My conclusions about MACE:
Little agreement among studies. In a string of about 10, none were identical.
Even looking at top journals, little agreement.
Have to separate cardiac and non-cardiac studies, where the cardiac ones get sophisticated about revacscularization, etc. We will use non-cardiac.
Have to separate RCT and retrospective. (RCT doesn’t use codes.)
No one, not even in top journals, validate MACE.
No one uses two codes for MACE or double checks inpatient status or troponin. But they may have specialized databases, such as inpatient only.
Seems like “3-point” MACE is good for us. It is defined as death (disagree on all cause or cardiovascular), MI, stroke (disagree on ischemic or hemorrhagic too).
Our definition was any of these:
MI with inpatient visit or with high troponin
Stroke ischemic and not hemorrhagic. We check that it is with inpatient visit or that it is with a head imaging study.
Death with a CV diagnosis.
Based on the above review, it would be reasonable to do all cause death, stroke without inpatient or imaging, or MI without inpatient or troponin. This would gather more patients and better match every other retrospective study of MACE I looked at. We are being more restrictive.
I don’t see anyone checking for two codes, especially because stroke and MI are episodes. For RA, I do see multiple codes used, but that’s a chronic disease.
And we could also validate, but that would be beyond the literature.