Preexisting—But Not New-Onset—A-fib Predicts Worse Long-term Outcomes After TAVR

Download this article's Factoid (PDF & PPT for Gold Subscribers)  

Patients with atrial fibrillation (A-fib) diagnosed before transcatheter aortic valve replacement (TAVR) face increased risks of stroke and death in the first year after the procedure compared with those without the arrhythmia at baseline, according to a registry study published online September 25, 2014, in the American Journal of Cardiology. However, contrary to previous data, new postprocedure A-fib does not appear to be prognostically significant.

Based on the findings, the study authors advocate factoring a prior diagnosis of A-fib into TAVR risk scores.

Lior Yankelson, MD, PhD, and colleagues of Tel Aviv Sourasky Medical Center (Tel Aviv, Israel), analyzed data from 380 patients who underwent TAVR—predominantly with the CoreValve bioprosthesis (Medtronic; Minneapolis, MN) delivered transfemorally—at their institution between September 2008 and April 2013 and were enrolled in the Tel Aviv Prospective Angiography Study registry. Dual antiplatelet therapy was prescribed for 6 months, with anticoagulants added as needed according to guidelines.
Overall, 118 patients (31%) had prior A-fib (19% paroxysmal and 12% permanent), while 31 (8.2%) experienced new-onset A-fib, defined as A-fib lasting at least 30 seconds during the procedure or the subsequent 30 days; 6 more patients developed A-fib after 30 days. CHADS2 and CHA2DS2VASc scores were similar between those with and without prior A-fib.
Over a mean follow-up of 528 days, there were 19 new episodes of stroke (5%), 6 occurring within 30 days and 18 within 1 year. Overall, 68 patients (20%) died, 12 within 30 days and 58 within 1 year.

Patients with new-onset A-fib were similar to those without the complication in terms of degree of aortic stenosis, comorbidities, and procedural features. However, the former had a slightly lower ejection fraction, which was the only predictor of new-onset A-fib (P = .04).

Overall, new-onset A-fib was not associated with higher cumulative rates of stroke or mortality at either 30 days or 1 year (table 1).

Table 1. Outcomes: New-Onset vs No A-fib


New-Onset A-fib

No A-fib

P Value


    30 Days

    1 Year











    30 Days

    1 Year











However, patients with new-onset A-fib had a longer average hospital stay than those with no A-fib (10.5 vs 7.5 days; P = .04).

Prior A-fib Worsens Prognosis

Compared with patients without A-fib, those with A-fib at baseline were older and had a higher rate of previously implanted pacemakers, worse kidney function, and a higher EuroSCORE. These patients experienced increased rates of stroke and mortality at 1 year and a trend toward more stroke at 30 days compared with those without prior A-fib (table 2).

Table 2. Outcomes: Prior vs No Prior A-fib


Prior A-fib

No Prior A-fib

P Value


    30 Days

    1 Year











    30 Days

    1 Year









< .001


Notably, however, the stroke rate was low, with only 1 event in each group.

On multivariate analysis adjusted for known stroke predictors, prior A-fib showed a trend toward increased stroke risk at 30 days (OR 8.7; P = .058) and was a predictor of stroke at 1 year (OR 5.9; P = .015). Similarly, multivariate Cox analysis showed that the absence of A-fib was associated with increased survival odds compared with baseline A-fib (HR 2.2; 95% CI 1.33-3.86) but not compared with new-onset A-fib (HR 1.5; 95% CI 0.58-4.08).  

Discrepancy Over Impact of New-Onset A-fib Explained

The link between chronic A-fib and stroke is well established for multiple types of procedures, Josep Rodés-Cabau, MD, of the Quebec Heart and Lung Institute (Quebec City, Canada), told TCTMD in a telephone interview. Moreover, new-onset A-fib has been tied to increased stroke and mortality in a prior study.  The fact that the latter type of A-fib was not implicated in the current analysis may be due in part to its relatively small sample size, he suggested.

But it may also be attributable to differences in how new-onset A-fib patients were managed with anticoagulation, he suggested, adding that although the authors say guidelines were followed, the guidelines are not entirely clear.

Dr. Rodés-Cabau observed that some physicians are reluctant to prescribe anticoagulants to elderly TAVR patients who have had a stroke after only a brief episode of A-fib. “But what we have learned—even though there’s no randomized evidence for it—is that probably these patients should receive anticoagulation because their CHADS2 scores are typically very high. If you are more aggressive in terms of anticoagulant therapy, maybe you can [reduce] the cardioembolic risk associated with A-fib. Now, we administer IV heparin or low molecular weight heparin, and stop one of the [post-TAVR] antiplatelet agents—usually clopidogrel, to minimize bleeding risk.”

In a telephone interview with TCTMD, Ron Waksman, MD, of MedStar Washington Hospital Center (Washington, DC), suggested another possible explanation for the current study’s failure to link new-onset A-fib to stroke and mortality: in some cases A-fib may have been transient, meaning the myocardium did not suffer a significant insult.

A Possible Role for LAA Closure?

To the extent that new-onset A-fib is prognostically important, he said, an alternative strategy that deserves study—especially if the Watchman device (Boston Scientific; Natick, MA) is approved—is closure of the left atrial appendage at the time of TAVR. This option would reduce the risk of ischemic stroke without increasing that of hemorrhagic stroke, as warfarin does, he added.

Drs. Waksman and Rodés-Cabau agreed that many instances of paroxysmal A-fib probably go undetected, and that what is termed ‘new-onset’ A-fib may in fact not be, blurring somewhat the distinction between chronic and new-onset cases.

“You don’t recognize [transient A-fib] until you focus on the ECG,” Dr. Waksman said, “but I think if patients do develop new-onset A-fib, maybe we can minimize [the consequences] by treating it more aggressively.”

Like the authors, he advocated including prior A-fib into the STS and EuroSCORE risk scores, noting that it is relevant not just to TAVR but to most interventions.

Yankelson L, Steinvil A, Gershovitz L, et al. Atrial fibrillation, stroke and mortality rates following transcatheter aortic valve implantation. Am J Cardiol. 2014;Epub ahead of print.

Related Stories:


Preexisting—But Not New-Onset—A-fib Predicts Worse Long-term Outcomes After TAVR

  • The paper contains no statement regarding conflicts of interest for Dr.Yankelson.
  • Dr. Rodés-Cabau reports serving as a consultant to Edwards Lifesciences and St. Jude Medical.
  • Dr. Waksman reports no relevant conflicts of interest.