Abstract 4141927: Risk of myocardial infarction in paroxysmal vs. non-paroxysmal atrial fibrillation: an individual patient-level data analysis of 71,466 patients from COMBINE AF

Background: Prior data suggest the MI risk may be higher with paroxysmal AF (PAF) vs. non-paroxysmal AF (non-PAF). Proposed mechanisms include tachycardia-induced oxidative stress (via LOX-1) with microvascular flow abnormalities, ischemia downstream of a fixed coronary obstruction, and plaque rupture. Methods: We compared MI rates in pts with PAF vs. non-PAF in COMBINE AF, a patient-level metanalysis of 4 RCTS of DOACs vs warfarin (ARISTOTLE, ENGAGE AF-TIMI 48, RE-LY,ROCKET AF). Secondary endpoints were ischemic stroke and CV death. Cox proportional-hazards models stratified by trial and adjusted for elements of the CHADS-VASc score were constructed. Sensitivity analyses were performed across subgroups, omitting pts on lower-dose DOAC regimens, and accounting for competing risk of death. Results: Of 71,466 pts, 16,609 (23%) had PAF at enrollment. Pts with PAF vs non-PAF were similar age (median 72 vs 72. P=0.15), but more likely women (43 vs 36%), with prior CAD (35 vs 31%), and on aspirin (41 vs 32%); but less likely Asian race (12 vs 15%) or with CHADS-VASc score > 4 (59 vs 60%), p < 0.002 for each. During >160,000 pt-yrs of follow-up, 1033 MIs occurred: 277 (1.67%) in pts with PAF vs 766 (1.40%) in pts with non-PAF, corresponding to rates of 0.81% and 0.70% per pt-year. The HR adj for MI with PAF vs non--paroxysmal AF was 1.17 [1.02-1.35], p=0.028 (Fig). Ischemic stroke occurred in 364 (2.19%) vs 1425 (2.60%) pts with PAF vs non--paroxysmal AF (HR adj 0.81 [0.72-0.91], p<0.001). CV death occurred in 625 (3.77%) pts with PAF vs 3027 (5.52%) with non-paroxysmal AF (HR adj 0.75 [0.68-0.81], p<0.001). No significant effect modification of pt characteristics on MI risk by AF pattern were present, although a trend was seen for higher MI risk with PAF vs non-PAF in pts with prior CAD (HR 1.34 [1.12-1.59]) vs no prior CAD (HR 0.96 [0.76-1.22]), P INT 0.06. Results were consistent across trials, by anticoagulant, excluding lower dose DOACs, and accounting for competing mortality risk. Conclusions: This individual patient-level metanalysis of 71,466 pts from COMBINE AF shows that the adjusted risk of MI is higher in pts with PAF than non--paroxysmal AF, while the adjusted risks of ischemic stroke and CV death were lower in pts with PAF.