Fatigue in myasthenia gravis: risk factors and impact on quality of life

Abstract Objectives Emerging evidence suggests that fatigue in myasthenia gravis (MG) is a relevant problem that negatively impacts activities of daily living (ADL). The relationship between fatigue and quality of life (QoL) has never been systematically explored in MG patients. The study aimed to assess the prevalence of fatigue and its relation to ADL and QoL as well as to identify factors associated with fatigue in MG. Material and Methods This was a cross‐sectional observational study in patients with confirmed diagnosis of MG independent of disease severity. Prevalence of fatigue was assessed using the Chalder Fatigue Scale (CFQ). Impact of fatigue on ADL and QoL was assessed by the MG activities of daily living profile (MG‐ADL) and the MG‐specific quality‐of‐life instrument (MG‐QoL), respectively. Association of fatigue with sociodemographics, clinical characteristics of MG, and comorbidities including mood and anxiety disorders as well as sleep disorders was investigated using multivariable logistic regression analyses. Results Overall, 200 MG patients were included. The observed rate of fatigue was 56.1%, of those 70.4% fulfilled the criteria of chronic fatigue (CF) with a duration of ≥6 months. Relevant fatigue was strongly associated to ADL and QoL. Factors associated with relevant fatigue were disease severity and depressive state. Furthermore, positive muscle‐specific tyrosine kinase (MuSK) antibody status showed a strong association with relevant fatigue. Conclusions MG patients have a high prevalence of fatigue which negatively impacts ADL and QoL. MG‐specific clinical characteristics are related to fatigue and might help to identify MG patients at risk for fatigue.

prevalence rates between 75% and 89% in MG patients (Kluger, Krupp, & Enoka, 2013) and qualitative data suggest that, in some patients, fatigue has a greater impact on daily living of MG patients than has muscle weakness (Barnett, Bril, Kapral, Kulkarni, & Davis, 2014;Zwarts, Bleijenberg, & van Engelen, 2008). However, overall literature on fatigue in MG is scarce and its impact on activities of daily living (ADL) and quality of life (QoL) has never been systematically explored. This might be due to various reasons. Fatigue is a complex, nonspecific, and highly subjective symptom and therefore difficult to evaluate and quantify (Krupp, LaRocca, Muir-Nash, & Steinberg, 1989;Norheim, Jonsson, & Omdal, 2011). The fluctuating and effortdependent nature of fatigue makes it even more difficult to separate fatigue from muscle fatigability in MG. We follow recent proposals to use the term fatigue to refer to subjective sensations of exhaustion and muscle fatigability to refer to objective changes in muscle performance (Kluger et al., 2013). Other related phenomena such as depression and sleep disorders need to be distinguished from fatigue and should therefore be included as covariates when assessing fatigue (Kluger et al., 2013). Finally, the understanding of the pathophysiology of fatigue is limited. MG is an autoimmune-mediated disease with autoantibodies directed against components of the postsynaptic muscular endplate (Szczudlik et al., 2014). The most likely confinement to the peripheral nervous system makes hypotheses on the pathophysiology of fatigue in MG particularly challenging.
The aim of the present study was to assess the prevalence of fatigue and its impact on ADL and QoL in a large cohort of MG patients as well as to identify factors associated with fatigue including MGspecific clinical characteristics as well as potential confounders such as mood and sleep disturbances (Elsais et al., 2013;Kluger et al., 2013).

| Patients
This was a cross-sectional, observational study performed at the certified Integrated Center for Myasthenia gravis (IMZ) of the Charité -Universitätsmedizin Berlin, Germany. Patients over the age of 18 years with confirmed diagnosis of myasthenia gravis were included independent of disease duration and severity (excluding myasthenic crisis). Patients were consecutively screened at the IMZ clinic between December 2012 and December 2013.
Sociodemographics (age, sex), current MG-specific medication (cholinesterase inhibitors, glucocorticoids, and long-term immunosuppressants), current comedication (antidepressants, NSAIDs, and opioids), and comorbidities (other immunopathies, cardiac insufficiency, and malignancies) were collected in a database. Clinical assessment was performed using the MGFA classification for disease classification (Jaretzki et al., 2000) and the QMG score for disease severity. Using the MGFA classification, patients were grouped into ocular (MGFA I) or generalized MG patients (MGFA II-IV) at the time of study inclusion. According to current recommendations, we used the MGFA classification by employing the most severely affected muscles to define the patient's MGFA class (the "maximum severity" designation being made historically; Jaretzki et al., 2000). Thereby, patients were only assigned to MGFA I if symptomatology was restricted to purely ocular symptoms throughout the patient's entire disease history (ocular MG).
Correspondingly, patients with generalized symptoms (affecting limb/ axial and/or bulbar/respiratory muscles) throughout their disease history were assigned to generalized MG. An exception to this approach was made in patients without any detectable myasthenic weakness under current medication who were assigned to "pharmacological remission."

| Questionnaires
All patients completed the self-assessment questionnaires at the IMZ. Completion of questionnaires required approximately 30 min.
If available, validated German versions of questionnaires were used, as in the case of the Chalder Fatigue Scale (CFQ; Martin, Gaab, Rief, & Brähler, 2010) and the Hospital Anxiety and Depression Scale (HADS; Hinz & Schwarz, 2001). The other questionnaires (Insomnia Severity Index, MG activities of daily living profile, and MG-specific quality-of-life instrument) were translated using forward-backward translation procedure by two physicians (SH, JR) who are both native German with excellent knowledge of English. The translations were reviewed by the principal investigator (AM). In case of discrepancies, a reconciled version was produced by a consensus conference.

| Chalder Fatigue Scale
For assessing fatigue, the CFQ presented by Chalder et al. (1993) was used. The CFQ is a self-administered questionnaire for measuring the severity of physical (CFQ-P) and mental (CFQ-M) fatigue within both clinical and epidemiological populations (Jackson, 2015).
It consists of 11 items that are answered on a 4-point scale ranging from the asymptomatic to maximum symptomatology, such as "less than usual," "no more than usual," "more than usual," and "much more than usual" offering the option of binary scoring (0 0 1 1) or Likert scoring (0 1 2 3). We used the binary scoring (total score 0-11) to define caseness in patients with a score of 4 points or higher as proposed by Chalder et al. and

| MG-specific quality-of-life instrument
MG-specific quality-of-life instrument (MG-QoL) is a 15-item questionnaire encompassing physical and psychological domains of MG to assess disease-specific QoL in MG patients (Burns, Conaway, Cutter, & Sanders, 2008). Rating consists of a 5-point scale ranging from 0 ("not at all") to 4 ("very much") as to the degree to which patients agree with the given statement summing up to a total score of 0-60 points.

| MG activities of daily living profile
Impact on daily living was assessed using the MG activities of daily living profile (MG-ADL; Wolfe et al., 1999). It is an eight-question survey of symptom severity, with each response graded from 0 (normal) to 3 (most severe). Questions include ocular, oropharyngeal, respiratory, and extremity functions. Total MG-ADL score ranges from 0 to 24.

| Hospital anxiety and depression scale
Hospital anxiety and depression scale is a self-assessment mood scale developed to identify cases of anxiety disorders and depression among patients in nonpsychiatric hospital clinics (Zigmond & Snaith, 1983). It is divided into an anxiety subscale (HADS-A) and a depression subscale (HADS-D) both containing seven items (Bjelland, Dahl, Haug, & Neckelmann, 2002) with a 0-to 3-point Likert response format summing up to a total score of 21 for each subscale. A cut-off score of 8 or higher on each subscale was used to define caseness of anxiety and depression, respectively (Bjelland et al., 2002). Symptoms of mood disorders likely to be present in somatic disorders (e.g., insomnia, fatigue, anorexia) are excluded in the HADS.

| Insomnia Severity Index
Sleep disturbances were assessed using the ISI. It is a seven-item selfreport instrument targeting the subjective symptoms and consequences of insomnia (Bastien, Vallieres, & Morin, 2001). Using the Likert scale, each item is rated between 0 to 4 points summing up to a total score range from 0 to 28. We used a cut-off score of 10 points to define caseness of sleep disturbances (Morin, Belleville, Belanger, & Ivers, 2011

| Statistical analysis
Descriptive analyses included reports of means and standard deviation, median and interquartile range (IQR), or absolute and relative frequencies depending on the scale and distribution of variables. We compared sociodemographic characteristics, questionnaire scores, and clinical characteristics between the myasthenia subgroups using Fisher's exact test, chi-square test, one-way ANOVA, or Kruskal-Wallis test where indicated (Tables 1 and 2). To analyze association of characteristics with clinical-relevant fatigue, we used t-test, Mann-Whitney test, Fisher's exact test, or chi-square test (Table 3). In multiple logistic regression analysis we included and tested variables that were associated with relevant fatigue in univariate analysis (p < .05).
As all six MuSK-ab-positive patients had relevant fatigue, the problem of quasi-separation appears resulting in nonexistence of OR and CI estimates. Therefore, Firth logistic regression was used (Stata/IC 13.1 command firthlogit), estimating ORs and CIs with penalized likelihood estimation methods (Heinze & Schemper, 2002). Only significant variables remained in the final model with the exception of MuSK-ab which was forced into the model. We used a two-sided significance level of alpha = 0.05. Linear regression was used to analyze association of relevant fatigue with ADL and QoL adjusted for MG type (one regression model for each outcome, Fig. 1 Age, year, mean (SD) 58 (17) 56 (17) 60 (15) 61 (17) .235 Age at disease onset, mean (SD) 48 (20) 45 (21) 53 (17)

| Ethics
The study was approved by the ethics committee of the Charité -Universitätsmedizin Berlin (EA1/281/10). All patients gave written informed consent in accordance with the Declaration of Helsinki in its currently applicable form.

| Fatigue prevalence and severity
Relevant fatigue (CFQ ≥ 4 points on the binary scale) was present in 110 (56.1%) of MG patients (

| Prevalence of mood, anxiety, and sleep disorders
Using a cut-off score of ≥8 points on the HADS, rates for depression and anxiety in the total sample are 19.6% and 27.8%, respectively (

| Factors associated with relevant fatigue: univariate and multivariate analysis
The following characteristics were associated with relevant fatigue in univariate analysis (Table 3): female sex, GMG, positive AchR or MuSK antibody status, medication with cholinesterase inhibitors or opioids, thymectomy, presence of other immunopathies, higher QMG score, higher HADS-D and HADS-A scores, and higher ISI score.
Multivariable logistic regression showed that higher QMG score and higher HADS-D score were significantly associated with relevant fatigue after adjustment for each other and for positive MuSK-ab status (Table 4). Positive MuSK-ab status showed a strong association with relevant fatigue, although due to the low case number this finding did not reach statistical significance.  Consistently, a higher QMG score was associated with a higher prevalence of relevant fatigue. It is noteworthy, that the QMG score is an   We included measures of mood and sleep disturbances as covariates in our analyses as recommended by other research groups (Elsais et al., 2013;Kluger et al., 2013). Depressive state was associated with higher prevalence of fatigue. The association between fatigue and depression is widely accepted and mainly attributed to the overlap in symptomatology between fatigue and depression (such as physical fatigue; Norheim et al., 2011). Rate of depressive state was 19.6% in our MG cohort and thereby comparable to those in the general German population (Hinz & Brahler, 2011). It has long been suggested that depression is common among MG patients (Chafetz, 1966;Meyer, 1966). However, more recent studies show inconsistent results. Although some studies report markedly higher depression rates of up to 33% in MG patients (Fisher, Parkinson, & Kothari, 2003;Magni et al., 1988), others have found no increased frequency of depression in MG patients over the general population (Doering, Henze, & Schussler, 1993;Tennant, Wilby, & Nicholson, 1986 -Pruszczyk et al., 2008;Niks, Kuks, & Verschuuren, 2007). Attempts have been made to better characterize MG patients by their antibody status and differences were found for patterns of involved muscles, disease severity, as well as treatment responses (Ohta et al., 2007). Interestingly, all of the MuSK-positive MG patients reported relevant fatigue. Due to the low case number, this finding needs to be interpreted with caution but MuSK-ab and their association with relevant fatigue seem to be a promising target for further investigations. Future studies on fatigue in MG should furthermore consider newly described antibodies such as anti-LRP4 and anti-Agrin antibodies (Gasperi et al., 2014).

| Fatigue and its association with quality of life and activities of daily living
Our study has several limitations that need to be acknowledged.
The study design did not include a healthy control group to compare our data with. However, the herein used CFS allows for comparisons of fatigue prevalence in the general population (Martin et al., 2010) and other MG cohorts (Elsais et al., 2013).  (Braley, Segal, & Chervin, 2015;Tang, Yu, & Yeh, 2010).
The test battery with five self-assessment questionnaires might have influenced the results. Patients were free to decide on the order of questionnaire completion, therefore, we cannot entirely exclude an overestimation of mental fatigue due to prior concentration on completion of other questionnaires. However, increasing emphasis has been placed on self-reporting instruments as they allow for an assessment of highly subjective conditions such as fatigue as well as for an evaluation of the impact of diseases such as QoL and ADL (Wolfe et al., 1999).
In conclusion, our study emphasizes the importance of fatigue in MG even in patients in pharmacological remission and its impact on ADL and QoL. MG-specific clinical characteristics seem to have

FUNDING INFORMATION
This study was supported by the Deutsche Forschungsgemeinschaft (Grant/Award Number: Exc 257).