Ted M. Burns Discusses the MG Composite Scale: A Valid and Reliable Outcome Measure for Myasthenia Gravis

June 25, 2010


Ted M. Burns Discusses the MG Composite Scale: A Valid and Reliable Outcome Measure for Myasthenia Gravis

Ted M. Burns, MD, Associate Professor of Neurology at the Department of Neurology of the University of Virginia, discusses his paper "The MG composite: a valid and reliable outcome measure for myasthenia gravis" that was recently published in Neurology (2010;74:1434–1440). He spoke with Jose´ Merino, MD, Science Editor of AAN.com

AAN.com: What is the MG composite scale?  What test items does it evaluate?

Burns: It’s a simple, 10-item scale that evaluates the functional domains most commonly affected by myasthenia gravis. It is easy to administer and interpret, and it requires no equipment. It takes about three minutes to complete.

AAN.com: Please describe how you and your colleagues developed the MG composite scale. Why did you include the specific test items?  How did you determine their relative weight?

Burns:  We selected test items that reflect the functional domains of most importance to myasthenia gravis patients and their physicians from three myasthenia gravis-specific ordinal scales: the Quantitative Myasthenia Gravis (QMG), Myasthenia Gravis-Activities of Daily Living (MG-ADL) and Myasthenia Gravis-Manual Motor Test (MG-MMT) scales. What we did was analogous to picking an all-star team in baseball; we selected the best performing test item for various "positions" on the body. For example, we selected one item for evaluating dysarthria, one for ptosis, one for neck strength, and so on.

We then assigned relative weights to each test item because we did not think that each item —and specifically each response option of each item— had the same value in terms of quality of life and risk to health and survival, among other things. For example, moderate eye closure weakness and moderate dysphagia are important features of the disease, but they have different quality of life implications.

We asked 36 experts in myasthenia gravis from nine countries with more than 600 years of combined experience with the disease to help us weight each possible response for each item of the MG composite. We decided that the worst score should be 50 points, and we gave each response for each of the 10 items a "relative weight."

The results were very interesting and, fortunately, made sense. For example, the MG specialists decided that severe ventilatory weakness was worth nine points, whereas severe neck weakness was worth only four points. And I think the relatively tight standard deviations of the weights of each response option suggest some consensus among experts from around the world and give support to the weighted scale.

AAN.com: What other outcome scales have been used in MG clinical trials?  How does the MG composite scale differ from these other outcome measures?

Burns: There are quite a few other scales and I’d direct anyone interested in all of them to a review paper about MG scales that I recently wrote for Muscle and Nerve. I think that the MG composite is unique for many reasons: it is weighted; it includes only test items that have been shown through careful analysis to be the best performing; it’s very easy to administer and to interpret; and it requires no equipment to score. It is also a hybrid of signs and symptoms, which is something that, if you think about it, makes great sense for myasthenia gravis.

This approach works particularly well for myasthenia gravis for two main reasons. First, the clinical manifestations are always evident to the patient, which is unlike some diseases, such as hypertension or diabetes mellitus. Second, the clinical manifestations fluctuate. Thus, a physician probably won’t witness all of the manifestations or the true extent of the manifestations experienced by the patient during the "snapshot" examination.

When we analyzed the test performance of all the items from the other scales we learned that the bulbar and respiratory test items taken from the history performed better than their counterparts from the "snapshot" examination. For example, if I see patient in clinic in the morning, I may not identify the fatigable dysarthria that occurs each afternoon, but you can bet that the patient and spouse will notice it. And when we analyzed the performance of the various items that assess the speaking domain, it turned out that the test item from the MG-ADL was the best performing item for speech—so that’s what we selected for the MG composite.

AAN.com: Please describe the methodology and key findings of your study.

Burns: The manuscript published in Neurology was an 11-center study (nine sites in the United States and two in Europe) that validated the MG composite and also our 15-item MG-specific quality of life (QOL)  item (the MG-QOL15). We enrolled 175 patients at these 11 sites. We completed the MG composite, MG-QOL15 and a few other simple myasthenia gravis-specific scales on two consecutive study visits. The key finding was the validation of the MG composite and MG-QOL15 in the everyday practice setting. We found that a three-point improvement in the MG composite score was optimal for signifying clinical improvement, and that patients also felt that such an improvement was clinically meaningful.

AAN.com: How did you select the gold standard for clinical improvement used to assess the sensitivity and specificity of various cut-off points?  How did you determine a cut-off point that was meaningful to patients?

Burns: Because there’s no gold standard for clinical improvement, we looked at several different outcome measures (clinician impression, MG-MMT and MG-ADL improvement >1 point, MG-QOL15 improvement and combinations of these scales) to define clinical improvement. Using these various indicators of clinical change, we were able to look at the sensitivity and specificity of the MG composite for detecting clinical change. When we crunched all that data it looked to us like a three-point cut-point was probably best for indicating clinical improvement.

AAN.com: In your study you establish a cut-off point of clinical improvement that can be used in routine clinical practice. Are other scales commonly used in this setting?

Burns: I find that the MG composite (MGC) and MG-QOL15 are very useful in clinic. The MG composite tells me the status of the 10 important functional domains that are commonly affected in myasthenia gravis. I find the total MGC score and the subscores for each of the 10 test items to be of paramount importance when I’m assessing the clinical status of my MG patients in clinic.

The three-point cut-point of the MGC is probably of secondary importance in everyday clinical practice, but I feel confident that a patient who has a three-point or more improvement in MGC has, in fact, experienced clinical improvement and that this improvement is very likely to be meaningful to the patient. This is because the test-retest reliability of a three-point change is very good: the patients in our study who had a three-point or more improvement in MGC had a mean improvement in the total MG-QOL15 score of greater than 10 points, and, again, because the sensitivity and specificity of a three-point or more change in MGC were quite good when we used other instruments to indicate clinical improvement.

I think that the three-point cut-point is more useful as an outcome measure in clinical trials, especially if one is comparing the proportions of patients who improve with two interventions such as a "new treatment" vs. a "standard of care" group.

AAN.com: Do you expect that the MG composite scale can be the main outcome measure in clinical trials?

Burns: I certainly hope so. I think the MG-QOL15 supplements the MG composite by providing an MG-specific QOL measure. But I believe the MG composite tells us the most about the clinical status of the MG patient.

AAN.com: Should neurologists use the MG composite in their practice to track changes in their patients?  Where can clinicians find the scale and instructions how to use it?

Burns: Yes. I use it every day in clinical practice and I find it very useful. It only takes about three to four minutes to administer and about three to four seconds to interpret. It doesn’t require any training or any equipment. It’s available as a Table in the Neurology paper or also in our 2008 Muscle and Nerve paper. Or I’d be happy to email it to anyone interested (email me at tmb8r@virginia.edu.)

Author Disclosure
Dr. Merino performed a one-time consultation with staff from Bell, Falla and Associates.