A cardinal sin in clinical development is to change trial design between Phase II and Phase III. But keeping the faith also is not a guarantee of success, as Connetics Inc. found when a Phase III study of its human recombinant relaxin hormone failed to repeat positive Phase II results in scleroderma.

The 239-patient Phase III trial of relaxin to treat diffuse systemic sclerosis, a disorder characterized by fibrotic changes in the skin, blood vessels, skeletal muscles and internal organs, missed its primary end point of a reduction in the Modified Rodnan Skin Score. Patients on placebo had mean reductions in skin scores of 4.9 after 24 weeks, while those on 10 m