© 2004 by European Society of Cardiology
Mortality of patients with implanted cardioverter/defibrillators in relation to episodes of atrial fibrillation
aBerufsgenossenschaftliche Kliniken Bergmannsheil Bochum, Germany; bBakken Research Center Maastricht, The Netherlands
Manuscript submitted 27 February 2003. Accepted after revision 23 November 2003.
*Corresponding author. Department of Cardiology, "Bergmannsheil Bochum", Bürkle-de-la-Camp-Platz 1, 44789 Bochum, Germany. Tel.: +49-234-302-6077; fax: +49-234-302-6084. E-mail address: thomas.deneke{at}ruhr-uni-bochum.de (T.Deneke).
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionConclusionsAppendixReferences |
|---|
AIMS: This study was undertaken to evaluate if a history of episodes of atrial fibrillation (AF) has an effect on survival in patients (pts) with implanted cardioverter defibrillators (ICD).
METHODS AND RESULTS: Retrospective analysis of baseline characteristics, medication and endpoints in pts receiving a dual-chamber ICD comparing pts without (group 1: n=291) and with a history of AF (group 2: n=68).
Pts in groups 1 and 2 did not differ with regard to baseline characteristics. Forty-one pts (8.7%) experienced a detected episode of AF.
During the study period (mean follow-up 9.5 months) 25 pts died (5.8% of group 1 vs 11.8% of group 2). Patient survival was found to be 97.9% vs 91.2% (n=279 vs 63) at 3 months after implantation, 97.5% vs 89.7% (n=244 vs 54) at 6 months, and 93.7% vs 86.5% (n=96 vs 28) at 12 months in group 1 vs group 2 (p=0.005).
The risk of death increased after a first episode of AF was experienced during follow-up (RR=5.0, p=0.0007). Independent predictors of mortality are NYHA class (p=0.002), history of AF (p=0.009 for predicting early mortality), and detected episodes of AF during follow-up (p=0.001).
CONCLUSION: ICD-patients with a history of AF had a higher than average risk of death early after implantation and after device-detected episodes of AF.
Key Words: recurrent atrial fibrillation, implantable defibrillators, mortality
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsAppendixReferences |
|---|
Atrial fibrillation (AF) can be found in up to 40% of patients (pts) with symptomatic heart failure [1–
3]. AF requires treatment for symptoms such as palpitations and tachycardias and is also associated with a 5-fold risk for thromboembolic complications if no anticoagulation therapy is prescribed [4–7].
The data on AF as an independent risk factor for impaired survival still remain controversial [2,8]. The presence of chronic AF in patients with left ventricular systolic dysfunction confers a greater risk of death in the SOLVED trial, but the V-HeFT I and II trials document no statistically significant difference in survival in patients with or without AF [9–13], even though it appears that "increasing pump failure begets AF" and "AF begets an increase in systolic dysfunction" [2].
In a study of patients receiving dual-chamber implantable defibrillators (ICD) (Medtronic model 7271 GEM DR) we observed that patients with a history of intermittent AF had a higher risk of dying than patients without a history of AF. We analyzed the mortality of these patients in greater detail in order to assess the influence of recurrent (episodes of atrial fibrillation that are either terminated spontaneously or can be converted to SR including chronic paroxysmal and chronic permanent) AF as a risk factor for mortality.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsAppendixReferences |
|---|
During the Clinical Evaluation Study of the Medtronic model 7271 GEM DR ICDTM, 359 patients were included in Europe (344 pts in 40 centres), Australia (12 pts in 2 centres) and New Zealand (3 pts in 1 centre). Patients were retrospectively divided into 291 patients without a history of AF (group 1) and 68 patients with a history of AF (group 2).
Patients were followed at least 6 months post-implantation applying a standard ICD follow-up with device interrogation and analysis of recorded arrhythmia episodes. Arrhythmia episodes detected by the ventricular rate criteria of the device were recorded. These include episodes that received a therapy from the device as well as episodes that did not receive a therapy when the PR LogicTM algorithm identified the rhythm as supraventricular. For all episodes the investigators indicated the appropriateness of the detection. Whenever the device classification was considered incorrect the investigator classified the rhythm. Supraventricular episodes are classified by the device as AF, atrial flutter (AFlut) or 1:1 supraventricular tachyarrhythmia based on cycle length, regularity and A–V sequences.
Information on concomitant medication (medication questionnaire with check boxes) was collected at baseline and at the 6 month follow-up visit.
Statistics
The continuous variables have been compared by means of a t-test, the categorial variables except NYHA class by means of a Fisher exact test and the NYHA class by means of a Mantel–Haenszel test for trend. A p-value of <0.05 was considered statistically significant.
The time to the first episode of AF or AFlut has been analyzed using Kaplan–Meier survival analysis method. To find baseline variables that predict the occurrence of episodes of supraventricular tachycardias a Cox proportional hazard model with backward variable selection was used.
Patient survival was estimated using Kaplan–Meier survival analysis and confidence limits have been calculated on cumulative hazard scale, from Greenwood's variance estimate and with a modified lower bound [14]. Differences in mortality were compared using the log-rank test (a p-value of <0.05 was considered significant). Cox proportional hazard models have been used to describe the relation between mortality and baseline variables. The estimated risk profile of the patient subgroups has been plotted using a smoothed hazard plot and history of AF was included in a Cox model as a time-varying risk factor [15]. The additional risk a patient has from the moment of a first detected episode of AF or AFlut onwards was analyzed by means of a Cox proportional hazard model with a time-varying co-variate.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsAppendixReferences |
|---|
In the overall patient collective 68 (18.9%) patients with a history of recurrent AF (group 2) were identified.
Patient baseline characteristics
There was no statistically significant difference in the patient baseline characteristics between the two groups (see Table 1). Indication for ICD implantation was survival of sudden cardiac death in 19.9% of patients in group 1 and 27.9% of patients in group 2, ventricular tachycardia in 43.3% and 33.8%, survived sudden cardiac death plus ventricular tachycardia in 32.0% and 36.8% and MADIT I criteria [16] in 4.8% and 1.5%, respectively (p>0.05).
|
Mortality
A total of 25 patients died during the follow-up period (9.5±3.8 months per patient) including 8 patients from group 2. Relative mortality was 5.8% in group 1 (17 deaths in 291 patients) and 11.8% in group 2 (8 deaths in 68 patients). Overall 6- and 12-month survival rates were 96.0% and 92.2%, respectively. Cause of death is reported in Table 2. Patients who died had a higher NYHA class (p=0.002) and, more frequently, heart failure as defined by NYHA class III/IV and an ejection fraction below 35% (p=0.02) (see Table 2).
|
Survival rates in groups 1 and 2, respectively, were at 1 month 99.0% (n=284) vs 95.6% (n=66), at 3 months 97.9% (n=279) and 91.2% (n=63), at 6 months 97.5% (n=244) and 89.7% (n=54) and at 12 months 93.7% (n=96) vs 86.5% (n=28). There is a significantly higher mortality in the AF group when follow-up is truncated after 3 months (p=0.005) but not with complete follow-up included (p=0.10) (see Fig. 1).
|
Risk factors for mortality
The Cox proportional hazard model has been fitted using backward variable selection. From baseline variables included in the hazard model only NYHA class was detected to be significantly related to mortality, with a relative risk of 3.1 (p=0.0008) for every increase in class. Baseline history of recurrent AF is not significantly associated with a proportional increase in mortality over the whole follow-up period.
Percentage of pacing was comparable in both groups as assessed at the first follow-up visit more than 6 months after implantation. Ventricular pacing was documented in 58.9±39.5% of the time in group 1 and 68.5±38.6% of the time in group 2 (p=0.075). Adding percentage ventricular pacing to the proportional hazard models reveals no qualitative difference and percentage ventricular pacing is found not to be a significant risk factor for mortality.
AF and mortality
We compared the risk of death as a function of time for patients in groups 1 and 2. Fig. 2 presents a smoothed hazard plot [15] displaying the time since implant vs risk of death.
|
Patients in group 2 had a higher than average risk of death shortly after implantation (first 5 months). At 3 months post-implant the risk for patients with intermittent AF (group 2) is approximately 5 times as high as for patients without intermittent AF (group 1). The risk decreases almost linearly over time until it is low from 7 months onwards. This reflects the fact that 7 out of the 8 deaths in group 2 occurred within 5 months after implantation, whereas in group 1 this was 6 out of 18.
A time-varying risk factor was defined for history of AF that was set to 1 at implant, which decreases linearly and is 0 from 7 months onwards. A Cox-model reveals NYHA class (p=0.0008) and the time-varying factor for history of intermittent AF (p=0.0007) as predictive factors of higher mortality.
AF episodes during follow-up
During the study period 244 episodes of AF (paroxysmal and persistent AF) or atrial flutter (AFlut) triggering the ventricular rate criteria of the ICD were detected in 41 patients (11%), 36.6% of these patients had a documented history of atrial fibrillation. A total of 22.1% of the patients in group 2 experienced episodes of AF or AFlut detected by the device and 8.3% of the patients in group 1 without a documented history of AF or AFlut (new-onset AF or AFlut). Predictive for the occurrence of atrial tachyarrhythmia episodes were female gender (relative risk 2.2, p=0.02) and history of atrial fibrillation (relative risk 2.8, p=0.002).
The Cox proportional hazard model with a time-varying co-variate reveals that patients, who had experienced episodes of atrial tachycardia during the study, had an increased risk of death from that moment onwards (relative risk 5.0, p=0.0007). Of the 25 patients who died during the study period, 7 (28%) had experienced episodes of AF during the trial. Two died of cerebral vascular accidents and 5 of heart failure.
Also in a multivariate analysis the time-varying variable of previously detected episodes of AF or AFlut was found to be a significant predictor of mortality (p=0.001).
Medication
A total of 71.8% of patients in group 1 and 73.5% of patients in group 2 took an antiarrhythmic drug (p=0.881) at baseline and 72.6% vs 79.7% (p=0.327) at the 6 month visit. Chronic heart failure medication (including digoxin, ace-inhibitor, diuretics) was comparable between the two groups (see Table 3).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsAppendixReferences |
|---|
The role of AF as a predictor of adverse clinical outcome in patients with heart disease is still under debate [17]
. Whereas chronic persistent AF has been interpreted as an independent risk factor for impaired survival in patients with left ventricular systolic dysfunction, the role of intermittent AF (chronic paroxysmal and persistent AF) has not been previously studied. In this retrospective analysis of patients with an established ICD indication and mainly mild left ventricular dysfunction (mean ejection fraction of 38.8±15.6%), intermittent AF pre-implantation is associated with a significantly increased risk of death in the first 5 months after ICD implantation. The risk of death decreases with duration post-implantation and is low 7 months post-implant. The risk stratification in this study can only be an estimation based on the limited experience of a total of 25 deaths. At the end of the follow-up period of 1 year the difference between the mortality rates in the two studied subgroups no longer reached statistical significance. Our data are in accordance with data documented by Dries et al. (SOLVED) or by Middlekauff et al., who found that chronic AF was independently associated with an increase in mortality in patients with impaired left ventricular function [2,3]. In patients without structural heart disease AF was found to be associated with excessive mortality independent of patient age or gender [16,18,19]. The relative risk of all-cause mortality is ranging from 1.3 to 2.6 in different trials [20,21]. In our study on patients with recurrent (paroxysmal and chronic persistent) AF we found up to 5-fold risk of death early after the inclusion decreasing slowly over time. A speculative mechanism may be that there is a group of patients at high risk of death and these patients do die early after the ICD implantation. Interestingly, Middlekauff and coworkers found AF to predict higher rates of sudden cardiac death only in patients with mild heart failure (documented by moderately elevated enddiastolic pressures under therapy) but not in patients with severe chronic heart failure.
Episodes of atrial tachyarrhythmia
A total of 8.3% of our patients without known intermittent AF at baseline were found to experience a detected (i.e. tachyarrhythmia triggering the ventricular rate algorithm of the device) episode of AF/AFlut during a mean follow-up of 9 months indicating the spontaneous occurrence rate of intermittent AF or AFlut in this patient group.
In our analysis the occurrence of detected episodes of AF or AFlut was significantly correlated with impaired survival. The relative risk of death added up, from the moment the first detected episode was documented, to be 5 times as high as in patients without such an episode. As found in a CHF-STAT sub-study patients with intermittent AF had a lower mortality when compared with those in permanent AF not convertible to SR by oral amiodarone therapy [22].
The occurrence of AF may be a marker of increased severity of left ventricular dysfunction indicating a degree of myocardial damage (fibrosis) beyond that indicated by measures of ejection fraction. Therefore, the occurrences of episodes of AF may themselves be only bystanding problems [2,22]. This phenomenon of increasing prevalence of AF with impairment of left ventricular function is well known and may contribute to the poor outcome of patients developing new-onset AF [3]. When considering our data though the question remains why the higher risk of these patients then diminishes during the follow-up? It may be explained by the hypothesis that the patients at highest risk of death die early in the follow-up phase.
All episodes analyzed in this study were atrial tachycardias triggering the ventricular rate criteria of the ICD. Episodes of AF with slow or normal ventricular response were not detected and were, therefore, not included in the analysis. In our study about two thirds of the patients were treated with at least 1 antiarrhythmic drug that is known to suppress AF and slow down ventricular response reducing the number and ventricular rate of atrial tachycardia episodes.
Mechanisms of increased mortality in patients with AF
As documented in this study, a history of intermittent AF appears as a time-varying risk factor of mortality in patients with ventricular arrhythmias. The mechanisms involved in the higher mortality in these patients were mainly increasing pump failure and thromboembolic cerebral accidents. Sudden cardiac death was ruled out as a cause of death after ICD implantation. Dries et al. discussed AF as not being associated with the risk for an arrhythmic death but as an independent risk factor of all-cause mortality and pump-failure death. On the other hand in a study by Middlekauff and coworkers AF was associated with an increase in mortality due to arrhythmic, unexpected death [18].
Thromboembolic events are known to occur more often in patients in AF compared with patients in SR [18]. In our study about 30% of the patients with a history of AF were anticoagulated. Even though this seems low it is comparable with data suggesting about one third of the patients with AF receive anticoagulation in the United States [23]. In our study 2 patients died due to cerebro-vascular complications in the group of patients with a history of AF. In the group of patients with a detected episode of AF or AFlut at follow-up about 30% were on anticoagulants at follow-up.
Antiarrhythmic medication
In this study patients without or with a history of AF did not differ with regard to antiarrhythmic therapy. Amiodarone was administered in one third and ß-blockers in more than 20% of the patients in both groups. There was no significant difference with regard to any medical therapy except aspirin given more often in patients without a history of AF.
Other risk factors of mortality
In addition to (1) the history of AF and (2) episodes of AF or AFlut during follow-up being independent time-varying risk factors of mortality, only higher NYHA class was significantly associated with higher mortality. NYHA class as a clinical parameter of severity of functional impairment is known to depend on individual patient characteristics. In our analysis NYHA class was comparable in the two groups and was mostly found to reflect mild ventricular dysfunction (NYHA class II) in 61% of all patients. Percentage ventricular pacing was high in the observed study population (almost 61% of the time in the overall group) but was found to be comparable in the two groups. Percentage ventricular pacing was found not to be a predictive factor for mortality in our patient cohort.
|
Conclusions |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsAppendixReferences |
|---|
- Intermittent atrial fibrillation occurs in about 20% of patients with an indication for ICD implantation and mild left ventricular dysfunction.
- A history of episodes of atrial fibrillation (chronic paroxysmal and persistent) in patients with an ICD indication is associated with a higher than normal risk of death early after the implantation. This risk seems to decrease over the duration of survival.
- An ICD documented episode of atrial tachycardia (AF or AFlut) during follow-up is associated with an additional risk of death from that moment on.
- NYHA class seems to be an independent predictor of mortality in patients with implanted ICDs.
Further prospective trials on larger number of patients are needed to stratify the course of risk of death with regard to history or documented episodes of atrial fibrillation and identify patients at higher risk. The effect of anticoagulation, ß-blockers and amiodarone needs to be evaluated in patients with recurrent atrial fibrillation.
|
Appendix |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsAppendixReferences |
|---|
|
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsAppendixReferences |
|---|
[1] Stevenson W.G. and Stevenson L.W. Atrial fibrillation in heart failure. N Engl J Med 1999; 341: 910–911.
[2] Dries D.L., Exner D.V., Gersh B.J., et al. 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: 695–703.
[3] Middlekauff H.R., Stevenson W.G., Stevenson L.W. Prognostic significance of atrial fibrillation in advanced heart failure: a study of 390 patients. Circulation 1991; 84: 40–48.
[4] The AFFIRM Investigators. Atrial fibrillation following investigation of rhythm management (AFFIRM) study design. Am J Cardiol 1997; 79: 1198–1202.[CrossRef][Web of Science][Medline]
[5] Nattel S. Newer developments in the management of atrial fibrillation. Am Heart J 1995; 130: 1094–1106.[CrossRef][Web of Science][Medline]
[6] Sopher S.M. and Camm A.J. Atrial fibrillation: maintenance of sinus rhythm versus rate control. Am J Cardiol 1996; 77: 24A–37A.[CrossRef][Medline]
[7] Stafford R.S. and Singer D.E. Recent national patterns of warfarin use in atrial fibrillation. Circulation 1998; 97: 1231–1233.
[8] Carson P.E., Johnson G.R., Dunkman W.B., et al. The influence of atrial fibrillation on prognosis in mild to moderate heart failure. The V-HeFT Studies. Circulation 1993; Suppl_VI VI 102–VI 110.
[9] Grogan M., Smith H.C., Gersh B.J., Wood D.W. Left ventricular dysfunction due to atrial fibrillation in patients initially believed to have idiopathic dilated cardiomyopathy. Am J Cardiol 1992; 69: 1570–1575.[CrossRef][Web of Science][Medline]
[10] Shinbane J.S., Wood M.A., Jensen D.N., et al. Tachycardia-induced cardiomyopathy: a review of animal models and clinical studies. J Am Coll Cardiol 1997; 29: 709–715.[Abstract]
[11] Clark D.M., Plumb V.J., Epstein A.E., Kay G.N. Hemodynamic effects of irregular sequence of ventricular cycle lengths during atrial fibrillation. J Am Coll Cardiol 1997; 30: 1039–1045.[Abstract]
[12] Skinner N.S., Mitchell J.H., Wallace A.G., Sarnoff S.J. Hemodynamic consequences of atrial fibrillation at constant ventricular rates. Am J Med 1964; 36: 342–350.[Medline]
[13] Stevenson W.G., Stevenson L.W., Middlekauff H.R., et al. Improving survival for patients with atrial fibrillation and advanced heart failure. J Am Coll Cardiol 1996; 28: 1458–1463.[Abstract]
[14] S-Plus 2000 guide to statistics 1999; Mathsoft Inc vol. 2: pp. 238–240.
[15] Allison P.D. Survival analysis using the SAS system—a practical guide 1995; SAS Institute Inc.
[16] Multicenter automatic defibrillator implantation trial (MADIT): design and clinical protocol. MADIT Executive Committee. Pacing Clin Electrophysiol 1991; 14: 920–927.[CrossRef][Medline]
[17] Fuster V., Rydén L.E., Asinger R.W., et al. ACC/AHA/ESC guidelines for the management of patients with atrial fibrillation: executive summary. A report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines and the European Society of Cardiology Committee for Practice Guidelines and Policy Conferences. Circulation 2001; 104: 2118–2150.
[18] Dries D.L., Rosenber Y.D., Waclawiw M.A., Domanski M.J. Ejection fraction and the risk of thromboembolic events in patients with systolic dysfunction and sinus rhythm: evidence for gender differences in the SOLVED trial. J Am Coll Cardiol 1997; 29: 1074–1080.[Abstract]
[19] Benjamin E.J., Wolf P.A., D'Agostino R.B., et al. Impact of atrial fibrillation on the risk of death. The Framingham Heart Study. Circulation 1998; 98: 946–952.
[20] Godtfredsen J. Atrial fibrillation: course and prognosis: a follow-up study of 1212 cases. In Kulbertus H.E., Olsson S.B., Schlepper M. (Eds.). Atrial fibrillation 1982; Molndal, Sweden AB Hassle pp. 134–145.
[21] Krahn A.D., Manfreda J., Tate R.B., et al. The natural history of atrial fibrillation: incidence, risk factors, and prognosis in the Manitoba follow-up study. Am J Med 1995; 98: 476–484.[CrossRef][Web of Science][Medline]
[22] Flegel K.M., Shipley M.J., Rose G. Risk of stroke in non-rheumatic atrial fibrillation. Lancet 1987; 1: 526–529.[CrossRef][Web of Science][Medline]
[23] Deedwania P.C., Singh B.M., Ellenbogen K., et al. for the Department of Veterans Affairs CHF-STAT Investigators. Spontaneous conversion and maintenance of sinus rhythm by amiodarone in patients with heart failure and atrial fibrillation. Observations from the Veterans Affairs Congestive Heart Failure Survival Trial of Antiarrhythmic Therapy (VHF-STAT). Circulation 1998; 98: 2574–2579.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||