Experiences with an adaptive design for a dose-finding study in osteoarthritis.

Aims Many new potential medicines fail in early clinical development. AZD1386 is a TRPV1 antagonist and was developed for treatment of osteoarthritis pain at AstraZeneca. Following preclinical studies to characterize the compound, translational studies and first time in man studies the challenge is great to select the right doses and study population for the first clinical studies on efficacy in phase II of the drug development process. Methods Different design alternatives in the planning of the dose-finding study for AZD1386 in OA were analysed with regard to statistical considerations, timelines and costs for the different options. Results Based on datasimulations and cost/benefit analysis a 4 week study with adaptive design with interim analysis at 2 weeks for possible dose adjustment or futility stop was chosen. The study was initiated with AZD1386 at to doses 30 and 90 mg and placebo. Based on the fact that the one-sided p-value for the primary analysis (comparison of 90 mg versus placebo) was 0.2891 and larger than 0.2, the data monitoring committee decided that the study was to be stopped after the interim analysis due to lack of efficacy. Conclusions The chosen adaptive design in this Phase IIa/b study enabled an early futility stop. This meant several advantages. Further exposure of patients of a non-efficacious drug could be stopped and resources could be re-allocated earlier as well as monetary savings.