The delayed-start study in Parkinson disease

Drs. Ahlskog and Uitti correctly point out the difficulty in identifying a disease-modifying therapy for Parkinson disease (PD).1 While there are many promising candidate drugs, endpoints employed in the past did not permit detection of a disease-modifying effect because of potential symptomatic or pharmacologic confounds.2 A validated biomarker would obviously be helpful, but one does not currently exist. The delayed-start design was proposed to address these problems.3 In this 2-period design, patients are randomized to receive placebo or active drug in the first period, while patients in both groups receive the active drug in the second period. Keys to defining a positive outcome are evidence that early treatment provides benefits at study end that cannot be achieved with delayed-treatment despite both groups receiving the same treatment, and evidence that the slopes of deterioration in the Unified Parkinson's Disease Rating Scale (UPDRS) in the 2 groups are not converging, indicating that the benefit is not readily explained by a symptomatic effect. There are many issues that must be addressed in designing a delayed-start study. The periods must be sufficiently long to permit disease modification to occur in period 1, and for the drug to reach its full symptomatic effect in period 2. Dropouts (especially differential dropouts between the groups) must be minimized as the primary analysis requires data from each of the 2 periods. Analytic methods must be used to account …