Population modeling of the relationship between ropinirole systemic exposure and efficacy in Parkinsonís disease
Debra Tompson, Ruth Oliver-Willwong, Chao Chen
GlaxoSmithKline, Harlow, UK
Objective: Ropinirole is a nonergoline dopamine D2 agonist used for the treatment of Parkinson's disease. The objective of this work was to explore the relationship between systemic exposure to ropinirole and the primary efficacy endpoints from two phase III studies.
Methods: In Study 1, efficacy of ropinirole 24 hour prolonged release (PR) was compared with immediate release (IR tid) at the same daily dose. The treatment phase was 36 weeks including a 12-week titration followed by three eight-week maintenance periods (IR-IR-PR, IR-PR-PR, PR-IR-IR, or PR-PR-IR). In Study 2, patients whose symptoms were not adequately controlled by L-DOPA were randomized to ropinirole PR or placebo for 24 weeks. Doses were individually titrated up to 24 mg/day. Sparse PK samples were collected in both studies. A one-compartment model with 1st-order absorption and 1st-order elimination was built using Phase II data. Dosing was via two depot compartments, for IR and PR respectively. The sparse data from the Phase III trials were added and parameters re-estimated. Logistic regression was conducted using NONMEM for the probability of a patient being a responder (>/= 30% reduction in UPDRS total motor score in Study 1 or >/=20% reduction in awake time spent "off" in Study 2) as a linear or hyperbolic function of AUC(0-24)ss derived from individual clearance prediction.
Results: The relationship between the probability of a patient being a UPDRS responder and AUC(0-24,ss) values was best described by a linear function. The logistic regression showed a flat probability of response (between 0.6 and 0.8) over the approximate five-fold exposure range and the relationship was comparable for both IR and PR. The relationship between the probability of a patient being an awake time "off" responder and AUC(0-24,ss) was also best described by a linear function. The probability of a response increased from 0.4 for placebo to ~0.9 at the high end of exposure range. Both the baseline and slope were reasonably well estimated (%RSE ~60% and ~25%, respectively).
Conclusions: Both analyses were limited to data at near top of response range. Although newly diagnosed patients receiving ropinirole monotherapy are unlikely to benefit further from higher doses, more advanced patients receiving the drug in combination with L-DOPA can achieve greater efficacy with higher doses.