What were the baseline and end-of-treatment levels of testosterone for patients in both arms of the study? Repros did not disclose.
Was the testosterone normalization endpoint reached with statistical significance? Repros did not disclose.
The study's sperm count endpoint is similarly messy.
Twenty of 134 men on Androxal reported sperm counts that fell 50% below their baseline levels compared to 2 of 47 men on placebo. While numerically worse, Repros said the Androxal was statistically non inferior to placebo but only just barely. To meet the study's sperm-count endpoint, Androxal could be no more than 20% worse than placebo. The reported statistical difference was 18.3%.Last March, Repros admitted that one of the clinical trials sites fabricated baseline sperm counts for patients. As a result, the company analyzed the study omitting patients from that troubled site. Here, the statistical worsening in sperm count reduction was 19.6%, meaning the study came within a whisper of failing on this endpoint. Repros also slipped in a change to the reporting from the first phase III study. Tuesday, the company said 16 of 113 Androxal-treated men reported sperm counts lower than 50% from baseline. Last March, it was 15 of 113 patients. Under the SPA agreement, Androxal must meet both primary endpoints in both phase III clinical trials in order for the study to be considered successful. On Tuesday, Repros said it "believes" the sperm-count endpoint was met. Why not just make a definitive statement? Repros CEO Podolski blamed his lawyers for wanting to be conservative. Repros must complete an Androxal safety study plus finish some preclinical work before submitting for FDA approval in the middle of next year. The company is also still working through a messy challenge to the patents for Androxal by a New York urologist who claims Podolski stole the idea for the drug from him. -- Reported by Adam Feuerstein in Boston. Follow @AdamFeuerstein