Original Article. Relationship of Sudden Cardiac Death to New- Onset Atrial Fibrillation in Hypertensive Patients With Left Ventricular Hypertrophy

Transcription

1 Original Article Relationship of Sudden Cardiac Death to New- Onset Atrial Fibrillation in Hypertensive Patients With Left Ventricular Hypertrophy Peter M. Okin, MD; Casper N. Bang, MD; Kristian Wachtell, MD, PhD; Darcy A. Hille, MSEMBA; Sverre E. Kjeldsen, MD, PhD; Björn Dahlöf, MD, PhD; Richard B. Devereux, MD Background Prevalent atrial fibrillation (AF) is associated with a higher sudden cardiac death (SCD) rate in some populations, and incident AF predicts increased mortality risk in the general population and after myocardial infarction. However, the relationship of SCD to new-onset AF is unclear. Methods and Results The relationship of SCD to new-onset AF was evaluated in 8831 hypertensive patients with electrocardiographic left ventricular hypertrophy with no history of AF, in sinus rhythm on their baseline electrocardiogram, randomly assigned to losartan- or atenolol-based treatment. During 4.7±1.1 years mean follow-up, new-onset AF occurred in 701 patients (7.9%) and SCD in 151 patients (1.7%). In univariate Cox analyses, new-onset AF was associated with a >4-fold higher risk of SCD (hazard ratio, 4.69; 95% CI interval, ; P<0.001). In multivariate Cox analyses adjusting for age, sex, race, diabetes mellitus, history of heart failure, myocardial infarction, ischemic heart disease, stroke, smoking, serum high-density lipoprotein cholesterol, creatinine, glucose, and urine albumin/creatinine ratio as standard risk factors, and for incident myocardial infarction, in-treatment use of digoxin, systolic and diastolic pressure, heart rate, QRS duration, Cornell voltage-duration product, and Sokolow-Lyon voltage left ventricular hypertrophy treated as time-varying covariates, new-onset AF remained associated with a >3-fold increased risk of SCD (hazard ratio, 3.13; 95% confidence interval, ; P<0.001). Conclusions Development of new-onset AF identifies hypertensive patients at increased risk of SCD. Clinical Trial Registration URL: Unique identifier: NCT (Circ Arrhythm Electrophysiol. 2013;6: ) Key Words: atrial fibrillation death, sudden, cardiac electrocardiography hypertension hypertrophy Sudden cardiac death (SCD) is a major public health issue, claiming to lives annually in the United States, 1,2 and may account for as much as 6% of total mortality in some communities. 3 The majority of sudden cardiac arrests occur in the community and are associated with a <10% survival rate. 1 Because SCD often presents without previous warning, 4,5 better identification of risk factors for SCD is warranted. 2,6 Clinical Perspective on p 251 Atrial fibrillation (AF) is a common arrhythmia 7,8 that has been increasing in prevalence. 8 The incidence of AF increases with age 7 and in patients with hypertension, left ventricular hypertrophy (LVH), coronary heart disease, and heart failure Prevalent and incident AF have been associated with an increased risk of all-cause mortality in population-based studies, postmyocardial infarction, 19,20 and in most but not all 27,28 analyses of patients with heart failure. However, less is known about the relationship of SCD to AF. Prevalent AF is associated with an increased risk of SCD in patients postmyocardial infarction 20,29,30 and in women with coronary artery disease, 31 but the relationship of SCD to AF in patients with heart failure has been inconsistent. 24,25,27 Moreover, in patients with implantable cardioverter-defibrillators (ICDs), a history of AF, 32 persistent AF at the time of ICD implantation, 33 and permanent but not persistent or paroxysmal AF, 34 have an increased risk of appropriate device therapies for ventricular arrhythmias. Hypertensive patients with LVH are at a high risk of SCD. 35 In hypertensive patients with electrocardiographic (ECG) LVH, serial assessment of ECG LVH, QRS duration, and heart rate have been demonstrated to stratify the risk of SCD beyond the information provided by standard risk factors Received August 2, 2012; accepted December 23, From the Greenberg Division of Cardiology, Weill Cornell Medical College, New York, NY (P.M.O., C.N.B., R.B.D.); the Department of Cardiology, Gentofte Hospital, Copenhagen, Denmark (K.W.); Merck Research Labs, West Point, PA (D.A.H.); University of Oslo, Ullevål Hospital, Oslo, Norway (S.E.K.); University of Michigan Medical Center, Ann Arbor, MI (S.E.K.); Sahlgrenska University Hospital/Östra, Göteborg, Sweden (B.D.). The online-only Data Supplement is available at Correspondence to Peter M. Okin, MD, Weill Medical College of Cornell University, 525 East 68th St, New York, NY pokin@med. cornell.edu 2013 American Heart Association, Inc. Circ Arrhythm Electrophysiol is available at DOI: /CIRCEP

2 244 Circ Arrhythm Electrophysiol April 2013 However, whether hypertensive patients with new-onset AF have an increased risk of SCD independent of risk factors for AF and SCD has not been examined. Therefore, the present study examined whether new-onset AF was associated with a higher risk of SCD in hypertensive patients with ECG LVH after adjusting for differences in the prevalence of SCD risk factors, treatment effects, in-treatment blood pressure, and for the previously demonstrated relation of black race and in-treatment ECG LVH, QRS duration, and heart rate to SCD in this population Methods Subjects The Losartan Intervention For Endpoint reduction in hypertension (LIFE) Study enrolled 9193 hypertensive patients with ECG LVH by Cornell voltage-duration product 41,42 or Sokolow-Lyon voltage criteria 31 on a screening ECG in a prospective, double-blind, randomized study that compared cardiovascular morbidity and mortality with losartan- as opposed to atenolol-based treatment, 39 as previously described A total of 362 patients had a history of AF (n=342) or AF on their LIFE baseline ECG (n=135), leaving 8831 patients without AF by history or baseline ECG in the present post hoc, retrospective analysis. Treatment Regimens Blinded treatment was begun with losartan 50 mg or atenolol 50 mg daily and matching placebo of the other agent, with a target pressure of 140/90 mm Hg or lower. During clinic visits at frequent intervals for the first 6 months and at 6-month intervals thereafter, study therapy could be up-titrated by addition of hydrochlorothiazide 12.5 mg, followed by increase in blinded losartan or atenolol to 100 mg daily. In patients whose blood pressure was still not controlled, additional open-label upward titration of hydrochlorothiazide and if necessary institution of therapy with a calcium channel blocker or additional other medications (excluding AT1- or β-blockers or ACE-inhibitors) was added to the double-blind treatment regimen. 39 Electrocardiography Study ECGs were obtained at baseline, at 6-months, and at yearly follow-up intervals until study termination or patient death and were interpreted as previously reported in detail Cornell product >2440 mm ms or Sokolow-Lyon voltage >38 mm were used to identify LVH End Point Determination New-onset AF was identified in a total of 701 patients, either from protocol-mandated in-study ECGs undergoing Minnesota coding at the ECG core laboratory (n=405) or by adverse event reports of AF by the investigators (n=572). 43 In patients who had new AF by both criteria, the earliest onset AF was taken as the time to new AF for this analysis. SCD was a prespecified secondary end point in LIFE and was defined as death that was sudden and unexpected, including observed arrhythmic deaths and those not attributable to myocardial infarction, intractable heart failure, or other identifiable cause, that occurred within 24 hours of symptom onset or when the subject was last seen alive if SCD was unwitnessed If an autopsy was performed in a patient who died suddenly and evidence of a recent myocardial infarction was found, the death was classified as secondary to myocardial infarction and not a SCD Each case was reviewed and verified by the End Point Committee who were blinded to study ECG LVH findings when classifying possible morbid events ,43 Statistical Analyses Data management and analyses were performed by the investigators using SPSS version 19.0 (IBM Inc, Armonk, NY), the free statistical software program R version ( and SAS statistical software package version 9.2 for PC (SAS Institute Inc, Cary, NC). Data are presented as mean±sd for continuous variables and proportions for categorical variables. Differences in mean values between patients grouped according to development of new AF were compared using unpaired t tests; comparison of proportions between groups was performed using χ 2 tests. The relation of new-onset AF to the risk of having SCD was assessed using Cox proportional hazards models with AF status as a time-varying covariate. 44 Time to SCD was measured from study entry in those without new AF and from time of first diagnosis of new-onset AF in patients with new AF to either SCD or when patients were censored for any other death or last in-study and known alive. To test the independence of new-onset AF as a predictor of SCD, new-onset AF was entered into a multivariable Cox model that also included as covariates other significant univariable predictors of SCD (Table I in the online-only Data Supplement) from among randomized treatment allocation, age, sex, race, diabetes mellitus, history of heart failure, myocardial infarction, ischemic heart disease, stroke, peripheral vascular disease, smoking, baseline body mass index, serum total and HDL cholesterol, creatinine, glucose, and urine albumin/creatinine ratio entered as standard risk factors; and incident myocardial infarction, in-treatment use of digoxin and antiarrhythmic drugs, diastolic and systolic blood pressure, heart rate, QRS duration, Cornell voltage-duration product, and Sokolow-Lyon voltage LVH entered as time-varying covariates by choosing the last measurement before SCD. Analyses were repeated stratifying the population by relevant subgroups by adding cross-product terms of new-onset AF and these subgroup variables into models in the total population. To account for the difference in all-cause mortality, odds ratio were determined by multivariable logit link model for the competing risk of death from any cause other than SCD, where a regression analysis of the cumulative incidence probabilities was calculated. 45 To illustrate the results of delayed-entry analyses, SCD rate was plotted as a function of new-onset AF using a univariate modified Kaplan Meier method, 46 implemented in SAS Release 8.2 on the WIN_PRO platform. Two-tailed P<0.05 was required for statistical significance. Results Patient Characteristics in Relation to Development of Atrial Fibrillation During mean follow-up of 4.7±1.1 years, new-onset AF occurred in 701 patients (7.9%). Clinical and demographic characteristics of patients in relationship to development of new AF are shown in Table 1. Hypertensive patients who developed new AF were older, more likely to be male, nonblack, have a history of ischemic heart disease, previous myocardial infarction, stroke and heart failure, had lower total cholesterol levels, greater albuminuria, and were less likely to be randomized to losartan-based treatment, but were similar with respect to other baseline characteristics. Blood pressure and ECG measurements at baseline and changes in these measurements between baseline and last in-study determination or the development of new-onset AF are shown in Table 2. Patients with new-onset AF had slightly higher mean baseline systolic pressure, lower baseline diastolic pressure, and greater reduction in systolic pressure but similar change in diastolic pressure. New-onset AF was associated with slightly lower mean baseline heart rate, slightly longer QRS duration, and more severe baseline ECG LVH by Cornell product and Sokolow-Lyon voltage criteria.

3 Okin et al Atrial Fibrillation and Sudden Cardiac Death 245 Table 1. Demographic and Clinical Characteristics in Relation to Development of New Atrial Fibrillation Variables No AF (n=8130) New AF (n=701) Age, y 66.6± ±6.5 <0.001 Sex, % male Race, % Black Treatment with Losartan, % Diabetes mellitus, % History of ischemic heart disease, % <0.001 History of myocardial infarction, % History of stroke, % History of peripheral vascular disease, % History of heart failure, % <0.001 Current smokers, % Body mass index, kg/m ± ± Serum glucose, mmol/l 6.00± ± Serum creatinine, μmol/l 86.5± ± Total cholesterol, mmol/l 6.06± ± High-density lipoprotein cholesterol, mmol/l 1.50± ± Uric acid, μmol/l 329±78 331± Urine albumin/creatinine ratio, mg/mmol 6.6± ± AF indicates atrial fibrillation. P Value Patients who developed AF had smaller reduction in mean heart rate, slightly greater increase in QRS duration, and less regression of LVH by Cornell product criteria, but had similar change in Sokolow-Lyon voltage compared with patients who did not develop AF. Digoxin and antiarrhythmic drug use at baseline and at year intervals during treatment in relation to development of new-onset AF are shown in Table 3. Patients who developed new-onset AF during the study were significantly more likely to be treated with digoxin or antiarrhythmic drugs at baseline and at each year of the study. Sudden Cardiac Death in Relation to Incident Atrial Fibrillation During mean follow-up of 4.7±1.1 years, SCD occurred in 151 patients (1.7%) and was significantly more frequent in patients Table 2. Baseline and Change From Baseline to Development of New Atrial Fibrillation or Last In-Study Measurement of Blood Pressure, Heart Rate, QRS Duration, and Electrocardiographic Left Ventricular Hypertrophy in Relation to Development of New Atrial Fibrillation Variables No AF (n=8130) New AF (n=701) Baseline measurements Systolic blood pressure, mm Hg 174±14 177±14 <0.001 Diastolic blood pressure, mm Hg 98±9 97±9 <0.001 Heart rate, bpm 74±11 73± QRS duration, ms 101±18 103± Cornell voltage-duration product, mm ms 2804± ± Sokolow-Lyon voltage, mm 29.8± ±11.6 <0.001 Change from baseline to last measurement* Systolic blood pressure, mm Hg 29±19 34±21 <0.001 Diastolic blood pressure, mm Hg 17±10 17± Heart rate, bpm 5±13 3±15 <0.001 QRS duration, ms 2±12 3± Cornell voltage-duration product, mm ms 204± ± Sokolow-Lyon voltage, mm 3.8± ± AF indicates atrial fibrillation; and bpm, beats per minute. *Change from baseline to last in-study measurement or last measurement before onset of new atrial fibrillation. P Value

4 246 Circ Arrhythm Electrophysiol April 2013 Table 3. Baseline and In-Treatment Use of Digoxin and Antiarrhythmic Drugs in Relation to Development of New Atrial Fibrillation Variables No AF (n=8130) New AF (n=701) P Value Digoxin use, % Baseline <0.001 Year <0.001 Year <0.001 Year <0.001 Year <0.001 Year <0.001 Antiarrhythmic drug use, % Baseline Year Year <0.001 Year <0.001 Year <0.001 Year <0.001 AF indicates atrial fibrillation. with new AF (25/701, 3.6%, 15.0/1000 patient-years) than in those without new AF (126/8130, 1.5%, 3.3/1000 patientyears, P<0.001). The relationship of SCD to new-onset AF is examined in Table 4 and the Figure. In univariate Cox analyses, new-onset AF was associated with a >4-fold higher risk of SCD (hazard ratio [HR], 4.69; 95% confidence interval [CI], ; P<0.001). In multivariate Cox analyses adjusting for statistically significant univariable predictors of SCD (see Table I in the online-only Data Supplement), including age, sex, race, diabetes mellitus, history of heart failure, myocardial infarction, ischemic heart disease, stroke, smoking, serum HDL cholesterol, creatinine, glucose and urine albumin/ creatinine ratio entered as standard risk factors, and for incident Table 4. Univariable and Multivariable Cox Regression and Competing Risk Analyses to Assess the Risk of Sudden Cardiac Death in Relation to the Development of New-Onset Atrial Fibrillation Model Hazard Ratio 95% CI P Value Univariable Cox <0.001 Multivariable Cox* <0.001 Alternative multivariable Cox** <0.001 Competing risk <0.001 CI indicates confidence interval; HDL, high-density lipoprotein; and LVH, left ventricular hypertrophy. *Adjusted for age, sex, race, diabetes mellitus, history of heart failure, myocardial infarction, ischemic heart disease, stroke, smoking, baseline serum high-density lipoprotein cholesterol, creatinine and glucose, and urine albumin/ creatinine ratio as standard risk factors, and for incident myocardial infarction, in-treatment digoxin use, diastolic and systolic blood pressure, heart rate, QRS duration, Cornell product, and Sokolow-Lyon voltage LVH as time-varying covariates. **Adjusted only for age, sex, race, diabetes mellitus, history of heart failure, myocardial infarction, ischemic heart disease, stroke, smoking, baseline serum HDL cholesterol, creatinine and glucose, and urine albumin/creatinine ratio as standard risk factors. myocardial infarction, in-treatment use of digoxin, systolic and diastolic pressure, heart rate, QRS duration, Cornell voltageduration product, and Sokolow-Lyon voltage LVH treated as time-varying covariates, new-onset AF remained associated with a >3-fold increased risk of SCD (HR, 3.13; 95% CI, ; P<0.001). Of note, although rate-corrected QT interval (QTc) was not routinely measured at baseline in the LIFE study, after additional adjustment for baseline QTc in a subset of 6657 patients, new-onset AF remained associated with SCD in the full multivariable Cox model (adjusted HR, 2.61; 95% CI, ; P=0.004). In parallel analyses in the entire population examining an alternative outcome of death attributable to coronary heart disease occurring <24 hours from onset of symptoms (n=78), new-onset AF remained associated with a >3-fold increased risk of this mortality end point in multivariable Cox analyses (HR, 3.45; 95% CI, ; P<0.001). In addition, new-onset AF remained associated with a >3-fold increased risk of SCD in an alternative multivariable Cox analysis that adjusted only for the baseline standard risk factors included in the full multivariable Cox model (Table 4). All-cause mortality occurred in 7.6% (618/8130) of patients who remained in sinus rhythm and in 16.6% (114/701) of patients who developed in-study new-onset AF. In competing risk calculations using a logit-link competing risk model with death from any cause other than SCD as a competing event, new-onset AF was independently associated with SCD (odds ratio, 5.58; 95% CI, ; P<0.001). In addition, newonset AF was independently associated with all-cause mortality in both univariate (HR, 2.30; 95% CI, ; P<0.001) and multivariate Cox analyses using the same covariates as above (HR, 1.50; 95% CI, ; P=0.001). Modified Kaplan Meier curves 45 comparing the rate of SCD according to development of new AF over the time course of the study (Figure) demonstrate that new-onset AF was associated with a greater risk of developing SCD as compared with the absence of AF, with development of new AF associated with an estimated 5% higher absolute incidence of SCD after 4 years of follow-up. The predictive value of new-onset AF for SCD in relevant subsets of the population is examined in Table 5. The association between SCD and new-onset AF was similar in men and women, patients above and below 65 years of age, patients with and without diabetes mellitus or a history of ischemic heart disease, myocardial infarction, or heart failure, and patients stratified by baseline HR of 84 beats per minute or the presence or absence of Cornell product LVH on their baseline ECGs, with slightly higher risk in black as opposed to nonblack patients and in patients <65 years of age. New-onset AF was associated with significantly greater risks of SCD among patients randomized to losartan- versus atenolol-based therapy and in patients with as opposed to without LVH by Sokolow- Lyon voltage on their baseline ECGs. New-onset AF remained a statistically significant predictor of SCD after also including the interaction terms between new-onset AF and both randomized treatment assignment and baseline LVH by Sokolow- Lyon voltage in the multivariable Cox model (P<0.001). When stratifying by treatment arm, new-onset AF was a significant predictor of SCD in patients randomized to losartan-based therapy (HR, 4.19; 95% CI, ; P<0.001) but not

5 Okin et al Atrial Fibrillation and Sudden Cardiac Death 247 Sudden Death (%) No. at risk new AFib no AFib new AFib no AFib Month Figure. Survival curves illustrating the rate of sudden cardiac death in relation to development of new-onset atrial fibrillation (AFib) over time. among patients randomized to atenolol-based treatment (HR, 1.92; 95% CI, ; P=0.143) in multivariable analyses. Similarly, when stratifying by the presence or absence of LVH by Sokolow-Lyon voltage at baseline, new-onset AF was a significant predictor in patients with LVH by Sokolow-Lyon voltage (HR, 4.80; 95% CI, ; P<0.001), but not among patients without Sokolow-Lyon LVH at baseline (HR, 2.07; 95% CI, ; P=0.100). Discussion These findings demonstrate that hypertensive patients with ECG LVH who develop new AF in the course of aggressive treatment to lower blood pressure have an increased risk of SCD. The increased risk of SCD associated with new-onset AF was independent of traditional risk factors for SCD and persisted after adjusting for the possible impact of incident myocardial infarction on SCD risk and after taking into account the possible relation of SCD to in-treatment blood pressure, digoxin use, and the known predictive values of race, in-treatment heart rate, QRS duration, and ECG LVH for SCD in this population These findings suggest that new-onset AF in hypertensive patients should be considered as a new risk factor for SCD and that additional evaluation of these patients should be considered at the time of new AF. Mortality and Atrial Fibrillation Previous studies have demonstrated a strong relationship of both prevalent and new-onset AF to mortality in population-based studies, and that hypertension remains the most common comorbid condition among patients of all ages who develop AF In addition, AF has also been related to increased mortality in patients after-myocardial infarction 19,20 and in most but not all studies of heart failure patients. However, the relationship of SCD to existing AF has been less well characterized, and there have been no studies to date examining the risk of SCD solely in relation to development of new AF. The presence of AF or atrial flutter at the time of presentation with an acute myocardial infarction has been associated with an increased risk of SCD 20,29,30 during follow-up that extends out to 7 years. 20 However, in the VALsartan In Acute myocardial infarction (VALIANT) trial, 30 the relative contribution of previous AF to the prediction of SCD waned with time, contributing 12.5% of the predictive power for SCD during the in-hospital period but only 2.9% of the model power during the period from 6 months to 3 years. 30 Because new-onset AF during follow-up was not ascertained in VALIANT, 30 whether consideration of new-onset AF would further improve risk stratification for SCD could not be determined. Among postmenopausal women with documented coronary artery disease, 31 the presence of AF on a baseline or in-study ECG performed at yearly follow-up was associated with an 1.9-fold increased risk of SCD in the overall study population and a >4-fold increased SCD risk in the subset of the population with echocardiographic left ventricular ejection fraction data. Although existing or new AF was treated as a time-varying covariate in these analyses, 31 there was no distinction as to whether prevalent or new AF during follow-up was predictive of SCD. In contrast, the relationship of SCD to AF in heart failure patients has been less consistent. Among 390 consecutive patients with advanced heart failure, 24 the presence of paroxysmal or chronic AF was associated with a significantly greater SCD than sinus rhythm during mean follow-up of 236±303 days (31 versus 18%, P=0.0013). In 8178 outpatients with heart failure, 25 prevalent AF was associated with an increased risk of SCD in patients 65 years old or younger, but not in patients aged 66 to 75 or >75 years old. Finally, the presence of AF on the first Holter monitor was not associated with any significant increase in SCD in patients with mild-to-moderate heart failure enrolled in the Veterans Affairs Vasodilator- Heart Failure Trial (V-HeFT) I (25 versus 33%, P<0.64) or V-HeFT II (20 versus 20%, P<0.68). 27 Additional support for the possible role of AF in SCD can be found in the relation of AF to appropriate device therapies for ventricular arrhythmias in patients with ICDs Among 400 patients with ICDs implanted for primary prevention of SCD, 32 a history of AF was associated with a 2.6-fold increased risk of ICD-detected ventricular arrhythmia during 15±12 month mean follow-up. During 20±14 month follow-up of 229 patients, 33 those with persistent AF at the time of ICD implantation had a higher rate of appropriate device therapy for ventricular arrhythmias than those in sinus rhythm (68 versus 38%, P=0.01) and AF was a significant predictor of both appropriate (RR, 1.8; 95% CI, ) and inappropriate device therapies (RR, 2.3; 95% CI, ) in multivariate logistic regression analysis. However, patients who developed new AF during follow-up were included in the sinus rhythm group, possibly underestimating the impact of total AF burden on appropriate ICD therapies and missing the opportunity to assess this outcome in relation to new AF. Finally, in 913 consecutive patients receiving an ICD at a single center, 34 permanent AF was associated with a 70% greater risk of death and a 220% greater risk of appropriate device therapies in multivariate Cox models during 833±394 days mean follow-up. However, patients with persistent or paroxysmal AF had no increased risk of appropriate device therapies in this population. 34

6 248 Circ Arrhythm Electrophysiol April 2013 Table 5. Univariable Cox Analyses to Assess the Predictive Value of Incident Atrial Fibrillation for Sudden Cardiac Death in Relevant Subgroups of the Study Population* Subgroup The current study builds on these previous observations to demonstrate that development of new AF identifies a subset of patients with an increased adjusted risk of SCD in a population of hypertensive patients who had no history of AF and who were in sinus rhythm on their baseline study ECG. The association of new-onset AF with SCD was independent of standard risk factors for AF and SCD and the potential confounding impact of greater use of digoxin and antiarrhythmic drugs in patients who developed new AF, persisted after adjusting for other baseline and in-treatment variables that may be associated with SCD Moreover, new-onset AF was an independent predictor of all-cause mortality and remained predictive of SCD when all-cause mortality was included as a competing SCD (n) Hazard Ratio 95% CI P Value for Subgroup Analysis P Value for Interaction** Sex Male (n=4022) Female (n=4809) Race White or other (n=8313) Black (n=518) <0.001 Treatment Atenolol (n=4392) Losartan (n=4439) <0.001 Age, y <65 years (n=3412) years (n=5419) History of congestive heart failure No (n=8702) Yes (n=129) History of ischemic heart disease No (n=7474) Yes (n=1357) History of myocardial infarction No (n=8304) Yes (n=527) Diabetes mellitus No (n=7725) Yes (n=1106) Baseline heart rate (37) <84 bpm (n=7046) bpm (n=1785) Cornell product LVH on baseline ECG No (n=2907) Yes (n=5924) Sokolow-Lyon voltage LVH on baseline ECG No (n=6976) Yes (n=1855) <0.001 bpm indicates beats per minute; CI, confidence interval; ECG, electrocardiographic; LVH, left ventricular hypertrophy; and SCD, sudden cardiac death. *The hazard ratios in each subgroup were calculated from separate univariable Cox analyses in each subgroup. **P values for interaction term in Cox models between incident atrial fibrillation as a delayed-entry covariate and the subgroup variable coded as absent or present. risk. Importantly, the predictive value of new-onset AF for SCD was similar in most subsets of the population examined (Table 5). The significantly greater risk of SCD associated with new-onset AF among patients randomized to losartanversus atenolol-based therapy is of note given the previously demonstrated lower incidence of AF among patients in the losartan arm of this study. 9 Given that there were only minor differences in clinical characteristics between losartan- and atenolol-treated patients who developed new AF in the LIFE study, 9 this finding suggests that losartan-treated patients who develop new AF may differ in some fundamental but heretofore unknown respect from atenolol-treated patients who develop AF that puts them at increased risk of SCD. However,

7 Okin et al Atrial Fibrillation and Sudden Cardiac Death 249 the possibility that this association is attributable to chance cannot be completely excluded. The markedly greater risk of SCD associated with new AF in patients with as opposed to without LVH by Sokolow-Lyon product at baseline may be in part explained by the greater prevalence of Sokolow-Lyon LVH in black hypertensive patients and the higher risk of SCD in black hypertensive patients. 38 Further study of new-onset AF and SCD according to race/ethnicity and use of losartan will be needed to more fully understand these relationships. The slightly lower risk in patients aged 65 years is of note given the absence of a significant relationship of AF to SCD in heart failure patients >65 years old. 25 Lastly, the similar performance of new-onset AF in patients characterized by baseline heart rate and Cornell product LVH is relevant given the strong association of higher in-treatment heart rate and Cornell product with SCD risk. 35,37 It is interesting to speculate about possible mechanisms via which new-onset AF could mediate an increased risk of SCD. AF produces irregular ventricular activation, leading to marked short-long-short sequences of RR intervals which can trigger ventricular arrhythmias. 33 Supporting this possibility, ventricular arrhythmia onset preceded by a typical shortlong-short interval sequence in which the short-coupled beat was suggestive of a ventricular origin was significantly more common before episodes of ventricular arrhythmias requiring ICD therapy in heart failure patients with persistent AF than in those in sinus rhythm (50 versus 22%, P=0.002). 33 Enhanced vulnerability of dogs and patients with AF to programmed ventricular stimulation-induced ventricular tachycardia that did not appear to be attributable to the higher heart in AF 47 suggests that an enhanced electric instability directly related to AF could play a role in the higher incidence of SCD in relation to AF. Pro-arrhythmia from therapies for AF could also play a role in the increased risk of SCD. However, adjustment for digoxin therapy did not attenuate the adjusted risk of SCD associated with new AF in the current study. AF may contribute directly to left ventricular systolic dysfunction as a result of rapid rates, irregularity of the ventricular rhythm, or loss of atrial systole. 48 Moreover, AF with rapid ventricular rates can be a cause of increased myocardial ischemia or tachycardia-induced cardiomyopathy. 48 However, the current study takes into account possible differences of in-treatment heart rate on SCD risk. 37 Finally, development of new AF may be a marker of additional or worse underlying structural heart disease which increases the risk of SCD. Limitations There are a number of potential limitations of the current study that warrant discussion. Because ECG detection of AF (but not adverse event reports of AF) required participants to survive to the next annual ECG, it is possible that some incident AF that occurred shortly before death may have gone undetected, resulting in an underestimation of the short-term SCD risk associated with new AF. In addition, because the LIFE study did not distinguish between paroxysmal and persistent or permanent AF, the risk of SCD to AF subtype could not be assessed. Although patients with active heart failure were excluded from LIFE 39 and we adjust for a history of heart failure in our multivariate analysis, we cannot exclude the possibility that new-onset AF identifies a subset of the population with a greater burden of structural heart disease and lower systolic function. 49 The absence of data on left atrial size and left ventricular size and function in the majority of patients and the small number of SCDs among patients in the echocardiographic substudy of LIFE who were free of AF at study baseline (n=10), unfortunately preclude a meaningful evaluation of whether new-onset AF remains predictive of SCD after further adjusting for these echocardiographic measures. Use of Cornell product and Sokolow-Lyon voltage criteria for LVH to select patients for the LIFE study increased the baseline risk of the study population; as a consequence, the present findings may not be representative of hypertensive populations with different or lower risk profiles. Lastly, cause of death may be difficult to ascertain and SCD is not necessarily the same as arrhythmic sudden death. Indeed, apparent SCD in this population could possibly reflect sudden death from embolic events including undetected stroke or pulmonary embolism. Further evaluation of the specific mode of sudden death would require data not acquired in this study. It is also possible that differences in definitions of SCD and types of SCD could in part be responsible for different associations observed between this and other studies. Conclusions and Implications Thus, in the setting of aggressive blood pressure lowering, onset of AF identifies a subset of the population at significantly increased risk of SCD. The increased risk of SCD associated with new-onset AF persists after controlling for other known and suspected risk factors for SCD and for the demonstrated predictive value of race and in-treatment heart rate, QRS duration, and ECG LVH for SCD in this population These findings suggest that hypertensive patients who develop new AF may warrant further evaluation for better identification of other potentially reversible risk factors for SCD that might merit therapy. Finally, these findings suggest that prevention of the development of new AF may be a therapeutic approach to reducing SCD in the hypertensive population. Further research will be necessary to determine whether therapies targeted specifically at reducing AF incidence can reduce the risk of SCD. Sources of Funding This work was supported in part by grant COZ-368 and an Investigator Initiated Study grant from Merck & Co Inc, West Point, PA. Disclosures Dr Okin has received grant support from Merck & Co Inc, and serves on a medical advisory board for GE Medical Systems and as a consultant to Novartis. Dr Bang receives grant support from the Danish Heart Association and Interreg IVA program. Dr Wachtell has received grant support from and served on a speakers bureau for Merck & Co Inc. D. A. Hille is used by Merck & Co Inc and owns stock or stock options in Merck & Co Inc. Dr Kjeldsen receives grant support from Pronova and Astra-Zeneca, receives honoraria from Astra-Zeneca, Takeda, Medtronic, and Nycomed, and serves as a consultant to Bayer, Medtronic, Serodeus, and Takeda. Dr Dahlöf has received grant support from Merck & Co Inc, Novartis, Pfizer, Krka, and Boehringer-Ingelheim, has an ownership interest in Mintage

8 250 Circ Arrhythm Electrophysiol April 2013 scientific, and serves as a consultant or on scientific advisory boards for Bayer, Boehringer-Ingelheim, Novartis, and Vicore Pharma. Dr Devereux has served as a consultant for on advisory boards for Merck & Co Inc and Novartis. References 1. Roger VL, Go AS, Lloyd-Jones DM, Benjamin EJ, Berry JD, Borden WB, Bravata DM, Dai S, Ford ES, Fox CS, Fullerton HJ, Gillespie C, Hailpern SM, Heit JA, Howard VJ, Kissela BM, Kittner SJ, Lackland DT, Lichtman JH, Lisabeth LD, Makuc DM, Marcus GM, Marelli A, Matchar DB, Moy CS, Mozaffarian D, Mussolino ME, Nichol G, Paynter NP, Soliman EZ, Sorlie PD, Sotoodehnia N, Turan TN, Virani SS, Wong ND, Woo D, Turner MB; American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Heart disease and stroke statistics 2012 update: a report from the American Heart Association. Circulation. 2012;125:e2 e Fishman GI, Chugh SS, Dimarco JP, Albert CM, Anderson ME, Bonow RO, Buxton AE, Chen PS, Estes M, Jouven X, Kwong R, Lathrop DA, Mascette AM, Nerbonne JM, O Rourke B, Page RL, Roden DM, Rosenbaum DS, Sotoodehnia N, Trayanova NA, Zheng ZJ. Sudden cardiac death prediction and prevention: report from a National Heart, Lung, and Blood Institute and Heart Rhythm Society Workshop. Circulation. 2010;122: Chugh SS, Jui J, Gunson K, Stecker EC, John BT, Thompson B, Ilias N, Vickers C, Dogra V, Daya M, Kron J, Zheng ZJ, Mensah G, McAnulty J. Current burden of sudden cardiac death: multiple source surveillance versus retrospective death certificate-based review in a large U.S. community. J Am Coll Cardiol. 2004;44: Adabag AS, Luepker RV, Roger VL, Gersh BJ. Sudden cardiac death: epidemiology and risk factors. Nat Rev Cardiol. 2010;7: Chugh SS. Early identification of risk factors for sudden cardiac death. Nat Rev Cardiol. 2010;7: Chugh SS, Reinier K, Teodorescu C, Evanado A, Kehr E, Al Samara M, Mariani R, Gunson K, Jui J. Epidemiology of sudden cardiac death: clinical and research implications. Prog Cardiovasc Dis. 2008;51: Go AS, Hylek EM, Phillips KA, Chang Y, Henault LE, Selby JV, Singer DE. Prevalence of diagnosed atrial fibrillation in adults: national implications for rhythm management and stroke prevention: the AnTicoagulation and Risk Factors in Atrial Fibrillation (ATRIA) Study. JAMA. 2001;285: Lloyd-Jones DM, Wang TJ, Leip EP, Larson MG, Levy D, Vasan RS, D Agostino RB, Massaro JM, Beiser A, Wolf PA, Benjamin EJ. Lifetime risk for development of atrial fibrillation: the Framingham Heart Study. Circulation. 2004;110: Wachtell K, Lehto M, Gerdts E, Olsen MH, Hornestam B, Dahlöf B, Ibsen H, Julius S, Kjeldsen SE, Lindholm LH, Nieminen MS, Devereux RB. Angiotensin II receptor blockade reduces new-onset atrial fibrillation and subsequent stroke compared to atenolol: the Losartan Intervention For End Point Reduction in Hypertension (LIFE) study. J Am Coll Cardiol. 2005;45: Krahn AD, Manfreda J, Tate RB, Mathewson FA, Cuddy TE. The natural history of atrial fibrillation: incidence, risk factors, and prognosis in the Manitoba Follow-Up Study. Am J Med. 1995;98: Verdecchia P, Reboldi G, Gattobigio R, Bentivoglio M, Borgioni C, Angeli F, Carluccio E, Sardone MG, Porcellati C. Atrial fibrillation in hypertension: predictors and outcome. Hypertension. 2003;41: Benjamin EJ, Levy D, Vaziri SM, D Agostino RB, Belanger AJ, Wolf PA. Independent risk factors for atrial fibrillation in a population-based cohort. The Framingham Heart Study. JAMA. 1994;271: Vaziri SM, Larson MG, Benjamin EJ, Levy D. Echocardiographic predictors of nonrheumatic atrial fibrillation. The Framingham Heart Study. Circulation. 1994;89: Psaty BM, Manolio TA, Kuller LH, Kronmal RA, Cushman M, Fried LP, White R, Furberg CD, Rautaharju PM. Incidence of and risk factors for atrial fibrillation in older adults. Circulation. 1997;96: Benjamin EJ, Wolf PA, D Agostino RB, Silbershatz H, Kannel WB, Levy D. Impact of atrial fibrillation on the risk of death: the Framingham Heart Study. Circulation. 1998;98: Miyasaka Y, Barnes ME, Bailey KR, Cha SS, Gersh BJ, Seward JB, Tsang TS. Mortality trends in patients diagnosed with first atrial fibrillation: a 21-year community-based study. J Am Coll Cardiol. 2007;49: Conen D, Chae CU, Glynn RJ, Tedrow UB, Everett BM, Buring JE, Albert CM. Risk of death and cardiovascular events in initially healthy women with new-onset atrial fibrillation. JAMA. 2011;305: Piccini JP, Hammill BG, Sinner MF, Jensen PN, Hernandez AF, Heckbert SR, Benjamin EJ, Curtis LH. Incidence and prevalence of atrial fibrillation and associated mortality among Medicare beneficiaries, Circ Cardiovasc Qual Outcomes. 2012;5: Jabre P, Roger VL, Murad MH, Chamberlain AM, Prokop L, Adnet F, Jouven X. Mortality associated with atrial fibrillation in patients with myocardial infarction: a systematic review and meta-analysis. Circulation. 2011;123: Berton G, Cordiano R, Cucchini F, Cavuto F, Pellegrinet M, Palatini P. Atrial fibrillation during acute myocardial infarction: association with all-cause mortality and sudden death after 7-year of follow-up. Int J Clin Pract. 2009;63: Dries DL, Exner DV, Gersh BJ, Domanski MJ, Waclawiw MA, Stevenson LW. Atrial fibrillation is associated with an increased risk for mortality and heart failure progression in patients with asymptomatic and symptomatic left ventricular systolic dysfunction: a retrospective analysis of the SOLVD trials. J Am Coll Cardiol. 1998;32: Wang TJ, Larson MG, Levy D, Vasan RS, Leip EP, Wolf PA, D Agostino RB, Murabito JM, Kannel WB, Benjamin EJ. Temporal relations of atrial fibrillation and congestive heart failure and their joint influence on mortality: the Framingham Heart Study. Circulation. 2003;107: Rivero-Ayerza M, Scholte Op Reimer W, Lenzen M, Theuns DA, Jordaens L, Komajda M, Follath F, Swedberg K, Cleland JG. New-onset atrial fibrillation is an independent predictor of in-hospital mortality in hospitalized heart failure patients: results of the EuroHeart Failure Survey. Eur Heart J. 2008;29: Middlekauff HR, Stevenson WG, Stevenson LW. Prognostic significance of atrial fibrillation in advanced heart failure: a study of 390 patients. Circulation. 1991;84: Baldasseroni S, Orso F, Fabbri G, De Bernardi A, Cirrincione V, Gonzini L, Fumagalli S, Marchionni N, Midi P, Maggioni AP; Italian Network on Congestive Heart Failure Investigators. Age-dependent prognostic significance of atrial fibrillation in outpatients with chronic heart failure: data from the Italian Network on Congestive Heart Failure Registry. Cardiology. 2010;116: Swedberg K, Olsson LG, Charlesworth A, Cleland J, Hanrath P, Komajda M, Metra M, Torp-Pedersen C, Poole-Wilson P. Prognostic relevance of atrial fibrillation in patients with chronic heart failure on long-term treatment with beta-blockers: results from COMET. Eur Heart J. 2005;26: Carson PE, Johnson GR, Dunkman WB, Fletcher RD, Farrell L, Cohn JN. The influence of atrial fibrillation on prognosis in mild to moderate heart failure. The V-HeFT Studies. The V-HeFT VA Cooperative Studies Group. Circulation. 1993;87(6 Suppl):VI102 VI Crijns HJ, Tjeerdsma G, de Kam PJ, Boomsma F, van Gelder IC, van den Berg MP, van Veldhuisen DJ. Prognostic value of the presence and development of atrial fibrillation in patients with advanced chronic heart failure. Eur Heart J. 2000;21: Pedersen OD, Abildstrøm SZ, Ottesen MM, Rask-Madsen C, Bagger H, Køber L, Torp-Pedersen C; TRACE Study Investigators. Increased risk of sudden and non-sudden cardiovascular death in patients with atrial fibrillation/flutter following acute myocardial infarction. Eur Heart J. 2006;27: Piccini JP, Zhang M, Pieper K, Solomon SD, Al-Khatib SM, Van de Werf F, Pfeffer MA, McMurray JJ, Califf RM, Velazquez EJ. Predictors of sudden cardiac death change with time after myocardial infarction: results from the VALIANT trial. Eur Heart J. 2010;31: Deo R, Vittinghoff E, Lin F, Tseng ZH, Hulley SB, Shlipak MG. Risk factor and prediction modeling for sudden cardiac death in women with coronary artery disease. Arch Intern Med. 2011;171: Deneke T, Lawo T, Reinecke J, Buttler C, Grewe PH, Shin DI, Gerritse B, Mügge A, Lemke B, Kloppe A. Predictors of sustained ventricular arrhythmia episodes in patients with primary ICD indication: male gender and AF in primary ICD prophylaxis. Herzschrittmacherther Elektrophysiol. 2011;22: Grönefeld GC, Mauss O, Li YG, Klingenheben T, Hohnloser SH. Association between atrial fibrillation and appropriate implantable cardioverter defibrillator therapy: results from a prospective study. J Cardiovasc Electrophysiol. 2000;11: Borleffs CJ, van Rees JB, van Welsenes GH, van der Velde ET, van Erven L, Bax JJ, Schalij MJ. Prognostic importance of atrial fibrillation in implantable cardioverter-defibrillator patients. J Am Coll Cardiol. 2010;55: Wachtell K, Okin PM, Olsen MH, Dahlöf B, Devereux RB, Ibsen H, Kjeldsen SE, Lindholm LH, Nieminen MS, Thygesen K. Regression of electrocardiographic left ventricular hypertrophy during antihypertensive

9 Okin et al Atrial Fibrillation and Sudden Cardiac Death 251 therapy and reduction in sudden cardiac death: the LIFE Study. Circulation. 2007;116: Morin DP, Oikarinen L, Viitasalo M, Toivonen L, Nieminen MS, Kjeldsen SE, Dahlöf B, John M, Devereux RB, Okin PM. QRS duration predicts sudden cardiac death in hypertensive patients undergoing intensive medical therapy: the LIFE study. Eur Heart J. 2009;30: Okin PM, Hille DA, Kjeldsen SE, Julius S, Dahlöf B, Devereux RB. Sudden cardiac death in relation to changing heart rate during treatment of hypertension: the LIFE study. J Am Coll Cardiol. 2010;55:A140, E Okin PM, Kjeldsen SE, Julius S, Dahlöf B, Devereux RB. Racial differences in sudden cardiac death among hypertensive patients during antihypertensive therapy: the LIFE study. Heart Rhythm. 2012;9: Dahlöf B, Devereux R, de Faire U, Fyhrquist F, Hedner T, Ibsen H, Julius S, Kjeldsen S, Kristianson K, Lederballe-Pedersen O, Lindholm LH, Nieminen MS, Omvik P, Oparil S, Wedel H. The Losartan Intervention For Endpoint reduction (LIFE) in Hypertension study: rationale, design, and methods. The LIFE Study Group. Am J Hypertens. 1997;10(7 Pt 1): Dahlöf B, Devereux RB, Kjeldsen SE, Julius S, Beevers G, de Faire U, Fyhrquist F, Ibsen H, Kristiansson K, Lederballe-Pedersen O, Lindholm LH, Nieminen MS, Omvik P, Oparil S, Wedel H; LIFE Study Group. Cardiovascular morbidity and mortality in the Losartan Intervention For Endpoint reduction in hypertension study (LIFE): a randomised trial against atenolol. Lancet. 2002;359: Okin PM, Roman MJ, Devereux RB, Kligfield P. Electrocardiographic identification of increased left ventricular mass by simple voltage-duration products. J Am Coll Cardiol. 1995;25: Sokolow M, Lyon TP. The ventricular complex in left ventricular hypertrophy as obtained by unipolar precordial and limb leads. Am Heart J. 1949;37: Okin PM, Wachtell K, Devereux RB, Harris KE, Jern S, Kjeldsen SE, Julius S, Lindholm L, Nieminen MS, Edelman JM, Dahlöf B. Regression of electrocardiographic left ventricular hypertrophy and decreased incidence of new-onset atrial fibrillation: the LIFE Study. JAMA. 2006; 296: Andersen PK, Gill R.D. Cox s Regression Model for Counting Processes: A Large Sample Study. Annals Statistics. 1982;10: Scheike TH. Analyzing competing risk data using R timesreg package. J Statistical Software. 2011;38: Snapinn SM, Jiang Q, Iglewicz B. Illustrating the impact of a time-varying covariate with an extended Kaplan-Meier estimate. The American Statistician. 2005;59: Somberg JC, Torres V, Keren G, Butler B, Tepper D, Kleinbaum H, Molnar J. Enhancement of myocardial vulnerability by atrial fibrillation. Am J Ther. 2004;11: Balasubramaniam R, Kistler PM. Atrial fibrillation in heart failure: the chicken or the egg? Heart. 2009;95: Wachtell K, Gerdts E, Aurigemma GP, Boman K, Dahlöf B, Nieminen MS, Olsen MH, Okin PM, Palmieri V, Rokkedal JE, Devereux RB. Intreatment reduced left atrial diameter during antihypertensive treatment is associated with reduced new-onset atrial fibrillation in hypertensive patients with left ventricular hypertrophy: The LIFE Study. Blood Press. 2010;19: Clinical Perspective Among 8831 hypertensive patients with electrocardiographic left ventricular hypertrophy with no history of atrial fibrillation (AF), in sinus rhythm on their baseline electrocardiograph, and randomly assigned to losartan- or atenolol-based treatment, new-onset AF developed in 701 patients (7.9%) and sudden cardiac death (SCD) in 151 patients (1.7%) during 4.7±1.1 years mean follow-up. In multivariate Cox analyses adjusting for age, sex, race, diabetes mellitus, history of heart failure, myocardial infarction, ischemic heart disease, stroke, smoking, serum high-density lipoprotein cholesterol, creatinine, glucose, and urine albumin/creatinine ratio as standard risk factors, and for incident myocardial infarction, in-treatment use of digoxin, systolic and diastolic pressure, heart rate, QRS duration, Cornell voltage-duration product, and Sokolow-Lyon voltage left ventricular hypertrophy treated as time-varying covariates, new-onset AF remained associated with a >3-fold increased risk of SCD (hazard ratio, 3.13; 95% confidence interval, ; P<0.001). These findings suggest that hypertensive patients who develop new AF may warrant further evaluation for better identification of other potentially modifiable risk factors for SCD that might merit therapy. These findings further suggest that prevention of the development of new AF may be a therapeutic approach to reducing SCD in the hypertensive population. Further research will be necessary to determine whether therapies aimed at reducing AF incidence can reduce the risk of SCD.