Nearly Half of CABG Patients Have New-Onset AF, but It Often Doesn’t Last

COPENHAGEN, Denmark—Almost half of patients with no history of arrhythmias who undergo CABG for three-vessel or left main coronary artery disease have postoperative atrial fibrillation (AF), but the burden of these new-onset arrythmias is low overall beyond 1 month, new prospective data show.

The findings should spur doubt about the need for oral anticoagulation in this population given their low stroke risk, said Florian E. M. Herrmann, MD (LMU University Hospital, Munich, Germany), who presented the data during a late breaking clinical science session at the 2025 European Association for Cardio-Thoracic Surgery (EACTS) meeting.

“The incidence of new onset AF within the first year after CABG is higher than previously reported,” he reported. “However, the AF burden is very low during the first year, especially later than 30 days after surgery. Given the known bleeding risk with oral anticoagulants, the very low AF burden questions current ESC guidelines recommending long-term oral anticoagulation. If oral anticoagulation is initiated for new-onset AF after CABG, the indication should be reassessed 30 days after surgery.”

The prevalence of AF following CABG has worried physicians for years, with the EXCEL trial estimating that about one in five patients develop it anew following their surgeries. Findings from the SEARCH-AF study show a similar rate with enhanced cardiac rhythm monitoring. The American Heart Association issued a scientific statement in 2023 emphasizing the need to monitor for acute AF, especially following cardiac surgery.

1-Year Follow-Up

The study, simultaneously published in JAMA, enrolled 198 patients (mean age 66 years; 87.4% male) undergoing CABG for three-vessel or left main CAD at one of two German centers between November 2019 and November 2023. None had a history of AF or other arrhythmias, and all patients were simultaneously implanted with a cardiac monitor (Reveal LINQ LNQ11; Medtronic) that regularly transfers heart rhythm information to a database.

At baseline, median CHA₂DS₂-VASc score was 3, 58% of patients received beta-blockers, and none received amiodarone prior to surgery. Median postoperative hospital length of stay was 10 days.

Among the 192 patients with 1 year of follow-up data, almost half (n = 95) reported new-onset AF detected by continuous monitoring (48% cumulative incidence) with a median AF burden of 0.07%, or 370 cumulative minutes spent in AF. The median AF burden decreased with time: 3.65% for days 1 through 7 (368 minutes), 0.04% for days 8 through 30 (13 minutes); and 0 for days 31 through 365 (0 minutes).

Three patients had an AF episode lasting longer than 24 hours following discharge, and none of these were detected with standard monitoring, which consisted of 2-5 days of postoperative telemetric monitoring and a 12-lead ECG prior to discharge.

The median time to incident AF episode was 3.3 days and the median episode length was 6 minutes. About two-thirds of incidents were asymptomatic, and another two-thirds were not detected by standard monitoring.

A similar proportion of patients in the study with and without new-onset AF experienced at least one MACCE within 1 year after surgery (7.4% vs 6.8%). There were also no statistical differences in terms of death (3.2% vs 1.0%) or stroke (0 vs 1.9%). At both discharge and 1 year, more patients with new-onset AF than those without AF events were on oral anticoagulation.

AF Burden Is Telling

Following the presentation, panelist Matthias Siepe, MD (University of Bern, Switzerland), commented that the AF rates observed in this CABG population are “surprisingly high,” adding that he “would’ve guessed one-third.” However, it’s not clear if other clinical endpoints will be “triggered” by this information given so much of it happens very early following surgery.

“That’s exactly the point,” Hermann said. “The study shows a higher incidence but a very low burden, especially after 30 days. So, while we expected that there would be much more atrial fibrillation, we actually showed the opposite—that there’s a low, low burden of atrial fibrillation. That should maybe let us question if we should really be anticoagulating so many patients.”

Beyond that, panelist Patrick Myers, MD (Lausanne University Hospitals, Switzerland), commented that the study’s strength lies in calculating AF burden above a simple binary, as it’s reported in many SAVR vs TAVI trials. “I absolutely love that you have this as an AF burden showing that it may be 40% or 50%, yes. But it’s actually of no consequence in 6 months.”

During the discussion, audience member Anders Jeppsson, MD, PhD (Sahlgrenska University Hospital, Gothenburg, Sweden), commented that the CHA₂DS₂-VASc score is one valid measure to use when planning oral anticoagulation in these patients, but noted that not only stroke but also bleeding risk should be included. The latter is possible to assess, “but there’s not yet any guidance,” he said.

But the main message from this study is that after 1 month, the AF burden is low or nonexistent Jeppsson agreed.

“So why should we treat them?” Jeppsson asked. “We know from other studies that if you initiate oral anticoagulation directly after surgery, then the patients still have it after 1 year, and many of them lifelong. And it also makes sense that we don’t see an effect. Study after study have now shown an increased bleeding risk, but no difference in stroke risk with oral anticoagulation. Now we know why: because the patients do not have any atrial fibrillation, only that bleeding risk.”