Neurol Sci (2009) 30:9–14

DOI 10.1007/s10072-008-0001-y

ORIGINAL ARTICLE

Neuropsychological assessment in myasthenia gravis

Emilia J. Sitek Æ Małgorzata M. Bilińska Æ
Dariusz Wieczorek Æ Walenty M. Nyka

Received: 22 February 2008 / Accepted: 10 October 2008 / Published online: 16 January 2009
Ó Springer-Verlag 2009

Abstract Neuropsychological studies in myasthenia Introduction

gravis (MG) were undertaken to prove the central nervous
involvement. However, they still produce contradictory Myasthenia gravis (MG) is an autoimmune disorder of
results. In the present study, a battery of cognitive mea- neuromuscular transmission characterized by increased
sures was administered to examine global cognitive fatigability of voluntary muscles [1]. Central nervous sys-
functioning, verbal learning, attention, executive function tem (CNS) involvement in MG has been suggested since
and motor performance. Analysis of partial scores in verbal the disease was described for the first time. At least some
learning and response fluency trials did not reveal fatigue of the subsequent studies, such as reports of increased
effect in MG patients. It was shown that in tasks requiring coexistence of MG and other neurological disorders such as
motor, and particularly oculomotor, involvement, the multiple sclerosis [2], Meige syndrome [3], and epilepsy
muscle fatigue could account for the deficits observed. [4], as well as reports of higher frequency of left-handed-
Thus, impaired performance on some cognitive measures ness in MG [5] seemed to confirm these preliminary
in MG should be interpreted as an effect of muscle fati- suggestions. As a result, further research [6] focused
gability rather than central nervous system involvement. mainly on the detection of electroencephalographic and
evoked response abnormalities [7, 8], rapid eye movement
Keywords Myasthenia gravis
Cognitive
Motor speed
sleep disturbance [9] and the presence of specific muscle
Fatigue
Neuropsychological assessment nicotinic acetylcholine receptor (AChR) antibodies in the
cerebrospinal fluid (CSF) [10]. With those methods used,
no clear evidence of CNS involvement has been found.
Due to scarcity of biological data supporting the afore-
mentioned hypothesis, some authors such as Tucker et al.
E. J. Sitek [11], resorted to cognitive studies as an alternative—
Department of Neurological and Psychiatric Nursing, although indirect—way to prove the presence of central
Medical University of Gdańsk, ul. Do Studzienki 38, cholinergic deficits in MG patients.
80-227 Gdańsk, Poland
Since Mertens et al.’s [12] first neuropsychological
E. J. Sitek (&) study of MG, four alternative explanations of the possible
Department of Neurology, St Adalbert Hospital, neuropsychological deficits in MG have been formulated.
Al. Jana Pawła II 50, 80-462 Gdańsk, Poland Neuropsychological impairment has been interpreted as (1)
e-mail: emiliasitek@amg.gda.pl
a central cholinergic deficit manifestation [11, 13–19], (2)
M. M. Bilińska
W. M. Nyka as a consequence of nocturnal respiratory problems [20],
Department of Neurology, Medical University of Gdańsk, (3) as a possible effect of nonspecific immunological pro-
ul. Debinki 7, 80-952 Gdańsk, Poland cesses [21], and (4) as a result of increased mental fatigue
[22].
D. Wieczorek
Department of Rehabilitation, Medical University of Gdańsk, The abundance of hypotheses generated or tested in
ul. Debinki 7, 80-952 Gdańsk, Poland those studies contrasted with their methodological

123
10 Neurol Sci (2009) 30:9–14

shortcomings such as insufficient sample size, lack of Patients averaged 47 years of age (SD 12; range 20–68),
common cognitive measures, lack of sample description in 12 years of education (SD 3; range 7–22), 8 years (SD
terms of the disease type and medication—reviewed in 7; range 0–29) since MG onset. According to Myasthenia
detail by Paul et al. [23] and Paul and Gilchrist [24]. Gravis Foundation of America (MGFA) Clinical Classifi-
The aim of the present study was to examine the cation, 4 patients were classified as Class 1, 17 patients as
cognitive function in MG patients through a detailed and Class II (IIA-15, IIB-2), 7 patients as Class III (IIIA-6, IIIB-
multidimensional analysis of performance on measures 1), 5 patients as IVB [25]. Nineteen patients were treated
frequently used to assess cognition in MG patients, such as with anticholinesterase inhibitors, 11 with both anticholin-
verbal learning tests and response fluency trials. We esterase inhibitors and glucocorticosteroids, two also took
hypothesized that there should be no impairment in global immunosuppressive medication (apart from aforementioned
cognitive functioning, memory and executive function, but therapy) and 1 received no pharmacological treatment. A
a possible fatigue-related decrement might appear in the total of 22 patients (67%) were thymectomized.
second part of performed tasks (4th and 5th trial in verbal Control group averaged 49 years of age (SD 12; range
learning test, the second part of each verbal fluency trial). 24–76), 13 years (SD 3; range 8–21) of education. Both
In some previous studies, motor fatigability could account groups were matched in terms of age (t = 0.7; P = 0.48),
for the observed deficient performance in cognitive testing sex (v2 = 0.65; P = 0.44) and years of education (t =
that was misattributed to CNS involvement. This hypothesis 0.97; P = 0.34), and presumed premorbid cognitive
was tested by means of comparing the results of Trail Making abilities (WAIS-R Similarities; Table 1).
Test (TMT) and Digit Span tests. If both test results were
deficient in MG group, it would indicate global attention or Procedure
working memory deficit. However, if the difference was
statistically significant only for TMT—which is dependent A neurological examination and a demographic interview
on near visual acuity and motor speed, it would clearly preceded the cognitive testing. The order of test adminis-
demonstrate that conclusions about CNS involvement in MG tration was the same for all patients: Mini Mental State
from previous studies relying on measures combining Examination (MMSE), Auditory Verbal Learning Test
cognitive, visual and/or motor requirements were method- (AVLT), Digit Span and Similarities from Wechsler Adult
ologically questionable. Amongst such measures are Benton Intelligence Scale Revised (WAIS-R), Trail Making Test,
Visual Retention Test [12] or Visual Reproduction from verbal fluency trials, Finger Tapping Test (FTT), Beck
Wechsler Memory Scale [11], Symbol Digit Modalities Test Depression Inventory (BDI) and MG Disability Scale.
(even in its oral version that still requires visual searching
and can be prone to dysarthria [21, 22]). Control of motor Measures
speed, with such methods as Finger Tapping Test, could
additionally clarify this interpretation. Global cognitive functioning was assessed by MMSE [26].
Instead of excluding patients with visual disturbance, as Similarities were used to control for premorbid cognitive
Paul et al. [23] recommended, we used methods that did function [27]. Verbal learning was tested with a modifi-
not rely on visual function or those in which the oculo- cation of AVLT [28] with five learning trials of a list of 15
motor involvement could be eliminated through the result words, and a delayed recognition trial, administered
analysis. 30 min after the last learning trial. In this trial, 30 words
were presented and the total score was computed by sub-
tracting false positive recognitions from the correctly
Methods recognized words. Taking into account a higher depression
level in MG patients, and the fact that spontaneous recall
Sample trials can be slightly impaired in depression probably due
to motivational problems [29, 30], delayed recall was
A sample of 33 individuals with MG and 30 healthy controls replaced in our study with delayed recognition, which is
volunteered to participate in the study. All participants gave supposed to be less prone to mood effect.
informed consent. The subjects who reported a history of Digit Span Test served as a measure of attention span
alcohol or drug abuse, or who had evidence of CNS lesions [31]. TMT Parts A and B [32] were used as a measure of
(confirmed by MRI or CT scans) were excluded from the attention dependent on psychomotor speed and visual
sample. All of the patients were recruited from Myasthenia function. Scaled score difference in TMT (B-A), that refers
Gravis Outpatient Clinic. Diagnosis was confirmed by to normative data, removes the speed and oculomotor factor
means of electromyography results, the presence of anti- from the test result and constitute an additional indicator of
nicotinic antibodies in serum and a positive Tensilon Test. working memory—more reliable than raw score.

123
Neurol Sci (2009) 30:9–14 11

Table 1 Performance on cognitive and mood measures in patients and controls

Measure Control group mean (SD) MG patients mean (SD) t/U P

Global cognitive functioning

Similarities 15.87 (5.40) 15.79 (5.54) 0.06 0.95 (ns)
MMSE 28.67 (1.45) 28.18 (1.40) 1.35 0.18 (ns)
Verbal learning and retention
AVLT—five trials (max. = 75) 47.83 (7.59) 46.97 (9.35) 0.40 0.69 (ns)
AVLT—confabulations 2.60 (2.74) 3.30 (3.08) -0.95 0.34 (ns)
AVLT-delayed recognition 13.70 (1.58) 13.61 (2.46) 0.18 0.86 (ns)
Motor performance
Tapping, dominant hand 46.13 (8.37) 37.72 (13.82) 2.94 \0.01
Tapping, non-dominant hand 41.33 (6.46) 34.88 (12.01) 2.63 0.01
Psychomotor speed
TMT A (s) 35.00 (10.07) 42.64 (15.92) -2.25 0.03
Working memory with a motor component
TMT B (s) 73.07 (25.93) 101.38 (59.83) -2.44 0.02
Working memory without motor component
TMT B–A (T score) 1.77 (6.94) 1.64 (9.59) -0.61 0.95 (ns)
Attention span
Digits forward 5.27 (1.05) 5.18 (0.88) 0.35 0.73 (ns)
Digits backward 4.03 (1.27) 3.85 (0.91) 0.67 0.51 (ns)
Executive function
Letter fluency, total 14.70 (4.10) 13.82 (4.30) 0.83 0.41 (ns)
Semantic fluency, total 22.40 (5.86) 20.61 (4.78) 1.34 0.19 (ns)
Mood
BDI 6.30 (5.95) 11.10 (6.29) -3.06 0.003

FTT served as a measure of motor performance [32]. variables, whilst r-Pearson was used for normally distrib-
Executive function was assessed by means of verbal flu- uted data.
ency. Two trials were administered, one for letter (K)
fluency and another one for semantic (animal) fluency.
Participants had 60 s to generate words according to the Results
given criteria. BDI [33] was used to assess mood. Finally,
the translation of MG disability Scale [34] was used as a In the MG disability scale individuals with MG averaged
subjective measure of daily function impairment in the 4.70 ± 3.12 (range 0–12). As shown in Table 1, MG
patient group. The scale contains six items assessing patients performed significantly poorer than the control
superior and inferior limb fatigability, swallowing, voice, subjects in FTT and TMT (parts A and B, respectively).
eyesight and breathing. The global score ranges from 0 to Other differences observed in cognitive performance were
18, where 18 indicates maximum severity. not statistically significant.
Time since MG onset was a period of time since the first MG patients exhibited an elevated depression score in
MG symptoms were noticed by patients, but not the time relation to control subjects. However, the mean score on
since MG diagnosis. We included this measure, apart from BDI for the patient group did not significantly exceed the
time since MG diagnosis, as the period between the cut-off score for depression.
appearance of symptoms and the MG diagnosis, averaging We did not find it possible in our sample to include
2 years (SD 3), ranged from 0 to 18 years in our sample. the medication dosage in the regression analysis as rec-
ommended by Paul et al. [23] as the patients were
Statistical analysis treated with different kinds of agents and the dose was
not stable, but depended on the fluctuation of symptoms.
Group comparisons were performed with the use of Student Therefore, we compared the cognitive performance of
two-tailed t test. Tau-B Kendall correlation coefficient was patients with (n = 19) and without steroid treatment
used to assess the strength of a relationship between ordinal (n = 14), using t test. No significant differences were

123
12 Neurol Sci (2009) 30:9–14

found on the same measures as those enumerated in indicates that TMT is loaded by motor and attention fac-
Table 1. tors. However, amongst measures related to TMT—Digit
As shown in Table 2, no fatigue effect was observed in Span backward and FTT—only the latter, that is motor
either AVLT learning trials or fluency tasks, as the final dependent differentiated between the groups as TMT did.
trials did not reveal decrements in MG group in compari- The results pattern supports the claim about intact attention
son with the control group. and working memory in MG.
Correlation analysis for BDI score, time since MG Paul et al. [20, 21] suggested that information process-
onset, time since MG diagnosis, age, MGFA Clinical ing speed was slowed as evidenced by MG group
Classification score and MG disability score was performed impairment on SDMT and measures requiring visual pro-
exclusively for the MG group. No statistically significant cessing that were time constrained, such as Visual
correlations were detected. Reproductions, but not on Rey Complex Figure Test that is
Subsequently, correlations of TMT with Digit Span and untimed. It could be argued that the observed impairment
FTT scores for MG patients were computed. TMT A raw may be dependent on increased visual demands (due to
score correlated with FTT dominant hand score (r = ophthalmoparesis or more subtle visual searching prob-
-0.37, P = 0.03). Similar correlation was revealed for lems) and is not a result of higher cognitive demands.
TMT B raw score (r = -0.41, P = 0.02), but not for TMT The unimpaired verbal learning found in our study is in
B–A scaled score difference. TMT A raw score correlated contradiction with the previous results [20, 21], where the
with Digit Span backward (r = -0.54, P = 0.001), as did differences were noted only for immediate recall trials. The
TMT B score (r = -0.63, P \ 0.001). TMT B–A scaled comparison of the groups of patients tested in our study and
score difference did not correlate with FTT. the aforementioned ones in terms of mean disease severity
and time since onset did not reveal large differences that
could account for the discrepancies in cognitive perfor-
Discussion mance. Two explanations are possible. First, the order of
test administration may have influenced the results. In our
In this study MG patients’ performance was comparable to study, AVLT was administered always at the beginning of
controls’ on measures of global cognitive functioning, the testing session, after MMSE, which may have limited
which contradicts Iwasaki et al.’s results [16, 35]. Indi- the fatigue impact and resulted in unimpaired patients’
viduals with MG exhibited significant difficulties only in performance. Paul et al. [20, 21] administered the test
tasks requiring motor speed and oculomotor accuracy, but battery in the counterbalanced order, which may indicate
eliminating the motor speed component showed working that a verbal learning test in a considerable number of
memory to be unimpaired, which confirms Bartel and patients was administered later in a test session than in our
Lotz’s results [18]. Furthermore, attention span was found study. Second, the previous studies tested either signifi-
to be intact, as in the previous studies [17, 21, 22]. cantly smaller samples of patients [11] (N = 12) or
Analysis of the relationships between a visually and controls [20, 21] (N = 18) than in our study.
motor dependent working memory measure (TMT) with a Delayed retention and recognition were deficient in
motor performance measure (FTT) and a working memory comparison to controls only in Tucker et al.’s study [11].
measure without motor component (Digit Span backward) However, our MG sample size is three times as big as

Table 2 Analysis of
Measure Control group MG patients t P
fatigability effect in AVLT and
mean (SD) mean (SD)
verbal fluency trials
Verbal learning and retention
AVLT, trial 1 5.27 (1.57) 5.39 (1.62) -0.32 0.75 (ns)
AVLT, trial 2 8.67 (1.92) 8.61 (2.50) 0.11 0.92 (ns)
AVLT, trial 3 10.43 (2.01) 10.21 (2.42) 0.39 0.70 (ns)
AVLT, trial 4 11.23 (2.01) 11.21 (2.46) 0.04 0.97 (ns)
AVLT, trial 5 12.23 (1.79) 11.55 (2.94) 1.11 0.27 (ns)
Executive function
Letter fluency 0–30 s 9.50 (2.43) 9.00 (2.68) 0.77 0.44 (ns)
Letter fluency 30–60 s 5.20 (2.52) 4.82 (2.46) 0.61 0.55 (ns)
Semantic fluency 0–30 s 14.10 (3.04) 13.45 (3.56) 0.77 0.44 (ns)
Semantic fluency 30–60 s 8.23 (3.55) 7.15 (2.83) 1.34 0.18 (ns)

123
Neurol Sci (2009) 30:9–14 13

Tucker et al’s one. Like in Paul et al.’s studies [21, 22] we 6. Keesey JC (1999) Does myasthenia gravis affect the brain?
have not found memory impairment on delayed measures. Review article. J Neurol Sci 170:77–89
7. Hokkanen E, Toivakka E (1969) Electroencephalographic find-
Lack of fatigue-related decrement in the last trials of ings in myasthenia gravis, 180 EEG recordings of 109 patients.
verbal learning task and the second part of each verbal Acta Neurol Scand 45:556–567
fluency trial in our study contradicts Paul et al.’s [21] 8. Jech R, Růžička E (1996) Brain stem auditory evoked potentials
interpretation of the mental fatigue sensitivity in MG. reflect central nervous system involvement in myasthenia gravis.
J Neurol 243:547–557
However, when comparing our data with Paul et al.’s [20, 9. Papazian O (1976) Rapid eye movement sleep alterations in
21], especially taking into account the order of test myasthenia gravis. Neurology 26:311–316
administration, it may be hypothesized that such an effect 10. Lefvert AK, Pirskanen R (1977) Acetylocholine-receptor anti-
would emerge if an additional learning trial was used. To bodies in cerebrospinal fluid of patients with myasthenia gravis.
Lancet 2:351–352
test the fatigue effect, we suggest that a verbal learning test 11. Tucker DM, Roeltgen DP, Wertheimer RI (1988) Memory
should be administered twice, the standard version at the dysfunction in myasthenia gravis: evidence for central choliner-
beginning of the test session and the alternative version at gic effects. Neurology 38:1173–1177
the end of it. Then, the greater decrement in patients than in 12. Mertens HG, Luetzenkirchen J, Hertel G (1976) Psychologic
problems in treatment of myasthenic patients [Psychologische
the control subjects would prove the elevated sensitivity to Probleme bei der Behandlung von Myastheniekranken]. Nerve-
mental fatigue in MG patients. narzt 47(8):517–519
Although we provided a larger sample than those pre- 13. Glennerster A, Palace J, Warburton D, Oxbury S, Newsom-Davis
viously reported in literature, our study has certain J (1996) Memory in myasthenia gravis: Neuropsychological tests
of central cholinergic function before and after effective immu-
limitations. It could not fully address the effect of medi- nologic treatment. Neurology 46:1138–1142
cation or mood on cognitive performance. The latter 14. Lewis SW, Ron MA, Newsom-Davis J (1989) Absence of central
shortcoming is due to unavailability of a depression functional cholinergic deficits in myasthenia gravis. J Neurol
measure with separate scores for mood and vegetative Neurosurg Psychiatry 52:258–326
15. Aarli JA, Gilhus S, Thoracius S, Johnsen HJ (1989) Recovery
components adapted for Polish conditions, such as Chicago from global amnesia during plasma exchange in myasthenia
Multiscale Depression Inventory. Inclusion of patients with gravis: report of a case. Acta Neurol Scand 80:351–353
visual impairment made it impossible to test visuospatial 16. Iwasaki Y, Kinoshita M, Ikeda K, Shiojima T, Kurihara T (1993)
functioning. Furthermore, the severity of symptoms was Neuropsychological function before and after plasma exchange in
myasthenia gravis. J Neurol Sci 114:223–226
assessed only by MGFA Clinical Classification score and a 17. Paradis CM, Lazar RM, Kula RW (1994) Cognitive function in
subjective measure. We did not find it possible to balance myasthenia gravis. Neuropsychiatry Neuropsychol and Behav
the patients’ number within each MG subgroup. However, Neurol 7:211–214
it can be argued that overrepresentation of patients classi- 18. Bartel PR, Lotz BP (1995) Neuropsychological test performance
and affect in myasthenia gravis. Acta Neurol Scand 91:266–270
fied as Class II corresponds well to the clinical reality. 19. Bohbot VD, Jech R, Bureš J, Ružička E (1997) Spatial and
Our results did not confirm CNS deficit in MG. MG nonspatial memory involvement in myasthenia gravis. J Neurol
patients’ cognitive performance seems to be unimpaired in 244:529–532
tasks independent of motor or visual function. 20. Stepansky R, Weber G, Zeitlhofer J (1997) Sleep apnea and
cognitive dysfunction in myasthenia gravis. Acta Med Austriaca
3:128–131
Acknowledgments We wish to thank Prof. Krzysztof Jodzio from 21. Paul RH, Cohen RA, Gilchrist JM, Aloia MS, Goldstein JM
University of Gdańsk, Institute of Psychology for his help in pre- (2000) Cognitive dysfunction in individuals with myasthenia
paring the testing procedure and supervising the research. gravis. J Neurol Sci 179:59–64
22. Paul RH, Cohen RA, Gilchrist JM (2002) Ratings of subjective
mental fatigue relate to cognitive performance in patients with
myasthenia gravis. J Clin Neurosci 9:243–246
References 23. Paul RH, Cohen RA, Zawacki T, Gilchrist JM, Aloia MS (2001)
What have we learned about cognition in myasthenia gravis? A
1. Kuks JBM, Oosterhuis HJGH (2003) Clinical presentation and review of methods and results. Neurosci Biobehav Rev 25:75–81
epidemiology of myasthenia gravis. In: Kaminski HJ (ed) Current 24. Paul RH, Gilchrist JM (2003) Psychological and social conse-
clinical neurology: myasthenia gravis and related disorders. quences of myasthenia gravis. In: Kaminski HJ (ed) Current
Humana Press, Totowa, pp 93–113 clinical neurology: myasthenia gravis and related disorders.
2. Aita JF, Snyder DH, Reichl W (1974) Myasthenia gravis and Humana Press, Totowa
multiple sclerosis: an unusual combination of diseases. Neurology 25. Task Force of the Medical Scientific Advisory Board of the
24:72–75 Myasthenia Gravis Foundation of America (2000) Myasthenia
3. Oosterhuis HJGH (1996) Acquired blepharoptosis. Clin Neurol gravis: recommendations for clinical research standards. Neurology
Neurosurg 98:1–7 55:16–23
4. Hokkanen E (1969) Myasthenia gravis, a clinical analysis of the 26. Folstein MF, Folstein SE, McHugh PR (1975) Mini mental state:
total material for Finland with special reference to endocrino- a practical method for grading the cognitive state of patients for
logical and neurological disorders. Ann Clin Res 1:94–108 the clinician. J Psychiatr Res 12:189–198
5. McManus IC, Naylor J, Booker BL (1990) Left-handedness and 27. Lezak MD (1995) Neuropsychological Assessment. Oxford
myasthenia gravis. Neuropsychologia 28:947–955 University Press, Oxford

123
14 Neurol Sci (2009) 30:9–14

28. Rey A (1964) L’examen clinique en psychologie. Presses 32. Reitan RM, Wolfson D (1985) The Halstead-Reitan neuro-
Universitaires de France, Paris psychological battery. Theory and clinical Interpretation.
29. Weingartner H, Cohen RM, Murphy DL, Martello J, Gerdt C Neuropsychology Press, Tucson
(1981) Cognitive processes in depression. Arch Gen Psychiatry 33. Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J (1961)
38:42–47 An inventory for measuring depression. Arch Gen Psychiatry
30. Lachner G, Engel RR (1994) Differentiation of dementia and 4:561–571
depression by memory tests—a meta-analysis. J Nerv Mental Dis 34. Romani A, Piccolo G, Bergamaschi R, Versino M, Cosi V (2002)
182:34 A reliability study of impairment and disability scales for
31. Brzeziński J, Gaul M, Hornowska E, Machowski A, Zakrzewska M myasthenia gravis patients. Funct Neurol 17:137–144
(1996) Skala Inteligencji D. Wechslera dla dorosłych—wersja 35. Iwasaki Y, Kinoshita M, Ikeda K, Takamiya K, Shiojima T
zrewidowana WAIS-R (PL)-podre˛cznik. [Wechsler Adult (1990) Cognitive dysfunction in myasthenia gravis. Int J Neurosci
Intelligence Scale-Revised. Manual]. Pracownia Testów Psycho- 54:29–33
logicznych PTP, Warszawa

123