Risk predictors of stroke and mortality after ablation for atrial fibrillation: The California experience 2005–2009 Uma N. Srivatsa, MBBS, MAS,* Beate Danielsen, PhD,† Ivan Anderson, MD,* Ezra Amsterdam, MD,* Nayereh Pezeshkian, MD, MPH,* Yingbo Yang, MD, PhD,* Richard H. White, MD† From the *Division of Cardiovascular Medicine, and †Department of Internal Medicine, UC Davis School of Medicine, Sacramento, California. BACKGROUND Ablation (ABL) is a second-line therapy for the management of atrial fibrillation (AF). Single-center studies have demonstrated the safety and efficacy of ABL; however, the low event rates render it difficult to assess predictors of major adverse events. OBJECTIVE The purpose of this study was to determine the population-based incidence of data and risk factors for both stroke o30 days and death after AF ABL. METHODS Patients (n ¼ 6207) identified as having undergone AF ABL between 2005 and 2009 by the California Ambulatory Surgery Database in 97 nonfederal hospitals in California were linked to the California Patient Discharge Database and to a master death registry. Data from these patients were analyzed for primary outcomes of 30-day death and ischemic stroke. Independent risk factors for these end-points were determined. RESULTS Mean patient age was 61.9 years, and the majority of the patients were men. Thirty-day mortality and stroke after ABL were 0.39% and 0.61%, respectively. Independent predictors of death were age Z80 years (odds ratio [OR] 8.2, 95% confidence interval [CI] 1.5–43) and heart failure (OR 9.2, 95% CI 3.0–28). Prior
Introduction Atrial fibrillation (AF) ablation (ABL) is recommended for symptomatic patients who have failed or are intolerant of antiarrhythmic drug therapy.1 Maintenance of sinus rhythm after ABL varies from 50% to 80%2,3 and is dependent on the type and duration of AF, as well as the method of detection.4 Because of its efficacy in patients who are refractory to medical therapy, ABL is an important therapeutic option in this group. However, it is a complex procedure usually performed in specialized centers, and it is associated with a wide variety of minor and major complications.4–8 In a worldwide survey of individual institutions, the frequency of procedure-related mortality and ischemic
Address reprint requests and correspondence: Dr. Uma Srivatsa, Division of Cardiovascular Medicine, Department of Internal Medicine, UC Davis School of Medicine, 4860 Y St, Suite 2820, Sacramento, CA 95817. E-mail address: [email protected].
1547-5271/$-see front matter B 2014 Heart Rhythm Society. All rights reserved.
stroke/transient ischemic attack/stroke was the only independent predictor for stroke (OR 6.3, 95% CI 3–13). CONCLUSION In our large population-based study, we found higher procedure-related mortality but comparable stroke rate after AF ABL than previously reported. Age Z80 years and heart failure was each independently associated with 48-fold increase in odds of death. The only significant predictor of stroke was prior stroke/ transient ischemic attack. These findings may aid in patient selection for AF ABL. KEYWORDS Atrial fibrillation; Ablation; Stroke; Mortality ABBREVIATIONS ABL ¼ ablation; AF ¼ atrial fibrillation; CAD ¼ coronary artery disease; CI ¼ confidence interval; CVD ¼ cardiovascular disease; DM ¼ diabetes mellitus; HCUP ¼ Healthcare Cost and Utilization Project; HF ¼ heart failure; HTN ¼ hypertension; OR ¼ odds ratio; PVD ¼ peripheral vascular disease; TIA ¼ transient ischemic attack (Heart Rhythm 2014;11:1898–1903) I 2014 Heart Rhythm Society. All rights reserved.
stroke was 0.05% and 0.28%, respectively.8 Similar low numbers reported from single-center studies make it difficult to identify risk factors for these events, and large populationbased studies that can provide these data are lacking. Therefore, we sought to determine the population-based incidence of periprocedural stroke and mortality in California residents who underwent ABLs for principal diagnosis of AF.
Methods California requires all nonfederal hospitals to submit diagnostic and discharge data for hospital-based inpatient and ambulatory procedures. Serial hospital, emergency department, and ambulatory surgery records for a specific individual can be temporally linked using an encrypted form of the social security number, called the record linkage number. These data can be linked to a master death registry. The strength of this comprehensive database is that events http://dx.doi.org/10.1016/j.hrthm.2014.07.017
Srivatsa et al
Atrial Fibrillation Ablation Mortality and Stroke
occurring in California hospitals other than the index hospital (where ABL was performed) can be identified. After obtaining institutional review board approval, we identified the records of all cases in California coded as having endovascular ABL (ICD-9-CM ¼ 37.34) coupled with a principal diagnosis of AF (427.31) performed between January 1, 2005, and December 31, 2009. We excluded all cases with the principal diagnosis of atrial flutter (427.32), supraventricular tachycardia (427.0), ventricular tachycardia (427.1), open surgical procedures (only procedure ¼ 37.34), and those who had pacemaker implant (37.80-37.87) plus sinoatrial node dysfunction (427.81) or AV block 426.0426.1). The primary outcomes were ischemic stroke or transient ischemic attack (TIA) (433.x1, 434.x1, 435, and 997.02 coupled with a stroke code) and mortality within 30 days of the ABL procedure. We performed a sensitivity analysis by repeating the analysis after excluding all patients who had received an implantable defibrillator (37.94-37.97), a cardiac device (V45.00), a pacemaker (V45.01), or had a history of defibrillator insertion (V45.02) and retested primary outcomes. This allowed exclusion of cases that underwent AV nodal ABL for AF. Annual volume of hospital ABLs for AF was calculated for the entire study period. We only analyzed the first ABL in all records, so the number of ABLs represents the total number of patients who underwent ABL. Preexisting comorbidities were defined using the Elixhauser comorbidity software from the Healthcare Cost and Utilization Project V3.7 (HCUP), which is based on ICD-9CM codes listed as present on admission.9 The software defines 29 specific comorbidities, and the terms that were included in this analysis were congestive heart failure (HF), hypertension (HTN), diabetes (DM), coronary artery disease (CAD), and peripheral vascular disease (PVD). CAD and PVD were considered together as cardiovascular disease (CVD). These comorbidities make up the components of the CHA2DS2VASc score.10
1899 Table 1 Population characteristics of atrial fibrillation ablation patients from nonfederal hospitals in California from 2005 to 2009 (N ¼ 6207) Population characteristic Age (years) Gender (%) Male Female Race/ethnicity (%) Non–Hispanic White Black Hispanic Asian/Pacific Islander Other Comorbidity (%) Heart failure Hypertension Diabetes Stroke/transient ischemic attack Cardiovascular disease CHADS2VASC score
61.93 ⫾ 12.11 4275 (68.9) 1932 (31.1) 5110 (82.3) 104 (1.7) 395 (6.4) 285 (4.6) 313 (5.0) 1210 (19.5) 596 (9.6) 964 (15.5) 397 (6.4) 1730 (27.9) 1.74 ⫾ 1.76
Values are given as mean ⫾ SD or no. (%).
high, intermediate, or low frequency. P o .05 was considered significant.
Results During the study period, 6207 ABL procedures were performed for AF in 97 nonfederal hospitals in California. Baseline characteristics of the study population are given in Table 1. Average age was 460 years, more than two thirds of patients were male, and more than three-fourths were Caucasians. Mean CHA2DS2VASc score was 1.74 ⫾ 1.76.
Periprocedural mortality
Statistical analysis
Mortality within 30 days of ABL was 0.39% and was significantly higher for patients Z80 years (9/445 [2.02%]) compared to lower age groups: o60 years (2/2458 [0.08%], P ¼ .001); 60–69 years (10/2010 [0.35%], P ¼ .008); and 70–79 years (6/1180 [0.35%], P ¼ .03; Figure 1). Thirty-day mortality did not differ between men (15/4202 [0.36%]) and women (9/1891 [0.48%], P ¼ .49) but was more frequent
SAS (version 9.3, Cary, NC) was used for all statistical analyses. Continuous variables are expressed as mean ⫾ SD. Categorical variables are presented as percentages. Univariate analysis was performed with the χ2 test for nominal variables, and the Fisher exact test was applied for outcomes o5 events per cell. Analysis was performed using multivariable logistic regression models to predict outcomes. The variables chosen for adjustment in our first model were age, gender, CAD, PVD, CAD, HTN, DM, HF, prior stroke/TIA, and hospital volume. For a second model, we used the CHA2DS2VASc score and hospital volume. The CHA2DS2VASc score includes HF [left ventricular ejection fraction r40 (1 point)], HTN (1 point), age [465 years (1 point); 475 years (2 points)], DM (1 point), prior TIA or stroke (2 points), CVD (1 point), and female gender (1 point).14 Hospital volume of ABL procedures was categorized as
Figure 1 Thirty-day death and 30-day stroke by age in years. Atrial fibrillation ablation patients, California, 2005–2009. Light gray bars and dark gray bars represent stroke and death, respectively, reported as percentage.
1900
Heart Rhythm, Vol 11, No 11, November 2014
Figure 2 Thirty-day death and 30-day stroke by CHADS2VASC score. Atrial fibrillation ablation patients, California, 2005–2009. Light gray bars and dark gray bars represent stroke and death, respectively, reported as percentage.
among patients with HF (19/1190 [1.6%]) than in those without (5/4903 [0.10%], P o .0001). In the multivariable analysis, the only predictors of 30-day mortality were age Z80 years (odds ratio [OR] 8.2, 95% confidence interval [CI] 1.54–43.3) and HF (OR 9.2, 95% CI ¼ 3.04 – 27.2). Thirty-day mortality in patients with a CHA2DS2VASc score of 0 or 1 was low (0.09%). Mortality increased stepwise with increasing score: 0.27% for score of 2, 1.05% for score of 3–5, and 1.19% for score Z6 (Figure 2 Table 2
and Table 2). This relationship between CHA2DS2VASc score and death mainly reflected the effect of age and HF, because sex, HTN, DM, prior stroke/TIA, and CVD were not significant predictors. Of the 97 hospitals, 37 performed Z50 procedures, 21 performed Z100, and 2 hospitals performed 4600 procedures per year during the study period; the rest performed o50 procedures per year. Mean number of procedures was 65 per year; 27.6% of the hospitals performed Z150 per year (high), 42.5% between 50 and 149 per year (intermediate), and 29.9% o50 per year (low). Although there was a trend toward higher 30-day mortality as hospital volume decreased (0.55% for r50 procedures per year, 0.39% for 50–150 per year, and 0.24% for hospitals performing 4150 per year), in a model that adjusted for the CHA2DS2VASc score, hospital volume was not a significant predictor of mortality (c-statistic 0.770 for this model).
Periprocedural stroke The 30-day incidence of stroke was 0.61% (Table 2). In the multivariable model, only a CHA2DS2VASc score Z6 was associated with a significantly higher risk of stroke (2%; OR 4.5, 95% CI 1.6–12.3, Figure 2). The only risk factor incorporated in the CHA2DS2VASc score that was a significant predictor in this cohort was prior stroke/TIA
Risk of death or stroke after ablation for atrial fibrillation 30-Day death
All atrial fibrillation ablation Age (years) 18–59 60–69 70–79 80þ Sex Female Male Race Non–Hispanic White Hispanic Other Heart failure Hypertension Diabetes Stroke/transient ischemic attack Cardiovascular disease Annual ablation volume 1–49 50–149 150þ CHADS2VASC score 0–1 2 3–5 6þ * Significant at the .05 confidence level. †Odds ratio (OR) adjusted for confounding characteristics.
30-Day stroke
%
OR†
%
OR†
0.39
NA
0.61
NA
0.08 0.35 0.51 2.02
Reference 3.53 3.46 8.16*
0.48 0.44 0.92 1.34
Reference 0.92 1.88 2.54
0.48 0.36
0.62 Reference
0.78 0.54
1.04 Reference
0.40 0.51 0.29 1.60 1.35 0.63 0.78 0.93
Reference 1.37 0.95 9.19* 1.27 0.73 0.91 1.62
0.63 0.51 0.57 0.99 0.50 0.52 2.77 0.52
Reference 0.89 0.87 1.91 0.47 0.72 6.28* 0.47
0.55 0.39 0.24
1.05 Reference 0.86
0.49 0.68 0.64
0.62 Reference 0.95
0.09 0.27 1.05 1.19
Reference 3.05 11.21 12.46
0.47 0.54 0.76 1.98
Reference 1.13 1.66 *
Srivatsa et al
Atrial Fibrillation Ablation Mortality and Stroke
(2.8% vs 0.5%, P o .0001, OR 6.3, CI 3.0–13.3). Hospital volume was not a predictor of 30-day incidence of ischemic stroke. Patients who had a stroke had a higher risk of 30-day mortality compared to those who did not have stroke (7.4% vs 0.12%, P ¼ .001). After excluding patients who had received any cardiac implantable device (n ¼ 984), HF remained the only significant multivariate predictor of mortality (3% vs 0.11%, P ¼ .006). For the outcome of TIA/stroke, patients with a prior history of TIA/stroke remained at higher risk (1.9% vs 0.47%, P .005).
Periprocedural complications Complications during the index admission for ABL procedure were assessed. Hemorrhage occurred in 2.08% of patients, hematoma in 3.66%, and cardiac tamponade in 6.67%. After adjustment for the presence of chronic comorbidities, both development of a hematoma and cardiac tamponade were associated with higher risk for stroke: 3.1% vs 0.48% (P ¼ .001) and 8% vs 0.51%, p o 0.0001, respectively, in those with and without complications. We analyzed the rates of periprocedural complications based on hospital volume for hematoma (high: 1.8%, intermediate: 1.4%, low: 1.9%) and for cardiac tamponade (high: 0.35%, intermediate: 0.45%, low: 0.59%) with no significant differences overall, although there was a slight trend toward a higher rate of tamponade in low-volume hospitals.
Discussion Atrial ABL for AF has been widely adopted as a second-line therapy for AF control. Pulmonary vein isolation, circumferential ABL, as well as elimination of complex fragmentation, ganglionic plexuses and rotors, all have been performed for treatment of AF, with varying success rates.11–15 Current data suggest that ABL may be more effective for AF control compared to antiarrhythmic drug therapy.16,17 In this regard, some investigators have reported that AF ABL is a potent first-line therapy.17 However, the benefits of ABL in reducing mortality and stroke have yet to be established in a prospective study and are under investigation in the ongoing Catheter Ablation Versus Antiarrhythmic Drug Therapy for Atrial Fibrillation (CABANA) trial.18,19 In this setting, assessment of the incidence and predictors of mortality and stroke risk after this elective procedure is important to aid patient selection. Our study has a large number of patients because it is a population study from 97 nonfederal hospitals in California. Our analysis used linked ambulatory surgery data, hospital discharge data, and a master death registry. Because there is no specific ICD-9-CM code for AF ABL,20 we identified the records of all cases in California coded as having endovascular ABL (ICD-9-CM ¼ 37.34) coupled with a principal diagnosis of AF (427.31) performed between January 1, 2005, and December 31, 2009. Because the patients (97%) were hospitalized, we could not use the CPT code to
1901 determine supraventricular tachycardia ABL. All cases with a principal diagnosis of atrial flutter were excluded to avoid including less complex atrial flutter ABL patients. We also excluded patients who had pacemaker implantation in order to eliminate those who underwent AV node ABL. Primary outcomes were ischemic stroke (433.x1, 434.x1, and 997.02) and mortality, both within 30 days of ABL. Validity of coding of stroke and comorbidities in the setting of AF has been established in several studies. In a cross-sectional study comparing ICD 9-CM coding to medical record data in patients with AF, the codes were found to have high specificity and positive predictive value for the 5 common risk factors of CAD, stroke/TIA, HF, DM, and HTN, which are part of the CHADS2VASC score.21 In another study, Quan et al22 assessed the validity of 32 clinical conditions based on ICD 9-CM and ICD-10 codes. After comparing them to the medical charts, they found these codes to be highly specific. The diagnosis of stroke based on the ICD-9 code has also been found to have a high positive predictive value for actual events in the Medicaid population. Periprocedural mortality and stroke after AF ABL have been reported to be uncommon: 0.05% and 0.28%, respectively, in a worldwide survey.8 However, the results could be biased because it is dependent on recall. Also, most of published data are from single-center studies; large multicenter data are not available pending results of the CABANA trial. Although event rates from the Medicare database have been published, they are not generalizable to a wider population.23 In the absence of prospective data showing that ABL improves long-term survival or that it prevents stroke, knowledge of short-term risk of morbidity and mortality associated with ABL is critically important. Clarification of the role of specific risk factors for death and stroke might modify patient selection for ABL in order to minimize the risk of these procedure-related adverse outcomes. In this large population-based study, we found the 30-day mortality rate associated with ABL was 0.39%, which is considerably higher than the 0.05% reported in a worldwide survey.8 Advancing age, particularly Z80 years, was a strong predictor of mortality within 30 days of the procedure. Approximately 40% of our patients were o60 years, and in this group only 5 of 4902 patients (0.08%) died within 30 days. Thereafter, mortality increased in a stepwise fashion at a rate of 2.02% in patients Z80 years. In the Medicare population, a similar trend showing higher mortality among octogenarians (1.8%) compared to those between 60 and 65 years (0.5%) has been reported.23 In our study, the female gender was adequately represented, constituting 30% of patients, very similar to the study by Shah et al. Although female gender has been associated with a higher complication rate and tamponade, we did not find an association between gender and mortality.24,25 We had a substantial number of patients (19.5%) with HF. Because we do not have data on ejection fraction, this represents both systolic and diastolic HF. Ours was the first study to show that HF was a strong predictor of mortality (1.6% vs 0.1%,
1902 P o .0001) after adjustment of risk factors and elimination of those who had any implantable device. The 30-day incidence of stroke in our study was 0.61%, which again was higher than the reported incidence of 0.28% in a worldwide survey but comparable to the 0.8% reported among Medicare recipients. We found that patients who had stroke at 30 days had higher mortality. Stroke was more frequent in women (1.04%) than in men (0.56%), but after adjusting for the previously noted risk factors, particularly age, female gender was not an independent predictor. Unlike the Medicare data,23 we found that the risk of stroke increased with age, although this finding was not statistically significant. Stroke may be related to recurrent AF, procedural/clinical characteristics, and inadequate anticoagulation in the early post-ABL period. Other reports suggest that the incidence of periprocedural stroke may be related to the type and intensity of anticoagulation.26,27 However, our database does not include this information. We used the CHA2DS2VASc score to predict the risk of stroke after ABL for AF. This score is reported to be a strong predictor of thromboembolic risk in patients with AF, and it has the highest negative predictive value for stroke among published risk stratification schemes.14 However, there are no prior data to validate the utility of the CHA2DS2VASc score for predicting risk of stroke after AF ABL. In our study cohort, the stroke rate was 4 times higher if the CHA2DS2VASc score was Z6 (1.98%) compared to a score r1 (0.47%, P ¼ .004). When we analyzed the components of the score, this was primarily driven by the presence of prior stroke/TIA, as there was a much higher risk among those with a prior stroke than in those without a prior history of stroke (2.8 % vs 0.5%, P o .0001), and this remained significant after elimination of those with devices. Hospital volume of ABL was highly skewed, with the 9 hospitals performing 4150 ABLs per year, accounting for 46% of all procedures. Twenty-eight hospitals performed 50–149 procedures per year, which accounted for 42% of all procedures. Sixty hospitals performed the remaining procedures, which accounted for only 11% of all procedures. Interestingly, there was no relationship between hospital volume and either 30-day mortality or 30-day incidence of stroke after controlling for CHA2DS2VASc score, unlike other studies.23,25 We did not find significant difference among hospitals based on caseload volume for either periprocedural hematoma or cardiac tamponade. This finding suggests that procedure- or operator-specific factors play a minor role in the occurrence of adverse outcomes compared to endogenous risk factors. The lack of difference could also be due to uniform patient selection and experienced operators working at different hospitals. In the study by Shah et al,25 which used the California patient discharge database and some methodologies similar to the present study, the 30-day mortality and stroke rates were 2.3% and 4.9%, respectively. They found an association of age 455 years, female gender, and procedure experience to correlate with inpatient complications and/or 30-day rehospitalizations. The current study was larger, and
Heart Rhythm, Vol 11, No 11, November 2014 we were able to identify risk factors specifically for 30-day stroke and mortality as well the effect of CHA2DS2VASc score on outcomes. In our study, the odds of 30-day stroke/ TIA risk was increased 6-fold if there was a periprocedural hematoma, and the risk was increased 21-fold if there was cardiac tamponade. These findings suggest that the risk of stroke was strongly related to the likely risk of having anticoagulation therapy withheld; however, we do not have the details of therapy. In the study by Deshmukh et al,24 a large number of patients who underwent AF ABLs between 2000 and 2010 were assessed for complications. This is a representative survey of HCUP, and the weights provided by the National Inpatient Sample were used to generate estimates of the number of admissions. They found an association of age, gender, and operator volume with complication rates. Hospital volumes between 50 and 100 per year was associated with fewer complications. However, although the number of ABLs increased steadily over time, there was no significant change in the yearly incidence of mortality or stroke after ABL. This study was based on a sample of 20% U.S. community hospitals. A minor variation in the estimate has the risk of amplifying the results. In our population-based study, although the total number of patients is less than this study, the chances of variability in the results is less likely because we accounted for every event in 30-day follow-up.
Study limitations Although this study reflects outcomes in the entire population of patients undergoing AF ABL in the nonfederal hospitals in the state of California, we did not have information regarding the specific manner in which ABL was achieved, duration of ABL, use of specific drugs, adherence to anticoagulation, or actual severity of comorbidities. These factors may play an important role in determining the incidence of adverse outcomes associated with ABL. Our data do not include patients who had events in hospitals outside California and thus could underestimate outcomes.
Conclusion To our knowledge, this is the first large population-based study to assess the risk of major adverse outcomes after AF ABL. The 30-day stroke rate was higher with CHADS2VASC score Z6, primarily reflecting the importance of prior stroke/TIA. Periprocedural hematoma and cardiac tamponade were associated with higher risk of 30-day stroke. The 30-day mortality was significantly higher for patients Z80 years and those with HF. Older individuals, patients with a prior history of stroke/TIA, and those with HF appear to be at greatest risk for major adverse outcomes after ABL for AF.
References 1. Calkins H. Catheter ablation to maintain sinus rhythm. Circulation 2012;125: 1439–1445.
Srivatsa et al
Atrial Fibrillation Ablation Mortality and Stroke
2. Jais P, Cauchemez B, Macle L, Daoud E, Khairy P, Subbiah R, Hocini M, Extramiana F, Sacher F, Bordachar P, Klein G, Weerasooriya R, Clementy J, Haissaguerre M. Catheter ablation versus antiarrhythmic drugs for atrial fibrillation: the A4 study. Circulation 2008;118:2498–2505. 3. Oral H, Veerareddy S, Good E, Hall B, Cheung P, Tamirisa K, Han J, Fortino J, Chugh A, Bogun F, Pelosi F Jr, Morady F. Prevalence of asymptomatic recurrences of atrial fibrillation after successful radiofrequency catheter ablation. J Cardiovasc Electrophysiol 2004;15:920–924. 4. Kuwahara T, Takahashi A, Kobori A, Miyazaki S, Takahashi Y, Takei A, Nozato T, Hikita H, Sato A, Aonuma K. Safe and effective ablation of atrial fibrillation: importance of esophageal temperature monitoring to avoid periesophageal nerve injury as a complication of pulmonary vein isolation. J Cardiovasc Electrophysiol 2009;20:1–6. 5. Pappone C, Rosanio S, Augello G, Gallus G, Vicedomini G, Mazzone P, Gulletta S, Gugliotta F, Pappone A, Santinelli V, Tortoriello V, Sala S, Zangrillo A, Crescenzi G, Benussi S, Alfieri O. Mortality, morbidity, and quality of life after circumferential pulmonary vein ablation for atrial fibrillation: outcomes from a controlled nonrandomized long-term study. J Am Coll Cardiol 2003;42:185–197. 6. Baman TS, Jongnarangsin K, Chugh A, Suwanagool A, Guiot A, Madenci A, Walsh S, Ilg KJ, Gupta SK, Latchamsetty R, Bagwe S, Myles JD, Crawford T, Good E, Bogun F, Pelosi F Jr, Morady F, Oral H. Prevalence and predictors of complications of radiofrequency catheter ablation for atrial fibrillation. J Cardiovasc Electrophysiol 2011;22:626–631. 7. Pappone C, Oral H, Santinelli V, Vicedomini G, Lang CC, Manguso F, Torracca L, Benussi S, Alfieri O, Hong R, Lau W, Hirata K, Shikuma N, Hall B, Morady F. Atrio-esophageal fistula as a complication of percutaneous transcatheter ablation of atrial fibrillation. Circulation 2004;109:2724–2726. 8. Cappato R, Calkins H, Chen SA, Davies W, Iesaka Y, Kalman J, Kim YH, Klein G, Packer D, Skanes A. Worldwide survey on the methods, efficacy, and safety of catheter ablation for human atrial fibrillation. Circulation 2005;111:1100–1105. 9. Steiner C, Elixhauser A, Schnaier J. The healthcare cost and utilization project: an overview. Eff Clin Pract 2002;5:143–151. 10. Lip GY, Nieuwlaat R, Pisters R, Lane DA, Crijns HJ. Refining clinical risk stratification for predicting stroke and thromboembolism in atrial fibrillation using a novel risk factor-based approach: the Euro Heart Survey on Atrial Fibrillation. Chest 2010;137:263–272. 11. Oral H, Chugh A, Yoshida K, Sarrazin JF, Kuhne M, Crawford T, Chalfoun N, Wells D, Boonyapisit W, Veerareddy S, Billakanty S, Wong WS, Good E, Jongnarangsin K, Pelosi F Jr, Bogun F, Morady F. A randomized assessment of the incremental role of ablation of complex fractionated atrial electrograms after antral pulmonary vein isolation for long-lasting persistent atrial fibrillation. J Am Coll Cardiol 2009;53:782–789. 12. Haissaguerre M, Jais P, Shah DC, Takahashi A, Hocini M, Quiniou G, Garrigue S, Le Mouroux A, Le Metayer P, Clementy J. Spontaneous initiation of atrial fibrillation by ectopic beats originating in the pulmonary veins. N Engl J Med 1998;339:659–666. 13. Haissaguerre M, Sanders P, Hocini M, Jais P, Clementy J. Pulmonary veins in the substrate for atrial fibrillation: the “venous wave” hypothesis. J Am Coll Cardiol 2004;43:2290–2292. 14. Nademanee K, McKenzie J, Kosar E, Schwab M, Sunsaneewitayakul B, Vasavakul T, Khunnawat C, Ngarmukos T. A new approach for catheter ablation of atrial fibrillation: mapping of the electrophysiologic substrate. J Am Coll Cardiol 2004;43:2044–2053. 15. Narayan SM, Krummen DE, Clopton P, Shivkumar K, Miller JM. Direct or coincidental elimination of stable rotors or focal sources may explain successful atrial fibrillation ablation: on-treatment analysis of the confirm trial (conventional
1903
16.
17.
18. 19.
20.
21.
22.
23.
24.
25.
26.
27.
ablation for af with or without focal impulse and rotor modulation). J Am Coll Cardiol 2013;62:138–147. Wilber DJ, Pappone C, Neuzil P, De Paola A, Marchlinski F, Natale A, Macle L, Daoud EG, Calkins H, Hall B, Reddy V, Augello G, Reynolds MR, Vinekar C, Liu CY, Berry SM, Berry DA. ThermoCool AFTI. Comparison of antiarrhythmic drug therapy and radiofrequency catheter ablation in patients with paroxysmal atrial fibrillation: a randomized controlled trial. JAMA 2010;303: 333–340. Cosedis Nielsen J, Johannessen A, Raatikainen P, Hindricks G, Walfridsson H, Kongstad O, Pehrson S, Englund A, Hartikainen J, Mortensen LS, Hansen PS. Radiofrequency ablation as initial therapy in paroxysmal atrial fibrillation. N Engl J Med 2012;367:1587–1595. Schmidt B, Bordignon S, Furnkranz A, Chun KR. Catheter ablation of atrial fibrillation to reduce stroke risk. Herz 2013;38:247–250. Cleland JG, Coletta AP, Buga L, Ahmed D, Clark AL. 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;12:623–629. Birman-Deych E, Waterman AD, Yan Y, Nilasena DS, Radford MJ, Gage BF. Accuracy of ICD-9-CM codes for identifying cardiovascular and stroke risk factors. Med Care 2005;43:480–485. Roumie CL, Mitchel E, Gideon PS, Varas-Lorenzo C, Castellsague J, Griffin MR. Validation of ICD-9 codes with a high positive predictive value for incident strokes resulting in hospitalization using Medicaid health data. Pharmacoepidemiol Drug Saf 2008;17:20–26. Quan H, Li B, Saunders LD, Parsons GA, Nilsson CI, Alibhai A, Ghali WA. IMECCHI Investigators. Assessing validity of ICD-9-CM and ICD-10 administrative data in recording clinical conditions in a unique dually coded database. Health Serv Res 2008;43:1424–1441. Piccini JP, Sinner MF, Greiner MA, Hammill BG, Fontes JD, Daubert JP, Ellinor PT, Hernandez AF, Walkey AJ, Heckbert SR, Benjamin EJ, Curtis LH. Outcomes of Medicare beneficiaries undergoing catheter ablation for atrial fibrillation. Circulation 2012;126:2200–2207. Deshmukh A, Patel NJ, Pant S, Shah N, Chothani A, Mehta K, Grover P, Singh V, Vallurupalli S, Savani GT, Badheka A, Tuliani T, Dabhadkar K, Dibu G, Reddy YM, Sewani A, Kowalski M, Mitrani R, Paydak H, Viles-Gonzalez JF. Inhospital complications associated with catheter ablation of atrial fibrillation in the United States between 2000 and 2010: analysis of 93 801 procedures. Circulation 2013;128:2104–2112. Shah RU, Freeman JV, Shilane D, Wang PJ, Go AS, Hlatky MA. Procedural complications, rehospitalizations, and repeat procedures after catheter ablation for atrial fibrillation. J Am Coll Cardiol 2012;59:143–149. Wazni OM, Beheiry S, Fahmy T, Barrett C, Hao S, Patel D, Di Biase L, Martin DO, Kanj M, Arruda M, Cummings J, Schweikert R, Saliba W, Natale A. Atrial fibrillation ablation in patients with therapeutic international normalized ratio: comparison of strategies of anticoagulation management in the periprocedural period. Circulation 2007;116:2531–2534. Lakkireddy D, Reddy YM, Di Biase L, Vanga SR, Santangeli P, Swarup V, Pimentel R, Mansour MC, D'Avila A, Sanchez JE, Burkhardt JD, Chalhoub F, Mohanty P, Coffey J, Shaik N, Monir G, Reddy VY, Ruskin J, Natale A. Feasibility and safety of dabigatran versus warfarin for periprocedural anticoagulation in patients undergoing radiofrequency ablation for atrial fibrillation: results from a multicenter prospective registry. J Am Coll Cardiol 2012;59: 1168–1174.
Introduction Atrial fibrillation (AF) ablation (ABL) is recommended for symptomatic patients who have failed or are intolerant of antiarrhythmic drug therapy.1 Maintenance of sinus rhythm after ABL varies from 50% to 80%2,3 and is dependent on the type and duration of AF, as well as the method of detection.4 Because of its efficacy in patients who are refractory to medical therapy, ABL is an important therapeutic option in this group. However, it is a complex procedure usually performed in specialized centers, and it is associated with a wide variety of minor and major complications.4–8 In a worldwide survey of individual institutions, the frequency of procedure-related mortality and ischemic
Address reprint requests and correspondence: Dr. Uma Srivatsa, Division of Cardiovascular Medicine, Department of Internal Medicine, UC Davis School of Medicine, 4860 Y St, Suite 2820, Sacramento, CA 95817. E-mail address: [email protected].
1547-5271/$-see front matter B 2014 Heart Rhythm Society. All rights reserved.
stroke/transient ischemic attack/stroke was the only independent predictor for stroke (OR 6.3, 95% CI 3–13). CONCLUSION In our large population-based study, we found higher procedure-related mortality but comparable stroke rate after AF ABL than previously reported. Age Z80 years and heart failure was each independently associated with 48-fold increase in odds of death. The only significant predictor of stroke was prior stroke/ transient ischemic attack. These findings may aid in patient selection for AF ABL. KEYWORDS Atrial fibrillation; Ablation; Stroke; Mortality ABBREVIATIONS ABL ¼ ablation; AF ¼ atrial fibrillation; CAD ¼ coronary artery disease; CI ¼ confidence interval; CVD ¼ cardiovascular disease; DM ¼ diabetes mellitus; HCUP ¼ Healthcare Cost and Utilization Project; HF ¼ heart failure; HTN ¼ hypertension; OR ¼ odds ratio; PVD ¼ peripheral vascular disease; TIA ¼ transient ischemic attack (Heart Rhythm 2014;11:1898–1903) I 2014 Heart Rhythm Society. All rights reserved.
stroke was 0.05% and 0.28%, respectively.8 Similar low numbers reported from single-center studies make it difficult to identify risk factors for these events, and large populationbased studies that can provide these data are lacking. Therefore, we sought to determine the population-based incidence of periprocedural stroke and mortality in California residents who underwent ABLs for principal diagnosis of AF.
Methods California requires all nonfederal hospitals to submit diagnostic and discharge data for hospital-based inpatient and ambulatory procedures. Serial hospital, emergency department, and ambulatory surgery records for a specific individual can be temporally linked using an encrypted form of the social security number, called the record linkage number. These data can be linked to a master death registry. The strength of this comprehensive database is that events http://dx.doi.org/10.1016/j.hrthm.2014.07.017
Srivatsa et al
Atrial Fibrillation Ablation Mortality and Stroke
occurring in California hospitals other than the index hospital (where ABL was performed) can be identified. After obtaining institutional review board approval, we identified the records of all cases in California coded as having endovascular ABL (ICD-9-CM ¼ 37.34) coupled with a principal diagnosis of AF (427.31) performed between January 1, 2005, and December 31, 2009. We excluded all cases with the principal diagnosis of atrial flutter (427.32), supraventricular tachycardia (427.0), ventricular tachycardia (427.1), open surgical procedures (only procedure ¼ 37.34), and those who had pacemaker implant (37.80-37.87) plus sinoatrial node dysfunction (427.81) or AV block 426.0426.1). The primary outcomes were ischemic stroke or transient ischemic attack (TIA) (433.x1, 434.x1, 435, and 997.02 coupled with a stroke code) and mortality within 30 days of the ABL procedure. We performed a sensitivity analysis by repeating the analysis after excluding all patients who had received an implantable defibrillator (37.94-37.97), a cardiac device (V45.00), a pacemaker (V45.01), or had a history of defibrillator insertion (V45.02) and retested primary outcomes. This allowed exclusion of cases that underwent AV nodal ABL for AF. Annual volume of hospital ABLs for AF was calculated for the entire study period. We only analyzed the first ABL in all records, so the number of ABLs represents the total number of patients who underwent ABL. Preexisting comorbidities were defined using the Elixhauser comorbidity software from the Healthcare Cost and Utilization Project V3.7 (HCUP), which is based on ICD-9CM codes listed as present on admission.9 The software defines 29 specific comorbidities, and the terms that were included in this analysis were congestive heart failure (HF), hypertension (HTN), diabetes (DM), coronary artery disease (CAD), and peripheral vascular disease (PVD). CAD and PVD were considered together as cardiovascular disease (CVD). These comorbidities make up the components of the CHA2DS2VASc score.10
1899 Table 1 Population characteristics of atrial fibrillation ablation patients from nonfederal hospitals in California from 2005 to 2009 (N ¼ 6207) Population characteristic Age (years) Gender (%) Male Female Race/ethnicity (%) Non–Hispanic White Black Hispanic Asian/Pacific Islander Other Comorbidity (%) Heart failure Hypertension Diabetes Stroke/transient ischemic attack Cardiovascular disease CHADS2VASC score
61.93 ⫾ 12.11 4275 (68.9) 1932 (31.1) 5110 (82.3) 104 (1.7) 395 (6.4) 285 (4.6) 313 (5.0) 1210 (19.5) 596 (9.6) 964 (15.5) 397 (6.4) 1730 (27.9) 1.74 ⫾ 1.76
Values are given as mean ⫾ SD or no. (%).
high, intermediate, or low frequency. P o .05 was considered significant.
Results During the study period, 6207 ABL procedures were performed for AF in 97 nonfederal hospitals in California. Baseline characteristics of the study population are given in Table 1. Average age was 460 years, more than two thirds of patients were male, and more than three-fourths were Caucasians. Mean CHA2DS2VASc score was 1.74 ⫾ 1.76.
Periprocedural mortality
Statistical analysis
Mortality within 30 days of ABL was 0.39% and was significantly higher for patients Z80 years (9/445 [2.02%]) compared to lower age groups: o60 years (2/2458 [0.08%], P ¼ .001); 60–69 years (10/2010 [0.35%], P ¼ .008); and 70–79 years (6/1180 [0.35%], P ¼ .03; Figure 1). Thirty-day mortality did not differ between men (15/4202 [0.36%]) and women (9/1891 [0.48%], P ¼ .49) but was more frequent
SAS (version 9.3, Cary, NC) was used for all statistical analyses. Continuous variables are expressed as mean ⫾ SD. Categorical variables are presented as percentages. Univariate analysis was performed with the χ2 test for nominal variables, and the Fisher exact test was applied for outcomes o5 events per cell. Analysis was performed using multivariable logistic regression models to predict outcomes. The variables chosen for adjustment in our first model were age, gender, CAD, PVD, CAD, HTN, DM, HF, prior stroke/TIA, and hospital volume. For a second model, we used the CHA2DS2VASc score and hospital volume. The CHA2DS2VASc score includes HF [left ventricular ejection fraction r40 (1 point)], HTN (1 point), age [465 years (1 point); 475 years (2 points)], DM (1 point), prior TIA or stroke (2 points), CVD (1 point), and female gender (1 point).14 Hospital volume of ABL procedures was categorized as
Figure 1 Thirty-day death and 30-day stroke by age in years. Atrial fibrillation ablation patients, California, 2005–2009. Light gray bars and dark gray bars represent stroke and death, respectively, reported as percentage.
1900
Heart Rhythm, Vol 11, No 11, November 2014
Figure 2 Thirty-day death and 30-day stroke by CHADS2VASC score. Atrial fibrillation ablation patients, California, 2005–2009. Light gray bars and dark gray bars represent stroke and death, respectively, reported as percentage.
among patients with HF (19/1190 [1.6%]) than in those without (5/4903 [0.10%], P o .0001). In the multivariable analysis, the only predictors of 30-day mortality were age Z80 years (odds ratio [OR] 8.2, 95% confidence interval [CI] 1.54–43.3) and HF (OR 9.2, 95% CI ¼ 3.04 – 27.2). Thirty-day mortality in patients with a CHA2DS2VASc score of 0 or 1 was low (0.09%). Mortality increased stepwise with increasing score: 0.27% for score of 2, 1.05% for score of 3–5, and 1.19% for score Z6 (Figure 2 Table 2
and Table 2). This relationship between CHA2DS2VASc score and death mainly reflected the effect of age and HF, because sex, HTN, DM, prior stroke/TIA, and CVD were not significant predictors. Of the 97 hospitals, 37 performed Z50 procedures, 21 performed Z100, and 2 hospitals performed 4600 procedures per year during the study period; the rest performed o50 procedures per year. Mean number of procedures was 65 per year; 27.6% of the hospitals performed Z150 per year (high), 42.5% between 50 and 149 per year (intermediate), and 29.9% o50 per year (low). Although there was a trend toward higher 30-day mortality as hospital volume decreased (0.55% for r50 procedures per year, 0.39% for 50–150 per year, and 0.24% for hospitals performing 4150 per year), in a model that adjusted for the CHA2DS2VASc score, hospital volume was not a significant predictor of mortality (c-statistic 0.770 for this model).
Periprocedural stroke The 30-day incidence of stroke was 0.61% (Table 2). In the multivariable model, only a CHA2DS2VASc score Z6 was associated with a significantly higher risk of stroke (2%; OR 4.5, 95% CI 1.6–12.3, Figure 2). The only risk factor incorporated in the CHA2DS2VASc score that was a significant predictor in this cohort was prior stroke/TIA
Risk of death or stroke after ablation for atrial fibrillation 30-Day death
All atrial fibrillation ablation Age (years) 18–59 60–69 70–79 80þ Sex Female Male Race Non–Hispanic White Hispanic Other Heart failure Hypertension Diabetes Stroke/transient ischemic attack Cardiovascular disease Annual ablation volume 1–49 50–149 150þ CHADS2VASC score 0–1 2 3–5 6þ * Significant at the .05 confidence level. †Odds ratio (OR) adjusted for confounding characteristics.
30-Day stroke
%
OR†
%
OR†
0.39
NA
0.61
NA
0.08 0.35 0.51 2.02
Reference 3.53 3.46 8.16*
0.48 0.44 0.92 1.34
Reference 0.92 1.88 2.54
0.48 0.36
0.62 Reference
0.78 0.54
1.04 Reference
0.40 0.51 0.29 1.60 1.35 0.63 0.78 0.93
Reference 1.37 0.95 9.19* 1.27 0.73 0.91 1.62
0.63 0.51 0.57 0.99 0.50 0.52 2.77 0.52
Reference 0.89 0.87 1.91 0.47 0.72 6.28* 0.47
0.55 0.39 0.24
1.05 Reference 0.86
0.49 0.68 0.64
0.62 Reference 0.95
0.09 0.27 1.05 1.19
Reference 3.05 11.21 12.46
0.47 0.54 0.76 1.98
Reference 1.13 1.66 *
Srivatsa et al
Atrial Fibrillation Ablation Mortality and Stroke
(2.8% vs 0.5%, P o .0001, OR 6.3, CI 3.0–13.3). Hospital volume was not a predictor of 30-day incidence of ischemic stroke. Patients who had a stroke had a higher risk of 30-day mortality compared to those who did not have stroke (7.4% vs 0.12%, P ¼ .001). After excluding patients who had received any cardiac implantable device (n ¼ 984), HF remained the only significant multivariate predictor of mortality (3% vs 0.11%, P ¼ .006). For the outcome of TIA/stroke, patients with a prior history of TIA/stroke remained at higher risk (1.9% vs 0.47%, P .005).
Periprocedural complications Complications during the index admission for ABL procedure were assessed. Hemorrhage occurred in 2.08% of patients, hematoma in 3.66%, and cardiac tamponade in 6.67%. After adjustment for the presence of chronic comorbidities, both development of a hematoma and cardiac tamponade were associated with higher risk for stroke: 3.1% vs 0.48% (P ¼ .001) and 8% vs 0.51%, p o 0.0001, respectively, in those with and without complications. We analyzed the rates of periprocedural complications based on hospital volume for hematoma (high: 1.8%, intermediate: 1.4%, low: 1.9%) and for cardiac tamponade (high: 0.35%, intermediate: 0.45%, low: 0.59%) with no significant differences overall, although there was a slight trend toward a higher rate of tamponade in low-volume hospitals.
Discussion Atrial ABL for AF has been widely adopted as a second-line therapy for AF control. Pulmonary vein isolation, circumferential ABL, as well as elimination of complex fragmentation, ganglionic plexuses and rotors, all have been performed for treatment of AF, with varying success rates.11–15 Current data suggest that ABL may be more effective for AF control compared to antiarrhythmic drug therapy.16,17 In this regard, some investigators have reported that AF ABL is a potent first-line therapy.17 However, the benefits of ABL in reducing mortality and stroke have yet to be established in a prospective study and are under investigation in the ongoing Catheter Ablation Versus Antiarrhythmic Drug Therapy for Atrial Fibrillation (CABANA) trial.18,19 In this setting, assessment of the incidence and predictors of mortality and stroke risk after this elective procedure is important to aid patient selection. Our study has a large number of patients because it is a population study from 97 nonfederal hospitals in California. Our analysis used linked ambulatory surgery data, hospital discharge data, and a master death registry. Because there is no specific ICD-9-CM code for AF ABL,20 we identified the records of all cases in California coded as having endovascular ABL (ICD-9-CM ¼ 37.34) coupled with a principal diagnosis of AF (427.31) performed between January 1, 2005, and December 31, 2009. Because the patients (97%) were hospitalized, we could not use the CPT code to
1901 determine supraventricular tachycardia ABL. All cases with a principal diagnosis of atrial flutter were excluded to avoid including less complex atrial flutter ABL patients. We also excluded patients who had pacemaker implantation in order to eliminate those who underwent AV node ABL. Primary outcomes were ischemic stroke (433.x1, 434.x1, and 997.02) and mortality, both within 30 days of ABL. Validity of coding of stroke and comorbidities in the setting of AF has been established in several studies. In a cross-sectional study comparing ICD 9-CM coding to medical record data in patients with AF, the codes were found to have high specificity and positive predictive value for the 5 common risk factors of CAD, stroke/TIA, HF, DM, and HTN, which are part of the CHADS2VASC score.21 In another study, Quan et al22 assessed the validity of 32 clinical conditions based on ICD 9-CM and ICD-10 codes. After comparing them to the medical charts, they found these codes to be highly specific. The diagnosis of stroke based on the ICD-9 code has also been found to have a high positive predictive value for actual events in the Medicaid population. Periprocedural mortality and stroke after AF ABL have been reported to be uncommon: 0.05% and 0.28%, respectively, in a worldwide survey.8 However, the results could be biased because it is dependent on recall. Also, most of published data are from single-center studies; large multicenter data are not available pending results of the CABANA trial. Although event rates from the Medicare database have been published, they are not generalizable to a wider population.23 In the absence of prospective data showing that ABL improves long-term survival or that it prevents stroke, knowledge of short-term risk of morbidity and mortality associated with ABL is critically important. Clarification of the role of specific risk factors for death and stroke might modify patient selection for ABL in order to minimize the risk of these procedure-related adverse outcomes. In this large population-based study, we found the 30-day mortality rate associated with ABL was 0.39%, which is considerably higher than the 0.05% reported in a worldwide survey.8 Advancing age, particularly Z80 years, was a strong predictor of mortality within 30 days of the procedure. Approximately 40% of our patients were o60 years, and in this group only 5 of 4902 patients (0.08%) died within 30 days. Thereafter, mortality increased in a stepwise fashion at a rate of 2.02% in patients Z80 years. In the Medicare population, a similar trend showing higher mortality among octogenarians (1.8%) compared to those between 60 and 65 years (0.5%) has been reported.23 In our study, the female gender was adequately represented, constituting 30% of patients, very similar to the study by Shah et al. Although female gender has been associated with a higher complication rate and tamponade, we did not find an association between gender and mortality.24,25 We had a substantial number of patients (19.5%) with HF. Because we do not have data on ejection fraction, this represents both systolic and diastolic HF. Ours was the first study to show that HF was a strong predictor of mortality (1.6% vs 0.1%,
1902 P o .0001) after adjustment of risk factors and elimination of those who had any implantable device. The 30-day incidence of stroke in our study was 0.61%, which again was higher than the reported incidence of 0.28% in a worldwide survey but comparable to the 0.8% reported among Medicare recipients. We found that patients who had stroke at 30 days had higher mortality. Stroke was more frequent in women (1.04%) than in men (0.56%), but after adjusting for the previously noted risk factors, particularly age, female gender was not an independent predictor. Unlike the Medicare data,23 we found that the risk of stroke increased with age, although this finding was not statistically significant. Stroke may be related to recurrent AF, procedural/clinical characteristics, and inadequate anticoagulation in the early post-ABL period. Other reports suggest that the incidence of periprocedural stroke may be related to the type and intensity of anticoagulation.26,27 However, our database does not include this information. We used the CHA2DS2VASc score to predict the risk of stroke after ABL for AF. This score is reported to be a strong predictor of thromboembolic risk in patients with AF, and it has the highest negative predictive value for stroke among published risk stratification schemes.14 However, there are no prior data to validate the utility of the CHA2DS2VASc score for predicting risk of stroke after AF ABL. In our study cohort, the stroke rate was 4 times higher if the CHA2DS2VASc score was Z6 (1.98%) compared to a score r1 (0.47%, P ¼ .004). When we analyzed the components of the score, this was primarily driven by the presence of prior stroke/TIA, as there was a much higher risk among those with a prior stroke than in those without a prior history of stroke (2.8 % vs 0.5%, P o .0001), and this remained significant after elimination of those with devices. Hospital volume of ABL was highly skewed, with the 9 hospitals performing 4150 ABLs per year, accounting for 46% of all procedures. Twenty-eight hospitals performed 50–149 procedures per year, which accounted for 42% of all procedures. Sixty hospitals performed the remaining procedures, which accounted for only 11% of all procedures. Interestingly, there was no relationship between hospital volume and either 30-day mortality or 30-day incidence of stroke after controlling for CHA2DS2VASc score, unlike other studies.23,25 We did not find significant difference among hospitals based on caseload volume for either periprocedural hematoma or cardiac tamponade. This finding suggests that procedure- or operator-specific factors play a minor role in the occurrence of adverse outcomes compared to endogenous risk factors. The lack of difference could also be due to uniform patient selection and experienced operators working at different hospitals. In the study by Shah et al,25 which used the California patient discharge database and some methodologies similar to the present study, the 30-day mortality and stroke rates were 2.3% and 4.9%, respectively. They found an association of age 455 years, female gender, and procedure experience to correlate with inpatient complications and/or 30-day rehospitalizations. The current study was larger, and
Heart Rhythm, Vol 11, No 11, November 2014 we were able to identify risk factors specifically for 30-day stroke and mortality as well the effect of CHA2DS2VASc score on outcomes. In our study, the odds of 30-day stroke/ TIA risk was increased 6-fold if there was a periprocedural hematoma, and the risk was increased 21-fold if there was cardiac tamponade. These findings suggest that the risk of stroke was strongly related to the likely risk of having anticoagulation therapy withheld; however, we do not have the details of therapy. In the study by Deshmukh et al,24 a large number of patients who underwent AF ABLs between 2000 and 2010 were assessed for complications. This is a representative survey of HCUP, and the weights provided by the National Inpatient Sample were used to generate estimates of the number of admissions. They found an association of age, gender, and operator volume with complication rates. Hospital volumes between 50 and 100 per year was associated with fewer complications. However, although the number of ABLs increased steadily over time, there was no significant change in the yearly incidence of mortality or stroke after ABL. This study was based on a sample of 20% U.S. community hospitals. A minor variation in the estimate has the risk of amplifying the results. In our population-based study, although the total number of patients is less than this study, the chances of variability in the results is less likely because we accounted for every event in 30-day follow-up.
Study limitations Although this study reflects outcomes in the entire population of patients undergoing AF ABL in the nonfederal hospitals in the state of California, we did not have information regarding the specific manner in which ABL was achieved, duration of ABL, use of specific drugs, adherence to anticoagulation, or actual severity of comorbidities. These factors may play an important role in determining the incidence of adverse outcomes associated with ABL. Our data do not include patients who had events in hospitals outside California and thus could underestimate outcomes.
Conclusion To our knowledge, this is the first large population-based study to assess the risk of major adverse outcomes after AF ABL. The 30-day stroke rate was higher with CHADS2VASC score Z6, primarily reflecting the importance of prior stroke/TIA. Periprocedural hematoma and cardiac tamponade were associated with higher risk of 30-day stroke. The 30-day mortality was significantly higher for patients Z80 years and those with HF. Older individuals, patients with a prior history of stroke/TIA, and those with HF appear to be at greatest risk for major adverse outcomes after ABL for AF.
References 1. Calkins H. Catheter ablation to maintain sinus rhythm. Circulation 2012;125: 1439–1445.
Srivatsa et al
Atrial Fibrillation Ablation Mortality and Stroke
2. Jais P, Cauchemez B, Macle L, Daoud E, Khairy P, Subbiah R, Hocini M, Extramiana F, Sacher F, Bordachar P, Klein G, Weerasooriya R, Clementy J, Haissaguerre M. Catheter ablation versus antiarrhythmic drugs for atrial fibrillation: the A4 study. Circulation 2008;118:2498–2505. 3. Oral H, Veerareddy S, Good E, Hall B, Cheung P, Tamirisa K, Han J, Fortino J, Chugh A, Bogun F, Pelosi F Jr, Morady F. Prevalence of asymptomatic recurrences of atrial fibrillation after successful radiofrequency catheter ablation. J Cardiovasc Electrophysiol 2004;15:920–924. 4. Kuwahara T, Takahashi A, Kobori A, Miyazaki S, Takahashi Y, Takei A, Nozato T, Hikita H, Sato A, Aonuma K. Safe and effective ablation of atrial fibrillation: importance of esophageal temperature monitoring to avoid periesophageal nerve injury as a complication of pulmonary vein isolation. J Cardiovasc Electrophysiol 2009;20:1–6. 5. Pappone C, Rosanio S, Augello G, Gallus G, Vicedomini G, Mazzone P, Gulletta S, Gugliotta F, Pappone A, Santinelli V, Tortoriello V, Sala S, Zangrillo A, Crescenzi G, Benussi S, Alfieri O. Mortality, morbidity, and quality of life after circumferential pulmonary vein ablation for atrial fibrillation: outcomes from a controlled nonrandomized long-term study. J Am Coll Cardiol 2003;42:185–197. 6. Baman TS, Jongnarangsin K, Chugh A, Suwanagool A, Guiot A, Madenci A, Walsh S, Ilg KJ, Gupta SK, Latchamsetty R, Bagwe S, Myles JD, Crawford T, Good E, Bogun F, Pelosi F Jr, Morady F, Oral H. Prevalence and predictors of complications of radiofrequency catheter ablation for atrial fibrillation. J Cardiovasc Electrophysiol 2011;22:626–631. 7. Pappone C, Oral H, Santinelli V, Vicedomini G, Lang CC, Manguso F, Torracca L, Benussi S, Alfieri O, Hong R, Lau W, Hirata K, Shikuma N, Hall B, Morady F. Atrio-esophageal fistula as a complication of percutaneous transcatheter ablation of atrial fibrillation. Circulation 2004;109:2724–2726. 8. Cappato R, Calkins H, Chen SA, Davies W, Iesaka Y, Kalman J, Kim YH, Klein G, Packer D, Skanes A. Worldwide survey on the methods, efficacy, and safety of catheter ablation for human atrial fibrillation. Circulation 2005;111:1100–1105. 9. Steiner C, Elixhauser A, Schnaier J. The healthcare cost and utilization project: an overview. Eff Clin Pract 2002;5:143–151. 10. Lip GY, Nieuwlaat R, Pisters R, Lane DA, Crijns HJ. Refining clinical risk stratification for predicting stroke and thromboembolism in atrial fibrillation using a novel risk factor-based approach: the Euro Heart Survey on Atrial Fibrillation. Chest 2010;137:263–272. 11. Oral H, Chugh A, Yoshida K, Sarrazin JF, Kuhne M, Crawford T, Chalfoun N, Wells D, Boonyapisit W, Veerareddy S, Billakanty S, Wong WS, Good E, Jongnarangsin K, Pelosi F Jr, Bogun F, Morady F. A randomized assessment of the incremental role of ablation of complex fractionated atrial electrograms after antral pulmonary vein isolation for long-lasting persistent atrial fibrillation. J Am Coll Cardiol 2009;53:782–789. 12. Haissaguerre M, Jais P, Shah DC, Takahashi A, Hocini M, Quiniou G, Garrigue S, Le Mouroux A, Le Metayer P, Clementy J. Spontaneous initiation of atrial fibrillation by ectopic beats originating in the pulmonary veins. N Engl J Med 1998;339:659–666. 13. Haissaguerre M, Sanders P, Hocini M, Jais P, Clementy J. Pulmonary veins in the substrate for atrial fibrillation: the “venous wave” hypothesis. J Am Coll Cardiol 2004;43:2290–2292. 14. Nademanee K, McKenzie J, Kosar E, Schwab M, Sunsaneewitayakul B, Vasavakul T, Khunnawat C, Ngarmukos T. A new approach for catheter ablation of atrial fibrillation: mapping of the electrophysiologic substrate. J Am Coll Cardiol 2004;43:2044–2053. 15. Narayan SM, Krummen DE, Clopton P, Shivkumar K, Miller JM. Direct or coincidental elimination of stable rotors or focal sources may explain successful atrial fibrillation ablation: on-treatment analysis of the confirm trial (conventional
1903
16.
17.
18. 19.
20.
21.
22.
23.
24.
25.
26.
27.
ablation for af with or without focal impulse and rotor modulation). J Am Coll Cardiol 2013;62:138–147. Wilber DJ, Pappone C, Neuzil P, De Paola A, Marchlinski F, Natale A, Macle L, Daoud EG, Calkins H, Hall B, Reddy V, Augello G, Reynolds MR, Vinekar C, Liu CY, Berry SM, Berry DA. ThermoCool AFTI. Comparison of antiarrhythmic drug therapy and radiofrequency catheter ablation in patients with paroxysmal atrial fibrillation: a randomized controlled trial. JAMA 2010;303: 333–340. Cosedis Nielsen J, Johannessen A, Raatikainen P, Hindricks G, Walfridsson H, Kongstad O, Pehrson S, Englund A, Hartikainen J, Mortensen LS, Hansen PS. Radiofrequency ablation as initial therapy in paroxysmal atrial fibrillation. N Engl J Med 2012;367:1587–1595. Schmidt B, Bordignon S, Furnkranz A, Chun KR. Catheter ablation of atrial fibrillation to reduce stroke risk. Herz 2013;38:247–250. Cleland JG, Coletta AP, Buga L, Ahmed D, Clark AL. 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;12:623–629. Birman-Deych E, Waterman AD, Yan Y, Nilasena DS, Radford MJ, Gage BF. Accuracy of ICD-9-CM codes for identifying cardiovascular and stroke risk factors. Med Care 2005;43:480–485. Roumie CL, Mitchel E, Gideon PS, Varas-Lorenzo C, Castellsague J, Griffin MR. Validation of ICD-9 codes with a high positive predictive value for incident strokes resulting in hospitalization using Medicaid health data. Pharmacoepidemiol Drug Saf 2008;17:20–26. Quan H, Li B, Saunders LD, Parsons GA, Nilsson CI, Alibhai A, Ghali WA. IMECCHI Investigators. Assessing validity of ICD-9-CM and ICD-10 administrative data in recording clinical conditions in a unique dually coded database. Health Serv Res 2008;43:1424–1441. Piccini JP, Sinner MF, Greiner MA, Hammill BG, Fontes JD, Daubert JP, Ellinor PT, Hernandez AF, Walkey AJ, Heckbert SR, Benjamin EJ, Curtis LH. Outcomes of Medicare beneficiaries undergoing catheter ablation for atrial fibrillation. Circulation 2012;126:2200–2207. Deshmukh A, Patel NJ, Pant S, Shah N, Chothani A, Mehta K, Grover P, Singh V, Vallurupalli S, Savani GT, Badheka A, Tuliani T, Dabhadkar K, Dibu G, Reddy YM, Sewani A, Kowalski M, Mitrani R, Paydak H, Viles-Gonzalez JF. Inhospital complications associated with catheter ablation of atrial fibrillation in the United States between 2000 and 2010: analysis of 93 801 procedures. Circulation 2013;128:2104–2112. Shah RU, Freeman JV, Shilane D, Wang PJ, Go AS, Hlatky MA. Procedural complications, rehospitalizations, and repeat procedures after catheter ablation for atrial fibrillation. J Am Coll Cardiol 2012;59:143–149. Wazni OM, Beheiry S, Fahmy T, Barrett C, Hao S, Patel D, Di Biase L, Martin DO, Kanj M, Arruda M, Cummings J, Schweikert R, Saliba W, Natale A. Atrial fibrillation ablation in patients with therapeutic international normalized ratio: comparison of strategies of anticoagulation management in the periprocedural period. Circulation 2007;116:2531–2534. Lakkireddy D, Reddy YM, Di Biase L, Vanga SR, Santangeli P, Swarup V, Pimentel R, Mansour MC, D'Avila A, Sanchez JE, Burkhardt JD, Chalhoub F, Mohanty P, Coffey J, Shaik N, Monir G, Reddy VY, Ruskin J, Natale A. Feasibility and safety of dabigatran versus warfarin for periprocedural anticoagulation in patients undergoing radiofrequency ablation for atrial fibrillation: results from a multicenter prospective registry. J Am Coll Cardiol 2012;59: 1168–1174.
