J NMD 2021 05 002
J NMD 2021 05 002
PII: S0960-8966(21)00127-9
DOI: https://doi.org/10.1016/j.nmd.2021.05.002
Reference: NMD 3996
Please cite this article as: S Birnbaum , R Porcher , P Portero , B Clair , S Demeret , B Eymard ,
M Gargiulo , E Louët , S Berrih-Aknin , R Le Panse , P Aegerter , JY Hogrel , T Sharshar , the
MGEX Study Group, Home-based exercise in autoimmune myasthenia gravis: a randomized con-
trolled trial, Neuromuscular Disorders (2021), doi: https://doi.org/10.1016/j.nmd.2021.05.002
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.
Louët E, Berrih-Aknin S, Le Panse R, Aegerter P, Hogrel JY, Sharshar T* and the MGEX
Study Group.
Abstract
The tolerance of exercise and its effects on quality of life in myasthenia gravis are not
currently backed up by strong evidence. The aim of this study was to determine whether
exercise as an adjunct therapy is well tolerated and can improve health-related quality of life
block randomization. Adults with stabilized, gMG, and no contra-indication to exercise, were
eligible. Participants received usual care alone or usual care and exercise. The exercise
home rowing program over 3 months. The primary endpoint was the change in HRQoL from
Of 138 patients screened between October 2014 and July 2017, 45 were randomly assigned to
exercise (n=23) or usual care (n=20). Although exercise was well tolerated, the intention-to-
treat analysis revealed no evidence of improved HRQoL compared to usual care (MGQOL-
15-F; mean adjusted between-groups difference of -0.8 points, 95%CI -5.4 to 3.7). Two
patients hospitalized for MG exacerbation were from the usual care group.
Quality of life
2
Introduction
agents [1]. These therapeutic strategies aim to reduce symptoms, improve functional status
and health-related quality of life (HRQoL); three dimensions which can be assessed with
specific scores [2]. Despite appropriate therapeutic management, individuals with MG still
have reduced HRQoL which is not improving despite therapeutic advances [3]. Therefore,
adjuvant therapy for improving HRQoL in MG is warranted. Whilst rest is indicated during
Exercise therapy has been shown to improve strength and functional capacity in various
dystrophies [5]. There are physiological and clinical arguments for its therapeutic use in MG.
interventional human studies have shown improved lower limb strength [7, 8], walking
capacity and overall function [9-11] and, exercise self-efficacy [11]. However, these studies
In addition, exercise has positive psychological and immunological effects, which are of
potential benefit to individuals with MG. Symptoms of depression and anxiety have been
effects have been induced by the release of myokines during repetitive muscle contractions
[15, 16]. Because of its pluripotent effects, it is conceivable that exercise could improve
HRQoL in MG. The objective of this study was to evaluate whether an unsupervised home
3
exercise program (ExP) could improve MG-related HRQoL of individuals with stabilized and
gMG, with reduced HRQoL. Secondary objectives were to assess exercise tolerance and the
Methods
Study design
comparing a home-exercise program to usual care (UC). Full details of the study rationale,
design, methods and statistical analysis have been previously published [17]. The overall
study duration for an individual participant was 36 weeks. Total intervention time was 12
weeks beginning at randomization at month 3 (M3), ending at month 6 (M6). This was
Writing of this manuscript follows the CONSORT guidelines [18]. Ethics approval to carry
out the study on human subjects was granted by the French regulatory board (Comité de
Protection des Personnes Ile de France XI under the CPP number #13064 on 13/12/2013 and
11/10/ 2013). The trial was conducted in accordance with the Helsinki Declaration and was
during regular outpatient consultations and study information was diffused to MG-patient
gMG (class II-III according to the Myasthenia Gravis Foundation of America (MGFA)
classification) were eligible. Six months of MG disease stability, determined by the treating
neurologist, and reduced HRQoL (MGQOL-15-F score ≥ 15), was required for inclusion.
4
Exclusion criteria included any contra-indication to exercise which precluded participation in
orthopedic condition, chronic pain, recent hospitalization for a serious medical or surgical
necessitating specific protection (Table S1). Written, informed consent was obtained from all
were randomized to either the intervention (UC + exercise) or control (UC) group with a 1:1
permuted blocks of randomly varying sizes. To ensure concealment, block sizes were not
disclosed. The randomization list was constructed prior to the beginning of the study by a
statistician with no clinical involvement in the study. Group allocation was revealed
software (CleanWeb), and was concealed until the moment of group allocation. Investigators,
evaluators and study personnel were all masked to group assignment for the duration of the
study. Only the physiotherapist performing the neuromuscular evaluations and managing the
Intervention
All participants received UC and evaluations as per the study protocol. The control group
were instructed to maintain their regular physical activity without starting a new exercise
program or increasing baseline activity. Baseline activity was measured over one week, prior
The intervention group was invited to participate in a 12-week home-based aerobic exercise
program (ExP) using a rowing ergometer (Concept2™), Table S2. The ExP consisted of three
5
40-minute sessions per week, i.e. a total of 36 sessions over the intervention period. A chest
heart rate (HR) monitor (Garmin®) was worn by participants and training HR was visualized
on a screen for direct feedback. Each 40-minute moderate-intensity rowing session consisted
of a 10 minute warm-up to reach each participants’ individual target HR (i.e. 70% of their
maximal HR [HRmax], using the equation 220-age as their HRmax), followed by a 20-minute
plateau phase of constant aerobic activity, rowing at their individual target HR, followed by a
5-minute power interval phase (5 sets of ten consecutive pulls at maximum effort performed
at the beginning of each minute, followed by regular intensity strokes for the remainder of
each minute) and finally, a 5-minute active cool-down period where rowing was continued at
a slow pace. Data (rowing distance and session duration) were recorded on an individual
logcard for each training session. Participants were advised to organize their training sessions
on alternate days. The initial 1-3 sessions took place at a different site (Rothschild Hospital,
Paris) to study visits, to maintain investigator blinding. These individualized sessions were
carried out by a trained physiotherapist to ensure feasibility, to teach participants the ExP,
how to use the HR monitor, the rowing machine (Concept2™) and how to perform the rowing
movement. Once the physiotherapist and participant were satisfied that the ExP could be
performed unsupervised at home, the rowing machine was then delivered to the participant’s
home for the duration of the intervention period (3 months). The Concept 2™ rowing
machine was selected as it is a stationary ergometer where one is seated and simultaneously
exercises upper and lower limbs and the trunk thus targeting the multiple muscles which can
be affected in MG. Adherence was defined as having completed at least 20 (frequency) 30-
Outcomes
The primary endpoint was the adjusted mean between-group difference in MG-specific
HRQoL, assessed with the French version of the Myasthenia Gravis Quality of Life
6
questionnaire (total score) [19] (MGQOL-15-F; range 0-60; higher scores represent lower
perceived HRQoL) between the ExP and UC group from M3 (baseline) to M6 (post-
intervention).
To assess tolerability of the ExP compared to UC, adverse events were recorded at each
monthly visit except serious adverse events such as MG crisis or MG exacerbation with
hospitalization which were notified immediately. MG crisis was defined as rapid clinical
exacerbation was defined as clinical deterioration with beyond usual impact on daily life, with
modification.
evaluated with the quantitative Myasthenia Muscle Score [20] (MMS; range 0-100; lower
scores represent greater symptom severity), impact of MG on daily function, evaluated with
the Myasthenia Gravis Activities of Daily Living scale [21] (MG-ADL; range 0-24, higher
scores represent greater impact of MG on function), physical capacity evaluated with the 6-
minute walk test (6MWT) for walking endurance [22, 23], maximal voluntary isometric
strength of elbow flexion and knee extension using a Biodex (Shirley, USA) and handgrip
strength using a MyoGrip and respiratory function (Forced Vital Capacity, maximal
MicroRPM). In addition, change in non-specific QoL (WHO-QOL Bref) [24, 25] and
psychological status; depression (Beck Depression Inventory) [26], anxiety (State Trait
7
Anxiety Inventory) [27] and self-esteem (Self-Esteem Inventory scale) [28] before and after
the ExP were assessed. Exploratory endpoints consisting of cytokine levels of immunological
Statistical analysis
Sample size calculations were described previously [17]. The primary analysis was conducted
in the intention-to-treat (ITT) population, where all randomized participants were analyzed in
the group they were allocated to irrespective of any protocol deviation. The primary outcome
was analyzed using analysis of covariance (ANCOVA), with changes in MGQOL-15-F as the
dependent variable, and randomization group, MGQOL-15-F score at baseline and center as
covariates. A similar model was used for secondary and exploratory outcomes defined as
changes from M3. For binary secondary outcomes, absolute risk difference and relative risks
were estimated. For the monthly MMS and MG-ADL scores, statistical tests were based on
linear mixed models on the ranked scores, with random subject intercept and slope, and a
separate fixed-effect for each time as well as interactions with the randomization arm from
Cytokines were analyzed after log-transformation and adjusted ratios of geometric means are
presented. For values under the lower limit of detection, ANCOVA was replaced by tobit
models, with the same formulation. For safety analyses, the numbers of participants with at
least one AE or SAE were compared by Fisher’s exact tests, and the number of AEs by
Poisson regression. Missing data were handled by multiple imputation by chained equations.
weight, MGQOL-15-F, MMS, MG-ADL, WHO-QOL, STAI, Beck, MINI, SEI, 6-minute
walking distance (6MWD), and adherence status in the intervention arm at M3, M6 and M9.
8
Imputations were performed separately in each randomization group as recommended [29].
Imputation of continuous variables was performed using predictive mean matching. Forty
imputed datasets were generated and analyzed. Results were then pooled. Physiological (e.g.
strength), biological (cytokines) and safety outcomes were not imputed. No allowance was
made for multiplicity for secondary endpoints. All tests were two-tailed with p values less
than 0.05 defined as significant. A compliers average causal effect (CACE) analysis was
carried out using an instrumental variable approach. Notification of AEs and visit tracking
Results
Participants
Figure 1 presents the study flow diagram. Between October 3rd, 2014 and July 7th, 2017, 138
patients were screened, of which 45 were enrolled in the study with 23 being randomly
assigned to the exercise intervention and 20 to the UC control group. Forty-one participants
(91%) completed the study. Two participants withdrew before randomization. One participant
refused to undertake the ExP but remained in the study. Two participants (control group) were
lost to follow-up, one following randomization (M3) and the other in their last month of the
study (between M8 and M9). Baseline characteristics are presented in Tables 1-2.
Intervention adherence
sessions were completed. Sixteen (70%) participants adhered to the ExP (Table S3). Reasons
for non-adherence were mainly work commitments. Reasons for missing occasional sessions
included: the flu, weekend away, menstrual pain/tiredness. Participant characteristics were
9
There was an adjusted between-groups mean difference of -0.8 points (95%CI -5.4 to 3.7) on
the MGQOL-15-F score at the end of the intervention period (M6) and an adjusted between-
groups mean difference of 3.9 points (95%CI -1.7 to 9.5) at follow-up (M9) (Table 2),
indicating that there was no difference in MG-QOL-15-F score between the two groups at
Exercise tolerance
Two participants from the control group necessitated hospitalization for an acute MG
exacerbation. No exacerbation was reported in the ExP group throughout the entire study
(Tables 3-4). There was no between-groups difference with regards to exercise tolerance
including cardiorespiratory indicators, muscular or articular pain (Table 3), nor concerning
non serious adverse events such as infection, allergic reaction, cephalalgia (Table 4). There
was no evidence of raised creatine phosphokinase levels with exercise (Table S5).
Secondary endpoints
There was a significant adjusted between-groups mean difference of -1.9 for the MG-ADL
(95%CI -3.0 to -0.9) and 27.7 m for the 6-minute walking distance (95%CI 7.2 to 48.1) in
favor of the ExP group at M6 but these differences were not maintained at M9 (Table 2).
Knee extension strength increased in the ExP group without reaching statistical significance
(8.8 Nm, %95 CI 0.1 to 17.5), Table 2. No between-groups difference was observed for the
MMS, elbow or grip strength, pulmonary function, general QoL (WHO-QOL BREF) or
psychological status (STAI, BDI-13, SEI) (Table 2, Table S6). Further, there was no between-
groups difference in prednisone dose -0.4 (95%CI -2.1 to 1.2) nor AChEi dose -0.1 (95%CI -
0.6 to 0.4) at M6, nor at M9 for prednisone -0.2 (95%CI -1.5 to 1.0) or, AChEi -0.0 (95%CI -
0.5 to 0.4).
Exploratory endpoints
10
There was no between-groups difference in adjusted ratios of geometric means from baseline
in cytokines IL-6; 1.09 (95%CI 0.97 to 1.23) nor TNF-α; 0.97 (95%CI 0.86 to 1.10), at M6,
nor at M9, IL-6; 1.07 (95%CI 0.95 to 1.21), TNF-α; 1.02 (95%CI 0.91 to 1.13).
Discussion
This randomized, clinical trial of individuals with mild-moderately severe, generalized and
stabilized MG, demonstrated that a 3-month unsupervised home exercise program did not
improve HRQoL (MGQOL-15-F) compared to UC. The exercise program was, however, well
tolerated, with improvement in the perceived impact of MG on daily life (MG-ADL score)
immunological markers.
We believed that HRQoL would be an appropriate primary endpoint for measuring the
pluripotent effects of exercise on both physical and psychological domains. Quality of life
measures provide information with regards to how satisfied one is with their current level of
functioning compared to what they perceive to be possible or ideal [30]. However, self-
subject to a response shift phenomenon whereby one’s internal reference can change over
time or as a result of an intervention (herein exercise), or placebo [31]. We were not able to
assess to what extent this phenomenon occurred. It is possible that the absence of a significant
change in MGQOL-15-F was due to the intervention itself. The frequency, mode and duration
of the exercise program may not have been sufficient to shift this complex construct nor, to
patients with an MGQOL-15-F score above 30 in order to have a greater chance of detecting a
benefit from exercise. It appeared that for the majority of patients with an MGQOL-15-F
11
score > 30, MG was not stabilized, which we considered a contra-indication to the moderate-
The improvement of walking endurance evaluated with the 6MWT and trend toward an
increase in knee extension strength indicates that our exercise program had an impact on
muscle performance, further confirmed by the CACE analyses (Table S8). Importantly, this
demonstrates a transfer from a rowing exercise to walking capacity, which we did not
specifically train. One may argue that an increase of 27.7m over six minutes is small however,
it is not only clinically relevant [32] but it should be noted that the distance walked in this
cohort was high prior to randomization (498.1±83.4m), making it more difficult to have such
a significant change [33]. Given that our exercise intervention was aerobic, these effects are
gains demonstrated by an increase in rowing distance for the same target heart rate.
Lower limb strength and functional gains (30 second chair stand test) have been demonstrated
in previous studies of exercise interventions in MG [7, 8]. We would like to emphasize that
any gain in strength for MG patients is important as weakness likely contributes to fatigue and
reduced daily function [34]. Consistent with our results, although the training programs in
these studies included upper limb training, no improvements were observed in the upper limb.
These studies were within-subject control (contralateral limb tested) [7] and repeated
measures design [8], each consisting of 11 participants. Thus it was important to confirm
Further, our exercise intervention was unsupervised and there was no specific strength
12
Tolerance and safety were thoroughly assessed throughout this study via clinical
examinations, vital parameters as well as biological indicators. One of the major findings in
this clinical trial is the tolerance of the exercise program. We did not observe any MG
with specific training in using the rowing machine and the individualized nature of the
program. We are not able to determine to what extent the high tolerance option may have
compromised efficiency. Only a trial comparing an exercise program with various intensities
and durations could address this issue. However, good tolerance is a major argument for
recommending this type of exercise program in a larger population. The rowing machine is a
adherence than a supervised programme. Adherence should be taken into account when
reported reasons were unrelated to the exercise program nor to their MG and baseline
characteristics between the two groups were similar (Table S4). The non-adherence in our
study is considered acceptable considering the intervention was home-based and unsupervised
and prior studies have shown similar compliance for exercise interventions [35]. Our
definition of non-adherence was arbitrary and the small number of participants limits
comparison between adherers and non-adherers, in determining the minimal exercise dosage
for efficacy. This RCT was designed with intention-to-treat (ITT) analyses however, CACE
analyses provides further insight into the effects of the exercise according to compliance. The
difference in the results of the CACE and ITT analysis suggests that it would be worth
optimizing adherence of the exercise intervention in future studies. This could possibly be
13
achieved with direct supervision or with remote monitoring using wearables, keeping in mind
This study presented some methodological limitations. Whilst our sample size could be
considered modest, there was an a priori calculated sample size, which is comparable to other
studies of exercise in rare disease [36, 37]. Despite the fact that MG is more prevalent in
women before the age of 50 (juvenile and early MG represents 86% of this cohort) [38], there
still remains disproportionately more women (40) than men (3) in this study, and all men were
randomly assigned to the control group. We note that this sex imbalance was also present in a
previous study suggesting possible selection bias which should be taken into account in future
confirmed in a larger population, notably to ensure the applicability and tolerance of our
exercise program particularly in men whom were largely lacking in this trial. Whilst it would
be of interest to assess whether the effects of exercise varies according to specific clinical or
immunological phenotypes of MG, our aim was to first evaluate the tolerance in all types of
MG. Due to the type of intervention, blinding of participants was not possible and some
improvement could be related to positive expectation bias particularly for those who were
At this stage more research is required to answer questions concerning the type of exercise,
intensity, duration and frequency and the impact of each aspect on individuals and subgroups
of MG. It may be important to consider current lifestyle and physical activity levels for better
need to be specifically assessed. Strengths of this study include the rigorous study design and
14
exercise using validated measures. Participant retention throughout the study was good with
Conclusion
This randomized clinical trial with intention-to-treat analyses is the first to demonstrate that a
3-month, home-based unsupervised exercise program is well tolerated and does not contribute
to exacerbations in adults with stabilized, gMG. This is a major result as benefit and risk of
exercise is a major concern for individuals with MG and for which no specific
15
Figure legend
Table legends
SD, standard deviation; IQR, interquartile range; BMI, body mass index; MGFA, Myasthenia
outcomesa.
a
Analyses of imputed data of primary efficacy endpoint (MGQOL-15-F: French version of the
15-item MG-specific quality of life scale). Secondary efficacy endpoints; symptom severity
(Myasthenia Muscle Score), limb strength, walking capacity (six-minute walking distance),
expiratory pressure). %pred, % predicted; SD, standard deviation; MD, mean difference; 95%
CI: 95% Confidence Interval; M3, Randomization, at month 3; M6, end of intervention
period, at month 6; M9, end of study at month 9. *Adjusted mean difference estimated by a
latent class mixed-effect models, and 95% confidence interval by bootstrap; **p-value
monthly visit.
16
RD: absolute risk difference; RR: relative risk; 95% CI: 95% Confidence Interval; M6, month
6; M9, month 9.
Table 4. Patient-reported adverse events throughout the entire study as reported by the
participant at each monthly visit. Serious adverse events reported at any time
62 adverse events were declared for 30 patients, among which 2 were considered a serious
adverse event. P values based on Fisher’s exact tests except * where P value is based on
Funding
This study was sponsored by the Assistance Publique – Hôpitaux de Paris « Direction de la
Recherche Clinique et de l’Innovation”. This study was funded through a grant from the
AOR12149), in collaboration with the Institute of Myology. The funders had no role in the
design of the study; nor data analysis, data interpretation, nor in the writing of this
manuscript.
Aknin, R. Le Panse, P. Aegerter, J.Y. Hogrel, T. Sharshar report no disclosures relevant to the
manuscript.
Data availability
Individual de-identified participant data may be shared for research/academic purposes once
all secondary analyses have been completed. Requests for access to data must be made to the
17
corresponding author and sponsor and should be accompanied by a methodologically sound
proposal. A signed data sharing agreement is required before access can be provided.
Acknowledgements
We sincerely thank all participants. We thank the APHP for the grant through the French
Ministry of Health grant program for Clinical Research (AOR12149). We also thank the
Sophie Hue, Mathieu Surenaud, IMRB luminex platform, Creteil. We would also like to
acknowledge the nurses and administration staff who played a crucial role in this study.
Guillaume Bassez, Anthony Behin, Sonia Berrih-Aknin, Francis Bolgert, Nawal Derridj Ait-
Younes, Yasmine Domingo, Mélinée Frenkian, Diane Friedman, Asmaa Jobic, Pascal
Laforêt, Isabelle Ledoux, Judith Mendelson, Sandra Misdrahi, Cécilia Orblin Bedos,
Benjamin Rohaut, Jacques Ropers, Elodie Soler, Philippe Thoumie, Frédérique Truffault,
Supplementary Information
Supplementary data Table S2. Description of exercise intervention according to Slade et al.
2016
Supplementary data Table S5. Maximum creatine phosphokinase values between month 3
18
Supplementary data Table S6. Analyses of secondary outcomes quality of life and
psychological status.
Supplementary data Table S7. Analysis of primary endpoint according to adherence status.
Supplementary data Table S8. Analysis of strength and walking capacity according to
adherence status.
CONSORT checklist
Declaration of interests
☒ The authors declare that they have no known competing financial interests or personal
relationships that could have appeared to influence the work reported in this paper.
References
[1] Sanders DB, Wolfe GI, Benatar M, Evoli A, Gilhus NE, Illa I, et al. International consensus
2016;87:419-25.
[2] Barnett C, Herbelin L, Dimachkie MM, Barohn RJ. Measuring Clinical Treatment Response in
[3] Boldingh MI, Dekker L, Maniaol AH, Brunborg C, Lipka AF, Niks EH, et al. An up-date on
health-related quality of life in myasthenia gravis -results from population based cohorts.
[4] Bankole LC, Millet GY, Temesi J, Bachasson D, Ravelojaona M, Wuyam B, et al. Safety and
[5] Siciliano G, Schirinzi E, Simoncini C, Ricci G. Exercise therapy in muscle diseases: open issues
and future perspectives. Acta myologica : myopathies and cardiomyopathies : official journal
19
[6] Cheng A, Morsch M, Murata Y, Ghazanfari N, Reddel SW, Phillips WD. Sequence of age-
associated changes to the mouse neuromuscular junction and the protective effects of
[7] Lohi EL, Lindberg C, Andersen O. Physical training effects in myasthenia gravis. Archives of
[8] Westerberg E, Molin CJ, Sporndly Nees S, Widenfalk J, Punga AR. The impact of physical
[9] Wong SH, Nitz JC, Williams K, Brauer SG. Effects of balance strategy training in myasthenia
[10] Rahbek MA, Mikkelsen EE, Overgaard K, Vinge L, Andersen H, Dalgas U. Exercise in
myasthenia gravis: A feasibility study of aerobic and resistance training. Muscle & nerve
2017;56:700-709.
[11] Westerberg E, Molin CJ, Lindblad I, Emtner M, Punga AR. Physical exercise in myasthenia
[12] Schuch FB, Vancampfort D, Rosenbaum S, Richards J, Ward PB, Stubbs B. Exercise improves
physical and psychological quality of life in people with depression: A meta-analysis including
of the anxiolytic effects of exercise for people with anxiety and stress-related disorders: A
[14] Herring MP, O'Connor PJ, Dishman RK. The effect of exercise training on anxiety symptoms
[15] Kruger K, Mooren FC, Pilat C. The Immunomodulatory Effects of Physical Activity. Current
20
[16] Gleeson M, Bishop NC, Stensel DJ, Lindley MR, Mastana SS, Nimmo MA. The anti-
inflammatory effects of exercise: mechanisms and implications for the prevention and
[17] Birnbaum S, Hogrel JY, Porcher R, Portero P, Clair B, Eymard B, et al. The benefits and
[18] Boutron I, Moher D, Altman DG, Schulz KF, Ravaud P. Extending the CONSORT statement to
[19] Birnbaum S, Ghout I, Demeret S, Bolgert F, Eymard B, Sharshar T, et al. Translation, cross-
cultural adaptation, and validation of the french version of the 15-item Myasthenia Gravis
[20] Sharshar T, Chevret S, Mazighi M, Chillet P, Huberfeld G, Berreotta C, et al. Validity and
reliability of two muscle strength scores commonly used as endpoints in assessing treatment
[21] Wolfe GI, Herbelin L, Nations SP, Foster B, Bryan WW, Barohn RJ. Myasthenia gravis activities
[22] Enright PL. The six-minute walk test. Respiratory care 2003;48:783-5.
[23] Salci Y, Karanfil E, Balkan AF, Kutukcu EC, Ceren AN, Ayvat F, et al. Functional exercise
capacity evaluated by timed walk tests in myasthenia gravis. Muscle & nerve 2019;59:208-
212.
[24] Skevington SM, Lotfy M, O'Connell KA. The World Health Organization's WHOQOL-BREF
quality of life assessment: psychometric properties and results of the international field trial.
A report from the WHOQOL group. Quality of life research : an international journal of
21
[25] Leplege A, Reveillere C, Ecosse E, Caria A, Riviere H. Propriétés psychométriques d'un nouvel
[26] Collet L, Cottraux J. [The shortened Beck depression inventory (13 items). Study of the
concurrent validity with the Hamilton scale and Widlocher's retardation scale]. Encephale
1986;12:77-9.
[28] Coopersmith S, Inventaire d’estime de soi. 1984, Paris: Edition du Centre de Psychologie
Appliquée.
[29] Sullivan TR, White IR, Salter AB, Ryan P, Lee KJ. Should multiple imputation be the method of
choice for handling missing data in randomized trials? Statistical methods in medical research
2016:962280216683570.
[30] Cella DF, Tulsky DS. Measuring quality of life today: methodological aspects. Oncology
[31] Bonnetain F, Fiteni F, Efficace F, Anota A. Statistical Challenges in the Analysis of Health-
Related Quality of Life in Cancer Clinical Trials. Journal of clinical oncology : official journal of
[32] Baert I, Freeman J, Smedal T, Dalgas U, Romberg A, Kalron A, et al. Responsiveness and
Repair 2014;28:621-31.
[33] Salci Y, Karanfil E, Fil Balkan A, Calik Kutukcu E, Ceren AN, Ayvat F, et al. Functional Exercise
Capacity Evaluated by Timed Walk Tests in Myasthenia Gravis. Muscle Nerve 2018.
[34] Dobkin BH. Fatigue versus activity-dependent fatigability in patients with central or
22
[35] Kelley GA, Kelley KS. Dropouts and compliance in exercise interventions targeting bone
2013;2013:250423.
[36] Bello-Haas VD, Florence JM, Kloos AD, Scheirbecker J, Lopate G, Hayes SM, et al. A
2007;68:2003-7.
[37] Gellen B, Messonnier LA, Galacteros F, Audureau E, Merlet AN, Rupp T, et al. Moderate-
Haematology 2018;5:e554-e562.
[39] Sackett DL. Why Did the Randomized Clinical Trial Become the Primary Focus of My Career?
Tables
Table 1. Demographic, disease characteristics, medical history and activity of participants at baseline i.e.
all untraineda.
23
Prednisone dose (mg) 3.2 (4.3) 3.2 (4.3) 3.1 (4.4)
Immunosuppressor 28 (65%) 13 (57%) 15 (75%)
Comorbidities
Thyroid dysfunction 7 (16%) 4 (17%) 3 (15%)
Osteoporosis 10 (23%) 6 (26%) 4 (20%)
Hypertension 8 (19%) 5 (22%) 3 (15%)
Type 2 diabetes mellitus 4 (9%) 2 (9%) 2 (10%)
Accelerometer data 37 22 15
Days worn 6 [5 to 7] 7 [5 to 7]
Average hours worn per day 14.2 (0.8) 14.5 (0.8)
Minimum hours worn per day 13.1 (12 to 15) 13.2 (12 to 16)
Maximum hours worn per day 15.3 (14 to 16) 15.5 (14 to 16)
Average activity (count per minute 446 (355 to 553) 566 (354 to 696)
per day)
a
Data are mean (SD) and range or n (%) or median [IQR].
SD, standard deviation; IQR, interquartile range; BMI, body mass index; MGFA, Myasthenia Gravis Foundation
of America classification; AChR, muscle nicotinic acetylcholine receptor; MuSK, muscle-specific tyrosine
kinase; AchE, acetylcholinesterase.
Table 2. Intention-to-treat analysis of the primary outcome. Analyses of secondary outcomesa.
Exercise Control
Outcome n Mean (SD) n Mean (SD) Adjusted MD (95% CI) p value
MGQOL-15-F score
M3 23 20.0 (8.8) 20 24.6 (8.7) … …
M6 23 18.8 (10.7) 20 23.4 (9.6) -0.8 (-5.4 to 3.7) 0.72
M9 23 19.9 (12.4) 20 20.2 (9.3) 3.9 (-1.7 to 9.5) 0.17
Myasthenia Muscle Score
M3 23 85.3 (10.6) 20 88.0 (11.5) … …
M6 23 85.4 (13.7) 18 87.1 (13.3) 1.1 (-4.9 to 6.8)* 0.58**
M9 21 85.7 (13.1) 18 88.5 (12.8) -1.1 (-6.3 to 4.3)* 0.80**
MG-ADL
M3 23 2.7 (2.6) 20 2.4 (2.2) … …
M6 23 1.6 (1.3) 19 3.4 (2.9) -1.9 (-3.0 to -0.9)* 0.005**
M9 22 2.5 (2.8) 18 2.8 (2.8) -0.3 (-2.0 to 1.4)* 0.35**
Handgrip strength (kg)
M3 23 23.8 (4.7) 20 26.4 (8.0) … …
M6 23 25.1 (5.8) 19 26.4 (8.2) 1.3 (-1.2 to 3.9) 0.31
M9 23 24.8 (4.9) 17 26.0 (9.5) 1.4 (-0.9 to 3.7) 0.25
Elbow flexion (Nm)
M3 23 36.8 (7.1) 20 39.2 (11.5) … …
M6 23 37.6 (8.3) 19 40.8 (11.7) -0.8 (-3.5 to 1.9) 0.57
M9 23 37.0 (8.2) 17 40.6 (14.9) -1.3 (-4.8 to 2.1) 0.45
Knee extension (Nm)
M3 23 107.5 (27.9) 20 115.5 (43.8) … …
M6 23 114.0 (33.3) 19 115.7 (45.9) 8.8 (0.1 to 17.5) 0.06
M9 23 109.3 (33.0) 17 118.1 (48.6) -4.0 (-18.6 to 10.6) 0.60
6-minute walking distance (m)
M3 22 502.1 (72.8) 20 493.8 (95.5) … …
M6 23 516.8 (77.8) 19 486.0 (88.2) 27.7 (7.2 to 48.1) 0.01
M9 23 508.0 (80.1) 16 499.0 (92.5) -7.5 (-29.8 to 14.9) 0.52
Forced Vital Capacity (%pred)
M3 23 82.1 (10.8) 20 87.4 (15.2) … …
M6 23 80.7 (10.6) 19 84.9 (15.0) 0.1 (-2.6 to 2.9) 0.93
M9 23 79.8 (13.7) 18 83.5 (15.3) 0.7 (-3.6 to 5.0) 0.75
Maximal Inspiratory Pressure (%pred)
M3 23 88.7 (33.1) 20 98.7 (28.9) … …
M6 22 90.7 (36.3) 19 98.5 (35.0) -0.7 (-11.2 to 9.8) 0.90
M9 23 92.7 (35.9) 17 93.4 (34.5) 5.3 (-6.1 to 16.6) 0.37
Maximal Expiratory Pressure (%pred)
M3 23 89.2 (30.1) 20 104.2 (31.4) … …
24
M6 23 97.0 (31.3) 19 103.0 (31.5) 2.6 (-10.7 to 15.9) 0.71
M9 23 96.6 (33.8) 17 101.5 (32.8) 4.4 (-7.6 to 16.5) 0.47
a
Analyses of imputed data of primary efficacy endpoint (MGQOL-15-F: French version of the 15-item MG-
specific quality of life scale). Secondary efficacy endpoints; symptom severity (Myasthenia Muscle Score), limb
strength, walking capacity (six-minute walking distance), pulmonary function (Forced Vital Capacity, Maximal
inspiratory pressure, Maximal expiratory pressure). %pred, % predicted; SD, standard deviation; MD, mean
difference; 95% CI: 95% Confidence Interval; M3, Randomization, at month 3; M6, end of intervention period,
at month 6; M9, end of study at month 9. *Adjusted mean difference estimated by a latent class mixed-effect
models, and 95% confidence interval by bootstrap; **p-value obtained by rank ANCOVA.
Table 3. Analysis of tolerance to exercise based on neurologist assessment at each monthly visit.
Exercise Control
n No. (%) n No. (%) RD (95% CI) RR (95% CI) p value
MG Exacerbations
No. (%) until M6 23 0 (0) 19 3 (16) -16% (-38 to 2) 0 (0 to 0.98) 0.084
No. (%) until M9 23 0 (0) 18 5 (28) -28% (-51 to -7) 0 (0 to 0.55) 0.011
Cardiac, respiratory, articular and
muscular symptoms up to M6
Cardio-thoracic pain — no. (%) 19 6 (32) 14 2 (14) 17% (-13 to 42) 2.21 (0.62 to 8.84) 0.42
ECG abnormalities — no. (%) 21 3 (14) 17 0 (0) 14% (-6 to 35) ∞ (0.70 to ∞) 0.24
Dyspnea exertional — no. (%) 23 5 (22) 19 7 (37) -15% (-40 to 12) 0.59 (0.23 to 1.51) 0.32
Articular pain 20 9 (45) 17 11 (65) -20% (-46 to 12) 0.70 (0.37 to 1.27) 0.32
Muscular pain 20 11 (55) 17 11 (65) -10% (-37 to 21) 0.85 (0.49 to 1.48) 0.74
RD: absolute risk difference; RR: relative risk; 95% CI: 95% Confidence Interval; M6, month 6; M9, month 9.
Table 4. Patient-reported adverse events throughout the entire study as reported by the participant at
each monthly visit. Serious adverse events reported at any time throughout the study and declared to the
sponsor.
25
26