Participant characteristics of existing exercise studies in persons with multiple sclerosis – a systematic review identifying literature gaps

Background: Exercise is a cornerstone in rehabilitation of persons with multiple sclerosis (pwMS), which is known to elicit beneficial effects on various symptoms and potential disease-modifying effect. However, it remains to be elucidated if the existing MS exercise literature covers the full age and disability span of pwMS. Objective: To systematically review MS exercise studies and provide a detailed mapping of the demographic and clinical characteristics of included pwMS. Methods: A systematic review of MS exercise studies were performed in MEDLINE and EMBASE. From the resulting MS exercise studies, mean sample characteristics were extracted. Results: 4576 records were identified, from which 202 studies were included. Of these, 166 studies (82.2%) enrolled pwMS aged 35-54 years, 10.9% enrolled pwMS <35 years, and 6.9% enrolled pwMS ≥55 years (only 1.5% enrolled pwMS ≥60 years). 118 studies (58.4%) reported Expanded Disability Status Scale (EDSS), with 88.1% of included pwMS having an EDSS between 2.0-6.5, while only one study enrolled pwMS with an EDSS  7.0. Finally, 80% of the studies included pwMS having a disease duration of 5-14.5 years. Conclusion: Exercise studies in pwMS included primarily middle-aged (35-54 years) pwMS having EDSS of 2.0-6.5 and a disease duration of 5-14.5 years. Few exercise studies were identified in young and older pwMS, in pwMS with mild disability and severe disability, and in pwMS having shorter or longer disease durations. These findings highlight the need for further investigation of exercise in these specific subgroups of pwMS as benefits of exercise might not generalize across subpopulations.


Introduction
Despite major advancements in medical treatment of MS over the past 25 years, rehabilitation is still a cornerstone in MS treatment (Feys et al., 2016). One of the most auspicious non-pharmacological rehabilitation interventions in MS is exercise (Dalgas et al., 2019), defined as structured and planned physical activities. Several exercise modalities are known to effectively improve muscular strength and aerobic capacity in persons with MS (pwMS) having mild to moderate disability, as well as mobility, fatigue, mood, and quality of life (Dalgas et al., 2019;Taul-Madsen et al., 2021). Furthermore, several studies have suggested that exercise might even postpone and slow down disease progression in MS thereby eliciting disease-modifying effects (Dalgas et al., 2019;Kjølhede et al., 2018;Wesnes et al., 2018). Consequently, recent reviews from our group proposed to prescribe tailored exercise as "medicine" to pwMS alongside conventional medical treatment, already at an early disease stage (Dalgas et al., 2019;Riemenschneider et al., 2018). Exercise is therefore considered an essential element of MS rehabilitation.
Even though exercise is a cornerstone of MS rehabilitation, duration and type of exercise researched across different age groups and disability levels have not been extensively mapped. Such a mapping is central when aiming to identify relevant literature gaps and clarify limitations of the existing exercise recommendations (Dalgas et al., 2008;Kalb et al., 2020;Kim et al., 2019;Latimer-Cheung et al., 2013). There have been attempts and preliminary studies exploring this issue (Lai et al., 2018;Rimmer et al., 2010), yet a comprehensive and systematic mapping of the MS populations enrolled in exercise studies is still pending in the existing literature. Therefore, the primary objective of the present systematic review is to perform a detailed mapping of the existing MS exercise literature in terms of participant age and disability level, as well as to categorize the applied interventions and study types.

Methods
This review is reported in accordance with the Preferred Reporting Items for Systematic Reviews guideline and preregistered at PROSPERO (CRD42022297563) (Page et al., 2021). Paper screening was at all stages done independently by two reviewers (JJF and TG) using Covidence (Covidence systematic review software, Veritas Health Innovation, Melbourne, Australia). Disagreement between the two reviewers was resolved by consensus, and a third reviewer (UD) was consulted if consensus could not be reached.

Definitions
Many previous reviews have applied the definition of exercise proposed by Caspersen et al. 1985, defining exercise as ‖physical activity that is planned, structured, repetitive, and purposive in the sense that improvement or maintenance of one or more components of physical fitness is an objective‖ (Caspersen et al., 1985). Caspersen et al. further define physical activity as -any bodily movement produced by skeletal muscles that expend energy‖ (Caspersen et al., 1985). As this definition implies that physical activity covers all activity where a person is not lying still, the following modification to Caspersen's definition of physical activity was made; -physical activity involves voluntary activation of the skeletal muscles, that results in movement of a body part in time and space". Therefore, the following definition of exercise was applied in the present study: -Exercise involves voluntary activation of the skeletal muscles, that results in movement of a body part in time and space in addition to being planned, structured, repetitive, and purposive in the sense that improvement or maintenance of one or more components of physical fitness is an objective‖.

Search
A systematic literature search was performed in MEDLINE/PubMed and EMBASE on November 11 th , 2021. The search was based on medical subject headings and included the (MeSH-) terms -exercise‖ OR -exercise therapy‖ (-Kinesiotherapy‖ in EMBASE) AND -multiple sclerosis‖ with a date range of year 1000 to November 11 th , 2021.

Inclusion criteria
The included studies: had to evaluate one or more of the following exercise interventions (which broadly cover the exercise spectrum): were reported in peer-reviewed papers were published in English, Danish, Swedish, Norwegian, or German.
had a description of the intervention that allowed characterization of the exercise modality and evaluated an exercise intervention where exercise constituted the main part of the intervention.
had to apply one of the following designs: Randomized controlled trial (RCT), nonrandomized controlled trial, intervention trial without control, or pilot RCT (i.e., RCTs specifically described as pilot studies by the authors).

Exclusion criteria
Studies were excluded if: -They evaluated acute exercise interventions or exercise interventions lasting < 2 weeks.
-They were outside study design restrictions -They reported data from a study that was previously published (only the first published paper from a study was included).
-Evaluated interventions outside the included exercise interventions, e.g., the following interventions:

Data Extraction
After screening, data extraction from the included studies was performed, extracting information on study design, intervention type, duration of intervention, study sample, and baseline characteristics.
The baseline characteristics extracted included age, gender, EDSS, height weight, body mass index (BMI), MS sub-type, and disease duration (i.e., time since diagnosis). After data extraction, all data points were checked to ensure data quality. In cases where studies only reported individual values, mean values, and standard deviations were calculated. This was also done if only separate groups were reported. A subgrouping based on EDSS and time since diagnosis was made to evaluate the distribution of exercise studies across disability and disease duration. EDSS was divided into ≤1.5, 2.0-2.5, 3.0-3.5, 4.0-4.5, 5.0-5.5, 6.0-6.5, and ≥7.0. Where EDSS scores represented mild disability (EDSS 0-2.5), moderate disability (EDSS 3.0-4.5), severe disability (EDSS 5-6.5), and very severe disability (EDSS 7+) (Kurtzke, 1983). Time since diagnosis was divided into groups spanning five years: 0-4.9 years, 5-9.9 years, 10-14.9 years, 15-19.9 years, and ≥20 years. Characteristics of each study included can be found in table 1 of the appendix.
Data from all groups included in RCTs were extracted, including control groups. The categorization of RCT groups were performed the following way: All groups in RCTs with a superiority design, comparing two or more exercise interventions, were reported as exercise groups, despite one of these groups being the methodological comparator. Further, RCTs including, in addition to the exercise group, a group receiving a non-exercise intervention (e.g. massage or scent impression), were categorized as having a -non-exercise‖ group although this is beside the applied exercise definition. Hence the number of intervention groups surpasses the number of studies included.

Statistics
All statistical analysis was undertaken using STATA software version 17 (STATA, IC 17, Stata Corp, College Station, TX, USA), and graphs were made using GraphPad Prism version 9.3.1 (GraphPad Software Inc., California, USA). Means and standard deviations (SD) were calculated.

Study inclusion
The systematic search yielded a total of 4576 records, which after the removal of duplicates, resulted in 3714 records. Screening identified 556 records deemed relevant for full-text inspection.
Of the 556 records, 5 records could not be retrieved, hence 551 records were assessed for eligibility at a full-text level. While screening the 551 records, one additional report was found from searching the reference lists. Of the 552 records, a total of 202 records were deemed eligible and subsequently included in the review (for further review details see Figure 1 and all included studies see Table 1). Table 2 summarizes population demographics, clinical characteristics, and study designs of the included studies. The identified studies covered a total of 8037 pwMS, having a mean EDSS of 3.7 ± 1.4 and a mean disease duration of 11 ± 4.1 years. Of these, information on MS-subtypes was reported in 5149 (64.1%) pwMS. The mean age across all studies was 44.7 ± 7.2 years, with more females (73.6%) than males being enrolled. The mean intervention group size across all studies was 17 participants (range: 5-67). Duration of exercise intervention ranged from 2-52 weeks with a mean of 10.7 ± 7.7 weeks. 70 studies had more than one intervention group, which explains why the total number of interventions (n=285) exceeded the total number of studies included in this review (n=202). Of the 202 included records, a total of 108 were RCTs (53.5%), 17 were controlled trials (8.4%), 40 were intervention studies without controls (19.8%), and 37 were pilot RCT studies (18.3%). Figure 3 shows the distribution of the different exercise modalities. Endurance (24.9%) and combined exercise (24.5%) constitutes 49.5% of the total number of interventions. The least prevalent modalities are balance exercise (7.0%), walking (7.3%), and body & mind exercise (9.2%).

Studies by disability level and disease duration
Data presented in tables 3 and 4 are based on the same studies as in table 2, yet here stratified by EDSS-groups (table 3) or disease duration (table 4). In total, 88.1% of the studies were conducted in pwMS with EDSS-scores between 2.0-6.5 (Table 3). A total of 13 studies (11.0%) have been conducted in patients with an EDSS of 1.0, and one single study (0.8%) was found reporting data from patients with an EDSS of 7.0

Discussion
The present review aimed to map patient characteristics, exercise modality, and study design of the existing MS exercise literature to identify potential knowledge gaps for certain subgroups of pwMS.
The main finding of the present review was that the majority of the existing studies (82.2%; n=166) had enrolled pwMS aged 35-54 years. It was further observed that pwMS with mild (<2.0) or severe (>6.5) disability scores (EDSS) only constitute 11.9% (n=14) of participants across all MS exercise studies reporting EDSS-data. Finally, pwMS with a short (<5yr) or long (>15yr) disease duration are rarely enrolled in MS exercise studies as those with a disease duration of 5-15-years account for 80% (n=100) of all studies. Altogether, these findings emphasize limitations when generalizing evidence of exercise effects in pwMS. It further highlights the need to expand exercise research to certain subgroups of pwMS, including both young and older, mild and severe disability, and short and long disease duration.
According to a newly published report from the Danish MS registry ) and a US population-based study (Wallin et al., 2019), only ~1.5% of pwMS are below 25 years of age, whereas ~49% are ≥55 years. Nonetheless, the present review observed that only 7% of the exercise studies enrolled pwMS ≥55 years and 1.5% ≥60 years. This leaves a marked discrepancy between the study population and the general population across the existing MS exercise studies. In addition, more than 80% of the existing studies have been conducted in an age span covering approximately 40% of the general MS population age span, further underlining the need to focus on specific subgroups in the years to come. For instance, merely no exercise studies exist in older pwMS. This is somewhat surprising given the widely recognized positive effects of exercise among otherwise healthy older adults (Davis and Hui, 2017;Di Lorito et al., 2021;Garatachea et al., 2015). The missing studies of exercise effect in older pwMS could be due to older pwMS being more likely to have substantial disability (see the following section for further discussion of the impact of disability), however this indicate that they would thus benefit more from exercise. Interestingly, a recent systematic review and meta-analysis by Edwards et al. (Edwards et al., 2021) found that the positive effects of exercise across modalities on participation (a complex component of quality of life) in pwMS were not affected by clinical variables. This suggests that exercise can potentially also improve quality of life even late in life, however, this warrants further investigations as no exercise studies have been performed in such populations. In addition, the systematic search revealed no exercise studies performed in the early stages of the disease course (<3 years after diagnosis). This is in line with recent reviews from our group highlighting the limited evidence on early exercise interventions, hence suggesting there could be a -window of opportunity‖ by introducing exercise at an early stage of the disease (Dalgas et al., 2019;Riemenschneider et al., 2018). However, shortly after the review process of the present systematic review, an RCT investigating 48 weeks of aerobic exercise in pwMS with a disease duration ≤2 years was published. Showing positive effects on physical function and brain microstructure in these newly diagnosed pwMS (Riemenschneider et al., 2022). Supporting the notion that if exercise is initiated before the disease causes irreversible neurological damage, this will offer a superior preventive approach as opposed to the more symptomatic treatment approach currently applied (Riemenschneider et al., 2018). Such an early phase approach is inspired by the medical field, where disease-modifying therapy is recommended to be initiated as early as possible in order to maximize neurological reserve as well as cognitive-and physical function . This would further implicate that exercise from an early stage should be prescribed as -medicine‖ (Dalgas et al., 2019).
Severe disabilities are infrequently included in MS exercise studies, highlighted by the finding of this present review. This could be due to the increasing difficulty carrying out exercise interventions in pwMS with higher EDSS (Kurtzke, 1983;Pilutti and Edwards, 2017). However, two studies have evaluated exercise in pwMS with higher disability levels (EDSS ranging from 5.0 to 8.0), showing that both endurance exercise and resistance exercise are feasible and able to induce physiological adaptations Skjerbaek et al., 2014). Furthermore, a review by Edwards and Pilutti investigated the effects of exercise in severely disabled pwMS  by including studies where some participants had an EDSS>6.0. Their findings showed promising evidence for the benefits of exercise, but some of the included pwMS had EDSS as low as 3.0, and very few of the included studies had solely enrolled participants with an EDSS >6.0 . As a consequence, the review does suggest that exercise is feasible in severe MS, whereas the conclusions regarding the effects are less robust (Pilutti et al., 2014), emphasizing the need for further studies. This is important as a disability score 7.5 is a major risk factor for death, being almost 4 times the rate of healthy controls (Sadovnick et al., 1992). Interestingly, data from healthy older adults having high physical activity levels show a substantially reduced mortality rate (Sherman et al., 1994), suggesting that structured physical activity (i.e., exercise) could be beneficial and perhaps even reduce the increased mortality rate of the severely disabled pwMS.
Related to the finding that the majority of pwMS included in exercise studies have mild disability, the majority of included pwMS have RRMS (71.8%), while only smaller samples of the patients have either SPMS (17.9%), PPMS (10.0%) or PRMS (0.3%). Although the observed distribution of MS phenotypes across exercise studies does resemble the generally young and middle-aged MS population quite well , it can still be questioned whether the existing results from exercise studies can be generalized across all MS phenotypes. At least, it must be considered that the existing evidence is mainly based on data from RRMS patients. Of notice, an increasing focus has been put on exercise and progressive MS in recent years .
Currently, two largescale exercise studies in progressive MS are being undertaken, the COGEx trial (Feinstein et al., 2020) and the CYPRO trial (NCT05229861), underlining that high-quality studies will appear for this subgroup of patients in the near future.

Clinical implications
In May 2020, the National MS Society in the US posted recommendations for exercise to all pwMS, based on a review from the same year (Kalb et al., 2020). These recommendations were published by some of the leading researchers and experts in the field and aimed to provide recommendations for clinicians treating pwMS (Kalb et al., 2020). The recommendations are categorized according to and 9.0) and were based on both evidence and expert consensus when necessary (Kalb et al., 2020). However, the present review finds only one exercise study investigating exercise in pwMS with EDSS 7.0. This entail that future exercise recommendation would benefit from more exercise studies in severely disabled pwMS, as for now expert opinion is the best resource in the development of recommendations for pwMS with EDSS ≥7.0. The same applies to old pwMS and pwMS with short (<4.5y) or long (>20y) disease duration. This underlines the need for further exercise studies in select MS subgroups, even though this introduces several challenges, in particular the most severely disabled pwMS.

Limitations
When interpreting the results from the present review, various limitations must be kept in mind.
First, this review follows the PRISMA guidelines for systematic reviews, except that no quality assessment of the included studies was performed. This was omitted as several different study designs were included, which makes it complex to facilitate via one single quality assessment tool.
Furthermore, the review does not evaluate the outcomes/effects of the included studies, however, it exclusively maps the demographic and clinical data of the pwMS included in these exercise studies.
Second, since exercise can be somewhat difficult to define, the applied definition could potentially exclude studies that would consider exercise studies by some. Third, it is acknowledged that the categorization of age groups, EDSS groups, and disease duration groups is based on average values reported by the identified studies. Consequently, some studies may have enrolled participants that are within the range of some of the sparsely investigated MS subgroups. Fourth, the distinction between RCT and pilot RCT was based solely on the author's description of the study design in the papers. Hence, an objective distinction between RCT and pilot RCT was not made, potentially leading to some RCT and pilot RCT being very similar, e.g., sample size.

Conclusion
The present review identified a large body of exercise studies (n=202) evaluating exercise in pwMS. These studies primarily including middle-aged (35-54y) pwMS having disability scores of 2.0-6.5 and a disease duration of 5-14.5 years. Exercise studies have mainly evaluated short-lasting (≤ 12 weeks) endurance exercise and combined exercise. Very few studies evaluating exercise interventions were identified in young (<35 y) or older (>54y) MS populations, as well as in pwMS with mild (<2.0) or severe (>6.5) disability scores (EDSS) and short (<5y) or long (>15y) disease duration. Our findings suggest that no well-founded evidence-based guidelines for exercise can be given to these specific groups of pwMS, encouraging future exercise studies and research in these under-investigated MS subpopulations.
Note: Data presented are Mean ± SD, percentages, or number of studies/participants. EDSS = Expanded Disability Status Scale; BMI = Body Mass Index; RR = Relapse-remitting multiple sclerosis; SP = Secondary progressive multiple sclerosis; PP = Primary progressive multiple sclerosis; PR = Progressive-relapsing multiple sclerosis; Non-Exercise is when a study has tested the effects of a non-training component (e.g., temperature, smell, sound etc.) *MS subtype Total are presented as the total number of participants in the 4 subtypes of MS.   Note: Data presented are Mean ± SD, percentages, or number of studies/participants. EDSS = Expanded Disability Status Scale; BMI = Body Mass Index; RRMS = Relapse-remitting multiple sclerosis; SPMS = Secondary progressive multiple sclerosis; PPMS = Primary progressive multiple sclerosis; PRMS = Progressive-relapsing multiple sclerosis; Non-Exercise = A exercise intervention study but also testing a non-training component (e.g. temperature, smell, sound etc.). *MS subtype total is presented as the total number of participants in the 4 subtypes of MS.