FutTheWuk,

In BETonMACE, there was a pre-specified statistical analysis plan that stated: "If superiority of apabetalone is established for the primary endpoint,a sequential gate-keeping approach will be used to control the overall Type I error rate in testing the key secondary endpoints." It is a hierarchal way to test multiple endpoints that requires the higher ranked endpoints to first pass statistical significance prior to lower ranked endpoints being "officially" tested.

Since BETonMACE failed the primary endpoint, the "gate" was never opened on the secondary endpoints and all subsequent tests were "officially" never done. The statistical tests can "unofficially" still be run (as presented at AHA and in the webcast); however, any results are considered exploratory and can't be used to formally prove or disprove a hypothesis. These p-values used in exploratory analyses may also be nominal p-values that may not be adjusted for multiple testing and other penalties. I'm going to stop there before anyone mistakes me for a biostatistician, which I am not!

BDAZ