Cognitive performance in patients with Myasthenia Gravis: an association with glucocorticosteroid use and depression

ABSTRACT. We investigated the cognitive performance of patients with Myasthenia Gravis (MG) through a cross-sectional study. A battery of cognitive assessments and self-report questionnaires regarding quality of life (QoL), sleep, and depression were applied. The sample consisted of 39 patients diagnosed with MG. The scores showed a predominance of cognitive impairment in the Montreal Cognitive Assessment screening test (MoCA) (66.7%) and in the immediate (59.0%) and recent memory (56.4%) tests. However, after the Poisson regression analysis with robust variance, it was found that patients diagnosed with depression had a prevalence ratio (PR) of 1,887 (CI 1,166‒3,054) for lower MoCA scores, PR=9,533 (CI 1,600‒56,788) for poorer phonemic verbal fluency scores, and PR=12,426 (CI 2,177‒70,931) for the Semantic Verbal Fluency test. Moreover, concerning a decline in short-term memory retention, patients using glucocorticosteroids (GC) and with Beck Depression Inventory scores indicating depression showed PR=11,227 (CI 1,736‒72,604) and PR=0.35 (CI 0.13‒0.904), respectively. No correlation was found between the QoL questionnaire and performance in cognitive tests. We found worse performance in tasks of memory and executive functions in MG patients. These are not associated with the length and severity of the disease. However, a significant prevalence ratio was found for poorer memory performance in patients diagnosed with depression and in those using GC.

D ementia constitutes a multifactorial process 1 that is always associated with cognitive decline and impaired functioning. As the disease progresses, people living with dementia experience, in addition to impaired cognitive functions, gradual dysfunction and loss of individual autonomies. Besides decline in memory and/or other cognitive domains, the criteria for diagnosis of dementia require loss of functional reserve and pejoration in functional status. 2 An important quality of life component from elderly people's perspective is functional indepen-dence. When older people show functional loss, they experience a variety of negative outcomes, such as higher rates of use of hospital services, institutionalization, and increased risk of death. 3 The progression of healthy aging to dementia must be considered a continuum, both in terms of the slow manifestation of the impairment of cognitive functions, as well as functional limitation. 4 Originally, mild cognitive impairment (MCI) was considered a condition in which someone has minor cognitive decline, not severe enough to interfere significantly with daily life and

INTRODUCTION
Myasthenia Gravis (MG) is an autoimmune disease caused by the destruction of nicotinic acetylcholine receptors (nAChRs) at the motor end plates found in striated muscles. Its incidence ranges from 1 to 9 cases per million of the general population. The prevalence rate of MG ranges from 15 to 179 cases per million of the worldwide population. There are no data regarding the national prevalence of the disease in the Brazilian population. 1,2 Although it is a predominantly muscular disease, cognitive impairment in patients with MG has been discussed in the literature. Some studies found cognitive decline in memory, [3][4][5][6][7][8][9] attention, executive functioning, 8,10 verbal fluency, 8,9 and planning tasks. 6,11 However, there are other studies 12-15 that found no difference in the cognitive performance of MG patients when compared to healthy controls.
In a recent systematic review and meta-analysis, 16 the authors described four main explanations for the cognitive deficits found among many MG patients: • central pathogenic antibody effect (Abs) against acetylcholine receptors (AChRs); • for some patients, a lack of certain protective factors such as age, disease severity, and type of treatment; • mood disturbances; • the possible effect of nonspecific immunological processes.
Therefore, while some studies have shown a tendency toward cognitive decline in patients with MG, conflicting results have also been published. Thus, considering this lack of clarity in the literature and absence of studies in the Brazilian population, the aim of this study was to investigate the cognitive performance of patients with MG and its association with clinical aspects and quality of life (QoL) in patients with MG.

Study design
This was a cross-sectional, exploratory study.

Subjects
Patients were recruited from a neuromuscular diseases outpatient clinic at a hospital in Porto Alegre, Brazil, which is a reference in the treatment of patients diagnosed with MG in the state. Diagnosis of the disease was confirmed by electromyography and/or by the presence of AChR/Musk/Striated Muscle antibodies. Patients were evaluated outside of crisis episodes and on medication, who had been stabilized for at least 6 months. Patients with a history of other primary neurological (e.g. transient ischemic attacks, cerebrovascular stroke or epilepsy), psychiatric (e.g. major depression), previous serious head injury or any sensory or motor disorder that would preclude psychological testing (such as blindness or deafness). We excluded The collection period took place between February 2017 and December 2018. All subjects were informed about our research objectives and signed an Informed Consent form. This study was approved by the Central Research Ethics Committee of the hospital, certificate of approval number 120399.

Procedures
All patients were assessed individually in a room at the hospital's research center. All patients were evaluated in the afternoon, between 12 and 4pm, and the last dose of pyridostigmine was not registered. The duration of a complete test per patient lasted an average of 40 minutes. This evaluation was always performed by the same previously trained researcher. All instruments and questionnaires used have been translated and validated to suit the Brazilian population.

Questionnaires
• Sociodemographic questionnaire: a structured questionnaire used to gather general patient data, such as age, gender, education, length of illness, age of diagnosis, initial symptoms, and marital status; • MG quality of life scale (MG-QOL 15): a self-report questionnaire specifically designed to assess the quality of life of patients with MG. It has 15 items, with scores ranging from 0 to 60 points. The higher the score, the worse the patients' perception of quality of life; 17 • Beck Depression Inventory (BDI): a self-assessment tool used to survey the intensity of depressive symptoms. 18 To determine each score, the values proposed by Gorenstein and Andrade 19 were used: less than 10 points -no depression or minimal depression; 10 to 18 points -mild to moderate depression; 19 to 29 points -moderate to severe depression; 30 to 63 points -severe depression; • Epworth Sleepiness Scale: an 8-item self-report questionnaire which assesses the likelihood of falling asleep in eight situations involving daily activities. The overall score ranges from 0 to 24; scores above 10 suggest excessive daytime sleepiness (EDS). 20 Motor scales In addition, patients who were prescribed antidepressants and/or had reports of depression in their medical records were quantified.

Statistical analysis
Statistical tests were selected according to the distribution of data provided by the Shapiro-Wilk test and histograms. Continuous variables were described using the terms minimum, maximum, mean and standard deviation. The scores found in the cognitive tests were described as percentages of normal and impaired, according to the cutoff points validated for the Brazilian population. Categorical variables were described by percentage and N.
Association analysis was performed between outcome (cognitive results categorized as normal or impaired) and contextual variables (e.g. medications, other associated diseases, clinical diagnosis of depression, gender, marital status, BDI score and Epworth score ) using the Fisher's exact test, except for the MG clinical classification, for which the Pearson's chi-square test was used. Subsequently, with associations established at p≤0.2, the Poisson regression with robust variance was used. The linearity of the quantitative variables was analyzed and it was found that the assumption of linearity was maintained. In addition, the presence of multicollinearity was assessed using the variance inflation factor (VIF) estimates, noting that the cutoff points are good (close to 1), indicating that the variables are not multicollinear. The statistical significance of the odds ratio indices was assessed using the Wald test. The model's adjustment was assessed using the Hosmer and Lemeshow test. Also, a Pearson correlation was performed between cognitive test scores, contextual variables, and questionnaires based on their gross values, except for the correlation between the MMSE and MoCA tests. For these two, the Spearman's correlation test was applied. The cognitive scores of patients with and without thymomas were compared, by means of the Student's t-test, to verify the possible influence of thymomas on cognitive performance. The statistical significance level adopted was p<0.05.

RESULTS
Eighty-eight patients with MG were initially included; of these, 49 patients were excluded for the following reasons: 39 did not come to the scheduled assessment date, eight did not want to participate in the study, one was hospitalized, and one was under 18 years of age. The final sample of this study consisted of 39 subjects diagnosed with generalized MG. Sociodemographic data are presented in Table 1.
Regarding the cognitive battery, there was a predominance of impairments, (according to the cutoff points specific for the Brazilian population) in the MoCA screening test (66.7%) and in the subtasks of immediate (59.0%) and recent memory (56.4%) in the RAVLT test. Regarding the self-perception questionnaires, 23.1% of patients presented scores which suggested EDS, based on data from the Epworth questionnaire, and 41.02% presented scores suggestive of depression, according to the BDI (Table 1).
Correlation analyses were performed between the gross scores of the cognitive tests, clinical variables, and questionnaire scores, as presented in Tables 2 and  3. Positive correlations were found between all cognitive tests and level of education, showing that the higher the education, the higher the test scores. Age correlat-  lower scores on the RAVLT subtask that assesses shortterm memory (A6). Therefore, there is a significant PR for the presence of cognitive deficits in patients with depression and who used GC. No association was observed between the MGFA clinical classification scores and cognitive tasks (Tables 4 and 5).

DISCUSSION
To the best of our knowledge, so far, this study has been the first to investigate cognitive performance in patients with MG in the Brazilian population. The results of this study showed worse performance in tasks related to memory in patients with MG. Moreover, this change was associated with depression and the use of GC. These data corroborate the findings of the systematic reviews by Mao et al. 16 and Paul et al., 1 in which the authors point out that, although there are several studies also pointing to a cognitive decline in MG, they did not exclude the possibility that cognitive function may have been affected by other aspects such as sleep apnea, depression and Type 1 drug use. In their review, Paul et al. 1 already mentioned the adverse effect of high doses of drugs, such as prednisone, as well as depression on the cognitive functioning of patients with MG. In addition, our results show cognitive decline in the same functions highlighted by other studies, such as memory 5-10 and executive functions. 8,10 There are few studies in the specialized literature that analyzed the interference of MG medication, depression and EDS on cognitive performance. Three studies 9-11 found a higher incidence of cognitive impairment in patients with depression or scores suggestive of depression in self-perception questionnaires. Three other studies describe an analysis of cognitive performance and the use of MG medication. Bartel and Lotz 29 published a paper on a possible association between medications and cognitive impairment. In a linear regression analysis, Marra et al. 14 found that longer treatment time with CG seemed to be correlated with better performance on attentional tasks and long-term verbal memory, contrary to the evidence in our sample. Interestingly, Jordan et al. 15 found no association at all between the use of acetylcholinesterase inhibitors and performance in cognitive tests.
Additionally, with our sample, we investigated whether the presence of thymomas could influence patients' cognitive performance. Our data corroborated other studies 14,15 which found that they did not influence cognitive performance.

Memory decline and glucocorticosteroid use
Regarding the association between impairments in short-term memory retention and the use of GC, several studies [30][31][32][33][34] have been found on other clinical populations (e.g. patients with asthma, rheumatoid arthritis, kidney transplants, and non-CNS systemic diseases), in which participants showed significantly worse performance in memory and attention tests when compared to the control group. Also, research 35,36 on healthy volunteers, with no history of systematically prescribed corticosteroid therapy, noted a significant reduction in performance in memory-related tasks after first-time prednisone use.
Nevertheless, it bears stressing that, while there is evidence of cognitive impairment after the use of GC  by MG patients, only two studies have investigated this association. Even so, data were not conclusive. Additionally, they investigated only prednisone 14 and AI. 15 The pathophysiology of the adverse effects of using synthetic GC is still unclear. 30,35,37 Therefore, the risk of memory impairment should be considered before starting treatment with GC and in monitoring patients with MG. Thus, with the other forms of monitoring already routinely performed on MG patients treated with GC, cognitive aspects should also be taken into consideration -such as memory -in addition to psychiatric symptoms such as depression 31 and the importance of a proactive approach on the part of health professionals who follow patients with MG, as these patients may not have self-perception of cognitive changes. Therefore, they should be investigated regardless of patients' complaints.

Losses in cognitive function associated with depression
In our sample, there was a significant prevalence ratio of depression in participants who showed cognitive impairment on screening tests, where verbal fluency and memory are concerned. These results corroborate the data in the literature that patients with depression present cognitive decline, regardless of severity (i.e. whether mildly or severely depressed). Studies have shown poorer performance in: memory tests (on tasks such as coding, retrieval, recall, and recognition), [37][38][39][40][41] sustained and/or selective attention, psychomotor deceleration (such as reaction time, information, writing, and drawing tasks), 38,39,42 verbal fluency, 37,39,40,43 and executive function. 38,39,41,42 The causes of these deficits can be explained by three different theories: 1) the stress hypothesis, which proposes that performance in stress tasks is disproportionately impaired in depressed patients when compared to their performance in automatic ones; 2) the cognitive velocity hypothesis, which states that depression is characterized by cognitive slowness and that this deceleration may be at the root of other cognitive impairments, and 3) the hypothesis of impairment of executive control functions which, in turn, is underlying to the hypothesis of effort. 44 The present study fortifies theories 1 and 2 of these authors. 44 Limitations of the study The cross-sectional and exploratory design of the study presented limitations, since it did not allow for analysis of the causal factors of the cognitive decline found in the sample. Thus, we identified a need for longitudinal studies that could explain whether cognitive impairment is due to the pathophysiology of the disease or associated with other clinical aspects.
Another limitation of the study is that it was carried out at a reference public hospital in the treatment of MG patients. This may have led more severe patients to our recruitment universe, as they require more specialized care. As such, the representative power of the sample may have been reduced. Moreover, we cannot analyze the correlation between antibodies and cognitive performance, as the test was not available in the public health system In addition, considering the known differences related to education and socioeconomic levels of other populations, the scarcity of studies on the Brazilian population with MG did not allow for a comparison between the scores found in this sample.
In this sample, participants with MG presented worse performance in tasks of executive function and immediate and recent memory. These are not associated with the time and severity of the disease. However, a significant prevalence ratio was found for poorer memory performance in patients diagnosed with depression and in those using GC. Regarding QoL, only the motor scale showed a positive correlation, suggesting that patients with a poorer perception of quality of life also suffered from more severe motor restrictions.