Biotech Values

[DewDiligence]

Trial-design info from last week’s Macugen panel:

An introductory presentation by Wiley Chambers, deputy director of the FDA department in charge of ophthalmic-drug review, contained technical details about the FDA’s requirements for pivotal-trial design in AMD. (These details by and large escaped media attention.)

1. In order to detect relative rare side effects, the FDA prefers that each of the two pivotal trials contain at least 500 patients.

2. At least one of the two pivotal trials must be a “superiority” trial. The other trial may be a "non-inferiority" trial but need not be.

3. In a non-inferiority trial, the therapy being tested must preserve at least half the treatment benefit of the comparator therapy at the 95% (two-sided) confidence level.

4. When multiple doses are tested in the same trial, the Hochberg adjustment will determine the threshold for statistical significance (#msg-3891544).

5. Acceptable primary endpoints are: a) The proportion of patients losing fewer than 15 letters (3 lines) relative to baseline (this was the primary endpoint in the phase-3 Macugen trials); or b) The proportion of patients gaining 15 or more letters relative to baseline; or c) The difference between the active arm and the control arm in the mean change of letters gained/lost relative to baseline. In case c), a minimum difference of 15 letters is required.

6. Safety and efficacy must be evaluated at two years; however, an application may be made to the FDA based on one-year data (as EYET has chosen to do).

7. Patients must be evaluated at least as often as every three months.

8. To account for patient drop-outs, the principal analysis employed by the FDA will be intent-to-treat/last-observation-carried-forward (LOCF). Intent-to-treat/worst-case and per-protocol analyses may be used at the FDA’s discretion to ascertain the robustness of the LOCF data.

--
Is all that perfectly clear?