Fatigue is associated with physical inactivity in people with multiple sclerosis despite different environmental backgrounds: merging and comparing cohorts from Turkey and Israel

Background Examining physical activity in people with multiple sclerosis (PwMS) from different environmental backgrounds might increase understanding and awareness of inactivity in PwMS. Objectives To compare the level of physical activity in PwMS in Israel and Turkey and to examine the relationship between the level of physical activity with common disease-related symptoms and demographical characteristics of PwMS. Methods Cross-sectional data collected by two centers were combined. The physical activity level was determined by the Godin Leisure-Time Exercise Questionnaire, and subsequently, classied into one of three subgroups: "active", "moderately active" and "insuciently active". Logistic regressions determined the risks of insuciently active PwMS, according to age, gender, body mass index, disability, impact of walking impairment, disease duration, type of MS and perceived fatigue. The analysis of variance test determined the differences between countries in terms of outcome variables. Results The study comprised 458 patients from Israel and 575 from Turkey; 68.2% Turkish PwMS were classied as insuciently active compared with 52.0% of Israeli PwMS. The percentage of insuciently active PwMS was signicantly higher in those categorized as fatigued compared to non-fatigued in the total cohort (72.4% vs. 51.9%, p<0.001) and in each country separately. Based on the regression analysis, fatigue was the main factor associated with the insuciently physically active group; odds ratio=1.968. Interpretation PwMS with increased fatigue tend to be physically inactive compared with the non-fatigued. This observation is supported by the merged data collected from two countries, Turkey and Israel, representing PwMS from different environmental backgrounds. the differences between countries in terms of MS type, gender distribution, physical activity level distribution, fatigue status and EDSS disability subgroups. Differences between countries in terms of age, disease duration, body mass index (BMI), EDSS, GLTEQ, MSWS-12 and MFIS were determined by the analysis of variance test. Univariate and multivariate logistic regression determined the risk of classication as insuciently active according to age, gender, BMI, EDSS, MSWS-12, disease duration, type of MS and perceived fatigue categories. All analyses were performed using the SPSS software program (IBM SPSS for Version 27.0 Armonk, NY, USA: IBM All reported P values were two-tailed. The level of signicance was set at P


Introduction
Multiple sclerosis (MS) is a demyelinating disease of the central nervous system (CNS) [1] with an estimated 2.8 million persons a icted globally [2,3]. The disease process produces a diversity of neuropathological changes in the CNS [1,4] typically affecting a wide range of neurological functions including mobility, cognition, vision, muscle strength, endurance, coordination, and sensation. There is a consensus amongst clinicians that physical activity is imperative for people with MS (PwMS). Emerging evidence has demonstrated that physical activity can improve walking performance, balance, cognition, fatigue, depression and the quality of life of PwMS [5]. Furthermore, Barry et al [6] demonstrated that physical activity and exercise are potential immunomodulatory therapies targeting innate signaling mechanisms, and modulating MS symptom development and progression. Nonetheless, regardless of the bene ts, up to 78% of PwMS do not participate in any meaningful physical activity [7]. Interestingly, the number of inactive PwMS contrasts with the fact that most PwMS believe that their health and function would improve if they would participate in physical activities [8].
The prevalence of physical inactivity in the general population varies between countries, even within countries of the same region. According to a pooled analysis of population-based surveys across 168 countries, the prevalence of people classi ed as physically inactive ranges from 5 to 67% [9]. Althoff et al [10] assessing physical activity via movement sensors, found disparities in physical activity participation between countries. The relatively large disproportions are explained by dissimilarities in environmental backgrounds including policies/guidelines of physical activity recommendations for adults between countries.
Only sparse data exist relating to the physical activity participation of PwMS from different environmental backgrounds. Reily et al [11] examined the prevalence of various comorbidities and modi able lifestyle factors in PwMS, including the level of physical activity performed in different global regions (Australia, Europe, and North America). The authors found a strong inverse relationship between the level of physical inactivity and the physical aspects of quality of life in the total cohort. They also observed that the distribution of physical activity levels varied between regions. Whereas, 29.9% of PwMS were classi ed at a high physical activity level in North America, only 22.2% were classi ed at that level in Europe [11].
Examining physical activity in PwMS across different countries from different environmental backgrounds could be bene cial. Firstly, by examining PwMS from different environmental backgrounds can help clarify the precise impact of disease-related factors as barriers of inactivity in PwMS. Secondly, in the event of signi cant differences in physical activity, health professionals from one country might offer (or ask) guidance/advice from health professionals from another country. In the current study, we merged two relatively large cohorts from Turkey and Israel (n = 1033). This decision was made since both countries are located in the same global region; characterize patients with different environmental backgrounds (organizational and social) and use similar measurement tools for assessing common MS disease symptoms. Therefore, the study objective was twofold: (1) to compare the level of physical activity in PwMS in Israel and Turkey (speci cally, Izmir) representing PwMS from different environmental backgrounds; (2) to examine the relationship between the level of physical activity with common disease-related symptoms and demographical characteristics of PwMS in the total cohort and in each country separately.

Study design and participants
Cross-sectional data collected by two centers were merged. All data were collected in accordance with the International Declaration of Helsinki. Institutional Ethics Committees in both countries approved the separate protocols (GOA-2959 (Turkey); SMC-5596 (Israel)). Written informed consent was obtained from the Turkish participants. The Sheba Institutional Review Board Ethics Committee approved the extraction of demographic, clinical, and physical activity data for analysis in addition to a full exemption from written or verbal consent from the study participants. Hence, individual data will not be made available in order to protect the participants' identity. Each patient's record was referenced by an anonymous code number to ensure con dentiality during the statistical analyses.
The cross-sectional bi-center study comprised 1033 PwMS (40.1±12.3 years old, 65.0% female); 458 participants from the Multiple Sclerosis Center, Sheba Medical Center, Tel-Hashomer, Israel and 575 participants recruited from the Multiple Sclerosis Unit, Dokuz Eylul University Hospital, Izmir, Turkey. Both samples used standardized criteria for an MS diagnosis [12]. Participants were >18 years old. There were no restrictions for MS subtypes. Disease severity was measured in both centers by the Expanded Disability Status Scale (EDSS) [13]. The EDSS score for inclusion was de ned at ≤ 6.5, equivalent to walking 20m with bilateral support. Common exclusion criteria were: relapse free for 90 days prior to the examination and an inability to understand and complete the patient-reported outcome measures due to linguistic, cognitive or other reasons.

Outcome variables
Godin Leisure-Time Exercise Questionnaire (GLTEQ) The GLTEQ, a self-administered measuring tool for assessing physical activity [14] has been validated for use in PwMS [15]. This tool contains three items which measure the frequency of strenuous (i.e., jogging), moderate (i.e., fast walking) and mild (i.e., easy walking) exercise for periods of >15 minutes during one's free time throughout a typical week. The weekly frequencies occurring over a 7-day period of strenuous, moderate, and mild activities are multiplied by 9, 5, and 3 metabolic equivalents, respectively, and summed to form a measure of total leisure activity. The GLTEQ is an appropriate, simple, and effective tool for measuring the change in physical activity in response to an intervention in the MS population [16].

12-item Multiple Sclerosis Walking Scale (MSWS-12)
The MSWS-12, a valid questionnaire assessing walking ability in PwMS, is the most widely used patient-reported measure of perceived limitation in walking attributable to MS. Many studies recommend the use of the MSWS-12 due to its psychometric properties [17][18][19]. Each of the 12 items are rated on a scale of 1 (not at all) to 5 (extremely). Items cover different aspects of walking function and quality such as the ability to walk, walking speed, ability to run, ability to climb and descend stairs, ability to stand, balance, endurance, smoothness of gait, need for support (in and outdoors), effort and concentration required.

Modi ed Fatigue Impact Scale (MFIS)
The MFIS [20], a self-reported, multidimensional 21-item questionnaire, determined the level of perceived fatigue in the PwMS study group. This instrument captured information relating to the effects of fatigue within the physical (9 items), psychosocial (2 items), and cognitive (10 items) domains over a four-week period. Participants rated the 21 items on a 5-point Likert-type scale from never (0) to always (4). The MFIS yielded three subscale scores and an overall score ranging from 0 to 84.

Data analysis
All data from each country were normally distributed according to the Kolmogorov-Smirnov test. Box plots determined outliers for each outcome. Data sets from each country were compared for heterogeneity by the Kruskal-Wallis Test. As for the level of physical activity, individuals from each data set were classi ed into one of three subgroups: "active", "moderately active" and "insu ciently active". Allocation was determined according to the GLTEQ. PwMS with scores ≥24 units were classi ed as active, 14 to 23 were considered moderately active and those who scored <14 were classi ed as insu ciently active. The selected cut-off points were aligned with Godin's recommendations [21] and de ned according to the literature [22]. Participants with disabilities were categorized as either "very mild" (EDSS 0-1.5), "mild" (EDSS 2.0-3.5), "moderate" (EDSS 4.0-5.5) or "severe" (EDSS 6.0-6.5). Perceived fatigue was categorized as "fatigued" (MFIS ≥38) or "non-fatigued" (MFIS<38). The chi-square test determined the differences between countries in terms of MS type, gender distribution, physical activity level distribution, fatigue status and EDSS disability subgroups. Differences between countries in terms of age, disease duration, body mass index (BMI), EDSS, GLTEQ, MSWS-12 and MFIS were determined by the analysis of variance test. Univariate and multivariate logistic regression determined the risk of classi cation as insu ciently active according to age, gender, BMI, EDSS, MSWS-12, disease duration, type of MS and perceived fatigue categories. All analyses were performed using the SPSS software program (IBM SPSS Statistics for Windows, Version 27.0 Armonk, NY, USA: IBM Corp). All reported P values were two-tailed. The level of signi cance was set at P <0.05.

Results
The total cohort's GLTEQ score was 13.3 ± 17.7. Signi cant differences in physical activity were observed between countries. Israeli PwMS participated in more leisure-time physical activities compared to the Turkish PwMS; GLTEQ = 16.4 ± 18.4 vs. 10.7 ± 16.6; p < 0.001. Moreover, 68.2% of Turkish PwMS were classi ed as insu ciently active compared with 52.0% of the Israeli PwMS. For comparison, we present data on the percentage of people classi ed as insu ciently active in the general population published by the World Health Organization [9]. The mean percentage in the Middle East and eastern Europe regions is 32.8% and 23.4%, respectively. The mean percentage in Turkey is 30.6%. To the best of our knowledge, no data exists as to the percentage of people classi ed as insu ciently active in Israel, although, the percentage is considered in the same range as countries classi ed with middle-high incomes (between 26 and 36%). Demographic and clinical characteristics of the study cohort are presented in Table 1.
Distribution of physical activity subgroups according to disability (EDSS), fatigue categories, MS type and gender are illustrated in Figs. 1-4. The percentage of the insu ciently active group was higher in Turkish PwMS compared to the Israeli PwMS for all disability levels, with the exception of severely disabled, which was identical. Furthermore, the percentage of insu ciently active people was signi cantly higher in PwMS categorized as fatigued compared to non-fatigued in the total cohort (72.4% vs. 51.9%, p < 0.001) and in each country separately. A higher percentage of females were classi ed as insu ciently active compared with males in the total cohort, 63.7% vs. 55.6%; p < 0.001.
Based on the multivariate logistic regression analysis of the total sample, being female and fatigued were the main factors associated with the insu ciently physically active group; odds ratio = 1.364, 1.968, respectively. Moreover, fatigue was the single signi cant risk factor for physical inactivity in each country. Table 2 presents the logistic regression analysis evaluating factors associated with insu ciently physically active. Figure 5 illustrates the contribution of fatigue on physical inactivity in terms of odds ratio according to country and in the total cohort.

Discussion
The main ndings of this research were the strong association between increased perceived fatigue with decreased participation in leisure time physical activities in PwMS. PwMS, categorized as fatigued, had a ~ 2.5-times higher risk of becoming insu ciently physically active compared to the PwMS who were non-fatigued. Importantly, a novelty of our study was the analyses of merged data from two relatively large samples collected from different environmental backgrounds. Although, several systematic reviews have investigated this topic, to the best of our knowledge, this is the rst study comparing physical activity behavior in PwMS according to country of origin. We decided to focus speci cally on Israel and Turkey as these two countries represent PwMS from the same global region but with different environmental backgrounds. In this context, it is worth mentioning a recent study exploring expert views on facilitators and barriers on long-term physical activity among PwMS [25]. The authors divided potential factors into two major categories: environmental and personal.
Environmental factors included organizational aspects, e.g availability, access and quality of physical activity options, health system services, health professionals, nancial resources, social and peer support. Personal factors related to the individual, i.e., age, gender, disease related factors, motivation, skills, goals, exercise history, knowledge and outcome experiences/expectancies.
Our study does not capture all potential facilitators and barriers of physical activity participation in PwMS, however, we do add signi cant information relating to the contribution of common disease-related factors on inactivity in PwMS. Regardless of the different environmental backgrounds, together with different rates of physical activity participation between the two cohorts, we found that perceived fatigue was a major contributor of inactivity in both cohorts of PwMS.
Worth noting, disability (represented by the EDSS), BMI, type of MS, disease duration and subjective impact of the disease on mobility (represented by the MSWS-12) were not signi cant factors for inactivity according to the multivariate logistic regression. Interestingly, our nding regarding the MSWS-12 indicates that when the goal is to encourage physical activity in PwMS, more effort should be invested in the aspect of reducing fatigue versus improving mobility.
The signi cant inversed relationship between perceived fatigue and participation in physical activity which was demonstrated in our study is in agreement with the MS literature [24][25][26][27]. Perceived fatigue has been found to decrease the likelihood of engaging in physical activity [28,29]. Due to the current study's design, we could not draw rm conclusions as to the cause-effect relationship between fatigue and physical activity, therefore, we reference several previous studies which may lend an insight into this complex relationship. Smith et al [29] investigated how fatigue in uences community exercise participation in PwMS by collecting information via a face-to face personal interview. Speci cally, the authors focused on how PwMS managed to engage in community exercise despite their fatigue. They also reported that what characterized PwMS who were able to maintain exercising, was the ability to readjust goalorientated behaviors on a continual basis. It is believed that the ability to adjust the PwMS' goals is linked to the levels of self-e cacy and should, therefore, be considered when encouraging PwMS to engage in meaningful exercise. Motl et al [30] found an association between increased fatigue, physical inactivity, and low self-e cacy. The authors investigated the intermediary role of disability, fatigue, mood, and self-e cacy between physical activity and quality of life in PwMS. In the present study, data were not collected relating to mood or self-e cacy. Nevertheless, we reinforced the existing knowledge by demonstrating that elevated fatigue is related to less participation in leisure time physical activities in PwMS, regardless of country (Israel and Turkey), health system or national physical activity guidelines.
The relationship between physical activity and fatigue has been observed in various clinical trials reporting on the e cacy of exercise programs on fatigue in PwMS. According to Andreasen et al's systematic review [31], physical activity and exercise training have the potential to generate a positive effect on PwMS fatigue. Turner et al [32] examined the e cacy of a telephone-based physical activity counselling intervention of 64 PwMS and found that participants in the intervention group signi cantly reduced perceived fatigue and depression upon termination of the program. In another randomized controlled trial, a signi cant reduction in fatigue was recorded following a 6-month behavioral intervention targeting lifestyle physical activity in 82 PwMS [33].
Despite the focus of our discussion on the impact of disease-related factors, speci cally fatigue on inactivity in PwMS, we nd it important to mention other factors as well. A complex range of factors in the micro and macro environments in uence the likelihood that an individual will be physically active. Factors in the macro environment include general socio economic, cultural, and environmental conditions, i.e., participation in leisure-time physical activity tend to be directly related to socioeconomic status [34]. Additionally, cultural background may affect the choice of leisure-time activities. Communities can strongly in uence people's levels of physical activity, particularly through social support systems [35]. According to a Eurobarometer survey, disparities exist throughout the European Union as to the extent to which people are acquainted with activity in their local areas [36]. Based on the present comparison between Israeli PwMS and Turkish PwMS, we believe that besides common disease features, future research studies examining physical activity in PwMS should consider macro environment factors.
Our study has many strengths such as the large, pooled sample of PwMS, however, there were several limitations which warrant attention. Firstly, physical activity was quanti ed solely by a self-reported questionnaire. Other tools such as accelerometers and pedometers are equally appropriate in measuring physical activity, although, they also have their limitations [37]. In this context, it is worth noting that in a recent validation study of the GLTEQ in PwMS, the scores primarily re ected moderate-to-vigorous physical activity rather than light-physical activity and sedentary behavior [38]. Secondly, the level of fatigue was solely estimated from data taken from the MFIS questionnaire without analyzing the test's subcategories, although, it is questionable whether additional fatigue measures would have led to different conclusions. Finally, our data did not include all potential factors related with long-term physical activity adherence in PwMS, as we merged only data that was collected at both MS centers.
In conclusion, participation in leisure-time physical activity is related to fatigue in PwMS. PwMS with increased fatigue tend to be physically inactive compared with the non-fatigued. This observation is supported by the merged data collected from two countries, Turkey and Israel, representing PwMS from different environmental backgrounds. Additional personal and environmental factors should be considered in future studies examining physical activity behavior in the PwMS population to further clarify this important topic.

Declarations
Author contributions ATO cconceptualized; formal analysis; writing -review and editing; visualization. TK: Writing-review and editing. SO Writing-review and editing; resources. AA writing-review and editing; resources. AK: Conceptualized; writing-original draft preparation; visualization.
Compliance with ethical standards Con icts of interest The authors declare that they have no con ict of interest.
Ethical approval The Institutional Ethics Committees in both countries approved the separate study protocols (GOA-2959 (Turkey); SMC-5596 (Israel). Written informed consent was obtained from the Turkish participants. The Sheba Institutional Review Board Ethics Committee approved the extraction of demographic, clinical, and physical activity data for analysis in addition to a full exemption from written or verbal consent from the study participants. Hence, individual data will not be made available in order to protect the participants' identity. Each patient's record was referenced by an anonymous code number to ensure con dentiality during the statistical analyses.

Variable
Total Sample       Contribution of fatigue on physical inactivity in terms of odds ratio.