Article Text

Download PDFPDF

Prevention of atrial fibrillation with omega-3 fatty acids: a meta-analysis of randomised clinical trials
  1. Tong Liu1,2,
  2. Panagiotis Korantzopoulos3,
  3. Michael Shehata1,
  4. Guangping Li2,
  5. Xunzhang Wang1,
  6. Sanjay Kaul1
  1. 1The Heart Institute, Cedars-Sinai Medical Center, Los Angeles, California, USA
  2. 2Department of Cardiology, Tianjin Institute of Cardiology, Second Hospital of Tianjin Medical University, Tianjin, People's Republic of China
  3. 3Department of Cardiology, University Hospital of Ioannina, Ioannina, Greece
  1. Correspondence to Dr Sanjay Kaul, Division of Cardiology, Cedars-Sinai Medical Center, 8700 Beverly Boulevard, Los Angeles, CA 90048, USA; kaul{at}


Context Previous randomised controlled trials (RCT) regarding n-3 PUFA supplementation for atrial fibrillation (AF) prevention have yielded conflicting results.

Objective A systematic review and meta-analysis of RCT was conducted to examine the role of n-3 PUFA in AF prevention.

Data Sources MEDLINE, Web of Science and Cochrane clinical trials database were searched until November 2010.

Study Selection Of 127 initially identified studies, 10 RCT with 1955 patients were finally analysed.

Data Extraction Two blinded reviewers extracted data independently to a predefined form. Disagreements were resolved through discussion and consensus.

Results n-3 PUFA had no significant effect on the prevention of AF (OR 0.81, 95% CI 0.57 to 1.15; p=0.24). There was significant heterogeneity among the studies (p=0.002, I2=65.0%). Subgroup analysis showed no significant beneficial effect of fish oils in any subset of population.

Conclusions No significant effects of n-3 PUFA supplementation on AF prevention were observed in this meta-analysis. A large-scale trial with higher doses and longer follow-up might be required to rule out the possibility of any treatment benefit.

  • Arrhythmias
  • atrial fibrillation
  • meta-analysis
  • omega-3 fatty acids

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Atrial fibrillation (AF) is a rapidly evolving epidemic with different underlying substrates and serious health consequences.1 2 A growing body of evidence indicates that, apart from the triggers, AF development and perpetuation depends on the electrical and structural remodelling of the atria.2 3 There seems to be an unmet need for the development of pharmacological therapies that target this pathological substrate without having proarrhythmic or other significant untoward effects.4 5 Several non-channel blocking agents with pleiotropic properties are under investigation, but currently there is no solid evidence for their clinical value in the setting of AF.4 5

Omega-3 polyunsaturated fatty acids (n-3 PUFA) and fish oils are believed to exhibit antiarrhythmic and other favourable pleiotropic properties. However, observational studies and clinical data regarding n-3 PUFA supplementation for AF prevention have yielded conflicting results. We therefore conducted a comprehensive meta-analysis of all available randomised controlled trials (RCT) examining the role of n-3 PUFA in AF prevention.


Inclusion criteria

The studies considered in this meta-analysis were RCT that evaluated the role of omega-3 fatty acids in the prevention of AF. The following inclusion criteria were utilised: (1) RCT with a parallel design; (2) comparison of omega-3 fatty acids with control; and (3) evaluating new-onset or recurrent AF as an outcome. We included published and unpublished studies without language restriction. No limits in the length of follow-up were set. For studies that were published only in abstract form, attempts were made to contact the authors for complete details.

Search strategies

We searched MEDLINE, the Web of Science and the Cochrane controlled Trials Register databases to November 2010 for all relevant studies using the following keywords: ‘omega-3 fatty acid’, ‘fish oil’, ‘n-3 fatty acid’ and ‘atrial fibrillation’. Titles and abstracts of all identified trials were assessed independently by two reviewers (TL and PK) and disagreements were resolved by consensus. In addition, a manual search was conducted using all review articles on this topic, bibliographies of original papers and abstracts from the scientific sessions of the American College of Cardiology, the American Heart Association, the European Society of Cardiology and the Heart Rhythm Society during the past 5 years. Finally, we contacted relevant experts and pharmaceutical companies to obtain unpublished data.

Quality assessment and data extraction

Two blinded reviewers (TL and PK) independently assessed the methodological quality of included RCT using the Jadad scale,6 which included randomisation, blinding, description of withdrawals and dropouts, methods of randomisation and double-blinding status. The quality scale ranges from 0 to 5 points, with a score of 2 or less qualifying as low quality and a score of 3 or more as a high quality study.

Two blinded reviewers (TL and PK) also extracted the data independently to a predefined form. The following data were collected from each study: (1) publication data: first author's last name, country of origin, year of publication; (2) study design; (3) characteristics of the study population: sample size, age, gender, diagnoses; (4) follow-up period, number of withdrawals and dropouts; (5) treatment modality: type, dose, duration of administration; (6) endpoint measurement: definition and detection methods of AF; whether AF endpoint was adjudicated by a blinded external committee; and (7) adverse events related to omega-3 fatty acids use. Disagreements were resolved through discussion and consensus.

Statistical analysis

Results of the AF outcome are expressed as OR with 95% CI for each study. Heterogeneity was first examined using the standard χ2 test of heterogeneity. As this test has limited power in the case of few studies, we considered the presence of significant heterogeneity at the 10% level of significance, and values of I2 exceeding 50% as an indicator of the significant impact of heterogeneity.7 If the χ2 test for heterogeneity was significant, a pooled effect was calculated with a random-effects model (DerSimonian–Laird method), which was used to take into account within-study and between-study variance. Otherwise, a fixed-effects Mantel–Haenszel model was used to derive the pooled estimates. Sensitivity analysis was performed in a random predefined manner to assess the impact of the following: (1) open-label studies; (2) intravenous omega-3 fatty acid preparation; (3) studies published only in abstract form; (4) low-quality studies (Jadad score ≤2); and checking the consistency of the overall effect estimate. We also performed prespecified subgroup analysis in different populations to explore the possible reasons for heterogeneity. Publication bias was evaluated using the funnel plot. Statistical significance for treatment effect was defined at p values less than 0.05. All analyses were performed using Review Manager version 4.2.


One hundred and twenty-seven records were identified by the primary literature search. However, after screening the titles and abstracts, 102 studies were excluded because they were either laboratory studies, review articles, or irrelevant to the study objective. We therefore retrieved 25 potentially relevant manuscripts for detailed review. Fifteen studies were further excluded for the following reasons: eight of the studies were observational studies,8–15 two were rationale and design reports on ongoing RCT,16 17 one RCT compared intervention with active control,18 and one did not evaluate clinical AF as an outcome,19 the last three were duplicate published abstracts. Consequently, the remaining 10 RCT were included for the meta-analysis (figure 1).

Figure 1

Flow diagram of the trial selection process. AF, atrial fibrillation; RCT, randomised controlled trial.

Of the 10 included RCT, six were published as full manuscripts20–23 24 25 and four as abstracts.24 26 27 28 Four papers published as manuscripts and one as an abstract evaluated new-onset AF following cardiac surgery,20–23 26 whereas two full-text papers and three abstracts examined the preventive effects of fish oil on recurrent AF either in the post-electrical cardioversion setting24 25 27 28 or in patients with symptomatic paroxysmal or persistent AF.29 The detailed characteristics of the RCT are presented in table 1. The follow-up and AF detection information are summarised in table 2. Overall, 1955 patients were included in our analysis (977 in the treatment group and 978 in the control group). The follow-up duration ranged from during hospitalisation to 14 days in five studies evaluating new-onset AF following cardiac surgery,20–23 26 and 0.5–1 year in five studies examining recurrent AF.24 25 27–29 The quality assessment and Jadad score evaluation of the 10 included studies are listed in table 3.

Table 1

Characteristics of RCT included in the meta-analysis

Table 2

Follow-up and AF detection in randomised controlled studies in the meta-analysis

Table 3

The interpretations of the Jadad score evaluation

The pooled estimate showed no significant effect of omega-3 fatty acids on the prevention of AF (OR 0.81, 95% CI 0.57 to 1.15; p=0.24). The results are consistent with treatment effects ranging from as much as a 43% odds reduction to a 15% odds increase in AF recurrence. There was significant heterogeneity among the studies (p=0.002, I2=65.0%; figure 2). We subsequently performed sensitivity and subgroup analyses to explore the source of this heterogeneity. None of the sensitivity analyses yielded significant results in favour of omega-3 fatty acids (table 4). Subgroup analysis showed that in five studies20–23 26 evaluating new-onset AF following cardiac surgery, there was no significant beneficial effect of fish oils (OR 0.78, 95% CI 0.48 to 1.27; p=0.32; pheterogeneity=0.06, I2=56.7%). Similarly, in the remaining five RCT24 25 27–29 that examined recurrent AF, there was no significant beneficial effect of fish oils (OR 0.83, 95% CI 0.48 to 1.45; p=0.51; pheterogeneity=0.004, I2=73.9%; figure 2). The funnel plot (figure 3) suggested that there was little publication bias, although the small number of studies makes interpretation difficult. Other sources of heterogeneity include differences in study populations and follow-up periods, different dosage, duration and type of omega-3 fatty acid (table 1)

Figure 2

Effects of omega-3 fatty acids on the development of atrial fibrillation. AF, atrial fibrillation.

Table 4

OR and heterogeneity tests for sensitivity analyses

Figure 3

Funnel plot of the meta-analysis.


This comprehensive meta-analysis of RCT failed to show any favourable effect of n-3 PUFA supplementation on the incidence of AF either postoperatively or after elective electrical cardioversion. Also, significant heterogeneity across the studies was evident.

Observational studies examining the association between dietary fish intake and incident AF have yielded conflicting results. The investigators of the Cardiovascular Health Study were the first to report an association between fish intake and reduced AF risk.8 However, subsequent studies such as the Danish Diet, Cancer, and Health Study,9 the Rotterdam Study,10 as well as the recently published Women's Health Initiative Study14 failed to show any benefit of increased fish consumption in the incidence of AF. These differences could be attributed to different definitions of AF, diverse methods of AF detection, as well as lower levels of fish consumption and low event rates in the three negative studies.14

n-3 PUFA represent active components of fish oils and are believed to exert beneficial pleiotropic effects. In addition, there is substantial evidence suggesting their favourable safety profile. Despite limited data regarding their efficacy in AF, the lack of cardiac safety concern has fuelled research in this field. In this context, several prospective randomised studies including FORomegaARD,16 OPERA and AFFORD involving nearly 4000 patients are underway to evaluate the efficacy of n-3 PUFA in this setting (table 5).

Table 5

Ongoing RCT on prevention of AF with omega-3 fatty acids

Atrial electrical and structural remodelling is widely recognised as a critical process in the pathophysiology of AF that contributes to the development and perpetuation of arrhythmia.2 3 The role of inflammation and oxidatve stress in the pathophysiology of atrial remodelling has recently received attention.30 31 There is evidence suggesting that fish or n-3 PUFA consumption attenuates inflammation and oxidative stress in humans.32 33 Apart from the anti-inflammatory and antioxidant properties, n-3 PUFA may modulate cardiac ion channels providing electrical stability to the cell membrane, decrease calcium overload and apoptosis, modulate ectopic electrical activity, reduce the development of myocardial fibrosis and favourably affect the autonomic tone promoting the vagal tone.34 All of these would be expected to exert a beneficial effect on AF.

Theoretically, postoperative AF represents an ideal model to test the efficacy of agents with anti-inflammatory properties given that the peak incidence of AF after cardiac surgery coincides with the peak concentration of inflammatory indices.35 In keeping with this assumption, statins, drugs with known anti-inflammatory properties, seem to exhibit greater antiarrhythmic effect in the setting of postoperative AF.36 The mechanisms involved in postoperative AF may be different from those underlying AF recurrences after cardioversion of persistent AF. In particular, electrical and structural remodelling is more prominent in persistent AF in which diseased atria often exhibit extensive fibrosis. On the other hand, postoperative AF may develop even in normal atria mainly due to perturbations of the sympathetic tone, oxidative reperfusion injury and immune and inflammatory activation related to the surgical procedure. However, the results of this meta-analysis show that n-3 PUFA do not have any meaningful effect on the incidence of postoperative AF. Of note, there seem to be important differences in the study design, route of administration, electrocardiographic monitoring and baseline characteristics between the five RCT examining postoperative AF (tables 1 and 2). For example, some investigators point out that significant effects of supplementation occur at relatively low baseline n-3 PUFA plasma concentrations.22 Also, administration through a nasogastric tube20 or intravenous administration21 may provide improved bioavailability. There is also controversy as to whether high-concentration circulating n-3 PUFA may have better antiarrhythmic potential compared with membrane incorporated n-3 PUFA.17 23 In fact, the complex interaction between circulating and tissue levels of n-3 PUFA and the resulting electrophysiological alterations are incompletely understood and may partly explain the variability in results across clinical trials.20–23 26 35 Interestingly, Metcalf et al37 have demonstrated that fish oil supplementation in humans causes a slow gradual increase in tissue levels of n-3 PUFA over a 2-month period, whereas plasma levels rise abruptly and remain constant thereafter. It could therefore be hypothesised that the short time course of administration before cardioversion or before cardiac surgery does not allow sufficient incorporation of n-3 PUFA into tissues.

Very recently, one large RCT in 663 patients with paroxysmal or persistent AF and no substantial structural heart disease reported that prescription-only omega-3 PUFA capsules (Lovaza; GlaxoSmithKline, Brentford, Middlesex, UK) even at the fairly high dose of 4 g/day for 6 months did not cut the risk of recurrence despite evidence of biological activity (increase in blood levels of docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA), and reduction in plasma triglycerides).29 Of note, almost half of the recurrences occurred within the first 2 weeks of follow-up,29 thereby suggesting insufficient time for the build up of tissue levels.

An important issue that needs further clarification is the relative impact of the underlying comorbidities on the effect of n-3 PUFA. It could be reasonable to assume that patients with an increased burden of comorbidities and significant structural heart disease would benefit the most from prophylactic treatment. In this context it is conceivable that n-3 PUFA would be beneficial if administered at an earlier stage to prevent irreversible electrical and structural remodelling, which might contribute to AF recurrences.

Finally, we should note that n-3 PUFA are found in fish oils, but are not actually ‘fish oils’. Important PUFA in nutrition included α-linolenacid, EPA and DHA. The most commonly administered n-3 PUFA are EPA and DHA, manufactured by phyto plankton, eaten by and concentrated in the tissues of fatty fish. n-3 PUFA used in clinical studies are not necessarily only derived from fatty fish but can come from non-marine sources (eg, DHA can be synthesised by algae). In the 10 RCT included in our meta-analysis, most of them used n-3 PUFA with an average EPA/DHA ratio of 1.2. Only one study24 from Germany selected 301.5 mg α-linolenacid as the interventional arm, which yielded a negative result in preventing AF recurrences after cardioversion of persistent AF. Two studies,27 28 published as abstracts, did not specify the type of n-3 PUFA used in the trials. It is unclear whether the benefit, if any, from n-3 PUFA resides in the EPA moiety, the DHA moiety, or both.


Our meta-analysis has important limitations. Half of the studies included in the present analysis have been published in the abstract form, and we were unable to have access to the full data. Although our results do not point to a specific positive or negative effect of n-3 PUFA on AF prevention, they stress the uncertainty and the heterogeneity of the published data, emphasising the need for further controlled studies in this field. Furthermore, we have to acknowledge that meta-analysis represents a retrospective research tool that is subject to the methodological deficiencies of the included studies. Finally, it is necessary to mention the relative limited power of our meta-analysis to find a treatment benefit. For example, the post-cardiac surgery evaluations were powered to detect implausibly high treatment effects (33–55% RRR), arguably driven by sample size considerations rather than expected real treatment effects. Given a baseline AF rate of 30–40% in the postoperative setting, a trial would require nearly 1000–1500 patients to detect a 25% RRR with n-3 PUFA treatment. However, the pooled sample size for the five trials in our meta-analysis was one-half to two-thirds of this required sample size. Therefore, the possibility that our pooled estimates represent a type II (false negative) error cannot be ruled out. A similar fate befell two recently published trials—the Alpha Omega Trial38 that enrolled 4837 patients with previous myocardial infarction (MI) and the OMEGA trial that enrolled 3804 patients 3–14 days post-acute MI.39 Low-dose supplementation with n-3 PUFA failed to reduce cardiovascular events, ventricular arrhythmias, or sudden cardiac death during a 12–40-month follow-up period in those trials. Both studies were underpowered to answer the question about the role of n-3 fatty acids in the post-MI setting definitively. This along with the more aggressive risk factor management in the current era might explain the discrepancy with previous studies (DART,40 JELIS,41 GISSI-P42 and GISSI-HF43 trials) that suggested n-3 fatty acids, particularly EPA, might be cardioprotective following MI. Furthermore, the dose of n-3 PUFA supplementation used in most of the studies might be insufficient (neutraceutical dose ranges from 3 to 4 g daily) to modify inflammation, or other risk or causal factors favourably.


Our meta-analysis of RCT did not demonstrate significant benefit of n-3 PUFA supplementation on AF prevention. However, given that fish oils and n-3 PUFA are well tolerated without major adverse effects, an adequately powered trial in the right clinical setting (post-operative AF) using higher doses such as the OPERA trial is warranted to rule out the possibility of any treatment benefit.


The authors are indebted to Dr Matthias Bayer (University of Giessen, Giessen, Germany), who kindly provided additional data on their published studies.



  • See Editorial, p 1032

  • Linked articles 221507.

  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Linked Articles