Recommendations for myasthenia gravis clinical trials

We read the recommendations for myasthenia gravis (MG) clinical trials1 with great interest. We would be highlight some further relevant issues.

Although composite measures may provide more comprehensive evaluation, the employment of a cutoff value for determining improvement or worsening is neither accurate nor sensitive enough to serve as a primary endpoint. The highly variable course of MG and the heterogeneity of disease severity may dilute some of the major changes in individual patients by between-group comparisons, particularly for those patients who have a relatively mild form of the disease. In this regard, we advise using a relative score that is based on the amount of change in the absolute outcome score of an individual patient during the observation period as compared with the initial absolute score. This analysis would make possible the determination of treatment effects for individual patients. In China, such an absolute and relative scoring system2 has been used for more than 15 years. The absolute scores were found to have excellent interobserver agreement2,3 and to correlate with the Quantitative MG scale (QMG),3 decremental responses on repetitive nerve stimulation, and parameters of single-fiber electromyography.4 Moreover, the absolute score was more sensitive to clinical change in severity than the QMG, especially in ocular MG patients, because more weight was given to extraocular muscles.3 The relative score calculated with QMG is not significantly different from that with the Chinese absolute score system.3 The relative score may provide a useful individualized evaluation of therapeutic effects and can be analyzed as a linear parameter. Furthermore, comparison of the proportion of patients in both treatment and placebo groups who meet prespecified efficiency criteria that are based on the relative score may provide us with another view of the treatment effects, even if between-group comparisons show no significant differences.

Although we agree that both timing and dose of cholinesterase inhibitor are covariates in the analysis, we propose performing scale evaluation after a sufficient time for elapsing of the effects of the last dose. Although there is a shifting focus from outcome measures of strength/fatigue to more global measures (e.g., MG-Composite), about half of the total scores derive from fatigability. Inclusion of the recommended latency from cholinesterase inhibitor administration to outcome assessment as a covariate will not fully reflect the treatment effects of tested therapies. Some patients may be reluctant to taper the dose of cholinesterase inhibitors for fear of symptom aggravation even though their disease severity has decreased, and different patients will respond differentially at a given time-point after the last dose due to the interindividual variability of pharmacokinetic parameters.5 This effect may appear in both the active treatment and control groups and lead to dilution of the true treatment effect. Because the duration is relatively short for treatment effects to be apparent in most MG trials and the concomitant immunomodulating treatment has more responses than formerly anticipated,1 increased sensitivity of the trial by excluding potential effects of cholinesterase inhibitors may increase the chance of identifying a true treatment effect.