RE: GTCB is trying to show non-inferiority...
"What we need to know is what percentage of HD patients given plasma derived ATIII had a symptomatic DVT. From the European trial of rATIII, we already know that zero patients out of 14 HD patients had a symptomatic DVT."
No, we can be fairly certain that the rate with plasma AT is near 0. But to get the inferiority delta you need to know what the rate with and w/o AT is historically (and then guess the SPA negotiated value).
For example. Suppose AT historically had 2% symptomatic DVT cases, and placebo comes in with 20%. Then the FDA would say that in order to meet the NI goal, rAT would have to be stat sig better than the comparison AT arm plus something like 10%. Thus, (wrt efficousy) the comparison arm data actually has very little effect on the actual outcome! What is much more important is the historical rate for non AT treated patients.
[OK, I pulled 10% out of my ass. It would be something less than 18% in this example, but given the already sketchy trial I doubt they FDA would be overly generous here].