderator: Total duration of the intervention (in weeks) Rationale: Recent systematic reviews and meta-analyses are inconclusive regarding the effect of intervention duration on participants' depressive episodes (Kandola et al., 2019; Singh et al., 2023). It is noted that the effectiveness of physical activity interventions may decrease with increasing duration of interventions. Longer duration interventions had smaller effects if compare with midand short duration, but the longest duration interventions still had positive effects (Singh et al., 2023). Within clinical populations interventions lasting 10–16 weeks lead to greater effects than interventions lasting 4–9 weeks (Lee et al., 2022; Rethorst et al., 2009; Singh et al., 2023). However, in the general population, interventions lasting 4–9 weeks resulted in significantly larger effects than interventions lasting 17–26 weeks, and interventions lasting 10–16 weeks also produced significantly larger effects than interventions lasting 16–26 weeks and more than 26 weeks (Rethorst et al., 2009). Considering the results of research, we hypothesized that the duration of the intervention is positively correlated with the beneficial effects of exercise until a certain duration. Results: The meta-regression, including 155 effect sizes, indicated a marginal relationship between the duration and the effect size of chronic exercises on overall depression level (β = -0.0001, R² = .00, p = .05456). This indicated that the longer the duration of the intervention in the studies, the larger the beneficial effects of chronic exercise on the overall depression level (see Figure 3). This result matches with that found by Yen et al. (2021) who observed that the total intervention duration had a significant effect on depressive outcomes, the longest duration leading to a greater efficiency against depressive outcomes (β = -0.255; p = 0.11). 8 Figure 3: regression line between standardized difference in means of level of depression as a function of duration of intervention in the RCT. Discussion: Our meta-regression did not find a significant relationship between the duration of intervention and the effect of chronic exercise on overall depression level. This result is inconsistent with earlier studies (Craft & Landers, 1998; North et al., 1990) according to which longer intervention resulted in larger decreases in depression scores, while also being consistent with more recent work on the lack of a linear relationship (Lee et al., 2022; Rethorst et al., 2009; Singh et al., 2023). One of the possible mechanisms explaining this effect may be the insufficient duration of the interventions, as well as the number of such interventions. Most exercise interventions for depression are around 16 weeks, and this period may only be sufficient to induce functional connectivity changes (Rethorst et al., 2009). Exercise must last for at least six months to induce structural changes (for example, in hippocampal volume), longer exercise interventions may induce structural changes that have a more pronounced impact on depressive symptoms. Also exercise produces several functional, but transient changes in neuroplasticity, such as elevations in BDNF circulation, improvements in cardiorespiratory fitness are necessary for more lasting structural changes and cognitive benefits (Kandola et al., 2019). 2.2.1.6. Moderator: Volume of the intervention Volume = Session frequency (NO sessions per week) * Session duration (in min) * Program duration (in weeks) Rationale: Among the number of studies that investigate the protective role of physical activity on depression, results are largely heterogeneous (Noetel et al., 2024; Singh et al., 2023; Wang et al., 2019). Among the findings is the ascertainment of the fact of a positive effect of physical activity on reducing the severity and symptoms of depression (Singh et al., 2023). Other studies have examined physical activity across different parameters, including frequency, duration, and volume, and have produced more complex results. Some findings 9 indicated that individuals who participated in physical activity with greater frequency were less likely to be depressed; some researches that lower frequency or PA with a lower volume is a protective factor (Wang et al., 2019). We hypothesize that chronic exercise interventions differ in their efficacy in reducing overall depression severity in older adults across interventions with different exercise characteristics (i.e., frequency of exercise sessions, duration of exercise session, and duration of intervention). Results: The meta-regression, including 141 effect sizes, indicated a linear relationship between the volume of intervention and the effect size of chronic exercise on overall depression level (β = -0.0001, R² = .00, p = .04). This indicated that the larger the volume of the intervention in the studies, the larger the effects of chronic exercise on the overall depression level (see Figure 4). Discussion: Our meta-regression indicated a significant relationship between intervention volume and the effect of chronic exercise on overall depression level. This result is consistent with some previous studies (Wang et al., 2019). At the same time the lack of consensus on this issue indicates that the relationship between physical activity and depression may be dependent on cultural and socio-demographic indicators (Olanrewaju et al., 2016), and also confirms the multidimensional nature of exercise training that consists of intensity, frequency, and duration of intervention (Falck et al., 2016). Figure 4: regression line between standardized difference in means of level of depression as a function of volume of intervention in the RCT. 2.2.1.7. Moderator: Cognitive demand of exercises Rationale: Two categories of exercise have been distinguished based on the cognitive load of the exercises included in the exercise programs. On the one hand, high cognitive demanding exercises include balance exercises, coordinative exercises, mind-body exercises, exergames and cognitive tasks. On the other hand, low cognitive demanding exercises include aerobic exercises and resistance exercises. Based on previous works, we know that depressed 16 complete-case analysis, the number of participants is the same at the beginning and at the end of the intervention for all the groups or the authors only analyzed data from participants who completed the study and had no missing data. Finally, in RCTs using a per-protocol analysis, only the participants who adhere to the treatment are included in the analysis to test the efficacy of the treatment. Most clinical researchers and statisticians agree that the primary analysis of data in a randomised controlled trial should compare patients according to the group to which they were randomly allocated, regardless of patients' compliance, crossover to other treatments, or withdrawal from the study. Such an analysis is referred to as an intention to treat analysis. Proponents argue that the intention to treat approach helps preserve prognostic balance in the study arms, limits inferences based on arbitrary or ad hoc sub-groups of patients in the trial, emphasizes greater accountability for all patients entered into the study and consequently minimizes the influence of withdrawals, non-compliers, and patients lost to follow up, is the most cautious approach and so minimizes type 1 error, and finally allows for the greatest generalizability (Fergusson et al., 2002). Critics say, however, that an intention to treat approach is too cautious and more susceptible to type II error (Sommer & Zeger, 1991; Rubin, 1998). Consequently, we can hypothesize that intention-totreat analyses have the tendency to underestimate the effect size (type II error) and perprotocol analysis to overestimate the effect size (type I error). Results: A subgroup analysis, including 162 effect sizes, revealed that the type of statistical analysis significantly explains the heterogeneity of the effect sizes, Q (2) = 8.724, p = .013. Pairwise comparisons revealed that RCTs using a complete-case analysis (n = 119; SMD = - 0.744; 95 % CI = [-0.873, -0.615]) led to a larger effect size than RCTs using an intention-totreat analysis (n = 32; SMD = -0.421; 95 % CI = [-0.603, -0.238]): Q (1) = 8.009, p = .005. RCTs using a per-protocol analysis (n = 11; SMD = -0.851; 95 % CI = [-1.339, -0.364]) obtained the largest mean effect size as expected, but it was not significantly different from the mean effect size obtained with the two other types of analysis. Discussion: Contrary to our hypothesis, we did not observe a significant difference between the effect sizes obtained in RCTs using an intention-to-treat analysis and those using a perprotocol analysis. In contrast, the RCTs using complete-case analysis, which were the most frequent category (73%), showed a larger effect size than RCTs using intention-to-treat analysis. Consequently, to have a clear idea of the efficacy of an exercise program it would be interesting to conduct the three analyses: one replacing missing values of people who dropped out (intention-to-treat analysis), one analyzing only participants with no missing value (complete-case analysis) and one analyzing only participants who adhere to the different interventions (per-protocol analysis). 2.2.3.3. Moderator: Study design Rationale: Two study designs, or more precisely two different methods of randomization, have been distinguished. In the parallel-group design, each participant is randomly assigned to a group, and all participants receive (or do not receive) an intervention. In the cluster design, pre-existing groups of participants (e.g., villages, nursing homes) are randomly selected to receive (or not receive) an intervention. This analysis is exploratory and we have no a priori hypothesis concerning this moderator. 17 Results: A subgroup analysis, including 162 effect sizes, revealed that the study design did not significantly explain the heterogeneity of the effect sizes, Q (1) = 0.082, p = .775. In other words, the study design does not significantly influence the effect size. RCTs using a parallelgroup design (n = 148; SMD = -0.674; 95 % CI = [-0.786, -0.562]) did not led to a significantly different effect size than RCTs using a cluster-design (n = 14; SMD = -0.720; 95 % CI = [-1.018, -0.422]). Discussion: The type of study design (parallel-group vs. cluster) does not seem to be an important determinant of the effect size of chronic exercise interventions on overall depression level. The parallel-group design remains the most used design in the research community (91%). 2.3. Conclusion The first results of the present meta-analysis clearly showed that exercise-based interventions are effective to reduce overall depression level leading to a moderate effect size (d = -0,675). According to moderator analyses, the most effective exercise program should include exercises with the following characteristics: high cognitive demand (e.g., exergames or mind-body exercise), vigorous or high intensity, supervised group-based. In addition, interventions with a large volume of training are more effective than interventions with a lower volume. Concerning, the population more responsive to exercise interventions with the aim to reduce overall depression level, the moderator analyses suggest that the higher the age and depressive symptoms at baseline, the smaller the effect size; i.e., with aging, it is more and more difficult to reduce depression through exercise intervention. Based on this resuls, evidence-based recommendations will be formulated. The present results will be shortly completed by several analyses: analysis of the interactions between pairs of moderators, analysis of the effect of the risk of bias on the effect size, analysis of the effect of publication bias on the effect size, analysis of the effect of chronic exercise on different symptoms of depression (e.g., apathy, fatigue, sleep). References Arrieta, H., Rezola-Pardo, C., Echeverria, I., Iturburu, M., Gil, S. M., Yanguas, J. J., ... & Rodriguez-Larrad, A. (2018). Physical activity and fitness are associated with verbal memory, quality of life and depression among nursing home residents: preliminary data of a randomized controlled trial. BMC Geriatrics, 18, 1-13. doi: 10.1186/s12877-018-0770y Bae, S., Jang, M., Kim, G. M., Yang, J. G., Thapa, N., Park, H. J., & Park, H. (2023). Nonlinear associations between physical function, physical activity, sleep, and depressive symptoms in older adults. Journal of Clinical Medicine, 12(18), 6009. doi: 10.3390/jcm12186009 Bo, A., Mao, W., & Lindsey, M. A. (2017). Effects of mind–body interventions on depressive symptoms among older Chinese adults: a systematic review and meta-analysis. International Journal of Geriatric Psychiatry, 32(5), 509–521. https://doi.org/10.1002/gps.4688 Borenstein, M. (2019). Common Mistakes in Meta-Analysis and How to Avoid Them. Biostat, Inc. 18 Borenstein, M. (2020). Research Note: In a meta-analysis, the I2 index does not tell us how much the effect size varies across studies. Journal of Physiotherapy, 66(2), 135-139. https://doi.org/10.1016/j.jphys.2020.02.011 Borenstein, M., Hedges, L. E., Higgins, J. P. T., & Rothstein, H. R. (2022). Comprehensive Meta-Analysis Version 4. In Biostat, Inc. www.Meta-Analysis.com Borenstein, M., Hedges, L. V., Higgins, J. P., & Rothstein, H. R. (2010). A basic introduction to fixed-effect and random-effects models for meta-analysis. Research Synthesis Methods, 1(2), 97-111. https://doi.org/10.1002/jrsm.12 Borenstein, M., Hedges, L. V., Higgins, J. P. T., & Rothstein, H. R. (2021). Introduction to Meta-Analysis (Second ed.). Wiley. Borenstein, M., Higgins, J. P., Hedges, L. V., & Rothstein, H. R. (2017). Basics of meta-analysis: I2 is not an absolute measure of heterogeneity. Research Synthesis Methods, 8(1), 5-18. https://doi.org/10.1002/jrsm.1230 Brooks, J. M., Titus, A. J., Bruce, M. L., Orzechowski, N. M., Mackenzie, T. A., Bartels, S. J., & Batsis, J. A. (2018). Depression and handgrip strength among US adults aged 60 years and older from NHANES 2011-2014. The Journal of Nutrition, Health and Aging, 22(8), 938943. doi: 10.1007/s12603-018-1041-5 Burke, S. M., Carron, A. V., Eys, M. A., Ntoumanis, N., & Estabrooks, P. A. (2006). Group versus individual approach? A meta-analysis of the effectiveness of interventions to promote physical activity. Journal of Sport & Exercise Psychology, 2, 19-35. https://core.ac.uk/reader/185423720 Chlapecka, A., Kagstrom, A., & Cermakova, P. (2020). Educational attainment inequalities in depressive symptoms in more than 100,000 individuals in Europe. European psychiatry, 63(1), e97. doi: 10.1192/j.eurpsy.2020.100 Colcombe, S., & Kramer, A. F. (2003). Fitness effects on the cognitive function of older adults: A meta-analytic study. Psychological Science, 14(2), 125-130. doi: 10.1111/14679280.t01-1-01430 Craft, L. L., & Landers, D. M. (1998). The Effect of Exercise on Clinical Depression and Depression Resulting from Mental Illness: A Meta-Analysis. Journal of Sport and Exercise Psychology, 20(4), 339–357. doi: 10.1123/JSEP.20.4.339 DerSimonian, R., & Laird, N. (1986). Meta-analysis in clinical trials. Controlled Clinical Trials, 7(3), 177-188. http://www.ncbi.nlm.nih.gov/pubmed/3802833 DerSimonian, R., & Laird, N. (2015). Meta-analysis in clinical trials revisited. Contemporary Clinical Trials, 45(Pt A), 139-145. https://doi.org/10.1016/j.cct.2015.09.002 Diniz, B. S., Sibille, E., Ding, Y., Tseng, G., Aizenstein, H. J., Lotrich, F., Becker, J. T., Lopez, O. L., Lotze, M. T., Klunk, W. E., Reynolds, C. F., & Butters, M. A. (2015) Plasma biosignature and brain pathology related to persistent cognitive impairment in late-life depression. Molecular Psychiatry, 20, 594–601. doi: 10.1038/mp.2014.76 Domènech-Abella, J., Mundó, J., Leonardi, M., Chatterji, S., Tobiasz-Adamczyk, B., Koskinen, S., Ayuso-Mateos, J. L., & Haro, J. M. (2018). The association between socioeconomic status and depression among older adults in Finland, Poland and Spain: A comparative 19 cross-sectional study of distinct measures and pathways. Journal of Affective Disorders, 241, 311-318. doi: 10.1016/j.jad.2018.08.077 Falck, R. S., McDonald, S. M., Beets, M. W., Brazendale, K., & Liu-Ambrose, T. (2016). Measurement of physical activity in older adult interventions: a systematic review. British Journal of Sports Medicine, 50(8), 464–470. Doi: 10.1136/BJSPORTS-2014-094413 Farmer, M. E., Locke, B. Z., Mościcki, E. K., Dannenberg, A. L., Larson, D. B., & Radloff, L. S. (1988). Physical activity and depressive symptoms: the NHANES I Epidemiologic Follow-up Study. American Journal of Epidemiology, 128(6), 1340-1351. doi: 10.1093/oxfordjournals.aje.a115087 Farrance, C., Tsofliou, F., & Clark, C. (2016). Adherence to community based group exercise interventions for older people: A mixed-methods systematic review. Preventive Medicine, 87, 155-166. https://www.sciencedirect.com/science/article/pii/S0091743516300147 Fergusson, D., Aaron, S. D., Guyatt, G., & Hébert, P. (2002). Post-randomisation exclusions: the intention to treat principle and excluding patients from analysis. BMJ, 325, 652-654. doi: 10.1136/bmj.325.7365.652 Girgus, J. S., Yang, K., & Ferri, C. V. (2017). The gender difference in depression: Are elderly women at greater risk for depression than elderly men? Geriatrics (Basel), 2(4), 35. doi: 10.3390/geriatrics2040035 Glaesmer, H., Riedel-Heller, S., Braehler, E., Spangenberg, L., & Luppa, M. (2011). Ageand gender-specific prevalence and risk factors for depressive symptoms in the elderly: a population-based study. International Psychogeriatrics, 23(8), 1294–1300. https://doi.org/10.1017/S1041610211000780 Hedges, L. V., & Olkin, I. (1985). Statistical methods for meta-analysis. Academic Press. Publisher description http://www.loc.gov/catdir/description/els032/84012469.html Hedges, L. V., & Vevea, J. L. (1998). Fixed and random-effects models in meta-analysis. Psychological Methods, 3(4), 486-504. https://doi.org/10.1037/1082-989X.3.4.486 Higgins, J. P. (2008). Commentary: Heterogeneity in meta-analysis should be expected and appropriately quantified. International Journal of Epidemiology, 37(5), 1158-1160. https://doi.org/10.1093/ije/dyn204 Higgins, J. P., & Thompson, S. G. (2002). Quantifying heterogeneity in a meta-analysis. Statistics in Medicine, 21(11), 1539-1558. https://doi.org/10.1002/sim.1186 Higgins, J. P., Thompson, S. G., Deeks, J. J., & Altman, D. G. (2003). Measuring inconsistency in meta-analyses. BMJ, 327(7414), 557-560. https://doi.org/10.1136/bmj.327.7414.557 Higgins, J. P. T., & Thomas, J. (2019). Cochrane Handbook for Systematic Reviews of Interventions (J. P. T. Higgins, J. Thomas, J. Chandler, M. Cumpston, T. Li, M. J. Page, & V. A. Welch, Eds. 2nd Edition. ed.). Wiley. Hinata, A., Kabasawa, K., Watanabe, Y., Kitamura, K., Ito, Y., Takachi, R., Tsugane, S., Tanaka, J., Sasaki, A., & Narita, I. (2021). Education, household income, and depressive symptoms in middle-aged and older Japanese adults. BMC public health, 21, 1-10. doi: 10.1186/s12889-021-12168-8 20 Horne, S. J., Topp, T. E., & Quigley, L. (2021). Depression and the willingness to expend cognitive and physical effort for rewards: A systematic review. Clinical Psychology Review, 88, 102065. https://doi.org/10.1016/j.cpr.2021.102065 IntHout, J., Ioannidis, J. P. A., Rovers, M. M., & Goeman, J. J. (2016). Plea for routinely presenting prediction intervals in meta-analysis. BMJ Open, 6(7), e010247. https://doi.org/10.1136/bmjopen-2015-010247 Kandola, A., Ashdown-Franks, G., Hendrikse, J., Sabiston, C. M., & Stubbs, B. (2019). Physical activity and depression: Towards understanding the antidepressant mechanisms of physical activity. Neuroscience and Biobehavioral Reviews, 107, 525–539. Doi: 10.1016/J.NEUBIOREV.2019.09.040 Klil-Drori, S., Klil-Drori, A. J., Pira, S., & Rej, S. (2020). Exercise Intervention for Late-Life Depression. The Journal of Clinical Psychiatry, 81(1). https://doi.org/10.4088/JCP.19r12877 Kong, J. Y., Hong, H., & Kang, H. (2022). Relationship between physical activity and depressive symptoms in older Korean adults: moderation analysis of muscular strength. BMC Geriatrics, 22(1), 884. doi: 10.1186/s12877-022-03610-6 Lacroix, A., Hortobagyi, T., Beurskens, R., & Granacher, U. (2017). Effects of supervised vs. unsupervised training programs on balance and muscle strength in older adults: a systematic review and meta-analysis. Sports medicine, 47, 2341-2361. doi: 10.1007/s40279-017-0747-6 Lee, Y. H., Kim, H., & Cho, H. (2022). The Effectiveness of Physical Activity Interventions on Depression in Korea: A Systematic Review and Meta-Analysis. Healthcare, 10(10). Doi: 10.3390/HEALTHCARE10101886 Leng et al. (2018). Effects of physical exercise on depressive symptoms in patients with cognitive impairment: A systematic review and meta-analysis. The Journal of Nervous and Mental Disease, 206(10), 809-823. doi: 10.1097/NMD.0000000000000887 Liu, Q., Ni, W., Zhang, L., Zhao, M., Bai, X., Zhang, S., Ding, Y., Yin, H., & Chen, L. (2023). Comparative efficacy of various exercise interventions on depression in older adults with mild cognitive impairment: A systematic review and network meta-analysis. Ageing Research Reviews, 91, 102071. doi: 10.1016/j.arr.2023.102071 Mack, M.; Badache, A.; Erden A.; Giannaki C.; Haider, S.; Kaltsatou, A.; Kömürcü Akik, B.; Netz, Y.; Pavlova, J.; Stavrinou, P.; Voelcker-Rehage, C. & Audiffren, M. (preprint). A metaanalysis of chronic exercise effects and moderating variables on depression severity and core symptoms in older adults. Preprint. https://sciety.org/articles/activity/10.21203/rs.3.rs-6795147/v1. Mann, T., Lamberts, R. P., & Lambert, M. I. (2013). Methods of prescribing relative exercise intensity: physiological and practical considerations. Sports Med, 43(7), 613-625. doi:10.1007/s40279-013-0045-x McPhee, J. S., French, D. P., Jackson, D., Nazroo, J., Pendleton, N., & Degens, H. (2016). Physical activity in older age: perspectives for healthy ageing and frailty. Biogerontology, 17, 567-580. doi: 10.1007/s10522-016-9641-0 21 Miller, K. J., Goncalves-Bradley, D. C., Areerob, P., Hennessy, D., Mesagno, C., & Grace, F. (2020). Comparative effectiveness of three exercise types to treat clinical depression in older adults: A systematic review and network meta-analysis of randomised controlled trials. Ageing Research Reviews, 58, 100999. doi: 10.1016/j.arr.2019.100999 Mitsui, T., Arii, Y., Tsukamoto, A., Taniguchi, K., Mabuchi, M., Shimizu, A., Sumitomo, N., & Maki, Y. K. (2021). Sociability-based fitness approach in Parkinson's disease: Comparison with conventional rehabilitation. European Journal of Neurology, 28(6), 1893-1900. https://pubmed.ncbi.nlm.nih.gov/33657674/ Mortazavi, S. S., Shati, M., Ardebili, H. E., Mohammad, K., Beni, R. D., & Keshteli, A. H. (2013). Comparing the effects of group and home-based physical activity on mental health in the elderly. International Journal of Preventive Medicine, 4(11), 1282-1289. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3883253/ Nebiker, L., Lichtenstein, E., Minghetti, A., Zahner, L., Gerber, M., Faude, O., & Donath, L. (2018). Moderating Effects of Exercise Duration and Intensity in Neuromuscular vs. Endurance Exercise Interventions for the Treatment of Depression: A Meta-Analytical Review. Front Psychiatry, 9, 305. doi:10.3389/fpsyt.2018.00305 Noetel, M., Sanders, T., Gallardo-Gomez, D., Taylor, P., Del Pozo Cruz, B., van den Hoek, D., . . . Lonsdale, C. (2024). Effect of exercise for depression: systematic review and network meta-analysis of randomised controlled trials. BMJ, 384, e075847. doi:10.1136/bmj-2023-075847 North, T. C., McCullagh, P., & Tran, Z. V. (1990). Effect of exercise on depression. Exerc Sport Sci Rev, 18, 379–415. Northey, J. M., Cherbuin, N., Pumpa, K.L., Smee, D. J., & Rattray, B. (2018). Exercise interventions for cognitive function in adults older than 50: A systematic review with meta-analysis. Br. J. Sports Med. 52, 154-160. doi: 10.1136/bjsports-2016-096587 Olanrewaju, O., Kelly, S., Cowan, A., Brayne, C., & Lafortune, L. (2016). Physical Activity in Community Dwelling Older People: A Systematic Review of Reviews of Interventions and Context. PLoS ONE, 11(12). doi: 10.1371/JOURNAL.PONE.0168614 Pagoto, S. L., McDermott, M. M., Reed, G., Greenland, P., Mazor, K. M., Ockene, J. K., ... & Ockene, I. (2013). Can attention control conditions have detrimental effects on behavioral medicine randomized trials? Psychosomatic Medicine, 75(2), 137-143. doi: 10.1097/PSY.0b013e3182765dd2 Poole, L., & Jackowska, M. (2018). The epidemiology of depressive symptoms and poor sleep: findings from the English Longitudinal Study of Ageing (ELSA). International Journal of Behavioral Medicine, 25(2), 151-161. doi: 10.1007/s12529-017-9703-y Ren, F. F., Chen, F. T., Zhou W. S., Cho, Y. M., Ho, T. J., Hung, T. M., & Chang, Y. K. (2021). Effects of Chinese mind-body exercises on executive function in middle-aged and older adults: A systematic review and meta-analysis. Frontiers in Psychology, 12, 656141. doi: 10.3389/fpsyg.2021.656141 Rethorst, C. D., Wipfli, B. M., & Landers, D. M. (2009). The antidepressive effects of exercise: a meta-analysis of randomized trials. Sports Medicine (Auckland, N.Z.), 39(6), 491–511. Doi: 10.2165/00007256-200939060-00004 22 Rhyner, K. T., & Watts, A. (2016). Exercise and depressive symptoms in older adults: A systematic meta-analytic review. Journal of Aging and Physical Activity, 24(2), 234–246. https://doi.org/10.1123/japa.2015-0146 Rice, K., Higgins, J. P. T., & Lumley, T. (2018). A re-evaluation of fixed effect(s) meta-analysis. Journal of the Royal Statistical Society: Series A (Statistics in Society), 181(1), 205–227. https://doi.org/10.1111/rssa.12275 Rubin, D. B. (1998). More powerful randomization-based p-values in double-blind trials with non-compliance. Statistics in Medicine, 17, 371-385. doi: 10.1002/(sici)10970258(19980215)17:3<371::aid-sim768>3.0.co;2-o Sanders, L. M. J., Hortobágyi, T., La Bastide-van Gemert, S., van der Zee, E. A., & van Heuvelen, M. J. G. (2019). Dose-response relationship between exercise and cognitive function in older adults with and without cognitive impairment: A systematic review and meta-analysis. PLoS ONE, 14(1), e0210036. https://doi.org/10.1371/journal.pone.0210036 Schlax, J., Jünger, C., Beutel, M. E., Münzel, T., Pfeiffer, N., Wild, P., Blettner, M., Kerahrodi, J. G., Wiltink, J., & Michal, M. (2019). Income and education predict elevated depressive symptoms in the general population: results from the Gutenberg health study. BMC public health, 19, 1-10. doi: 10.1186/s12889-019-6730-4 Schuch, F. B., Vancampfort, D., Rosenbaum, S., Richards, J., Ward, P. B., Veronese, N., Solmi, M., Cadore, E. L., & Stubbs, B. (2016). Exercise for depression in older adults: a metaanalysis of randomized controlled trials adjusting for publication bias. Brazilian Journal of Psychiatry, 38(3), 247-254. doi: 10.1590/1516-4446-2016-1915 Singh, B., Olds, T., Curtis, R., Dumuid, D., Virgara, R., Watson, A., Szeto, K., O’Connor, E., Ferguson, T., Eglitis, E., Miatke, A., Simpson, C. E. M., & Maher, C. (2023). Effectiveness of physical activity interventions for improving depression, anxiety and distress: An overview of systematic reviews. British Journal of Sports Medicine, 57(18) 1203–1209. Doi: 10.1136/bjsports-2022-106195 Sommer, A., & Zeger, S. L. (1991). On estimating efficacy from clinical trials. Statistics in Medicine, 10, 45-52. doi: 10.1002/sim.4780100110 Vinckier, F., Jaffre, C., Gauthier, C., Smajda, S., Abdel-Ahad, P., Le Bouc, R., Daunizeau, J., Fefeu, M., Borderies, N., Plaze, M., Gaillard, R., & Pessiglione, M. (2022). Elevated effort cost identified by computational modeling as a distinctive feature explaining multiple behaviors in patients with depression. Biological Psychiatry: Cognitive Neuroscience and Neuroimaging, 7, 1158–1169. https://doi.org/10.1016/j.bpsc.2022.07.011 Wang, R., Bishwajit, G., Zhou, Y., Wu, X., Feng, D., Tang, S., Chen, Z., Shaw, I., Wu, T., Song, H., Fu, Q., & Feng, Z. (2019). Intensity, frequency, duration, and volume of physical activity and its association with risk of depression in middleand older-aged Chinese: Evidence from the China Health and Retirement Longitudinal Study, 2015. PLoS ONE, 14(8). doi: 10.1371/JOURNAL.PONE.0221430 Xie, Y., Wu, Z., Sun, L., Zhou, L., Wang, G., Xiao, L., & Wang, H. (2021). The Effects and Mechanisms of Exercise on the Treatment of Depression. Frontiers in Psychiatry, 12, 705559. doi:10.3389/fpsyt.2021.705559 23 Yen, H. Y., & Chiu, H. L. (2021). Virtual reality exergames for improving older adults’ cognition and depression: a systematic review and meta-analysis of randomized control trials. Journal of the American Medical Directors Association, 22(5), 995-1002. doi: 10.1016/j.jamda.2021.03.009 Zhao, X., Huang, X., Cai, Y., Cao, T., & Wan, Q. (2022). The relative effectiveness of different combination modes for exercise and cognitive training on cognitive function in people with mild cognitive impairment or Alzheimer’s disease: a network meta-analysis. Aging & Mental Health, 26(12), 2328–2338. https://doi.org/10.1080/13607863.2022.2026879 24 Appendix 1 # Authors Depression scale 1 Ahmed (2019) GDS-30 2 Aibar-Almazan et al. (2019) HADS 3 Albinet et al. (2016) GDS-30 4 An et al. (2020) GDS-30 5 Ansai & Rebelatto (2015) GDS-15 6 Antunes et al. (2005) GDS-30 7 Belza et al. (2002) CESD-11 8 Bernard et al. (2015) BDI 9 Bernardelli et al. (2019) GDS-30 10 Bonura (2008), Bonura et al. (2014) GDS-30 11 Bouaziz et al. (2019) GADS-18 12 Brăilescu et al. (2017) HAMD-17 13 Brenes et al. (2007) HAMD-17 & GDS-15 14 Buman et al. (2011) CESD-10 15 Casas-Herrero et al. (2022) GDS-15 16 Chang et al. (2020) BDI & HAMD-17 17 Chang et al. (2021) GDS-15 18 Chen et al. (2008) CESD-20 19 Chen et al. (2021) GDS-15 20 Cheng et al. (2012) GDS-15 21 Chin et al. (2022) HADS & PHQ-9 22 Choi & Sohng (2018) GDS-15 23 Chou et al. (2004) CESD-20 24 Clegg et al. (2014) GDS-15 25 Cordes et al. (2021) CESD 26 Cugusi et al. (2015) BDI-II 27 Deus et al. (2021) BDI 28 Dong et al. (2013) GDS-15 29 Eggermont et al. (2009) GDS-30 30 Fakhari (2017) BDI-II 31 Farzane & Koushkie Jahromi (2022) HADS 32 Ferreira et al. (2015) GDS 33 Flegal et al. (2007) / Oken et al. (2006) CESD-10 34 Frih et al. (2017) HADS 35 Furtado et al. (2021) CESD-20 36 Gary et al. (2004), Gary (2006) GDS-15 37 Gary et al. (2010) HAMD-17 38 Ge et al. (2022) GDS-30 39 Gleeson et al. (2017) GDS-5 40 Haboush et al. (2006) HAMD-17 & GDS-30 41 Hashimoto et al. (2015) SDS 42 Hembree (2000) BDI 43 Henskens et al. (2018) CSDD 25 44 Hsu et al. (2016) GDS-15 45 Irwin et al. (2014) IDS-30 46 Jeong et al. (2019) GDS-15 47 Jeong et al. (2021) GDS-30 48 Jin et al. (2019) GDS-15 49 Jolly et al. (2009) HADS 50 Kerse et al. (2010) GDS-15 51 Kim et al. (2019) GDS-15 52 Kraiwong et al. (2021) PHQ-9 53 Krawcyk et al. (2019) MDI 54 Krishnamurthy & Telles (2007) GDS-15 55 Lai et al. (2006) GDS-15 56 Langoni et al. (2019) GDS-15 57 Laredo-Aguilera et al. (2018) GDS-15 58 Lee et al. (2013) GDS-15 59 Lee et al. (2018) BDI 60 Liao et al. (2018) GDS-30 61 Lima et al. (2019) HAMD-17 62 Lin et al. (2007) GDS-15 63 Lin et al. (2020) HADS 64 Lincoln et al. (2011) GDS-30 65 Lippke et al. (2022) CESD-20 66 Liu et al. (2018) GDS-30 67 Ma et al. (2018) CESD-10 68 Maci et al. (2012) CSDD 69 Makizako et al. (2020) GDS-15 70 Martínez et al. (2015) GDS-5 71 Martínez-Velilla et al. (2021) GDS-15 72 McNeil et al. (1991) BDI 73 Morris et al. (2017) CSDD 74 Motamedi et al. 2021) GDS-15 75 Netz et al. (1994) GDS-30 76 Niemi et al. (2016) PHQ-9 77 Nolte et al. (2015) PHQ-9 78 Pakkala et al. (2008) CESD-20 79 Parvin et al. (2020) GDS-15 80 Phansuea et al. (2020) GDS-30 81 Picelli et al. (2016) BDI 82 Piraux et al. (2021) CESD-20 83 Prakhinkit et al. (2014) GDS-30 84 Ravari et al (2021) BDI-II 85 Redwine et al. (2020) BDI-IA 86 Requena Hernández et al. (2008) GDS-30 87 Rica et al. (2020) BDI 88 Rolland et al. (2007) MADRS 89 Roswiyani et al. (2020) BDI-II