Endurance exercise improves walking distance in MS patients with fatigue


Prof. Dr Christian Dettmers, Kliniken Schmieder Konstanz, Eichhornstr. 68, 78464 Konstanz, Germany.
Tel.: +49 7531 986 3537
Fax: +49 7531 986 3155
e-mail: c.dettmers@kliniken-schmieder.de


Objectives –  Effects of endurance training in multiple sclerosis (MS) patients complaining of motor fatigue.

Materials and methods –  Thirty MS patients complaining of fatigue with low to moderate disabilities randomly allocated to the intervention (thrice weekly 45-min intervals of endurance exercise) or control treatment (three 45-min episodes of stretching, balance training and coordination), both as ‘add-on’ therapy for 3 weeks during inpatient rehabilitation.

Results –  Maximal walking distance before intervention averaged 1043 ± 568 and 1163 ± 750 m in the two groups. The intervention group increased its maximal walking distance by 650 ± 474 m. The control group extended its walking distance by 96 ± 70 m.

Conclusions –  The present data confirm a strong effect of endurance exercise on maximal walking distance. Remarkably, there were no parallel improvements on the Modified Fatigue Impact Scale, the Beck Depression Inventory and the Hamburg Quality of Life Questionnaire for MS.


To date no one has disputed the fact that fatigue is a frequent symptom in MS patients (75–90%) (1) despite the uncertainty regarding its pathophysiological basis, diagnostic criteria and tools for assessments. In a sample of over 500 MS patients and over an observation period spanning 2 years, 38% had persistent fatigue and 37% sporadic fatigue (2). It is experienced as the worst symptom in 28–40% of MS patients (3, 4). Patients, doctors and therapists are often convinced that fatigue is not easy to overcome or to change. In addition, there are limited trials, which seek to ameliorate fatigue through exercise, especially through endurance exercise (5–7). Patients have been often educated not to provoke relapses by intense training (8). Generally, physiotherapists also focus their therapy on the neurological deficit, and not on endurance or strength training (9). Thus far, physical exercise has not been central to MS patients’ rehabilitation.

Recently, an increasing number of studies tried to investigate the effect of regular exercise in MS patients on physical fitness, quality of life and potentially also on relapse frequency (10–15). These studies support the idea that regular exercise should help to improve on maximal walking distance in patients with fatigue.

As such, we were interested in finding out whether regular training, thrice weekly for a period of 45 min for 3 weeks, might improve on patients’ performance. This, in turn, would be compared with a control intervention which would avoid cardiovascular training.



The study was approved by the local ethics committee at the University of Konstanz. All patients gave informed consent. The study was performed in agreement with the declaration of Helsinki.


The study was conducted over a 4-month period, at an inpatient rehabilitation unit with 180 beds and with a yearly intake of approximately 400 MS patients. Rehabilitation programs generally last 5 weeks on an average.

In order to differentiate between motor fatigue and other neurological deficits, only patients with mild to moderate MS with an Expanded Disability Status Scale (EDSS) <4.5 were selected (compare Table 1). Between July and September 2007, 30 patients (corresponding to about 30% of all patients admitted to the hospital in this time frame) were consecutively recruited, who complained of fatigue and whose maximal walking distance was considerably reduced (<2500 m, compare Table 1). Main selection criteria were that the decrease in their walking distance could not be explained through the extent of their pareses, ataxia or spasticity. Patients with permanent, serious leg weakness (degree of paresis <4 according to the Medical Research Council), ataxia or spasticity even at rest were excluded from the study. Patients with a relapse or corticoid treatment within the last 3 months were also excluded. Symptomatic medical treatment or disease-modifying drugs were not altered during the course of the study. Further exclusion criteria included prominent cognitive deficits, major depression or insufficient motivation to participate in additional training along with the normal rehabilitation program.

Table 1.   Clinical characteristics of patients
 Intervention groupControl group
Age (years)45.8 ± 7.9 (29–60)39.7 ± 9.1 (16–52)
Secondary progressive22
Primary progressive03
EDSS2.6 ± 1.2 2.8 ± 0.7
Duration since first manifestation10.7 ± 7.310.5 ± 6.7
Duration since diagnosis8.0 ± 5.96.1 ± 4.3
Completely (partially) retired for health reasons2 (1)3 (4)
MFIS36.8 ± 17.441.8 ± 20.3
MFIS motor17.4 ± 8.322.0 ± 7.3
BDI9.9 ± 6.89.9 ± 5.5
HAQUAMS114.0 ± 15.3113.9 ± 10.5
Maximal walking distance (m), 1st test1043 ± 5681163 ± 750
Maximal walking distance (m), 2nd test1693 ± 9781260 ± 794


A master’s degree student in Physical Education conducted both interventions (16). He had received additional training from physiotherapists and neurologists regarding treatment and training of MS patients. Both interventions lasted for thrice weekly 45-min sessions. Intervention training consisted of warming up, mild strength training, repetitive endurance exercise, followed by relaxation and feedback. The trainer tried to camouflage training difficulties by including games and other playful elements. An example of this training was very similar to a biathlon. Patients walked several rounds. Then they had to throw balls at cans. Depending on their successful hits they might have to walk extra rounds. Another playful element consisted of ‘playing cards’. Patients had to collect cards from different corners of the hall. Depending on the cards they found, they might have to collect more cards from various places in order to complete a certain set of cards. There were a small number of participants per training session (three to five). This enabled an individual training plan. Generally, patients were trained to keep their own comfortable speed and not to compete too hard with other individuals. They were expected to slow down their pace when they felt tired, but they were asked not to stand still. In contrast to training in healthy volunteers, demanding tasks were performed at the beginning and not at the end of the training. Pulse frequency was not systematically recorded and documented. Sporadic controls always indicated pulse frequencies below 140 per minute.

Control training consisted of warming up, sensory training, stretching, balance, coordination training and periods of relaxation. Any training involving the heart and circulation was avoided. Training sessions lasted 3 weeks.


Patients were recruited between the second or fourth day of their inpatient rehabilitation program. Maximal walking distance was determined on the fifth or sixth day respectively. Questionnaires were also completed on the sixth day. Patients were assigned to one of the two groups according to a randomly ordered list. The list was only accessible to the therapist who was not involved in patient selection. Training started on the eighth day and continued for 3 weeks until the end of the fourth week. Maximal walking distance was measured, and questionnaires were repeated on completion of the training sessions. These included Modified Fatigue Impact Scale (MFIS) (17, 18), Fatigue Scale for Motor and Cognition (FSMC) (19), Beck Depression Inventory (BDI) (20) and the Hamburg Quality of Life Questionnaire in Multiple Sclerosis (HAQUAMS) (21). The MFIS is a standardized questionnaire to assess fatigue on a scale from 0 to 84. Thirty-eight is the cut-off for fatigue (18). The FSMC ranging from 0 (no symptom of fatigue) to 50 (maximum fatigue symptoms) is a newly developed questionnaire which differentiates more precisely between motor and cognitive fatigue. The evaluation study with a cohort of 300 enrolled patients and volunteers is not yet published. The cut-off is expected to be around 25 for each modality. The BDI is a common evaluation form in treatment studies in patients with depression. The HAQUAMS is a questionnaire developed with special emphasis on MS patients’ needs.

Walking parameters

Patients selected for the study were invited to make themselves comfortable with the treadmill. The patients determined a comfortable walking pace for themselves. According to a standardized protocol, the actual measurement took place on a different occasion starting at 09:00 hours, after patients’ breakfast and without their participation in any of the training sessions. When subsequent measurements of walking distance were completed, the treadmill was set at each patient’s predetermined ‘comfortable’ walking pace. Patients were asked to indicate, when they were ready to stop walking and needed to have a break. The amount of exertion corresponded to level 16 or 17 on the Borg scale. Walking distance and time were recorded, while walking speed was set on a constant, individual speed as indicated previously. The assessor was trained not to encourage patients in either direction. He was not allowed to encourage patients to continue on the treadmill once they started to complain about fatigue; neither was he expected to indicate when they had attained the walking distance of the previous test. Patients were not allowed to read the distance they had moved on the treadmill. They were not informed as to how far they had reached. The assessor was aware of patients’ allocation to treatment and control groups.

Outcome parameters

The primary outcome parameter was walking distance as determined on the treadmill. Time allocated for walking was also recorded to determine patients’ different walking speeds. Patients were also asked to complete four questionnaires (MFIS, FSMC, BDI and HAQUAMS).


We measured the two parameters, walking distance s and walking time t, to depict changes in patients’ walking ability. While there is a difference in the distribution of the absolute values between distance and time, the distribution of the relative values are identical. Each subject was allowed to choose a comfortable walking speed. Thereafter, pre- and post-training measurements were performed with these constant speeds. Since walking distance s and walking time t are related via the well-known relation s = vsubjectt we took measurements for changes in relative walking ability (rWA) as follows:


A t-test and an F-test on the pretraining data were carried out to determine homogeneity in the two training groups.


At the time of selection, patients were characterized as follows: Mean Extended Disability Status Scale (EDSS) was 2.6 ± 1.2 and 2.8 ± 0.7, respectively, at the border between low and moderate disability (see Table 1 for patients’ details).

Average disease duration was 10.7 ± 7.3 and 10.5 ± 6.7 years, respectively, for the intervention and control groups. Two patients in the intervention group and three in the control group were retired. In the intervention group, one patient had partial retirement status because of the disease and similarly four in the control group. MFIS was 36.8 ± 17.4 and 41.8 ± 20.3 for both groups not significantly different and indicative of fatigue in most instances (cut-off of the MFIS is 38). Beck Depression Inventories were identical and normal in both groups (9.9 ± 6.8 and 9.9 ± 5.5). The Hamburg Quality of Life Questionnaire for MS (HAQUAMS) was almost identical in both groups (114.0 ± 15.3 and 113.9 ± 10.5). Maximal walking distance was determined on the treadmill with 1043 ± 568 (intervention group) and 1163 ± 750 m (control group) (Fig. 1, Table 1).

Figure 1.

 Walking speed of participants of both groups. Ordered according to increasing speed from left to right. There is no statistically significant difference between the two groups’ mean value (P = 0.890) as well as standard deviation (P = 0.668).


The overall acceptance rate was very high. One patient in the intervention group dropped out because training was too demanding. He was replaced with the recruitment of a 16th patient. No one dropped out in the control group. A total of 15 patients in each group completed the study.

The rWA improved significantly (P < 10−4) in the intervention group by 66% in comparison with 12 % in the control group (see Figures 2–4).

Figure 2.

 Maximal walking distance before (left column of each pair) and after (right column of each pair) the intervention. Patients are ordered according to increasing walking distance.

Figure 3.

 Maximal walking distance before (left column of each pair) and after (right column of each pair) 3 weeks in the control training. Patients are ordered according to increasing walking distance. Please note that the maximal walking distance is not as different and ‘unpredictable’ as many patients believe, but appear to be a pretty constant phenomenon.

Figure 4.

 Increase in relative walking distance after the training for the intervention group (left column of each pair) receiving endurance exercise. The control group (right column of each pair) had received sensory training, balance training, stretching and relaxation but avoiding any kind of circulation training.

Only two patients in the intervention group (patients 2 and 12) did show a somewhat small increase. The improvements were rather homogenous. Walking distance extended for 650 ± 474 m compared with 96 ± 70 m in the control group (Table 2). Walking time increased by 11 min in the intervention group compared with 1 min in the control group. The differences between both groups regarding walking distance and walking time were highly significant (P = 0.001) (Table 2).

Table 2.   Increase in walking distance and time (mean ± standard deviation)
 Intervention groupControl group
Walking distance (m)650 ± 47497 ± 70
Walking time (min)11.3 ± 61.3 ± 1

In both groups, 9 and 10 patients completed questionnaires after training. Interestingly, the MFIS and the FSMC scores did not demonstrate similar changes as the extension of the walking distance. The relative change was very similar in both groups. Improvement reached significant levels in the control group, but not in the intervention group (Table 3). Moreover, BDI and HAQUAMS demonstrated more often improvements in single patients of the control group than in the intervention group (Table 3).

Table 3.   Description of changes of questionnaires
 Intervention groupControl group
  1. Due to the incomplete data set no statistical tests were performed.

Motor MFIS


The major result of the present study is that low-level endurance training, over a 3-week period, in MS patients with motor fatigue improves on walking distance by 66%. The studies’ limitations are outlined before discussing the results in the light of current literature.

For economical and logistic reasons, it was not possible to document precisely why 70% of the MS patients admitted to our hospital did not qualify for the study. Some of them were wheel chair bound or dependent on walking sticks which disqualified them for the treadmill test. The majority had serious ataxia, paresis or spasticity which made it difficult to distinguish, whether the reduction in the walking distance was due to the primary deficit (paresis, ataxia and spasticity) or due to fatigue. This does not imply that these patients might not also have some additional fatigue. For clarity, we selected those patients in whom the reduction in the maximal walking distance was clearly the result of motor fatigue. Other common reasons for exclusion were serious cognitive deficits, major depression or lack of motivation. In our opinion, 30% of patients recruited for our study does not reflect a highly selected minority of patients, but fairly represents those patients with mild to moderate MS (EDSS <4.5) and motor fatigue.

Unfortunately, the return rate of completed forms was insufficient. As the second treadmill test was performed shortly before patient discharge, fully completed forms for only 19 of 30 patients were collected. We do not see any obvious systematic bias between those who returned their questionnaire and those who did not. Nonetheless, we reject any statistical evaluation and simply describe the results. Although the incomplete return rate does not facilitate drawing definite conclusions, the completed questionnaires raise important questions for future studies.

The major methodological limitation is that the assessor of the primary outcome parameter was not blind to the allocation of patients to the treatment groups. This clearly limits the value of the study. On the other hand, the assessor was trained and repetitively reassured that he must maintain a neutral stance and not encourage patients in either direction.

With respect to actual literature on the subject, it is the consensus of opinion that fatigue in MS patients is not of peripheral or muscular origin – although patients frequently relate their fatigue to their legs or their muscles. The metabolic change in muscles during voluntary exercise and fatigue was even smaller than in controls suggesting the lack of central drive or muscle activation (22). Thus, fatigue is of central origin. Its precise pathophysiology is not known. Alterations in immune system activation, central nervous system dysregulation, cytokines and neuroendocrine dysregulation may be involved (23) possibly causing a ‘use-dependent conduction block’. Smith et al. described a temporary change in sensory symptoms during exercise, but no development of relapses or permanent deficits (12).

The increase in 66% of the maximal walking distance is a remarkable observation. The question arises how can we explain such a pronounced and quick improvement in endurance? A partial explanation may be that previously MS patients were told to be very careful with or even restrict physical exercise, as they were prone to provoking a relapse (8). As such the cardiovascular system is often in a worse condition than it should be. Primary motor fatigue – although a phenomenon of and within the nervous system – obviously has harmful effects on the cardiovascular system. Serious fatigue might also cause some patients to avoid demanding tasks. This systematic under-demanding of the cardiovascular system – comparable with the learned non-use of the affected arm in stroke patients – may be called learned under-use or adopting a saving or protective attitude leading to a vicious circle of fatigue and deconditioning. Deconditioning may be partially reversed through training. This might explicate the rapid and profound improvements in our study.

This explanation is in line with recent studies that investigated the relationship between exercise, fitness and quality of life in MS patients. In a meta-analysis of the effect of exercise training and quality of life, it was concluded that exercise training is associated with a small improvement in quality of life among individuals with MS (5). Similar conclusions were drawn in a recent review (8). They also stated that MS patients not only suffer from paresis but also have disuse of the muscles and circulation. Other authors also identified deconditioning as a substantial cause of fatigue (24).

Historically, general advice for MS patients with inflammatory disease has been to avoid exhaustion in order not to provoke a relapse or progression of the disease by overexertion (5). A variety of studies addressed the phenomenon of deconditioning in MS. Treadmill training was investigated in 16 adults with MS (10). The intervention consisted of 12 sessions of 30 min treadmill training. Comfortable walking speed and walking endurance increased, while oxygen consumption decreased at rest and during walking. Reduced respiratory muscle function was suggested to play a strong role in the reduced aerobic capacity of 25 MS patients (13). An investigation in the knee extensors in 15 MS patients demonstrated muscle mass of lower quality (i.e. reduced force/unit muscle mass) than in controls (25). Weekly exercise classes with the use of a stationary bicycle along with home exercise showed significant improvement in measures of fatigue compared with waiting list control group (26). In a very small randomized controlled study, it was demonstrated that peak oxygen consumption increased after a brief moderate exercise program thrice a week for 60 min over 5 weeks (6). Petajan et al. recommended regular physical training on an outpatient basis (27). Mostert and Kesselring also demonstrated an improvement in physical fitness through regular training (28).

A secondary observation in the control group is that the walking distance at the second point in time (after a 3-week period) was almost identical to the first measurement (Fig. 3). This is remarkable. If patients with fatigue are questioned as to how far they can reach, some of them might say: ‘it depends’. By that they mean, it varies, it is different every day, and it is unpredictable. The second measurement in our control group standardized at 09:00 hours without previous exertion was very similar to the first measurement in all instances. This observation argues for the reduction in walking distance to be a very predictable and constant, pathophysiological phenomenon. Again this reiterates an organic basis for primary motor fatigue possibly resembling a use-dependent conduction block.

Surprisingly the clinical scores in our study did not show similar trends as the improvements in the rWA. They did not demonstrate any trend parallel to physical improvement. Although the questionnaires were incompletely filled in after finishing training sessions, there was almost a negative trend, as the control patient group had more patients than the intervention groups, who showed an improvement on the depression scale (BDI), quality of life (HAQUAMS) and fatigue scale (MFIS) (compare Table 3). Although physical fitness was improved, mental state, mood and quality of life did not show similar trends. Possibly training in the intervention group was much more demanding and thus fatigue scales did not improve. Similar trends, however, have been reported in two studies cited above investigating short-term effects of treadmill training and exercise sessions (10, 12). Both studies revealed unchanged reported fatigue scales.

Most of the reviews describe a positive long-term effect of sports and exercise on physical fitness in MS patients (5, 10, 13). Some of them describe an improvement in quality of life (5, 13). Heesen et al. even discuss whether sports should be regarded as a disease-modifying treatment through modulation of the immune system (29). These studies advocate regular physical exercise in MS patients. This appears to be in contradiction to ‘energy conservation courses’. It has been shown that energy conservation courses decreased fatigue impact, improved some aspects of quality of life and increased self-efficacy among persons with MS (30, 31). However, it is not advisable to encourage MS patients to avoid exercise. So, it appears to be a difficult compromise between ‘energy conservation training’ and ‘endurance training’, which demands a different approach in each individual case.


(1) Systematic low-level endurance training is able to improve rWA in MS patients with motor fatigue. Therapists and patients should be informed about this option in order to select appropriate patients. (2) Endurance training is not necessarily expected to reduce the subjective feeling of fatigue. (3) Motor fatigue is an organic phenomenon.


This study was supported by Kliniken Schmieder Stiftung and the Lurija Institute. We greatly appreciate the advice of Katharina Iris Penner, PhD, for evaluation of the FSMC and Patricia T. Alleyne-Dettmers, PhD, linguistic anthropologist, who edited the manuscript. We also appreciate the constructive and careful criticisms of two reviewers.