Factors predisposing to ventricular proarrhythmia during antiarrhythmic drug therapy for atrial fibrillation in patients with structurally normal heart Chin-Yu Lin, MD,*† Yenn-Jiang Lin, MD,*† Li-Wei Lo, MD,*† Yun-Yu Chen, MPH,*‡ Eric Chong, MBBS,§ Shih-Lin Chang, MD, PhD,*† Fa-Po Chung, MD,*† Tze-Fan Chao, MD,*† Yu-Feng Hu, MD,*† Ta-Chuan Tuan, MD,*† Jo-Nan Liao, MD,*† Yaoting Chang, MD,*† Kuo-Liong Chien, MD, PhD,‡ Chuen-Wang Chiou, MD,*† Shih-Ann Chen, MD*† From the *Division of Cardiology, Department of Medicine, Taipei Veterans General Hospital, Taipei, Taiwan, †Institute of Clinical Medicine, and Cardiovascular Research Center, National Yang-Ming University, Taipei, Taiwan, ‡Institute of Epidemiology and Preventive Medicine College of Public Health, National Taiwan University, Taipei, Taiwan, and §Division of Cardiology, Department of Medicine, Alexandra Hospital, Jurong Health, Singapore. BACKGROUND Ventricular arrhythmia (VA) can occur during propafenone therapy in atrial fibrillation (AF) patients with structurally normal heart. OBJECTIVE The purpose of this study was to evaluate the incidence and characteristics of propafenone-associated VAs in AF patients with structurally normal heart. METHODS We studied and compared the risk of new-onset VAs between AF patients with structurally normal heart taking and those not taking propafenone in a nationwide longitudinal cohort in Taiwan (n ¼ 127,197 since 2000). We then investigated the association between propafenone and VA in AF patients with structurally normal heart in a single-center database (n ¼ 396). RESULTS In the nationwide cohort, 102 patients (0.008% per patient-year) developed ventricular tachycardia (VT)/ventricular fibrillation (VF) during a follow-up period of 9.8 ⫾ 3.5 years. After multivariate Cox regression analysis, propafenone treatment was a significant risk factor for new-onset VT/VF with a hazard ratio (HR) of 3.59 (95% confidence interval [CI] 1.30–9.89, P ¼ .0136). Propafenone treatment offered protection against ischemic stroke with HR 0.649 (95% CI 0.55–0.77, Po.001). In the single-center

Introduction Atrial fibrillation (AF) is the most common sustained cardiac arrhythmia in general population. It is associated with increased risk of stroke, heart failure, ventricular fibrillation (VF), and mortality.1–3 The principles of AF management are rhythm control, rate control, and prevention of thromboembolism. Class IC antiarrhythmic drugs offer effective Address reprint requests and correspondence: Dr. Shih-Ann Chen, Division of Cardiology, Taipei Veterans General Hospital, 201 Sec 2, Shih-Pai Road, Taipei, Taiwan. E-mail address: [email protected].

1547-5271/$-see front matter B 2015 Heart Rhythm Society. All rights reserved.

study using ECG and medical records, the presence of inferior J wave, wider QRS, and old age were independent risk factors for VA after adjustment for clinical, biochemical, and echocardiographic variables. CONCLUSION Albeit with low incidence, propafenone therapy for AF was associated with new-onset VA in the nationwide longitudinal cohort study in Taiwan. Old age, presence of inferior lead J wave, and wider QRS on ECG were significant risk factors in our single-center study. KEYWORDS Atrial fibrillation; Propafenone; Proarrhythmia ABBREVIATIONS ACC ¼ American College of Cardiology; AF ¼ atrial fibrillation; AHA ¼ American Heart Association; CI ¼ confidence interval; EHRA ¼ European Heart Rhythm Association; ERS ¼ early repolarization syndrome; HR ¼ hazard ratio; NHI ¼ National Health Insurance; NHIRD ¼ National Health Insurance Research Database; OR ¼ odds ratio; VA ¼ ventricular arrhythmia; VF ¼ ventricular fibrillation; VT ¼ ventricular tachycardia (Heart Rhythm 2015;0:1–11) rights reserved.

I

2015 Heart Rhythm Society. All

therapy for the treatment of symptomatic AF. They are listed among the Class I guideline recommendations for treatment of AF patients with structurally normal heart with Class IC drugs (level of evidence A).4,5 Class IC antiarrhythmic drugs can be safely administered in patients without significant structural heart disease. The European consensus document recommends that patients at risk for ischemic heart disease undergo exercise stress test first before starting Class IC antiarrhythmic drug treatment (Class I).3 In clinical practice, ventricular arrhythmia (VA) can still occur in AF patients with structurally normal heart while they are undergoing http://dx.doi.org/10.1016/j.hrthm.2015.04.018

2 propafenone treatment. We evaluated the relationship and characteristics of propafenone-associated VA in AF patients with structurally normal heart.

Heart Rhythm, Vol 0, No 0, Month 2015 pairs 4:1 with identical propensity scores using a 0.01 caliper width for age, gender, hypertension, diabetes mellitus, chronic kidney disease, chronic lung disease, and hyperlipidemia (Table 1).6

Methods Study design

Part I: Follow-up

The study consisted of 2 parts. First, we studied and compared the risk of new-onset ventricular tachycardia (VT)/VF between AF patients with structurally normal heart taking and those not taking propafenone in a nationwide cohort in Taiwan. Second, we investigated the association between clinical, biochemical, ECG, and echocardiographic characteristics and the incidence of VT/VF in AF patients with structurally normal heart in a single-center database. We obtained Institutional Review Board (IRB) approval to conduct the retrospective study without requiring informed consent from the patients (NTUH-REC No. 201305044W, VGH-IRB No. 2013-08-002AC1). We also obtained permission for the rights for data linking from the National Research Institute for the Department of Health, and the Health Promotion Administration, Ministry of Health and Welfare.

The follow-up data were retrieved from NHIRD. The primary end-points of this study were incidences of VT/ VF, all-cause mortality, hemorrhagic stroke, and ischemic stroke. In order to ensure diagnostic accuracy, we defined patients with VT or VF only if the diagnosis was listed as the discharge diagnosis. Potential incident events and all deaths were investigated in detail based on initial identification through International Classification of Diseases diagnostic codes and the database of the CCHIA, which was validated previously.7–10

Part I: Propafenone and risk of VT/VF in the nationwide cohort We used the National Health Insurance Research Database (NHIRD) from the Ministry of Health and Welfare, Taiwan, to study the relationship between propafenone and new-onset VT/VF. The Taiwan’s National Health Insurance (NHI) program currently enrolls 23 million people, which covers 99% of the country’s population. The NHIRD contains data on utilization of all NHI resources, including outpatient visits, hospital care, prescribed medications, and the National Death Registry. The insurance claim database can be used for studies of natural diseases and for clinical research in realworld clinical settings. In the study cohort dataset, the patients' original identification numbers were encrypted to protect their privacy. The encrypting procedure was consistent so that linkage of the claims to the same patient was feasible within the NHI database and the patient can be followed continuously. From January 1, 2000, to December 31, 2001, patients from the NHIRD who were 18 years or older diagnosed with AF and without past history of VT/VF or structural heart disease were selected into the study population. Patients with new-onset coronary artery disease during the study period were deleted to avoid confounding by silent ischemic heart disease. We simultaneously excluded patients who had experienced a stroke, transient ischemic attack, valvular heart disease, and heart failure before the enrollment date. The study involved 127,197 participants who met the inclusion criteria. Among them, 2577 patients taking propafenone were assigned to the treatment group; the remaining patients were assigned to control group. We used propensity analysis and matching technique to minimize confounding from propafenone therapy in this cohort study. We matched

Part II: Clinical characteristics and occurrence of VT/VF in the single-center cohort This retrospective, longitudinal study included patients with symptomatic AF (European Heart Rhythm Association [EHRA] score II–IV) from the Taipei Veterans General Hospital from September 2002 to November 2013. Exclusion criteria were age o18 years or 480 years, previous implantation of pacemaker or defibrillator, presence of left atrial thrombus, active infection or sepsis, pregnancy, unstable angina, prior myocardial infarction, valvular heart disease, prior history of VT/VF, life expectation o12 months, mental disease, inability to follow-up, or coexisting diseases for which antiarrhythmic treatment was contraindicated. The study population involved a total of 1215 patients. Patients were treated according to the physician’s recommendations and concurrent guidelines. Class III drugs (amiodarone and dronedarone) were recommended for patients with structural cardiomyopathy, and Class IC (propafenone) plus diltiazem or beta-blockers were recommended for patients with structurally normal heart. Among these patients, 396 were taking propafenone, 281 betablocker, 6 mexiletine, 413 amiodarone, 24 dronedarone, 16 procainamide, and the remaining were taking calcium channel blocker. No patients in our study group were taking sotalol or flecainide. Patients not taking propafenone were excluded from the study. Collected variables included basic physical examination data, medication record, comorbidities, ECG, and echocardiography parameters before antiarrhythmic treatment. Medical history, physical examination, and 12-lead ECGs were recorded for all participants. Echocardiography was performed in all patients to exclude underlying structural heart disease. Participants were followed-up at regular clinic visits. Baseline cardiovascular comorbid conditions were ascertained by medical history, physical examination, physician report, and medical record review.

Part II: AF management principle The treatment of AF was in accordance to the American College of Cardiology (ACC)/American Heart Association

Lin et al

Baseline characteristics of overall NHIRD cohort and propensity-matched cohorts

Variable Age (years) Male gender HTN DM CKD COPD Hyperlipidemia CHA2DS2VASC score 0 1 Z2 Medication Antiplatelet Warfarin Quinidine Procainamide Mexiletine Beta-blocker Amiodarone Sotalol Verapamil Diltiazem

Control group (without propafenone) (n ¼ 124,620)

%

Study group (with propafenone) (n ¼ 2577)

%

65 ⫾ 13 70379 49226 13376 12761 3741 5834

56.47 39.5 10.73 10.24 3 4.68

64 ⫾ 13 1476 1069 236 204 38 110

57.28 41.48 9.16 7.92 1.47 4.27

31,489 52,477 40,654

25.3 42.1 32.6

703 1069 805

27.3 41.5 31.2

6360 1489 93 60 129 2100 1916 22 837 2223

5.1 1.19 0.07 0.05 0.1 1.69 1.54 0.02 0.67 1.78

0 0 0 0 0 45 0 0 0 0

0 0 0 0 0 1.75 0 0 0 0

P value

Control group (without propafenone) (n ¼ 10,308)

.99 .42 .04 .01 .00 o.0001 .33

64 ⫾ 13 5907 4275 941 148 816 435

.06

2815 4279 3214

o.0001 o.0001 .27 .64 .12 .06 o.0001 4.99 o.0001 o.0001

557 142 7 3 9 172 183 4 75 177

%

Study group (with propafenone) (n ¼ 2577)

%

P value

57.31 41.47 9.13 1.44 7.92 4.25

64 ⫾ 13 1476 1069 236 38 204 110

57.28 41.48 9.16 1.47 7.92 4.27

4.99 .98 .99 .96 .88 .99 .97

27.3 41.5 31.2

703 1069 805

27.3 41.5 31.2

5.4 1.38 0.07 0.03 0.09 1.67 1.78 0.04 0.73 1.72

0 0 0 0 0 45 0 0 0 0

0 0 0 0 0 1.75 0 0 0 0

VT/VF in Structurally Normal Heart AF Patient

Table 1

.99

o.001 o.001 .36 4.99 .22 .58 o.001 .059 o.001 o.001

CKD ¼ chronic kidney disease; COPD ¼ chronic obstructive pulmonary disease; DM ¼ diabetes mellitus; HTN ¼ hypertension; NHIRD ¼ National Health Insurance Research Database.

3

4

Heart Rhythm, Vol 0, No 0, Month 2015

(AHA)/EHRA guideline for AF management. An AF episode was defined as AF documented by ECG monitoring and having a duration of at least 30 seconds or, when less than 30 seconds, was present continuously throughout the ECG monitoring tracing. The presence of subsequent episodes of AF required sinus rhythm to be documented on ECG monitoring between AF episodes. Paroxysmal AF was defined as recurrent AF (42 episodes) that terminated spontaneously within 7 days. Episodes of AF of “48 hours’ duration that were terminated by cardioversion” were classified as paroxysmal AF episodes.11 The prescription of propafenone was according to the ACC/AHA/EHRA guidelines for AF management.3,12 Rhythm control therapy was indicated in patients with clinical symptoms of AF (EHRA score II–IV).13 All patients taking propafenone were free of heart failure, coronary heart disease, left ventricular hypertrophy, hypertensive heart disease, and history of VA.

Part II: Follow-up and event ascertainment Patients were followed-up at 1, 3, and 6 months, and then every 3 months after the initiation of propafenone treatment. Twelve-lead ECG was recorded at every visit. Twenty-four– hour Holter monitoring was performed every 12 months or whenever patients complained of palpitation. The primary end-point of the study was occurrence of VA. VAs were recorded on 24-hour Holter reports, ECG reports, defibrillation records in ambulance transfer sheets, and VA records in emergency room transfer sheets from other hospitals. The follow-up period was from the beginning of the registry to November 2013. After the incidence of VA, patients were followed-up monthly and underwent Holter monitoring every 3 months for half a year.

Statistical analysis All analyses were performed using SPSS statistical software (version 20, IBM Corporation, Armonk, NY, USA). Quantitative data are expressed as mean ⫾ SD. The Student t test was used to compare continuous variables. The χ2 test with Table 2

Results National longitudinal cohort study A total of 127,197 patients were studied. During mean followup duration of 9.8 ⫾ 3.5 years, 102 patients (0.08% of the study population, incidence 0.008% per patient-year) developed VT/ VF. Table 1 lists the baseline characteristics of 1:4 propensitymatched populations. There were 2557 and 10308 eligible subjects for analysis in the propafenone and control groups, respectively. The propafenone group was associated with a higher rate of VT/VF occurrence compared to control group (0.27% vs 0.08%, Po.001; Table 2). Mean age was 64 years. Approximately 57% of patients were male. Baseline characteristics were similar between the 2 groups with respect to age, sex, and comorbidities after propensity matching. More patients were taking antiplatelet and warfarin in the control group.

Cardiovascular end-points During median follow-up of 9.8 years (interquartile range 9.0–12.0 years), there were 15 cases of VA in the propensitymatched cohort. Kaplan–Meier curves suggested propafenone use increased the hazard of VA occurrence (hazard ratio [HR] 3.59, 95% CI 1.30–5.89, Po.001). In addition, there were 4398 mortality cases, 1067 new ischemic stroke cases, and 106 hemorrhagic stroke cases. Propafenone patients had higher risk for mortality (HR 1.096, 95% CI 1.019–1.179, P ¼ .0132) but lower risk for ischemic strokes (HR 0.649, 95% CI 0.545–0.773, Po.001), and lower risk for hemorrhagic strokes (HR 0.471, 95% CI 0.252–0.879, P ¼ .0181).

Clinical mortality and morbidity in AF patients with or without propafenone in the overall cohort and propensity-matched cohort Overall

Variable

Yates correction was used to analyze nonparametric data. For comparison within groups in which the expected number in any of the 4 cells was o5, a Fisher exact test was used. Multivariate analysis was performed with logistic regression and Cox regression hazard model to determine the independent predictors of VA. Variables selected for multivariate analysis were parameters with Po.1 in the univariate model. Odds ratio (OR) was significant when 95% confidence interval (CI) exceeded 1.0 and Po.05 (considered significant).

Propensity-Matched

Control group (without propafenone) (n ¼ 124,620) %

VT/VF 95 HF 7660 Ischemic stroke 10,877 Hemorrhagic stroke 901 Mortality 45,945 MI 2010 HF 1745 CV death 11,295 Cancer 3066

0.08 6.15 8.73 0.72 36.87 1.61 1.4 9.06 2.46

Study group (with propafenone) (n ¼ 2577) % 7 164 146 11 924 44 30 223 55

Control group (without propafenone) P value (n ¼ 10,308) %

0.27 .00 8 6.36 .65 640 5.67 o.0001 921 0.43 .08 95 35.86 .29 3474 1.71 .71 167 1.16 .31 123 8.65 .47 909 2.13 .29 231

0.08 6.21 8.93 0.92 33.7 1.62 1.19 8.82 2.24

Study group (with propafenone) (n ¼ 2577) % 7 164 146 11 924 44 30 223 55

P value

0.27 .02 6.36 .77 5.67 o.001 0.43 .01 35.86 .04 1.71 .75 1.16 .9 8.65 .79 2.13 .74

AF ¼ atrial fibrillation; CV ¼ cardiovascular; HF ¼ heart failure; MI ¼ myocardial infarction; VF ¼ ventricular fibrillation; VT ¼ventricular tachycardia.

Lin et al

VT/VF in Structurally Normal Heart AF Patient

The Kaplan–Meier curves of new-onset VT/VF-free survival, all survival, ischemic stroke-free survival, and hemorrhagic stroke-free survival between patients taking and those not taking propafenone are shown in Figure 1. Propafenone treatment led to significantly higher incidence of VT/VF and lower incidence of ischemic stroke with log-rank Po.05. Crude stroke incidence in the propafenone group was 5.67 per 100 person-years compared with 8.93 in the nonpropafenone group (Po.001). Crude stroke incidence risk

5 was significantly lower in the propafenone group compared with the non-propafenone group in the low, moderate, and high CHA2DS2-VASC score risk groups (Figure 2).

Clinical characteristics and occurrence of VT/VF in the single-center cohort All 396 patients in this retrospective cohort study received propafenone treatment for atrial arrhythmia. Surface ECGs

Figure 1 Kaplan–Meier survival curves by prescription with or without propafenone in a nationwide cohort. A: Kaplan–Meier ventricular tachycardia/ ventricular fibrillation (VT/VF-free survival curve in patients with or without propafenone. B: Kaplan–Meier survival curve in patients with or without propafenone. C: Kaplan–Meier ischemic stroke-free survival curves in patients with or without propafenone. D: Kaplan–Meier hemorrhagic stroke-free survival curves in patients with or without propafenone.

6

Heart Rhythm, Vol 0, No 0, Month 2015

Figure 2 Crude incidence rates of stroke by CHA2DS2-VASC score. Stroke crude incidence per 100 people among atrial fibrillation patients with or without propafenone treatment according to CHA2DS2-VASC score risk assessment for stroke. CHA2DS2-VASC ¼ congestive heart failure, hypertension, age Z75 years, diabetes mellitus, and previous stroke or transient ischemic attack, vascular disease, age 65–74 years, and sex category.

were available for all. After mean follow-up of 4.4 years, the proportion of persistent AF patients was 15.2%. The proportion of patients free from sustained episodes of paroxysmal AF was 37.0 %. For safety concern, echocardiography was performed in all patients, treadmill testing in 4%, thallium scans in 5%, and coronary angiography in 12%. None of the patients had previously documented VA based on the medical records. Overall, 6 patients had documentation of VA between 2002 and 2013 in Taipei Veterans General Hospital. Table 3 lists the baseline patient profiles, ECGs, and echocardiographic Table 3

characteristics, highlighting the comparison between patients with and those without VA. In univariate study, patients experiencing VA were older (P ¼ .003), and their ECGs showed wider QRS duration (P ¼ .0042), longer PR interval (P ¼ .011), and presence of J waves (Po.001). Older age, wider QRS duration, and presence of J wave were independently associated with increased incidence of VA in multivariate analysis (Table 3). Table 4 summarizes the detailed profiles of all patients with VA. All patients underwent electrophysiologic studies and echocardiography, which showed preserved left ventricular ejection fraction in

Baseline characteristics and univariate and multivariate analyses for the presence of ventricular arrhythmia

Variable

Ventricular arrhythmia

Univariate

Multivariate

General characteristic

Yes (n ¼ 6)

No (n ¼ 390)

P value

P value

OR

95% CI

Age (years) Men PAF Hypertension Diabetes mellitus CAD history Hyperlipidemia Thyroid disease Smoker Height (cm) BMI (kg/m2) SBP (mm Hg) DBP (mm Hg) CHA2DS2-VASC Basic biochemical parameter ALT (U/L) Creatinine (mg/dL) Echocardiographic parameter LAD (mm) LVDd (mm) LVEF (%) Electrocardiography QRSd (ms) PR (ms) QT (ms) QTc (ms) J wave (N/%)

66.8 ⫾ 11.8 6 (66.7) 5 (83.3) 1 (16.7) 0 (0) 1 (16.7) 2 (25.9) 0 (0) 3 (50) 164.6 ⫾ 11.9 25.5 ⫾ 4.0 128.2 ⫾ 11.0 71.4 ⫾ 5.9 1.2 ⫾ 1.2

52.9 ⫾ 11.1 276 (70.8) 330 (84.6) 173 (44.4) 35 (9.0) 73 (18.8) 100 (25.8) 6 (1.7) 105 (27.2) 166.9 ⫾ 12.6 24.9 ⫾ 3.7 125.9 ⫾ 15.6 78.6 ⫾ 11.3 1.1 ⫾ 1.0

.003 .827 .931 .236 .441 .894 .675 .396 .215 .679 .713 .745 .159 .846

.037 — — — — — — — — — — — — —

1.174 — — — — — — — — — — — — —

1.009–1.365 — — — — — — — — — — — — —

19.6 ⫾ 7.0 0.99 ⫾ 0.27

25.5 ⫾ 15.7 0.96 ⫾ 0.22

.403 .135

— —

— —

— —

38.3 ⫾ 6.6 50.3 ⫾ 49.2 56.0 ⫾ 2.8

37.9 ⫾ 5.9 48.2 ⫾ 5.1 59.3 ⫾ 7.3

.678 .434 .311

— — —

— — —

— — —

.016 .9 — — .011

1.123 1.003 — — 29.358

1.022–1.234 0.953–1.0057 — — 2.187–394.074

120.3 ⫾ 26.0 193.7 ⫾ 36.1 402.3 ⫾ 51.7 439.7 ⫾ 24.5 4 (66.7)

91.6 165.6 394.7 434.2 16

⫾ 10.5 ⫾ 26.5 ⫾ 39.9 ⫾ 25.7 (5)

.042 .011 .646 .606 o.001

Values are given as no. of events (%) or mean ⫾ SD unless otherwise indicated. ALT ¼ alanine transaminase; BMI = body mass index; CAD ¼ coronary artery disease; CI ¼ confident interval; DBP ¼ diastolic blood pressure; LAD ¼ left atrial diameter; LVDd ¼ left ventricular end-diastolic dimension; LVEF ¼ left ventricular ejection fraction; OR ¼ odds ratio; PAF ¼ paroxysmal atrial fibrillation; QRSd ¼ QRS duration; SBP ¼ systolic blood pressure. * OR was obtained by multivariable logistic regression for age, QRS duration, PR interval, and presence of J wave.

32 52 56 42 PAF, AFL M 58 6

F 86 5

M 77 4

AFL ¼ atrial flutter; CAG ¼ coronary angiography; DC ¼ direct cardioversion; EPS ¼ electrophysiologic study; ERS ¼ early repolarization syndrome; LA ¼ left atrium; LPAF ¼ long-lasting permanent atrial fibrillation; LVIDd ¼ left ventricular internal diameter end diastole; LVIDs ¼ left ventricular internal diameter end systole; LVEF ¼ left ventricular ejection fraction; NSVT ¼ nonsustained ventricular tachycardia; PAF ¼ paroxysmal atrial fibrillation; PVC ¼ premature ventricular complex; QRSd ¼ QRS duration; RVOT ¼ right ventricular outflow tract; VF ¼ ventricular fibrillation; VT ¼ ventricular tachycardia.

450 218 363

2

120

336

421 256 51

12.6 No inducible VT 10.1 No inducible VT 28 52 31 LPAF, rare PVC

43

24 56 46 PAF (ECG)

52

2

120

356

404 180 —

98

398

449 164

5.9 No inducible 1129 VT 9.8 No inducible 450 VT 22 56 M 60 3

F 61 2

M 58

Patent

Spontaneous PAF, AFL Polymorphic termination VT/VF Patent Ablation PAF (ECG) Polymorphic NSVT and syncope Patent DC shock PAF, AFL Polymorphic (ECG) VT/VF Insignificant DC shock PAF, AFL, Monomorphic rare VT, cardioPVC genic shock Insignificant Spontaneous LPAF, rare NSVT þ termination PVC syncope Patent DC shock PAF, AFL Polymorphic VT/VF

PAF (ECG)

38

43

33 58 35 PAF, frequent PVCs

52

2

96

404

474 452 168 120 —

176 41 PAF

58

54

34

6.7 No inducible 7 VT 10.9 RVOT VT 1812

2

168

468

440

VT/VF in Structurally Normal Heart AF Patient

1

Baseline Holter Treatment Age No. (years) Sex CAG

Table 4

Case series with ventricular arrhythmia

Ventricular arrhythmia

Postmedication LA LVEF LVIDs Holter (mm) (%) LVIDd (mm) (mm) E/E' EPS

Exposure ERS QRSd PR QT QTc days type (mm) (mm) (mm) (mm)

Lin et al

7 all. Three of these patients (50%) received cardioversion during VA. All of them underwent electrophysiologic studies, and VT/VF was noninducible in 5 of 6 (83.3%). The last patient was diagnosed with right ventricular outflow tract VT and underwent ablation. Neither of the patients developed further VT/VF episodes nor died after propafenone was stopped.

VA events Figure 1 shows the Kaplan–Meier estimates of cumulative incidence of first ventricular proarrhythmic event within 5 years for the 396 patients exposed to propafenone and for patients with or without J wave. After adjustment of significant covariates (defined as Po.05), presence of J wave was strongly associated with ventricular proarrhythmia (Table 3). Figure 3 shows an example of an ECG recorded before propafenone treatment and Holter recording of VT/ VF during treatment. In the 396 patients, 20 (5.1%) had documented J waves on 12-lead ECG. All of the J waves observed were inferior type (early repolarization syndrome [ERS] type 2). In multivariable logistic regression analysis, the presence of J waves was significantly associated with ventricular proarrhythmia, independent from other significant covariates, along with age, QRS duration, and PR interval (Table 3). Estimated OR and 95% CI for VA was 29.358 (2.187–394.074, P ¼ .011]. VA was observed in 2 (0.5%) from the J-wave negative group and 4 (20%) from the J-wave positive group (P o.001; Figure 4).

Discussion Main finding First, propafenone therapy associated VA had a low incidence of 0.008% per year. Propafenone treatment was a risk

Figure 3 Kaplan–Meier curve of ventricular arrhythmia-free survival in patients with or without J wave.

8

Heart Rhythm, Vol 0, No 0, Month 2015

Figure 4 Patient 2. A: Twelve-lead ECG before propafenone revealed J wave in inferior leads. B: Holter recording of atrial fibrillation/atrial flutter (AFL) with variable conduction transferring into 1:1 conduction while taking propafenone. C: Holter recording of AFL with 1:1 conduction that degenerated into ventricular fibrillation (VF) while taking propafenone. VT ¼ ventricular tachycardia.

factor for new-onset VA but a protecting factor against stroke. Second, older age, presence of inferior lead J wave, and wider QRS complex might be associated with VA in AF patients taking propafenone despite having a structurally normal heart.

Propafenone, risk factor, and VA AF was associated with increased risk for VF in a previous report.2 Propafenone is a Class IC antiarrhythmic drug that

has been commonly used for AF treatment.3 In the present guidelines, we avoid giving Class IC antiarrhythmic drugs to patients with structural heart disease. However, as shown in our study, their use can cause complications even in patients with structurally normal heart. We report a strong association between Class IC antiarrhythmic drugs and ventricular proarrhythmia in structurally normal heart patients, especially when concurrent J waves were present on baseline ECG.

Lin et al

VT/VF in Structurally Normal Heart AF Patient

In previous studies, propafenone was found to induce polymorphic VT in Brugada patients with normal structural heart among a broad spectrum of ERS.14 Samaan et al15 reported an AF patient taking propafenone who suffered from complications of wide QRS complex and syncope after heavy alcohol consumption. Femenia et al16 and Jastrzebski17 reported 3 patients with paroxysmal AF and atrial flutter taking propafenone for rhythm control. In 2 patients, slowing of the flutter cycle length resulted in a change from well-tolerated 2:1 atrioventricular conduction to 1:1 conduction with symptoms of palpitation and hypotension. In the other case, organization of atrial fibrillation to atypical atrial flutter and use-dependent left bundle branch block resulted in fast broad QRS tachycardia that mimicked VT.16,17 In our case series, most VAs were irregular and polymorphic. Unstable hemodynamic change was recorded in most of the patients. These characteristics favored ventricular origin of arrhythmia.

Relationship between J wave, Class IC antiarrhythmic drugs, and VA Type 2 J wave displays early repolarization pattern predominantly in the inferior or inferolateral leads and is associated with a higher level of ventricular arrhythmic risk in previous reports.18,19 Our study revealed high OR of 29.358 (95% CI 2.187–394.074) for VA in the type 2 ERS group. To the best of our knowledge, our study is the first to demonstrate a relationship between type 2 ERS and proarrhythmic effect of Class IC antiarrhythmic drugs in patients with a structurally normal heart.

Propafenone and VA Because the J wave is related to ventricular proarrhythmia in AF patients taking propafenone, it is reasonable to generate a hypothesis that propafenone treatment may cause VA in patients with AF. A previous report found an increased incidence of sudden cardiac death and proarrhythmic events in AF patients treated with flecainide.20 In comparison to our study, the study did not include the J wave as a variable. In addition, detailed univariate/multivariate analyses with/without the presence of VA were not performed. To the best of our knowledge, there are no other clinical studies evaluating VAs in AF patients with structurally normal heart receiving propafenone. Our study investigated this issue in a nationwide longitudinal cohort, which included adults free from coronary vascular disease, heart failure, and valvular heart disease. Therefore, our results could confirm that propafenone treatment is an important risk factor for VA in AF patients. The current study was the first to demonstrate a close relationship between propafenone and new-onset VA in a longitudinal cohort. Additional studies are needed to confirm the proarrhythmic effect of propafenone in patients with J waves.

9

Propafenone and stroke The present analysis showed reduced risk of stroke in patients with paroxysmal AF receiving propafenone treatment. Dronedarone has been the only antiarrhythmic drug that demonstrated reduction in stroke risk in patients with AF.21,22 Whether the stroke reduction was due to the maintenance of sinus rhythm or other factors is unclear. Prevention of arrhythmia recurrence could reduce stroke risk. Although this theory might be reasonable at first impression, previous trials evaluating antiarrhythmic drugs failed to demonstrate stroke prevention despite reduction in AF recurrence. In previous studies comparing rhythm and rate control therapies, stroke rates were similar between rate control and rhythm control groups. These data conflict with our result. Three of 4 meta-analyses showed similar thromboembolic outcomes between the rate and rhythm control strategies (total 13,760 patient-years in 5 trials), whereas 1 metaanalysis showed that rate control strategy was associated with a strikingly lower risk of thromboembolic stroke (total 215,467 patient-years in 8 trials).23–26 In 1 large populationbased observational study, compared with rate control therapy, rhythm control therapy was associated with lower rates of stroke/transient ischemic attack among AF patients, particularly among those with CHAD2 score of 1 at maximum 8.2 years of follow-up (total 115,340 patientyears).27 We propose several reasons to explain why we obtained different results from other studies. First, our study was a population-based observational study that excluded high-risk patients. The difference in study populations and the larger sample size in our study compared with other studies could account for the difference in stroke outcome. Second, a total of 124,6531 patient-years of follow-up was obtained in our study, which is far longer than any previous studies, and this may explain the difference in our result. Third, observational studies are designed to identify associations rather than causality; therefore, our results can only imply an association between rhythm control therapy and stroke prevention. To the best of our knowledge, this is the first nationwide cohort study demonstrating the stroke-protecting effect of antiarrhythmic drugs apart from dronedarone.

Anticoagulant drugs in Taiwan The low prevalence of use of anticoagulant and antithrombotic agents in our nationwide cohort population has been noted. First, antithrombotic therapy was underused in Asia as well as in Taiwan, as reported in previous population study.28 Second, the nationwide cohort population in our study excluded patients with multiple morbidities such as heart failure, ischemic stroke, ischemic heart disease, thyroid dysfunction, and valvular heart disease. The mean CHADS2 score was relatively low at 0.67 (CHA2DS2-VASC score 1.04) in the study group. The data were retrieved from the 2000 to 2001 database. There was no indication for antithrombotic drug use when the CHADS2 score was o1.

10

Propafenone prescription and rate control medication The low prevalence of prescribing rate control agents in our nationwide cohort population has been noted. There is general evidence that treating AF patient with propafenone in the absence of beta-blockers or calcium channel blockers may organize AF into atrial flutter with 1:1 conduction.2,20 Before beginning therapy with these agents, a beta-blocker or calcium channel blocker should be given to prevent rapid AV conduction or 1:1 AV conduction if atrial flutter develops. The result suggests that rate control medication might be underused before initiation of Class IC medication.

Heart Rhythm, Vol 0, No 0, Month 2015 registration and claim datasets from the NHI might underestimate the usage of rate control medication.

Conclusion In AF patients with structurally normal heart and no previous history of VA, rhythm control therapy with propafenone was associated with a higher incidence of VAs. The incidence was low (0.008% per patient-year) based on the nationwide longitudinal cohort study in Taiwan. In our single-center study, older age, presence of inferior lead J waves, and wider QRS complex were identified as risk factors for VA.

Clinical implications One interesting finding from the present study was that propafenone could play a part in ventricular proarrhythmia in AF patients without structural heart disease, albeit an extremely low incidence. In addition to the known risk factors for propafenone-related VA, such as ischemic heart disease, heart failure, and valvular heart disease, several novel risk factors were recognized in our study. These factors include old age, wide QRS complex, and presence of J wave on surface ECG. Our results suggest that the physician should be more careful when starting propafenone in patients with structurally normal heart, particularly in elderly patients with J wave and wide QRS duration.

Study limitations There were several limitations to our study. First, the ECG and echocardiographic characteristics were not available in the nationwide dataset. We defined structurally normal heart patients by excluding patients with diagnosis of ischemic heart disease, valvular heart disease, heart failure, and congenital heart disease according to International Classification of Diseases codes. Second, the occurrence of VT/VF was based on the diagnostic code documented by physicians responsible for patients, and this was not verified further. However, VT/VF should be highly specific diagnoses. In Taiwan, cardiologist referral was routinely made after diagnosis of VT/VF. This is a common limitation that has been noted in previous studies. To ensure the accuracy of diagnosis, we defined patients as having VT/VF only when it was listed as a discharge diagnosis. The diagnostic accuracy of arrhythmia using this definition has been validated.7,10 Third, we stopped propafenone administration before electrophysiologic study. For safety reasons, we could not repeat electrophysiologic study in patients during propafenone treatment to evaluate the association between propafenone and VA further. Fourth, exercise testing is recommended for patients receiving Class IC antiarrhythmic drugs in order to assess the proarrhythmic risk. In our single-center study, all patients underwent echocardiography, but only 10% underwent stress testing and 12% underwent coronary angiography. Sixth, the rate control medication was prescribed temporarily and therefore was not used in the long term. The cross-section of retrospective analysis using the

References 1. Sherman DG, Kim SG, Boop BS, Corley SD, Dimarco JP, Hart RG, Haywood LJ, Hoyte K, Kaufman ES, Kim MH, Nasco E, Waldo AL. Occurrence and characteristics of stroke events in the Atrial Fibrillation Follow-up Investigation of Sinus Rhythm Management (AFFIRM) study. Arch Intern Med 2005;165: 1185–1191. 2. Bardai A, Blom MT, van Hoeijen DA, van Deutekom HW, Brouwer HJ, Tan HL. Atrial fibrillation is an independent risk factor for ventricular fibrillation: a largescale population-based case-control study. Circ Arrhythm Electrophysiol 2014;7: 1033–1039. 3. Camm AJ, Kirchhof P, Lip GY, et al. Guidelines for the management of atrial fibrillation: the Task Force for the Management of Atrial Fibrillation of the European Society of Cardiology (ESC). Europace 2010;12:1360–1420. 4. Pritchett EL. Management of atrial fibrillation. N Engl J Med 1992;326: 1264–1271. 5. Sopher SM, Camm AJ. Atrial fibrillation: maintenance of sinus rhythm versus rate control. Am J Cardiol 1996;77:24A–37A. 6. Garrido MM, Kelley AS, Paris J, Roza K, Meier DE, Morrison RS, Aldridge MD. Methods for constructing and assessing propensity scores. Health Serv Res 2014;49:1701–1720. 7. Chao TF, Hung CL, Chen SJ, Wang KL, Chen TJ, Lin YJ, Chang SL, Lo LW, Hu YF, Tuan TC, Chen SA. The association between hyperuricemia, left atrial size and new-onset atrial fibrillation. Int J Cardiol 2013;168:4027–4032. 8. Lu TH, Lee MC, Chou MC. Accuracy of cause-of-death coding in Taiwan: types of miscoding and effects on mortality statistics. Int J Epidemiol 2000;29: 336–343. 9. Hsieh CY, Chen CH, Li CY, Lai ML. Validating the diagnosis of acute ischemic stroke in a National Health Insurance claims database. J Formos Med Assoc 2015;114:254–259. 10. Tamariz L, Harkins T, Nair V. A systematic review of validated methods for identifying ventricular arrhythmias using administrative and claims data. Pharmacoepidemiol Drug Saf 2012;21(Suppl 1):148–153. 11. Calkins H, Kuck KH, Cappato R, et al. 2012 HRS/EHRA/ECAS Expert Consensus Statement on Catheter and Surgical Ablation of Atrial Fibrillation: recommendations for patient selection, procedural techniques, patient management and follow-up, definitions, endpoints, and research trial design. Europace 2012;14:528–606. 12. Camm AJ, Lip GY, De Caterina R, Savelieva I, Atar D, Hohnloser SH, Hindricks G, Kirchhof P. 2012 focused update of the ESC Guidelines for the management of atrial fibrillation: an update of the 2010 ESC Guidelines for the management of atrial fibrillation—developed with the special contribution of the European Heart Rhythm Association. Europace 2012;14:1385–1413. 13. Kirchhof P, Auricchio A, Bax J, et al. Outcome parameters for trials in atrial fibrillation: executive summary. Eur Heart J 2007;28:2803–2817. 14. Karaca M, Dinckal MH. Monomorphic and propafenone-induced polymorphic ventricular tachycardia in Brugada syndrome: a case report. Acta Cardiol 2006;61:481–484. 15. Samaan RA, Sobamowo HO, Tamburrino F, Grodman R, Isber N. Syncope, widened QRS interval, and left ventricular systolic depression: coincident with propafenone therapy for atrial fibrillation. Tex Heart Inst J 2010;37:476–479. 16. Femenia F, Palazzolo J, Arce M, Arrieta M. Proarrhythmia induced by propafenone: what is the mechanism? Indian Pacing Electrophysiol J 2010;10: 278–280. 17. Jastrzebski M. Proarrhythmic effect of propafenone in patients with atrial fibrillation and atrial flutter. Kardiol Pol 2008;66:1221–1224. 18. Antzelevitch C, Yan GX. J wave syndromes. Heart Rhythm 2010;7:549–558.

Lin et al

VT/VF in Structurally Normal Heart AF Patient

19. Junttila MJ, Sager SJ, Tikkanen JT, Anttonen O, Huikuri HV, Myerburg RJ. Clinical significance of variants of J-points and J-waves: early repolarization patterns and risk. Eur Heart J 2012;33:2639–2643. 20. Almroth H, Andersson T, Fengsrud E, Friberg L, Linde P, Rosenqvist M, Englund A. The safety of flecainide treatment of atrial fibrillation: long-term incidence of sudden cardiac death and proarrhythmic events. J Intern Med 2011;270:281–290. 21. Dagres N, Varounis C, Iliodromitis EK, Lekakis JP, Rallidis LS, AnastasiouNana M. Dronedarone and the incidence of stroke in patients with paroxysmal or persistent atrial fibrillation: a systematic review and meta-analysis of randomized trials. Am J Cardiovasc Drugs 2011;11:395–400. 22. Freemantle N, Lafuente-Lafuente C, Mitchell S, Eckert L, Reynolds M. Mixed treatment comparison of dronedarone, amiodarone, sotalol, flecainide, and propafenone, for the management of atrial fibrillation. Europace 2011;13: 329–345. 23. Caldeira D, David C, Sampaio C. Rate vs rhythm control in patients with atrial fibrillation and heart failure: a systematic review and meta-analysis of randomised controlled trials. Eur J Intern Med 2011;22:448–455.

11 24. de Denus S, Sanoski CA, Carlsson J, Opolski G, Spinler SA. Rate vs rhythm control in patients with atrial fibrillation: a meta-analysis. Arch Intern Med 2005;165:258–262. 25. Kumana CR, Cheung BM, Cheung GT, Ovedal T, Pederson B, Lauder IJ. Rhythm vs. rate control of atrial fibrillation meta-analysed by number needed to treat. Br J Clin Pharmacol 2005;60:347–354. 26. Testa L, Biondi-Zoccai GG, Dello Russo A, Bellocci F, Andreotti F, Crea F. Rate-control vs. rhythm-control in patients with atrial fibrillation: a meta-analysis. Eur Heart J 2005;26:2000–2006. 27. Tsadok MA, Jackevicius CA, Essebag V, Eisenberg MJ, Rahme E, Humphries KH, Tu JV, Behlouli H, Pilote L. Rhythm versus rate control therapy and subsequent stroke or transient ischemic attack in patients with atrial fibrillation. Circulation 2012;126:2680–2687. 28. Lin LJ, Cheng MH, Lee CH, Wung DC, Cheng CL, Kao Yang YH. Compliance with antithrombotic prescribing guidelines for patients with atrial fibrillation–a nationwide descriptive study in Taiwan. Clin Ther 2008;30:1726–1736.

CLINICAL PERSPECTIVES This study found 2 important findings. (1) Class IC antiarrhythmic drugs might cause ventricular arrhythmia (VA) in AF patients with structurally normal heart. (2) Rhythm control medications might decrease stroke incidence in long-term follow-up. Physicians might prevent VA in high-risk patients. More clinical studies are needed to confirm the association between J wave and VA in AF patients with structurally normal heart.