Portal de Periódicos da CAPES

Stroke and Cardiovascular Events After Ablation or Antiarrhythmic Drugs for Treatment of Patients With Atrial Fibrillation

2018; Elsevier BV; Volume: 121; Issue: 10 Linguagem: Inglês

10.1016/j.amjcard.2018.01.043

1879-1913

Moussa Mansour, E. Kevin Heist, Rahul Agarwal, T. Jared Bunch, Edward Karst, Jeremy N. Ruskin, Srijoy Mahapatra,

Cardiac electrophysiology and arrhythmias

Catheter ablation and antiarrhythmic drugs (AADs) are the most common rhythm-control strategies for atrial fibrillation (AF). Data comparing the rate of stroke and cardiovascular events between the treatment strategies are limited. Therefore, this observational study uses claims data to compare rate of cardiovascular hospitalization and stroke for patients with AF treated with ablation or AADs. Patients in the MarketScan dataset with AF between January 2010 and December 2014 were categorized in the ablation group if an atrial catheter ablation was performed, or in the AAD group if a relevant AAD was prescribed for AF but no ablation was performed. One year of history was required, and the index event was selected as the most recent ablation or AAD prescription closest to January 1, 2013. A 2:1 propensity score match was performed for age, gender, co-morbidities, and total medical cost in the year before index event. Outcomes included thromboembolic event (ischemic stroke, transient ischemic attack, or systemic embolism) and all cardiovascular hospitalizations. Of the 164,639 patients in the AAD group, 29,456 were matched to the 14,728 ablation patients. There were no significant differences in age (64 ± 10 in both groups), gender (58% male), or CHA2DS2-VASc score (3.2 ± 1.3). Risk of hospitalization with primary diagnosis of thromboembolic event was 41% greater in the AADs group (p < 0.001), and cardiovascular hospitalizations were 13% more likely (p < 0.001). In conclusion, patients treated with catheter ablation of AF have lower risk of thromboembolic events and cardiovascular hospitalizations than a matched cohort of patients managed with AADs. Catheter ablation and antiarrhythmic drugs (AADs) are the most common rhythm-control strategies for atrial fibrillation (AF). Data comparing the rate of stroke and cardiovascular events between the treatment strategies are limited. Therefore, this observational study uses claims data to compare rate of cardiovascular hospitalization and stroke for patients with AF treated with ablation or AADs. Patients in the MarketScan dataset with AF between January 2010 and December 2014 were categorized in the ablation group if an atrial catheter ablation was performed, or in the AAD group if a relevant AAD was prescribed for AF but no ablation was performed. One year of history was required, and the index event was selected as the most recent ablation or AAD prescription closest to January 1, 2013. A 2:1 propensity score match was performed for age, gender, co-morbidities, and total medical cost in the year before index event. Outcomes included thromboembolic event (ischemic stroke, transient ischemic attack, or systemic embolism) and all cardiovascular hospitalizations. Of the 164,639 patients in the AAD group, 29,456 were matched to the 14,728 ablation patients. There were no significant differences in age (64 ± 10 in both groups), gender (58% male), or CHA2DS2-VASc score (3.2 ± 1.3). Risk of hospitalization with primary diagnosis of thromboembolic event was 41% greater in the AADs group (p < 0.001), and cardiovascular hospitalizations were 13% more likely (p < 0.001). In conclusion, patients treated with catheter ablation of AF have lower risk of thromboembolic events and cardiovascular hospitalizations than a matched cohort of patients managed with AADs. Atrial fibrillation (AF) is the most common sustained arrhythmia, with an estimated 2.7 to 6.1 million cases in the United States.1Go A.S. Hylek E.M. Phillips K.A. Chang Y. Henault L.E. Selby J.V. Singer D.E. Prevalence of diagnosed atrial fibrillation in adults.JAMA. 2001; 285: 2370-2375Crossref PubMed Scopus (5189) Google Scholar, 2Wyndham C.R.C. Atrial fibrillation: the most common arrhythmia.Texas Heart Inst J. 2000; 27: 257-267PubMed Google Scholar, 3Rahman F. Kwan G.F. Benjamin E.J. Global epidemiology of atrial fibrillation.Nat Rev Cardiol. 2014; 11: 639-654Crossref PubMed Scopus (436) Google Scholar AF is associated with 5-fold increase in stroke risk,4Wolf P.A. Abbott R.D. Kannel W.B. Atrial fibrillation as an independent risk factor for stroke: the Framingham Study.Stroke. 1991; 22: 983-988Crossref PubMed Scopus (5740) Google Scholar and is treated by rate control or rhythm control strategies, along with anticoagulation to prevent formation of thrombus.5January C.T. Wann L.S. Alpert J.S. Calkins H. Cleveland J.C. Cigarroa J.E. Conti J.B. Ellinor P.T. Ezekowitz M.D. Field M.E. 2014 AHA/ACC/HRS guideline for the management of patients with atrial fibrillation.Circulation. 2014; 130: 2071-2104Crossref PubMed Scopus (1482) Google Scholar, 6Calkins H. Hindricks G. Cappato R. Kim Y.-H. Saad E.B. Aguinaga L. Akar J.G. Badhwar V. Brugada J. Camm J. Chen P.-S. Chen S.-A. Chung M.K. Nielsen J.C. Curtis A.B. Davies D.W. Day J.D. D'Avila A. de Groot N.M.S.N. Di Biase L. Duytschaever M. Edgerton J.R. Ellenbogen K.A. Ellinor P.T. Ernst S. Fenelon G. Gerstenfeld E.P. Haines D.E. Haissaguerre M. Helm R.H. Hylek E. Jackman W.M. Jalife J. Kalman J.M. Kautzner J. Kottkamp H. Kuck K.H. Kumagai K. Lee R. Lewalter T. Lindsay B.D. Macle L. Mansour M. Marchlinski F.E. Michaud G.F. Nakagawa H. Natale A. Nattel S. Okumura K. Packer D. Pokushalov E. Reynolds M.R. Sanders P. Scanavacca M. Schilling R. Tondo C. Tsao H.M. Verma A. Wilber D.J. Yamane T. 2017 HRS/EHRA/ECAS/APHRS/SOLAECE expert consensus statement on catheter and surgical ablation of atrial fibrillation.Heart Rhythm. 2017; 14: e275-e444Abstract Full Text Full Text PDF PubMed Scopus (1076) Google Scholar Despite benefits in retrospective studies with rhythm control approaches,7Tsadok M.A. Jackevicius C.A. Essebag V. Eisenberg M.J. Rahme E. Humphries K.H. Tu V, J. Behlouli H. Pilote L. Rhythm versus rate control therapy and subsequent stroke or transient ischemic attack in patients with atrial fibrillation: clinical perspective.Circulation. 2012; 126: 2680-2687Crossref PubMed Scopus (99) Google Scholar rate control strategies have yielded similar outcomes in randomized prospective trials.8Wyse D.G. Waldo A.L. DiMarco J.P. Domanski M.J. Rosenberg Y. Schron E.B. Kellen J.C. Greene H.L. Mickel M.C. Dalquist J.E. Corley S.D. A comparison of rate control and rhythm control in patients with atrial fibrillation.N Engl J Med. 2002; 347: 1825-1833Crossref PubMed Scopus (3672) Google Scholar, 9Roy D. Talajic M. Nattel S. Wyse D.G. Dorian P. Lee K.L. Bourassa M.G. Arnold J.M.O. Buxton A.E. Camm A.J. Connolly S.J. Dubuc M. Ducharme A. Guerra P.G. Hohnloser S.H. Lambert J. Le Heuzey J.-Y. O'Hara G. Pedersen O.D. Rouleau J.-L. Singh B.N. Stevenson L.W. Stevenson W.G. Thibault B. Waldo A.L. Rhythm control versus rate control for atrial fibrillation and heart failure.N Engl J Med. 2008; 358: 2667-2677Crossref PubMed Scopus (1241) Google Scholar As a consequence, rate control approaches remain more common in patients with asymptomatic AF. In part because of long-term side effects and intolerance to antiarrhythmic drugs (AADs), catheter ablation has evolved to become a rhythm control treatment for selected patients with symptomatic drug-refractory AF.10Reddy V.Y. Dukkipati S.R. Neuzil P. Natale A. Albenque J.P. Kautzner J. Shah D. Michaud G. Wharton M. Harari D. Mahapatra S. Lambert H. Mansour M. Randomized, controlled trial of the safety and effectiveness of a contact force-sensing irrigated catheter for ablation of paroxysmal atrial fibrillation: results of the TactiCath Contact Force Ablation Catheter Study for Atrial Fibrillation (TOCCASTAR) S.Circulation. 2015; 132: 907-915Crossref PubMed Scopus (269) Google Scholar, 11Sarabanda A.V. Bunch T.J. Johnson S.B. Mahapatra S. Milton M.A. Leite L.R. Bruce G.K. Packer D.L. Efficacy and safety of circumferential pulmonary vein isolation using a novel cryothermal balloon ablation system.J Am Coll Cardiol. 2005; 46: 1902-1912Crossref PubMed Scopus (188) Google Scholar, 12Natale A. Reddy V.Y. Monir G. Wilber D.J. Lindsay B.D. McElderry H.T. Kantipudi C. Mansour M.C. Melby D.P. Packer D.L. Nakagawa H. Zhang B. Stagg R.B. Boo L.M. Marchlinski F.E. Paroxysmal A.F. Catheter ablation with a contact force sensing catheter.J Am Coll Cardiol. 2014; 64: 647-656Crossref PubMed Scopus (350) Google Scholar, 13Malchano Z.J. Neuzil P. Cury R.C. Holmvang G. Weichet J. Schmidt E.J. Ruskin J.N. Reddy V.Y. Integration of cardiac CT/MR imaging with three-dimensional electroanatomical mapping to guide catheter manipulation in the left atrium: implications for catheter ablation of atrial fibrillation.J Cardiovasc Electrophysiol. 2006; 17: 1221-1229Crossref PubMed Scopus (129) Google Scholar, 14Mansour M. Karst E. Heist E.K. Dalal N. Wasfy J.H. Packer D.L. Calkins H. Ruskin J.N. Mahapatra S. The impact of first procedure success rate on the economics of atrial fibrillation ablation.JACC Clin Electrophysiol. 2017; 3: 129-138Crossref PubMed Scopus (28) Google Scholar, 15Bunch T.J. Crandall B.G. Weiss J.P. May H.T. Bair T.L. Osborn J.S. Anderson J.L. Muhlestein J.B. Horne B.D. Lappe D.L. Day J.D. Patients treated with catheter ablation for atrial fibrillation have long-term rates of death, stroke, and dementia similar to patients without atrial fibrillation.J Cardiovasc Electrophysiol. 2011; 22: 839-845Crossref PubMed Scopus (291) Google Scholar The ongoing Catheter Ablation versus Antiarrhythmic Drug Therapy for Atrial Fibrillation multicenter randomized controlled trial compares outcomes of upfront catheter ablation versus pharmacological management of AF.16Cleland J.G.F. Coletta A.P. Buga L. Ahmed D. Clark A.L. Clinical trials update from the American College of Cardiology meeting 2010: DOSE, ASPIRE, CONNECT, STICH, STOP-AF, CABANA, RACE II, EVEREST II, ACCORD, and NAVIGATOR.Eur J Heart Fail. 2010; 12: 623-629Crossref PubMed Scopus (61) Google Scholar To understand contemporary real-world treatment patterns, we undertook a retrospective observational study to compare outcomes in patients receiving catheter ablation or AADs for AF. This retrospective observational study used MarketScan administrative claims (IBM Truven Health, Ann Arbor, MI) from January 2010 to December 2014. The dataset includes persons in the United States with private insurance or Medicare supplemental insurance, and contains records of inpatient hospitalizations, outpatient visits, and prescription drugs. Records include dates of service, diagnosis and procedure codes, and number of days and type of prescription. Patients with AF (International Classification of Diseases, Ninth Revision, Clinical Modification diagnosis code 427.31) and at least 1 year of linked drug coverage were extracted from MarketScan. Those treated with rhythm control (ablation or AADs) after at least 1 year of enrollment were placed in 2 groups:•Ablation: Catheter ablation of atrial arrhythmia, defined as Current Procedural Terminology outpatient procedure code 93651, 93654, or 93656 or International Classification of Diseases, Ninth Revision procedure code 37.34 with accompanying primary AF diagnosis, or•AADs: Prescription or infusion of class Ia, Ic, or III AAD, but no catheter ablation of AF observed throughout enrollment period. For the AAD group, the index event was identified as the AAD prescription that occurred after 1 year of coverage and was closest to January 1, 2013; for the ablation group, the index event was identified as the most recent atrial outpatient or inpatient ablation after 1 year of coverage. Additional criteria required for inclusion were at least 2 encounters with AF diagnosis by time of index event, no atrial, ventricular or atrioventricular node ablation before first 2 encounters with AF diagnosis, age 18 years or older at time of index event, no more than 2 months' gap in insurance coverage in the year before index event, linked prescription drug coverage in the year of index event, and CHA2DS2-VASc score of ≥2. The supplementary materials contain further details on diagnosis and procedure codes used for patient selection. Baseline characteristics including age, gender, co-morbidities, CHA2DS2-VASc score, history of cardioversion, and anticoagulant usage were identified in the period up to the index event. Total medical cost was determined as the sum of reimbursement from all sources in the year before index event, and categorized into low, medium, and high terciles with cutoffs from the ablation group. A 2:1 propensity score match17Caliendo M. Kopeinig S. Some practical guidance for the implementation of propensity score matching.J Econ Surv. 2008; 22: 31-72Crossref Scopus (2836) Google Scholar was used to identify a cohort of patients prescribed AADs with similar baseline characteristics to those receiving an ablation. Variables used for matching were age, gender, history of hypertension, diabetes mellitus, heart failure, vascular disease, stroke, cardioversion, usage of anticoagulants, and terciles of total medical cost in the previous year. Usage of additional prescriptions for AADs and oral anticoagulants was calculated in the year before and the year after index event for the matched subgroup of patients with at least 1 year follow-up before and after the index event. The index event from the AAD group was excluded from the calculation of prescription usage, as were any medications used within 3 months after index event in the ablation group. The fraction with prescriptions covering at least 180 days in the year before and after index event was also determined. The primary outcome was hospitalization for thromboembolic event, defined as an inpatient hospitalization with a primary diagnosis code of ischemic stroke, transient ischemic attack (TIA), or systemic embolism. Freedom from hospitalization was compared between the propensity-matched AAD and ablation groups. Additional outcomes included hospitalizations with a primary diagnosis code for stroke of any kind (ischemic or hemorrhagic), hemorrhagic stroke, major bleeding excluding hemorrhagic stroke, and cardiovascular disease. Finally, all hospitalizations that contained any diagnosis code for cardiovascular disease were determined. Use of all diagnosis codes accounts for hospitalizations with primary cardiovascular diagnosis, as well as those with secondary involvement of cardiovascular disease or its treatment. Because of potential side effects of AADs such as amiodarone,18Shukla R. Jowett N.I. Thompson D.R. Pohl J.E. Side effects with amiodarone therapy.Postgrad Med J. 1994; 70: 492-498Crossref PubMed Scopus (32) Google Scholar, 19Guccione P. Paul T. Garson A. Long-term follow-up of amiodarone therapy in the young: continued efficacy, unimpaired growth, moderate side effects.J Am Coll Cardiol. 1990; 15: 1118-1124Crossref PubMed Scopus (98) Google Scholar, 20De Ferrari G.M. Dusi V. Drug safety evaluation of dronedarone in atrial fibrillation.Expert Opin Drug Saf. 2012; 11: 1023-1045Crossref PubMed Scopus (21) Google Scholar we assessed the prevalence of thyroid dysfunction, pulmonary toxicity, optic neuritis, liver disease, and bradycardia in the AAD group in the year before index event, the year beginning on the day of index event, and combined. For the ablation group, we noted the prevalence of hospitalization within 30 days before the index event and 30 days beginning on the day of index event, because of thromboembolic event, vascular complication or hematoma, perforation or tamponade, phrenic nerve palsy, and pneumothorax or hemothorax. Subgroups were formed in the ablation group with previous prescription for warfarin or a direct oral anticoagulant (DOAC). A 2:1 propensity score match was used to match patients in the AAD group with the same type of anticoagulant to each ablation subgroup. Hospitalization because of thromboembolic event was compared between those treated with AADs or ablation for each type of anticoagulant, and a multiway comparison determined risk of hospitalization for thromboembolic event as a function of previous oral anticoagulant and therapy for AF. Continuous variables were summarized using mean and standard deviation and compared using t tests. For categorical variables, number and percentage were reported and compared with a chi-square test. Freedom from hospitalization was determined using the Kaplan-Meier method, and compared using a log-rank test. A multivariable Cox proportional hazards model was constructed to determine the risk of hospitalization as a function of therapy for AF. Covariates in the model were age, gender, history of hypertension, diabetes, heart failure, vascular disease, previous stroke, previous cardioversion, and oral anticoagulant usage. Event rates were computed by dividing total number of events by total patient-years, and compared between the 2 groups assuming a Poisson process. The level of significance was 0.05. R and Matlab were used for statistical analysis and data visualization. Of 1,829,712 patients with an AF diagnosis from 2010 to 2014, there were 337,935 with linked drug data treated with AADs or catheter ablation. After applying additional criteria (Figure 1), there were 14,728 patients in the ablation group and 164,639 patients in the AAD group. The 2:1 propensity match yielded a cohort of 29,456 AAD and 14,728 ablation patients. Patients in both groups were aged 64 ± 10 years (p = 0.15) and 58% were male (p = 0.43). The only significant differences after matching were a slightly higher rate of previous stroke in the ablation group as compared with the AAD group (17% vs 16%; p = 0.002) and slightly lower prevalence of hypertension (92% vs 93%; p < 0.001) as shown in Table 1.Table 1Basic characteristicsVariableAblation (n = 14,728)AADs (n = 164,639)p valueAADs (2:1 Matched) (n = 29,456)p valueAge (years)64.1 ± 9.772.3 ± 11.4<0.00164.0 ± 10.00.154Women6,135(42%)79,333(48%)<0.00112,386(42%)0.429Men8,593(58%)86,305(52%)17,070(58%)Hypertension13,496(92%)149,147(91%)<0.00127,292(93%)<0.001Diabetes mellitus5,505(37%)64,461(39%)<0.00111,123(38%)0.433Heart failure5,370(36%)70,467(43%)<0.00110,806(37%)0.645Vascular disease2,966(20%)41,304(25%)<0.0015,828(20%)0.381Prior stroke2,506(17%)35,791(22%)<0.0014,676(16%)0.002CHA2DS2-VASc3.2 ± 1.44.1 ± 1.7<0.0013.2 ± 1.30.542Years after index event1.3 ± 1.01.2 ± 0.9<0.0011.3 ± 0.90.365Prior cardioversion7,122(48%)31,698(19%)<0.00113,840(47%)0.007Prior oral anticoagulant12,962 (88%)102,931(63%)<0.00124,956(88%)0.740AADs = antiarrhythmic drugs; CHA2DS2-VASc = congestive heart failure, hypertension, age, diabetes, stroke, vascular risk factors, age, sex. Open table in a new tab AADs = antiarrhythmic drugs; CHA2DS2-VASc = congestive heart failure, hypertension, age, diabetes, stroke, vascular risk factors, age, sex. In the year before index event, AAD usage was slightly lower and anticoagulant usage was slightly higher in the ablation group (77% vs 80% and 86% vs 81%, p < 0.001). During the year after the index event, AAD usage in patients followed a full year was much higher in the AAD group (67% vs 45%; p < 0.001), whereas anticoagulant usage was similar (70% vs 73%; p < 0.001). See Table 2.Table 2Medications before and after index event in ablation and AADs groupsMedicationYear Before Index Event*Includes only patients with at least 1 year of follow-up after index event.Year After Index Event*Includes only patients with at least 1 year of follow-up after index event.Ablation (n = 7,649)AADs (n = 16,903)p valueAblation†First 3 months were blanked. (n = 7,649)AADs (n = 16,903)p valueAny AADs5,879 (77%)13,673 (80%)<0.0013,446 (45%)11,332 (67%) 180 days)3,306 (43%)9,089 (53%)<0.0011,919 (25%)9,852 (58%)<0.001Amiodarone1,761 (23%)5,232 (31%)<0.001969 (13%)3,936 (23%)<0.001Dronedarone1,674 (22%)2,520 (15%)<0.001432 (6%)1,529 (9%)<0.001Sotalol1,561 (20%)3,106 (18%)<0.001883 (12%)2,736 (16%)<0.001Any anticoagulants6,587 (86%)13,682 (81%)<0.0015,371 (70%)12,375 (73%) 180 days)3,736 (49%)8,065 (47%)<0.0013,675 (48%)9,846 (58%)<0.001Warfarin4,269 (56%)8,498 (50%)<0.0012,853 (37%)7,548 (44%)<0.001DOAC3,272 (43%)5,353 (32%)<0.0012,715 (35%)5,389 (32%)<0.001AADs = antiarrhythmic drugs; DOAC = direct oral anticoagulants.* Includes only patients with at least 1 year of follow-up after index event.† First 3 months were blanked. Open table in a new tab AADs = antiarrhythmic drugs; DOAC = direct oral anticoagulants. Prevalence of potential side effects of AADs is listed in Table 3. Thyroid dysfunction occurred in 25% of patients, followed by bradycardia (22%), pulmonary toxicity (16%), and liver disease (7%). Only 0.2% had a diagnosis of optic neuritis. Table 4 lists potential complications of ablation observed in the 30 days before and after the index event. Vascular complication or hematoma occurred in 3% of patients, whereas thromboembolic event was observed in just 0.1%.Table 3Prevalence of primary or secondary diagnosis codes of possible side effects of AADs such as amiodarone, in subgroup of the AAD cohort with 1 year of follow-up after index event (n = 16,903)Type of eventDuring 1 year before index eventDuring 1 year after index eventCombinedThyroid dysfunction3,033(18%)3,359(20%)4,257(25%)Pulmonary toxicity1,734(10%)1,381(8%)2,677(16%)Optic neuritis15(<1%)18(<1%)29(<1%)Liver disease725(4%)714(4%)1266(7%)Bradycardia2,462(14%)2,005(12%)3805(22%)AADs = antiarrhythmic drugs. Open table in a new tab Table 4Prevalence of primary or secondary diagnosis codes of known side effects of ablation in 30 days before and after ablation, n = 14,728Type of eventDuring 30 days before index eventDuring 30 days after index eventp valueThromboembolic event*Only inpatient hospitalization with primary codes are counted.≤10(≤0.07%)17(0.12%)0.010Bleeding*Only inpatient hospitalization with primary codes are counted.≤10(≤0.07%)69(0.47%)<0.001Vascular complication/ Hematoma49(0.33%)447(3.04%)<0.001Perforation/Tamponade≤10 (≤0.07%)107(0.73%)<0.001Phrenic nerve palsy12(0.08%)85(0.58%)<0.001Pneumothorax/ Hemothorax≤10 (≤0.07%)23(0.16%)<0.001* Only inpatient hospitalization with primary codes are counted. Open table in a new tab AADs = antiarrhythmic drugs. Risk of hospitalization for thromboembolic event was significantly higher in the AAD group than in the ablation group (hazard ratio [HR] 1.41, 95% confidence interval [CI] [1.15, 1.73], p < 0.001). Event risk was also higher for stroke of any kind (HR 1.40, 95% CI [1.16, 1.69], p < 0.001). No significant difference was found in risk of hemorrhagic stroke or bleeding (Table 5 and Figure 2). There was also an elevated rate of hospitalization for thromboembolic event and stroke of any kind in the AAD group compared with the ablation group (Table 6).Table 5Risk of hospitalization for patients using AADs, compared with ablationType of eventUnivariate HR [95% CI] p valueMultivariable HR [95% CI] p valueThromboembolic event1.41 [1.15–1.72] < 0.0011.41 [1.15–1.73] < 0.001Stroke of any kind1.40 [1.17–1.69] < 0.0011.40 [1.16–1.69] < 0.001Hemorrhagic stroke1.34 [0.89–2.01] 0.1641.33 [0.89–2.00] 0.166Bleeding1.14 [0.97–1.35] 0.1041.13 [0.96–1.33] 0.132Cardiovascular hospitalization (primary diagnosis code)1.03 [0.98–1.09] 0.2501.03 [0.98–1.09] 0.246Cardiovascular hospitalization (any diagnosis code)1.13 [1.09–1.17] < 0.0011.13 [1.09–1.17] < 0.001AADs = antiarrhythmic drugs; CI = confidence interval; HR = hazard ratio. Open table in a new tab Table 6Hospitalization rate per 100 yearsType of eventAblation Mean[95% CI]AADs Mean[95% CI]p valueThromboembolic event0.73 [0.61, 0.85]1.07 [0.97, 1.17]<0.001Stroke of any kind0.87 [0.74, 1.01]1.24 [1.13, 1.36]<0.001Hemorrhagic stroke0.18 [0.13, 0.25]0.23 [0.18, 0.28]0.248Bleeding1.21 [1.06, 1.37]1.37 [1.26, 1.49]0.101Cardiovascular hospitalization (primary diagnosis code)14.82 [14.27,15.37]16.01 [15.61, 16.41]<0.001Cardiovascular hospitalization (any diagnosis code)33.08[32.26,33.89]40.35[39.72,40.99]<0.001AADs = antiarrhythmic drugs; CI = confidence interval. Open table in a new tab AADs = antiarrhythmic drugs; CI = confidence interval; HR = hazard ratio. AADs = antiarrhythmic drugs; CI = confidence interval. There was a higher risk of cardiovascular hospitalization in the AAD group than in the ablation group (HR 1.13, 95% CI [1.09, 1.17], p < 0.001). However, if only the hospitalizations with a primary cardiovascular diagnosis were considered, there was no statistical difference (Table 5 and Figure 3). Rate of cardiovascular hospitalizations in the AAD group was 40.35 hospitalizations per 100 patient-years, compared with 33.08 in the ablation group (p < 0.001). Restricting to hospitalizations with a primary cardiovascular diagnosis code, the hospitalization rate was still higher for AADs (16.01 per 100 years) than ablation (14.82, p < 0.001, Table 6). In the subgroup prescribed DOACs, risk of hospitalization for thromboembolic event was higher for those prescribed AADs than for those undergoing ablation (HR 1.47; p = 0.030). The difference for the subgroup using warfarin did not reach statistical significance (HR 1.32; p = 0.053, Table 7). The multiway comparison analyzed risk for hospitalization because of thromboembolic event for each subgroup with those treated with ablation and DOACs, and found greatest risk in those prescribed AADs and warfarin (HR 1.64; p = 0.003), followed by nonsignificant increases for AADs and DOACs (HR 1.43; p = 0.051), as well as ablation and warfarin (HR 1.19; p = 0.367). In other covariates, risk of thromboembolic event was higher in older patients (additional risk of 2% per year, p < 0.001), females, and those with diabetes mellitus, heart failure, and previous stroke (Table 8).Table 7Multivariable regression indicating risk of hospitalization because of thromboembolic event categorized by previous oral anticoagulant typeVariablePrior warfarin6,591 Ablation patients13,182 AADs patientsHR [95% CI] p valuePrior DOACs6,371 Ablation patients12,742 AADs patientsHR [95% CI] p valueAADs vs. ablation1.32 [1.00, 1.73] 0.0531.47 [1.04, 2.09] 0.030Age (per year)1.02 [1.01, 1.03] 0.0031.01 [0.99, 1.03] 0.227Male gender0.86 [0.67, 1.12] 0.2580.73 [0.54, 0.99] 0.045Hypertension1.39 [0.85, 2.29] 0.1930.96 [0.55, 1.67] 0.889Diabetes1.29 [1.00, 1.67] 0.0541.71 [1.26, 2.32] < 0.001Heart failure1.51 [1.17, 1.96] 0.0021.29 [0.94, 1.77] 0.111Vascular disease1.00 [0.73, 1.35] 0.9801.01 [0.68, 1.49] 0.956Prior stroke2.28 [1.73, 3.01] < 0.0012.32 [1.65, 3.26] < 0.001Prior cardioversion0.92 [0.71, 1.19] 0.5241.00 [0.74, 1.36] 0.987AADs = antiarrhythmic drugs; CI = confidence interval; DOAC = direct oral anticoagulants; HR = hazard ratio. Open table in a new tab Table 8Multivariable regression indicating risk of hospitalization because of thromboembolic event, for each subgroup by rhythm-control strategy and type of anticoagulantVariableHR [95% CI] p valueAblation and prior DOAC (reference)1.00Ablation and prior warfarin1.19 [0.81–1.75] 0.367AADs and prior DOAC1.43 [1.00–2.05] 0.051AADs and prior warfarin1.64 [1.18–2.29] 0.003Age (years)1.02 [1.01–1.03] < 0.001Male gender0.79 [0.65–0.96] 0.018Hypertension1.13 [0.79–1.63] 0.502Diabetes1.46 [1.20–1.77] < 0.001Heart failure1.41 [1.16–1.71] < 0.001Vascular disease0.96 [0.76–1.21] 0.702Prior stroke2.35 [1.91–2.89] < 0.001Prior cardioversion1.03 [0.85–1.25] 0.729AADs = antiarrhythmic drugs; CI = confidence interval; DOAC = direct oral anticoagulants; HR = hazard ratio. Open table in a new tab AADs = antiarrhythmic drugs; CI = confidence interval; DOAC = direct oral anticoagulants; HR = hazard ratio. AADs = antiarrhythmic drugs; CI = confidence interval; DOAC = direct oral anticoagulants; HR = hazard ratio. In prospective randomized controlled trials, although catheter ablation was shown to reduce recurrence of AF and increases quality of life as compared with AADs,21Shi L. Heng R. Liu S. Leng F. Effect of catheter ablation versus antiarrhythmic drugs on atrial fibrillation: a meta-analysis of randomized controlled trials.Exp Ther Med. 2015; 10: 816-822Crossref PubMed Scopus (28) Google Scholar, 22Jaïs P. Cauchemez B. Macle L. Daoud E. Khairy P. Subbiah R. Hocini M. Extramiana F. Sacher F. Bordachar P. Klein G. Weerasooriya R. Clémenty J. Haïssaguerre M. Catheter ablation versus antiarrhythmic drugs for atrial fibrillation: clinical perspective.Circulation. 2008; 118: 2498-2505Crossref PubMed Scopus (651) Google Scholar, 23Morillo C.A. Verma A. Connolly S.J. Kuck K.H. Nair G.M. Champagne J. Sterns L.D. Beresh H. Healey J.S. Natale A. (RAAFT-2) Radiofrequency ablation vs antiarrhythmic drugs as first-line treatment of paroxysmal atrial fibrillation.JAMA. 2014; 311: 692-700Crossref PubMed Scopus (414) Google Scholar no significant reduction in stroke or TIA and all-cause mortality was found.21Shi L. Heng R. Liu S. Leng F. Effect of catheter ablation versus antiarrhythmic drugs on atrial fibrillation: a meta-analysis of randomized controlled trials.Exp Ther Med. 2015; 10: 816-822Crossref PubMed Scopus (28) Google Scholar This was possibly because of the small number of patients enrolled in these studies, low event rates for carefully managed study patients, and selection bias. Therefore, in this study we set out to study a large-scale claims database retrospectively to compare effects of ablation and AADs treatment in a matched cohort of patients with AF. We found that patients managed using AADs were hospitalized 41% more for a thromboembolic cause and 13% more for a cardiovascular cause than patients treated with catheter ablation. There was a nonsignificant trend of increased bleeding risk of 13% for the AAD group over the observed follow-up period. Those undergoing ablation experienced an elevated risk of bleeding in the first few months after ablation, but risk was lower in the ablation group than in the AAD group beyond 6 months until the end of follow-up. This nonproportional hazard may be part of the reason for failure to reach of statistical significance. Complications of the ablation procedure were found to be infrequent (<3%) in real-world experience, whereas multiple conditions that could be attributable to or exacerbated by amiodarone were present in a sizeable fraction of patients. The multiway comparison found that patients receiving DOACs with catheter ablation had the lowest incidence of thromboembolic events leading to ischemic stroke, TIA, or systemic embolism, whereas use of AADs and warfarin was associated with higher incidence. It may not be surprising to report a lower event rate in patients anticoagulated with DOACs compared with warfarin,24Ruff C.T. Giugliano R.P. Braunwald E. Hoffman E.B. Deenadayalu N. Ezekowitz M.D. Camm A.J. Weitz J.I. Lewis B.S. Parkhomenko A. Yamashita T. Antman E.M. Comparison of the efficacy and safety of new oral anticoagulants with warfarin in patients with atrial fibrillation: a meta-analysis of randomised trials.Lancet. 2014; 383: 955-962Abstract Full Text Full Text PDF PubMed Scopus (3315) Google Scholar but also suggests that benefit of catheter ablation may be additive to that of DOACs as a method of anticoagulation. These findings also underscore the potential adverse consequences of polypharmacy in typically elderly AF patients on warfarin that results in altered efficacy and safety and have an impact on morbidity and mortality.25Focks J.J. Brouwer M.A. Wojdyla D.M. Thomas L. Lopes R.D. Washam J.B. Lanas F. Xavier D. Husted S. Wallentin L. Alexander J.H. Granger C.B. Verheugt F.W.A. Polypharmacy and effects of apixaban versus warfarin in patients with atrial fibrillation: post hoc analysis of the ARISTOTLE trial.BMJ. 2016; 353: i2868Crossref PubMed Scopus (71) Google Scholar, 26Ionescu-Ittu R. Abrahamowicz M. Jackevicius C.A. Essebag V. Comparative effectiveness of rhythm control vs rate control drug treatment effect on mortality in patients with atrial fibrillation.Arch Intern Med. 2012; 172: 997-1004Crossref PubMed Scopus (115) Google Scholar Ablation of AF may be associated with overall beneficial cardiovascular effects as underscored by lower risk of cardiovascular hospitalization in the ablation group. However, risk of hospitalization with a primary diagnosis code for cardiovascular disease was not significantly different. It is possible that a portion of the hospitalizations associated with AADs may have been partially attributable to side effects, and thus not coded with a primary diagnosis of cardiovascular disease. On the other hand, side effects of ablation were rare. The most common type of event was vascular complication or hematoma, which occurred in only 3% of patients undergoing ablation. Although this analysis was not able to attribute causality for a side effect, it was found that 25% of patients who were on AADs also had thyroid dysfunction. Even if AADs were not the cause, the cohort of patients with AF may have numerous co-morbidities, and ablation may offer a reduced risk profile compared with AADs in many patients. This result concurs with retrospective observational study conducted using medical claims data from 2005 to 2009; Reynolds and colleagues showed that catheter ablation reduced the risk of stroke or TIA by about 38%.27Reynolds M.R. Gunnarsson C.L. Hunter T.D. Ladapo J.A. March J.L. Zhang M. Hao S.C. Health outcomes with catheter ablation or antiarrhythmic drug therapy in atrial fibrillation.Circ Cardiovasc Qual Outcomes. 2012; 5: 171-181Crossref PubMed Scopus (89) Google Scholar The present work found a similar reduction in incidence of thromboembolic events, but included an order of magnitude greater sample size from 2010 to 2014 than the 801 matched pairs of patients treated by ablation or AADs in 2005 to 2009.27Reynolds M.R. Gunnarsson C.L. Hunter T.D. Ladapo J.A. March J.L. Zhang M. Hao S.C. Health outcomes with catheter ablation or antiarrhythmic drug therapy in atrial fibrillation.Circ Cardiovasc Qual Outcomes. 2012; 5: 171-181Crossref PubMed Scopus (89) Google Scholar Moreover, the present study also captures the effect of introduction of DOACs in the years since 2009. Finally, the present analysis also considered outcomes of bleeding, hospitalization for cardiovascular disease, and prevalence of known side effects of AADs and ablation. This was a retrospective, observational study that may not be able to account for all possible confounding factors. Patients were matched in the AAD and the ablation groups on baseline demographics, but there may have been other reasons for selecting candidates for ablation or use of AADs. Reliance on claims for hospitalization with a primary diagnosis code of stroke or TIA as a clinical outcome undercounts the true incidence of the event, because in many cases there could be an outpatient visit, emergency department visit that is treated without hospital admission, or even an event undiagnosed at the time of occurrence. However, using primary or secondary diagnosis codes has the potential to overstate the likelihood of the event, because previous occurrences in the past may be reported. In the present study, we required a primary diagnosis code for specific event outcomes but also presented diagnosis of cardiovascular disease in any diagnosis code. The cohort treated with ablation remains less than one-tenth the size treated with AADs, despite ablation performing better in similar patients who were only treated with AADs. This suggests that a durable device therapy may be preferable in many patients than one that requires continual adherence to medical therapy. Lasting benefit without long-term exposure to side effects of medications also warrants a consideration for catheter ablation, independent of results from the AFFIRM trial.8Wyse D.G. Waldo A.L. DiMarco J.P. Domanski M.J. Rosenberg Y. Schron E.B. Kellen J.C. Greene H.L. Mickel M.C. Dalquist J.E. Corley S.D. A comparison of rate control and rhythm control in patients with atrial fibrillation.N Engl J Med. 2002; 347: 1825-1833Crossref PubMed Scopus (3672) Google Scholar Indeed, our analysis observed a decrease in medication usage after ablation, with 77% of patients prescribed AADs before ablation compared with 45% thereafter. Long-term prospective randomized trials such as Catheter Ablation vs Anti-arrhythmic Drug Therapy for Atrial Fibrillation Trial and the Early Treatment of Atrial Fibrillation for Stroke Prevention Trial will ultimately define the role of catheter ablation compared with management by medication alone.16Cleland J.G.F. Coletta A.P. Buga L. Ahmed D. Clark A.L. Clinical trials update from the American College of Cardiology meeting 2010: DOSE, ASPIRE, CONNECT, STICH, STOP-AF, CABANA, RACE II, EVEREST II, ACCORD, and NAVIGATOR.Eur J Heart Fail. 2010; 12: 623-629Crossref PubMed Scopus (61) Google Scholar, 28Aliot E. Brandes A. Eckardt L. Elvan A. Gulizia M. Heidbuchel H. Kautzner J. Mont L. Morgan J. Ng A. Szumowski L. Themistoclakis S. Van Gelder I.C. Willems S. Kirchhof P. The EAST study: redefining the role of rhythmcontrol therapy in atrial fibrillation: EAST, the Early treatment of Atrial fibrillation for Stroke prevention Trial.Eur Heart J. 2015; 36: 255-256Crossref PubMed Scopus (26) Google Scholar M. Mansour has received compensation for services from Abbott, Boston Scientific, Biotronik, Medtronic; has equity interests/stock options in NewPace Ltd.; has research grants from Boehringer Ingelheim, Pfizer, Abbott, Biosense Webster, Boston Scientific. E.K. Heist has received compensation for services from Boston Scientific, Abbott, Biotronik, Pfizer; has research grants from Boston Scientific, Abbott. R. Agarwal, E. Karst and S. Mahapatra receive salary from Abbott. T.J. Bunch has no disclosures. J.N. Ruskin has received compensation for services from Advanced Medical Education, Cardiome, Daiichi Sankyo, Gilead Sciences, InCarda Therapeutics, Laguna Medical, Medtronic, Pfizer; has stock options/equity interests in NewPace Ltd, InfoBionic and Portola Pharmaceuticals. Abbott provided funding for this study. Funding: Abbott provided funding for this study. The following is the supplementary data to this article: Download .docx (.01 MB) Help with docx files Appendix S1Supplementary materials.