Learning from the past: reflections on recently completed myasthenia gravis trials.

Recently competed clinical trials of therapeutics for myasthenia gravis have varied widely in design, but also perhaps in less explicit ways. We explore ways in which these design characteristics may have influenced recruitment and results, as well as the implications for forthcoming studies. Trial eligibility criteria may inadvertently select for incident versus prevalent cases or patients with relatively mild versus more severe disease. Trial enrichment with patients who have relatively mild disease may limit the sensitivity of the trial to detect a therapeutic effect. Enrichment for patients with more severe disease may introduce confounds caused by regression toward the mean. Overly narrow eligibility may limit the generalizability of results. An exclusive focus on incident cases may hamper recruitment, as may many other factors, such as access to the experimental therapeutic treatment outside of the trial or following completion of the double-blind treatment period. We illustrate how other design characteristics (e.g., treatment duration, strategy for steroid tapering, selection of the primary outcome, and principal analytic approach) may affect the sensitivity of a trial to demonstrate therapeutic effects. Finally, we consider the importance of placebo effects, being careful to differentiate these from therapeutic effects observed in the placebo group, and discuss how the use of combined outcome measures may minimize placebo effects.