Thanks, not trolling and appreciate your insight on this. The thing is these liver data are worse than ph2 (9.7% vs 8% >3x ULN; 3.3% vs unspecified >5x).

The issue in the data would seem to be that there’s a potentially larger than expected number of patients for whom the liver issues are *not* transient. You noted that the study protocol specified discontinuation for unresolved ALT: “If a second measurement of ALT was greater than 5 times ULN or a second measurement of bilirubin was greater than 2 times ULN, study medication was permanently discontinued.”

So now we know the size of that non-transient group was 35, about 3% of treated patients (not sure what the other 10 incremental discontinuations are due to).

I guess my question, which affects approvability, is how do we know these patients weren’t sicker/more likely to experience cardiac events than the control population? Liver disease generally is associated with increased MACE risk. Since the trial design systematically excluded patients who had persistent liver reactions, could that affect the relative risk of the remaining group?