Author + information
- Received May 24, 2007
- Revision received July 10, 2007
- Accepted July 24, 2007
- Published online November 27, 2007.
- Stefan H. Hohnloser, MD, FACC⁎,⁎ (, )
- Dimitri Pajitnev, MD⁎,
- Janice Pogue, PhD†,
- Jeff S. Healey, MD†,
- Marc A. Pfeffer, MD, FACC‡,
- Salim Yusuf, MD, FACC†,
- Stuart J. Connolly, MD, FACC†,
- ACTIVE W Investigators
- ↵⁎Reprint requests and correspondence:
Dr. Stefan H. Hohnloser, J. W. Goethe University, Department of Cardiology, Division of Electrophysiology, Theodor-Stern-Kai 7, 60590 Frankfurt, Germany.
Objectives Our goal was to determine the risk of stroke or non-cerebral embolism associated with paroxysmal compared with sustained atrial fibrillation (AF).
Background The risk of stroke and non-cerebral embolism and the efficacy of oral anticoagulation (OAC) in paroxysmal AF as compared with sustained AF are not precisely known.
Methods The ACTIVE W (Atrial Fibrillation Clopidogrel Trial With Irbesartan for Prevention of Vascular Events) was a trial comparing OAC to combined antiplatelet therapy with aspirin and clopidogrel for prevention of vascular events in 6,706 AF patients. The incidence of thromboembolic events and major bleeds were compared in patients with paroxysmal AF (n = 1,202) and persistent or permanent AF (n = 5,495).
Results Patients with paroxysmal AF were younger, had a shorter AF history, more hypertension, and less valvular disease, heart failure, and diabetes mellitus than patients with sustained AF. At baseline, patients with paroxysmal AF had a CHADS2(cardiac failure, hypertension, age, diabetes, stroke [doubled]) risk score of 1.79 ± 1.03 compared with 2.04 ± 1.12 in patients with sustained AF (p < 0.00001). The annualized risk of stroke or non-central nervous system (CNS) systemic embolism was 2.0 in paroxysmal AF compared with 2.2 in sustained AF (relative risk 0.87, 95% confidence interval [CI] 0.59 to 1.30, p = 0.496). After adjusting for confounding baseline variables, the relative risk was 0.94 (95% CI 0.63 to 1.40, p = 0.755). The incidence of stroke and non-CNS embolism was lower for patients treated with OAC irrespective of type of AF. There were more bleedings of any type in patients receiving clopidogrel plus aspirin, irrespective of the type of AF.
Conclusions Patients with paroxysmal AF treated with aspirin plus clopidogrel or OAC have a similar risk for thromboembolic events than patients with sustained AF. This risk can be significantly lowered with OAC. (The ACTIVE W trial; http://www.clinicaltrials.gov/ct/show/NCT00243178; NCT00243178)
Atrial fibrillation (AF) is associated with increased mortality and morbidity, mainly as a result of thromboembolic complications. Several risk stratification models of different complexity have been introduced to identify AF patients who benefit from oral anticoagulation (OAC) (1). In none of these models, the type of AF—paroxysmal versus persistent or permanent—has emerged as an independent predictor of thromboembolic events. The most recent guidelines on therapy of AF recommend the use of OAC, therefore, for patients with stroke risk factors irrespective of the type of AF (2). However, this recommendation is based on relatively weak data. In fact, only one retrospective analysis has specifically addressed the risk of stroke in patients with paroxysmal versus chronic AF (3). This study in 460 patients with paroxysmal AF who were enrolled in the SPAF (Stroke Prevention in Atrial Fibrillation) trials found similar incidences of thromboembolic events for both types of AF. This analysis was based on the SPAF trials conducted more than 15 years ago and was limited to patients treated with aspirin. The definition of “intermittent” AF as used by Hart et al. (3) is different from the definition of paroxysmal AF as used in the current guidelines (2). More importantly, treatments for AF and for underlying cardiovascular diseases have markedly changed, for instance regarding therapy for arterial hypertension or current international normalized ratio (INR) management. Finally, Hart et al. (3) used only data from patients treated with aspirin and not with OAC. Accordingly, a contemporary study is needed to confirm the stroke risk of patients with paroxysmal as compared with sustained AF. The present study, which is based on data from 6,706 patients, enrolled in the largest AF trial completed (ACTIVE W [Atrial Fibrillation Clopidogrel Trial With Irbesartan for Prevention of Vascular Events]) aims to answer 2 questions: 1) in patients taking OAC or aspirin plus clopidogrel, does paroxysmal AF carry a similar or lower stroke risk compared with persistent or permanent AF? and 2) is there a difference in efficacy and safety of OAC or combined antithrombotic therapy using aspirin plus clopidogrel in patients with paroxysmal AF?
Details of the ACTIVE W study have been recently published (4,5). In brief, patients with electrocardiographically documented AF and at least 1 risk factor for stroke (age ≥75 years; on treatment for hypertension; prior stroke, transient ischemic attack, or non-central nervous system [CNS] systemic embolus; left ventricular ejection fraction <45%; peripheral arterial disease; if patients were between the ages of 55 to 74 years requirement to have either diabetes mellitus requiring drug therapy or previous coronary artery disease) were randomized to receive open-label OAC (target INR 2.0 to 3.0) or combined antiplatelet therapy with aspirin (75 to 100 mg/day) and clopidogrel (75 mg/day). The primary outcome of the main trial was the first occurrence of one of the following vascular events: stroke, non-cerebral systemic embolism, myocardial infarction, or vascular death. Major bleeding was the most important safety outcome and was defined as any bleeding requiring transfusion of at least 2 U of red blood cells or equivalent of whole blood, or as bleeding associated with death, drop in hemoglobin of at least 5 g/dl, significant hypotension with need of inotropic support, intraocular bleeding, bleeding requiring surgery, or symptomatic intracranial hemorrhage. Patients were enrolled between June 2003 and December 2004. In August 2005, the Data Safety Monitoring Board recommended stopping the ACTIVE W study for strong evidence for superiority of OAC over the combined antiplatelet therapy.
Types of AF
Patients were classified as having paroxysmal, persistent, or permanent AF (2,6) by local investigators. Patients with permanent AF had electrocardiogram-documented AF at the time of enrollment and no evidence of sinus rhythm in the 6 months before randomization. Patients with paroxysmal or persistent AF may not have been in AF at the time of randomization, but had electrocardiogram-documented AF on 2 separate occasions, at least 2 weeks apart, in the 6 months before randomization.
Continuous variables were compared between patients with paroxysmal versus persistent/permanent AF using Student ttest. Because the present analysis aims to determine the thromboembolic risk associated with the 2 specific types of AF, the combination of all-cause stroke and non-CNS systemic embolism were considered as the primary outcome variable. Total and major bleeds were the main safety outcomes. Kaplan-Meier curves were constructed for time to event and were compared by log-rank tests before and after adjusting for baseline variables. A 2-sided p value <0.05 was considered statistically significant.
A total of 6,706 patients were randomized, of whom 1,202 (18%) had paroxysmal, 891 patients persistent, and 4,604 permanent AF. The 5,495 patients (82% of the entire study population) with persistent or permanent AF were analyzed together as patients with sustained AF. Patients with paroxysmal AF were younger, had a shorter history of the arrhythmia, more hypertension as the underlying cardiovascular problem, and less valvular disease, signs of heart failure, and diabetes mellitus compared with patients with sustained AF (Table 1).The prior use of OAC was lower in paroxysmal AF (64.8% vs. 79.5%, p < 0.00001) (Table 2).Patients with paroxysmal AF had a CHADS2(cardiac failure, hypertension, age, diabetes, stroke [doubled]) risk score of 1.79 ± 1.03 compared with 2.04 ± 1.12 in patients with sustained AF (p < 0.00001). Table 3shows that the lower mean CHADS2risk score in paroxysmal AF was due to younger age and the presence of less structural heart disease in these patients.
Patients were followed for median of 1.3 years until the ACTIVE W study was stopped after the recommendation of the independent Data Safety Monitoring Board (5).
Stroke and non-CNS systemic embolism
In 1,202 patients with paroxysmal AF, 25 patients had strokes and 4 non-CNS systemic embolic events occurred compared with 136 and 20, respectively, in 5,495 patients with sustained AF. The annual rate of the combined end point of stroke or non-CNS systemic embolism was 2.0 per 100 patient-years in paroxysmal AF compared with 2.2 in sustained AF (relative risk [RR] 0.87, 95% confidence interval [CI] 0.59 to 1.30, p = 0.50) (Fig. 1).After adjusting for prognostically important baseline variables (age, hypertension, heart failure, valvular heart disease, and diabetes mellitus), the RR was 0.94 (95% CI 0.63 to 1.40, p = 0.76) (Table 4).In the sustained AF group, baseline CHADS2risk score was 2.0 ± 1.1 in patients without compared to 2.7 ± 1.3 in patients with subsequent stroke or systemic embolism (p < 0.0001). In paroxysmal AF, the respective numbers were 1.8 ± 1.0 versus 2.1 ± 1.0 (p = 0.066). There was no interaction of AF type and stroke/non-CNS embolism.
When the incidence of stroke and non-CNS systemic embolism was analyzed according to treatment allocation, there were no significant differences in event rates between patients with paroxysmal versus sustained AF (Table 4).
Efficacy and safety of randomized therapy
The incidence of stroke and non-CNS embolism was reduced by OAC in patients with both types of AF (Table 4). For patients with sustained AF, the RR for stroke or non-CNS systemic embolism was 2.09 for patients treated with aspirin plus clopidogrel versus OAC (95% CI 1.50 to 2.93, p = 0.0000). For patients with paroxysmal AF, the RR was 1.61 (95% CI 0.76 to 33.42, p = 0.211) (Fig. 2).There was no significant interaction according to type of AF, indicating that OAC reduces stroke or non-CNS systemic embolism to a similar extent in both groups.
As detailed in Table 5,there was significantly more total bleeding in patients receiving therapy with clopidogrel plus aspirin, irrespective of the type of AF. The incidence of major bleedings as defined per protocol was similar for both treatment allocations.
The present study is the largest to examine the incidence of stroke and non-CNS embolism in patients with paroxysmal compared with those with sustained AF and on therapy with either OAC or clopidogrel and aspirin. Data were obtained from the ACTIVE W study, the largest single trial ever completed in AF comprising 6,706 patients. The most important finding of the present analysis is that paroxysmal AF carries the same risk for thromboembolic complications as sustained AF. This contemporary study thus confirms the only prior study by Hart et al. (3) and results of a meta-analysis (7). The second major finding is that therapy with OAC is similarly more effective compared with aspirin and clopidogrel in both forms of the arrhythmia.
Risk of thrombolembolic complications related to type of AF
Historical reports indicate that paroxysmal AF accounts for 35% to 66% of all AF (8), whereas newer studies using currently applied definitions of AF types found that approximately one-quarter of all AF is of the paroxysmal type (9,10). In line with previous findings, ACTIVE W patients with paroxysmal AF were younger and had less associated structural heart disease (i.e., valvular heart disease, congestive heart failure) compared with patients with sustained AF. This suggests that patients with paroxysmal AF are at an earlier stage of the arrhythmia (i.e., less electrical and structural remodeling of the atria) and might have a lower risk for thromboembolic events (11). In fact, the CHADS2score was significantly lower in patients with paroxysmal compared with those with sustained AF. Despite this, the risk of subsequent thromboembolic events was similar for ACTIVE W patients with both types of the arrhythmia.
Current guidelines recommendations on the use of OAC in paroxysmal AF are based on weak data. Only 1 prior study has aimed to evaluate the risk for thromboembolic complications depending on the type of AF. Hart et al. (3) analyzed the data from the SPAF I to III trials and found similar risk for patients with paroxysmal and sustained AF. There are, however, several important differences between this study (3) and the ACTIVE W study. First, more than 3 times as many patients with paroxysmal AF were enrolled in our study than in all 3 SPAF trials. Second, in the SPAF trials the diagnosis of “intermittent” AF required documentation of sinus rhythm and of at least 2 AF episodes 3 to 12 months before study entry (3). In contrast, the ACTIVE W study relied solely on investigator-based definition of the type of AF. Third, the ACTIVE W study evaluated the risk for thromboembolic complications in a population in which the overall stroke risk on warfarin was much lower than in previous studies. For example, the risk for ischemic stroke and non-CNS embolism on warfarin was 2.3% in the SPAF I trial (12) similar to risk for all-cause strokes plus non-CNS embolism in warfarin-treated patients enrolled in the SPORTIF III (Stroke Prevention Using the Oral Direct Thrombin Inhibitor Ximelagatran in Patients With Nonvalvular Atrial Fibrillation) trial (13). In contrast, the incidence of all-cause strokes plus non-CNS embolism was much lower in ACTIVE W patients on OAC (1.5% per year) similar to that observed in other recent anticoagulation trials (i.e., 1.2% per year in SPORTIF V ). These lower event rates may reflect improvements in the management of treatable stroke risk factors such as hypertension or heart failure along with better INR management. Thus, the present data represent the most up-to-date analysis of risk for stroke and non-CNS embolism in patients with paroxysmal AF versus patients with sustained AF. Our findings confirm the observations made by Hart et al. (3) and provide strong evidence that paroxysmal AF is risky, even if AF is not sustained. This may explain observations made in other contemporary trials such as AFFIRM (Atrial Fibrillation Follow-up Investigation of Rhythm Management) (15). In the AFFIRM trial, the prevalence of ischemic stroke was 7.1% in patients in whom rhythm control was attempted compared with 5.5% in those who were randomized to rate control. Based on the present data, it is possible that episodes of (asymptomatic) paroxysmal AF in the rhythm control arm for which patients did not receive anticoagulation therapy may account for this finding.
The results from the ACTIVE W study allow no firm conclusion as to what amount of paroxysmal AF increases risk for stroke and non-CNS embolism to that associated with sustained AF. However, this issue is currently being evaluated in a prospective clinical trials (ASSERT [ASymptomatic atrial fibrillation and Stroke Evaluation in pacemaker patients and the atrial fibrillation Reduction atrial pacing Trial]) in which pacemaker memory is used to precisely determine the number of episodes and duration of paroxysmal AF. This arrhythmia burden will be linked to stroke risk during 2.5 years of follow-up (16).
There is a relative paucity of data derived from controlled clinical trials regarding the benefit–risk ratio of OAC in patients with paroxysmal AF. From the 5 primary prevention trials on OAC in AF, 2 enrolled exclusively patients with permanent AF (17,18); in the remaining 3 trials, the prevalence of paroxysmal AF was 7% (19), 16% (20), and 34% (12). Thus, before the ACTIVE W study, only data derived from meta-analysis indicated that aspirin was less effective in stroke prevention than OAC in patients with paroxysmal and sustained AF (21). The present data demonstrate that the risk for stroke and non-CNS embolism was significantly higher in patients allocated to combined antiplatelet therapy compared to OAC. This also applied to patients with paroxysmal AF, although due to the smaller sample size, the power to detect treatment differences was somewhat smaller. The lack of a significant interaction according to type of AF shows that OAC is equally effective in patients with paroxysmal compared with those with sustained AF. Based on our data, 68 patients with sustained AF need to be treated with OAC to prevent one thromboembolic event compared with 135 patients with paroxysmal AF.
This study is not a natural history paper. Thus, comparisons between stroke incidence of patients with paroxysmal compared with sustained AF may be confounded by the treatment variable. Accordingly, our analysis describes the outcomes of treated patients. It is probably not correct to assume that paroxysmal AF is a greater risk factor for stroke than sustained AF based on the observation that both types of AF have similar event rates, although paroxysmal AF had a lower baseline CHADS2risk score. This is because the study has much greater power to detect a difference in CHADS2score than to detect a difference in outcome rates.
The present analysis of data obtained in the ACTIVE W study, the largest trial on OAC in nonvalvular AF, convincingly demonstrates similar risks for stroke and non-CNS embolism in patients with paroxysmal compared with sustained AF despite initially lower CHADS2scores in the first group. In patients generally tolerant to OAC, this therapy is more effective than combined antiplatelet therapy, with similar risks for major bleeding events.
This study was supported by Sanofi-Aventis and Bristol-Myers Squibb.
- Abbreviations and Acronyms
- atrial fibrillation
- confidence interval
- central nervous system
- international normalized ratio
- oral anticoagulation
- relative risk
- Received May 24, 2007.
- Revision received July 10, 2007.
- Accepted July 24, 2007.
- American College of Cardiology Foundation
- Lip G.Y.H.,
- Boos C.J.
- Fuster V.,
- Rydén L.E.,
- Cannom D.S.,
- et al.
- Hart R.G.,
- Pearce L.A.,
- Rothbart R.M.,
- et al.
- The ACTIVE Steering Committee
- The ACTIVE Writing Group on behalf of the ACTIVE Investigators
- Levy S.,
- Camm A.J.,
- Saksena S.,
- et al.
- Levy S.,
- Maarek M.,
- Coumel P.,
- et al.
- Camm A.J.,
- Savelieva I.,
- Bharati S.,
- et al.
- Stroke Prevention in Atrial Fibrillation Investigators
- Executive Steering Committee on behalf of the SPORTIF III Investigators
- Connolly S.J.,
- Laupacis A.,
- Gent M.,
- et al.