Background. Atrial fibrillation (AF) and atrial flutter (AFL) are associated with increased risk of stroke and mortality after transcatheter aortic valve replacement (TAVR). Many episodes of new-onset AF/AFL (NOAF) occur after hospital discharge and may not be clinically apparent. Pacemakers can detect subclinical episodes of rapid atrial rate, which correlate with electrocardiographically documented AF. Methods. From 2012 to 2017, patients who underwent pacemaker implantation after TAVR were reviewed, and pacemaker data from device checks were analyzed for detection of NOAF. Patients with prior AF/AFL were excluded. Secondary outcomes were mortality and ischemic stroke. Results. A total of 172 patients underwent TAVR and pacemaker implantation, and 95 were without pre-existent AF/AFL. Over a median follow-up of 15 months, a total of 24 patients had NOAF (25%), of which 10 patients (10.5%) had manifest NOAF detected on electrocardiography, and 14 patients (14.7%) had subclinical NOAF first identified on device interrogation. The cumulative incidence of mortality was 16.7% for NOAF and 15.5% for normal sinus rhythm (P=.83). The cumulative incidence of stroke was 12.5% for NOAF and 1.4% for normal sinus rhythm (P=.04). Subclinical NOAF patients were less likely to be started on anticoagulation compared with manifest NOAF patients (70% vs 15.3%, respectively; P=.02). Conclusion. Subclinical NOAF is common after TAVR, usually occurs months after hospital discharge, and is associated with lack of anticoagulation therapy and increased risk of stroke. Prolonged surveillance of subclinical NOAF may be warranted after TAVR.
Key words: TAVR, TAVI, new-onset atrial fibrillation, new-onset atrial flutter, post-TAVR atrial fibrillation, post-TAVR atrial flutter
Transcatheter aortic valve replacement (TAVR) is an established treatment for severe aortic stenosis (AS) in both high- and moderate-risk patients.1,2 Although the incidence of new-onset atrial fibrillation/atrial flutter (NOAF) is less with TAVR relative to surgical AVR,3 it has been identified as an independent predictor of mortality and stroke after TAVR.
The incidence of NOAF after TAVR is dependent on the method of surveillance, length of observation, and access route.4,5 In the Society of Thoracic Surgeons/American College of Cardiology Transcatheter Valve Therapy (STS/ACC TVT) registry, the incidence of NOAF was 8%. However, studies that used continuous and prolonged electrocardiographic (ECG) monitoring reported a much higher incidence (20%-30%).6-10 Transapical approach is an independent predictor of NOAF. 10
Understanding the incidence and consequences of postprocedural NOAF is important to provide optimal clinical care for patients undergoing TAVR. Under-diagnosis (ie, subclinical NOAF) carries an increased risk of systemic embolism and stroke.4,5 Patients with post-TAVR pacemaker implantation present a unique opportunity to investigate otherwise undetected atrial arrhythmias. We aimed to identify the actual cumulative incidence of NOAF in patients who required pacemaker insertion after TAVR using pacemaker device interrogation.
Patient selection. Consecutive patients without prior pacemakers who underwent TAVR at Allina Health system in Minneapolis, Minnesota from January 2012 to May 2017 were retrospectively identified by querying the electronic medical record system. We identified patients who received a permanent pacemaker implant within 1 month of TAVR. Patients were included if they had pre- and post-TAVR ECGs, and had at least one pacemaker interrogation performed. Exclusion criteria included age <18 years, incarceration, pregnancy, or opting out of research as indicated by Allina Health’s release of information policy. Institutional review board approval was obtained for data collection, follow-up, and data analysis.
Data collection. The following variables were collected for all patients: baseline demographics; periprocedural TAVR data (during hospital admission); ECGs; indications for pacemaker implant and type of device; date of the implant and last follow-up; pacemaker device checks, computed tomography (CT) features; and incidence of stroke and mortality. Pre- and post-TAVR ECGs were reviewed to confirm indication for pacemaker implant and the presence or absence of atrial fibrillation (AF) or atrial flutter (AFL). All available pacemaker interrogations were examined to assess the detection of arrhythmias. All baseline and follow-up data were collected from the electronic medical records.
Pre-existing AF/AFL was defined as any single episode of AF or AFL detected on an ECG or rhythm strip that was recorded in the patient’s electronic medical record. All post-procedure ECGs and all device check information were interrogated for evidence of NOAF. All pacemaker checks were reviewed by an electrophysiologist, and the NOAF events were confirmed by the investigators. We classified NOAF into two groups: (1) manifest NOAF was identified first on ECG during routine clinical care; and (2) subclinical NOAF was first detected on pacemaker device checks (at least one episode of 30 seconds of AF or AFL confirmed by an electrophysiologist). All cerebrovascular events were evaluated by a neurologist and confirmed through neuroimaging techniques.
Statistical methods. Descriptive statistics are displayed as means ± standard deviations for continuous variables and numbers (percentages) for categorical variables. When continuous variables have skewed distributions (time to event data, accuracy and frequency of events), data are summarized as median (interquartile range [IQR], defined as 25th-75th percentiles). Categorical variables were analyzed using Pearson’s Chi-square or Fisher’s exact tests. Continuous variables were tested for normality with the Shapiro-Wilk test and were analyzed using the two-sample t-test for normally distributed variables, or Mann-Whitney test for non-normally distributed variables.
Time to the first occurrence of ischemic stroke was estimated and compared between patients with NOAF vs normal sinus rhythm (NSR) and patients with subclinical vs manifest NOAF using the cumulative incidence competing-risk method and the K-sample test developed by Gray.11 The Kaplan-Meier method was used to calculate estimates of survival and the difference between groups was compared with the log-rank test. A P-value <.05 was considered significant, and P-values were two-sided. All statistical calculations and plots were performed with EZR 1.37 (Saitama Medical Center, Jichi Medical University, 2018).
Patient characteristics. The study flow diagram is illustrated in Figure 1. A total of 720 patients without prior pacemaker underwent TAVR during the study period, of which 172 patients (23.9%) underwent pacemaker implantation within 1 month of the procedure. Indications for pacemaker implantation included new left bundle-branch block, new right bundle-branch block, complete heart block, sinus node dysfunction, and a group for all other less-frequent indications (such as AF with a slow ventricular response).
Of the 172 patients who had pacemakers after TAVR, a total of 77 had pre-existing AF/AFL and were excluded from this analysis. Of the remaining 95 patients, a total of 24 (25.2%) had NOAF over a median follow-up time of 443 days (IQR, 271-709 days) post TAVR. Mean age of the study cohort was 80 ± 8 years, median STS score was 4.6% (IQR, 3%-6%), and 55% were males. Baseline characteristics of NOAF patients vs patients who remained in NSR are illustrated in Table 1.
All patients had evidence of at least two episodes of NOAF separated by at least 4 weeks during their follow-up period, except 1 patient who had a single event of NOAF detected in the postprocedural ECG with no recurrent episodes. Manifest NOAF was detected in 10 patients (10.5%), while subclinical NOAF was detected in 14 patients (14.7%). The cumulative incidence of NOAF is illustrated in Figure 2.
Procedural characteristics are shown in Table 2. Patients with NOAF had significantly more valve over-sizing compared with patients who remained in NSR (15.4 ± 9% vs 11.6 ± 7.7%; P=.048). There was no difference in the indication for pacemaker implantation between the two groups. The study results are summarized in a visual abstract (Figure 3).
Subclinical AF and AFL. Fourteen patients had NOAF first detected on device checks (14.5%), and only two of these episodes were detected within the first 4 weeks of TAVR (2.1%). The median time to development of subclinical NOAF was 66 days. None of those patients had AF/AFL recognized clinically during follow-up, except 1 patient who had a single ECG with AF 16 months after detection of NOAF on the device check.
One patient was already on anticoagulation for a mechanical mitral valve, and 2 patients were started on anticoagulation (one for NOAF, and the second as routine post-TAVR anticoagulation). One patient was taken off warfarin 4 months after TAVR; he then suffered from an ischemic stroke 4 months later. One patient, who was not started on anticoagulation, suffered an ischemic stroke associated with subacute valve thrombosis 8 months after TAVR. He was started on anticoagulation after the stroke.
Manifest AF or AFL. Ten patients had NOAF detected first on ECG (10.5%). Six of these events were detected within the first 4 days and 9 events were detected within the first 4 weeks of TAVR. The median time to development of manifest NOAF was 3 days. None of these patients were on anticoagulation before TAVR. Seven patients (70%) were started on anticoagulation after TAVR (1 patient for deep vein thrombosis and 6 patients for AF/AFL); of these patients, 1 was taken off warfarin 2 months after TAVR and had an ischemic stroke 10 months later.
Prior AF or AFL. Seventy-seven patients had prior AF/AFL. Forty-eight patients were already on anticoagulation (62%) and 6 patients (7.8%) were started on anticoagulation after TAVR. Of the patients who were on anticoagulation, 1 patient had an ischemic stroke immediately after TAVR, and another patient had a hemorrhagic stroke. Of the patients who were not on anticoagulation, 1 patient had an ischemic stroke, and another had a hemorrhagic stroke. The cumulative incidence of ischemic strokes in patients with prior AF/AFL was 2.6% over a median follow-up period of 14 months.
Patients who remained in NSR. A total of 71 patients remained in NSR at follow-up. Of those, only 1 patient – who was not on anticoagulation – had an ischemic stroke 3 months after TAVR.
Initiating anticoagulation in NOAF. After the exclusion of patients who were already on anticoagulation, only 10 patients out of 23 (43.5%) were started on anticoagulation. The percentage of patients who were started on anticoagulation was significantly lower in the subclinical NOAF group (15.3% vs 70.0%; P=.02).
Mortality. Over a median follow-up of 443 days, the cumulative incidence of mortality was 16.7% in the NOAF group and 15.5% in the NSR group. There was no difference in survival rate between patients with NSR vs those with NOAF (log rank P=.83), or patients with manifest NOAF vs those with subclinical NOAF (log rank P=.80) (Supplemental Figures S1 and S2).
Stroke. The cumulative incidences of ischemic stroke in the NOAF and NSR groups were 12.5% and 1.4% over a median follow-up period of 443 days, respectively. The cumulative incidence of stroke was significantly higher in patients with NOAF (P=.04) (Supplemental Figure S3). The cumulative incidence of stroke in patients with subclinical AF/AFL was higher compared with patients who remained in NSR (P=.04) (Supplemental Figure S4). There was no difference in the cumulative incidence of stroke between subclinical and manifest NOAF patients (P=.73) (Supplemental Figure S5).
The findings of our study using pacemaker data to estimate the incidence of NOAF after TAVR can be summarized as follows: (1) subclinical NOAF is more common than clinical NOAF (it represents more than 50% of NOAF episodes after TAVR); (2) subclinical NOAF presents months after hospital discharge (median time to presentation, 66 days), whereas clinical NOAF presents before hospital discharge (median time to presentation, 3 days); (3) both subclinical and clinical NOAF are strongly associated with increased incidence of stroke at follow-up; and (4) patients with subclinical NOAF are less likely to be started on anticoagulation compared with manifest NOAF patients (15.3% vs 70%, respectively; P=.02). Taken together, our study findings may provide an opportunity to improve outcomes of TAVR patients through adequate risk stratification and optimization of medical therapy.
The incidence of AF/AFL after TAVR has been inconsistent in different reports, with a range from 7% to 15% at 30 days using clinical follow-up information alone.7,9,10 These discrepancies are due to the inconsistencies in the definitions, study methods, and duration of AF surveillance. The actual incidence of subclinical NOAF after TAVR is likely under-reported in clinical studies without prolonged electrocardiographic surveillance. Vora et al reported an 8.4% incidence of NOAF after TAVR using the STS/ACC TVT registry, which consists primarily of inpatient data linked to Medicare claims, without prolonged ECG monitoring.10 In the MARE (Ambulatory Electrocardiographic Monitoring for the Detection of High-Degree Atrioventricular Block in Patients With New-Onset Persistent Left Bundle-Branch Block After Transcatheter Aortic Valve Implantation) study, 17% of patients with new-onset left bundle-branch block who were monitored for a year developed NOAF.12 In our study, we report a 25% cumulative incidence of NOAF over a median follow-up of 15 months. Our study has higher power and longer follow-up duration. Other factors that make our study different are the presence of ventricular pacing in a subset of patients implanted with pacemakers, and different valve systems used in our cohort; thus, our results are more reflective of a real-world experience.
In our cohort, less than one-half of the NOAF events were discovered during routine clinical care, and most of these events were detected early after the procedure (60% within 4 days and 90% within 4 weeks of TAVR). On the other hand, subclinical NOAF events were common (14% of our cohort) and presented after hospital discharge (85% were detected on device checks beyond 4 weeks of the procedure). The immediate postprocedural period carries a high risk of developing rhythm disturbances (45% of events were detected in the first 4 weeks), and most of these events (82%) are recognized during routine inpatient clinical follow-up. Less evident is the high risk of subclinical NOAF that persists beyond 4 weeks of the procedure, which could justify longer-term surveillance.
Recent publications have shown that stroke rates in TAVR patients range from 2.0%-2.8%.13-15 Approximately one-half of these strokes occur several days after TAVR.1 Pre-existing and new-onset AF has been identified as a predictor of cerebrovascular events and mortality after TAVR.6,7,10,16-22 Subclinical NOAF has been previously studied in patients who did not undergo TAVR and received pacemakers. Healy et al reported an incidence of 10% subclinical AF/AFL events over 3 months of follow-up, and these events were associated with a higher risk of ischemic strokes over 2.5 years of follow-up.23 In a recent meta-analysis by Mahajan et al,24 subclinical AF/AFL was associated with increased risk of stroke compared with patients who remained in NSR, but lower risk than those with clinical AF/AFL. However, the association between subclinical AF/AFL and the risk of stroke in post-TAVR patients is not clear due to lack of data. In our study, subclinical AF/AFL patients has higher cumulative incidence of stroke compared with patients who remained in NSR.
In our cohort, only 43% of patients who were diagnosed with NOAF were started on anticoagulation, which is slightly lower than the rate reported in the BRAVO-3 (Effect of Bivalirudin on Aortic Valve Intervention Outcomes-3) trial, in which 58.7% of prior and new-onset AF/AFL patients were discharged on oral anticoagulants,21 but slightly higher than reported by Vora et al,10 as only 28.9% of their patients were discharged on anticoagulation. It has been reported that mortality rates were highest among patients who were not discharged on anticoagulation.10 In our cohort, subclinical NOAF patients were less likely to be started on anticoagulation compared with manifest NOAF patients. All patients who had ischemic strokes were not on anticoagulation at the time of the stroke. The potential benefits of anticoagulation need to be balanced against the risks of bleeding in a cohort of elderly patients with multiple comorbidities and frailty. A preliminary single-center analysis suggests this approach is as safe as dual-antiplatelet therapy.25
To our knowledge, this study is the first to use pacemaker interrogation data to detect the actual incidence of new-onset subclinical AF in post-TAVR patients. Since NOAF is associated with worse outcomes, prolonged surveillance for subclinical AF/AFL may be reasonable in patients after TAVR. There is a need to emphasize the importance of adherence to anticoagulation guidelines in TAVR patients with NOAF. The need for long-term monitoring with implantable-loop monitors to obtain the actual incidence of subclinical NOAF and understand the potential cerebrovascular events risk would help achieve optimal medical care and improve outcomes for those patients.
Study limitations. The present study has several limitations. First, our analysis is limited to a selected cohort of patients (23%) who were implanted with a pacemaker after TAVR. However, the overall incidence of NOAF (25%) is in line with the MARE study (17%), which had broader inclusion criteria. Second, the sample size is small in statistical terms, which precludes more robust analysis such as regression models to identify predictors of NOAF. Third, due to the small number of events, quantitative analysis of AF burden and its relationship to stroke risk was not performed. Finally, the risk of NOAF increases with age and might not be related to TAVR. However, the clinical impact of our findings remains relevant to the care of these patients.
Subclinical NOAF detected via pacemaker interrogations is common after TAVR, usually occurs months after hospital discharge, and is associated with lower use of anticoagulation therapy and increased risk of stroke. Prolonged surveillance of subclinical NOAF may be warranted after TAVR.
This article was published with permission from J Invasive Cardiol. 2019;31(7):E177-E183.
Acknowledgments. The authors would also like to thank and acknowledge the Abbott Northwestern Hospital Foundation for their ongoing support of research and the MHIF internship program.
Disclosure: Dr. Garcia reports consultant income from Surmodics, Osprey Medical, Medtronic, Edwards Lifesciences, and Abbott; grant support from Edwards Lifesciences and the VA Office of Research and Development. Dr. Sorajja reports consulting and speaking income from Abbott Vascular, Edwards Lifesciences, Medtronic, and Boston Scientific; equity and consulting for Pipeline Technologies and Admedus. Dr. Sengupta reports grant support from Medtronic.
- Leon MB, Smith CR, Mack M, et al. Transcatheter aortic-valve implantation for aortic stenosis in patients who cannot undergo surgery. N Engl J Med. 2010;363:1597-1607.
- Reardon MJ, Van Mieghem NM, Popma JJ, et al. Surgical or transcatheter aortic-valve replacement in intermediate-risk patients. N Engl J Med. 2017;376:1321-1331.
- Thourani VH, Kodali S, Makkar RR, et al. Transcatheter aortic valve replacement versus surgical valve replacement in intermediate-risk patients: a propensity score analysis. Lancet. 2016;387:2218-2225.
- Benjamin EJ, Wolf PA, D’Agostino RB, Silbershatz H, Kannel WB, Levy D. Impact of atrial fibrillation on the risk of death: the Framingham Heart Study. Circulation. 1998;98:946-952.
- Miyasaka Y, Barnes ME, Bailey KR, et al. Mortality trends in patients diagnosed with first atrial fibrillation: a 21-year community-based study. J Am Coll Cardiol. 2007;49:986-992.
- Amat-Santos IJ, Rodés-Cabau J, Urena M, et al. Incidence, predictive factors, and prognostic value of new-onset atrial fibrillation following transcatheter aortic valve implantation. J Am Coll Cardiol. 2012;59:178-188.
- Barbash IM, Minha S, Ben-Dor I, et al. Predictors and clinical implications of atrial fibrillation in patients with severe aortic stenosis undergoing transcatheter aortic valve implantation. Catheter Cardiovasc Interv. 2015;85:468-477.
- Smith CR, Leon MB, Mack MJ, et al. Transcatheter versus surgical aortic-valve replacement in high-risk patients. N Engl J Med. 2011;364:2187-2198.
- Tarantini G, Mojoli M, Windecker S, et al. Prevalence and impact of atrial fibrillation in patients with severe aortic stenosis undergoing transcatheter aortic valve replacement: an analysis from the SOURCE XT prospective multicenter registry. JACC Cardiovasc Interv. 2016;9:937-946.
- Vora AN, Dai D, Matsuoka R, et al. Incidence, management, and associated clinical outcomes of new-onset atrial fibrillation following transcatheter aortic valve replacement: an analysis from the STS/ACC TVT registry. JACC Cardiovasc Interv. 2018;11:1746-1756.
- Gray RJ. A class of K-sample tests for comparing the cumulative incidence of a competing risk. Ann Stat. 1988;16:1141-1154.
- Rodés-Cabau J, Urena M, Nombela-Franco L, et al. Arrhythmic burden as determined by ambulatory continuous cardiac monitoring in patients with new-onset persistent left bundle branch block following transcatheter aortic valve replacement: the MARE study. JACC Cardiovasc Interv. 2018;11:1495-1505.
- Mack MJ, Brennan JM, Brindis R, et al. Outcomes following transcatheter aortic valve replacement in the United States. JAMA. 2013;310:2069-2077.
- Thomas M, Schymik G, Walther T, et al. One-year outcomes of cohort 1 in the Edwards SAPIEN aortic bioprosthesis European outcome (SOURCE) registry. Circulation. 2011;124:425-433.
- Möllmann H, Bestehorn K, Bestehorn M, et al. In-hospital outcome of transcatheter vs. surgical aortic valve replacement in patients with aortic valve stenosis: complete dataset of patients treated in 2013 in Germany. Clin Res Cardiol. 2016;105:553-559.
- Yankelson L, Steinvil A, Gershovitz L, et al. Atrial fibrillation, stroke, and mortality rates after transcatheter aortic valve implantation. Am J Cardiol. 2014;114:1861-1866.
- Stortecky S, Buellesfeld L, Wenaweser P, et al. Atrial fibrillation and aortic stenosis: impact on clinical outcomes among patients undergoing transcatheter aortic valve implantation. Circ Cardiovasc Interv. 2013;6:77-84.
- Nombela-Franco L, Webb JG, de Jaegere P, et al. Timing, predictive factors and prognostic value of cerebrovascular events in a large cohort of patients undergoing transcatheter aortic valve implantation. Circulation. 2012;126:3041-3053. Epub 2012 Nov 13.
- Nuis R-J, Van Mieghem NM, Schultz CJ, et al. Frequency and causes of stroke during or after transcatheter aortic valve implantation. Am J Cardiol. 2012;109:1637-1643.
- Chopard R, Teiger E, Meneveau N, et al. Baseline characteristics and prognostic implications of pre-existing and new-onset atrial fibrillation after transcatheter aortic valve implantation: results from the FRANCE-2 registry. JACC Cardiovasc Interv. 2015;8:1346-1355.
- Hengstenberg C, Chandrasekhar J, Sartori S, et al. Impact of pre-existing or new-onset atrial fibrillation on 30-day clinical outcomes following transcatheter aortic valve replacement: results from the BRAVO 3 randomized trial. Catheter Cardiovasc Interv. 2017;90:1027-1037.
- Furuta A, Lellouche N, Mouillet G, et al. Prognostic value of new onset atrial fibrillation after transcatheter aortic valve implantation: a FRANCE 2 registry substudy. Int J Cardiol. 2016;210:72-79.
- Healey JS, Connolly SJ, Gold MR, et al. Subclinical atrial fibrillation and the risk of stroke. N Engl J Med. 2012;366:120-129.
- Mahajan R, Perera T, Elliott AD, et al. Subclinical device-detected atrial fibrillation and stroke risk: a systematic review and meta-analysis. Eur Heart J. 2018;39:1407-1415.
- Gurevich S, Oestreich B, Kelly RF, et al. Routine use of anticoagulation after transcatheter aortic valve replacement: initial safety outcomes from a single-center experience. Cardiovasc Revasc Med. 2018;19:621-625.