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

Nearly one-third of patients undergoing transcatheter aortic valve replacement (TAVR) experience new-onset atrial fibrillation (A-fib), increasing the risk of stroke and systemic embolism, according to a study published online December 14, 2011, ahead of print in the Journal of the American College of Cardiology. More than half the cases occurred either during the procedure or within the first 2 days.

Investigators led by Josep Rodés-Cabau, MD, of Laval University (Quebec City, Canada), analyzed prospective data on 138 patients with severe symptomatic aortic stenosis who underwent TAVR at their institution between May 2007 and May 2011. All procedures were performed with the balloon-expandable Edwards Sapien or Sapien XT aortic valve devices (Edwards Lifesciences, Irvine, CA).

Procedural success was 93.5%, and the 30-day mortality rate was 7.3%. Eight patients (5.8%) suffered an ischemic stroke following the procedure.

A-fib Appears Early

New A-fib was observed in a total of 44 patients (31.9%) within 30 days, with a median onset of 48 hours after the procedure. In fact, up to 40% of A-fib episodes occurred either during (36.3%) or within 24 hours (4.6%) of TAVR. Of all A-fib episodes, 22.7% resolved spontaneously; the other cases required pharmacological (IV amiodarone; 43.2%) or electrical (34.1%) cardioversion. Only 1 patient remained in A-fib at discharge.

Anticoagulation with IV heparin and warfarin was initiated immediately after A-fib diagnosis in 34 of 44 patients. Those who experienced new-onset A-fib had longer hospital stays than those who did not (9 days vs. 6 days; P = 0.0004).

Predictors of A-fib

Multivariate analysis identified 2 independent predictors of new-onset A-fib:

• Left atrial size (OR 1.21 for each 1 mm/m² increase; 95% CI 1.09-1.34; P = 0.0001)
• Transapical approach (OR 4.08; 95% CI 1.35-12.31; P = 0.019)

An atrial size of 27 mm/m² or larger as measured by echocardiography was determined to be the best cutoff for prediction of A-fib within 30 days of TAVR, yielding 67% sensitivity and 61% specificity, with an area under the ROC curve of 0.71 (95% CI 0.62-0.80; P = 0.0001). The incidence of new-onset A-fib in patients with both predictive factors (an atrial size of 27 mm/m² or larger undergoing transapical TAVR) was 51% compared with 12% among patients with neither factor.

A-fib Increases Cardioembolic Events

At 30 days, stroke tended to be more common in patients with new-onset A-fib than those without the condition. The incidence of stroke or systemic embolism was higher in the new-onset A-fib group. However, mortality rates were similar (table 1).

Table 1. Events at 30 Days

^a After adjustment for baseline differences between the groups.

Over a median follow-up of 12 months, rates of stroke and composite of stroke or systemic embolism remained higher in patients with new-onset A-fib, while there were no differences between the A-fib and no-A-fib groups for systemic embolism or death (table 2).

Table 2. Cumulative Events 1 Year

^a After adjustment for baseline differences between the groups.

There were 4 early strokes (≤ 24 hours after TAVR) and 5 late strokes (> 24 hours after TAVR). Only 1 of the 4 early strokes occurred in a patient with new-onset A-fib, while all 5 late strokes occurred in patients with the condition. Interestingly, among new-onset A-fib patients, the incidence of stroke or systemic embolism at 30 days was 40% in those who did not receive immediate anticoagulation compared with 2.9% in those who did receive such therapy (P = 0.008).

In a telephone interview with TCTMD, Dr. Rodés-Cabau said the study highlights that A-fib following TAVR, which has been distinctly under-reported compared with the surgical literature, is more frequent than previously thought. 

He acknowledged that the incidence of A-fib in the study may be somewhat skewed by the fact that the condition was tracked meticulously and includes brief episodes that occurred during the procedure. In addition, postoperative management is especially important for transapical patients (who comprised about 72% of the cohort), Dr. Rodés-Cabau noted, adding that a considerable proportion of the cases analyzed come from his institution’s early TAVR experience. The incidence of A-fib among more recent patients in this cohort, who benefitted from better postop management and pain control, would almost certainly be lower, he suggested.

Nonetheless, “it is interesting to see that even when you have short episodes of A-fib, the risk of cardioembolic events increases,” Dr. Rodés-Cabau said. In cardiac surgery, it is not clear whether or not this kind of postop A-fib requires anticoagulation, he noted, but the TAVR population tends to be much sicker.

Even Brief A-fib Often Warrants Anticoagulation

Dr. Rodés-Cabau described how his practice has changed as a result of the current findings.

“When a patient has an episode of A-fib, even it if is short, we look at him more carefully [in terms of cardioembolic risk],” he said. “In general, we start anticoagulation right away with heparin and warfarin [because] almost invariably these TAVR patients are very old and have a high CHADS [stroke risk] score. And it is not uncommon for them to have hypertension, diabetes, and a previous stroke.” If there are no recurrent episodes of A-fib, anticoagulation is stopped after several weeks, Dr. Rodés-Cabau reported.

A confounding factor is that the patients were already on clopidogrel and aspirin, and adding an anticoagulant increases bleeding risk, he noted. “So we stop the [clopidogrel] and aspirin if the patient has no coronary disease, or just the [clopidogrel] if they do have coronary disease,” he said. 

A-fib a Prime Suspect in Post-TAVR Stroke 

Perhaps the most provocative study finding, Dr. Rodés-Cabau said, is that A-fib is an important player in the risk of cardioembolic events after TAVR. Despite widespread concern over the incidence of stroke associated with the procedure, the reason for the risk is not well characterized, he noted. The current study sheds light on the issue by suggesting that strokes that occur during the subacute, or postprocedural, period have a distinct mechanism not directly related to the prosthetic valve or implantation but instead mediated through A-fib.

Dr. Rodés-Cabau stressed that these findings cannot be extrapolated to TAVR with CoreValve, and underlined that the study is hypothesis-generating and requires confirmation in larger series. But, he added, hopefully its insights can at least steer research in a direction that will help reduce the risk of cardioembolic events after TAVR. 

In an accompanying editorial, Bernard Iung, MD, and colleagues of Bichat Hospital (Paris, France) echo the same theme. “The potential contribution of postprocedural [A-fib] to the occurrence of stroke after [TAVR] highlights the need to improve the identification of patients at risk, to test pharmacological approaches to reduce the incidence of [A-fib], and to conduct randomized trials to refine antithrombotic therapy in this setting,” they write. 

2. Iung B, Himbert D, Vahanian A. Atrial fibrillation following transcatheter aortic valve implantation: Do we underestimate its frequency and impact? J Am Coll Cardiol. 2011;Epub ahead of print.

• Dr. Iung reports receiving consulting fees from Abbott, Bayer, Boehringer Ingelheim, Servier, and Valtech and speaker’s fees from Edwards Lifesciences, Sanofi-Aventis, and St. Jude Medical.

Related Stories:

• Smaller Delivery System Means Fewer Complications with Aortic Valve Implantation
• Myocardial Injury Common with TAVI, But Clinical Import Less Certain
• Consensus TAVI Endpoint Definitions Raise Standard of Research in New Field