MG-ADL: STILL A RELEVANT OUTCOME MEASURE
SRIKANTH MUPPIDI, MD,
GIL I. WOLFE, MD,
MARK CONAWAY, PhD,
and TED M. BURNS, MD
and the MG COMPOSITE AND MG-QOL15 STUDY GROUP
Department of Neurology, University of Texas Southwestern Medical Center, Dallas, Texas, USA
Department of Public Health Sciences, University of Virginia, Charlottesville, Virginia 22908, USA
Department of Neurology, University of Virginia, Charlottesville, Virginia, USA
Accepted 15 April 2011
ABSTRACT: Introduction: The aim of this analysis was to
examine the performance of the Myasthenia Gravis–specific
Activities of Daily Living scale (MG-ADL) during a multicenter
scale validation study. Methods: Consecutive MG patients were
assessed with several MG outcome measures, including the
MG-ADL. Statistical tests included descriptive analysis, Pearson
correlation, and sensitivity/specificity. Results: Eighty-seven
patients completed the MG-ADL, MG Composite (MGC), and
MG 15-item Quality of Life scale (MG-QOL15) on the first visit,
and 76 returned for the second visit. At the first visit, there was
a strong positive correlation between MG-ADL and MGC (r¼
0.85, P<0.0001) and MG-QOL15 (r¼0.76, P<0.0001). Cor-
relation of the delta MG-ADL score and physician impression of
change between the visits was strong (r¼0.70, P<0.0001). A
2-point improvement in the MG-ADL best predicted clinical
improvement. Test–retest analysis demonstrated a high reliabil-
ity coefficient. Conclusions: The MG-ADL correlates strongly
with newer, validated MG outcome measures. A 2-point
improvement in the MG-ADL indicates clinical improvement.
The MG-ADL is useful as a research tool and in routine clinical
Muscle Nerve 44: 727–731, 2011
Multiple outcome measures have been developed
for evaluating the clinical status of patients with
myasthenia gravis (MG).
These outcome measures
differ in ease of use, ease of interpretation, what is
being measured, how it is being measured, and
who interprets and reports the symptoms and
impairments (i.e., patient-reported vs. physician-
The Myasthenia Gravis–speciﬁc
Activities of Daily Living scale (MG-ADL) was devel-
oped in the late 1990s to assess the status of symp-
toms and activities in MG.
It is an 8-item, patient-
reported questionnaire that can be completed in
2–3 minutes with no need for specialized equip-
ment or training (Table 1).
We recently studied and validated two new MG
outcome measures, the MG Composite (MGC) and
MG 15-item Quality of Life (MG-QOL15) scales, in
a prospective, multicenter study.
As part of this
scale validation study, several centers collected MG-
ADL data. In this analysis, we evaluate the perform-
ance of the MG-ADL.
A multicenter, scale validation observational study
was recently completed in patients with MG to
assess the validity and reliability of the MGC and
MG-QOL15. The uniform entry criteria for all cen-
ters were as follows: (1) patients were at least 17
years of age with a diagnosis of MG based on clini-
cal, serological, and electrodiagnostic testing; and
(2) the treating physician scheduled the patient to
be seen for follow-up within 6 months. The study
protocol did not mandate any changes in treat-
ment or management; this was left to the discre-
tion of the treating physician.
MG patients were enrolled at 11 centers. Prior to
the study, investigators at the centers were asked to
perform either the MG-ADL or the MG Manual
Motor Test, in addition to other measures
included in the study. Patients at ﬁve centers com-
pleted the MG-ADL and other outcome measures
at two consecutive visits. At both visits, demo-
graphic data and disease features, including
antibody status, duration of MG, and MG Founda-
tion of America (MGFA) class, were also docu-
mented. A neuromuscular-trained neurologist
documented the physician impression of change at
the second visit using a 7-point scale: markedly
improved; moderately improved; slightly improved;
unchanged; slightly worsened; moderately wors-
ened; or markedly worsened. The study protocol
did not provide guidance or instruction on the
care of the patients. The protocol was approved by
the institutional review board at each study center,
and informed consent was obtained from all
Test–retest reproducibility of the MG-ADL was
performed on 26 patients in a separate study at
the University of Virginia. The MG-ADL was com-
pleted by each patient during the ﬁrst clinic visit.
Patients were asked to complete a second MG-ADL
at home within 2–4 days of the clinic visit and
return it in a self-addressed, stamped envelope. We
did not call patients to remind them to complete
the scale or assist them during the completion of
the second MG-ADL. The 2–4-day time period was
chosen to minimize the likelihood that the patient
would recall answers from the ﬁrst MG-ADL and to
minimize the likelihood the disease would have
changed signiﬁcantly. The protocol was approved
Abbreviations: AChR-Ab, acetylcholine receptor antibodies; MG, myas-
thenia gravis; MG-ADL, Myasthenia Gravis–specific Activities of Daily Living
scale; MGC, MG Composite scale; MGFA, Myasthenia Gravis Foundation
of America; MG-QOL15, 15-item Myasthenia Gravis Quality of Life scale;
MuSK-Ab, muscle-specific tyrosine kinase antibodies; QMG, Quantitative
Myasthenia Gravis scale; ROC, receiver operating characteristic
Correspondence to: T. M. Burns; e-mail: email@example.com
C2011 Wiley Periodicals, Inc.
Published online 15 October 2011 in Wiley Online Library
(wileyonlinelibrary.com). DOI 10.1002/mus.22140
Key words: MG outcome measures, MG scales, MG-ADL, MGC and MG-
QOL15, myasthenia gravis
MG-ADL MUSCLE & NERVE November 2011 727
by the institutional review board at the University
of Virginia, and informed consent was obtained
from all subjects.
We analyzed the correlations between the MG-
ADL, MGC, and MG-QOL15 at both visits and also
the change in scores between the visits. Both Pear-
son and Spearman correlation analyses were used;
however, only the Pearson correlation is reported
here, as there was no signiﬁcant difference
between these correlation parameters. Receiver
operating characteristic (ROC) curve and sensitiv-
ity/speciﬁcity analyses were performed based on
the physician impression of change and the MG-
QOL15 score at the second visit. SAS version 9.2
(SAS Institute, Inc., Cary, North Carolina) software
was used for statistical analysis.
A total of 87 patients were enrolled from the ﬁve
participating centers. Seventy-six patients returned
for the second visit. The mean duration between
visits was 129 days (range 25–198 days). The mean
age was 58.1 years, and 49% were women. The
MGFA classiﬁcations at visit 1 were as follows: class
I, 18% (16 of 87); class IIa, 22% (20 of 87); class
IIb, 9% (8 of 87); class IIIa, 25% (22 of 87); class
IIIb, 6% (6 of 87); and class IV, 4% (4 of 87).
Eleven patients were in remission. Seventy-four
patients were seropositive for acetylcholine recep-
tor antibodies (AChR-Ab), 4 for muscle-speciﬁc ty-
rosine kinase antibodies (MuSK-Ab), and 7 were
seronegative. Antibody status was unknown in 2
patients. At visit 1, 73 patients were established fol-
low-up patients, 8 patients were new clinic patients,
and 6 patients were hospitalized.
Absolute scores and change in scores for the
MG-ADL, MGC, and MG-QOL15 for the ﬁrst and
second visits are presented in Table 2.
Pearson correlations were performed between
the MG-ADL, MGC, and MG-QOL15 scores at each
visit and on the change in scores between visits
(Table 3). All of these values revealed robust posi-
tive correlations between the MG-ADL and the
other two MG scales (P<0.0001). Correlation
coefﬁcients for the change in scores were also high
(P<0.0001). A high correlation also existed
between the MG-ADL, MGC, MG-QOL15, and phy-
sician impression of change (P<0.0001; Table 4).
Of note, in only 1 of 76 patients seen in follow-up
was the change in MG-ADL score and the physi-
cian impression of change discordant.
Based on previously published data,
improvements in physician impression of change
Table 1. MG Activities of Daily Living (MG-ADL) Profile.
Grade 0 1 2 3
(0, 1, 2, 3)
1. Talking Normal Intermittent
slurring of nasal speech
Constant slurring or nasal,
but can be understood
Difficult to understand
2. Chewing Normal Fatigue with solid food Fatigue with soft food Gastric tube
3. Swallowing Normal Rare episode of choking Frequent choking
changes in diet
4. Breathing Normal Shortness of
breath with exertion
Shortness of breath at rest Ventilator dependence
5. Impairment of ability
teeth or comb hair
None Extra effort, but no rest
Rest periods needed Cannot do one
of these functions
6. Impairment of ability
from a chair
None Mild, sometimes
7. Double vision None Occurs, but not daily Daily, but not constant Constant
8. Eyelid droop None Occurs, but not daily Daily, but not constant Constant
MG-ADL score (items
Tab le 2 . Descriptive data for MG-ADL, MGC, and
MG-QOL15 for both visits and physician impression of
change at second visit.
NMean SD Median
MG-ADL 87 4.89 3.54 5
MGC 87 8.89 6.87 8
MG-QOL15 total 87 20.8 15.27 20.0
MG-ADL 76 3.59 3.3 3
MGC 76 6.26 5.8 5
MG-QOL15 total 76 15.14 13.88 9
Change in scores (visit 2-visit 1)
MG-ADL 76 1.22 3.0 1.0
MGC 76 2.61 6.25 0.5
MG-QOL15 total 76 5.32 11.7 1.5
Physician impression of change 76 0.69 2.1 0
728 MG-ADL MUSCLE & NERVE November 2011
and in the MG-QOL15 were thought to be the best
indicators of clinical improvement. Figure 1 dem-
onstrates the ROC curve for the MG-ADL using
this indicator as the ‘‘gold standard’’ for clinical
improvement. The ROC area under the curve was
0.90, suggesting high accuracy. Sensitivity/speciﬁc-
ity analysis performed with various cut-off points
for the change in MG-ADL suggested that a 2-point
reduction in the scale best predicted improvement
in MG clinical status (Table 5).
Test–Retest Reproducibility of MG-ADL. A total of
26 patients completed the MG-ADL during the
clinic visit, and 20 repeated the MG-ADL at home
within 1 week. The mean MG-ADL score was 3.6
(SD: 63.2) at the clinic visit for the 20 patients
who returned the second MG-ADL; mean MG-ADL
was 7.3 (SD: 63.2) for the 6 patients who did not
return the second MG-ADL. The test–retest reli-
ability coefﬁcient was 93.7%, with a lower 95% con-
ﬁdence interval at 87.3%. The MG-ADL scores
were within 2 points 85% of the time (17 of 20
instances) and never differed by >3 points.
Responsiveness of MG-ADL. The mean improve-
ment in ADL score in patients who improved
based on the gold standard (improvement in MG-
QOL15 score plus improvement in physician
impression of change score) was 3.88 (SD: 62.72).
The standardized mean response was 3.88/2.72 ¼
1.43. The effect size was 1.21 (mean change in
MG-ADL in patients who improved divided by the
SD from the test–retest data ¼3.88/3.2). Both of
these ﬁndings indicate excellent responsiveness of
MG-ADL in patients who improved between the
ﬁrst and second visits.
The MG-ADL was designed to assess relevant symp-
toms and the functional performance of daily activ-
ities that are sometimes impaired by MG. The 8
items in the scale were derived from symptom-
based components of the original 13-item QMG.
In the original description, the MG-ADL was found
to have good correlation with the QMG and was
recommended as a secondary endpoint in clinical
Since then, it has been widely applied to
both routine practice and clinical research. The
inclusion of the MG-ADL in our multicenter scale
validation study of the MGC and MG-QOL15 pro-
vided an opportunity to further study its perform-
ance. Our analysis suggests that a 2-point improve-
ment in MG-ADL score optimally indicates clinical
improvement and has acceptable reliability. The
correlation between the change in MG-ADL and
the physician assessment of global change also sug-
gested that the MG-ADL possesses excellent
responsiveness to clinical change. The correlations
Table 3 . Correlation coefficients between MG-ADL, MGC, and MG-QOL15.
Visit 1 0.846 (<0.0001) 0.763 (<0.0001) 0.734 (<0.0001)
Visit 2 0.869 (<0.0001) 0.775 (<0.0001) 0.647 (<0.0001)
Change in scores (second visit-first visit) 0.753 (<0.0001) 0.671 (<0.0001) 0.547 (<0.0001)
Data expressed as Pearson’s R (P-value).
Tab le 4 . Correlation coefficients between MG-ADL, MGC,
MG-QOL15, and physician impression of change.
MGQOL 15 vs.
0.703 (<0.0001) 0.770 (<0.0001) 0.554 (<0.0001)
Data expressed as: Pearson’s R (P-value).
FIGURE 1. ROC curve for MG-ADL change with improvement
in physician global impression of change score plus improve-
ment in MG-QOL15 score as the ‘‘gold standard.’’ An area under
the curve of 0.90 suggests high accuracy.
MG-ADL MUSCLE & NERVE November 2011 729
between the MG-ADL and both the MGC and MG-
QOL15 were also high.
The construct validity, responsiveness to clinical
improvement, simplicity, and ease-of-use of the
MG-ADL are attractive attributes for use in the
clinic and in clinical trials. In addition to the favor-
able properties observed in this analysis, the MG-
ADL was recently found to be a sensitive outcome
measure of responsiveness to change in a clinical
trial of mycophenolate mofetil in MG.
more, because it does not necessarily require face-
to-face contact, the MG-ADL might potentially be
useful as one way to follow patients outside of the
clinic visit (along with, e.g., the MG-QOL15). This
potential additional use of the MG-ADL is sup-
ported by our acceptable test–retest reproducibility
results, which were obtained by having patients
ﬁrst complete the MG-ADL in the clinic followed
by completion of the second MG-ADL at home
without assistance from any care provider. This
potential remote use of the MG-ADL merits fur-
Our analysis demonstrates the validity of the
MG-ADL in a representative sample of the MG
population. In contrast to most clinical trials, our
study included patients seen for MG regardless of
clinical disease severity, many of whom would not
be eligible for clinical trials (e.g., those in remis-
sion). The mean MG-ADL score at ﬁrst visit in this
study was lower than for some clinical trials,
likely because we included mildly affected patients
and patients in remission. Our sample population
is similar in many characteristics, such as for sero-
logical status and MGFA class distribution, to other
reported MG cohorts.
There are a few limitations to our analysis. Our
multicenter study was primarily designed to assess
the validity of the MGC and MG-QOL15, and ini-
tial power calculations were based on this analysis.
Another limitation stems from the fact that the
MG-ADL and MGC share 4 items, which inﬂates
the correlations between these two scales. How-
ever, similarly signiﬁcant correlations were
observed between the MG-ADL and MG-QOL15,
where no test items are shared. Finally, only 5 of
11 centers chose to record MG-ADL scores, and
therefore the sample size was 50% of the original
We do not believe this created a sig-
niﬁcant selection bias, however, as all sites enrolled
patients using identical entry criteria.
This project was supported by Myasthenia Gravis Foundation of
America grant (to Ted Burns) and the North and Central Texas
Clinical and Translational Science Initiative KL2RR024983 (to Sri-
The MG Composite and MG-QOL15 Study Group
includes: Ted M. Burns (Principal Investigator,
University of Virginia); Mark Conaway (University
of Virginia); Guillermo Solorzano (University of
Virginia); Maria E. Farrugia (Institute of Neurolog-
ical Sciences, Glasgow, UK); Donald B. Sanders
(Duke University); Janice M. Massey (Duke Univer-
sity); Vern C. Juel (Duke University); Lisa D. Hob-
son-Webb (Duke University); Bernadette Tucker-
Lipscomb (Duke University); Carlo Antozzi (Neu-
rological Institute Foundation ‘‘Carlo Besta,’’ Mi-
lan, Italy); Renato Mantegazza (Neurological
Institute Foundation ‘‘Carlo Besta,’’ Milan, Italy);
David Lacomis (University of Pittsburgh); Elliot
Dimberg (Mayo Clinic, Jacksonville); Srikanth
Muppidi (University of Texas Southwestern); Gil I.
Wolfe (University of Texas Southwestern); Mazen
M. Dimachkie (University of Kansas); Richard J.
Barohn (University of Kansas); Mamatha Pasnoor
(University of Kansas); April L. McVey (University
of Kansas); Laura Herbelin (University of Kansas);
Tahseen Mozaffar (University of California, Irvine);
Vinh Q. Dang (University of California, Irvine);
Sandhya Rao (University of California, Irvine);
Robert Pascuzzi (Indiana University); Riley Snook
(Indiana University); and Anthony A. Amato (Brig-
ham and Women’s Hospital and Harvard Medical
1. Burns TM. History of outcome measures for myasthenia gravis. Mus-
cle Nerve 2010;42:5–13.
2. Barohn RJ. Standards of measurements in myasthenia gravis. Ann
NY Acad Sci 2003;998:432–439.
3. Burns TM, Conaway M, Sanders DB. The MG Composite: a valid and
reliable outcome measure for myasthenia gravis. Neurology 2010;74:
4. Burns TM, Conaway MR, Cutter GR, Sanders DB. Construction of an
efﬁcient evaluative instrument for myasthenia gravis: the MG com-
posite. Muscle Nerve 2008;38:1553–1562.
5. Burns TM, Conaway MR, Cutter GR, Sanders DB. Less is more, or
almost as much: a 15-item quality-of-life instrument for myasthenia
gravis. Muscle Nerve 2008;38:957–963.
6. Farrugia ME, Robson MD, Clover L, Anslow P, Newsom-Davis J, Ken-
nett R, et al. MRI and clinical studies of facial and bulbar muscle
involvement in MuSK antibody-associated myasthenia gravis. Brain
7. Gajdos P, Sharshar T, Chevret S. Standards of measurements in my-
asthenia gravis. Ann NY Acad Sci 2003;998:445–452.
8. Mantegazza R, Antozzi C, Peluchetti D, Sghirlanzoni A, Cornelio
F. Azathioprine as a single drug or in combination with steroids
in the treatment of myasthenia gravis. J Neurol 1988;235:449–
Tab le 5 . Sensitivity and specificity for various levels of
improvement in MG-ADL score with improvement in physician
impression of change score plus improvement in MG-QOL15
score as ‘‘gold standard’’.
change in score
730 MG-ADL MUSCLE & NERVE November 2011
9. Padua L, Evoli A, Aprile I, Caliandro P, Batocchi AP, Punzi C, et al.
Myasthenia gravis outcome measure: development and validation of
a disease-speciﬁc self-administered questionnaire. Neurol Sci 2002;
10. Sanders DB, Tucker-Lipscomb B, Massey JM. A simple manual
muscle test for myasthenia gravis: validation and comparison with
the QMG score. Ann NY Acad Sci 2003;998:440–444.
11. Wolfe GI, Herbelin L, Nations SP, Foster B, Bryan WW, Barohn RJ.
Myasthenia gravis activities of daily living proﬁle. Neurology 1999;52:
12. Barohn RJ, McIntire D, Herbelin L, Wolfe GI, Nations S, Bryan WW.
Reliability testing of the quantitative myasthenia gravis score. Ann
NY Acad Sci 1998;841:769–772.
13. Muscle Study Group.A trial of mycophenolate mofetil with predni-
sone as initial immunotherapy in myasthenia gravis. Neurology 2008;
14. Wolfe GI, Barohn RJ, Sanders DB, McDermott MP. Comparison of
outcome measures from a trial of mycophenolate mofetil in myasthe-
nia gravis. Muscle Nerve 2008;38:1429–1433.
15. Mantegazza R, Beghi E, Pareyson D, Antozzi C, Peluchetti D, Sghir-
lanzoni A, Cosi V, et al. A multicentre follow-up study of 1152
patients with myasthenia gravis in Italy. J Neurol 1990;237:339–344.
16. Grob D, Brunner N, Namba T, Pagala M. Lifetime course of myas-
thenia gravis. Muscle Nerve 2008;37:141–149.
17. Oosterhuis HJ. Studies in myasthenia gravis. 1. A clinical study of
180 patients. J Neurol Sci 1964;38:512–546.
MG-ADL MUSCLE & NERVE November 2011 731