Pomp,

The data are blinded, so the the two groups are not the apabetalone and control groups. I believe the two groups for which the 'change' has been observed are the intital scores and the subsequent scores for those patients for which (at least) two MoCa scores had been obtained at the the time the preliminary analysis was obtained.  In the cognition study, patients have a MoCa test annually.

What the sentence means, I believe, is that 1) given the observed change in MoCa score between the initial and subsequent assessments for 108 patients in the study, 2) the observed standard deviation in the scores at the time of the subsequent assessments, and 3) making the assumption that approximately half the 108 pateints are in each arm of the study, they have calculated, a difference in MoCa scores between the two groups of 2 (or more) points would be signficant.  They don't know at this point if there is a difference between the groups, but they now know that they have the ability to detect (and establish significnace for) a change of that magnitude if it exists.  Likely, the power to detect an even smaller difference between the apabetalone and placebo groups will increase as more patients have multiple MoCa scores and the duration of the trial increases.

They may be excited because the mean (or median) MoCa score did not decrease much (or even stayed constant or increased) in the subsequent assessments but the standard deviation of scores increased.  This would be consistent with (but not prove) the possibility that the apabetalone treatment was working to slow the delcine in cognition expected in old, sick patients but the placebo group was still declining at the expected rate.  Thus, the standard deviation in MoCa scores for the entire group combined would increase as a difference in median score developed between the two arms over time.

Hope this helps,

Jupe