Heart CManuscripaccepted O

See paThe wo

Research ipital.

*CorreE-mail

0002-9149http://dx.do

Impact of Atrial Fibrillation on Outcomes in Patients WhoUnderwent Transcatheter Aortic Valve Replacement

Abhishek Maan, MD, E. Kevin Heist, MD, PhD, Jonathan Passeri, MD, Ignacio Inglessis, MD,Joshua Baker, MD, Leon Ptaszek, MD, PhD, Gus Vlahakes, MD, Jeremy N. Ruskin, MD,

Igor Palacios, MD, Thoralf Sundt, MD, and Moussa Mansour, MD*

Transcatheter aortic valve replacement (TAVR) has emerged as an alternative treatment

enter,t receivctoberge 225rk wasn Atria

spondinaddres

/14/$ -i.org/1

for surgical high-risk patients with severe aortic stenosis. The aim of this study was todetermine the impact of atrial fibrillation (AF) on procedural outcomes. Data from 137patients who underwent TAVR using Edwards SAPIEN valve were reviewed. Thepredictors of new-onset atrial fibrillation (NOAF) after the procedure were analyzed. Inaddition, the post-TAVR clinical outcomes and adverse events were compared according tothe presence and absence of preprocedural and postprocedural AF. Previous AF waspresent in 49% of the patients who underwent TAVR. After the procedure, NOAF wasdetected in 21% of patients, and the cumulative incidence of post-TAVR AF was 60%. AfterTAVR, 50% of all the episodes of NOAF occurred in the initial 24 hours after the proce-dure. Transapical approach was observed to an important predictor of NOAF (adjustedodds ratio [OR] 5.05, 95% confidence interval [CI] 1.40 to 18.20, p[ 0.013). The compositeoutcome of all-cause mortality, stroke, vascular complications, and repeat hospitalization in1 month after TAVR was significantly higher in patients with previous AF (33 of 67 vs 19 of70, adjusted OR 2.60, 95% CI 1.22 to 5.54, p [ 0.013) compared with patients who did nothave previous AF. The presence of post-TAVR AF led to a prolongation in the duration ofintensive care unit stay by an average of 70 hours (95% CI 25 to 114.7 hours, p [ 0.002).Similarly, post-TAVR AF also led to the prolongation in the hospital stay by an average of6.7 days (95% CI 4.69 to 8.73 days, p <0.0005). In conclusion, our study demonstrates thatthe presence of AF before TAVR is an important predictor of the composite end point ofall-cause mortality, stroke, vascular complications, and repeat hospitalization in 1 monthafter the procedure. AF after TAVR is more likely to be encountered with the transapicalapproach and is associated with a prolongation of intensive care unit and hospitalstay. � 2015 Elsevier Inc. All rights reserved. (Am J Cardiol 2015;115:220e226)

Transcatheter aortic valve replacement (TAVR) is a rela-tively new procedure that has evolved as an alternativetreatment option for patients with severe aortic stenosis (AS)who are at a high surgical risk.1e3 Atrial fibrillation (AF) is themost common cardiac arrhythmia with an increased preva-lence especially in the elderly population.4,5 Because of therelatively older age of patients with severe AS, the prevalenceof AF in this patient population tends to be higher. Thepresence of severe AS also leads to left ventricular hyper-trophy and a state of an increased afterload; both these path-ophysiological processes also mediate the development ofAF.6e8 Both the presence and development of AF have beenobserved to be associated with a higher incidence of adverseclinical outcomes and increased mortality in the setting ofsurgical aortic valve replacement and cardiothoracic

Massachusetts General Hospital, Boston, Massachusetts.ed August 14, 2014; revised manuscript received and14, 2014.for disclosure information.supported in part by the Deane Institute for Integrativel Fibrillation and Stroke at Massachusetts General Hos-

g author: Tel: (617) 726-5557; fax: (617) 724-1241.s: mmansour@mgh.harvard.edu (M. Mansour).

see front matter � 2015 Elsevier Inc. All rights reserved.0.1016/j.amjcard.2014.10.027

surgeries.9e12 AF carries an increased risk of thromboembolicand vascular complications compared with patients in sinusrhythm (SR).13 AF also leads to an atrioventricular dyssyn-chrony, which further adversely affects cardiac function inAS.14 The aim of this study was to determine the impact of AFon procedural outcomes.

Methods

The study population consisted of 137 consecutive pa-tients with severe AS who under underwent TAVR atMassachusetts General Hospital between June 2008 andOctober 2012. The procedure was performed using theballoon-expandable valve (Edwards SAPIEN, SAPIEN XT;Edward Lifesciences, Irvine, California). The preproceduralrisk was calculated using the Society of Thoracic Surgeonsrisk score for Prediction of Mortality15 and the logisticEuroSCORE.16,17 The study design is briefly explained inFigure 1. Data on baseline characteristics on all patientswere collected retrospectively. Preprocedural and post-procedural electrocardiograms (ECGs) were analyzed forrhythm, various intervals, and conduction abnormalities.Data on various echocardiographic parameters were ob-tained from the transthoracic echocardiograms that wereobtained before and after TAVR.

www.ajconline.org

Figure 1. Study design and various study subsets.

Valvular Heart Disease/Impact of AF in Patients Who Underwent TAVR 221

Patients were evaluated for the presence of AF before theprocedure. The presence of AF was confirmed on ECGs andthe documentation from inpatient and outpatient medicalrecords. Pre-TAVR rhythm was documented on all patientswho underwent the procedure. All the study patients weremonitored by continuous telemetry until the day of discharge.New-onset atrial fibrillation (NOAF) was defined as anyepisode of AF lasting >30 seconds in the patients withoutany history of this arrhythmia.18,19 In the patients whodeveloped NOAF, collected data included the timing ofarrhythmia occurrence, requirement of cardioversion, and theuse antithrombotic therapy. The patients were divided into 2groups: post-TAVR AF (all patients with AF �30 secondspost-TAVR) and post-TAVR SR.

Univariate analyses were performed for comparison be-tween various subgroups (previous AF vs no previous AF,post-TAVR AF vs post-TAVR SR). Categorical variablesbetween the 2 subgroups were compared using the Fisher’sexact test, and continuous variables were compared using the2-tailed unpaired Student’s t test. After the univariate ana-lyses, 4 logistic regression models were prepared for each ofthe adverse outcomes (NOAF, major vascular complication,composite outcome of all-cause mortality, stroke, vascularcomplications, and repeat hospitalization in 1 month afterTAVR, and all-cause 1-year mortality). We also prepared 2linear regression models (each for duration of intensive careunit [ICU] stay and duration of hospital stay, respectively) toinvestigate the impact of post-TAVR AF on these outcomes.The logistic regression models were all tested for the “good-ness of fit” of the logistic model with the HosmereLemeshowtest. All statistical tests were performed with STATA 9.2software (Stata Corp LP, College Station, TX).

Results

The median age was 85 years (range 55 to 96). PreviousAF was present in 67 patients (49%). There was no difference

in the baseline clinical variables in the patients with or withoutprevious AF, except that the use of warfarin was significantlymore common in the patients with previous AF (59.7% vs1.4%; p <0.0001). The baseline characteristics of studypatients and the comparison between the patients according tothe presence of previous AF are listed in Table 1. ECGperformed at the time of TAVR revealed the presence of AFin 60% (40 of 67) of the patients with previous AF. Thecomparison of various echocardiographic parameters be-tween the 2 groups of patients (previous AF vs withoutprevious AF) is listed in Table 1. TAVRwas performed usingthe Edwards SAPIEN (Edwards Lifesciences) balloon-expandable prosthesis with the 2 available sizes (23 and26 mm). Transapical approach was used in most patients(81 of 137; 59%).

To determine the incidence of NOAF, the 67 patients witha history of AF were excluded. In the remaining 70 patients,NOAF occurred in 21 patients (21 of 70; 30%). A total of50% of all the episodes of NOAF occurred in the initial24 hours after TAVR (Figure 2). Electrical cardioversionwas performed in 3 patients who developed NOAF, and 8patients converted from AF back to sinus rhythm afterreceiving amiodarone. The comparison of baseline clinical,echocardiographic, and procedural characteristics of thepatients categorized according to the occurrence of NOAF islisted in Table 2. In the multivariate analysis, transapicalapproach of TAVR was observed to be an important pre-dictor of NOAF after the procedure (adjusted odds ratio [OR]5.05, 95% confidence interval [CI] 1.40 to 18.20; p¼ 0.013).

The overall 1-year mortality in this study was 12%(16 patients). The standard definition of “improvement infunctional status and health-related quality of life after theprocedure” was used for determining the success ofTAVR.2,20,21 According to this definition, the proceduralsuccess of TAVR was 95%. Based on the predischargeechocardiogram, there was a significant decrease in the mean(50.73� 16.29 to 11.67� 4.52 mm Hg; p <0.001) and peak

Table 1Baseline characteristics of the study population and comparison according to the presence of prior AF

Baseline variables Study population(N ¼ 137)

No Prior AF(N ¼ 70)

Prior AF(N ¼ 67)

P value

Age (years) 84.18 � 6.83 83.92 � 6.96 84.46 � 6.72 0.65Male 65 (47%) 30 (43%) 35 (52%) 0.31Body mass index (kg/m2) 26.70 � 5.90 26.40 � 5.87 27 � 5.94 0.55Active smokers 7 (5%) 5 (7%) 2 (3%) 0.44STS-PROM score (%) 6.88 � 3.82 6.49 � 3.18 7.29 � 4.38 0.23Logistic EuroSCORE (%) 14.33 � 12.24 13.67 � 12.01 15.02 � 12.52 0.52Hypertension 109 (80%) 54 (77%) 55 (82%) 0.53Hyperlipidemia 95 (69%) 49 (70%) 46 (69%) 1.00Diabetes Mellitus 47 (34%) 27 (39%) 20 (40%) 0.37Coronary artery disease 99 (72%) 48 (69%) 51 (76%) 0.35Congestive heart failure 74 (54%) 34 (39%) 40 (60%) 0.23Chronic obstructive pulmonary disease 39 (28%) 18 (26%) 21 (31%) 0.57Peripheral vascular disease 42 (31%) 24 (34%) 18 (27%) 0.36Pulmonary hypertension 38 (28%) 14 (20%) 24 (36%) 0.06Serum creatinine (mg/dl) 1.33 � 0.47 1.32 � 0.49 1.34 � 0.45 0.80Cerebrovascular disease 25 (18%) 15 (21%) 10 (15%) 0.38Carotid artery disease 31 (23%) 20 (29%) 11 (16%) 0.10Previous ProceduralPercutaneous coronary intervention 51 (37%) 23 (33%) 28 (42%) 0.29Prior pacemaker implantation 27 (20%) 12 (17%) 15 (22%) 0.52Balloon valvuloplasty 24 (18%) 13 (19%) 11 (16%) 0.83Coronary artery bypass grafting 55 (40%) 24 (34%) 31 (46%) 0.17Prior Medication h/oAspirin 113 (82%) 64 (91%) 49 (73%) 0.007Clopidogrel 28 (20%) 16 (23%) 12 (18%) 0.53Warfarin 41 (30%) 1 (1%) 40 (60%) <0.0001Beta-blockers 90 (66%) 44 (63%) 46 (69%) 0.74Diuretics 108 (79%) 56 (80%) 52 (78%) 0.84ACEi/ARBs 55 (40%) 27 (39%) 28 (42%) 0.73Statins 106 (79%) 55 (79%) 51 (76%) 0.84Echocardiographic parametersEjection fraction (%) 55.29 � 17.10 55.10 � 17.95 56.25 � 13.87 0.89LVEF < 40% 32 (23%) 19 (27%) 13 (19%) 0.32LV EDD (mm) 44.58 � 7.14 43.81 � 7.16 45.54 � 7.11 0.16Severe mitral regurgitation 8 (6%) 4/70 (6%) 4/67 (6%) 1.00Moderate mitral regurgitation 60 (44%) 24/66 (36%) 36/63 (57%) 0.022Mean aortic gradient (mm Hg) 50.73 � 16.29 54.36 � 16.56 46.94 � 15.22 0.0072Peak aortic gradient (mm Hg) 89.60 � 26.53 92.20 � 25.51 80.75 � 26.49 0.011Left atrial size (mm) 43.76 � 6.71 41.40 � 6.37 46.24 � 6.17 <0.0001Interventricular septum (mm) 13.38 � 2.09 13.40 � 2.14 13.36 � 2.05 0.91Aortic valve area (cm2) 0.60 � 0.13 0.58 � 0.13 0.60 � 0.13 0.27

ACEi ¼ Angiotensin converting enzyme; ARBs ¼ Angiotensin receptor blockers; LVEDD ¼ Left ventricular end diastolic diameter; PPM ¼ Permanentpacemaker; STS-PROM ¼ Society of Thoracic Surgery-Predictor of Mortality score.

222 The American Journal of Cardiology (www.ajconline.org)

aortic gradients (89.60 � 26.53 to 23.18 � 8.57 mm Hg;p <0.001) after the procedure. There was no significantdifference in 1-year post-TAVR mortality in the patients ac-cording to the presence of previous AF (15% in patients withprevious AF vs 9% in patients without previous AF; OR 1.87,95% CI 0.64 to 5.48, p ¼ 0.29) on univariate analysis. Theunivariate comparison of clinical outcomes, adverse events,and echocardiographic outcomes in patients grouped ac-cording to the presence of previous AF is described in Table 3.The incidence of the combined end point of all-cause 1-yearmortality, stroke, vascular complications, and repeat hospi-talization in 1-month post-TAVR was also analyzed. Onmultivariate analysis using logistic regression, previous AFwas detected to be an important predictor of this combinedend point (adjusted OR 2.60, 95%CI 1.22 to 5.54, p¼ 0.013).

The cumulative incidence of overall AF after TAVR wasalso investigated. For this purpose, the patients with a historyof AF were also included. The cumulative incidence ofoverall post-TAVR AF was defined as the combination ofpatients who had previous AF and persisted in AF throughouttheir pre- and post-TAVR hospital course and patients whodid not have previous AF but developed NOAF after TAVR.According to this definition, the cumulative incidence ofpost-TAVRAFwas found to be 60% (82 of 137 patients). Oncomparison of baseline clinical and echocardiographicparameters, there were no significant differences in thepost-TAVR AF versus post-TAVR SR group. The presenceof post-TAVR AF was associated with a significantly longerstay in the hospital compared with the patients in SR afterTAVR (16.21 � 8.06 days vs 9.17 � 4.29 days; p <0.001).

Figure 2. Timing of occurrence of new-onset atrial fibrillation after trans-catheter aortic valve replacement.

Table 2Baseline clinical, echocardiographic and procedural variables in patients according to the development of new-onset AF

Variables New onset AF (n ¼ 21) SR (n ¼ 49) Odds ratio and 95% CI (Univariate analysis) P value

Age (years) 84.48 � 6.01 83.69 � 7.38 0.67Female 15 (71%) 27 (55%) 2.04; 0.68-6.13 0.29Body mass index (Kg/m2) 25.62 � 5.8 26.73 � 5.9 0.47Active smokers 1 (5%) 4 (8%) 0.56; 0.06-5.36 1.00Hypertension 16 (76%) 38 (78%) 0.93; 0.28-3.10 1.00Hyperlipidemia 13 (62%) 36 (73%) 0.59; 0.20-1.74 0.35Diabetes mellitus 6 (29%) 21 (43%) 0.53; 0.18-1.61 0.29Coronary artery disease 13 (62%) 35 (71%) 0.65; 0.22-1.91 0.57H/o Percutaneous coronary intervention 6 (29%) 17 (35%) 0.75; 0.25-2.30 0.78H/o Coronary artery bypass grafting 7 (33%) 17 (35%) 0.94; 0.32-2.78 1.00Cerebrovascular disease 6 (29%) 9 (18%) 1.78; 0.54-5.85 0.35Congestive heart failure 10 (48%) 24 (49%) 0.95; 0.34-2.64 1.00Peripheral Vascular disease 6 (29%) 18 (37%) 0.69; 0.22-2.10 0.59H/o Balloon aortic valvuloplasty 4 (19%) 9 (18%) 1.05; 0.28-3.87 1.00Chronic obstructive Pulmonary disease 5 (24%) 13 (27%) 0.87; 0.26 -2.84 1.00Serum creatinine (mg/dl) 1.34 � 0.48 1.32 � 0.49 0.86EGFR (<60 ml/min) 15 (71%) 27 (55%) 2.04; 0.68-6.13 0.29Prior Medication h/oAspirin 19 (90%) 45 (92%) 0.84; 0.14-5.01 1.00Clopidogrel 6 (29%) 10 (20%) 1.56; 0.48-5.05 0.54Beta-Blockers 12 (57%) 32 (65%) 0.71; 0.25-2.02 0.59ACEi/ARBs 7 (33%) 20 (41%) 0.73; 0.25-2.12 0.60Statins 17 (81%) 38 (78%) 1.23; 0.34-4.42 1.00Echocardiographic parametersEjection fraction (%) 55.95 � 18.71 54.71 � 17.80 0.79Ejection fraction (<40%) 6 (29%) 13 (26%) 1.11; 0.35-3.46 1.00Left ventricular end diastolic diameter (mm) 43.43 � 8.27 43.98 � 6.72 0.77Severe Mitral regurgitation 3 (14%) 1 (2%) 8.0; 0.78-82.03 0.077Mean Aortic Gradient (mm Hg) 52.28 � 13.51 55.24 � 17.76 0.49Aortic valve area (cm2) 0.58 � 0.14 0.57 � 0.12 0.87Left atrial size (mm) 40.47 � 7.39 41.77 � 5.94 0.43Right ventricular systolic Pressure (mm Hg) 50.57 � 15.16 47.35 � 11.27 0.33Transapical approach 18 (86%) 21 (43%) 8.0; 2.1-30.77 0.0013

h/o ¼ history of.

Valvular Heart Disease/Impact of AF in Patients Who Underwent TAVR 223

The univariate comparison of clinical outcomes and adverseevents after TAVR in the patients grouped according to theoccurrence of post-TAVR AF is described in Table 4.

Using amultivariate analysis and a linear regressionmodel,the presence of post-TAVR AF was observed to be a strongpredictor of the prolongation in the hospital stay. The presenceof AF after TAVR prolonged the hospital stay by an average

of 6.7 days (95% CI 4.69 to 8.73, p<0.0005). Similarly, post-TAVR AF was also an important determinant of the durationof post-TAVR ICU stay. The presence of AF after TAVR ledto prolongation of ICU course by an average of 70 hours (95%CI 25.1 to 114.7, p ¼ 0.002). In a logistic regression model,post-TAVR AF was also an important predictor of vascularcomplications after TAVR (adjusted OR 10.5, 95% CI 1.21 to91.75; p¼ 0.033). Patients with post-TAVRAF also showed astatistical trend toward increased all-cause mortality at 1 year(adjusted OR 2.90, 95% CI 0.86 to 9.85; p ¼ 0.087).

In addition to preprocedural and postprocedural AF,several other baseline and procedural predictors of adverseoutcome after TAVR were also indentified. Patients who hada history of percutaneous coronary intervention were also atincreased risk for the composite of clinical outcomes(described previously; adjusted OR 2.10, 95% CI 0.98 to

4.50; p ¼ 0.055). Patients who underwent the procedurethrough a transapical approach were less likely to developpost-TAVR vascular complications (adjusted OR 0.055, 95%CI 0.00064 to 0.47; p ¼ 0.008). In the linear regressionmodel, an elevation in serum creatinine at baseline was alsoobserved to prolong the hospital stay by an average of2.8 days (95% CI 0.005 to 5.64; p¼ 0.05). A greater baseline

Table 3Clinical outcomes in patients undergoing TAVR according to the presence of prior h/o AF

Clinical Outcomes No h/o AF (n ¼ 70) Pre-existing AF (n ¼ 67) OR with 95 % CI(Univariate Analysis)

p value

AF at the time of procedure 0/70 40 (60%) 207.65; 12.33-3498.3 <0.0001Procedural success 67 (96%) 63 (94%) 0.71; 0.15-3.28 0.71Valve related complication (embolization, damage to

the valve and requirement of 2nd valve)4 (6%) 2 (3%) 0.51; 0.089-2.869 0.68

Post procedural AF 21 (30%) 61 (91%) 23.72; 8.88-63.36 <0.001Myocardial Infarction 1 (1%) 1 (1%) 1.05; 0.06-17.07 1.00Development of new conduction block 22 (31%) 22 (33%) 1.07; 0.52-2.19 1.00Implantation of new pacemaker 20 (27%) 11 (16%) 0.49; 0.21-1.13 0.10Major vascular complication 2 (3%) 6 (9%) 3.34; 0.65-17.20 0.16Cerebrovascular events (all) 4 (6%) 3 (5%) 0.77; 0.17- 3.60 1.00Stroke 2 (3%) 3 (5%) 1.60; 0.26- 9.90 0.68Repeat hospitalization in 1 month 9/67 (13%) 14/61 (23%) 0.52; 0.21-1.31 0.18Death within 1 month 3 (4%) 6 (9%) 2.20; 0.53-9.17 0.32Death after 1 month 3/67 (5%) 4/61 (7%) 1.50; 0.32-6.98 0.71Overall Mortality 6/70 (9%) 10/67 (15%) 1.87; 0.64-5.48 0.29Overall Mortality secondary to cardiac causes 1/70 (1%) 7/67 (10%) 8.05; 0.96- 67.35 0.031Length of hospitalization (days) 11.12 � 6.36 15.67 � 8.12 0.0005Composite of clinical outcomes (All-cause mortality,

stroke, vascular complications and repeathospitalization in 1 month)

19(27%) 33 (49%) 2.61; 1.28-5.31 0.009

(All-cause mortality, stroke and vascular complications) 11 (16%) 20 (30%) 2.28; 0.995-5.233 0.066LVEF prior to discharge from the hospital (%) 57.51 � 15.07 56.25 � 13.87 0.62Mean Ao gradient (mm Hg) 12.45 � 4.56 (n ¼ 67) 10.86 � 4.41 (n ¼ 63) 0.0452Peak Ao gradient (mm Hg) 24.99 � 8.83 (n ¼ 67) 21.27 � 7.90 (n ¼ 63) 0.0126D Mean Aortic gradient (mm Hg) 41.91 � 16.32 (n ¼ 67) 36.27 � 14.19 (n ¼ 63) 0.0372D Peak Aortic gradient (mm Hg) 66.93 � 23.89 (n ¼ 67) 59.97 � 24.92 (n ¼ 63) 0.1071

Table 4Comparison of clinical and adverse outcomes in patients with post-procedural AF and SR

Clinical Outcome/adverse Events Post-procedural AF (n¼82) Post-procedural SR (n¼55) OR with 95 % CI P value

Procedural success 76 (93%) 54 (98%) 0.23; 0.03-2.00 0.24Valve related complications (embolization, damage

to the valve, requirement of 2nd valve)2 (2%) 4 (7%) 0.32; 0.06-1.81 0.22

Myocardial infarction 1 (1%) 1 (2%) 0.67; 0.04-10.90 1.00Development of new conduction block 25 (30%) 19 (35%) 0.83; 0.40-1.72 0.709Requirement of PPM implantation 19 (23%) 9 (16%) 1.54; 0.64-3.72 0.39Major vascular complications 7 (9%) 1 (2%) 5.04; 0.60-42.19 0.144Cerebrovascular events (all) 4 (5%) 3 (5%) 0.89; 0.19-4.14 1.00Stroke 4 (5%) 1 (2%)Repeat hospitalization in 1 month 16/75 (21%) 7/53 (13%) 1.78; 0.68-4.69 0.350Death within 1 month 7 (9%) 1 (2%) 5.04; 0.60-42.19 0.144Death after 1 month 5 (6%) 3 (5%) 1.12; 0.26-4.92 1.00Overall mortality 12 (15%) 4 (7%) 2.19; 0.67-7.17 0.278Overall mortality secondary to cardiac causes 7 (9%) 1 (2%) 5.04; 0.60-42.19 0.144Length of hospitalization (days) 16.21 � 8.06 9.17 � 4.29 <0.0001Composite of clinical outcomes

(All-cause mortality, stroke, vascular complicationsand repeat hospitalization in 1 month)

39 (48%) 13 (24%) 2.93; 1.37-6.26 0.0068

224 The American Journal of Cardiology (www.ajconline.org)

left ventricular ejection fraction before the procedure wasassociated with a shortening of hospital stay by an average of0.10 days (95% CI 0.17 to 0.03; p ¼ 0.006).

Discussion

This study resulted in 4 important findings. First, thepresence of previous AF was an important predictor ofthe composite of all-cause mortality, stroke, vascular

complications, and repeat hospitalization in 1 month.Second, NOAF after TAVR was detected in 30% of thepatients who did not have a history of AF. Third, the useof transapical approach was an important predictor ofdevelopment of NOAF after TAVR. Fourth, the presenceof post-TAVR AF in patients was a strong predictor ofduration of post-TAVR ICU and overall hospital stayand demonstrated a trend toward increased 1-yearmortality.

Valvular Heart Disease/Impact of AF in Patients Who Underwent TAVR 225

In our study, the presence of previous AF was found tobe an important predictor of the composite outcomes of all-cause mortality, stroke, vascular complications, and repeathospitalizations within 1 month after TAVR. The findings inthis study are consistent with the epidemiologic observationsof greater cumulative and cardiovascular mortality in patientswith AF.22 From a mechanistic standpoint, the presence ofAF may cause a loss of atrioventricular synchrony that canlead to an impairment of cardiac function. The manifestationsof such adverse hemodynamic changes could be moreprofound in the patients with severe AS.

This study demonstrated that the incidence of NOAF inpatients without a history of this arrhythmia was 30%. Theincidence of NOAF within a 30-day follow-up period afterTAVR was reported to be 15% in the Placement of AorticTranscatheter Valve (PARTNER) trial.2 This relativelyhigher incidence of NOAF in our study could be attributed toa greater number of patients who underwent TAVR throughthe transapical approach (81 of 37 patients; 59%) comparedwith 33% of patients in the PARTNER trial who underwentTAVR through this approach. In a single-center based studyon 138 patients by Santos et al,23 the incidence of NOAF wasreported to be 31%. Another recently published study byTanawuttiwat et al24 reported a 42% incidence of NOAF in123 patients. Both these studies identified the transapicalapproach as an important procedural risk factor for thedevelopment of NOAF. Similar to the findings in thesestudies, the use of transapical approach was a strong predictorfor the occurrence of NOAF after TAVR in our study.

The use of transapical approach to perform TAVR entailsa left mini-thoracotomy and a direct puncture of left ven-tricular apex. The predisposition to develop NOAF usingtransapical approach in the setting of TAVR seems similar tothe setting of thoracotomy.25 The development of NOAF inpatients who underwent thoracotomy seems to be mediatedby multiple mechanisms; local trauma during thoracotomycan potentially lead to an injury to the sympathovagal fibersinnervating the sinus node that can sensitize the atrialmyocardium to catecholamines and, thereby, promote theinitiation of AF. Local trauma as a result of direct myocardialpuncture of left ventricular apex during the transapicalapproach can also lead to pain response and ventilator re-striction that can further potentiate the hyperadrenergic stateand lead to NOAF after TAVR. The occurrence of NOAFrelatively early in the postoperative course both in the settingof TAVR and after thoracotomy also may implicate periop-erative inflammation as a relevant risk factor.26

In this study, AF after TAVR was identified to be animportant predictor of the duration of post-TAVR ICUand hospital stay. Patients with post-TAVR AF were alsoobserved to have a trend toward increased all-cause mortal-ity. We believe that the AF after TAVR could possiblyindicate a relatively permanent form of AF that persiststhroughout the periprocedural and postprocedural course.The findings of an increased duration of ICU stay in thesepatients might be because of the loss of “atrial kick” thatfurther leads to an impairment of left ventricular systolicfunction. These adverse hemodynamic changes might lead toa greater requirement for hemodynamic support and, there-fore, contribute to a prolonged ICU stay. The presence ofpost-TAVR AF could also represent the coexistence of left

ventricular dysfunction and impaired vascular function inthese patients. The pathophysiological changes mediatedby AF and the greater incidence of coexisting co-morbidconditions in patients with AF might explain the findings inour study.

The findings of increased duration of ICU and overallhospital stay in the patients with AF are also relevant from ahealth-care economic perspective. To the best of ourknowledge, this study is the first to investigate the impact ofall 3 aspects of periprocedural AF (previous, new-onset, andpost-TAVR AF) on clinical and adverse outcomes afterTAVR. Our study is limited by its single-center, retrospec-tive, and nonrandomized design. We believe that the ECGsand telemetry strips might be limited for detection of transientepisodes of AF. We also acknowledge that the assessment ofechocardiographic parameters is also subjected to interob-server variability.

Disclosures

Dr. Kevin Heist: (all modest in amount): Biotronik(research grant, honoraria), Boston Scientific (research grant,consultant, honoraria), Medtronic (honoraria), Sanofi(consultant), Sorin (consultant, honoraria), St. Jude Medical(research grant, consultant, honoraria); Dr. Passeri: EdwardLifesciences (proctoring role); Ptaszek: Mast Therapeutics(consultant), St. Jude Medical (consultant), World CareClinical (consultant); Dr. Ruskin: Advanced MedicalEducation (consultant), Astellas/Cardiome (consultant,significant), Biosense Webster (consultant, modest) andFellowship Support (significant), Boston Scientific-Fellowship Support (significant), Bristol Myers Squibb(consultant), CardioInsight-Scientific Advisory Board(modest), CryoCath-Scientific Steering Committee (nocompensation), InfoBionic-Scientific Advisory Board andEquity, Medtronic (consultant, modest) and FellowshipSupport (significant), Pfizer-Consultant and ScientificSteering Committee (modest), Portola-Consultant and equity(modest), Sanofi-Consultant (modest), St. Jude Medical-Fellowship Support (significant); Dr. Mansour: Biosense-Webster (consultant, research grant), Boston Scientific(research grant), St. Jude (consultant), Cardiofocus (researchgrant), Endosense (research grant), MC10 (research grant),St. Jude Medical (consultant, research grant), VoyageMedical (research grant); Dr. Sundt: Thrasos (SteeringCommittee). Dr. Maan reports nothing to disclose.

1. Tamburino C, Capodanno D, Ramondo A, Petronio AS, Ettori F,Santoro G, Klugmann S, Bedogni F, Maisano F, Marzocchi A, Poli A,Antoniucci D, Napodano M, De Carlo M, Fiorina C, Ussia GP.Incidence and predictors of early and late mortality after transcathetervalve implantation in 663 patients with severe aortic stenosis. Circu-lation 2011;123:299e308.

2. Smith CR, Leon MB, Mack MJ, Miller DC, Moses JW, Svensson LG,Tuzcu EM, Webb JG, Fontana GP, Makkar RR, Williams M, Dewey T,Kapadia S, Babaliaros V, Thourani VH, Corso P, Pichard AD, BavariaJE, Herrman HC, Akin JJ, Anderson WN, Wang D, Pocock SJ;PARTNER Trial Investigators. Transcatheter versus surgical aorticvalve-replacement in high-risk patients. N Engl J Med 2011;364:2187e2198.

3. Thomas M, Schymik G, Walther T, Himbert D, Leferve T, Treede H,Eggebrecht H, Rubino P, Colombo A, Lange R, Schwarz RR, WendlerO. One-year outcomes of cohort 1 in the Edward SAPIEN AorticBioprosthesis European Outcome (SOURCE) registry: the European

226 The American Journal of Cardiology (www.ajconline.org)

registry of transcatheter aortic valve implantation using the EdwardsSAPIEN valve. Circulation 2011;124:425e433.

4. Chugh SS, Blackshear JL, Shen WK, Hammill SC, Gersch BJ.Epidemiology and natural history of atrial fibrillation: clinical impli-cations. J Am Coll Cardiol 2001;37:371e378.

5. Norberg J, Bäckström S, Jansson JH, Johansson L. Estimating theprevalence of atrial fibrillation in a general population using validatedelectronic health data. Clin Epidemiol 2013;5:475e481.

6. Lee JS, Shim CY, Wi J, Juong B, Ha JW, Lee MH, Pak HN. Leftventricular diastolic dysfunction is closely associated with mechanicalfunction of the left atrium in patients with paroxysmal atrial fibrillation.Circ J 2013;77:697e704.

7. Chrispin J, Jain A, Soliman EZ, Guallar E, Alonso A, Heckbert SR,Bluemke DA, Lima JA, Nazarian S. Association of electrocardio-graphic and imaging surrogates of left ventricular hypertrophy withincident atrial fibrillation: the Multi-Ethnic Study of Atherosclerosis.J Am Coll Cardiol 2014;63:2007e2013.

8. Bang CN, Dalsgaard M, Greve AM, Kober L, Gohlke-Baerwolf C, RayS, Rossebo AB, Egstrup K, Wachtell K. Left atrial size and function aspredictors of new-onset of atrial fibrillation in patients with asymp-tomatic aortic stenosis: the Simvastatin and Ezetimibe in AorticStenosis study. Int J Cardiol 2013;168:2322e2327.

9. Levy F, Garayalde E, Quere JP, Ianetta-Peltier M, Peltier M, TriboulloyC. Prognostic value of preoperative atrial fibrillation in patients withsevere aortic stenosis and low ejection fraction having aortic valvereplacement. Am J Cardiol 2006;98:809e811.

10. Saxena A, Dinh DT, Reid CM, Smith JA, Shardey GC, Newcomb AE.Does preoperative atrial fibrillation portend a poorer prognosis inpatients undergoing isolated aortic valve replacement? A multicentreAustralian study. Can J Cardiol 2013;29:697e703.

11. Ngaage DL, Schaff HV, Barnes SA, Sundt TM III, Mullany CJ,Dearani JA, Daly RC, Orszulak TA. Prognostic implications ofpreoperative atrial fibrillation in patients undergoing aortic valvereplacement: is there an argument for concomitant arrhythmia surgery?Ann Thorac Surg 2006;82:1392e1399.

12. Ngaage DL, Schaff HV, Mullany CJ, Barnes S, Dearani JA, Daly RC,Orszulak TA, Sundt TM III. Does preoperative atria fibrillation influ-ence early and late outcomes of coronary artery bypass grafting?J Thorac Cardiovasc Surg 2007;133:182e189.

13. Dulli DA, Stanko H, Levine RL. Atrial fibrillation is associated withsevere acute ischemic stroke. Neuroepidemiology 2003;22:118e123.

14. Levy F, Debry N, Labescat AL, Meimoun P, Malaquin D, MarechauxS, Rusinaru D, Jeu A, Ennezat PV, Castel AL, Tribouilloy C.Echocardiographic prediction of postoperative atrial fibrillation afteraortic valve replacement for aortic stenosis: a two-dimensional speckletracking left ventricular longitudinal strain multicentre pilot study. ArchCardiovasc Dis 2012;105:499e506.

15. Online STS Risk Calculator- Society of Thoracic Surgeons. Availableat: http://riskcalc.sts.org/STSWebRiskCalc273/de.aspx. Accessed onAugust 08, 2014.

16. Roques F, Michel P, Goldstone AR, Nashef SA. The logisticEuroSCORE. Eur Heart J 2003;24:882e883.

17. Roques F, Nashef SA, Michel P, Gauducheau E, de Vincentiis C,Baudet E, Cortina J, David M, Faichney A, Gabrielle F, Gams E,

Harjula A, Jones MT, Pintor PP, Salamon R, Thulin L. Risk factors andoutcome in European cardiac surgery: analysis of the EuroSCOREmultinational database of 19,030 patients. Eur J Cardiothorac Surg1999;15:816e822.

18. Anderson JL, Halperin JL, Albert NM, Bozkurt B, Brinidis RG, CurtisLH, DeMets D, Guyton RA, Hochman JS, Kovacs RJ, Ohman EM,Pressler SJ, Sellke FW, Shen WK, Wann LS, Curtis AB, EllenbogenKA, Estes NA III, Ezekowitz MD, Jackman WM, January CT, LoweJE, Page RL, Slotwiner DJ, Stevenson WG, Tracy CM, Fuster V,Ryden LE, Cannom DS, Crinjs HJ, Curtis AB, Ellenbogen KA, LeHeuzey JY, Kay GN, Olsson SB, Prystowsky EN, Tamargo JL, WannS. Management of patients with atrial fibrillation (compilation of 2006ACCF/AHA/ESC and 2011 ACCF/AHA/HRS recommendations): areport of the American College of Cardiology/American HeartAssociation Task Force on Practice Guidelines. J Am Coll Cardiol2013;61:1935e1944.

19. El-Chami MF, Kligo PD, Elfstrom KM, Halkos M, Thourani V,Lattouf OM, Delurgio DB, Guyton RA, Leon AR, Puskas JD.Prediction of new-onset atrial fibrillation after cardiac revascularizationsurgery. Am J Cardiol 2012;110:649e654.

20. Arnold SV, Spertus JA, Lei Y, Green P, Kirtane AJ, Kapadia S,Thourani VH, Herrmann HC, Beohar N, Zajarias A, Mack MJ, LeonMB, Cohen DJ. How to define a poor outcome after transcatheter aorticvalve replacement: conceptual framework and empirical observationsfrom the Placement of Aortic Transcatheter Valve (PARTNER) trial.Circ Cardiovasc Qual Outcomes 2013;6:591e597.

21. Reynolds MR, Magnuson EA, Lei Y, Leon MB, Smith CR, SvenssonLG, Webb JG, Babaliaros VC, Bowers BS, Fearon WF, Herrmann HC,Kapadia S, Kodali SK, Makkar RR, Pichard AD, Cohen DJ; Placementof Aortic Transcatheter Valves (PARTNER) investigators. Health-related quality of life after transcatheter aortic valve replacement ininoperable patients with severe aortic stenosis. Circulation 2011;124:1964e1972.

22. Kannel WB, Abbott RD, Savage DD, McNamara PM. Epidemiologicfeatures of chronic atrial fibrillation: the Framingham study. N Engl JMed 1982;306:1018e1022.

23. Amat-Santos JJ, Rodes-Cabau J, Urena M, DeLarochelliere R,Doyle D, Bagur R, Villeneuve J, Cote M, Nombela-Franco L,Philippon F, Pibarot P, Dumont E. Incidence, predictive factors, andprognostic value of new-onset atrial fibrillation following trans-catheter aortic valve replacement. J Am Coll Cardiol 2012;59:178e188.

24. Tanawuttiwat T, O’Neill BP, Cohen MG, Chintakanan O, HeldmanAW, Martinez CA, Alfonso CE, Mitrani RD, Macon CJ, Carrillo RG,Williams DB, O’Neill WW, Myerburg RJ. New-onset atrial fibrillationafter aortic valve replacement: comparison of transfemoral, transapical,transaortic and surgical approaches. J Am Coll Cardiol 2014;63:1510e1519.

25. Amar D. Prevention and management of perioperative arrhythmiasin the thoracic surgical population. Anesthesiol Clin 2008;26:325e335.

26. Amar D, Goenka A, Zhang H, Park B, Thaler HT. Leuokocytosis andincreased risk of atrial fibrillation after general thoracic surgery. AnnThorac Surg 2006;82:1057e1061.