1410

ORIGINAL ARTICLE

Predicting Exercise Capacity Through Submaximal Fitness

Tests in Persons With Multiple Sclerosis
Ayse Kuspinar, MSc, Ross E. Andersen, PhD, Shang Yuan Teng, MSc, Miho Asano, MSc,
Nancy E. Mayo, PhD
ABSTRACT. Kuspinar A, Andersen RE, Teng SY, Asano ULTIPLE SCLEROSIS IS an autoimmune demyelinat-
M, Mayo NE. Predicting exercise capacity through submaxi-
mal fitness tests in persons with multiple sclerosis. Arch Phys
M 1
ing disease of the CNS. It is the leading cause of
2
neurologic disability in young adults, affecting 3 times as
Med Rehabil 2010;91:1410-7. many women as men.3 Although its exact etiology remains
unknown, evidence indicates that both genetic and environ-
Objective: To estimate, for persons with multiple sclerosis mental factors are involved with the disease onset.4
(MS), the extent to which peak oxygen consumption (VO2peak) Typically, MS is associated with multiple impairments of
can be predicted by the results on submaximal tests. muscle, sensation, coordination, and balance.4 These and the
Design: Cross-sectional study. stress of a serious diagnosis can lead to a decrease in physical
Setting: Three MS clinics in the Greater Montreal region, activity5 that will have a concomitant effect on exercise capac-
Canada. ity, leading to a vicious cycle of inactivity, deconditioning, and
Participants: A center-stratified random sample of 135 further disability.
women and 48 men was drawn (N⫽183). A subgroup of 59 The current criterion standard for testing of exercise capacity
subjects with MS, who were able to perform the step test, was is a maximal exercise test that determines VO2peak. Although
selected from this sample to complete the maximal exercise it is the most accurate measure of physical fitness, it is used
test. only in specialized laboratory settings because of its cost,
Interventions: Not applicable. equipment, and need for a trained assessor.6 Furthermore, in
Main Outcome Measure: VO2peak. persons with MS, neuromuscular symptoms such as fatigue,
Results: In this sample (mean age ⫾ SD, 39 ⫾ 9y; median pain, and weakness may hinder them from safely and success-
fully completing the test.6
Expanded Disability Status Scale⫽1.5), the mean VO2peak ⫾
Because maximal exercise testing is impractical and unfea-
SD was 27.6 ⫾ 7.3mL·kg⫺1·min⫺1. This value is considerably sible in clinical settings, several submaximal tests have been
low when compared with healthy persons, ranking below the developed as alternate means of measuring fitness. Compared
25th percentile for both men and women. In a multivariate with maximal tests, submaximal exercise tests are clinically
regression analysis, the step test and grip strength were iden- useful because they are functional, easy to administer, and
tified as the only significant predictors of VO2peak. When inexpensive, and they do not require subjects to exercise to
combined with body weight, grip strength and the step test exhaustion.
explained 74% of the variance in VO2peak. In 1981, the most comprehensive study of fitness ever un-
Conclusions: Patients with MS with a mild degree of dis- dertaken in Canada was a nationwide survey of more than
ability exhibit marked reductions in exercise capacity. Also, in 15,000 people. The data obtained from this study were com-
persons with MS, submaximal tests are good predictors of piled to produce the CPAFLA. The CPAFLA evaluates the 5
exercise capacity. These measures may be used in clinical components of fitness: aerobic fitness, body composition, flex-
settings to help assess and monitor maximum oxygen con- ibility, strength, and endurance. It is a standardized battery of
sumption and in research to evaluate the effect of exercise- submaximal tests with established norms to help interpret and
related interventions. Furthermore, they will allow people with evaluate fitness levels of individuals.7
MS to self-monitor their exercise capacity and be more actively Functional walk tests are also submaximal tests that are
engaged in taking charge of their fitness level. simple and easy to administer. McGavin and colleagues intro-
Key Words: Multiple sclerosis; Physical fitness; Rehabili-
tation. List of Abbreviations
© 2010 by the American Congress of Rehabilitation
Medicine CHAMPS Community Healthy Activities Model Program
for Seniors
CNS central nervous system
CPAFLA Canadian Physical Activity Fitness and
Lifestyle Appraisal
From the School of Physical and Occupational Therapy, Faculty of Medicine EDSS Expanded Disability Status Scale
(Kuspinar, Teng, Asano, Mayo), Department of Kinesiology and Physical Education,
EQ-5D EuroQol-5D
Faculty of Education (Andersen), McGill University; Division of Clinical Epidemi-
ology, Royal Victoria Hospital (Mayo), Montreal, QC, Canada. MAS Modified Ashworth Scale
Supported by the Canadian Institutes of Health Research (grant no. 174178). mCAFT Modified Canadian Aerobic Fitness Test
No commercial party having a direct financial interest in the results of the research MS multiple sclerosis
supporting this article has or will confer a benefit on the authors or on any organi-
6MWT six-minute walk test
zation with which the authors are associated.
Reprint requests to Ayse Kuspinar, MSc, Royal Victoria Hospital, Division of RER respiratory exchange ratio
Clinical Epidemiology, 687 Pine Ave West, Ross 4.29, Montreal, Quebec, H3A 1A1 RPE rate of perceived exertion
Canada, e-mail: ayse.kuspinar@mail.mcgill.ca. VAS visual analog scale
0003-9993/10/9109-00284$36.00/0 VO2peak peak oxygen consumption
doi:10.1016/j.apmr.2010.06.005

Arch Phys Med Rehabil Vol 91, September 2010

 EXERCISE CAPACITY IN MULTIPLE SCLEROSIS, Kuspinar
duced the twelve-minute walk test to assess functional walking
capacity in persons with chronic bronchitis.8 The distance
ambulated in 12 minutes was recorded. However, after Butland
et al9 demonstrated that similar results could be obtained in less
time, the duration of the test was reduced to 6 minutes. While
several other measures of walking capacity are available, the
6MWT is recognized as one of the most accepted and well
established tests used in clinical settings and in research.
A similar noticeable finding is that distance walked in 6
minutes is reduced in persons with MS compared with healthy
subjects. Based on an assessment of 30 subjects who were
diagnosed with MS and had moderate disability, Savci et al10
observed that the distance walked by them was significantly
reduced (P⬍.001) compared with healthy control subjects of
similar age, height, and weight. Subjects with MS ambulated
on average 380m, whereas the control group ambulated almost
double the distance at 618m. Chetta et al11 used a sample of 11
patients with MS with mild to moderate level of disability
compared with 10 healthy controls. Patients with MS ambu-
lated on average 193m less than their age-matched and sex-
matched controls. They walked on average ⫾ SD, 384⫾42m,
whereas the control group walked 577⫾56m. Recently, Gold-
man et al12 used a modified 6MWT that emphasized speed in
40 subjects with MS and 20 healthy controls. Both groups were
similar in age (mean⫽41y), height, and weight. When they
compared the group of patients with MS altogether with the
healthy subjects, they found that subjects with MS ambulated
significantly less and at a slower speed (P⬍.001) than the
control group. Hence, the authors decided to divide the patients
with MS into 3 groups according to level of their disabilities
and then compare each subgroup with healthy subjects. Pa-
tients in the mild group walked on average ⫾ SD, 603⫾48.5m,
slightly less than the control group, who walked on average
620⫾49.1m. The moderate and severe disability groups walked
on average ⫾ SD, 507⫾103m and 389⫾77.7m, respectively.
The distance walked in 6 minutes was significantly reduced
with increasing disability and lower than the control group,
indicating that the 6MWT may be a good indicator of disabil-
ity.
Experiments have demonstrated that the 6MWT correlates
with VO2peak in persons with respiratory13 and cardiac dis-
eases,14 but this association has not been assessed in persons
with MS. However, Savci10 evaluated the relationship between
the 6MWT and forced vital capacity, a measure of pulmonary
function, and found a modest correlation between the 2
(r⫽.36).
The hypothesis was that by combining statistical data from
the series of submaximal tests, an equation that will estimate
the fitness level of patients with MS can be formulated. This
equation may be used in clinical settings to help assess and
monitor exercise capacity, as well as in research to evaluate the
effect of exercise related interventions. Furthermore, it will
allow people with MS to self-monitor their exercise capacity
and be more actively engaged in taking charge of their fitness
level.
As such, the purpose of this study was to estimate, for
persons with MS, the extent to which maximal exercise capac-
ity can be predicted by the results on submaximal tests.
METHODS
Participants
This cross-sectional study was incorporated in a pilot study
of the life-impact of the New MS. The available study popu-
lation consisted of both men and women who had been regis-
tered after 1994 at the MS clinics of the Montreal Neurological
1411
Hospital, Centre Hospitalier de l’Université de Montréal, and
Clinique Neuro Rive-Sud. A center-stratified random sample
was drawn, and the sample was composed of 135 women and
48 men. A subsample of 21 men and 39 women who were able
to perform the step test and who did not have a history of
cardiac, pulmonary, orthopedic, or other impairments that
would prevent participation in exercise testing was selected
from the pilot study sample. Subjects who had severe cognitive
impairments or a relapse in the preceding month were excluded
from participating in this cross-sectional study.
Procedure
The ethics committees of each participating hospital ap-
proved the study. Potential subjects were informed of the study
by letter from the physician in charge of the MS clinic. A
research coordinator subsequently contacted the subject to as-
certain the subject’s interest in participating. For those agree-
ing, a questionnaire package with the consent form was mailed,
and an appointment for the evaluation was arranged. All sub-
jects attended 2 successive evaluation sessions, and a consent
form was signed prior to the initial evaluation. At the first
session, the submaximal fitness tests were carried out, and at
the second session, the maximal exercise test was performed
with repeat testing of the push-ups, partial curl-ups, grip test,
and jump test in order to ensure that there were no important
changes in patients’ physical status. Subjects also rated their
perceived health at both evaluations using the feeling thermom-
eter of the EQ-5D to confirm that patients’ perceived health
status did not change. Subjects were instructed to indicate their
health condition on a scale from 0 to 100, the former showing
the worst imaginable health state, the latter marking the best
imaginable health state.
Measurement
Subject characteristics. Personal factors such as age, sex,
height, weight, and smoking status were recorded on the day of
testing. Duration of the disease, type of MS, and patient’s score
on the EDSS were determined from medical charts and re-
corded at the last medical visit. Balance was assessed using a
MS-specific balance scale called Equi-Scale, created using
Rasch modeling from the Tinetti Performance Oriented Bal-
ance Scale and the Berg Balance Scale.15 Spasticity of the
lower and upper limbs was assessed using the MAS. Physical
activity was evaluated using the modified CHAMPS Physical
Activity Questionnaire. Because it captures activity including
quiet recreation, socialization, and physical activity, it avoids
social desirability bias in people who may not do a lot of
physically demanding activity. For people with disability, it is
closer to the concept of participation. Fatigue and muscle pain
were measured with a 10-cm VAS with anchors of no fatigue
to severe fatigue and no pain to worst pain imaginable.
Exercise capacity. VO2peak was determined using an in-
cremental graded cycle ergometer test.a The person was prop-
erly seated on the bicycle with an upright posture, hands
positioned on the handlebars, and seat adjusted for adequate
knee extension. Heart rate was recorded each minute. Oxygen
uptake and RER were continuously measured. Perceived exer-
tion was evaluated using the Borg RPE 6 –20 scale. All persons
started the test at a minimal workload of 10W with a gradual
increase of 10W a minute. They were instructed to pedal at a
constant frequency of 60rpm throughout the test. The test was
terminated when the subject was not capable of maintaining a
pedaling frequency of at least 45rpm. VO2peak was expressed
as an absolute value (L/min) or relative to body weight
(mL · kg–1 · min–1).
Arch Phys Med Rehabil Vol 91, September 2010
1412 EXERCISE CAPACITY IN MULTIPLE SCLEROSIS, Kuspinar
The Modified Canadian Aerobic Fitness Test. This is a
multistage step test that assesses submaximal exercise capacity.
Standardized instructions and procedures as outlined in the
CPAFLA manual were used. Subjects were asked to perform a
series of stepping sequences on a double 20.3-cm step in time
with a musical cadence.16-18 Starting stage or stepping se-
quence was determined according to the patient’s age. Each
stage lasted 3 minutes and became progressively more chal-
lenging. Subjects completed all the stages necessary to achieve
85% of their age-predicted maximum heart rate.19 An oxygen
uptake value for the final stepping stage was taken from the
Canadian Standardized Test of Fitness Operations Manual.20
The mCAFT has been shown to have a high degree of reliabil-
ity21 and validity for predicting exercise capacity.16 The pub-
lished regression equation by Weller et al16 has an r2 of .83 for
a sample of 153 participants.
The Six-Minute Walk Test. The 6MWT was performed in
an enclosed 10-m to 15-m corridor. Individuals were instructed
to walk as far as possible in 6 minutes at their own pace. They
could rest at any time during the test but were encouraged to
resume walking as soon as they were ready to do so. The
number and duration of rests as well as the total distance
ambulated were recorded. Standardized instructions and en-
couragements were used.22 The 6MWT has high reliability and
moderate validity.23
Vertical Jump Test. The vertical jump test, also taken from
the CPAFLA, was administered to evaluate lower extremity
power. In order to measure initial reach height, while the feet
are flat on the floor, subjects reached as high as possible with
the fingers and elbow of the dominant hand fully extended. The
participants then jumped as high as possible, touching the wall
at the peak height of their jump. The vertical jump height was
determined by subtracting the initial reach height from the jump
height. The maximum height jumped out of 3 trials was recorded
in centimeters. Peak leg power was then estimated from height
jumped through the equation reported by Sayers et al.24
Partial curl-ups. The partial curl-ups are also a test of
muscle endurance. During this test, subjects were instructed to
execute as many consecutive curl-ups as possible at a rate of
25/min for a maximum of 1 minute, following the cadence
provided on a metronome. From a supine position with knees
bent at 90o, subjects were asked to slowly curl up the upper
spine until the middle finger tips of both hands have reached
the 10-cm mark on the mat, and then to return slowly to the
mat.19,24,25 The guidelines set out by the CPAFLA were used.
Push-ups. The push-ups are a test of muscle endurance
that evaluates an individual’s ability to perform repetitive con-
tractions over time.19 Subjects were instructed to perform as
many consecutive push-ups to fatigue without any time limit.
For anchor points, men used their toes, women their knees. The
test was terminated when subjects were seen to strain forcibly
or were using compensatory techniques.25 Guidelines were
implemented according to the CPAFLA manual. Norm-refer-
enced age-specific and sex-specific averages are available for
this test.
Grip Strength. Grip strength was measured using the Jamar
dynamometer.c Standardized instructions and positioning were
used, and 3 consecutive trials for each hand were recorded.26,27
Grip strength has excellent reliability26,28-30 with a measure-
ment error of approximately 2kg.31
Data Analysis
The primary outcome measure was VO2peak from the graded
cycle ergometer test. The predictor variables were the mCAFT,
the 6MWT, grip strength, push-ups, partial curl-ups, vertical
jump height, age, height, and weight.
Arch Phys Med Rehabil Vol 91, September 2010
Descriptive statistics were used to characterize the partici-
pants and verify the distribution of variables. Patients’ charac-
teristics and maximal and submaximal test scores were sum-
marized using the mean ⫾ SD.
Spearman and Pearson correlation coefficients were used for
categoric and continuous variables, respectively. Unlike the
Pearson correlation, the Spearman correlation is rank-ordered
and distribution-free. Therefore, the Spearman correlation
rather than Pearson was used for the push-ups and the curl-ups,
because they were not normally distributed.
The Student t test was used to compare results between the
first and second evaluation on the push-ups, curl-ups, grip
strength, and jump test. Muscle pain, fatigue, and perceived
health status were also reassessed during the second session to
ensure that patients’ physical status had not changed since the
initial visit. Proportions of people who performed the same,
better, or worse on the second evaluation compared with the
first were calculated.
The potential for confounding from sex, muscle pain, fa-
tigue, balance, disease severity, and spasticity was tested using
a simple linear regression. Each potential confounding variable
was assessed for its relationship with the outcome (VO2peak)
and with each of the predictor variables (age, height, weight,
6MWT, mCAFT, grip strength, push-ups, curl-ups, jump
height). Those variables found to have a significant effect on
both the outcome and one of the explanatory variables were
added to the final predictive model.
Forward stepwise multiple linear regression was performed
to identify the best predictive model of exercise capacity.
Because the mCAFT is a predictor of VO2peak,16-18,32 simple
linear regression of the outcome with the mCAFT was per-
formed first. After that, the predictor variable that most in-
creased the r2 was selected. This procedure was repeated until
no additional variability in the outcome was explained.
Regression diagnostics, including the Shapiro-Wilk statistics
and residual-by-predicted plots, were generated to verify the
assumptions of normality, homoscedasticity, and linearity. Re-
gression diagnostics were also used to check for collinearity
and outliers.
All statistical analysis was carried out using the Statistical
Analysis Systems Version 9.1.b
RESULTS
Patients
Initially, a total of 60 subjects with MS were evaluated.
However, because of technical problems with calibration of the
cycle ergometer, the results of 1 subject were distorted, which
left us with 59 subjects to be analyzed. Patient characteristics
are presented in table 1. The average age was 39 years, 66%
were women, and the median EDSS score was 1.5. Overall,
58% of our sample was receiving immunomodulatory therapy
at the time of the study. Furthermore, 10% of the sample was
previously on therapy, and one quarter was never treated with
disease-modifying agents. According to the CHAMPS Physical
Activity Questionnaire, 28% of our sample was sedentary, and
only 12% performed any type of vigorous activity during a
typical week.
Table 2 shows the results on the tests of exercise capacity, first
for the submaximal tests, the 6MWT, and the step test, and then
for the graded maximal test. Distance walked during the 6MWT
was 569m, and estimated oxygen uptake for the mCAFT was
1.6L/min. The mean VO2peak was 27.6 mL·kg⫺1·min⫺1 or 1.9L/
min. At peak exercise, the mean heart rate was 170 beats per
minute and the RER was 1.2. The average RPE on the Borg
scale was 19.
 EXERCISE CAPACITY IN MULTIPLE SCLEROSIS, Kuspinar 1413

Table 1: Demographic and Clinical Characteristics of Table 3: Performance of Participants on Fitness Test Compared
Study Participants With Healthy Persons
Participants Measure Mean Percentile ⫾ SD
Characteristic (n⫽59)
6MWT 81⫾12
Age (y), mean ⫾ SD 39⫾8.8 Curl-ups 61⫾38
Sex, no. (%) Grip strength 49⫾32
Male 20 (34) mCAFT 45⫾24
Female 39 (66) Jump test 29⫾21
Type of MS,* no. (%) VO2peak 23⫾24
Definite MS 55 (93) Push-ups 16⫾19
Clinically isolated syndrome 4 (7)
EDSS score,* median (range) 1.5 (0–3.5)
Height in cm,† mean ⫾ SD 169.4⫾9.1
Weight in kg,† mean ⫾ SD 71.2⫾15.5
BMI in kg/m2, mean ⫾ SD 24.8⫾4.8 SD between the 2 evaluations was 16⫾7. The average values
Current smoking status,† no. (%) for these tests are presented for the 2 time points and compared
Yes 15 (25) using a paired t test. Jump height was significantly lower;
No 44 (75)
push-ups and grip were statistically higher. The proportions of
MS treatment, no. (%)
people who stayed the same, improved, and declined are also
Currently on DMT 34 (58)
presented. There was no consistent trend for improvement on
No longer on DMT 6 (10)
Never on DMT 15 (25)
the second evaluation because for 2 of the tests, the proportion
No information 4 (7)
of people who declined was greater than the proportion who
Physical activity level, no. (%) improved (jump height and curl-ups).
Sedentary (⬍3.3 METS) 17 (29)
Moderate (3.3 to ⬍7 METS) 35 (59) Potential Predictor Variables to VO2peak
Vigorous (ⱖ7 METS) 7 (12) The relationships among sex, pain, fatigue, present smoking
status, disease severity, and spasticity were assessed using a
NOTE. EDSS rating of 4.0 indicates need for a cane for ambulation.
Abbreviations: BMI, body mass index; DMT, disease-modifying ther- simple linear regression with the results of the submaximal
apy. tests. Current smoking status, treated as a binary variable
*Obtained from medical charts.
†
(currently smoking or not), was found to be associated with the
Self-reported by participant. jump test (P⫽.04), explaining approximately 7% of the vari-
ance in jump height, but was not associated with VO2peak.
Spasticity, measured by the MAS, was treated as a binary
Table 3 illustrates how the participants performed on fitness variable (presence or absence of spasticity). It explained 13%
tests in comparison with a normative sample obtained from the of the variability in the 6MWT (P⫽.006). Sex was significantly
Canadian Standardized Test of Fitness Operations Manual.20 associated with most of the predictor variables and the out-
For the 6MWT, because there are no norms, percent walked of come. A confounding variable must be associated with both the
age-predicted distance using an already developed multiple explanatory variable (here, submaximal tests) and the outcome
linear regression equation was calculated.33 As shown in table (here, VO2peak) but must not be in the causal pathway. There-
3, subjects ambulated at 80% of their age-predicted and sex- fore, sex was only included in the multiple linear regression
predicted distance. For the curl-ups and grip strength, subjects model as a confounding variable.
were at the 61st and the 49thh percentiles, respectively. As for
the jump test and the push-ups, they ranked, when compared Multiple Linear Regression Analyses
with healthy persons, below the 30th percentile.
Table 5 displays the results of the multiple linear regressions
Comparison of Health Status and Physical Performance developing the predictive model for absolute VO2peak from the
Across Time submaximal tests. Age, height, weight, and sex were also
included in the model. When all of these variables were mod-
Table 4 presents the values on the push-ups, curl-ups, grip
eled together, the mCAFT, grip strength, and body weight
strength, and jump test at both the initial and repeat evaluations
emerged as significant predictors. When sex was controlled for,
prior to the graded bicycle test. The average number of days ⫾
these 3 variables explained 74% (P⬍.001) of the variability in
absolute VO2peak (L/min).
Table 2: Exercise Capacity of Study Participants (nⴝ59)
In table 6, the same group of variables, excluding body
weight, was modeled together with relative VO2peak (mL·
Variable Mean ⫾ SD kg⫺1·min⫺1 ) as the outcome. This time, when sex was con-
6MWT (m) 569⫾89.7 trolled for, the mCAFT and the 6MWT surfaced as significant
Step test (L/min) 1.6⫾0.3 predictors, explaining 55% (P⬍.001) of the variability. The
VO2peak (L/min) 1.9⫾0.6 standardized parameter estimates, unstandardized parameter
VO2peak (mL·kg⫺1·min⫺1) 27.6⫾7.3 estimates, cumulative r2 by step, and their P values are pre-
Maximal heart rate (beats/min) 170.9⫾16.2 sented in the tables. Applying the published regression equa-
Estimated peak heart rate (220 – age) 180.5⫾8.7 tion from Weller et al16 to our data yielded an r2 of .38. The
Peak RPE (Borg 6–20) 19.3⫾1.0 difference between the published r2 of .8316 for a healthy
Peak RER 1.2⫾0.1 sample and our sample (.38) yields a shrinkage value of .45,
much greater than the threshold value of 0.1 for intersample
NOTE. Results were obtained using cycle ergometry. reliability.34

Arch Phys Med Rehabil Vol 91, September 2010

1414 EXERCISE CAPACITY IN MULTIPLE SCLEROSIS, Kuspinar

Table 4: Fitness Test Scores, Fatigue, and Perceived Health Status at First and Second Evaluation
Participant Participant Participant
Variable Mean ⫾ SD Time 1 Mean ⫾ SD Time 2 Stayed Same (%) Improved (%) Declined (%)

Jump height (cm) 26.1⫾10.2 25.0⫾10.3* 22 23 55

Curl-ups (no.) 15.9⫾10.7 15.6⫾10.7 73 10 17
Push-ups (no.) 5.7⫾6.6 6.5⫾6.9* 53 33 13
Grip strength (kg) 75.7⫾24.3 78.5⫾24.5* 56 34 10
Fatigue (0–100) 2.5⫾2.3 1.9⫾2.1 73 18 8
Health today (0–100) 80.3⫾12.6 82.0⫾15.1 88 8 3

*P⬍.05.

DISCUSSION patients with MS and mild disability (mean EDSS⫽2.75;

The average VO2peak ⫾ SD of this sample was 27.6⫾
7.3mL·kg⫺1·min⫺1 . VO2peak was successfully reached for
this sample, because the mean RER was 1.2 and the mean
maximum heart rate was 170 beats per minute, which is
approximately 94% of the age-predicted maximum heart
rate. This value is low compared with healthy individuals,
ranking below the 25th percentile for both men and women.7
Interestingly, despite having mild levels of disability (me-
dian EDSS score⫽1.5), this sample demonstrated marked
reductions in exercise capacity.
Bjarnadottir et al35 found similar results in their study on
the effects of moderate exercise in patients with mild MS
with an average EDSS of 2.0. They conducted a randomized
controlled trial with a control group of 10 subjects with MS
and an intervention group of 6 subjects with MS. Their mean
recordings at baseline were 27.3 mL·kg⫺1·min⫺1 for the con-
trol group and 23.4mL·kg⫺1·min⫺1 for the intervention group.
Although they studied a small sample of patients, their findings
are very similar to those of ours. Bjarnadottir35 also evaluated
exercise capacity for men and women separately. For their
male participants, the average VO2peak was 33mL·kg⫺1·min⫺1,
and for their female participants it was 21 mL·kg⫺1·min⫺1.
Likewise, in our study, the mean VO2peak ⫾ SD for the male
participants was 32⫾6.4 mL·kg⫺1·min⫺1, and for the female
participants it was 25.7⫾6.6mL·kg⫺1·min⫺1.
Romberg et al36 studied a larger group of patients in a
cross-sectional study on the extent to which exercise capacity
can be predicted by disability and physical activity. Their
sample consisted of 34 men and 61 women with average EDSS
scores of 3 and 2.2, respectively. The average VO2peak value ⫾
SD for men was 27⫾5.2 mL·kg⫺1·min⫺1, and for women it
was 21.7⫾5.5 mL·kg⫺1·min⫺1. It is not surprising that their
findings were lower than ours, because their sample consisted
of patients with slightly greater levels of disability.
In 2008, Morrison et al37 evaluated perceived exertion in
patients with MS compared with healthy controls while
undergoing a graded maximal exercise test. A total of 12
range, 0 –3) and average age ⫾ SD, 38.3⫾4.9y were in-
cluded in the study. The average VO2peak ⫾ SD recorded
during a graded maximal cycle ergometer test was
22.9⫾6.2mL·kg⫺1·min⫺1 for this group of patients. A total
of 12 age-matched and sex-matched sedentary controls was
used as a comparison group, with a mean VO2peak ⫾ SD of
25.7⫾5.3mL·kg⫺1·min⫺1 The results of their control group
were almost the same as those of the MS group because the
authors specifically selected sedentary individuals with low
levels of physical activity. Our sample of patients with MS
obtained similar VO2peak results as their MS group and
sedentary controls, confirming that patients with MS indeed
have very low levels of exercise capacity or equal to that of
highly sedentary individuals. Compared with norms7 that
were measured using a treadmill, they would in fact rank, as
mentioned, below the 25th percentile. However, in concor-
dance with previous studies in MS, a cycle ergometer was
used to determine VO2peak, a test that is relatively safer and
easier to administer because the subjects are in a seated
position.
This exceptionally low VO2peak value found in MS brings
forth the question why this is so. First, only 12% of our patient
population reported performing any type of vigorous physical
activity on a weekly basis. Moreover, 28% of the sample
reported performing no moderate or vigorous activity at all.
Finally, almost half our sample would be classified as over-
weight or obese (body mass index⬎24.9m/kg2), and it is well
known that adiposity is inversely associated with aerobic ca-
pacity.
Although our patients had mild disability levels, they still
demonstrated significantly reduced exercise capacity. Further-
more, they also demonstrated a considerable decrease in mus-
cle strength and endurance. As far as the push-ups and the jump
test were concerned, they ranked, compared with norms, below
the 30th percentile. They had slightly better results for grip
strength and curl-ups, ranking at the 50th and 60th percentiles,
respectively. According to the EDSS classification system, our

Table 5: Multiple Linear Regression Model for Absolute VO2peak

Unstandardized Standardized Cumulative
Coefficients Coefficients R2 by Step P

B SE

Step test (L/min) 1.091 .180 .31 .64 ⬍.001

Grip strength (kg) 0.009 .003 .22 .72 .004
Body weight (kg) 0.009 .004 .14 .74 .02
Sex 0.067 .180 NA NA .71

NOTE. Outcome variable: VO2peak in L/min. Total R2⫽.74; P⬍.001. Intercept is ⫺1.176; SE⫽.589; P⫽.05. Standardized coefficient ⫽ ß ⫻ 1 SD;
used to standardize the measurement scale of the different variables.
Abbreviation: NA, not applicable.

Arch Phys Med Rehabil Vol 91, September 2010

 EXERCISE CAPACITY IN MULTIPLE SCLEROSIS, Kuspinar
Table 6: Multiple Linear Regression Model for Relative VO2peak
Unstandardized
Coefficients
B SE
Step test (L/min) 16.3 3.24
6MWT (m) 0.02 0.008
Sex 1.78 1.95
NOTE. Outcome variable: VO2peak in mL·kg⫺1·min⫺1 . Total R2⫽.55; P⬍.001. Intercept is ⫺11.83; SE⫽6.41; P⫽.734. Standardized coefficient ⫽
ß ⫻ 1 SD; used to standardize the measurement scale of the different variables.
Abbreviation: NA, not applicable.
sample would be categorized as fully ambulatory with minimal
neurologic signs. This raises an issue with the EDSS as a
comprehensive indicator of MS severity because the partici-
pants, even with minimal disability, had very poor exercise
capacity. As mentioned in the review by Dalgas et al 38 these
reductions in physical capacity may be a result of the disease
process per se (changes in the CNS) or the increased physical
inactivity levels found in these patients.
Functional walking capacity, measured by the 6MWT cor-
related weakly with absolute VO2peak (r⫽.25) and relative
VO2peak (r⫽.48). The correlation of this submaximal test
against VO2peak has been evaluated for different diseases such
as cardiac and pulmonary conditions. In persons with end-stage
lung disease or pulmonary hypertension, the coefficients are
.7339 and .7040, respectively. Also, in persons with chronic
cardiac failure, this value is .88.14 However, other studies, like
ours, failed to demonstrate strong concurrent validity of the test
against VO2peak. Weak correlation coefficients of .37 and .48
were reported for persons with intermittent claudication41 and
for persons with congestive heart failure,42 respectively. Our
findings would suggest that the 6MWT is a measure of func-
tional walking capacity as indicated by Eng et al43 for stroke
rather than being a measure of exercise capacity. Then again,
the lack of correlation between the 6MWT and VO2peak may
have been found because a self-paced 6MWT was used rather
than the modified 6MWT that maximizes effort and speed, as
was validated by Goldman et al.12 The modified 6MWT, in-
stead of the self-paced 6MWT, may have been a better measure
of walking capacity because our subjects had mild disability
and were able to complete the test without need for rest.
Unfortunately, at the time this study commenced in 2007, the
authors were not yet aware of the modified 6MWT, because it
was published in 2008. However, the distance ambulated at 2
and 6 minutes was recorded in this sample, and subjects’
walking speed at the 2 time points was found to be identical. At
2 minutes, their average speed ⫾ SD was 1.6 ⫾ .24 m/s, and
at 6 minutes it was 1.6 ⫾ .26 m/s. Had the modified test of
Goldman12 been used, a decline in gait speed between the 2
time points may have been observed.
Because of a concern that the submaximal and maximal
exercise tests could lead to excessive fatigue and an exacerba-
tion of symptoms, the assessments were conducted on different
days. The submaximal fitness tests (grip strength, push-ups,
curl-ups, jump test) were administered a second time before
maximal exercise testing to ensure that there were no changes
in patients’ physical status compared with the initial evalua-
tion. Table 4 shows that there were no important average
differences on the fitness test scores between the 2 evaluations,
although some differences reached statistical significance.
There was improvement and deterioration on each of the mea-
sures, suggesting that differences were from random variation
rather than learning. For example, for the push-ups, 33% of the
1415
Standardized Cumulative
Coefficients R2 by Step P
4.89 .49 ⬍.001
1.79 .55 .008
NA NA .367
sample performed at least 1 extra push-up during the second
visit, and 13% performed at least 1 push-up less than the initial
visit. As for grip strength, taking into consideration measure-
ment error, 34% of the sample demonstrated improvements in
grip strength, and 10% showed deterioration on the second day
of testing. A better measure to confirm that participants did not
have a relapse was the EQ-5D VAS of perceived health status.
Subjects were asked to provide a global rating of their current
health status from 0 to 100, where 0 was marked by the worst
imaginable health state and 100 as the best imaginable health
state. Because there were no significant differences in patients’
ratings of their health status between the 2 sessions, and 88%
of patients rated their health status similar to their initial visit,
the data from the 2 different testing days can be combined.
For the step test, there were 7 participants who were not able
to complete all of the stages that were necessary to attain 85%
of their age-predicted maximum heart rate. Although they all
performed at least 1 stage of the test, because of problems with
coordination and motor fatigue of the lower extremities, they
could not continue further. The multiple linear regression anal-
ysis was carried out first with these patients included in the
model, and then excluded from the model. Interestingly, the
same amount of variance was explained with them excluded
from the model, indicating that the step test was still a good
predictor of VO2peak, even when patients were not able to
reach 85% of the maximum capacity.
Two models of multiple linear regression were performed,
one with relative VO2peak and another with absolute VO2peak.
For VO2peak relative to body weight, the mCAFT alone ex-
plained 48% of the variability in the outcome. The 6MWT
surfaced as the only predictor variable, when combined with
the mCAFT, that explained the most additonal variability in
VO2peak. Together, the mCAFT and the 6MWT described 55%
of the variance in the outcome.
In the multiple linear regression analysis with absolute
VO2peak, the mCAFT, grip strength, and body weight emerged
as the only significant predictors in the model. Altogether,
these 3 variables explained 74% of the variability in VO2peak
(L/min). The need for a new regression equation to predict
VO2peak in people with MS is demonstrated by the low inter-
sample reliability between healthy subjects and our sample
(shrinkage⫽.45).34
Interestingly, among all the fitness tests that were performed
in this study, in addition to the mCAFT, grip strength was the
only submaximal test that was able to explain any further
variability in VO2peak. In the predictive model in table 5, the
standardized beta, which puts the different variables on equiv-
alent measurement scales, should be used to infer relative
strength of a predictor. The standardized beta for grip strength
indicates that people with a 1 SD difference in grip strength
differ by 0.22L/min (adjusted for sex), while the regression
coefficent for 1 SD of the step test was .33. Therefore, grip
Arch Phys Med Rehabil Vol 91, September 2010
1416 EXERCISE CAPACITY IN MULTIPLE SCLEROSIS, Kuspinar
strength was not stronger than the mCAFT, but was neverthe-
less important.
Grip strength is not only a convenient and simple measure of
upper limb function but also a good indicator of overall health
and strength. As has been shown, it can differentiate between
high and low levels of health and detect persons at risk for
developing physical disability.44,45 Moreover, several studies
have shown that poor grip strength is a predictor of all causes
of mortality in middle-age and elderly individuals.46-48 There-
fore, based on the previous studies that have proven grip
strength to be an indicator of overall strength, physical func-
tion, and mortality, it is not suprising that it is associated with
VO2peak, especially because VO2peak itself is an indicator of
physical disbility and mortality. Undoubtedly, the push ups,
curl-ups, and jump test were also regarded as important mea-
sures of fitness, but when modeled together with grip strength
and the mCAFT, they failed to explain any additional variabil-
ity in VO2peak.
Study Limitations
Because this study was cross-sectional in design, subjects
were assessed at 1 point in time, and hence the data may not
necessarily reflect all time points. Also, the EDSS scores were
not recorded on the day of testing; instead, they were taken
from subjects’ medical charts during the last medical visit.
Furthermore, our inference is to those people who can perform
the fitness tests; the study of exercise capacity for the larger
sample of 183 would probably be lower, because 25% of the
entire sample could not do the mCAFT. Moreover, VO2peak
values obtained from cycle ergometry were compared against
normative VO2peak values obtained from treadmill testing.
This might not be a valid comparison, and the extent to which
VO2peak is reduced in this sample may have been overesti-
mated. However, in line with prior studies in MS, a cycle
ergometer was used to measure VO2peak because this test is
generally safer and easier to administer.
CONCLUSIONS
The mCAFT, grip strength, 6MWT, and body weight are
strong predictors of exercise capacity. These measures may be
used in clinical settings to help assess and monitor exercise
capacity, and in research to evaluate the effects of exercise-
related interventions.
Acknowledgments: We thank the New MS investigative team:
Nancy Mayo, PhD, Pierre Duquette, MD, Francois Grand’Maison, MD,
Yves Lapierre, MD, Lisa Koski, PhD, and Laury Chamelian, MD.
References
1. Noseworthy JH, Lucchinetti C, Rodriguez M, Weinshenker BG.
Multiple sclerosis. N Engl J Med 2000;343:938-52.
2. O’Connor P. Key issues in the diagnosis and treatment of multiple
sclerosis: an overview. Neurology 2009;59:S1-33.
3. Orton SM, Herrera BM, Yee IM Valdar W, Ramagopalan SV,
Sadovnick AD, Ebers GC. Sex ratio of multiple sclerosis in
Canada: a longitudinal study. Lancet Neurol 2006;5:932-6.
4. White LJ, Dressendorfer RH. Exercise and multiple sclerosis.
Sports Med 2004;34:1077-100.
5. Motl RW, McAuley E, Snook EM. Physical activity and multiple
sclerosis: a meta-analysis. Mult Scler 2005;11:459-63.
6. Noonan V, Dean E. Submaximal exercise testing: clinical appli-
cation and interpretation. Phys Ther 2000;80:782-807.
7. Niemen DC. Exercise testing and prescription: a health-related
approach. 6th ed. New York: McGraw-Hill Higher Education;
2007.
Arch Phys Med Rehabil Vol 91, September 2010
8. McGavin CR, Gupta SP, McHardy GJ. Twelve-minute walking
test for assessing disability in chronic bronchitis. BMJ 1976;1:
822-3.
9. Butland RJ, Pang J, Gross ER, Woodcock AA, Geddes DM. Two-,
six-, and 12-minute walking tests in respiratory disease. Br Med J
(Clin Res Ed) 1982;284:1607-8.
10. Savci S, Inal-Ince D, Arikan H, et al. Six-minute walk distance as
a measure of functional exercise capacity in multiple sclerosis.
Disabil Rehabil 2005;27:1365-71.
11. Chetta A, Rampello A, Marangio E, et al. Cardiorespiratory re-
sponse to walk in multiple sclerosis patients. Respir Med 2004;
98:522-9.
12. Goldman MD, Marrie RA, Cohen JA. Evaluation of the six-
minute walk in multiple sclerosis subjects and healthy controls.
Mult Scler 2008;14:383-90.
13. Rejeski WJ, Foley KO, Woodard CM, Zaccaro DJ, Berry MJ.
Evaluating and understanding performance testing in COPD pa-
tients. J Cardiopulm Rehabil 2000;20:79-88.
14. Riley M, McParland J, Stanford CF, Nicholls DP. Oxygen con-
sumption during corridor walk testing in chronic cardiac failure.
Eur Heart J 1992;13:789-93.
15. Tesio L, Perucca L, Franchignoni FP, Battaglia MA. A short
measure of balance in multiple sclerosis: validation through Rasch
analysis. Funct Neurol 1997;12:255-65.
16. Weller IM, Thomas SG, Gledhill N, Paterson D, Quinney A. A
study to validate the modified Canadian Aerobic Fitness Test. Can
J Appl Physiol 1994;20:211-21.
17. Weller IM, Thomas SG, Corey PN, Cox MH. Prediction of max-
imal oxygen uptake from a modified Canadian Aerobic Fitness
Test. Can J Appl Physiol 1993;18:175-88.
18. Jette M, Campbell J, Mongeon J, Routhier R. The Canadian Home
Fitness Test as a predictor for aerobic capacity. Can Med Assoc J
1976;114:680-2.
19. CSEP Health & Fitness Program’s Health-Related Appraisal &
Counselling Strategy. The Canadian Physical Activity, Fitness and
Lifestyle Approach. 3rd ed. Ontario: Canadian Society for Exer-
cise Physiology; 2004.
20. Canadian Standardized Test of Fitness operations manual. 3rd ed.
Ottawa: Fitness Canada; 1986.
21. Weller IM, Corey PN. A study of the reliability of the Canada
Fitness Survey questionnaire. Med Sci Sports Exerc 1998;30:
1530-6.
22. ATS statement: guidelines for the six-minute walk test. Am J
Respir Crit Care Med 2002;166:111-7.
23. Finch E, Brooks D, Stratford PW, Mayo NE. Physical rehabilita-
tion outcome measures: a guide to enhanced clinical decision
making. Ontario: Lippincott Williams & Wilkins; 2002.
24. Sayers SP, Harackiewicz DV, Harman EA, Frykman PN, Rosen-
stein MT. Cross-validation of three jump power equations. Med
Sci Sports Exerc 1999;31:572-7.
25. Payne N, Gledhill N, Katzmarzyk PT, Jamnik VK, Keir PJ.
Canadian musculoskeletal fitness norms. Can J Appl Physiol
2000;25:430-42.
26. Mathiowetz V, Weber K, Volland G, Kashman N. Reliability and
validity of grip and pinch strength evaluations. J Hand Surg [Am]
1984;9:222-6.
27. Desrosiers J, Hebert R, Bravo G, Dutil E. Comparison of the
Jamar dynamometer and the Martin vigorimeter for grip strength
measurements in a healthy elderly population. Scand J Rehabil
Med 1995;27:137-43.
28. Paltamaa J, West H, Sarasoja T, Wikstrom J, Malkia E. Reliability
of physical functioning measures in ambulatory subjects with MS.
Physiother Res Int 2005;10:93-109.
29. Peolsson A, Hedlund R, Oberg B. Intra- and inter-tester reliability
and reference values for hand strength. J Rehabil Med 2001;33:
36-41.
 EXERCISE CAPACITY IN MULTIPLE SCLEROSIS, Kuspinar
30. Bohannon RW, Schaubert KL. Test-retest reliability of grip-
strength measures obtained over a 12-week interval from commu-
nity-dwelling elders. J Hand Ther 2005;18:426-7, quiz.
31. Schreuders TA, Roebroeck ME, Goumans J, van Nieuwenhuijzen
JF, Stijnen TH, Stam HJ. Measurement error in grip and pinch
force measurements in patients with hand injuries. Phys Ther
2003;83:806-15.
32. Cox MH, Thomas SG, Weller IM, Corey P. Reliability and va-
lidity of a fitness assessment for epidemiological studies. Can
J Sport Sci 1992;17:49-55.
33. Gibbons WJ, Fruchter N, Sloan S, Levy RD. Reference values for
a multiple repetition 6-minute walk test in healthy adults older
than 20 years. J Cardiopulm Rehabil 2001;21:87-93.
34. Kleinbaum DG, Kupper LL, Miller KE, Nizam A. Applied re-
gression analysis and other multivariable methods. 3rd ed. Pacific
Grove: Brooks/Cole Publishing Co; 1998.
35. Bjarnadottir OH, Konradsdottir AD, Reynisdottir K, Olafsson E.
Multiple sclerosis and brief moderate exercise: a randomised
study. Mult Scler 2007;13:776-82.
36. Romberg A, Virtanen A, Aunola S, Karppi SL, Karanko H,
Ruutiainen J. Exercise capacity, disability and leisure physical
activity of subjects with multiple sclerosis. Mult Scler 2004;10:
212-8.
37. Morrison EH, Cooper DM, White LJ, et al. Ratings of perceived
exertion during aerobic exercise in multiple sclerosis. Arch Phys
Med Rehabil 2008;89:1570-4.
38. Dalgas U, Stenager E, Ingemann-Hansen T. Multiple sclerosis and
physical exercise: recommendations for the application of resistance-,
endurance- and combined training. Mult Scler 2008;14:35-53.
39. Cahalin L, Pappagianopoulos P, Prevost S, Wain J, Ginns L. The
relationship of the 6-min walk test to maximal oxygen consump-
1417
tion in transplant candidates with end-stage lung disease. Chest
1995;108:452-9.
40. Miyamoto S, Nagaya N, Satoh T, et al. Clinical correlates and
prognostic significance of six-minute walk test in patients with pri-
mary pulmonary hypertension: comparison with cardiopulmonary
exercise testing. Am J Respir Crit Care Med 2000;161:487-92.
41. Montgomery PS, Gardner AW. The clinical utility of a six-minute
walk test in peripheral arterial occlusive disease patients. J Am
Geriatr Soc 1998;46:706-11.
42. Hendrican MC, McKelvie RS, Smith T, et al. Functional capacity
in patients with congestive heart failure. J Card Fail 2000;6:214-9.
43. Eng JJ, Chu KS, Dawson AS, Kim CM, Hepburn KE. Functional
walk tests in individuals with stroke: relation to perceived exertion
and myocardial exertion. Stroke 2002;33:756-61.
44. Rantanen T, Masaki K, Foley D, Izmirlian G, White L, Guralnik
JM. Grip strength changes over 27 yr in Japanese-American men.
J Appl Physiol 1998;85:2047-53.
45. Rantanen T, Guralnik JM, Foley D, et al. Midlife hand grip strength
as a predictor of old age disability. JAMA 1999;281:558-60.
46. Sasaki H, Kasagi F, Yamada M, Fujita S. Grip strength predicts
cause-specific mortality in middle-aged and elderly persons. Am J
Med 2007;120:337-42.
47. Gale CR, Martyn CN, Cooper C, Sayer AA. Grip strength, body
composition, and mortality. Int J Epidemiol 2007;36:228-35.
48. Ali NA, O’Brien JM, Jr, Hoffmann SP, et al. Acquired weakness,
handgrip strength, and mortality in critically ill patients. Am J
Respir Crit Care Med 2008;178:261-8.
Suppliers
a. Racer Mate Inc, 3016 Northeast Blakeley St, Seattle, WA 98105-
4012.
b. SAS Institute Inc, 100 SAS Campus Dr, Cary, NC 27513-2414.
c. Jamar Technologies, 151 Keith Valley Rd, Horsham, PA 19044.
Arch Phys Med Rehabil Vol 91, September 2010