A trial of mycophenolate mofetil with prednisone as initial immunotherapy in myasthenia gravis

Study design.

This was a multicenter randomized, double-blind, placebo-controlled, parallel group trial of MMF and moderate-dosage prednisone (20 mg/day) compared to moderate-dosage prednisone alone in subjects with mild to moderate MG. Subjects were followed for 12 weeks on blinded medication and for an additional 24 weeks on open-label MMF.

Subjects.

Eighty subjects aged ≥18 years with acquired, autoimmune, acetylcholine receptor antibody positive generalized MG of mild to moderate severity, who had not been previously treated with immunosuppressive drugs other than corticosteroids, were recruited at 13 sites in the United States. Subjects had to have had a positive edrophonium test or electrophysiologic evidence of a neuromuscular junction disorder (either an abnormal repetitive nerve stimulation study or abnormal single fiber EMG). Subjects were required to have been taking a constant dosage (if any) of pyridostigmine for at least 2 weeks prior to study entry and to have symptom severity that would, in the judgment of the site investigator, justify initiation of immunosuppressive treatment. Subjects were excluded if they had weakness affecting only ocular or periocular muscles (Myasthenia Gravis Foundation of America [MGFA] Class I),⁶ severe weakness of oropharyngeal or respiratory muscles (MGFA Class IVB), or thymoma at time of enrollment or in the past. Subjects could not have received corticosteroids, plasma exchange, or IV immunoglobulin within 90 days of randomization, or thymectomy within 12 months of randomization. This project was approved by the Investigational Review Boards of all participating institutions, and all patients gave written consent to participate.

Interventions.

All subjects began 20 mg/day prednisone and were randomized to MMF (2.5 g/day [1,250 mg BID]) or placebo tablets of identical appearance. The prednisone dosage was kept constant during the first 12 weeks. If adverse events that could be due to MMF occurred, the dosage of blinded medication could be reduced to a minimum of 1,500 mg/day (750 mg BID) at the discretion of the site investigator, and upon concurrence of the medical monitor. The daily pyridostigmine dosage (if any) was kept constant for at least 2 weeks before study entry and during the first 12 weeks, unless pyridostigmine-related adverse events dictated a change. The pyridostigmine dose was held for at least 6 hours prior to performance of the study assessments.

At the 12-week visit, subjects could receive open-label MMF, 2.5 g/day, while continuing 20 mg/day prednisone and pyridostigmine (if any). The MMF could be increased to a maximum of 3 g/day at the discretion of the site investigator and the dosages of prednisone and pyridostigmine were reduced thereafter if warranted by the clinical response. There was no prespecified rate or frequency of prednisone dosage reduction during the open-label phase.

Randomization and enrollment.

Subjects were randomly assigned with equal allocation to MMF or placebo. The randomization plan was stratified by center and by severity of disease at time of randomization (quantitative MG [QMG] score ≤10 or >10). Blocking was also included to promote balance in the number of subjects assigned to the two treatments within each stratum. The computer-generated randomization plan was developed by a programmer in the Muscle Study Group (MSG) Biostatistics Center; only this programmer and the central pharmacist at the University of Rochester who packaged and labeled the drug had access to the treatment assignments. To enroll a subject, the site coordinator sent a randomization form to the randomization administrator, who verified the subject’s eligibility, assigned the next available randomization number, and sent this information to the site coordinator and site pharmacist.

Subject assessment.

Subjects who met the entry criteria were randomized after completing the following baseline assessments of disease activity: the QMG score,⁷ an MG-specific manual muscle testing (MMT) score,⁸ the SF-36v2 quality of life questionnaire,⁹ and an MG-related activities of daily living (MG-ADL) questionnaire.¹⁰ Site personnel included a neurologist investigator, a coordinator, and an evaluator. The QMG score was determined by the evaluator, who was blinded to all other assessments of disease activity. The MMT score was determined by the investigator and the MG-ADL score was determined by the coordinator—both were blinded to the QMG score. The ADL questionnaire was administered weekly by telephone for the first 3 weeks of the double-blind and open-label phases, to ensure safety and to detect evidence of initial improvement. Subjects returned for safety evaluations and assessment of MG disease severity at weeks 4, 8, 12, 16, 20, 28, and 36. If the telephone assessment suggested worsening that was of concern to the subject or the investigator, the subject was seen by the investigator earlier than the next scheduled visit. Serum for measurement of acetylcholine receptor antibody (AChR-Ab) level was drawn at study entry and at weeks 12 and 36; all serum samples were assayed at the end of the study, using the same antigen batch.

At the end of the double-blind and open-label phases (weeks 12 and 36), the investigator made a global judgment about the subject’s response to study treatment, based on the subject’s symptoms and pyridostigmine dosage, and the results of the examination at that visit, but blinded to the QMG score. The potential responses were as follows: (a) markedly improved—MGFA categories Pharmacologic Remission or Minimal Manifestations⁶; (b) improved—less than (a), but the investigator felt the subject was definitely better than at study entry; (c) unchanged; and (d) worse.

Treatment failures.

Patients who deteriorated to a degree that other immunosuppressive therapies needed to be initiated were considered treatment failures and withdrawn from the study. At each clinic visit, the site investigator completed a questionnaire to ascertain if any of the following criteria for possible treatment failure had been met: (a) progressive weakness of respiratory or limb muscles sufficient to significantly impair activities of daily living; (b) progressive weakness of oropharyngeal muscles sufficient to impair adequate nutrition or constitute a significant risk of aspiration; (c) the subject was unwilling to continue in the study because of progressive or continuing disability; (d) the subject’s physician felt that it would not be in the best interest of the subject to continue in the study due to MG symptoms; (e) forced vital capacity (FVC) <30% of predicted; and (f) ADL score increase of more than 3 points from baseline. If any of these conditions were met, the investigator contacted the study medical monitor within 24 hours to discuss the need to withdraw the subject from the study. If the site investigator determined that immediate intervention was needed to assure the subject’s safety, the subject was withdrawn prior to consultation with the study medical monitor.

Safety monitoring.

A Safety Monitoring Committee consisting of two neurologists and a biostatistician reviewed safety data every 4 months throughout the study, paying particular attention to treatment failures. No interim analyses were performed for early evidence of efficacy.

Outcome variables.

The primary outcome variable for efficacy was the change in the QMG score from baseline to week 12. Secondary efficacy outcome variables included changes from baseline to week 12 in the myasthenic MMT score, MG-ADL score, FVC, and SF-36v2 Physical and Mental Component Summary scores. Other efficacy outcomes included treatment failure, global assessment of response, and change in AChR-Ab level from baseline to week 12. Measures of safety included adverse events, laboratory test abnormalities, and changes in laboratory test results and vital signs.

Sample size determination.

Based on a previous clinical trial,¹¹ the SD of the 12-week change in total QMG score was estimated to be 4.0. A sample size of 56 subjects would provide 80% power to detect a difference of 3 points between the MMF and placebo groups using a t test and a two-tailed 5% significance level. This was inflated to 80 subjects to account for an anticipated 15% attrition rate, assuming that attrition rates and mean responses in subjects who withdrew would be comparable between the treatment groups.¹²

Statistical analysis.

The MSG Coordination and Biostatistics Centers (Rochester, NY) provided data management and biostatistical support for this trial.

The primary statistical analyses involved fitting an analysis of covariance model with 12-week change in total QMG score as the dependent variable and treatment group, center, and baseline QMG score as independent variables. The significance of the difference between the adjusted group means and a 95% CI for this difference (treatment effect) was obtained using this model. Similar analyses were performed for the other continuous secondary outcome variables, including laboratory test results and vital signs. Changes from baseline to other time points were examined descriptively. A Wilcoxon rank sum test was used to compare the treatment groups with regard to the global assessment of response to study treatment. Adverse events were assessed descriptively; where warranted, two-tailed Fisher exact tests were performed to compare groups with regard to the incidence of particular adverse events and laboratory test abnormalities. The mean cumulative prednisone dosage during the open-label phase and the mean prednisone dosage at week 36 were compared between the two treatment groups using Wilcoxon rank sum test.

The primary analyses were performed according to the intention-to-treat principle and included all randomized subjects. For subjects who withdrew from the trial, data from the last available visit were carried forward up through week 12 for purposes of statistical analysis of the continuous efficacy outcomes. For the global assessment of response to study treatment, subjects who withdrew from the trial were considered to be in the “worse” category for purposes of analysis. Sensitivity analyses, e.g., using other methods of imputation or including only subjects who completed 12 weeks of follow-up, revealed no major impact of the imputation strategy on the results. No data imputation was performed for the analysis of the data from the open-label phase.

Blinded phase.

Seventy subjects (87.5%) successfully completed the blinded phase (figure 1). Ten subjects were withdrawn during the 12-week blinded phase, four on MMF and six on placebo (table e-1 on the Neurology ® Web site at www.neurology.org). Eight of these were withdrawn because of MG progression requiring alternative therapy (treatment failure), three on MMF and five on placebo. One subject on MMF withdrew at the subject’s request, and one on placebo withdrew at the physician’s request. Two subjects on placebo had MG progression that did not require alternative therapy or withdrawal.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1 Subject flow and reasons for withdrawal during the blinded phase.

Although the total QMG score improved over time in both treatment groups, the extent of improvement in the MMF group was similar to that in the placebo group at week 12 (treatment effect = −0.37, 95% CI −2.36 to 1.61, p = 0.71) (figure 2). The treatment groups did not significantly differ with respect to change over time in any of the secondary outcome variables for efficacy, including change in AChR-Ab level (table 2). The global assessment of response to study treatment also reflected similar improvement in both treatment groups (p = 0.77, table e-2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2 Mean change in quantitative myasthenia gravis score over time by treatment group

Vertical bars indicate one SEM.

Open-label phase.

Sixty-eight subjects entered the open-label phase at week 12, 37 from the MMF group and 31 from the placebo group of the blinded phase. Two subjects were withdrawn during this phase of the study, one because of MG progression that required alternative therapy, and another who died from cardiac and renal failure. The prednisone dosage by time curves (figure e-1) and the mean prednisone dosage at week 36 (MMF: 10.9 ± 12.2 mg/day, n = 33; placebo: 8.1 ± 8.7 mg/day, n = 28; p = 0.26) were similar for subjects from both treatment groups of the blinded phase, thus there was no apparent greater steroid-sparing effect in subjects who had received MMF during the initial 12 weeks. The QMG score during the open-label phase was not different between patients in the two arms (figure e-2).

The frequencies of some adverse events appeared to be somewhat higher in the MMF-treated group than in the placebo group but none of the differences reached significance (table 3). Details of adverse events and compliance can be found in the appendix.