Influence of strength training and multicomponent training on the functionality of older adults: systematic review and meta-analysis

– Physical exercise is an important option to maintain functional independence in older adults, however, it is not clear which type of exercise is the most benefic: strength or multicomponent. The objective was to verify the effectiveness of strength training and multi- component training on functionality of healthy older patients through a systematic review with meta-analysis. Registration number: CRD42017071887. Two independent evaluators searched Pubmed, Web of Science, PEDro, Cochrane and Lilacs databases. Of the 1434 studies found, 32 clinical trials that investigated the effects of strength training only and/or combined with other modalities (multicomponent) in older adults and evaluated the Timed up and Go (TUG), sit-to-stand (STS) and/or Berg Balance Scale (BBS) were included. The methodological quality was evaluated with the Downs & Black scale. Data analysis was performed with the Software Review Manager. It was verified improvement in all the investigated outcomes when perform- ing multicomponent training in comparison to control groups. Strength training, compared to control groups, showed benefit only for sit-to-stand test. Studies comparing the two trainings found no difference between them. The not high average score in the methodological quality assessment of the included studies is a limitation of the present study. In conclusion, both types of training were effective to improve functionality and are good strategies of training for older individuals. However, as the comparison between the two types of training was performed in few studies, it is not possible to infer which is more effective for the functionality, suggesting the realization of new clinical trials.


INTRODUCTION
Currently, the elder population is increasing at higher rates 1 . Throughout the aging process many changes occur in the neuromuscular, somatosensory, visual, vestibular systems, besides the body and bone compositions 2 . Additionally, the brain volume reduces at a rate of 5% per decade after the age of 40 3 . As a consequence, physical abilities decline, such as strength and balance, compromising basic and instrumental activities of daily life; besides, the mental health of older adults 2 .
The ability to maintain postural control is important for orthostatic position, ambulation, and for daily life activities. Postural imbalance in elderly subjects may indicate reduced ability to recover from small postural disorders, which may be a fall predictive factor 4 . It is important to detect the risks of falling in order to prevent them 5,6 .
In order to assess the risks of falling, validated tests and scales are used, such as the Timed Up and Go (TUG), which is a quick and practical test to evaluate the physical mobility, gait speed and balance of individuals, and has already been described in the literature as a predictor of falls and an instrument for assessment of the dynamic balance 7 . Another widely used test is the Sit-to-Stand Test (STS), as a parameter to determine muscle strength, speed of performance of lower members strength and the functional ability to sit and stand up from the chair 8 . This test coul be performed in two ways: the number of repetitions performing during 30 seconds (30STS), or the time to performing 5 repetitions (5STS). Additionally, the Berg Balance Scale (BBS) is also used to evaluate the dynamic balance 9 .
Physical exercise has been proved to be beneficial in the physical performance of older subjects, and is a strategy for the prevention of falls 10,11 . Among several exercises modalities, strength training is a widely used method for the prevention of diseases associated with the development of co-morbidity and mortality in aging population, as well as improvement of physical performance through muscular strength. These exercises are based on the principle of overload, assuming that loads must be applied gradually, leading to an increased effort during the training sessions, aiming to provoke a disorder of cellular homeostasis and, in response, an adaptation to this stress. These adaptations occur at the muscular, neural and cellular levels 12 .
Strength exercises increase muscle mass in the elderly subjects of both genders, minimizing and even reversing sarcopenia. This process is characterized by low muscle strength, low muscle quantity or quality and low physical performance 13 . Strength exercises can also reduce the frequency of falls, thus contributing to the maintenance of independence and a better quality of life for the older individuals 14 .
More recently, studies have shown that multicomponent exercises, which include an approach of different physical abilities may also be beneficial in maintaining and improving the functionality of the older population [15][16][17] . Besides minimizing sarcopenia, they contribute to the maintenance of balance, cardio-respiratory performance, flexibility, and are therefore recommended by the ACSM to increase and maintain the functionality of the older subjects 18 .
However, it is unclear which of the two types of training may be best for the functionality of the elderly population. In order to demonstrate the evidence found in the literature on the subject, the objective of the present study was to conduct a systematic review with meta-analysis of clinical trials that investigated the effectiveness of strength training or multicomponent training on the functionality in older subjects, performing comparisons with a control group and between the two types of training.

Protocol and registration
This study was registered in the PROSPERO under the number CRD42017071887. The present review follows the PRISMA statement.

Design
This study consists of a systematic review with meta-analysis of randomized controlled trials. This review includes studies that analyzed the effect of a training involving strength and/or multicomponent exercises (strength training combined with other exercises) on functional capacity of older subjects, evaluated by the Timed up and Go (TUG), sit-to-stand (5STS and/or 30 STS), and Berg Balance Scale tests (BBS). The following comparisons were performed: strength training versus control group, multicomponent training versus control group (no exercise), and multicomponent training versus strengh training.

Search Strategy
Systematic searches were performed in Pubmed, Web of Science, PEDro, Cochrane and Lilacs databases. The results of the search performed in each database were exported for further analysis.
The Mesh and keywords used in the search were related to the subjects (older individuals), type of training (strength training) and evaluated outcomes (TUG, STS and BBS). The terms used were combined through the Boolean operators "AND" and "OR", and a complete search example, carried out in the Pubmed database, are presented in supplementary document 1.

Eligibility Criteria
To be included in the present review, the articles should have been randomized controlled trials and have been performed only with healthy older adults 60 years or over residing in the community or in asylum institutions. Studies carried out with hospitalized older people or with some pathology were not included. The training performed should be only of strength or strength combined with other exercise modalities (multicomponent), with a total duration of at least 8 weeks and a weekly frequency of at least two sessions supervised by professionals. Studies that performed concurrent training supplementation and studies that did not have a control group which did not perform any type of exercise were not included. Regarding the outcomes, articles that evaluated one or more of the following tests were included: TUG, BBS and 30-second sit-to-stand test (30STS) or five repetitions (5STS). Those which did not present sufficient data regarding the results to perform the meta-analysis or that made adaptations in the tests (eg, inclusion of a double task) were excluded. Articles written in English, Portuguese or Spanish, which were published since 2006 were considered.

Studies selection
After the search in each database and export of the results, the selection of potential studies for inclusion in the present systematic review was carried out by two independent reviewers (ECWM and ECG) by reading their titles and abstracts, excluding only those that clearly did not fill the eligibility criteria. Articles that were selected by at least one of the reviewers, remained for the next stage of analysis.
Then, duplicated articles (found in more than one database) were excluded and the full texts were searched for articles that remained in the list after the initial selection. Two independent reviewers applied the eligibility criteria therein, and discrepancies were resolved by consensus.

Quality assessment and risk of bias
The methodological quality and risk of bias within and across the studies included in the present review was assessed using a PEDro scale for the evaluation of randomized controlled trials 19 .

Data extraction and analysis
In order to characterize the studies included in the systematic review, a standard form was used to organize data related to authorship, year of publication, information about study participants, characteristics of the training performed and methodology for evaluating the outcomes.
To analyze the effects of the trainings performed, through metaanalysis, the results found in each study (number of participants before and after training, mean and standard deviation of each outcome before and after training and results found) were included in a standard form.
For studies where there were more training groups, consisting of other types of training, besides strength training, only the data of the groups that performed strength exercises were included in the review. When there were two types of strength training in the same study, only data from the group that performed the type of training that was closest to a conventional strength training were used. In addition, when more than one type of concurrent training was performed, data from the group that performed strength training before the other exercises were used.
In the quantitative analysis, since the outcomes analyzed were continu-ous, the values of the weighted mean difference were used for the analysis with a 95% confidence interval. Delta values were used in the meta-analysis (difference between the pre and post-mean of each outcome). The results of the meta-analysis are presented in the unit of the investigated outcomes. A random effects model was used, and the heterogeneity was assessed through the Cochran Q test (p≤0.05, indicating the existence of heterogeneity) and the inconsistency test (I², low heterogeneity: I² ≤ 25%; moderate: 26% < I² ≤ 50%; and high: I² > 50%). When moderate or high heterogeneity between the studies were found, sensitivity analyzes were performed, considering gender, total training program duration, mean age, number of subjects and characteristics of the exercises. A meta-analysis was performed separately for each outcome and for each comparison (control versus strength training, control versus multicomponent training, and multicomponent training versus strength training), and a level of significance was considered as p<0.05. The analyses were performed using the Review Manager software, version 5.3 (Cochrane Collaboration).

Articles search and selection
After searching the defined databases using the search strategy, 1434 articles were found. Of these, 453 articles were selected by reading titles and abstracts and the full text was analyzed. After exclusion of duplicated articles, the eligibility criteria were applied, and 28 articles were included in the present systematic review with meta-analysis ( Figure 1). had not a longitudinal design or were not a paper; 12 included individuals younger than 60 years or not community or geriatric dwellers; 37 due to training characteristics; 84 did not measure the outcomes of interest or had insufficient data; 91 did not include a control group that not performed exercise; 3 were not related to the aim of the review; 1 included individuals with some disease; 9 were published in other languages; 67 published before 2006; 1 with repeated data; 4 not randomized

Quality Assessment
The results of the methodological quality analysis of the studies included in the present review are presented in Supplementary document 2, individually for each study. The scale PEDro used allows a score of 0 to 11 points 19 . The overall mean of included studies was 6.93 points.

Characteristics of included studies
The main characteristics of the studies included in the present review regarding the participants and the details of the training performed are presented in Supplementary document 2.

Effects of interventions
The effect of the interventions performed was assessed by the meta-analysis.
Comparisons were made between control group and strength training, between control group and multicomponent training (which included strength training and some training associated with it) and between multicomponent training and strength training. They are presented in Figures 2 to 5.

Timed up & go (TUG)
In the comparison of the effect of multicomponent training group to control group, it was observed that the multicomponent training significantly improved (p<0.001) the TUG test performance, reducing the time needed to perform the test in -1.48 seconds (95% CI -2.09; -0.88, I² = 0%) compared to the control group ( Figure 2a). When comparing the effect of the strength training groups to control groups, strength training did not improve significantly (p=0.14) the TUG performance [95% confidence interval (CI) -1.10, 0.15, I² = 63%] ( Figure  2b). Due to the fact that high heterogeneity occurred for this analysis, sensitivity analyses were performed. Through these analyses, it was not possible to identify which factor was related to the high heterogeneity, once that doing the analysis separately for studies that included only women or men or both, studies with different duration or different exercises did not reduce the heterogeneity.
None of the included studies compared the effect of a multicomponent workout with a strength training for this outcome.

Sit-to-stand (30 seconds) -30STS
Comparing the effect of a multicomponent training with control group, it was verified that the multicomponent training significantly improved (p<0.001) the performance of the 30STS test, increasing the number of repetitions performed in 1.79 repetitions (IC 95% 0.76, 2.83, I² = 52%) (Figure 3a).
In a sensitivity analysis, it was observed that when studies that included a balance component in the training (in addition to aerobic and strength exercises which were performed in all studies), were withdrawn from the analysis, the heterogeneity changed from moderate (52%) to low (0%), while still maintaining the significant difference between the control group and the multicomponent group. However, the improvement without these studies was lower (0.89 repetitions), showing that studies which also included balance exercises in the training program presented better results.  Strength training compared to the control group significantly improved the performance of the 30STS test, increasing the number of repetitions performed in 2.51 repetitions (95% CI 0.98; 4.03; I² = 0%) (Figure 3b).
For the 30STS test only one study examined the effects of a strength training versus multicomponent workout. 8 Therefore, no meta-analysis was performed. It can be observed from the change between pre and post training that the group that trained only strength increased 3.00 repetitions in the test and that who trained strength and other exercises increased 2.90 repetitions, on average, after the training, both groups had a very similar improvement.
Comparing strength training group to the control group, it was found that strength training did not improve significantly (p=0.16) the performance in the 5STS test (95% CI -1.90, 0.30, I² = 15%) (Figure 4b).

Berg Balance Scale
It was verified that multicomponent training group significantly increased the performance of the Berg Balance Scale by 3.17 points (95% CI 0.77, 5.57, I² = 0%) compared to the control group ( Figure 5).
None of the included studies compared the effect of a multicomponent workout with a strength training for this outcome, nor of a strength training group comparison to a control group.

DISCUSSION
The goal of this systematic review with meta-analysis was to investigate the effects of strength training and multicomponent training on the functionality and risk of falls in older subjects, evaluated by the TUG, STS and BBS tests. The results of the meta-analysis showed that the multicomponents training brought significant benefits to all the outcomes investigated, and that strength training only benefited performance in the 30-second sit-tostand test, when compared to control groups. Few studies compared the two types of training, and in those which made this comparison, no differences were found between them when considering the subjects' functionality.
In a descriptive analysis of the studies included in the present review, it was verified that most of the studies that evaluated the effect of multicomponent training or strength training on the functionality used the TUG test, since it was found in 13 of the 28 articles 9, 21, 22, 25, 26, 33-36, 38, 40, 41, 43 . The STS test was found in 19 studies, and 14 of them used the evaluation verifying the maximum number of repetitions that older subjects performed in 30 seconds of test (30STS) 1, 8, 9, 20, 23, 27, 28, 30-32, 36, 37, 39, 41 and five studies used the assessment by checking the time the elders needed to perform the five times sit-to-stand task (5STS) 22,24,29,35,44 . The Berg Balance Scale (BBS) was used in only three of the studies included in the review 9, 42,43 . In addition, it was verified that most of studies used only one of the cited tests, and only 8 of them performed more than one of these functional tests 9,22,35,36,41,43 .
Regarding the results of the studies included in the meta-analysis, it was verified that the intervention performed with multicomponent training, the session being composed of strength, balance, aerobic and stretching exercises performed at least twice a week for a minimum duration of 8 weeks showed benefits to the healthy older adults in the four outcomes analyzed: TUG, 30STS, 5STS and BBS, when compared to the control group. This demonstrates that multicomponent training is a great strategy to improve functional capacity and decrease the risk of falls in older subjects.
These results are aligned to previous studies, where multicomponent training had positive effects on the functionality of older adults [15][16][17] . Similarly, the results of the meta-analysis with all the studies corroborate with the individual results of the studies, which have found significant improvements on the functionality 35,41,43,44 .
On the other hand, strength training when performed as a single training modality demonstrated benefits in relation to the control group only in the 30STS test. For the other outcomes (TUG and 5STS) this type of training did not provide benefits. Additionally, regarding BBS, none of the studies investigated the effect of strength training on this parameter.
Other parameters that are related to the process of sarcopenia and the risk of falls, in addition to the functional performance evaluated in the present review, are the quantity of muscle mass and the strength, that can improve after both strength 26 and multicomponent 15 training, and lead to positive effects on functional performance 13 .
Few studies have compared strength training with multicomponent training. For the TUG outcome, the study by Fahlman et al. 8 found improvements in both types of training on this variable. For 30STS and BBS none of the included studies compared the effect of the two types of training. It was possible to perform meta-analysis only for the 5STS outcome, where there was no statistically significant difference in relation to the two types of training 24,44 . As only two studies could be included in this comparison, the results should be looked carefully.
In one of the studies of this comparison 44 , the authors applied strength training for one group and multicomponent training (strength, aerobic and balance exercises) to another for 16 weeks and verified that the group which participated of multicomponent training has significant improvement in 5STS, which did not occur for the strength training group. Based on this, the authors concluded that for older population, multicomponent training appears to be more beneficial and has fewer adverse events, as suggested by Gillespie et al. 45 and recommendations from the ACSM 18 .
In the study of Locks et al. 24 , where there was a strength training group and another one that performed strength exercises combined with stretching exercises, developed during 12 weeks, none of the groups had significant improvement after training in 5STS.
In addition, the evaluation of the methodological quality of the included studies showed that, on average, they scored 6.93 points. The scale used 19 has a maximum score of 11 points. Thus, it is observed that the studies did not present high scores of methodological quality, which may influence the results found, being a limitation for the present study.

CONCLUSIONS
Both strength and multicomponents exercises are good strategies to improve the functionality and decrease the risk of falls in older subjects, since both presented significant benefits when compared to the control group. However, multicomponent training demonstrated improvement in all outcomes analyzed in the present meta-analysis, while strength training showed benefit only for the 30STS outcome. As the comparison between the two types of training was performed in few studies, it is not possible to infer which is more effective for the functionality, suggesting the realization of new clinical trials.

Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. This study was funded by the authors.

Ethical approval
Ethical approval was obtained from the local Human Research Ethics Committee -Federal University of Santa Maria and the protocol (no. 55992216.1.0000.5346) was written in accordance with the standards set by the Declaration of Helsinki.