Feasibility of a randomized single-blind crossover trial to assess the effects of the second-generation slow-release dopamine agonists pramipexole and ropinirole on cued recall memory in idiopathic mild or moderate Parkinson’s disease without cognitive impairment

BACKGROUND: The aim was to assess the feasibility of a single-centre, single-blind, randomized, crossover design to explore the effects of two slow-release dopamine agonists, ropinirole and pramipexole, on cued recall in Parkinson’s disease. As the design required a switch from the prescribed agonist (pramipexole-to-ropinirole, or ropinirole-to-pramipexole), the primary objectives were to (a) examine the efficacy of processes and procedures used to manage symptoms during the washout period and (b) to use cued recall estimates to inform a power calculation for a definitive trial. Secondary objectives were to assess consent and missing data rates, acceptability of clinical support for the OFF sessions, experience of the OFF sessions and of agonist switching, barriers-to-participation for patients and informal caregivers. METHODS: Patients were randomized in a 1:1 ratio to two treatment arms and stabilized on each agonist for 6 weeks. The arms differed only in the sequence in which the agonists were administered. Cued recall was assessed ON medication and, following a washout period resulting in 93.75% agonist elimination, OFF medication. RESULTS: A total of 220 patients were screened: 145 were excluded and 75 invitations to participate were sent to eligible patients. Fifty-three patients declined, 22 consented and 16 completed the study. There were no serious adverse events, and rates of non-serious adverse events were equivalent between the agonists. Using the largest standard deviation (SD) of the ON–OFF difference cued recall score (inflated by ~25% to give a conservative estimate of the SD in a definitive trial) and assuming an effect of at least 10% of the observed range of OFF medication cued recall scores for either agonist to be clinically important, a main trial requires a sample size of just under 150 patients. The consent and missing data rates were 29 and 27% respectively. The washout period and the preparation for the OFF sessions were acceptable, and the sessions were manageable. The experience of switching was also manageable. Barriers to participation included concerns about disease stability, side effects, research process, carer workload and accessibility of the information sheet. CONCLUSIONS: This study presented challenges to recruitment both in design and execution, and while it was a major aim of the study to assess this, evaluation of these challenges provided the opportunity to explore how they could be overcome for future studies. TRIAL REGISTRATION: EudraCT 2012-000801-64