OUP user menu

Editor in chief

Roberto Ferrari

Associate Editors

Francisco Fernández-AvilésJeroen BaxMichael BöhmThomas F. LüscherFrank Ruschitzka

5.640

Published on behalf of

Upstream therapies to prevent atrial fibrillation

Paul Dorian, Bramah N. Singh
DOI: http://dx.doi.org/10.1093/eurheartj/sun033 H11-H31 First published online: 2 September 2008

Abstract

Atrial fibrillation (AF) is the most common sustained arrhythmia in the western world. It is associated with increased morbidity and mortality and decreased quality of life. The absence of a clear benefit of a rhythm-control strategy over a rate-control strategy observed in recent trials may be due to the fact that none of the available membrane-acting antiarrhythmics is entirely satisfactory. In addition, ablative therapy is available only for a small number of patients. Besides research efforts to improve the efficacy and safety of conventional antiarrhythmic agents, therapies directed ‘upstream’ of the electrical aspects of AF, towards the underlying anatomical substrate (atrial remodelling), have emerged as potential new pharmacological therapies. Potential upstream therapies for AF comprise a variety of agents such as those targeting the renin–angiotensin system [angiotensin-converting enzyme inhibitors (ACEI) and angiotensin receptor blockers (ARB)], statins, steroids, and N-3 polyunsaturated fatty acids. On the basis of suggestive experimental data, early phase clinical studies have been conducted and have provided exciting information on the potential of upstream therapy for the prevention of AF across a broad spectrum of cardiovascular patient groups. In some of these groups, such as patients with hypertension or heart failure, data may be considered to be sufficient to support the use of ACEI or ARB, at least in combination with membrane-acting antiarrhythmics. However, in most clinical settings examined, the evidence appears to be insufficient to drive changes in therapy management, and additional data from large-scale, randomized, double-blind, placebo-controlled trials with adequately defined endpoints are still needed. Numerous such trials are ongoing, reflecting the intense scientific interest in this field. The data derived from these trials may add to our understanding of the complex mechanisms that lead to AF and its maintenance, and may provide the necessary substantive evidence clarifying the benefit-to-risk ratio of these new therapeutic approaches.

Keywords
  • Atrial fibrillation
  • Upstream therapy
  • Angiotensin-converting enzyme inhibitors
  • Angiotensin receptor blockers
  • Statins
  • Steroids
  • N-3 polyunsaturated fatty acids
  • Fish oil

Introduction

Atrial fibrillation (AF) is the most common sustained arrhythmia in the western world, and its prevalence is still increasing with advancing age of the population and with improved survival of patients with cardiovascular diseases.1 It is now recognized as a serious cardiac disorder, closely associated with increased long-term risk of stroke, congestive heart failure and mortality, independent of the severity of underlying disease, as well as with decreased quality of life.1 Thus, prevention of thromboembolic complications and rate control or restoration and maintenance of sinus rhythm using antiarrhythmic drugs or, to a lesser extent, ablative procedures have been the mainstay of treatment of AF for the last decades. However, none of the antiarrhythmic drugs is entirely optimal in terms of efficacy and safety. In addition, ablative therapy is available only for a modest number of patients. Besides research efforts to improve class III antiarrhythmic agents, such as dronedarone2 and other newer classes of antiarrhythmic agents, such as atrial repolarization-delaying agents,3 attention is now being directed ‘upstream’ of the electrical aspects of AF, i.e. towards the underlying anatomical substrate, and potentially new pharmacological therapies for this condition have emerged.

Upstream therapies for AF comprise a variety of agents, such as those targeting the renin–angiotensin system [angiotensin-converting enzyme inhibitors (ACEI) and angiotensin receptor blockers (ARB)], statins, steroids, and N-3 polyunsaturated fatty acids. The growing interest in evaluating these agents in patients with AF is based on the recognition that pathologic remodelling within the atria (the substrate) plays a critical role in promoting and maintaining AF. The development and progression of AF is a multiple-component complex process still incompletely understood, but resulting in both electrical and structural changes of the atrial myocardium, which include the development of atrial fibrosis.46 It is well known that age, structural heart disease, hypertension, and enlarged atria are strongly associated with AF development.79 In recent clinical trials in AF, hypertension requiring therapy was present in as many as 50–70% of patients.1013 AF itself aggravates or even produces atrial electrical and functional changes, cell death, and fibrosis that self-perpetuate its maintenance, a phenomenon well illustrated by the adage ‘AF begets AF’.4,14 Experimental and clinical data suggest that inflammation,15,16 increased oxidative stress,17 and the renin–angiotensin–aldosterone system6,18,19 play an important role in the development of atrial remodelling and fibrosis associated with persistent AF (Figure 1).6,20 Inflammation seems to be involved during the early phase of persistent AF (electrical remodelling), while the renin–angiotensin system (RAS) may influence both the electrical and structural remodelling seen during later stages of AF development, thereby being a common pathway for the development of AF regardless of its primary cause. In addition, experimental studies suggest that angiotensin-II has several pro-inflammatory properties and that inflammation is able to stimulate angiotensin-II production, supporting the concept of a reciprocal stimulatory relationship between the two processes.6,16,18 Thus, the mechanisms by which upstream therapies may prevent or reduce AF include (i) the prevention of structural remodelling, by reduction of fibrosis, inflammation and oxidative stress, (ii) the improvement of haemodynamics, including lowering of blood pressure and reducing left ventricular (LV) and left atrial (LA) wall stress, and (iii) the prevention of coronary artery disease.21

Figure 1

Multiple pathways leading to atrial remodelling that predisposes the atria to atrial fibrillation. Numerous compounds can inhibit these pathways (X = blocks the pathway).20 CCB, calcium channel blocker.

However, most of the data supporting the beneficial clinical effect of upstream therapies in patients with AF are derived from observational, non-randomized, retrospective studies or post hoc analyses of large randomized trials not primarily designed with AF as a pre-specified endpoint. In addition, some of these studies have generated conflicting results.

Angiotensin-converting enzyme inhibitors and angiotensin receptor blockers

Inhibition of the RAS has been shown to reduce morbidity and mortality in patients with congestive heart failure (CHF), acute myocardial infarction (MI) with LV dysfunction, and is effective therapy in hypertensive patients. There is now a growing body of experimental and clinical evidence that ACEI and ARB may have some beneficial effect in preventing AF through several possible mechanisms, including prevention of LA dilatation and atrial fibrosis, regression of LV hypertrophy, reduction of inflammation and oxidative stress, modulation of sympathetic nerve activity, and direct modulation of ion-channel function, all potentially acting in mechanistic synergy with membrane-acting antiarrrhythmics.6,18

Several clinical studies have evaluated ACEI and ARB for the prevention of new-onset AF in patients with CHF or MI (primary prevention; Table 1),2227 as well as hypertension (primary prevention; Table 2).2833 Other studies evaluated ACEI and ARB for the prevention of the recurrence of AF after cardioversion (secondary prevention; Table 3) or ablation,3443 or recurrence in paroxysmal AF (secondary prevention; Table 4),4447 and for the prevention of AF after cardiothoracic surgery.4850 In a recent meta analysis of 11 of these trials (including a total of 56 308 patients), Healey et al.51 demonstrated that ACEI and ARB reduced the risk of AF by 28% [95% confidence interval (CI): 15–40%, P = 0.0002] across a broad range of patients. Reduction in AF was similar between the two classes of drugs (ACEI: 28%, P = 0.01; ARB: 29%, P = 0.00002). However, this meta analysis highlighted the significant differences in treatment effect between individual trials (test for heterogeneity: P = 0.00001), indicating that the beneficial effect of RAS inhibition varied according to the clinical situations considered. Specifically, RAS inhibition was found to be most effective in preventing AF in trials in patients with CHF (RR: 0.56; 95% CI: 0.37–0.85) and in those having undergone cardioversion for AF (RR: 0.52; 95% CI: 0.35–0.79), achieving non-significant results in trials in patients with hypertension (RR: 0.88; 95% CI: 0.66–1.19) or after MI (RR: 0.73; 95% CI: 0.43–1.26).51 Studies evaluating ACEI or ARB in the context of the prevention of AF after cardiothoracic surgery4850 are too few, retrospective, and/or observational, and yielded inconsistent results and thus will not be discussed in the present review.

View this table:
Table 1

Clinical trials of renin–angiotensin system inhibition to prevent new-onset atrial fibrillation in patients with congestive heart failure or acute myocardial infarction (primary prevention)

ReferenceStudy designPatientsInterventionMain findings
Pedersen et al.22Post hoc analysis of the TRACE study1577 patients, age >18 years, with AMI + LVEF ≤ 36% and NSR at entryTrandolapril vs. placebo Follow-up: 2–4 yearsNew-onset AF: 2.8% in the trandolapril group vs. 5.3% in the control group (P < 0.05)
Adjusted RR (95% CI): 0.45 (0.26–0.76)
Pizzetti et al.23Post hoc analysis of the GISSI-3 study17 749 patients with AMI and no chronic AFLisinopril vs. no lisinopril Follow-up: 4 yearsNo difference in new-onset AF vs. control (8.2 vs. 8.7%)
Vermes et al.24Post hoc analysis of patients included in one centre of the SOLVD study374 patients, mean age 57 years, with left ventricular dysfunction (asymptomatic or mild CHF + LVEF ≤ 35% = prevention arm) or overt CHF (treatment arm) and NSR at entryEnalapril vs. placebo Mean follow-up: 2.9 yearsNew-onset AF: 5.4% in the enalapril group vs. 24% in the placebo group (P < 0.0001). HR (95% CI): 0.22 (0.11–0.44)
Enalapril effect more marked in the prevention arm
Alsheikh-Ali et al.25Post hoc analysis of the SOLVD study6797 patients, mean age 60 years, with left ventricular dysfunction (asymptomatic or mild CHF+LVEF ≤ 35% = prevention arm) or overt CHF (treatment arm) (6.4% with AF at entry)Enalapril vs. placebo Mean follow-up: 34 monthsReduced incidence of hospitalization with atrial tachyarrhythmia in the enalapril group: RR (95% CI): 0.64 (0.48–0.85)
In patients without AF at study entry: RR (95% CI): 0.64 (0.46–0.88)
Maggioni et al.26Post hoc analysis of the Val-HeFT study4395 patients ≥18 years with chronic symptomatic HF (NYHA II–IV; median LVEF: 27%) with NSR at entryValsartan vs. placebo Mean follow-up: 23 monthsAF reported in 5.1% of patients with valsartan vs. 8.0% with placebo (P = 0.0002) Adjusted HR (95% CI): 0.63 (0.49–0.81)
Beneficial effect of valsartan observed whether or not patients received ACEI or amiodarone
Ducharme et al.27CHARM study. new-onset AF: pre-specified secondary endpoint6379 patients, mean age 66 years, with symptomatic CHF (NYHA II–IV) and reduced or preserved LV systolic function with NSR at entryCandesartan vs. placebo Median follow-up: 37.7 monthsNew-onset AF: 5.6% in the candesartan group vs. 6.7% in the placebo group (P < 0.05) Adjusted OR (95% CI): 0.80 (0.65–0.99); P = 0.039
  • AF, atrial fibrillation; AMI, acute myocardial infarction; CHF, congestive heart failure; CI, confidence interval; HR, hazard ratio; LV, left ventricular; LVEF, left ventricular ejection fraction; NSR, normal sinus rhythm; NYHA, New York Heart Association; OR, odds ratio; RR, relative risk.

View this table:
Table 2

Clinical trials of renin–angiotensin system inhibition to prevent new-onset atrial fibrillation in patients with hypertension (primary prevention)

ReferenceStudy designPatientsInterventionMain findings
Hansson et al.28CAPPP study: randomized, open-label. AF: pre-specified secondary endpoint (analysis based on adverse event reports)10 985 patients with hypertension aged 25–66 years, 0.6% and 0.3% with a history of AF and CHF, respectivelyCaptopril vs. diuretics ± beta-blocker Mean follow-up 6.1 yearsNo difference in new-onset AF: 2.1% in the captopril group vs. 2.5% in the control group (NS)
Hansson et al.29STOP-2 study: randomized, open label. New-onset AF: pre-specified secondary endpoint (analysis based on adverse event reports)6628 patients with hypertension, aged 70–84 years, 4.7 and 1.9% with a history of AF and CHF, respectivelyEnalapril/lisinopril vs. CCB vs. diuretics ± beta-blocker Study duration: 4 yearsNo difference in new-onset AF vs. CCB and vs. diuretics ± beta-blocker
L'Allier et al.30Retrospective longitudinal cohort study using integrated medical and pharmacy data from an administrative database of 8 million people in the USA10 926 patients ≥18 years with hypertension who filed a prescription of an ACEI or a CCB, matched on a propensity score (previous history of AF in 2.4%; prevalence of CHF: 3.7%)ACEI vs. CCB Average follow-up: 4.5 yearsAdjusted HR (95% CI) in favour of ACEI for incidence of new-onset AF: 0.85 (0.74–0.97) Absolute incidence of AF at 7 years: 19.4% with ACEI vs. 22.1% with CCB Adjusted HR (95% CI) in favour of ACEI for AF-related hospitalization: 0.74 (0.62–0.89)
Wachtell et al.31Planned secondary analysis of the randomized, double-blind LIFE study8851 patients aged 55–80 years, with hypertension+ECG LVH and NSR and no history of AF at entry.Losartan vs. atenolol Mean follow-up: 4.8 yearsReduced incidence of new-onset AF in the losartan group (3.5 vs. 5.3%): RR (95% CI): 0.67 (0.55–0.83); P < 0.001 vs. atenolol, with parallel reduction in stroke, despite similar BP reduction
Salehian et al.32Post hoc analysis of the HOPE study8335 patients at high cardiovascular risk (47% with hypertension), ≥55 years, without known heart failure or LV systolic dysfunction (LVEF >40%), with NSR at entryRamipril vs. placebo Mean follow-up: 4.5 yearsNo difference in new-onset AF (2.0%) vs. placebo (2.2%) RR (95% CI): 0.92 (0.68–1.24)
Schmieder et al.33Planned secondary analysis of the randomized, double-blind VALUE study13 760 patients, ≥50 years, with hypertension at high cardiovascular risk with NSR at entry (a small proportion with LVH at entry)Valsartan vs. amlodipine Follow-up: up to 5 yearsReduced incidence of new-onset AF with valsartan (3.7%) vs. 4.3% with amlodipine, despite lower blood pressure on amlodipine
HR (95% CI): 0.843 (0.713–0.997); P = 0.0455 (adjusted HR: P = 0.0521)
  • ACEI, angiotensin-converting enzyme inhibitor; AF, atrial fibrillation; ARB, angiotensin receptor blockers; CCB, calcium channel blockers; CHF, congestive heart failure; CI, confidence interval; HR, hazard ratio; LVEF, left ventricular ejection fraction; LVH, left ventricular hypertrophy; NS, not significant; NSR, normal sinus rhythm; RR, relative risk.

View this table:
Table 3

Clinical trials of renin angiotensin system inhibition to prevent recurrence of atrial fibrillation after cardioversion (secondary prevention)

ReferenceStudy designPatientsInterventionMain findings
Van den Berg et al.34Randomized, double-blind18 patients (aged 21–80 years), with CHF (NYHA II–III) and chronic AF, after successful cardioversionLisinopril vs. placebo started 6 weeks before cardioversion and continued for 6 weeks thereafterMaintenance of SR (Holter recording): 71% (5/7) with lisinopril vs. 36% (4/11) with placebo (NS)
Madrid et al.35Randomized, open-label154 patients (>18 years) with a history of an episode of persistent AF >7 days (median duration: 6 months)Irbesartan + amiodarone vs. amiodarone, started 3 weeks before cardioversion and continued for 2 months thereafter. Median follow-up: 254 daysSignificant reduction in the risk of AF recurrence at 2 months: adjusted HR (95% CI): 0.19 (0.04–0.86); P = 0.031. Benefit maintained at the end of follow-up: P = 0.007
Ueng et al.36Randomized, open-label145 patients (>18 years) with chronic AF >3 months (median duration: 34 months)Enalapril + amiodarone vs. amiodarone started 4 weeks before cardioversion Median follow-up: 270 daysLower rate of immediate recurrence of AF with enalapril (4.3% vs. 14.7%, P = 0.067) with a higher probability of remaining in SR at 4 weeks (84.3 vs. 61.3%, P = 0.002). Adjusted HR (95% CI): 0.37 (0.12–1.15); P = 0.041. Benefit maintained at the end of follow-up: P = 0.021, observed mainly in patients with LA diameter >40 mm
Zaman et al.37Observational, non-randomized retrospective analysis47 patients with persistent AF (between 48 h and 6 months) undergoing cardioversionACEI (for 6–12 months before inclusion, mainly for hypertension, and continued for the 1 year follow-up) vs. no ACEIThe number of defibrillation attempts required for successful cardioversion was decreased in the ACEI group (P < 0.001). Adjusted RR (95% CI) for hospital admissions: 0.14 (0.03–0.8); P = 0.03
Van Noord et al.38Observational, non-randomized, retrospective analysis107 patients with persistent (≥24 h) AF undergoing cardioversionACEI (pre-treated patients before onset of AF for hypertension or CHF) vs. no ACEICardioversion successful in 96% of patients under ACEI vs. 80% of the patients not pre-treated (P = 0.04). After 1 month of follow-up 49% of the pre-treated patients and 50% of those not pre-treated with ACEI were still in SR (NS)
Madrid et al.39Randomized, open-label90 patients (>18 years) with a history of an episode of persistent lone AF >7 days (median duration of AF: 9–10 months)Irbesartan 300 mg + amiodarone vs. irbesartan 150 mg + amiodarone vs. amiodarone started 3 weeks before cardioversion and continued for 2 months thereafter. Follow-up: 12 monthsPatients remaining free of AF recurrence at 12 months: irbesartan 300 mg + amiodarone: 77%; irbesartan 150 mg + amiodarone: 65%; amiodarone: 52% (P = 0.001 vs. irbesartan 300 mg + amiodarone)
Tveit et al.40Randomized, double-blind (CAPRAF study)137 patients (≥18 years) with persistent AF (>48 h; lone AF in 50%), and no strong indication for ACEI or ARBCandesartan for 3–6 weeks before and 6 months after cardioversion vs. placebo in patients with successful cardioversionRecurrence of AF at 6 months: 71% with candesartan vs. 65% with placebo
  • ACEI, angiotensin-converting enzyme inhibitor; AF, atrial fibrillation; ARB, angiotensin receptor blockers; CHF, congestive heart failure; CI, confidence interval; HR, hazard ratio; NYHA, New York Heart Association; RR, relative risk; SR, sinus rhythm.

View this table:
Table 4

Clinical trials of renin angiotensin system inhibition to prevent progression of atrial fibrillation (secondary prevention)

ReferenceStudy designPatientsInterventionMain findings
Murray et al.44Retrospective analysis of the AFFIRM studyPatients randomized to the rhythm control arm who were in SR421 patients treated with ACEI or ARB vs. 732 (40% with hypertension) not treated (at both baseline and 2 month follow-up visits)Risk of AF recurrence in the treated group did not differ from the risk observed the non-treated group: adjusted HR = 0.91, 95% CI = 0.77–1.09
Treated patients more likely had hypertension, diabetes, CAD, and reduced LVEF and duration of qualifying AF episode ≥2 days compared to patients not treatedIn patients with CHF or impaired LV function, RAS blockade was associated with a lower risk of AF recurrence
Hirayama et al.45Observational, retrospective analysis95 patients with paroxysmal AF (spontaneously terminated within 7 days)42 patients treated with ACEI for hypertension throughout the treatment period and follow-up (mean 8.3 years) vs. 53 not treated. Patients with CHF, acute or history of MI, cardiomyopathy, and cardiac surgery<6 months were excludedThe 5 year probability of persistence of paroxysmal AF without progression to chronic AF (>3 months duration) was 88.3% in the ACEI group, but 47.5% in the non-ACEI group (P = 0.002) Similar findings when the analysis was restricted to hypertensive patients
Fogari et al.46Randomized, double-blind222 hypertensive patients with NSR but with a history of recent (<6 months) paroxysmal AF and receiving amiodarone for at least 8 weeks before randomizationLosartan + amiodarone vs. amlodipine + amiodarone Follow-up: 1 year Patients with CHF, AMI <6 months, LA size >6 cm, and cardiac surgery<6 months were excludedDespite similar antihypertensive effect, at least one ECG-documented episode of AF was reported in 12.1% of the losartan-treated patients vs. 36.8% of amlodipine-treated patients (P < 0.01). This benefit was already apparent at the 12-week follow-up visit
Yin et al.47Randomized, open-label177 patients with lone paroxysmal AFLosartan + amiodarone vs. perindopril + amiodarone vs. amiodarone alone Follow-up: 24 months Patients with any structural heart disease, hypertension, LA size >40 mm, and LVEF < 50% were excludedAF recurrence: amiodarone alone: 41%; losartan + amiodarone: 19% (P = 0.006 vs. amiodarone alone); perindopril + amiodarone: 24% (P = 0.04 vs. amiodarone alone). Adjusted RR (95% CI): 0.36 (0.17–0.75) for losartan and 0.39 (0.20–0.78) for perindopril
LA diameter smaller in the losartan and perindopril groups vs. amiodarone alone (P < 0.001 for both comparisons)
No difference between losartan and perindopril
  • ACEI, angiotensin-converting enzyme inhibitor; AF, atrial fibrillation; AMI, acute myocardial infarction; ARB, angiotensin receptor blockers; CAD, coronary artery disease; CHF, congestive heart failure; CI, confidence interval; LA, left atrium; LVEF, left ventricular ejection fraction; MI, myocardial infarction; NSR, normal sinus rhythm; RAS, renin angiotensin system; RR, relative risk.

Clinical trials of renin–angiotensin system inhibition to prevent new-onset atrial fibrillation in patients with structural heart disease (primary prevention)

The first evidence of a beneficial effect of RAS inhibition in the prevention of new-onset AF came from a post hoc analysis of the TRAndolapril Cardiac Evaluation (TRACE) trial in patients with LV dysfunction after MI. In the subset of 1577 patients who were in normal sinus rhythm (NSR) at randomization, trandolapril was associated with a 55% reduction in the risk of new-onset AF compared with placebo after a 2–4 year follow-up (Table 1).22 Conversely, in the GISSI-3 trial, no significant reduction in new-onset AF was found, after a 4-year follow-up period, with 6 weeks of lisinopril therapy initiated within 24 h of MI.23 Interestingly, in the TRACE trial, all patients had LV dysfunction, whereas in the GISSI-3 trial, the majority of patients had no evidence of CHF at the time of randomization.

Retrospective analyses of the databases of the Studies Of Left Ventricular Dysfunction (SOLVD) and The Valsartan Heart Failure Trial (Val-HeFT) and a pre-specified secondary analysis of the Candesartan in Heart Failure-Assessment of Reduction in Mortality and Morbidity (CHARM) trial, all involving patients with LV dysfunction, showed that ACEI and ARB could prevent new-onset AF (Table 1).2427 In addition, analysis of the whole database of the SOLVD trial showed that the enalapril treatment was associated with a 36% reduction in the incidence of hospitalization with atrial tachyarrhythmia.25 It is worth noting that in the Val-HeFT and CHARM trials, valsartan and candesartan were evaluated on top of optimal heart failure therapy, including ACEI and beta-blockers.26,27 CHARM was in fact a three component trial: CHARM-Alternative, which enrolled patients with LV ejection fraction (LVEF) ≤0.40 not treated with ACE inhibitors because of prior intolerance, CHARM-Added, which enrolled patients with LVEF ≤0.40 treated with an ACE inhibitor, and CHARM-Preserved, which enrolled patients with LVEF >0.40 and allowed ACE inhibitor treatment in patients meeting specific eligibility criteria.27 Although there was no heterogeneity regarding the beneficial effect of candesartan in preventing AF across the three populations (P = 0.57), the benefit seemed to be driven by the results in patients with low LVEF. The odds reductions for new onset AF were 0.686 (5.6 vs. 8.0%), 0.856 (6.8 vs. 7.9%), and 0.894 (4.4 vs. 4.9%) for the CHARM-Alternative, CHARM-Added, and CHARM-Preserved components, respectively. For the two low (≤0.40) LVEF trials (Alternative and Added pooled), the odds ratio was 0.779. The reduction in the risk of AF among patients in the candesartan group was also consistent across subgroups, including New York Heart Association (NYHA) class, and use of antiarrhythmic agents, ACE inhibitors, and beta-blockers.

Collectively, these studies (TRACE, SOLVD, Val-HeFT, and CHARM) show that treatments with ACEI and/or ARB are associated with a reduction in the incidence of new-onset AF in patients with LV dysfunction.

Clinical trials of renin–angiotensin system inhibition to prevent new-onset atrial fibrillation in patients with hypertension (primary prevention)

Results obtained in trials evaluating ACEI or ARB in patients with hypertension are less consistent (Table 2).2833 The early CAptopril Prevention Project (CAPP) and Second Swedish Trial in Old Patients with Hypertension-2 (STOP-2) studies failed to show a significant reduction in new-onset AF with ACEI.28,29 However, although the occurrence of AF was a pre-specified secondary endpoint in these studies, the analysis was based on adverse event reports, which may account for the low incidence of AF observed and may have underestimated this incidence. Conversely, in a retrospective longitudinal cohort study of 10 926 newly diagnosed hypertensive patients from an administrative US database, ACEI therapy for at least 6 months was associated with lower incidences of new-onset AF and AF-related hospitalizations compared with calcium channel blockers (pair-matched patients), after an average follow-up of 4.5 years.30 The effect was greater in patients with a prior history of AF.

In two studies, a beneficial effect of ARB for the prevention of new-onset AF was shown to be independent of the antihypertensive effect.31,33 In a planned secondary analysis of the Losartan Intervention For Endpoint (LIFE) reduction in hypertension study, losartan was associated with a reduced incidence of new-onset AF compared with atenolol, despite similar blood pressure reduction.31 Patients receiving losartan tended to stay in sinus rhythm longer (P = 0.057) than those receiving atenolol. Similarly, in a planned secondary analysis of the Valsartan Antihypertensive Long-term Use Evaluation (VALUE) trial, new-onset AF was less frequent in the valsartan-treated patients than in the amlodipine-treated patients, despite lower blood pressure levels in the amlodipine group after up to 5 year follow-up.33 In VALUE, the incidence of new-onset AF was 3.7% with valsartan and 4.3% with amlodipine, and that of persistent AF was 1.4% with valsartan vs. 2.0% with amlodipine [unadjusted hazard ratio (HR): 0.683; 95% CI: 0.525, 0.889; P = 0.0046; adjusted HR: P = 0.0047]. Importantly, the LIFE study showed that patients presenting with new-onset AF had two-, three- and five-fold increased rates of cardiovascular events, stroke and hospitalization, respectively, all being reduced by losartan compared to atenolol.

It has been argued that positive results were seen in LIFE because the study included hypertensive patients with LV hypertrophy, i.e. at higher risk of AF, whereas the CAPP and STOP-2 studies enrolled patients with hypertension alone.51 The negative findings of a post hoc analysis of the Heart Outcomes Prevention Evaluation study database, showing no difference in new-onset AF between ramipril and placebo, seem to corroborate the hypothesis that the benefit of ACEI or ARB is likely to be greater in patients at high risk of AF, i.e. those with structural heart disease.32 HOPE included patients who were at high risk of cardiovascular events, but those with LV dysfunction or a history of heart failure were excluded; less than 50% had hypertension, and less than 10% had LV hypertrophy. On the other hand, hypertensive patients included in the VALUE study had similar risk factors to patients included in HOPE, notably a comparable prevalence of LV hypertrophy, and patients with CHF requiring ACEI therapy were excluded.33 The significant differences in favour of valsartan in the risk of new-onset AF were maintained after an adjustment for pre-defined potential confounding risk and disease factors such as age, history of coronary artery disease, and LV hypertrophy.

Since the results of both LIFE and VALUE are based on planned secondary analyses of new-onset AF, and both studies used ECG analysis data, these two studies suggest that ARB are beneficial in preventing new-onset AF in hypertensive patients. However, considering the negative results of other studies, further data from randomized clinical trials in hypertensive patients with normal LV function are required before any definite conclusion can be drawn.

Clinical trials of renin–angiotensin system inhibition to prevent recurrence of atrial fibrillation after cardioversion or ablation (secondary prevention)

Several studies have addressed the question as to whether ACEI or ARB may prevent the recurrence of AF after cardioversion (mainly electrical) (Table 3),3440 a situation where the rate of recurrence (usually clustered within the first month after cardioversion) is much higher than in patients who have never experienced AF.414 Again, results were not consistent across studies.

In an early small, randomized, double-blind, placebo-controlled study in 18 patients with chronic AF and CHF (NYHA classes II–IV), pre-treatment with lisinopril was shown to reduce the rate of AF recurrence within the first 6 weeks after successful electrical cardioversion, although the degree of reduction did not reach statistical significance, undoubtedly due to the small sample size.34 In the landmark randomized, open-label study by Madrid et al.35 in 154 patients with persistent AF, irbesartan added to amiodarone, started 3 weeks before cardioversion and continued for 2 months thereafter, reduced the risk of AF recurrence at the end of a median follow-up of 254 days, compared to amiodarone alone. Most of the benefit of irbesartan occurred during the first 2 months after conversion. In addition, no patient had recurrence during a time window of 1 h after cardioversion. In a similarly designed study in 145 patients with chronic AF, Ueng et al.36 demonstrated that the addition of enalapril to amiodarone started 4 weeks prior to electrical cardioversion was more effective in maintaining sinus rhythm than amiodarone alone at 4 weeks (P = 0.002) and at the end of a median follow-up of 270 days (P = 0.021). As in the Madrid study, there was a trend towards a lower rate of immediate recurrence of AF in the combination group (4.3 vs. 14.7%, P = 0.067), and most of the benefit of enalapril occurred during the first 4 weeks after cardioversion. In addition, a LA dimension of more than 40 mm was the only clinical parameter predicting relapse into AF, and the treatment with enalapril significantly reduced the recurrence of AF in these patients, suggesting potential reversibility of structural remodelling following cardioversion.36 In both the studies, there was a similar proportion of hypertension and LV hypertrophy and there was no significant difference in blood pressure between the two groups either at randomization or after the follow-up. Two observational non-randomized retrospective studies in patients with persistent AF undergoing cardioversion showed that pre-treatment with ACEI was associated with a reduction in the number of defibrillation attempts required for successful cardioversion37 and an increase in the rate of successful cardioversion,38 although in the latter study there was no difference in the rate of patients in sinus rhythm at 1 month follow-up. In a subsequent study in patients with lone AF, Madrid et al.39 demonstrated a dose-dependent effect of the addition of irbesartan pre-treatment (150 vs. 300 mg/day) to amiodarone, compared with amiodarone alone (Table 3). However, none of these studies were placebo-controlled, and their results have been contradicted by those of the recently published double-blind, placebo-controlled Candesartan in the Prevention of Relapsing Atrial Fibrillation (CAPRAF) study in patients with persistent lone AF undergoing electrical cardioversion.40 In this study, the recurrence rate of AF at 6 month follow-up was similar in the group pre-treated with candesartan and in the placebo group (Table 3). Several theories have been put forward to explain these contradictory findings, including differences in underlying diseases, a bridging effect of amiodarone in the previous studies, and the relatively low dose of candesartan (8 mg daily) used in the Tveit study.52,53 Thus, further studies are required before ACEI or ARB alone or in combination with membrane-acting antiarrhythmic agents can be recommended for the prevention of the recurrence of AF after successful cardioversion.

Three studies were conducted in patients undergoing ablative procedures, but they were retrospective and yielded inconsistent results. In a study of 196 patients undergoing ablation for atrial flutter, the use of ACEI/ARB and diuretics was significantly associated with lesser development of AF after a median follow-up of 2.2 years.41 In another study of 234 patients (aged 23–80 years; 71.8% men) with drug-resistant paroxysmal or persistent AF who underwent ablation, statin use, ACEI or ARB use, and their combined use, did not significantly influence ablation outcome.42 Finally, in a study of 177 patients (mean age: 56 years) who underwent ablation for paroxysmal or persistent AF, the administration prior to ablation (at least 1 month before this) and for the duration of follow-up (mean: 13.8 months) of ACEI (n = 31) and ARBs (n = 18) did not reduce the recurrence of AF, although a trend was observed in favour of ARBs (HR: 0.17; 95% CI: 0.02–1.34) that did not reach statistical significance (P = 0.09) probably because of the small sample size.43

Clinical trials of renin–angiotensin system inhibition to prevent recurrence of atrial fibrillation (secondary prevention)

Four studies examined the therapeutic possibility of using ACEI or ARB in the setting of paroxysmal AF in patients in sinus rhythm at inclusion (Table 4).4447

In a retrospective analysis of the rhythm-control arm of the AFFIRM database,44 Murray et al. found no effect of ACEI or ARB treatment in preventing the recurrence of AF, except in patients with a history of HF with reduced LVEF (P = 0.02 for ACEI and 0.04 for ARBs), in accordance with the results obtained in primary prevention trials in patients with LV dysfunction.22,2427 It was noted that the reduced number of patients with significant LV dysfunction and a much higher use of beta-blockers in AFFIRM compared with previous studies could account for the lack of clinical benefit of ACEI or ARB for the trial as a whole.44

Two studies, one retrospective observational45 and one randomized, double-blind,46 found that ACEI and ARB could prevent the progression of AF in hypertensive patients with paroxysmal AF (Table 4). In the Fogari study, patients did not necessarily have LV hypertrophy, and those with CHF, acute MI within the last 6 months and a LA size of more than 6 cm were excluded.46 Despite the similar antihypertensive effect of the two treatment, at least one ECG-documented episode of AF was reported in 12.1% of the losartan + amiodarone-treated patients vs. 36.8% of amlodipine + amiodarone-treated patients (P < 0.01), a benefit that was already apparent at the 12 week follow-up visit.

The study by Yin et al.47 has several important features (Table 4; Figures 2 and 3). First, the beneficial effect of adding ACEI or ARB to amiodarone to prevent AF recurrence was demonstrated in patients with lone paroxysmal AF, i.e. patients without any structural heart disease, suggesting a direct antiarrhythmic effect of RAS inhibition. Second, losartan and perindopril had similar efficacy in preventing AF recurrence in this group of patients. Finally, although the LA size was comparable in the three treatment groups at baseline, losartan and perindopril inhibited LA enlargement in patients with lone paroxysmal AF.

Figure 2

Kaplan–Meier estimates of atrial fibrillation (AF) recurrence-free survival after randomization. AF episodes within 14 days of randomization were excluded from the analysis. Group 1, amiodarone; Group 2, amiodarone + losartan; Group 3, amiodarone + perindopril.47

Figure 3

Change in left atrial diameter. Error bars indicate SEM. Numbers below the curves are P-values for pairwise comparison among the three groups at each time point. Group 1, amiodarone; Group 2, amiodarone + losartan; Group 3, amiodarone + perindopril.47

Clinical implications

Taken together, these studies suggest a clinical beneficial effect of RAS inhibition for the prevention of AF in a wide range of clinical situations. However, several of these trials had negative results. Such inconsistencies may be related to differences in study methodology such as design (retrospective/observational/prospective/randomized), sample size, duration of follow-up, duration of treatment, and the method used to document AF. In addition, differences in clinical situations may account for these conflicting results.

In patients with CHF and MI with LV dysfunction, RAS inhibition is clearly indicated on the basis of the demonstrated benefit in terms of morbidity and mortality, and the prevention of AF, if any, would just represent an additional benefit. It appears reasonable to consider ACEI or ARB as first-line therapy in patients with hypertension, particularly those with LV hypertrophy or other risk factors pre-disposing to AF. The 2007 Guidelines for the Management of Arterial Hypertension of the European Society of Hypertension (ESH) and the European Society of Cardiology (ESC) indicate that ACEI or ARB may be preferable for both primary and secondary prevention of AF, although it is acknowledged that confirmation from ongoing trials is desirable.54 The 2006 guidelines for the management of patients with AF from the American College of Cardiology (ACC)/American Heart Association (AHA) Task Force on Practice Guidelines and the ESC Committee for Practice Guidelines state that the results of the available studies, together with the favourable safety profile of ACEI and ARB compared with antiarrhythmic drugs, suggest a role for these agents for the prevention of initial or recurrent episodes of AF associated with hypertension, MI, HF, or diabetes mellitus.1

Finally, the role of ACEI and ARB in preventing AF recurrence in patients with lone AF (i.e. in patients without hypertension or LV dysfunction), is yet to be determined. Three studies in this patient population are available, two showing a beneficial effect of ACEI or ARB on top of amiodarone,39,47 and one showing no effect of ARB alone vs. placebo.40

Anti-inflammatory agents

Numerous clinical studies have evaluated the role of inflammation in the pathogenesis of AF by measuring serum or plasma levels of vascular inflammatory markers such as C-reactive protein and interleukin (IL)-6.15,16 This role was initially suggested by the high incidence of AF observed after cardiac surgery. In a first trial in patients undergoing cardiopulmonary bypass, a relationship was shown between postoperative C-reactive protein levels, levels of C-reactive protein-complement complexes, and incidence of postoperative atrial arrhythmias.55 In addition, atrial biopsies from patients with lone AF showed histological changes consistent with myocarditis.56 In non-operative patients, C-reactive protein appears to correlate with the AF burden: C-reactive protein levels were found to be higher in patients with AF compared with controls in sinus rhythm,5759 higher in patients with persistent AF compared with patients with paroxysmal AF,57,59 and higher in patients with symptomatic AF compared with asymptomatic AF.59 In addition, a large population-based study showed that an elevated baseline C-reactive protein in patients in sinus rhythm predicted an increased risk for developing AF in the future.58 High C-reactive protein levels also predicted recurrence of AF after successful electrical cardioversion.60 It was, therefore, postulated that anti-inflammatory therapy such as steroids could prevent the occurrence of AF.

This hypothesis was first tested in a double-blind study of 104 patients with normal LV function who had experienced persistent AF and who were converted to sinus rhythm with amiodarone or electrical cardioversion.59 Patients were randomized to receive low-dose glucocorticoid therapy (16 mg methylprednisolone for 4 weeks tapered to 4 mg for 4 months) or placebo. Propafenone (450 mg daily dosage) was administered to all patients who were cardioverted into NSR and was maintained during the entire follow-up period (median: 23.65 months). Compared with placebo, methylprednisolone reduced recurrent AF from 50 to 9.6% and permanent AF from 29 to 2%. Survival distributions for methylprednisolone and for placebo were significantly different for both recurrent and permanent AF (P < 0.001) (Figure 4). A multivariate Cox analysis showed that the risk of recurrent AF was increased by approximately seven times (P = 0.006), and the risk of permanent AF by about 12 times (P = 0.0006) for an increase of 1 mg/dL in plasma levels of average C-reactive protein concentration during follow-up. In the placebo group, patients who did not have recurrent AF had a significantly decreased C-reactive protein compared with patients who did (P < 0.001). Furthermore, the C-reactive protein concentration at the time of AF recurrence was increased compared to the most recent C-reactive protein concentration during follow-up in patients who did not experience a recurrence of AF (P < 0.001). C-reactive protein was increased in patients who developed permanent AF compared with those who did not (P < 0.001). Methylprednisolone significantly lowered C-reactive protein by an average of 80% in the first month, and this reduction was maintained for the duration of the study (P < 0.001, Figure 4). Importantly, the relative risk of recurrent and permanent AF in patients in the glucocorticoid group was similar to the relative risk of patients in the placebo group with the same level of C-reactive protein concentration. This suggests that the clinically important reduction in recurrent and permanent AF was probably a consequence of the reduction of C-reactive protein caused by methylprednisolone.

Figure 4

Main results of a study of methylprednisolone to prevent recurrence of permanent atrial fibrillation (AF) in patients successfully converted to sinus rhythm. Survival functions for each group (glucocorticoid and placebo) for the primary terminal event, recurrent AF (A) and expanded endpoint, permanent AF (B). The glucocorticoid group shows significantly (P < 0.001) better survival for both primary and expanded endpoints; (C) percentage change from baseline in C-reactive protein concentration during follow-up in the glucocorticoid and placebo groups.59

The hypothesis was further tested in a study of 102 patients (aged 67 ± 11 years; 58 men; none with LV dysfunction) who underwent successful electrical cardioversion and were followed up for 1 year.61 Antiarrhythmic therapy began before cardioversion and continued during follow-up or until AF recurrence. Patients were divided into four groups according to C-reactive protein quartiles. Patients in the lowest C-reactive protein quartile (<1.9 mg/L) had significantly lower rates of AF recurrence (4 vs. 33% at 3 months in the other three groups combined, P = 0.007, and 28 vs. 60% at 1 year, P = 0.01). Cox regression analyses using relevant patient characteristics and associated treatments as covariates showed that only C-reactive protein was independently associated with AF recurrence during follow-up (HR: 4.98; 95% CI: 1.75–14.26; P = 0.003).

Finally, three studies examined the effect of steroids for the prevention of AF in the context of cardiac surgery.6264 In an early randomized, double-blind, placebo-controlled study in 300 patients undergoing coronary revascularization surgery, dexamethasone (4 mg/mL) administered after the induction of anaesthesia (with a second dose on the morning after surgery) failed to decrease the incidence of AF (27 vs. 32%) in the postoperative period.62 A second double-blind study was conducted in 88 patients undergoing elective coronary artery bypass grafting (CABG) and free of documented or suspected arrhythmias before surgery. Patients were randomized to receive either 1 g of methylprednisolone before surgery and 4 mg of dexamethasone every 6 h for 1 day after surgery or placebo.63 Postoperative AF occurred in 21% of the patients in the steroid group, compared with 51% of the 43 patients in the placebo group (P = 0.003). Minor but not major postoperative complications were more frequent in the steroid group (P = 0.01). Finally, a double-blind, randomized multicentre trial in three university hospitals in Finland included 241 patients without prior history of AF or flutter and scheduled to undergo first CABG surgery and/or aortic valve replacement.64 Patients were randomized to receive either 100 mg hydrocortisone or placebo as follows: the first dose in the evening of the operative day, then one dose every 8 h during the next 3 days. All patients received oral metoprolol (50–150 mg/day) titrated to heart rate. The incidence of postoperative AF was significantly lower in the steroid group (30%) than in the placebo group (48%; adjusted HR: 0.54; 95% CI: 0.35–0.83; P = 0.004). Patients receiving hydrocortisone did not experience higher rates of superficial or deep wound infections, or other major complications. The conflicting results observed between the two latter studies and the study by Halvorsen et al. may be explained by the difference in the steroid dosage regimen. In the study by Halvorsen et al., patients received only two doses of steroids, whereas in the two other studies steroids were administered for 24–72 h. Overall, and considering the positive results of two other studies,65,66 steroids seem to prevent the occurrence of postoperative AF after CABG and/or valve replacement.

Taken together, these findings indicate that inflammation indeed plays a role in the development and maintenance of AF, and that steroids are able to prevent the occurrence of AF after successful cardioversion and after cardiac surgery. However, because of their adverse effects, particularly during long-term administration, the long-term routine use of steroids cannot be recommended and there is a need for alternative anti-inflammatory agents.

Statins

HMG CoA-reductase inhibitors (statins) may represent such an alternative. The mechanisms by which these drugs exhibit favourable effects in the primary and secondary prevention of coronary artery disease are poorly understood. Statins are thought to have pleiotropic (multiple) effects, including anti-inflammatory and anti-oxidant properties (Figure 5),67 and thus represent another potential upstream therapy for the prevention of AF.

Figure 5

Possible mechanisms of antiarrhythmic effects of statins against atrial fibrillation.67

Several clinical studies have evaluated statins for the prevention of new-onset AF and AF progression (Table 5),6874 for the prevention of the recurrence of AF after cardioversion7579 (Table 6) or ablation,42,43 and for the prevention of AF after thoracic surgery (Table 7).8087 As for ACEI and ARB, most studies were retrospective and similarly yielded mixed results.

View this table:
Table 5

Clinical trials of statins to prevent new-onset and progression of atrial fibrillation (primary and secondary prevention)

ReferenceStudy designPatientsInterventionMain findings
Young-Xu et al.68Prospective observational cohort449 patients (40–87 years) with chronic stable angina, without CHF, and history of AFStatins vs. no statins Average follow-up: 5 yearsNew-onset AF: 9% in statin users vs. 15% in non-statin users Adjusted OR (95% CI): 0.37 (0.18–0.76)
Merckx et al.69Retrospective218 patients on statins and 449 matched controls in SR but with increased risk for the development of AF (LA ≥ 40 mm and LVH: wall thickness ≥10 mm) without LV dysfunctionStatins vs. no statins Mean follow-up: 6.5 yearsNew-onset AF: 7.3% in statin users vs. 11.6% in non-statin users Lower incidence of AF especially in patients >65 years, and independent of the presence of ischaemic heart disease
Dernellis et al.70Randomized, single-blind80 patients with proven paroxysmal AF and C-reactive protein levels between 0.8 and 13 mg/LAtorvastatin vs. placebo Follow-up: 4–6 months of therapyParoxysmal AF completely resolved in 65% of atorvastatin-treated patients vs. 10% with placebo. C-reactive protein levels were lower in the atorvastatin group at study end. The benefit was independent of cholesterol levels and use of ACEI/ARB
Amit et al.71Retrospective264 patients (mean age: 71 years) with permanent pace maker, at high risk for AF, but with NSR at entryStatins vs. no statins Median follow-up: 359 daysNew-onset AF: 10.5 events/100 patient-years with statin vs. 9.8% without statins (P = 0.51)
Hanna et al.72Retrospective25 268 patients with CHF (LVEF ≤ 40%), of whom 28% had AF at or prior to enrolmentLipid-lowering drug use vs. no useAF prevalence: 25.1% in treated-patients, 32.6% in untreated hyperlipidaemic patients and 32.8% in patients without hyperlipidaemia (P < 0.001 for both comparisons)
Adabag et al.73Retrospective13 783 patients with CHD without known AFStatins vs. no statins Average follow-up: 4.8 yearsNo difference in new-onset AF incidence with statin treatment (HR: 1.0: 95% CI: 0.88–1.14, P = 0.9).
However, AF was less common among statin-treated patients with CHF (HR: 0.57; 95% CI: 0.33–1.00, P = 0.04)
Ramani et al.74Retrospective1526 patients with acute coronary syndrome without known AFStatins vs. no statins at time of admissionSignificant decrease in new-onset AF or atrial flutter at time of admission in statin users: adjusted OR: 0.57; 95% CI: 0.39–0.83 (P < 0.01) after correcting for age, race, diabetes mellitus, chest pain, and use of ACEI
  • ACEI, angiotensin-converting enzyme inhibitor; AF, atrial fibrillation; ARB, angiotensin receptor blockers; CHD, coronary heart disease; CHF, congestive heart failure; CI, confidence interval; HR, hazard ratio; LVEF, left ventricular ejection fraction; LVH, left ventricular hypertrophy; NSR, normal sinus rhythm; OR, odds ratio.

View this table:
Table 6

Clinical trials of statins to prevent recurrence of atrial fibrillation after cardioversion (secondary prevention)

ReferenceStudy designPatientsInterventionMain findings
Siu et al.75Retrospective62 patients with lone persistent AF (mean age 61 years) having undergone successful electrical cardioversionStatins before and after cardioversion (n = 10) vs. no statins (n = 52) Mean follow-up: 44 monthsAt the end of follow-up, the statin group had fewer patients with recurrent AF than the control group (Kaplan–Meier analysis 40 vs. 84%, P = 0.007)
RR = 0.31; 95% CI: 0.103–0.905, P = 0.032 (multivariable Cox regression analysis)
Tveit et al.76Randomized, open-label114 AF patients (36%: lone AF) scheduled for electrical cardioversionPravastatin for 3 weeks before and 6 weeks after cardioversion vs. no treatmentPravastatin did not reduce the rate of AF recurrence (35 vs. 33% at 6 weeks)
Colivicchi et al.77Retrospective851 hypertensive patients with persistent AF having undergone successful cardioversionStatins (n = 294) vs. no statins Follow-up: 1 year1 year unadjusted AF recurrence rate: 50.6% in the statin group and 58.5% in the no-statin group
After adjustment (propensity to receive statin therapy+all other potential confounders), statin therapy remained associated with a lower risk of AF recurrence at 1 year (HR, 0.75; 95% CI 0.63–0.89. P = 0,032)
Ozaydin et al.78Randomized open-label48 patients (mean age: 62 years) with AF lasting >48 h scheduled for electrical cardioversionAtorvastatin started 48 h before cardioversion and continued during 3-month follow-up vs. no atorvastatinAF recurrence: 12.5% in the treated group vs. 45.8% in the non-treated group (P = 0.01) Adjusted RR (95% CI): 0.19 (0.052–0.72); P = 0.01 C-reactive protein levels 48 h after cardioversion decreased significantly in the treated group but not in the non-treated group
Garcia-Fernandez et al.79Randomized, open-label54 patients with persistent AF undergoing electrical cardioversionAtorvastatin started before cardioversion and continued up to the end of follow-up vs. no treatmentNo difference between the two groups in terms of AF recurrence and markers of inflammation, including C-reactive protein
Follow-up: 3 months
  • AF, atrial fibrillation; CI, confidence interval; HR, hazard ratio; RR, relative risk.

View this table:
Table 7

Clinical trials of statins to prevent recurrence of atrial fibrillation after thoracic surgery

ReferenceStudy designPatientsInterventionMain findings
Auer et al.80Prospective cohort (SPPAF study)253 patients (mean age 65 years) with no CHF or LV dysfunction undergoing cardiac surgeryStatin use before surgery vs. no statin useStatins significantly decreased the rate of postoperative AF, from 45.9–32.8% (P < 0.05 after adjustment for age, heart valve surgery, and antiarrhythmic drug use)
Amar et al.81Prospective cohort131 patients (mean age 73 years) undergoing major lung or oesophageal resection, in NSR prior to surgeryStatin use before surgery vs. no statin useStatin use was associated with a three-fold decrease in the odds of developing postoperative AF (11 vs. 29%; OR = 0.26; 95% CI = 0.08–0.82; P = 0.022 (stepwise logistic regression), independent of C-reactive protein or IL-6 levels
Marin et al.82Prospective cohort234 patients undergoing CABG (mean age 65 years) with previous permanent/persistent AF/flutterStatin use for a median duration of 31 days before surgery vs. no statin useStatin use was associated with a decrease in postoperative AF (OR = 0.52; 95% CI = 0.28–0.96, P = 0.038) (multivariate analysis)
Patti et al.83Randomized, double-blind (ARMYDA-3)200 patients undergoing cardiac surgery (mean age 66 years) without previous history of AF and previous or current treatment with statinsAtorvastatin vs. placebo starting 7 days before surgery and continued until hospital dischargeAtorvastatin reduced the incidence of postoperative AF (35 vs. 57%, P = 0.003); OR = 0.39; 95% CI = 0.18 to 0.85; P = 0.017 (multivariate analysis). Length of stay was longer in the placebo vs. atorvastatin group (P = 0.001)
Chello et al.84Randomized, double-blind40 patients scheduled for CABG (mean age: 65 years) in NSR at entryAtorvastatin vs. placebo started 3 weeks before surgeryPre-treatment with atorvastatin reduced the incidence of postoperative AF (2 vs. 5%) and significantly reduced cytokine release and neutrophil adhesion to the venous endothelium
Ozaydin et al.85Prospective cohort362 patients (mean age 61 years) undergoing first elective CABG, without AF and CHF NYHA III–IV at entryStatin use for a mean duration of 2.7 months before surgery vs. no statin usePostoperative AF was less frequent (8.2 vs. 16.8%) and its duration was shorter in the statin group compared to the non-statin group (P = 0.03 and 0.0001, respectively)
Lertsburapa et al.86Prospective cohort from the AFIST I, II, and III studies555 patients (mean age 67.8 years) undergoing cardiothoracic surgery, 6.1% with a history of AF, 12.6% a history of CHF, 84.0% received postoperative beta-blocker, and 44.1% prophylactic amiodaroneStatin use before surgery vs. no statin useStatin use was associated with a reduction in postoperative AF: 27.8 vs. 36.6%; multivariate analysis: OR = 0.60; 95% CI: 0.37–0.99
Virani et al.87Retrospective study4044 patients without history of AF undergoing cardiac surgeryStatin use before surgery vs. no statin usePostoperative AF occurred in 31.4% in both the statin and non-statin groups
  • AF, atrial fibrillation; CABG, coronary artery bypass graft; CHF, congestive heart failure; CI, confidence interval; LV, left ventricular; NSR, normal sinus rhythm; NYHA, New York Heart Association; OR, odds ratio.

Clinical trials of statins to prevent new-onset atrial fibrillation and atrial fibrillation progression (primary and secondary prevention)

The first study to assess the effects of statins for the prevention of new-onset AF was a prospective observational cohort study in 449 patients with chronic stable angina at high risk of AF, but no CHF or history of previous AF.68 Statin therapy was used by 59% of the patients during the study period. After an average follow-up of 5 years, statin therapy was associated with a significant reduction in the risk of new-onset AF (9 vs. 15%), an association that remained significant after adjustment for potential confounders, including age, hypertension, LV systolic function, occurrence of HF or acute ischaemic events, and pre- and post-treatment cholesterol levels. Moreover, there was a dose-response relationship between length of statin therapy and reduction of new-onset AF. Additional analyses showed that the incidence of AF did not differ between users of non-statin cholesterol-lowering agents and non-users (RR: 0.92; 95% CI: 0.56–1.50).

A series of retrospective studies of various sizes have been performed since then, yielding conflicting results (Table 5). In a study in 218 patients at high risk of AF, Merckx et al.69 found that the benefit of statins on new-onset AF seemed to be restricted to patients older than 65 years. Similarly, in 13 783 patients with coronary heart disease, statin therapy was associated with a reduced risk of new-onset AF only in patients with CHF.73 Amit et al.71 could not demonstrate any benefit in 264 patients with a permanent pacemaker, whereas Ramani et al.74 showed that the use of statins was associated with a reduced risk of new-onset AF or flutter at time of admission in 1526 patients with acute coronary syndrome.

In a cross-sectional analysis of 25 268 patients included in the multicentre Guidant-sponsored ADVANCENT Heart Failure Registry (28% with AF at or prior to enrolment), Hanna et al.72 assessed the impact of lipid-lowering drug use (mainly statins) on AF prevalence in patients with reduced LVEF. AF prevalence was 25.1% in patients taking lipid-lowering drugs compared with 32.6% in untreated hyperlipidaemic patients and 32.8% in patients without hyperlipidaemia (P < 0.001 for both comparisons). In the multivariable analysis, lipid-lowering drug use remained significantly associated with reduced odds of AF (OR: 0.69; 95% CI: 0.64–0.74), an effect that was larger than that of ACEI/ARB (OR: 0.85; 95% CI: 0.79–0.92) or beta-blockers (OR: 0.95; 95% CI: 0.88–1.02). The beneficial impact of lipid-lowering drugs on AF risk was independent of their effects on the lipid profile.

While there are no published prospective, randomized, controlled studies to assess the efficacy of statins in the prevention of new-onset AF, Dernellis and Panaretou70 performed a randomized, single-blind, placebo-controlled study of atorvastatin for 4–6 months in 80 patients with paroxysmal AF, using C-reactive protein levels as a guide to dosage adjustment. At the end of the treatment period, paroxysmal AF completely resolved in 65% of atorvastatin-treated patients vs. 10% of placebo-treated patients (P = 0.001). C-reactive protein levels were lower in the atorvastatin group at study end. The benefit was independent of cholesterol levels and use of ACEI/ARB.

Overall, data supporting the use of statins for the primary and secondary prevention of AF are limited, obtained mainly from retrospective studies that generated conflicting results. Thus, additional well-conducted studies are needed before any conclusion can be drawn.

Clinical trials of statins to prevent recurrence of atrial fibrillation after cardioversion or ablation (secondary prevention)

Siu et al.75 were the first to study statins for the prevention of the recurrence of AF after successful electrical cardioversion in a retrospective study of 62 patients with lone persistent AF. After a 44-month follow-up, statin-treated patients (n = 10) had less recurrent AF than the control group (n = 52) (40 vs. 84%, P = 0.0007). The use of statins was associated with a significant decrease in the risk of arrhythmia recurrence on multivariable Cox regression analysis (RR: 0.31; 95% CI: 0.103–0.905; P < 0.032). The benefit of statin therapy was already apparent during the first few months after cardioversion. The benefit observed in this small early study was confirmed in another retrospective study of 851 hypertensive patients with persistent AF having undergone successful cardioversion (Table 6).77 However, three prospective randomized, open-label studies were conducted evaluating pravastatin76 or atorvastatin78,79 vs. no treatment for the prevention of AF recurrence after cardioversion, and only one confirmed the beneficial effect of statins in this setting (Table 6).78 Two studies were performed to evaluate the effect of statins for the prevention of AF recurrence after ablative procedures.42,43 Both failed to show benefit of statin therapy.

Overall, the number of studies in the clinical settings of cardioversion and ablation is limited. Factors that could explain the negative results obtained in the majority of randomized studies include small sample size, insufficient duration of statin therapy/follow-up, inclusion of patients at a too advanced stage of AF disease, and, of course, a true lack of efficacy of statins in these contexts. Further studies are needed to clarify whether some subgroups of selected patients might benefit from statin therapy.

Clinical trials of statins to prevent recurrence of atrial fibrillation after thoracic surgery (secondary prevention)

Postoperative AF is a common complication after cardiothoracic surgery, with incidences consistently reported to range between 27 and 40%, and is associated with greater postoperative resource use and poorer outcomes.48 Since perioperative inflammation is thought to play a major role in its development, statins have been evaluated in this setting in several trials (Table 7).8087

Based on observational evidence suggesting that patients who have undergone previous statin therapy have a lower incidence of postoperative AF,8082 Patti et al. performed a well-conducted, randomized, double-blind trial in 200 patients undergoing elective cardiac surgery with cardiopulmonary bypass.83,8890 Patients had no history of previous AF and were not being treated with statins. They were randomized (regardless of patient cholesterol levels) to receive atorvastatin (40 mg/day) or placebo, starting 8 days before surgery and continued until hospital discharge. Atorvastatin significantly reduced the incidence of postoperative AF (35 vs. 57%, P = 0.003; OR: 0.39; 95% CI: 0.18–0.85; P = 0.017 by multivariate analysis). The incidence of postoperative AF in the placebo group was remarkably high in this study, probably reflecting underutilization of beta-blockers and amiodarone (66% of patients on beta-blockers and only one patient in each group receiving amiodarone before surgery; no information is available regarding these therapies after surgery). Nevertheless, logistic regression analysis revealed that beta-blocker therapy, atorvastatin, and a combination of these were independently associated with a reduction in the risk of postoperative AF. The incremental benefit of statins in patients already receiving beta-blockers amounted to a 90% risk reduction. Although this treatment effect occurred irrespective of age, sex, presence of diabetes mellitus, or chronic obstructive pulmonary disease, atorvastatin had no benefit in patients with LA enlargement or those undergoing valve surgery, i.e. those potentially more prone to develop postoperative AF. However, LA enlargement (LA diameter ≥40 mm) was not a predictor of higher incidence of postoperative AF in this study. Length of hospital stay was longer in the placebo group than in the atorvastatin group (P = 0.001). Peak C-reactive protein levels were lower in patients without AF (P = 0.01), irrespective of randomization assignment. The incidence of major adverse cardiac and cerebrovascular events at 30 days was similar in the two groups. Another study from the same group showed that atorvastatin significantly reduced cytokine release and neutrophil adhesion to the venous endothelium, an effect associated with a reduction in the incidence of postoperative AF.84

The beneficial effect of statins in preventing postoperative AF was confirmed in two cohort studies in patients undergoing CABG85 or cardiothoracic surgery (mainly CABG).86 In the latter cohort from the Atrial Fibrillation Suppression Trials I, II, and III, in which postoperative beta-blockers and amiodarone were used in 84.0 and 44.1% of patients, respectively, statin therapy before surgery had a positive and independent effect on postoperative AF rates. Higher intensity statin dosing (equivalent of ≥40 mg of atorvastatin) seemed to be associated with the greatest reductions in risk (adjusted OR: 0.45; 95% CI: 0.21–0.99). Only one of the published studies reported negative results.87 Although this was a retrospective study, it should be noted that a large proportion (34%) of patients underwent valve surgery.

Taken together, these findings indicate that statin therapy initiated before surgery is associated with a decreased risk in postoperative AF, at least in patients undergoing CABG.

Meta analyses of statin studies

Two meta analyses of trials evaluating statins for the prevention of AF have recently been published.91,92

In the first one, which included six randomized clinical trials (RCTs) and 10 observational studies (7041 patients), the analysis of RCTs showed no significant effect of statins on AF development (RR: 0.76; 95% CI: 0.55–1.05; P = 0.09), and indicated significant heterogeneity between individual studies (P = 0.0008) probably because of differences in AF detection methodology.91 Indeed, a subgroup analysis showed that in the three RCT studies which used Holter or continuous monitoring for more than 24 h for AF detection, statin use was associated with an overall 50% relative risk reduction (P < 0.00001), without statistical heterogeneity. The analysis of the observational studies demonstrated that statin use reduced the relative risk for AF by 23% (95% CI: 0.70–0.85; P < 0.00001) without significant differences between the trials (P = 0.08). This favourable effect was greatest in the postoperative patients (RR: 0.61; 95% CI: 0.49–0.76; P < 0.0001). The authors concluded that the meta analysis suggests that statins may be effective in AF prevention especially in the postoperative setting.91

In the second meta analysis, which included six RCTs (3357 patients), the use of statins was significantly associated with a decreased risk of AF compared with the control group (OR: 0.39; 95% CI: 0.18–0.85; P = 0.02).92 The benefit of statin therapy seemed to be more marked in secondary prevention of AF (OR: 0.33; 95% CI: 0.10–1.03; P = 0.06) than for new onset or postoperative AF (OR: 0.60; 95% CI: 0.27–1.37; P = 0.23). The authors concluded that the use of statins was significantly associated with a decreased risk of incidence or recurrence of AF in patients in sinus rhythm with a history of previous AF, or undergoing cardiac surgery or having experienced an acute coronary syndrome.92

Meta analyses are very useful for revealing treatment effects that are otherwise undetectable due to low rates of events, and represent a remarkable tool for generating hypotheses. However, heterogeneity across trials is one of their limitations, and treatment effect may differ according to the clinical setting. As stated in the 2006 guidelines for the management of patients with AF issued by the ACC/AHA Task Force on Practice Guidelines and the ESC Committee for Practice Guidelines, although the use of statins may protect against AF for primary prevention, this use is still inadequately explored.1 The available data regarding the effect of statins for the prevention of AF recurrence after cardioversion or ablative procedures are still insufficient to recommend their use. Finally, although statins seem to offer protection against postoperative AF, this is mainly after CABG surgery, in a population of coronary artery disease patients in whom statins are almost universally already indicated. Larger RCTs with long-term follow-up in different clinical settings are clearly needed.

N-3 polyunsaturated fatty acids

N-3 polyunsaturated fatty acids (PUFAs), of which oily fish are an important source, are believed to possess anti-inflammatory properties, antiarrhythmic effects, and antifibrotic effects and to have potential in the prevention and treatment of AF.16,9396 Few controlled studies, however, have been published on the therapeutic use of PUFAs in AF; three are population-based cohort studies evaluating the impact of PUFAs on the occurrence of new-onset AF and one is a randomized, open-label study in patients undergoing cardiac surgery (Table 8).97100

View this table:
Table 8

Clinical trials of n−3 (omega−3) polyunsaturated fatty acids to prevent atrial fibrillation

ReferenceStudy designPatientsInterventionMain findings
Mozaffarian et al.97Population-based, prospective cohort (Cardiovascular Health Study)4815 subjects (≥65 years)Dietary intake assessment Follow-up: 12 yearsConsumption of tuna and other broiled or baked fish was inversely associated with the incidence of AF, with 28% lower risk with intake 1 to 4 times/week (P = 0.005), and 31% lower risk with intake ≥5 times per week (P = 0.008), compared with <1 intake per month
Calo et al.98Randomized, open-label160 patients (mean age 65 years) undergoing elective CABG, with NSR at entryPUFAs for at least 5 days before surgery and up to discharge vs. controlPostoperative AF developed in 33.3% of patients of the control group vs. 15.2% of the PUFA group (15.2%) (P = 0.013). OR = 0.35; 95% CI = 0.16–0.76; P = 0.013)
PUFAs reduced the length of hospital stay
Frost et al.99Population-based prospective cohort (Danish Diet, Cancer and Health study)47 949 subjects (mean age 56 years)Dietary intake assessment Mean follow-up: 5.7 yearsConsumption of n−3 fatty acids from fish was not associated with a reduction in risk of AF or flutter
Brouwer et al.100Population-based prospective cohort. (Rotterdam study)5184 subjects (mean age 67 years)Dietary intake assessment Mean follow-up: 6.4 yearsIntake of very long chain n-3 PUFA or fish was not related to the incidence of AF
  • AF, atrial fibrillation; CABG, coronary artery bypass graft; NSR, normal sinus rhythm; OR, odds ratio; PUFA, polyunsaturated fatty acid.

The Cardiovascular Health Study was the first prospective study evaluating the impact of fish oil consumption on the risk of new-onset AF in 4815 adults aged at least 65 years.97 At 12-year follow-up, consumption of tuna or other broiled or baked fish, but not fried fish or fish sandwiches (fish burgers), one to four times per week was associated with a 28% lower risk of new-onset AF (HR: 0.72; 95% CI: 0.58–0.91; P = 0.005) compared with an intake less than once per month. However, the two other prospective cohort studies, the Danish Diet, Cancer and Health study in 47 949 subjects aged 50–60 years99 and the Rotterdam study in 5184 subjects (mean age 67 years),100 could not demonstrate that the consumption of PUFAs was associated with a reduction in the risk of new-onset AF after 5.7 and 6.4 years of follow-up, respectively. In the Danish study, the quintile with the greatest intake of fish oils even demonstrated significantly more episodes of AF than the lowest quintile (adjusted HR of 1.44).

A proof-of-concept study in 160 patients undergoing elective CABG surgery showed that pre-operative PUFA administration reduced the incidence of postoperative AF when compared with no therapy (Table 8).98 In this randomized, open-label study, patients on active treatment received 2 g PUFA for 5 days before surgery and until hospital discharge. In addition, treatment with fish oils reduced the length of hospitalization from 8.2 to 7.3 days (P = 0.017). PUFA therapy was also associated with a very low incidence of adverse events.

In summary, although the concept of preventing AF with PUFA is attractive and based on a large body of experimental data, the results of available studies are still equivoqual. Placebo-controlled trials of PUFA are in progress.

Conclusion

On the basis of promising experimental data, numerous retrospective and prospective clinical studies have been undertaken. These have provided information on the potential of upstream therapy for the prevention of AF across a broad spectrum of cardiovascular patient groups. In some of these groups, such as high-risk patient populations, the clinical use of these agents is very tempting; the rationale for using ACEI or ARB, at least in combination with membrane-acting antiarrhythmics, is particularly compelling. However, in most clinical settings examined, the evidence appears to be insufficient to drive changes in management, particularly with respect to statins and PUFA therapy, and additional data from large-scale, randomized, double-blind, placebo-controlled trials with adequately defined endpoints are still needed. Numerous such trials are ongoing (Table 9),101104 reflecting the intense scientific interest in this field. The data derived from these trials are likely to add to our understanding of the complex mechanisms that may lead to AF and its maintenance. It may provide the necessary evidence clarifying the benefit-to-risk ratio of these new therapeutic approaches.

View this table:
Table 9

Ongoing clinical trials of upstream therapies to prevent atrial fibrillation

Title of Registered Clinical TrialSponsorStatus
ACEI/ARB studies
A Prospective, Randomized, Double-Blind Placebo-Controlled Study to Determine the Efficacy of 8 mg/Day Oral Perindopril to Prevent the Recurrence of Atrial Fibrillation in Patients With Essential Hypertension (CTAF-2)Montreal Heart InstituteNot yet recruiting
The Use of DIOVAN to Reduce Post-Cardioversion Recurrence of Atrial Fibrillation Trial (DRAFT Trial) Randomized, Double-Blind, Placebo Controlled StudyIntermountain Health Care Inc.Recruiting
Trial to Investigate the Efficacy of Olmesartan in Paroxysmal Atrial FibrillationGerman AF NetworkRecruiting
GISSI-AF- Use of Valsartan an Angiotension II AT1-Receptor Blocker in the Prevention of Atrial Fibrillation Recurrence - Randomized, double blind, placebo controlled studyGruppo di Ricerca GISSIOngoing, not recruiting
ACTIVE—Substudy evaluating the effects of irbesartan on recurrences of paroxysmal AFSanofi-aventisOngoing, not recruiting
ONTARGET / TRANSCEND (AF: secondary endpoint)—Randomized, Double-Blind, Active ControlBoehringer IngelheimOngoing, not recruiting
A Randomized Clinical Trial of Prevention of Post Ablation Recurrence of Atrial Fibrillation With Angiotensin Receptor Blockers (valsartan)—Randomized, Double-Blind, Placebo ControlQueen Elizabeth II Health Sciences CentreNot yet recruiting
Electrophysiological Effects of Irbesartan in Patients With Paroxysmal Atrial Fibrillation (AF) Randomized, Double-Blind, Placebo Control,Sanofi-aventisCompleted
A Retrospective Analysis of the Use of ACE-Inhibitors and Angiotensin Receptor Blockers on Post-Operative Atrial FibrillationEmory UniversityOngoing, not recruiting
Irbesartan for the Prevention of Atrial Arrhythmias and Cardiac Electrical Remodeling in Patients With Hypertension, Permanent Pacemakers and Risk Factors for Developing Atrial FibrillationConnolly, Stuart, M.D.Not yet recruiting
Impact of Irbesartan on Oxidative Stress and C-Reactive Protein Levels in Patients With Persistent Atrial Fibrillation - Randomized, Double-Blind, Placebo ControlUniversity of MagdeburgNot yet recruiting
RAAS, Inflammation, and Post-Operative AF - Randomized, Double-Blind, Placebo ControlVanderbilt UniversityRecruiting
Left Ventricular Reverse Remodelling After Aortic Valve Replacement in Severe Valvular Aortic Stenosis—Effect of Blockade of the Angiotensin-II Receptor—Randomized, Open Label, Active ControlOdense University HospitalRecruiting
Comparison of Two Treatments to Decrease Inflammation and Atrial Fibrillation Following Cardiopulmonary Bypass SurgeryNHLBIRecruiting
PUFA studies
Arrhythmia Prevention with an Alpha-Linolenic Enriched DietParis HospitalsCompleted
Use of Fish Oils to Prevent Mechanical Stunning and Atrial Remodeling Due to Atrial Arrhythmia Randomized, Single Blind, Placebo ControlMelbourne HealthRecruiting
Use of Fish Oils to Reduce the Frequency and Duration of Episodes of Atrial Fibrillation in Patients with Paroxysmal Atrial Fibrillation—Randomized, Single Blind, Placebo ControlMelbourne HealthRecruiting
Evaluation of Efficacy and Safety of Omacor (Omega-3-Acid Ethyl Esters) in Recurrent, Symptomatic AFReliant PharmaceuticalsRecruiting
Omega-Three Fatty Acids for the Prevention of Atrial Fibrillation After Cardiac Surgery Randomized, Double-Blind, Placebo ControlUniversity of IowaRecruiting
Omega 3 Fatty Acids and Atrial Fibrillation—Randomized, Double-Blind, Placebo ControlUniversity of DundeeRecruiting
Fish Oil for Atrial Fibrillation - Effect and Mechanisms - Randomized, Double-Blind, Placebo ControlNHLBIRecruiting
Randomized, Prospective, Placebo-Controlled, Multi-Center Study to Test the Efficacy of n−3 PUFA for the Maintenance of Normal Sinus Rhythm in Patients With Persistent Atrial FibrillationFundacion GESICANot yet recruiting
Corticosteroids and statin studies
Feasibility of Corticosteroid in the Prevention of Atrial Fibrillation After Cardiac Surgery: A Prospective, Double-Blind Randomized Multicenter Trial (intravenous hydrocortisone)Kuopio University HospitalCompleted
Paroxysmal Atrial Fibrillation: Role of Inflammation, Oxidative Stress Injury and Effect of Statins Randomized, Double Blind (Subject, Investigator), Placebo ControlUniversity of CalgaryRecruiting
The Use of Statins Following a Left Atrial Catheter Ablation Procedure to Prevent Atrial Fibrillation Randomized, Double Blind (atorvastatin)Mayo Clinic
Atorvastatin Therapy for the Prevention of Atrial Fibrillation (SToP-AF)—Randomized, Double-Blind, Placebo ControlEmory UniversityRecruiting
Atorvastatin Decreases Atrial Fibrillation Burden in Patients With Paroxysmal Atrial Fibrillation Randomized, Open Label, Active ControlNational Taiwan University HospitalRecruiting
The Effects of Atorvastatin on the Occurrence of Postoperative Atrial Fibrillation After Off-Pump Coronary Artery Bypass Grafting Surgery—Randomized, Open LabelSamsung Medical CenterCompleted
The Impact of Short-Term High Intensity Statin Therapy vs. Low Intensity Statin Therapy on Inflammatory Mediators in Cardiothoracic Surgery Patients: the Atrial Fibrillation Suppression Trial IV (AFIST IV) Pilot StudyHartford HospitalRecruiting

Funding

Funding for editorial support was provided by sanofi-aventis.

Conflict of interest: Dr Dorian has received consulting fees and research support from sanofi-aventis.

References

View Abstract