Yes it is. Take Lipitor for example. On Lipitor's label it speaks about cases of patients being over 3x ULN. Ranges of APPROVED DOSES range from approximately <1% to 3%. In my opinion, Repros believes that efficacy is equivalent for all 3 doses so why create noise/distractions to dampen the efficacy signal by taking a dose forward that is ~4% when the other 2 doses are below 1%. The only "beef" that anyone could have that I can see is POSSIBLY that a more rigorous look at the data could have teased this out after phase 2 data (phase 2 is done in part to chose phase 3 dose).
The FDA does not do a benefit/risk analysis, but determines efficacy and safety. This might be just semantics for some, but this is a statutory requirement for the FDA.
Agreed that what you say is true by statute. BUT it is very much not true in practice. At least one instance of this has explicitly surfaced recently - see VNDA FDA oops after the initial rejection.
PS Tentatively (i.e. from memory and without having reconstructed the whole story) agree that there is a concern that as dose is reduced endometrial thickness goes up and so does breakthrough bleeding. I would not necessarily rate this a drug killer however - it would depend upon frequency and magnitude. However again agree that this is something they will need to address head on at some point (as they admirably did for the dose related liver tox issue).