Introduction

Multiple sclerosis (MS) is a chronic and progressive disease, which is a demyelinating disorder attacking the central nervous system (CNC), in young adults [1,2,3,4,5,6]. MS is now found to be more prevalent in women than men (3:1). The onset of this disease is between 20 and 40 years of age [7,8,9,10]. The National MS Society in the USA announced that MS affects approximately 2.5 million people around the world, and nearly 200 new cases are diagnosed each week in the USA [11]. According to the Iran MS Society, there are more than 40,000 diagnosed MS cases in Iran, while only 9000 of them have been registered. The main cause of the disease is still unknown but it is known as an autoimmune disease or inflammatory disorder [12]. Most common symptoms of MS are as follows: sensory problems, weakness, low muscle tone, vision problems, cognitive impairment, fatigue, organ temper, voiding dysfunction, forgetfulness, strabismus, and so forth. MS disease influences all aspects of individual and public life of patients [12]. Studies have shown that the quality of life (QOL) in MS patients is significantly lower than that of healthy persons and other people with chronic diseases like epilepsy, diabetes, rheumatoid arthritis, and irritable bowel syndrome [13,14,15,16,17,18,19]. Since MS is a chronic disease and there is not a certain cure for it, a decrease in intensity of clinical signs and, in turn, protection of QOL in the patients in the appropriate level are the main goals of clinical cares [13,14,15,16,17,18]. As pharmacotherapy has side effects and high costs, non-medicated ways such as exercise therapy can be considered suitable [13,14,15,16,17,18]. In the past, doing sport activities was not recommended for the MS patients because of increasing body temperature and causing fatigue, which intensify the disease [20,21,22]. Considering the potential of sport activities and rehabilitation in reducing different physical and mental problems in these patients, various studies on different sports like aerobics and endurance exercises, yoga, hydrotherapy, and so on with different protocols (frequency, intensity, duration, and type of activity) have been conducted in both Iran and the world [13, 17, 23,24,25,26,27,28,29,30,31,32,33,34,35,36]. The results of a few studies show that different exercises with different protocols can be effective or ineffective in some physical dimensions such as reducing muscle spams, fatigue, and pain; can be beneficial in increasing the power of the patients, flexibility of muscles, joints, and bone moves; and can be ineffective for some mental dimensions like improvement in mood and self-confidence or decreasing depression [23, 34, 37]. Recently, aggregation of the results extracted from review articles has changed the guidelines to present suggestions regarding training methods for these patients [31]. Up to now, few review articles have investigated the effects of exercise on total QOL in MS patients [28, 31]. A descriptive research with the aim of developing a guideline in this area found that, despite lack of evidence with high quality, there is sufficient evidence in order to develop physical activities in these patients [31]. The current research has not evaluated quantitatively the size of the effect and included all random and non-random trials [31]. To our knowledge, only in one review the size of the exercise effect on quality of life of patients has been reported, which has been reported little, but significant [28]. It should be noted that this study was comprised of both experimental and non-experimental trials, which is considered a limitation [28].

The primary purpose of our study was to estimate the effect size of exercise therapy on each of mental and physical dimensions of the QOL in MS patients in Iran only by using the results of randomized controlled trials (RCTs) as the best evidence. The second aim was to evaluate the effect of some selected theoretically or practically significant factors on this estimation (such as frequency, duration, type of exercise, type of questionnaire, articles’ quality, and so on).

Methods

Search strategy, inclusion and exclusion criteria

Relevant studies were identified by a computer-aided search of PubMed, Scopus, Google scholar and Persian medical databases including IranMedex, Irandoc, Magiran, Scientific Information Database (SID), and Medlib. In addition, supplementary searches were performed manually by reviewing the reference lists of the relevant articles. The following comprehensive combination of keywords in English and Persian were used: “multiple sclerosis,” “MS,” “exercise,” “training,” “physical activity,” “quality of life,” “randomized trial,” and “randomized controlled trial.” The inclusion criteria for the articles were (1) being conducted in English or Persian in Iran, (2) focusing on only patients with MS, (3) having the randomized controlled trial (RCT) design, (4) having either an exercise intervention of at least 3 weeks or a comparison condition lacking exercise therapy (i.e., with usual care or activities), and (5) measuring at least one of the outcomes of physical or mental factors influencing QOL before and after exercise therapy. Studies were excluded if they (1) included exercise as one part of a multicomponent intervention but did not include the additional component in a comparison condition, (2) compared exercise only with an active treatment (e.g., massage therapy alone or another mode of exercise), (3) provided incomplete data or not available data. In this review, a wide range of exercise modalities such as aerobic, endurance training, resistance training, aquatics, and yoga or a mixtue of them with at least 3 weeks duration were all included under the umbrella term of exercise. A comprehensive search of electronic databases was conducted from January 2008 to July 2015.

Selection of studies and data extraction

All publications identified by the search strategy were downloaded into EndNote (version X7), and repeated cases were deleted. A reviewer screened publications for inclusion according to the pre-specified criteria. If it was clear from the title or abstract that a study did not meet the inclusion criteria, it was excluded. If it was unclear from the abstract whether a study met the selection criteria, the full paper was retrieved. Excluded studies were reviewed twice by a second reviewer. One reviewer extracted data from all included studies in a structured form. All extracted data were checked by the second reviewer. The following data were extracted from the included studies: first author, publication date, study design, number of participants, patient characteristics (e.g., age, sex, expanded disability status scale (EDSS)), type of experimental and control intervention, characteristics of the exercise intervention (e.g., exercise mode, number of sessions, time per session, frequency in weeks, relative intensity, and adherence), time and scale of outcome assessment, and mean and standard deviation of the outcome measures in experimental and control intervention groups.

Quality assessment of studies

Independently, two reviewers rated the quality of each included study according to the Physiotherapy Evidence Database (PEDro) assessment tool, and any disagreement was resolved by discussion [38]. The scale evaluates the methodological aspects of RCTs such as having specified eligibility criteria, random allocation, concealment of allocation, comparability of groups at baseline, blinding of patients, investigators or outcome assessors, having completed follow-up for more than 85% of participants, analysis by intention to treat, between-group statistical comparisons, and complete outcome data (report of both point estimates and measures of variability at least for one outcome measure). The maximum possible score of the PEDro scale was 11, and higher scores indicated better methodological quality (Table 1).

Table 1 Quality score of papers according to the PEDro scale
Full size table

Statistical analysis

Effect sizes were calculated by subtracting the mean change in the comparison condition from the mean change in the experimental condition, divided by the pooled standard deviation of baseline scores [49]. Effect sizes and 95% confidence interval (CI) were calculated for physical and mental QOL separately and are shown in the forest plots. The heterogeneity of the studies was checked visually by the Galbraith plots and assessed by the Cochrane chi-squared test and I2 statistics. Since the studies were heterogenous, a random-effect model was used to calculate the mean effect size. To adjust the mean effect size for the sample size of each study, Hedge’s adjusted G was used [49]. To assess the hypotheses about the variation in the effect size by the moderators, variables such as sex, type of exercise, program length, and quality of study were assessed in the meta-regression model and subgroup analyses were conducted based on the significant moderators. According to the quality score, studies were divided into two groups: high quality and low quality, as the quality scores were ≥6 or not. To examine the risk of publication bias, the egger regression test was used. Moreover, a sensitivity analysis was performed to assess the contribution of each study to the overall effect. All statistical analyses were performed using STATA 11 (Stata Corp LC, Texas, USA).

Results

Figure 1 shows the flowchart depicting the search strategy. By using the electronic searching of the databases, 2367 publications (published from January 2008 to June 2015) were identified. After removing duplicated records, screening the articles related to the QOL in MS patients and hand-searching of the screened references, 21 articles were retrieved. Seven of 21 publications were excluded because the full texts were not available in two studies; the control group was missing in two studies [50, 51]; only some aspects of QOL were investigated in one study [34]; and one study was published in two reports and, then, aggregated [52], and there were not enough data on QOL measurement in one study [53].

Fig. 1
figure 1

Flowchart depicting the search strategy

Full size image

Participants

Thirteen RCTs involving a total of 535 patients (279 patients in the intervention group and 256 patients in the control group) were included in this study. Individual study sizes ranged between 21 and 90 patients. Only in five studies, men entered into the studies. In total, 53 (10%) patients were male (30 patients in intervention group and 23 patients in control group). The mean (standard deviation) of patients’ age was 33.75 years (2.79). Six studies were published in English and seven studies in Persian languages. The EDSS score was available in eight studies with a mean (sd) of 2.24 (0.49).

Outcome measures

Most of the studies (n = 7, 54%) measured the patients’ QOL by the Multiple Sclerosis Quality of Life-54 questionnaire; three (23%) studies used Short-Form Health Survey-36 (SF-36) instrument; and three remaining studies were assessed QOL by using the SF-8 QOL, WHOQOL-BREF (World Health Organization Quality of Life-short version), and FAMS.Version2 (functional assessment of multiple sclerosis), respectively. Characteristics of the included studies [18, 24, 39,40,41,42,43,44,45,46,47,48, 54] are shown in Table 2. The PEDro scores of paper quality range from 4 to 8 points, whereas 54% of the papers have the score of 6 or more (Table 2).

Table 2 Patients and outcome characteristics of included studies
Full size table

Interventions

A wide variety of types of exercise were used in the studies but the aquatic exercise was the most common one (Table 3) [18, 41, 44, 46]. The mean (sd) length of the exercise was 9.57 weeks (2.50) and usually, the training sessions were done thrice per week (range 2–3 per week). The time of exercise ranged from 20 to 75 min per session (with 30–40 min per session in 57% of the interventions) and mean (sd) of 47 min (16). Proportion of drop-outs reported in the studies varied from none to almost 32%. None of the studies reported any adverse events for patients during the exercises.

Table 3 Exercise characteristics of included studies
Full size table

Estimated effects of the interventions

Overall, 34 effect sizes were retrieved from the 13 studies because in some studies, there was more than one effect size per study. In one study [39], two different moods of exercise outcome were reported, and in three studies [18, 40, 45], the outcome was measured in different times following the exercise. Twelve effects were derived for overall QOL, and 11 effects were derived for mental and physical subscales of QOL (Fig. 2). In one study, the quality of work life was measured [41]. The effect size of each intervention has been shown in Fig. 2. As the effect sizes of mental, physical, and overall QOL were heterogeneous (Q = 19.95, df = 10, P = .030 with I2 = 49.9% for mental QOL, Q = 21.77, df = 10, P = .016 with I2 = 54.1% for physical QOL, and Q = 29.49, df = 11, P = .002 with I2 = 62.7% for overall QOL), the random effect models were used to calculate the mean effect size. The mean effect sizes of exercise therapy on mental, physical, and overall QOL were 1.021 (95%CI 0.712–1.331, P < .001), 1.040 (95%CI 0.730–1.349, P < .001), and 0.846 (95%CI 0.508–1.184, P < .001), respectively. The Galbraith plot identified that Kargarfard et al.’s [18] study is the most heterogeneous on the estimation of training effect on mental and physical QOL; however, exclusion of this study did not significantly alter the results; and with the study excluded, the effect sizes of exercise therapy on mental and physical QOL were 0.941 (95%CI 0.672–1.210) and 0.914 (95%CI 0.699–1.128), respectively. Furthermore, three effects of (Hassanpour-dehkordi et al. [39] and Eftekhari et al. [47]) studies were the most heterogeneous on the estimation of training effect on overall QOL and with exclusion of these studies, the mean effects estimation (0.596 with 95%CI 0.310–0.882) decreased. To describe the heterogeneity, variation of the effects was tested according to some selected variables as moderators. However, none of the moderators were related to the mental and physical QOL effect size (details of the results are not shown here but P > .1 was set for all regression coefficients); the mean effect size of overall QOL was influenced by the length of the exercise (P = .049) and questionnaire type (P = .023). Therefore, the effect size of exercise on overall QOL on an 8-week exercise was lower than a 12-week-exercise (effect size 0.393 (95%CI 0.054–0.732, P = .023) vs. 1.122 (95%CI 0.668–1.575, P < .001)). Also, the effect sizes of exercise on overall QOL assessed by SF-36 or SF-8 QOL was higher than others (effect size of 1.173 (95%CI 0.726–1.619, P < .001) for SF-36 or SF-8 QOL questionnaire vs. .379 (95%CI 0.058–0.700, P = .021) for other assessment tools, respectively); albeit, there was a significant relation between questionnaire type and the length of the exercise (r = .657, P = .020).The egger regression test revealed no publication bias for exercise effects on mental and overall QOL (P = .313 and P = .500, respectively), but the test detected a small evidence of publication bias for trials assessing exercise effects on physical QOL (P = .047).

Fig. 2
figure 2figure 2

a, b, c Individual and overall effect sizes (with 95% confidence interval) for mental, physical, and overall quality of life, respectively (mental, physical and overall QOL were heterogeneous so Q = 19.95, df = 10, P = .030 with I2 = 49.9% for mental QOL, Q = 21.77, df = 10, P = .016 with I2 = 54.1% for physical QOL, and Q = 29.49, df = 11, P = .002 with I2 = 62.7% for overall QOL were calculated)

Full size image

Discussion

Exercise therapy can be introduced as a practical and safe method for patients with chronic diseases like MS with a mild to moderate disability. Several review articles have assessed the effect of doing exercises on different physical and mental problems of MS patients such as spasticity, mobility, balance, walking, fitness, fatigue, depression, and health-related quality of life [14, 25, 26, 28, 29, 35, 55,56,57,58,59,60,61,62,63,64]. Based on these studies, some guidelines and recommendations have been published [31, 65,66,67]. However, these guidelines do not accord with each other and suffer from a few drawbacks in methodology (like not being evidence-based, lack of evaluating the quality of the reviewed evidence or not setting a robust, standardized guideline development process). Also, these guidelines just review fitness outcomes descriptively and not quantitatively. Generally, it cannot be concluded that doing exercises can undoubtedly enhance health-related QOL in individuals with MS [30, 31]. The main objective of the present study was quantitative evaluation of the effect of doing exercises on overall QoL as well as physical and mental dimensions in MS patients based on the conducted studies in Iran. Unlike previous studies, in our investigation, we considered only randomized clinical trials (RCTs). Approximately half of the studies included in this study were RCTs with good quality (6 or more) [67]. The results of Iranian RCTs showed that 90% of the patients were women and different kinds of exercise interventions including aerobic; yoga; and aquatic, stretching, endurance, and combinatorial exercises have been used. In general, the length of the exercise was about 10 weeks and training sessions were 3 times per week. Moreover, in half of these studies, the MS specific tool was used for recording QOL. We found a dramatic improvement in all aspects of QOL. In other words, the mean effect sizes indicated that doing exercise makes a basic improvement in effect (about one SD) in all mental, physical, and overall QOL of patients with MS. In the review article by Cruickshank et al., the effect of the endurance exercise on the QOL in MS patients has been reported [64]. In this study, the aggregated size of the effects was not calculated. However, a significant improvement in QOL was claimed in two of three studies entered in this review article. Similar results were reported in the review article by Kjølhede et al. [55]. The effect size was calculated only in one review study, which was significant but small in magnitude [28]. The reason for this difference in the results of this study and our study can be attributed to the methodological differences of the two studies. In our study, we used only interventional studies which are stronger in terms of evidence for deduction but in the study by Motl et al. [28], effects of both interventional and observational studies were aggregated. Also, in our study, in more than half of the studies, the measurements were evaluated with a MS-specific tool, MSQOL-54, while in Motl et al.’s [28] study, the measures of QOL were quite diverse ranging from generic measures of QOL (e.g., Medical Outcomes Study Short Form-36) through measures of fatigue as a component of QOL (e.g., Fatigue Severity Scale). It is notable that a study showed that the path of exercise effectiveness on QOL is indirect due to influence of exercises on fatigue, pain, social support, and self-efficacy in individuals with MS [68]. Similar to Motl et al.’s [68] study, results of our study showed that the effect of overall QOL differed based on the type of the questionnaire. Our study illustrated that using general tools like SF8 or SF36 for measurement of QOL shows a higher effect of exercise, whereas in Motl et al., effect sizes estimated based on MS-specific tools were higher and significant. These results stated that lack of accuracy and specificity of general tools as well as lack of record and cover of all special dimensions of QOL in individuals with special diseases like MS may affect the results in both direction of over or underestimation. In addition, our data showed that a longer period of exercise (3 vs. 2 months) influences the overall QOL more. This result is consistent with that of other studies, as the time of 3-month programs is among the logical underpinnings of the exercise initiation. Also, it is worth mentioning that Motl et al. [28] explained that exercise interventions, shorter than 3 months, have a higher effect size than interventions with a period of 3 months or more. This can be described by difficulty in sustaining exercise adherence rates in the long term. However, we could not investigate this pattern because of insufficient data. In addition to the type of measuring tools and the duration of exercise program, Motl et al. [28] concluded that the kind of exercise (aerobic against non-aerobics or combination exercises), and time of each session (more than 90 min) are related to a higher effect on QOL. The authors stated that the effect of the aerobic exercise on QOL is significant, but the effect of different non-aerobic exercises (like yoga, endurance and so on) is not significant. Nonetheless, Cruickshank et al. [64] explained that the endurance exercises are effective in QOL. In our study, we did not find any relationship between the type, duration, and time of exercise per session and QOL. It may be attributed to the heterogeneity of exercise programs performed in Iranian studies and also insufficient data. Thus, considering that some of exercises are only effective in fitness parameters and do not increase the QOL and most of the present studies are descriptive, more studies are required to compare the effect of different exercises on QOL in MS individuals. In addition to the limitations of performed initial studies such as self-reporting of QOL, lack of concealment, and blinded outcome assessment, the data on clinical courses of individuals with MS (such as relapsing-remitting against progressive or different stages of disease remission) had not been reported in the included studies. Also, despite trying to collect all published studies, we observed a weak effect of publication bias in the estimation of exercise effect on physical QOL. Furthermore, it should be noted that in the present study, only eight EDSS studies were reported and the mean EDSS score was ≤4, indicating that most patients in the studies had a low disability status. Also, since the effect of exercise on men population may be different in terms of physiological features and most of the present studies have included women, more studies on men are needed. In conclusion, as no complications for the participants in exercise programs have been reported (even though this issue was not clearly mentioned by some studies), it can be stated that doing exercises is definitely helpful for improving the QOL in MS patients. And, since the rate of doing exercises in these individuals is low [27], performing interventions to enhance the exercise behavior of MS patients can be helpful. Furthermore, to develop a guideline for physical activity of these patients and to determine the properties of the most suitable and effective exercise programs such as exercise type, frequency, length, and intensity, more studies are needed.

Conclusions

The overall effect of exercise therapy interventions on QOL among individuals with MS was studied by using the meta-analysis procedures. The mean effect size of overall QOL was higher for 12 vs. 8 weeks length of the exercise and was lower for the MS-specific questionnaires. Based on the Iranian studies, there is a strong evidence of exercise therapy effect on QOL in patients with MS, but due to the heterogeneity of the studies, more studies are needed to establish the effective exercise program.