Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Aug 1.
Published in final edited form as: Rehabil Psychol. 2022 Jul 14;67(3):421–429. doi: 10.1037/rep0000454

Association of Disease Outcomes with Physical Activity in Multiple Sclerosis: A Cross-sectional Study

Trinh L T Huynh 1, Stephanie L Silveira 2, Brenda Jeng 1, Robert W Motl 3
PMCID: PMC10187883  NIHMSID: NIHMS1889186  PMID: 35834207

Abstract

Purpose:

This study examined combinations of disease outcomes (i.e., walking, cognition, and symptoms) as correlates of physical activity subgroups (insufficiently active vs. sufficiently active) in persons with multiple sclerosis (MS).

Design:

This study included 213 participants who completed walking and cognitive functional tests and self-report measures of symptoms and physical activity. Multivariate analysis of variance and discriminant function analysis identified combinations of MS outcomes associated with physical activity.

Results:

The sample had a mean age of 49.6 years (SD = 13.2), a 3:1 female:male ratio, and Patient Determined Disease Steps median (IQR) score of 1 (3). Multivariate analysis of variance demonstrated that MS outcome clusters were significantly associated with physical activity, namely walking (i.e., Six-Minute Walk, Timed Up and Go, and MS Walking Scale), Pillai’s Trace V = .16, F(3,180) = 11.43, η2 = .16); cognition (i.e., Symbol Digits Modalities Test, California Verbal Learning Test-Second Edition, Brief Visuospatial Memory Test-Revised), Pillai’s Trace V = .04, F(3,204) = 2.79, η2 = .04); and symptoms (i.e., fatigue, anxiety, depression, and pain), Pillai’s Trace V = .16, F(4,199) = 9.30, η2 = .16). Discriminant function analysis indicated a significant discriminant function of walking endurance and walking limitations, depression, fatigue, and processing speed was associated with physical activity subgroups.

Conclusions:

The findings identified walking endurance and limitations, processing speed, depression, and fatigue as primary correlates of physical activity differences in persons with MS. These results may inform researchers and healthcare providers to consider these factors in behavior change interventions and clinical practices for promoting adequate physical activity levels in MS.

Keywords: mobility limitation, cognitive dysfunction, symptoms, physical activity, multiple sclerosis

Introduction

Multiple sclerosis (MS) is a prevalent and debilitating immune-mediated, neurodegenerative disease of the central nervous system (Dobson & Giovannoni, 2019). There are nearly 1 million adults living with MS in the United States (U.S.) and 2.8 million worldwide (The Multiple Sclerosis International Federation [MSIF], 2020). Persons with MS experience multiple debilitating outcomes, including walking and cognitive dysfunctions as well as symptoms of pain, fatigue, depression, and anxiety (Gilmour et al., 2018). Those MS disease outcomes may have secondary influences on health behaviors, including physical activity (Lenz & Pugh, 2018).

Physical activity and exercise have yielded improvements in walking, cognition, depression, and fatigue among persons with MS (Charron et al., 2018), yet those individuals typically do not engage in sufficient physical activity for health benefits (Casey et al., 2018). For example, one study on the epidemiology of physical activity in MS reported that only 26.8% of the study sample sufficiently engaged in physical activity for health benefits (Motl et al., 2015), and persons with MS were 2.3 times less likely to engage in sufficient physical activity compared with healthy controls (Motl et al., 2015). The low level of participation in physical activity among persons with MS supports research identifying factors explaining variation in this health behavior, as these might inform the design and delivery of behavior change interventions.

Of note, the associations between MS disease outcomes, such as walking and cognitive dysfunctions, and symptoms of fatigue, pain, anxiety, and depression, with physical activity have been separately examined in multiple studies, yet those studies have had limitations and/or inconsistencies. Regarding walking outcomes, one study examining the impact of ambulatory ability on physical activity in MS and reported that participants who had lower ambulatory limitations demonstrated a higher level of physical activity (Cavanaugh et al., 2011). However, that study was conducted using a small sample size of 22 participants with MS. Regarding symptoms of depression, anxiety, fatigue, and pain, one study reported that fatigue and depressive symptoms were not significantly correlated with physical activity behavior (Beckerman et al., 2010), yet another study reported that fatigue had a negative relationship with physical activity even when adjusting for age, MS type, depression, and anxiety (Rietberg et al., 2011). Regarding cognitive functions, one study on physical activity and cognitive functions in MS reported that physical activity was associated with cognitive processing speed (pr = .35) (Motl et al., 2011). Additionally, an integrative review of physical activity and cognitive function in adults with MS reported that high levels of physical activity were associated with better cognitive performance (Morrison & Mayer, 2017); yet, this may not be interpreted as evidence for the reverse hypothesis that cognitive function may be explaining physical activity levels in MS.

There further is limited evidence regarding a comprehensive examination of MS disease outcomes as correlates of physical activity levels in persons with MS. We are aware of two studies examining the association of combinations of specific outcomes as correlates of physical activity in MS. The first study examined the associations of difficulty walking, fatigue, pain, and depression with physical activity in 133 individuals with MS (Motl et al., 2008). That study indicated that overall symptoms (r = −.50), fatigue (r = −.26), and walking difficulty (r = −.46) were inversely associated with levels of physical activity. The second study examined modifiable factors influencing physical activity in 103 participants with MS (Beckerman et al., 2010). That study indicated that disease severity, disability pension, and childcare responsibilities were significantly associated with physical activity levels in MS. Although those studies included a wide range of disease outcomes and other potential physical activity determinants, cognitive functions were still missing or included under disability status rather than stand alone as specific disease outcomes. We believe that more research regarding the associations of specific disease outcomes, including cognitive functions, would provide more robust evidence regarding factors associated with physical activity behavior in MS.

The current study investigated a comprehensive set of specific MS disease outcomes as explanatory factors for categorical levels of health-promoting physical activity (i.e., insufficiently active, sufficiently active) in persons with MS. To that end, this study examined whether there were significant combinations of disease outcomes, including walking, cognitive function, and symptoms of fatigue, depression, anxiety, and pain, that contributed to the differences in health-promoting levels of physical activity in persons with MS. It was hypothesized that a linear combination of walking, cognitive function, and symptoms of fatigue, depression, anxiety, and pain would correlate with group membership based on levels of health-promoting physical activity in persons with MS. We further hypothesized that walking capacity/impairment variables would be the strongest explanatory factors for physical activity in MS among the variables of interest, as walking is the most common type of physical activity in MS (Weikert et al., 2011).

Method

Participants

This secondary analysis was conducted using data from a cross-sectional study of physical function, cognition, and MS disease outcomes across age groups in persons with MS (Sikes et al., 2019a; Sandroff et al., 2021). The study sample was recruited via advertisements distributed by mail or electronically in the local community, the National MS Society and local chapters, MS clinics, the University Medical Center MS database, University and community collaborators, and word-of-mouth between 2017–2020. Participants contacted the research laboratory and staff by telephone or email for further information about the study. Participants were eligible to participate based on five inclusion criteria: (1) age between 20 and 79 years; (2) definite diagnosis of MS; (3) ambulatory with or without assistance; (4) relapse-free for the past 30 days; and (5) willingness to complete the testing procedures. Those who were wheelchair users were excluded from the study.

Outcome Measures

Demographic and Clinical Characteristics

Demographic and clinical characteristics were collected using questionnaires regarding age, MS types, disease duration, sex, race, marital status, and other socioeconomic information. Disability status was measured by the Patient Determined Disease Steps (PDDS). The PDDS is a self-report measure of disability with nine ordinal response options consisting of ranging between 0 (normal) and 8 (bedridden) (Hohol et al., 1999) and has been validated in persons with MS (Learmonth et al., 2013).

Physical activity levels

Physical activity levels were assessed using the Godin Leisure-Time Exercise Questionnaire (GLTEQ; Godin & Shephard, 1985) and this scale has been applied extensively in persons with MS (Sikes et al., 2019b). The applications include 8 psychometric studies, 21 studies examining rates of physical activity, 24 studies examining correlates of physical activity, 28 studies examining consequences of physical activity, and 15 physical activity intervention studies in MS. The GLTEQ requires that participants report the number of 15-minute bouts of mild, moderate, and strenuous physical activity engaging in during a typical week. The GLTEQ responses were computed into health contribution scores (GLTEQ-HCS) by multiplying the moderate and strenuous bouts by weights of 5 and 9, respectively, and summing these scores (units); a higher score represents a higher physical activity level (Godin, 2011). The GLTEQ-HCS scores yields three categories of the volume of physical activity and health outcomes, including Active (substantial benefits) (HCS ≥ 24 units), Moderately active (some benefits) (HCS of 14–23 units), and Insufficiently Active (less substantial or low benefits) (HCS < 14 units) (Godin, 2011). The GLTEQ-HCS cutoff of 14 units was used to categorize participants based on the health benefits of physical activity, namely insufficiently active (less substantial or low benefits, < 14 units) and sufficiently active (substantial and some benefits, ≥ 14 units). There is evidence supporting the validity of the GLTEQ-HCS score in MS (Motl et al., 2018).

Symptoms

Fatigue

The fatigue severity was assessed using the 9-item Fatigue Severity Scale (FSS). The items were rated on a scale from 1 = strongly disagree to 7 = strongly agree based on the severity and impact of fatigue on persons with MS in a 7-days period. The FSS scores were the mean of item scores and ranged between 1–7, with higher scores representing more fatigue (Krupp et al., 1989).

Depression and anxiety

The two sub-scales of the 14-item Hospital Anxiety and Depression scale (HADS) were used to assess anxiety and depressive symptoms. Items were rated on a scale from 0 = Not at all to 3 = Most of the time. Sub-scores were the sums of item scores and ranged between 0 and 21, with higher scores representing higher levels of anxiety and depressive symptoms (Honarmand & Feinstein, 2009).

Pain

Pain was assessed using a 2-item subscale of the 36-item Medical Outcomes Study Short-Form Health Survey (SF-36). One of the two items was rated on a five-point scale from 1 = Not at all to 5 = Extremely, and the other item was rated with a six-point scale from 1 = None to 6 = Very severe. The 2-item scores were averaged to create the composite score for bodily pain; higher scores represented less bodily pain (Ware & Sherbourne, 1992).

Cognitive functions

Processing speed

The visual information processing speed of the participants was assessed using the Symbol Digits Modalities Test (SDMT), which was validated in persons with MS (Benedict et al., 2017). Participants were instructed to say the digit (from 1 to 9) paired with the symbol according to the symbol-digit pairings in the key as quickly as possible. The outcome was the total number of correct responses in 90 seconds (Smith, 1982).

Verbal learning and memory

The California Verbal Learning Test-Second Edition (CVLT-II) was utilized to assess verbal learning and memory in this study and is validated in persons with MS (Houtchens et al., 2007). Participants were instructed to immediately recall a list of 16 words provided by the assessors. The assessment included five trials, and the number of correct responses from each trial was summed for a total score out of 80.

Visual learning and memory

The visual learning and memory of the participants were assessed using the Brief Visuospatial Memory Test-Revised (BVMT-R), which is also validated among persons with MS (Benedict, 1997). Participants were instructed to recall and draw six figures displayed on a sheet of paper presented to them for 10 seconds. The participants had three trials to draw the figures as accurately as possible and in the correct position on the page. Each figure received a score of 0, 1, or 2 based on the accuracy of the figure and the position of the figure on the page. Scores were summed across trials for a total score out of 36 (Delis et al., 2000).

Walking

Walking endurance

Walking endurance of the participants was assessed using the Six-Minute Walk (6MW) obtained by measuring the farthest distance walked by participants in six minutes (Goldman et al., 2008).

Ambulatory function and leg strength

The Timed Up and Go (TUG) test assessed ambulatory function and leg strength of participants by measuring the average time to complete two trials of TUG (Sebastião et al., 2016).

Walking limitations

The 12-item MS Walking Scale (MSWS-12) was used to assess walking limitations perceived by persons with MS. Items were rated with a fivepoint scale from 1 = Not at all to 5 = Extremely. The total scores of 12 items ranged between 12 and 60 and were then transformed into a scale from 0 to 100, with higher scores indicating higher walking limitations due to MS (Hobart et al., 2003).

Procedures

This study was not preregistered and was approved by the University Institutional Review Board. All participants provided written informed consent. This study included one 90-minute visit at a university research laboratory. During the visit, data were collected from questionnaires (i.e., demographic and clinical information, GLTEQ, HADS, FSS, SF-36, and MSWS-12), cognitive tests (i.e., SDMT, CVLT-II, and BVMT-R), and physical tests (6MW and TUG). Participants received remuneration upon completion of the study.

Data Analysis

SPSS version 27 (IBM, Corp., Armonk, NY, USA) was utilized to analyze the data. The data were presented as means and standard deviations in text and tables unless otherwise noted.

We initially examined normality using skewness and visualization of histograms. The completeness of data and outliers were examined using descriptive statistics. The potential differences between participants included in the primary analyses and those excluded due to missing data were examined using independent samples t-tests and Chi-square tests as appropriate. Participants were categorized into two subgroups based on the health benefits of physical activity reflected through the GLTEQ-HCS (i.e., insufficiently active [GLTEQ-HCS < 14 units or less substantial or low benefits] and sufficiently active [GLTEQ-HCS ≥ 14 units or substantial and some benefits]) (Godin, 2011; Motl et al., 2018). The differences in descriptive characteristics of the physical activity subgroups at baseline were examined using independent samples t-tests and Chi-square tests as appropriate.

Multivariate analysis of variance (MANOVA) was then conducted to investigate the interactions between physical activity subgroups and sex, age, and disability scores to ensure that it would be appropriate to continue with one-way MANOVA analysis using physical activity subgroup as the independent variable (Field, 2018). Then, three separate MANOVAs were performed to examine the potential combinations of dependent variables (i.e., three outcome clusters mentioned in the Outcome Measures section: symptoms, cognitive functions, and walking) associated with the group differences based on physical activity levels.

Discriminant function analysis (DFA) was then performed to identify significant predictors in each category of variables that distinguished physical activity subgroups. One final DFA examined the strongest two predictors per category of variables in the previous analyses for examining the primary contribution of variables associated with membership in physical activity subgroups.

Transparency and openness

We reported how we determined our sample size, data exclusions, and all measures in the study, and we also follow JARS (Kazak, 2018). There is not analytic code associated with this study. De-identified data from this study are not available in a public archive and will be made available (as allowable according to institutional IRB standards) by emailing the corresponding author. This study was not formally pre-registered. Data were analyzed using SPSS version 27 (IBM, Corp., Armonk, NY, USA).

Results

Participant characteristics

Two hundred thirteen participants with MS who provided consent for participation and completed the GLTEQ were included in this study. The participant characteristics were presented in Table 1. On average, the study sample was 49.6 years of age (SD = 13.2) and was diagnosed with MS for 12.9 years (SD = 9.6). The sample consisted of 75% female and 25% male. Participants were predominantly represented by 64% White/Caucasian and 32% Black/African American. Relapsing-remitting MS was a predominant MS type that accounted for 92% of the sample. The majority of participants reported mild-to-moderate disability with a PDDS median (IQR) of 1 (3). The participants mostly identified as married (60%), employed (56%), college graduate or more (56%), and having an annual household income of $40,000 or more (72%). Of note, participants were excluded from the MANOVA analyses due to incomplete questionnaires or cognitive/walking function data (See Table 2). With cognitive and symptom variables, the percentages of missing data were not significant (< 5% of the dataset); however, with walking variables, the percentages of missing data were more than 10%. There were no significant differences in demographic and clinical characteristics between participants included in the data analysis and those excluded based on missing data, except for education (p = .04).

Table 1.

Participant characteristics for the overall sample and subgroups differing in physical activity levels

GLTEQ-HCS PA Subgroups
Variable, units (n) Full Sample
n = 213		 Insufficiently Active
n = 100		 Sufficiently Active
n = 113
Age, years (213), M (SD) 49.6(13.2) 50.2(12.6) 49.0 (13.7)
MS Duration, years (213), M (SD) 12.9(9.6) 12.6(10.0) 13.2(9.3)
GLTEQ-HCS (213), M (SD) 22.4(25.5) 1.7(3.6) 40.8(22.2)
PDDS (186), Mdn (IQR) 1 (3) 2 (4) 1 (2)
Type of MS (206), n (%)
 RRMS 189 (92) 87 (91) 102 (93)
 SPMS 11 (5) 4 (4) 7 (6)
 PPMS 6 (3) 5 (5) 1 (1)
Sex (212), n (%)
 Female 160 (75) 79 (79) 81 (72)
 Male 52 (25) 20 (21) 32 (28)
Marital Status (213), n (%)
 Married 127 (60) 59 (59) 68 (60)
 Other Status 86 (40) 41 (41) 45 (40)
Employed (210), n (%)
 Yes 117 (56) 48 (49) 69 (62)
 No 93 (44) 50 (51) 43 (38)
Race (212), n (%)
 Caucasian 137 (64) 59 (59) 78 (70)
 African American 67 (32) 39 (39) 28 (25)
 Other 8 (4) 2 (2) 6 (5)
Education (213), n (%)
 Less than College 94 (44) 54 (54) 40 (35)
 College Graduate or more 119 (56) 46 (46) 73 (65)
Annual Household Income (208), n (%)
 Less than $40,000 58 (28) 33 (33) 25 (23)
 $40,000 or greater 150 (72) 66 (67) 84 (77)
Open in a new tab

Note. PA = Physical Activity; SD = Standard Deviation; GLTEQ-HCS = Godin Leisure-Time Exercise Questionnaire – Health Contribution Score; IQR= Interquartile Range; PDDS= Patient Determined Disease Steps; RRMS= Relapsing-Remitting Multiple Sclerosis; PPMS= Primary Progressive Multiple Sclerosis; SPMS= Secondary Progressive Multiple Sclerosis.

Table 2.

Results of MANOVAs for physical activity subgroups on variable groups for walking, symptoms, and cognition outcomes

Predictors Insufficiently active Sufficiently Active F
Walking
Pillai’s Trace V = .16, F(3,180) = 11.43, η2 = .16		 n = 81 n = 103
 6MW 1301.08 (395.08) 1627.76 (423.92) 28.58***
 MSWS 41.15 (29.31) 22.21 (25.09) 22.27***
 TUG 9.46 (4.28) 7.52 (4.28) 9.31*
Symptoms
Pillai’s Trace V = .16, F(4,199) = 9.30, η2 = .16		 n = 96 n = 108
 Fatigue 5.00 (1.49) 3.97 (1.55) 23.77***
 Depression 7.15 (4.26) 4.35 (3.46) 26.66***
 Anxiety 7.48 (4.98) 5.63 (3.05) 10.47***
 Pain 59.43 (26.29) 71.04 (23.37) 11.16***
Cognition
Pillai’s Trace V = .04, F(3,204) = 2.79, η2 = .04		 n = 97 n = 111
 CVLT-II 43.26 (9.12) 46.70 (11.00) 5.94*
 SDMT 47.16 (10.75) 50.96 (13.45) 4.94*
 BVMT-R 20.95 (7.21) 21.86 (6.88) 0.88
Open in a new tab

Note. 6MW = Six-Minute Walk; MSWS = Multiple Sclerosis Walking Scale; TUG = Timed Up and Go; CVLT-II = California Verbal Learning Test-Second Edition; SDMT = Symbol Digits Modalities Test; BVMT-R = Brief Visuospatial Memory Test-Revised.

***

p < .001

*

p < .05

We did not undertake an a priori power analysis for sample size estimation; however, a G-power estimation with MANOVA as the final analysis indicated that the sample size of 196 was sufficient to detect the difference between groups with a small-to-moderate effect size (f2 = 0.05), 80% power, and ɑ = .05.

Physical activity levels

The details regarding the physical activity subgroups of the participants are provided in Table 1. The mean GLTEQ-HCS of the sample was 22.4 (SD = 25.5). The study sample was categorized into two subgroups based on GLTEQ-HCS: insufficiently active (n = 100, 46.9%) with a mean score of 1.7 (SD = 3.6) and sufficiently active (n = 113, 53.1%) with a mean score of 40.8 (SD = 22.2). The physical activity subgroups based on GLTEQ-HCS were not significantly different regarding demographic and clinical characteristics, including disease duration (p = .63), MS type (p = .29), employment status (p = .06), marital status (p = .69), race (p = .26), education (p = .06), and income (p = .45).

Multivariate test for physical activity levels and sex, age, and disability interactions

MANOVA analyses were conducted to ensure sex, age, and disability were not confounding factors influencing the relationships between the variables of interest and physical activity levels in the study sample. The interactions between GLTEQ-HCS physical activity subgroups and sex, age, and disability were not significant for all three groups of variables (p > .05). These non-significant interactions between physical activity levels and sex, age, and disability indicated that it was appropriate to conduct one-way MANOVAs with only physical activity subgroups as the independent variable.

One-way MANOVA with DFA

The physical activity subgroups were compared on each of the three groups of dependent variables (i.e., walking, symptoms, and cognition) using three separate MANOVAs. The findings are presented in Table 2. Using Pillai’s Trace, there was at least a significant combination of walking variables (i.e., 6MW, TUG, MSWS) that differed by physical activity subgroups, V = .16, F(3,180) = 11.43, p < .001 with a medium effect size (η2 = .16). Regarding symptoms of depression, fatigue, anxiety, and pain, there was at least one significant combination that differed by physical activity subgroups, V = .16, F(4,199) = 9.30, p < .001, with a medium effect size (η2 = .16). Regarding cognition (i.e., SDMT, CVLT-II, and BVMT-R), using Pillai’s Trace, the analysis indicated a significant combination that differed by physical activity subgroups, V = .04, F(3,204) = 2.79, p = .04 and a small effect size (ƞ2 = .04). The univariate tests revealed significant effects of all outcome variables between physical activity subgroups, except for cognitive function measured by BVMT-R, F(1, 206) = 0.88, p = .35. According to the F ratios in Table 2, the largest differences in GLTEQ-HCS defined physical activity subgroups involved 6MW, MSWS, fatigue, depression, CVLT-II, and SDMT variables.

The MANOVAs were followed with DFA to explore the nature of the multivariate relationships between walking, cognition, symptoms and physical activity subgroups (Table 3). Regarding walking, one significant discriminant function emerged and explained 16% of the variation in physical activity subgroups with eigenvalue = 0.19, canonical R2 = 0.16, p < .001. The structure matrix of correlations between predictors demonstrated the significant function order with 6MW (.91), MSWS (−.81), and TUG (−.52) based on coefficients exceeding .30 and the structure matrix of correlations of predictors (Comrey & Lee, 1992). Pooled within-group correlations among the target variables presented all medium-large correlation coefficients rs = −.78 - .64. One significant discriminant function emerged from the symptoms of depression, anxiety, pain and fatigue explaining 16% of the variation in physical activity subgroups with eigenvalue = 0.19, canonical R2 = 0.16, p < .001. The significant function order was depression (.84), fatigue (.79), pain (−.54), and anxiety (.53). Pooled within-group correlations among the predictors presented low to medium correlations with rs = −.36 - .58. Regarding cognition, a significant discriminant function emerged from the analysis in order of CVLT-II (.84) and SDMT (.76), which accounted for 4% of the variation in physical activity levels with eigenvalue = 0.04, canonical R2 = 0.04, p < .05. Table 3 indicated that the correlations among the cognitive variables are medium with rs = .51 - .60. Overall, the Functions at Group Centroids analysis demonstrated significant discrimination between the insufficiently active and sufficiently active groups based on the three disease outcome clusters.

Table 3.

Results of DFA of walking, symptoms, and cognition outcomes for physical activity subgroup classification

Predictors Correlations of Predictor Variables with Discriminant Function Pooled Within-Group Correlations Among Predictor Variables
Walking 1 6MW MSWS TUG
 6MW .91 - −.71 −.78
 MSWS −.80 - .64
 TUG −.52 -
Symptoms 1 Depression Fatigue Pain Anxiety
 Depression .84 - .34 −.47 .58
 Fatigue .79 - −.41 .34
 Pain −.54 - −.36
 Anxiety .53 -
Cognition 1 CVLT-II SDMT BVMT-R
 CVLT-II .84 - .51 .55
 SDMT .76 - .60
 BVMT-R .32 -
Open in a new tab

Notes. Walking: Note: Canonical Correlation = 0.40; Eigenvalue = 0.19; Functions at Group Centroids: Insufficiently active (−0.49) and Sufficiently active (0.39). Symptoms: Canonical Correlation = 0.40; Eigenvalue = 0.19; Functions at Group Centroids: Insufficiently active (−0.46) and Sufficiently active (0.41). Cognition: Canonical Correlation = 0.20; Eigenvalue = 0.04; Functions at Group Centroids: Insufficiently active (−0.22) and Sufficiently active (0.19). Variables reported in the table were based on coefficients exceeding 0.30 (Comrey & Lee, 1992). DFA = Discriminant Function Analysis; 6MW = Six-Minute Walk; MSWS = Multiple Sclerosis Walking Scale; TUG = Timed Up and Go; CVLT-II = California Verbal Learning Test-Second Edition; SDMT = Symbol Digits Modalities Test; BVMT-R = Brief Visuospatial Memory Test-Revised.

An overall DFA for the strongest predictors per category of variables that emerged from the previous three analyses was then performed to further understand the nature of the multivariate relationships between walking, cognition, symptoms and physical activity subgroups (Table 4). The significant discriminate function, which emerged from this analysis, accounted for 19% of the variation in physical activity subgroups with eigenvalue = 0.24 and canonical R2 = 0.19. Based on the structure matrix of correlations between predictors and discriminant functions and coefficients exceeding .30 (Comrey & Lee, 1992), 6MW (.81), MSWS (−.66), depression (−.65), fatigue (−.61), and SDMT (.31) had the most meaningful contribution to distinguish physical activity subgroups. The group centroids indicated that the function effectively separated the insufficiently active group (−.54) from the sufficiently active group (.43).

Table 4.

Overall results of Discriminant Analysis of disease outcomes for Physical Activity Subgroup Classification

Correlations of Predictor Variables with Discriminant Function Pooled Within-Group Correlations Among Predictor Variables
Predictors 1 6MW MSWS Depression Fatigue SDMT CVLT-II
 6MW .81 - −.72 −.21 −.23 .44 .37
 MSWS −.66 - .34 .43 −.44 −.35
 Depression −.65 - .37 −.17 −.22
 Fatigue −.61 - −.14 −.01
 SDMT .31 - .51
 CVLT-II .29 -
Open in a new tab

Note. Canonical Correlation = 0.44; Eigenvalue = 0.24; Functions at Group Centroids: Insufficiently active (-0.54) and Sufficiently active (0.43). Variables reported in the table were based on coefficients exceeding 0.30 (Comrey & Lee, 1992). 6MW = Six-Minute Walk; MSWS = Multiple Sclerosis Walking Scale; CVLT-II = California Verbal Learning Test-Second Edition; SDMT = Symbol Digits Modalities Test.

Discussion

This study examined whether linear combinations of three distinct groups of disease outcomes (i.e., walking, cognition, and symptoms) were differentially associated with physical activity subgroups in persons with MS. Overall, one linear combination emerging from each group of variables was significantly associated with the physical activity subgroups. Regarding walking outcomes, walking endurance (6MW), perceived/self-report walking limitations (MSWS), and ambulatory function and leg strength (TUG) were associated with physical activity levels. Walking endurance and perceived/self-report walking limitations had the most significant associations with physical activity levels. Regarding cognition, verbal learning and memory (CVLT-II) and processing speed (SDMT) were associated with physical activity levels of participants. Regarding symptoms of depression, anxiety, pain, and fatigue, MANOVA and DFA indicated a combination of depression and fatigue that were associated with the physical activity of participants. The final DFA noted that physical activity subgroups were collectively associated with walking endurance (6MW), perceived/self-reported walking limitations (MSWS-12), depression, fatigue, and processing speed, and those variables explained 19% of the variation in physical activity levels of the study sample. Of note, a large portion of the variation in physical activity (81%) is not attributed to those disease outcome variables. Based on a review of studies examining determinants of physical activity in MS (Learmonth & Motl, 2016), much of the remaining variation in physical activity could be explained by the presence of comorbidities, demographic factors, social and environmental context, interpersonal factors such as knowledge and self-regulation skills to perform physical activity, self-efficacy, perceived barriers and benefits, enjoyment, and outcome expectations, among other factors. Such findings highlight that walking, cognition, and symptoms of depression and fatigue should be considered in future research and clinical practices for changing physical activity in persons with MS, yet more research is warranted to investigate physical activity correlates extensively, including the targeted disease outcomes and other factors noted above.

This study suggested the disease outcome regarding walking, cognition, and symptoms could separately differentiate physical activity subgroups. Participants who reported insufficient physical activity were characterized by less walking endurance, verbal learning memory, processing speed, more walking limitations, fatigue, depression, anxiety, and pain compared to individuals with sufficient physical activity. Indeed, the findings are consistent with previous evidence on physical activity correlates (Motl et al., 2008; Cavanaugh et al., 2011). Of note, walking endurance, walking limitations, processing speed, visual learning and memory, depression, and fatigue emerged as the most significant predictors of physical activity group membership within the outcome cluster. Therefore, future studies should consider those variables when developing physical activity interventions that will bring the most benefits to persons with MS. Specifically, these outcomes could be utilized in the inclusion/exclusion criteria when selecting study participants for physical activity behavior change studies in MS. Furthermore, different components of physical activity interventions or clinical practices could tailor varying levels and types of disease outcomes in MS. For example, interventionists and healthcare professionals in rehabilitation could utilize different modalities of physical activities or exercise based on different levels of fatigue, physical and cognitive functions in MS, which would, in turn, create the most meaningful progress in physical activity behavior change. Besides that, the association between depression and physical activity would be telltale evidence to urge researchers and/or healthcare providers such as psychologists or counselors in a rehabilitation team to focus more on addressing depression, among other factors, in order to maximally promote physical activity, health, participation, and quality of life in persons with MS.

This study further examined a comprehensive contribution of walking, cognition, and symptom variables to the physical activity group membership of the study sample. The final DFA indicated that walking and symptom variables had strong correlations with the discriminant function compared to cognition variables. Therefore, physical activity group membership of the study sample could be differentiated by walking (i.e., walking endurance, walking limitations), symptoms (i.e., depression, fatigue), and cognition (i.e., processing speed) in that order. To date, we are aware of a limited number of studies examining if these factors collectively and differentially correlate with physical activity in MS. Our findings reflect some levels of similarities and differences with some previous studies on physical activity correlates with MS. For example, one study identified severe levels of disability, depression, and fatigue as risk factors for being inactive in MS (Reider et al., 2017). However, self-report cognitive status did not emerge as a predictor of physical activity in that study (Reider et al., 2017). Another study examining demographic, disease-related, cognitive-behavioral, and environmental correlates of physical activity highlighted an important role of walking and depression in physical activity behavior in MS (Beckerman et al., 2010). That study reported disease-related outcomes (i.e., disability status predominantly determined by walking ability, depression, and quality of life) explained 28.3% variance in physical activity. Compared with our study, that study emphasized a stronger influence of demographic factors on physical activity (Beckerman et al., 2010). Collectively, such findings emphasized the importance of walking and symptoms of depression and fatigue in promoting physical activity behavior with MS. Yet, more research is warranted to better understand the extensive impact of these disease outcomes and demographic features, among other factors, on physical activity in MS.

We note that walking performance/limitations was the most significant contributor for explaining physical activity behavior in MS; this is consistent with other research (Motl et al, 2008; Cavanaugh et al., 2011). This could influence physical activity behavior of persons with MS, as walking is the most common physical activity type reported by this population (Weikert et al., 2011). It is challenging to modify walking ability, but we might examine correlates of physical activity across degrees of walking dysfunction in MS. This could inform future behavior change research on other potentially modifiable factors (such as health belief related factors and physical activity knowledge related to other physical activities in addition to walking) (Kasser & Kosma, 2012; CDC, 2022) that could potentially support persons with MS to overcome the barrier of walking difficulty and be more physically active.

Of note, cognitive functions (i.e., processing speed, verbal learning and memory) demonstrated a low level of association with physical activity subgroups when integrated into the discriminant function along with walking and symptoms. Previous studies documented that physical activity could positively predict cognitive performance; however, there was limited evidence on the reverse direction from cognitive function to physical activity level. For example, a study on physical activity and cognition in MS reported a significant association between physical activity and cognitive processing speed but not composite learning and memory (Motl et al., 2011). Another study in the general population indicated that high intensity of physical activity might be associated with better cognitive performance (Brown et al., 2012). On the other hand, a study on cognitive and physical disability in MS reported that visual and verbal learning and memory correlated with the physical dimension of health-related quality of life (Hoogs et al., 2011). Although physical activity was not the main outcome of the study, its relationship with cognitive functions was indirectly reflected via the physical component of quality of life in MS. This preliminary evidence emphasized the need for further research to better understand the linkage between cognition and physical activity in MS.

There are several limitations to this study. One of the limitations is the self-reported nature of the physical activity and other questionnaires measuring walking limitations, fatigue, depression, anxiety, and pain. Self-reported measurements might have a risk of reporting bias and be influenced by the fitness level of participants (Ogonowska-Slodownik et al., 2021; Shook et al., 2016). However, these questionnaires, particularly the GLTEQ, have been applied and validated in previous studies of persons with MS (Sikes et al., 2019b); see a comprehensive review of the validity and extensive use of the GLTEQ in MS research (Sikes et al., 2019b). It is important to notice that the GLTEQ itself is not influenced by social desirability biases (Motl et al., 2005). Furthermore, the causal relationship between physical activity and MS outcomes in this study is precluded by the cross-sectional study design. Therefore, further research, for example, longitudinal studies, could provide more robust evidence regarding the changes in physical activity levels and the management of the MS outcomes. The sample of this study had a mean disease duration of 12.9 years (i.e., > 10 years) and a low disability status (PDDS median = 1). This could also limit the generalization of the results among the entire population with MS. It is suggested that clinical researchers and interventionists should examine the combinations of the MS outcomes as correlates of physical activity behavior in other MS subpopulations.

Overall, this study provided a comprehensive view of the associations between MS consequences and symptoms with physical activity levels in persons with MS. The findings suggest that there were associations between MS manifestations, including walking mobility, cognitive functions, and depression and fatigue with physical activity levels in persons with MS. These findings should be considered in future research that develops and delivers interventions for promoting physical activity in persons with MS.

Supplementary Material

Supplemental Material

Impact and implications.

  • Physical activity levels are reduced in persons with MS and are explained by MS disease outcomes. This study contributes to the existing literature by examining MS disease outcomes that could comprehensively explain physical activity subgroups in persons with MS.

  • The combinations of walking, symptoms, and cognitive variables contribute to differentiating the subgroups of this health behavior in MS differentially. These preliminary findings indicate that disease outcomes of walking endurance, walking limitations, depression, fatigue, and processing speed should be considered when designing and delivering physical activity behavioral interventions for persons with MS. Rehabilitation psychologists might simultaneously targeting these outcomes for improving physical activity levels in MS.

Acknowledgments

This study was funded by the National Multiple Sclerosis Society [Grant number: CA-1708-29059], and supported, in part, by the Eunice Kennedy Shriver National Institute of Child Health & Human Development of the National Institutes of Health [Grant numbers: NIH F32HD101214 and F31HD101281]. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Footnotes

The authors have no conflicts of interest to disclose. We have full control of all data that the journal may review if requested. Materials and analysis code for this study are available by emailing the corresponding author.

References

  1. Beckerman H, De Groot V, Scholten MA, Kempen JCE, & Lankhorst GJ. (2010). Physical activity behavior of people with multiple sclerosis: Understanding how they can become more physically active. Physical Therapy, 90(7), 1001–1013. [DOI] [PubMed] [Google Scholar]
  2. Benedict RH. (1997). Brief visuospatial memory test--revised: Professional manual. Lutz (FL): Psychological Assessment Resources, Inc. [Google Scholar]
  3. Benedict RH, DeLuca J, Phillips G, LaRocca N, Hudson LD, Rudick R, & Multiple Sclerosis Outcome Assessments Consortium. (2017). Validity of the Symbol Digit Modalities Test as a cognition performance outcome measure for multiple sclerosis. Multiple Sclerosis, 23(5), 721–733. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Brown BM, Peiffer JJ, Sohrabi HR, Mondal A, Gupta VB, Rainey-Smith SR, Taddei K, Burnham S, Ellis KA, Szoeke C, Masters CL, Ames D, Rowe CC, Martins RN, & AIBL research group. (2012). Intense physical activity is associated with cognitive performance in the elderly. Translational Psychiatry, 2(11), e191. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Casey B, Coote S, Hayes S, & Gallagher S. (2018). Changing physical activity behavior in people with multiple sclerosis: A systematic review and meta-analysis. Archives of Physical Medicine and Rehabilitation, 99(10), 2059–2075. [DOI] [PubMed] [Google Scholar]
  6. Cavanaugh JT, Gappmaier VO, Dibble LE, & Gappmaier E. (2011). Ambulatory activity in individuals with multiple sclerosis. Journal of Neurologic Physical Therapy, 35(1), 26–33. [DOI] [PubMed] [Google Scholar]
  7. CDC. (2022, January 04). Physical Activity for People with Disability. https://www.cdc.gov/ncbddd/disabilityandhealth/features/physical-activity-for-all.html
  8. Charron S, McKay KA, & Tremlett H. (2018). Physical activity and disability outcomes in multiple sclerosis: A systematic review (2011–2016). Multiple Sclerosis and Related Disorders, 20, 169–177. [DOI] [PubMed] [Google Scholar]
  9. Comrey AL, & Lee HB. (1992). Interpretation and application of factor analytic results. In Comrey AL & Lee HB (Eds.), A First Course in Factor Analysis (pp. 240–262). Psychology Press. [Google Scholar]
  10. Delis DC, Kramer JH, Kaplan E, & Ober BA. (2000). California Verbal Learning Test (2 edition). Pearson. [Google Scholar]
  11. Dobson R, & Giovannoni G. (2019). Multiple sclerosis – a review. European Journal of Neurology, 26(1), 27–40. [DOI] [PubMed] [Google Scholar]
  12. Field A (2018). Discovering statistics using IBM SPSS statistics (5 edition). SAGE Publications Inc. [Google Scholar]
  13. Gilmour H, Ramage-Morin PL, & Wong SL. (2018). Multiple sclerosis: Prevalence and impact. Health Reports, 29(1), 3–8. [PubMed] [Google Scholar]
  14. Godin G, & Shephard RJ. (1985). A simple method to assess exercise behavior in the community. Canadian Journal of Applied Sport Sciences, 10(3):141–146. [PubMed] [Google Scholar]
  15. Godin G. (2011). The Godin-Shephard leisure-time physical activity questionnaire. Health & Fitness Journal of Canada, 4(1). [Google Scholar]
  16. Goldman MD, Marrie RA, & Cohen JA. (2008). Evaluation of the six-minute walk in multiple sclerosis subjects and healthy controls. Multiple Sclerosis, 14(3), 383–390. [DOI] [PubMed] [Google Scholar]
  17. Hobart JC, Riazi A, & Lamping DL. (2003). Measuring the impact of MS on walking ability: The 12-Item MS Walking Scale (MSWS-12). Neurology, 60(1), 31–36. [DOI] [PubMed] [Google Scholar]
  18. Hohol MJ, Orav EJ, & Weiner HL. (1999). Disease steps in multiple sclerosis: A longitudinal study comparing disease steps and EDSS to evaluate disease progression. Multiple Sclerosis, 5(5), 349–354. [DOI] [PubMed] [Google Scholar]
  19. Honarmand K, & Feinstein A. (2009). Validation of the Hospital Anxiety and Depression Scale for use with multiple sclerosis patients. Multiple Sclerosis, 15(12), 1518–1524. [DOI] [PubMed] [Google Scholar]
  20. Hoogs M, Kaur S, Smerbeck A, Weinstock-Guttman B, & Benedict RH. (2011). Cognition and physical disability in predicting health-related quality of life in multiple sclerosis. International Journal of MS Care, 13(2), 57–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Houtchens M, Benedict R, Killiany R, Sharma J, Jaisani Z, Singh B, Weinstock-Guttman B, Guttmann C, & Bakshi R. (2007). Thalamic atrophy and cognition in multiple sclerosis. Neurology, 69(12), 1213–1223. [DOI] [PubMed] [Google Scholar]
  22. Kasser SL, & Kosma M. (2012). Health beliefs and physical activity behavior in adults with multiple sclerosis. Disability and health journal, 5(4), 261–268. [DOI] [PubMed] [Google Scholar]
  23. Kazak AE. (2018). Editorial: Journal article reporting standards. American Psychologist, 73(1), 1–2. [DOI] [PubMed] [Google Scholar]
  24. Krupp LB, LaRocca NG, Muir-Nash J, & Steinberg AD. (1989). The fatigue severity scale: Application to patients with multiple sclerosis and systemic lupus erythematosus. Archives of Neurology, 46(10), 1121–1123. [DOI] [PubMed] [Google Scholar]
  25. Learmonth YC, Motl RW, Sandroff BM, Pula JH, & Cadavid D. (2013). Validation of patient determined disease steps (PDDS) scale scores in persons with multiple sclerosis. BMC Neurology, 13(1), 37. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Learmonth YC, & Motl RW. (2016). Physical activity and exercise training in multiple sclerosis: A review and content analysis of qualitative research identifying perceived determinants and consequences. Disability and Rehabilitation, 38(13), 1227–1242. [DOI] [PubMed] [Google Scholar]
  27. Lenz ER, & Pugh LC. (2018). Theory of unpleasant symptoms. In Smith M & Liehr P, Eds. Middle Range Theory for Nursing, 4th ed (pp. 179–214). Springer. [Google Scholar]
  28. Morrison JD,. & Mayer, L.. (2017). Physical activity and cognitive function in adults with multiple sclerosis: An integrative review. Disability and Rehabilitation, 39(19), 1909–1920. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Motl RW, McAuley E, & DiStefano C. (2005). Is social desirability associated with self-reported physical activity? Preventive Medicine, 40(6), 735–739. [DOI] [PubMed] [Google Scholar]
  30. Motl RW, Snook EM, & Schapiro RT. (2008). Symptoms and physical activity behavior in individuals with multiple sclerosis. Research in Nursing & Health, 31(5), 466–475. [DOI] [PubMed] [Google Scholar]
  31. Motl R, Gappmaier E, Nelson K, & Benedict R. (2011). Physical activity and cognitive function in multiple sclerosis. Journal of Sport & Exercise Psychology, 33(5), 734–741. [DOI] [PubMed] [Google Scholar]
  32. Motl RW, McAuley E, Sandroff BM, & Hubbard EA. (2015). Descriptive epidemiology of physical activity rates in multiple sclerosis. Acta neurologica Scandinavica, 131(6), 422–425. [DOI] [PubMed] [Google Scholar]
  33. Motl RW, Bollaert RE, & Sandroff BM. (2018). Validation of the Godin Leisure-Time Exercise Questionnaire classification coding system using accelerometry in multiple sclerosis. Rehabilitation Psychology, 63(1), 77–82. [DOI] [PubMed] [Google Scholar]
  34. Ogonowska-Slodownik A, Morgulec-Adamowicz N, Geigle PR, Kalbarczyk M, & Kosmol A. (2021). Objective and self-reported assessment of physical activity of women over 60 years old. Ageing International. [Google Scholar]
  35. Reider N, Salter AR, Cutter GR,Tyry T, & Marrie RA. (2017). Potentially modifiable factors associated with physical activity in individuals with multiple sclerosis. Research in Nursing & Health, 40(2), 143–152. [DOI] [PubMed] [Google Scholar]
  36. Rietberg MB, Van Wegen EE, Uitdehaag BM, & Kwakkel G. (2011). The association between perceived fatigue and actual level of physical activity in multiple sclerosis. Multiple Sclerosis, 17(10), 1231–1237. [DOI] [PubMed] [Google Scholar]
  37. Sandroff BM, Silveira SL, Baird JF, Huynh T, & Motl RW. (2021). Cognitive processing speed impairment does not influence the construct validity of Six-Spot Step Test performance in people with multiple sclerosis. Physical Therapy, 101(2). [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Sebastião E, Sandroff BM, Learmonth YC, & Motl RW. (2016). Validity of the Timed Up and Go test as a measure of functional mobility in persons with multiple sclerosis. Archives of Physical Medicine And Rehabilitation, 97(7), 1072–1077. [DOI] [PubMed] [Google Scholar]
  39. Shook RP, Gribben NC, Hand GA, Paluch AE, Welk GJ, Jakicic JM, Hutto B, Burgess S, & Blair SN. (2016). Subjective estimation of physical activity using the International Physical Activity Questionnaire varies by fitness level. Journal of Physical Activity and Health, 13, 79–86. [DOI] [PubMed] [Google Scholar]
  40. Sikes EM, Cederberg KL, Baird JF, Sandroff BM, & Motl RW. (2019a). Self-efficacy and walking performance across the lifespan among adults with multiple sclerosis. Neurodegenerative disease management, 9(5), 267–275. [DOI] [PubMed] [Google Scholar]
  41. Sikes EM, Richardson EV, Cederberg KJ, Sasaki JE, Sandroff BM, & Motl RW. (2019b). Use of the Godin leisure-time exercise questionnaire in multiple sclerosis research: A comprehensive narrative review. Disability and Rehabilitation, 41(11), 1243–1267. [DOI] [PubMed] [Google Scholar]
  42. Smith A. (1982). Symbol Digits Modalities Test (SDMT). Manual (Revised). Western Psychological Services. [Google Scholar]
  43. The Multiple Sclerosis International Federation (MSIF). (2020, April 24). Atlas of MS 2020, 3rd erdition. https://www.msif.org/wp-content/uploads/2020/10/Atlas-3rd-Edition-Epidemiology-report-EN-updated-30-9-20.pdf. [Google Scholar]
  44. Ware JE Jr., & Sherbourne CD. (1992). The MOS 36-item short-form health survey (SF-36). I. conceptual framework and item selection. Medical Care, 30(6), 473–483. [PubMed] [Google Scholar]
  45. Weikert M, Dlugonski D, Balantrapu S, & Motl RW. (2011). Most common types of physical activity self-selected by people with multiple sclerosis. International Journal of MS Care, 13(1), 16–20. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplemental Material
NIHMS1889186-supplement-Supplemental_Material.pdf (163.6KB, pdf)

RESOURCES