Peter, I found this reference to the discrepancy in the carrier vs noncarrier groups in your paper:
>A number of reasons might account for these findings: first, greater efficacy was found in subjects who completed the study and a greater proportion of noncarriers were completers; second, more advanced A pathology in APOE 4 carriers23 may have affected the clinical response; finally, the differences observed in these exploratory analyses could be due to chance.<
There might be a correlation between carrier status/and/or progression of AD with frequency of VE, 10 of the 12 VE cases occured in carriers and (interestingly) 11 of the 12 cases were at the 1mg dose.
58% of the carriers completed 71% of the placebo completed 77% of the noncarriers completed