Topcoin and Bear,

I think the true sample size for the cognition experiment is the 467 patients who completed the initital MoCa test (and who comprise 19% of the total BoM patient population - those who are over 70, although at least one 69-yr-old appears to have been included).  Of these, 53% (n=246) showed impaired cognition at the beginning of the study (average MoCa=22, range from 7 to 25).  The remaining 221 (47%) patients are those who did not show impairment at the beginning of the trial but are still part of the cognition experiment (all of these had MoCa scores of 26 or greater by definition, and, if you do the arithmentic, you can calculate that their median was about 28).

You can see from the initial data that the range in MoCa was great (from 7-30).  People with mild cognitive impairment have values around 22, those with AD have values of around 16 (or worse, presumably).  The change in MoCa score will be evaluated in all patients in the experiment, regardless if they showed impairment at the beginning or not.  Hopefully, about half each of the 'impaired' and 'normal' older patients were allocated to each of the placebo and apabetalone arms.  It would be great if the apabetalone treatment actually improved the MoCa score in impaired individuals, but a more realistic hope would be that the patients on apabetalone showed a smaller decrease in MoCa over time.