Skip main navigation

Effect of Long-Term Marine ɷ-3 Fatty Acids Supplementation on the Risk of Atrial Fibrillation in Randomized Controlled Trials of Cardiovascular Outcomes: A Systematic Review and Meta-Analysis

Originally publishedhttps://doi.org/10.1161/CIRCULATIONAHA.121.055654Circulation. 2021;144:1981–1990

Abstract

Background:

Some, but not all, large-scale randomized controlled trials (RCTs) investigating the effects of marine ɷ-3 fatty acids supplementation on cardiovascular outcomes have reported increased risks of atrial fibrillation (AF). The potential reasons for disparate findings may be dose-related.

Methods:

The MEDLINE and Embase databases were searched for articles and abstracts published between January 1, 2012, and December 31, 2020, in addition to a meta-analysis of large cardiovascular RCTs published in 2019. RCTs of cardiovascular outcomes of marine ɷ-3 fatty acids that reported results for AF, either as a prespecified outcome, an adverse event, or a cause for hospitalization, with a minimum sample size of 500 patients and a median follow-up of at least 1 year were included. RCTs specifically examining shorter-term effects of ɷ-3 fatty acids on recurrent AF in patients with established AF or postoperative AF were not included. The hazard ratio (HR) for the reported AF outcomes within each trial was meta-analyzed using random effects model with Knapp-Hartung adjustment and evaluated a dose-response relationship with a meta-regression model.

Results:

Of 4049 screened records, 7 studies were included in the meta-analysis. Of those, 5 were already detected in a previous meta-analysis of cardiovascular RCTs. Among the 81 210 patients from 7 trials, 58 939 (72.6%) were enrolled in trials testing ≤1 g/d and 22 271 (27.4%) in trials testing >1 g/d of ɷ-3 fatty acids. The mean age was 65 years, and 31 842 (39%) were female. The weighted average follow-up was 4.9 years. In meta-analysis, the use of marine ɷ-3 fatty acid supplements was associated with an increased risk of AF (n=2905; HR, 1.25 [95% CI, 1.07–1.46]; P=0.013). In analyses stratified by dose, the HR was greater in the trials testing >1 g/d (HR, 1.49 [95% CI, 1.04–2.15]; P=0.042) compared with those testing ≤1 g/d (HR, 1.12 [95% CI, 1.03–1.22]; P=0.024; P for interaction <0.001). In meta-regression, the HR for AF increased per 1 g higher dosage of ɷ-3 fatty acids dosage (HR, 1.11 [95% CI, 1.06–1.15]; P=0.001).

Conclusions:

In RCTs examining cardiovascular outcomes, marine ɷ-3 supplementation was associated with an increased risk of AF. The risk appeared to be greater in trials testing >1 g/d.

Clinical Perspective

What Is New?

  • In this updated meta-analysis of 7 randomized controlled trials including 81 200 patients, marine ɷ-3 fatty acid supplementation was associated with a significant increased risk of atrial fibrillation (AF) compared with placebo.

  • The incremental risk of AF associated with ɷ-3 fatty acid appeared to be greater in trials testing >1 g/d of ɷ-3 fatty acid supplementation.

What Are the Clinical Implications?

  • The potential risk of developing AF should be discussed with the patients when prescribing marine ɷ-3 supplementation, especially when prescribing a higher dosage.

  • Postmarketing surveillance for AF along with systematic ascertainment of AF outcomes in future trials of marine ɷ-3 supplementation will be needed to better define the risk-benefit ratio across ɷ-3 formulations.

Editorial, see p 1991

Marine ɷ-3 fatty acids supplements may have a beneficial effect on the risk of atherosclerotic cardiovascular events1–4; however, concerns have also been raised regarding potential off target adverse effects on atrial fibrillation (AF) risk within these trials. The Reduction of Cardiovascular Events with Icosapent Ethyl-Intervention Trial (REDUCE-IT) reported a decrease of 25% in major cardiovascular events with the use of 4 g of icosapent ethyl over a median follow-up of 4.9 years.2 On the basis of these results, the 2019 European Society of Cardiology and European Atherosclerosis Society of dyslipidemia guidelines recommended the use of 4 g of icosapent ethyl in patients with established cardiovascular disease with triglycerides between 135 to 499 mg despite statin treatment.5 However, REDUCE-IT also reported an increase in a prespecified tertiary outcome of AF hospitalization in those randomized to active treatment compared with placebo (3.1% versus 2.1%; P=0.004).2 Subsequently, several randomized controlled trials (RCTs) of marine ɷ-3 fatty acid supplementation have reported results for AF, but AF case numbers have generally been small.6,7 The study with the largest number of AF events, VITAL (Vitamin D and OmegA-3 Trial) Rhythm, did not find a significant increased risk of incident AF with 1 g/d of marine ɷ-3 fatty acids (460 mg of eicosapentaenoic acid and 380 mg of docosahexaenoic acid [DHA]) compared with placebo (HR, 1.09 [95% CI, 0.96–1.24]; P=0.19).8 In light of these conflicting findings, there is a need to summarize the overall effect of marine ɷ-3 fatty acid supplements on AF across these studies and explore whether marine ɷ-3 fatty acid dose might account for the seemingly disparate results.

Methods

Selection Criteria and Search Strategy

This meta-analysis was registered to PROSPERO (CRD42021234291). The PRISMA guidelines (Preferred Reporting Items for Systematic Review and Meta-Analyses) were followed for this systematic review and meta-analysis.9 All the data from the original articles that were extracted for this meta-analysis are publicly available. The authors declare that all supporting data are available within the article and its Supplemental Material. B.G. and O.T.A. searched MEDLINE and Embase for all randomized, controlled, double-blind, cardiovascular outcome trials of marine ɷ-3 fatty acid supplements of >500 patients with a minimum follow-up of at least 1 year1,10 that reported AF events as primary, secondary, exploratory, or safety (adverse events or cause of hospitalization) outcomes. Trials that specifically tested the effect of ɷ-3 fatty acids on postoperative AF or recurrent AF in patients with established AF were excluded because these hypotheses regarding potential short-term antiarrhythmic effects of ɷ-3 fatty acids had been addressed in previous meta-analyses.11,12 On the basis of the results of these trials, marine ɷ-3 supplements are not recommended for these indications.5,13

The search was begun in 2012 after the last systemic review on RCTs specifically examining ɷ-3 fatty acids and atrial fibrillation.11 Records were searched between January 1, 2012, and December 31, 2020, without any language restrictions. Data on available AF outcomes were also extracted from original trials that were screened in a previous meta-analysis of randomized controlled trials of marine ɷ-3 fatty acids and atherosclerotic cardiovascular outcomes.1 The VITAL Rhythm study that was presented at the American Heart Association annual meeting in 2020 and published in 2021 by our group was also included.8 B.G. and O.T.A. screened titles, abstracts, and full texts of articles identified in this search and assessed the risk of bias using the Cochrane tool. B.G. and O.T.A. extracted the data for eligible studies using a standardized data form for aggregated study level, and discrepancies were resolved by consensus. Because this meta-analysis was based on data extracted from previously published research, the data and study materials are available in the public domain. For further details on the algorithm used for literature search, see the Supplemental Methods.

Annual Event Calculations

The annual event rate was estimated by dividing the total number of events by the product of the total number of persons and the median follow-up duration in each trial population. In 1 trial (OMEMI [Omega-3 Fatty Acids in Elderly With Myocardial Infarction]), the median follow-up was not provided, and the maximal follow-up duration completed by 97.8% of patients was used.7 The annual event rate was estimated by dividing the total number of events with the product of the total number of persons and the median follow-up duration in each trial according to the weight assigned in the primary random effect model (Figure 1).14

Figure 1.

Figure 1. Effect of marine ɷ-3 fatty acids supplements on the risk of atrial fibrillation events using Knapp-Hartung adjustment for random effect model. ASCEND indicates A Study of Cardiovascular Events in Diabetes; GISSI-HF, Gruppo Italiano per lo Studio della Sopravvivenza nell’Insufficienza Cardiaca; HR, hazard ratio; OMEMI, Omega-3 Fatty Acids in Elderly With Myocardial Infarction; REDUCE-IT, Reduction of Cardiovascular Events With Icosapent Ethyl-Intervention Trial; RP, Risk and Prevention Study; STRENGTH, Long-Term Outcomes Study to Assess Statin Residual Risk Reduction With EpaNova in High Cardiovascular Risk Patients With Hypertriglyceridemia; and VITAL, Vitamin D and Omega-3 Trial.

Data Analysis

Outcomes from each trial were selected to target the definition of AF events used within each trial. The HR or risk ratio (RR) and 95% CIs were extracted as reported in original articles. When the original HR or RR was not available, the RR was calculated using the cumulative incidence in each group (csi stata command). A random effects DerSimonian-Laird meta-analysis was used for the primary analysis to account for heterogeneity across included trials that may be a result of marine ɷ-3 fatty acids dosage, follow-up duration, and study population (primary versus secondary prevention, general population versus elderly with myocardial infarction, and varying prevalence of preexisting AF). The Knapp-Hartung adjustment was used to account for uncertainty in the between-study variance estimate.15 The heterogeneity was assessed using Cochran’s Q statistic, and Higgins and Thompson’s I2, as well as average dispersion in effect sizes τ2. Because the Knapp-Hartung estimates are conservative when the heterogeneity is low with a small number of studies,16,17 a fixed effects model was used when no heterogeneity was observed (I2=0%) as a sensitivity analysis. The risk of bias was assessed according to the Cochrane tool used to assess bias in randomized clinical trials. Publication bias was assessed with a funnel plot,18 as well as the trim and fill method.19 Because the test for a publication is not advised for fewer than 10 studies, Egger’s test was not performed.20

In a secondary analysis, RCTs were stratified by low dose (≤1 g/d) versus high dose (>1 g/d) of marine ɷ-3 supplements.15 Exploratory analyses examining the linear association between the dosage of ɷ-3 fatty acids and the hazard ratio for AF were performed using meta-regression where the intercept was set at 0 to reflect the clinical assumption that as the dose goes to 0, the lipid effect goes to 0, and the treatment effect goes to 1 (neutral effect) assuming linearity.21 In a sensitivity analysis, a constant term was included. The corresponding HRs with 95% CIs and P values per 1 g increase in ɷ-3 fatty dosage acids are reported. In addition to the dose analyses, 4 additional sensitivity analyses stratified RCTs by whether AF was a prespecified outcome, baseline AF was excluded, hospitalization for AF was the only AF outcome, and DHA was included in the study intervention.1,10 Statistical analysis was done with Stata 16.0 using the meta suite commands.

Role of the Funding Sources

The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the article; or decision to submit the article for publication. All authors had full access to all the data in the study, and the corresponding author had final responsibility for the decision to submit for publication.

Results

After screening 4059 records (Figure S1 for the PRISMA flowchart), data from 7 articles were included in the systematic review and meta-analysis. Of those, 5 trials were already included in a previous meta-analysis of cardiovascular RCTs.1 All trials included reported obtaining approval from institutional review committees and that all participants gave informed consent (Table 1).2,6–8,22–24 Among the 81 210 patients from 7 trials, 58 939 (72.6%) were enrolled in trials testing low-dose marine ɷ-3 fatty acids (≤1 g/d) and 22 271 (27.4%) in trials testing high-dose of marine ɷ-3 fatty acids (>1 g/d). Trials testing low doses of marine ɷ-3 fatty included the VITAL Rhythm study,8 ASCEND (A Study of Cardiovascular Events in Diabetes),22,23 GISSI-HF (Gruppo Italiano per lo Studio della Sopravvivenza nell’Insufficienza Cardiaca),24 and the RP (Risk and Prevention Study; Table 2).25 Trials testing a high dose of marine ɷ-3 fatty acids were REDUCE-IT,2 STRENGTH (Long-Term Outcomes Study to Assess Statin Residual Risk Reduction With EpaNova in High Cardiovascular Risk Patients With Hypertriglyceridemia),6 and OMEMI.7 The weighted mean age was 65 years, and 31 842 (39.2%) were female (Table 1). The weighted median follow-up was 4.9 years. All the trials met the criteria for low risk of bias according to the Cochrane tool for assessing risk of bias in randomized clinical trials (Table S1).

Table 1. Study Characteristics

TrialsSample sizeAF at baseline included in the analysisMean age, yFemale patients, n (%)White patients, n (%)Intervention armControl armMedian follow-up, yTotal AF events
(% per year)
Type of AF event
VITAL825 119No6712 757 (51)17 425 (71)1 g/d of ɷ-3 fatty acids
(460 mg EPA and 380 mg DHA)
Olive oil5.3900 (0.7)New-onset AF
ASCEND*22,2315 480*Yes635796 (37)14 935 (96)1 g/d of ɷ-3 fatty acids
(460 mg EPA and 380 mg DHA)
Olive oil7.4301 (0.3)Main article: patient-reported AF (without excluding preexisting AF); research letter: new-onset AF
STRENGTH613 078Yes62.54568 (35)10 723 (82)4 g/d of ɷ-3 fatty acids
(EPA and DHA)
Corn oil3.5230 (0.5)New-onset AF
RP2512 505Yes644818 (39)NA1 g/d of ɷ-3 fatty acids
(EPA and DHA <85% in ratio from 0.9:1 to 1.5:1)
Olive oil5205 (0.3)Hospitalization for AF (without excluding preexisting AF)
REDUCE-IT28179Yes642357 (29)7379 (90)4g per day of icosapent ethyl (ethyl ester of EPA)Mineral oil4.9374 (0.9)New-onset or worsening AF events (without excluding preexisting AF).
GISSI-HF245835No661252 (21)NA1 g/d of ɷ-3 fatty acids (850–882 mg of EPA and DHA as ethyl esters at an average ratio of 1.2:1)Olive oil3.9852 (3.7)New-onset AF
OMEMI71014No75294 (29)1012 (100)1.8 g/d of ɷ-3 fatty acids (930 mg EPA and 660 mg DHA)Corn oil243 (2.1)New-onset AF
Total81 2106531 842 (39)4.92905 (1.3)

AF indicates atrial fibrillation; ASCEND, A Study of Cardiovascular Events in Diabetes; DHA, docosahexaenoic acid; EPA, eicosapentaenoic acid; GISSI-HF, Gruppo Italiano per lo Studio della Sopravvivenza nell’Insufficienza Cardiaca; NA, not available; OMEMI, Omega-3 Fatty Acids in Elderly With Myocardial Infarction; REDUCE-IT, Reduction of Cardiovascular Events With Icosapent Ethyl-Intervention Trial; RP, Risk and Prevention Study; STRENGTH, Long-Term Outcomes Study to Assess Statin Residual Risk Reduction With EpaNova in High Cardiovascular Risk Patients With Hypertriglyceridemia; and VITAL, Vitamin D and Omega-3 Trial.

* Patient-reported AF adverse events were extracted from the original publication of ASCEND trial. The post hoc research letter reporting a more comprehensive assessment of AF (N=1177) based on electronic health records was used as sensitivity analysis.

† The total sample size of the trial was 1014, but 759 were included in the AF analysis after excluding those with prevalent AF.

‡ Only total duration of the trial was reported.

Table 2. Summary of AF Events Outcome Assessment and Reported Treatment Effect

StudyOutcome assessmentReported treatment effect
VITAL8Incident AF cases were identified through self-reported diagnosis and linkage with claims data for the Centers for Medicare and Medicaid Services. An end point committee consisting of cardiologists reviewed medical records and confirmed AF events according to predefined criteria (AF events adjudicated).469/12 542 in the experimental arm vs 431/12 577 in the control arm.
Hazard ratio was 1.09 (95% CI, 0.96–1.24; P=0.19).
ASCEND (main article)22Patient-reported AF was included as an exploratory vascular outcome, but only primary and secondary outcomes of the trial were adjudicated centrally by clinicians (AF events were not adjudicated).166/7740 (2.1%) in the experimental arm vs 135/7740 (1.7%) in the control arm.
Rate ratio was 1.23 (95% CI, 0.98–1.54).
ASCEND (research letter)23More comprehensive assessment of AF events, using additional data extracted from linked electronic health records. AF events were not adjudicated. AF diagnoses in hospital episodes before randomization were used to define previously known AF. Arrhythmia outcomes considered are AF among participants without any previously known AF.Among the remaining 99%, AF was recorded from either electronic health records or participant reports in 1177 participants, compared with 287 by self-report alone. AF occurred in 7.7% of participants in the experimental arm and in 7.6% in the placebo arm, with a nonsignificant rate ratio of 1.02 (95% CI, 0.91–1.15).
STRENGTH6New-onset investigator-reported AF was prespecified as a tertiary outcome, but only the primary and secondary outcomes of the trial were centrally adjudicated by a core laboratory (AF events not adjudicated).144/6539 (2.2%) in the experimental arm vs 86/6539 (1.3%) in the placebo arm.
Hazard ratio was 1.69 (95% CI, 1.29–2.21; P<0.001).
RP25AF reported as a reason for hospital admission for cardiovascular disease, but only the primary outcome was adjudicated (AF events nonadjudicated).113/6239 (1.8%) in the experimental arm vs 92/6266 (1.5%) in the placebo arm.
Hazard ratio was 1.22 (95% CI, 0.93–1.61; P=0.15).
REDUCE-IT2REDUCE-IT reported the rates for the prespecified adjudicated tertiary outcome of hospitalization for AF or flutter. The rates were significantly higher in the icosapent ethyl group compared with the placebo group (3.1% vs 2.1%; P=0.004). Because the required numbers to derive treatment effect were not provided in the article, data were extracted for patient-reported treatment adverse AF events (AF unadjudicated) defined as new-onset or worsening AF since initiation of drug therapy.215/4089 (5.3%) AF events in the experimental arm vs 159/4090 (3.9%) in the placebo arm (P=0.003).
The calculated risk ratio was 1.35 (95% CI, 1.11–1.65; P=0.003)
GISSI-HF24Incident AF was reported in a population without baseline AF. New AF events during the trial were defined as follows: AF in the ECGs taken at each visit during the trial, as an event occurring between visits causing or worsening HF/hospital admission, or as an event occurring while in hospital. The adjudication was not done.444/2921 (15.2%) new AF events in the experimental arm vs 408/2914 (14.0%) in the control arm.
Hazard ratio was 1.10 (95% CI, 0.96–1.25; P=0.19).
OMEMI7New AF was a prespecified secondary outcome, defined as a standard 12-lead ECG recording or a single-lead ECG tracing >30 showing no discernible repeating P waves and irregular rate ratio intervals. The assessment for AF events was done with clinical records and ECGs taken at study visits. In addition, patients were screened with ambulant single-lead rhythm monitoring for for 30 seconds twice daily for 14 days. All outcomes were adjudicated centrally by an independent end point committee of experienced clinicians blinded to treatment allocation (AF events adjudicated).28/387 (7.2%) in the experimental arm vs 15/372 (4.0%) in the placebo arm restricted in those with no previous AF.
Hazard ratio was 1.84 (95% CI, 0.98–3.45; P=0.06).

AF indicates atrial fibrillation; ASCEND, A Study of Cardiovascular Events in Diabetes; GISSI-HF, Gruppo Italiano per lo Studio della Sopravvivenza nell’Insufficienza Cardiaca; OMEMI, Omega-3 Fatty Acids in Elderly With Myocardial Infarction; REDUCE-IT, Reduction of Cardiovascular Events With Icosapent Ethyl-Intervention Trial; RP, Risk and Prevention Study; STRENGTH, Long-Term Outcomes Study to Assess Statin Residual Risk Reduction With EpaNova in High Cardiovascular Risk Patients With Hypertriglyceridemia; and VITAL, Vitamin D and Omega-3 Trial.

Four studies reported on new-onset AF events or excluded patients with prevalent AF from the analysis. The ASCEND trial excluded participants with established cardiovascular disease or those treated with anticoagulants, whereas 2 trials (REDUCE-IT and RP) did not specify whether events were new-onset. A post hoc research letter from the ASCEND trial reported AF events after excluding those with preexisting AF at baseline.23 Five studies reported HR and 95% CI for AF events, whereas REDUCE-IT reported the rates for the prespecified adjudicated tertiary outcome of hospitalization for AF or flutter, and the ASCEND trial reported the rate ratio of AF events. In the REDUCE-IT trial, the rates were significantly higher in the icosapent ethyl group compared with the placebo group (3.1% versus 2.1%; P=0.004); however, the exact numbers were not provided in the original article to derive the effect size. Instead, the published treatment-emergent adverse event rates, defined as an event that first occurs or worsens in severity on or after the date of dispensing study drug, were used for AF outcomes (215/4089 [5.3%] in the experimental arm versus 159/4090 [3.9%] in the placebo arm; P=0.003) in the meta-analysis to calculate the RR in the REDUCE-IT trial (1.35 [95% CI, 1.11–1.65]; P=0.003). In RP, AF events were listed as a reported reason for cardiovascular disease hospitalization.25 In GISSI-HF, the analysis reporting new-onset AF events was a post hoc analysis in an ancillary article.24 In the ASCEND trial, the original publication of the main trial presented patient-reported adverse outcomes caused by AF in the overall population with a rate ratio Rhythm,22 whereas a post hoc research letter used a more comprehensive review of electronic health records in patients without known AF (N=15 374, 99% of the population).23 For the ASCEND trial, data from the original article were used in the primary analysis,22 and data from the research letter were included in a sensitivity analysis.23

The AF outcomes were prespecified in the methods of the VITAL, OMEMI, and STRENGTH trials. For the remaining trials, the assessment of AF events was not prespecified in the methods of the original article. The extracted AF outcomes were centrally adjudicated by a panel of clinicians in the VITAL Rhythm and OMEMI trials, whereas in the REDUCE-IT, ASCEND, STRENGTH, GISSI-HF, and RP trials, the AF outcomes were not.

Of the 81 210 participants included in this analysis, 2905 (3.6%) patients had an AF outcome during the trial period, of which 2258/2905 (77.8%) occurred in trials testing a low-dose of ɷ-3 fatty acids supplements and 647/2905 (22.2%) occurred in trials testing a high-dose of ɷ-3 fatty acids. The pooled HR for the association between marine ɷ-3 fatty acids and AF was 1.25 (95% CI, 1.07–1.46; P=0.013; Figure 1). The heterogeneity across the studies was moderate (I2=54.57%). The heterogeneity within the groups decreased substantially in analysis stratified by dose (I2<0.01% for low-dose and I2=9.90% for high-dose marine ɷ-3 fatty acids). The pooled HR for AF events compared with placebo was higher in trials testing a high-dose (HR, 1.49 [95% CI, 1.04–2.15]; P=0.042) than in those testing a low-dose of marine ɷ-3 fatty acids (HR, 1.12 [95% CI, 1.03–1.22]; P=0.024; Figure 2; P for interaction <0.001) with Knapp-Hartung adjustment. In the meta-regression model examining the linear association between ɷ-3 fatty acid dosage and risk of AF events, the HR was 1.11 (95% CI, 1.06–1.15; P=0.001; Figure 3) per 1 g increase of ɷ-3 fatty acids (residual heterogeneity, I2=0.00%; P=0.54) over a range of dosage from 1 g/d to 4 g/d. The sensitivity analysis including the constant term gave similar results (HR, 1.09 [95% CI, 1.01–1.18]; P=0.030 per 1 g increase of ɷ-3 fatty acids dosage).

Figure 2.

Figure 2. Effect of marine ɷ-3 fatty acids supplements on the risk of atrial fibrillation events stratified by low dose (≤1 g/d) versus high dose (>1 g/d) using Knapp-Hartung adjustment for random effect model. ASCEND indicates A Study of Cardiovascular Events in Diabetes; GISSI-HF, Gruppo Italiano per lo Studio della Sopravvivenza nell’Insufficienza Cardiaca; HR, hazard ratio; OMEMI, Omega-3 Fatty Acids in Elderly With Myocardial Infarction; REDUCE-IT, Reduction of Cardiovascular Events With Icosapent Ethyl-Intervention Trial; RP, Risk and Prevention Study; STRENGTH, Long-Term Outcomes Study to Assess Statin Residual Risk Reduction With EpaNova in High Cardiovascular Risk Patients With Hypertriglyceridemia; and VITAL, Vitamin D and Omega-3 Trial.

Figure 3.

Figure 3. Regression of ɷ-3 fatty acids dosage and risk for atrial fibrillation events in 7 randomized controlled trials using Knapp-Hartung adjustment for random effect model. ASCEND indicates A Study of Cardiovascular Events in Diabetes; GISSI-HF, Gruppo Italiano per lo Studio della Sopravvivenza nell’Insufficienza Cardiaca; HR, hazard ratio; OMEMI, Omega-3 Fatty Acids in Elderly With Myocardial Infarction; REDUCE-IT, Reduction of Cardiovascular Events With Icosapent Ethyl-Intervention Trial; RP, Risk and Prevention Study; STRENGTH, Long-Term Outcomes Study to Assess Statin Residual Risk Reduction With EpaNova in High Cardiovascular Risk Patients With Hypertriglyceridemia; and VITAL, Vitamin D and Omega-3 Trial.

In sensitivity analysis using data from the post hoc publication of the ASCEND trial, the risk of AF persisted in the pooled association between marine ɷ-3 fatty acids and AF events (HR, 1.20 [95% CI, 1.01–1.43]; P=0.039; Figure S2). The estimates remained statistically significant with Knapp-Hartung adjustment among trials testing a high dose of ɷ-3 fatty acids (>1 g/d) with an HR of 1.49 (95%CI 1.04–2.15, P=0.042), but not in trials testing low-dose (≤1 g/d) with an HR of 1.08 (95% CI, 0.99–1.17; P=0.077; Figure S3). Because the heterogeneity was absent in the subgroup of trials with low-dose marine ɷ-3 fatty acids, the application of the fixed effects model yielded a marginally significant increased risk of AF (HR, 1.075 [95% CI, 1.003–1.153]; P=0.042). In the meta-regression model for the association of ɷ-3 fatty acid dosage with the risk of AF events, the HR was 1.10 (95% CI, 1.05–1.14; P=0.002; Figure S4) per 1 g increase of ɷ-3 fatty acids dosage (residual heterogeneity, I2=0.00%; P=0.46).

No significant interaction (P=0.36; Figure S5) was observed between studies that prespecified AF outcomes (HR, 1.42 [95% CI, 0.70–2.90]) versus studies that did not prespecify AF outcomes (HR, 1.19 [95% CI, 1.02–1.38]). Sensitivity analysis with studies that excluded baseline AF or qualified AF events as new-onset yielded similar results (HR, 1.26 [95% CI, 0.85–1.87]) compared with studies that did not exclude baseline AF (HR, 1.28 [95% CI, 1.11–1.48]) with no significant interaction (Figure S6; P=0.92). The effect of marine ɷ-3 fatty acid supplements on the risk of AF events after excluding REDUCE-IT (the only trial without DHA) remained significant (HR, 1.23 [95% CI, 1.02–1.49]; P=0.038; Figure S7). Additional sensitivity analysis excluding the RP trial, which only reported AF hospitalizations, yielded similar results (HR, 1.26 [95% CI, 1.04–1.52]; P=0.028) with no significant interaction (P=0.85). A funnel plot suggested the presence of a slight publication bias possibly related to the dosage of ɷ-3 (Figure S8), because studies with higher dosage were more likely to be on the right side (greater treatment effect). The “trim and fill” method imputed 1 study on the left side of the funnel plot (Figure S9), and the effect size after imputation (HR, 1.23 [95% CI, 1.10–1.39]) was similar to the observed effect size (HR, 1.25 [95% CI, 1.11–1.40]) using the DerSimonian-Laird approach.

Discussion

This meta-analysis adds new evidence regarding the risk of AF in patients taking marine ɷ-3 fatty acid supplements. The risk of AF was significantly more pronounced in trials testing high doses of marine ɷ-3 fatty supplements (>1 g daily) compared with placebo versus low doses of marine ɷ-3 fatty supplements (≤1 g daily) versus placebo. The association appeared to have a dose relationship with a 10% to 11% higher relative risk of AF events per 1 g increased in daily supplementation. The stratification of the trials by the dosage significantly attenuated the heterogeneity across trials; however, no trial has directly compared different dosages (eg, high versus low versus placebo) of marine ɷ-3 supplements on the risk of AF events.

This meta-analysis of large cardiovascular trials with different doses and formulations of ɷ-3 fatty acid supplementations (icosapent ethyl, carboxylic acid formulations, regular eicosapentaenoic acid/DHA combinations) with >2900 AF events provided adequate power to assess small-to-moderate risk of AF and an examination of dose effect relationship at the study level. Before the publication of the REDUCE-IT trial, no safety concerns were reported with ɷ-3 fatty acid supplements. The results suggesting an increased risk for AF hospitalizations and adverse events observed in REDUCE-IT were then replicated with other AF outcomes in 2 other studies. In the STRENGTH trial, 4 g of marine ɷ-3 fatty acids was associated with an increased risk of AF events (2.2% versus 1.3%; HR, 1.69 [95% CI, 1.29–2.21]; P<0.001).6 In the OMEMI trial, older patients with recent myocardial infarction (≥75 years) who were randomized to 1.8-g supplements also had a higher risk of AF events detected clinically or by electrocardiographic monitoring when compared with placebo, although the result did not reach statistical significance (7.2% versus 4.0%; HR, 1.94 [95% CI, 0.98–3.45]; P=0.06). In the VITAL trial, the largest AF primary prevention randomized trial that tested 1 g of ɷ-3 fatty acids, the findings were neutral (HR, 1.09 [95% CI, 0.96–1.24]; P=0.19) and did not support the use of supplemental for the primary prevention of AF. Although the formulations across trials were different (icosapent ethyl, carboxylic acid formulations, regular eicosapentaenoic acid/DHA combinations), these results were consistent when the REDUCE-IT trial, which tested only eicosapentaenoic acid, was excluded. However, these data do not exclude the possibility that specific ɷ-3 formulations used across trials may have differing effects on AF risk.

The 2019 European Society of Cardiology guidelines for dyslipidemia have now integrated icosapent ethyl as second-line treatment in addition to statin for high-risk patients with high triglyceride values.5 According to the 2012 National Health Interview Survey, fish oil supplements are the natural product most commonly taken by adults. About 7.8% of patients reported using marine ɷ-3 fatty supplements, corresponding to approximately 18.8 million people in the United States (https://www.nccih.nih.gov/health/omega3-supplements-in-depth). These results suggest that both the benefits and risks of marine ɷ-3 supplementation should be discussed with the patients, especially when prescribing a higher dosage. The risk-benefit ratio may not only vary according to dose or formulation but also differ according to the patient characteristics. In the OMEMI trial, the absolute risk difference for developing AF (>3%) was the highest among those 75 years or older. However, the investigators of the OMEMI trial also used monitoring to capture AF events; thus, it is likely that more asymptomatic, subclinical cases were detected.7 Because this meta-analysis pooled aggregate-level trial data, and not individual participant data, this report is unable to undertake subgroup analysis by age or other patient-level characteristics. To better understand the risk-benefit ratio in some high-risk subgroups (eg, elderly, patients with cardiac morbidity), future trials testing ɷ-3 fatty acid supplements will need to include systematic, prespecified ascertainment and adjudication of AF outcomes, which can be facilitated with ECG monitoring via new devices (eg, smart phone) and adjudication by independent clinicians.

This comprehensive meta-analysis encompasses recently published large-scale trials of ɷ-3 fatty supplementation; however, there are limitations to be considered. First, differences exist in the AF outcome assessment between trials. In ASCEND (main article), REDUCE-IT, and RP trials, participants with preexisting AF were not systematically excluded from the analysis, whereas in other trials, the analysis reported new-onset AF events or excluded participants with preexisting AF. Sensitivity analysis with studies that excluded baseline AF yielded similar results compared with studies that did not exclude baseline AF with no significant interaction. Second, the HR with 95% CI were provided in each trial, except REDUCE-IT, where the RR was calculated on the basis of the number of patients with AF in each of the treatment arms. The ASCEND trial reported only rate ratio. Because the incidence rate of AF was uncommon, the use of RR is an appropriate proxy of HR, and this limitation should not affect the results. Third, not every large cardiovascular clinical trial testing ɷ-3 fatty acid supplementation reported AF outcomes (eg, JELIS [Japan EPA Lipid Intervention Study] and ORIGIN [Outcome Reduction With an Initial Glargine Intervention]). This limitation should not bias the current conclusions, because trials that hypothesized the superiority of ɷ-3 fatty acid supplements on atherosclerotic cardiovascular outcomes are unlikely to have intentionally underreported neutral safety outcomes for AF events. Of note, the quality of the trials included in this meta-analysis was high with a low risk of bias. Fourth, the lack of variability in dosage among trials, particularly at the lower dose range, limited the ability to definitive test the linearity of the relationship with meta-regression. Fifth, participants who are included in clinical trials might not be representative of those seen in everyday practice.

In conclusion, in this meta-analysis incorporating data from 7 large-scale RCTs, ɷ-3 fatty acid supplementation was associated with a risk of AF, especially in trials testing a higher dose of ɷ-3 fatty acids. Because the benefit of ɷ-3 fatty acids also appears to be dose-dependent, the associated risk of AF should be balanced against the benefit on atherosclerotic cardiovascular outcomes.

Article Information

Supplemental Material

Supplemental Methods

Table S1

Figures S1–S9

Nonstandard Abbreviations and Acronyms

AF

atrial fibrillation

ASCEND

A Study of Cardiovascular Events in Diabetes

DHA

docosahexaenoic acid

EAS

European Atherosclerosis Society

EPA

eicosapentaenoic acid

ESC

European Society of Cardiology

GISSI-HF

Gruppo Italiano per lo Studio della Sopravvivenza nell’Insufficienza Cardiaca

HR

hazard ratio

OMEMI

Omega-3 Fatty Acids in Elderly With Myocardial Infarction

RCT

randomized controlled trial

REDUCE-IT

Reduction of Cardiovascular Events with Icosapent Ethyl-Intervention Trial

RP

Risk and Prevention Study

RR

risk ratio

STRENGTH

Long-Term Outcomes Study to Assess Statin Residual Risk Reduction With EpaNova in High Cardiovascular Risk Patients With Hypertriglyceridemia

VITAL

Vitamin D and Omega-3 Trial

Disclosures Dr Djousse reported receiving grants from the National Institutes of Health and received an investigator-initiated award from Novartis. Dr Albert reported receipt of grants from St Jude Medical, Abbott, and Roche Diagnostics. Dr Manson reported receiving grants from Mars Symbioscience. The other authors report no conflicts.

Footnotes

https://www.ahajournals.org/journal/circ

Supplemental Material, the podcast, and transcript are available with this article at https://www.ahajournals.org/doi/suppl/10.1161/CIRCULATIONAHA.121.055654.

Continuing medical education (CME) credit is available for this article. Go to http://cme.ahajournals.org to take the quiz.

For Sources of Funding and Disclosures, see page 1989.

Correspondence to: Christine M. Albert, MD, MPH, Department of Cardiology, Smidt Heart Institute, Cedars-Sinai Medical Center, 127 South San Vincente Blvd, AHSP 3100, Los Angeles, CA 90048. Email

References

  • 1. Hu Y, Hu FB, Manson JE. Marine omega-3 supplementation and cardiovascular disease: an updated meta-analysis of 13 randomized controlled trials involving 127 477 participants.J Am Heart Assoc. 2019; 8:e013543. doi: 10.1161/JAHA.119.013543LinkGoogle Scholar
  • 2. Bhatt DL, Steg PG, Miller M, Brinton EA, Jacobson TA, Ketchum SB, Doyle RT, Juliano RA, Jiao L, Granowitz C, et al; REDUCE-IT Investigators. Cardiovascular risk reduction with icosapent ethyl for hypertriglyceridemia.N Engl J Med. 2019; 380:11–22. doi: 10.1056/NEJMoa1812792CrossrefMedlineGoogle Scholar
  • 3. Dietary supplementation with n-3 polyunsaturated fatty acids and vitamin E after myocardial infarction: results of the GISSI-Prevenzione trial. Gruppo Italiano per lo Studio della Sopravvivenza nell’Infarto miocardico.Lancet. 1999; 354:447–55. doi: 10.1016/S0140-6736(99)07072-5CrossrefMedlineGoogle Scholar
  • 4. Yokoyama M, Origasa H, Matsuzaki M, Matsuzawa Y, Saito Y, Ishikawa Y, Oikawa S, Sasaki J, Hishida H, Itakura H, et al; Japan EPA Lipid Intervention Study (JELIS) Investigators. Effects of eicosapentaenoic acid on major coronary events in hypercholesterolaemic patients (JELIS): a randomised open-label, blinded endpoint analysis.Lancet. 2007; 369:1090–1098. doi: 10.1016/S0140-6736(07)60527-3CrossrefMedlineGoogle Scholar
  • 5. Mach F, Baigent C, Catapano AL, Koskinas KC, Casula M, Badimon L, Chapman MJ, De Backer GG, Delgado V, Ference BA, et al; ESC Scientific Document Group. 2019 ESC/EAS Guidelines for the management of dyslipidaemias: lipid modification to reduce cardiovascular risk.Eur Heart J. 2020; 41:111–188. doi: 10.1093/eurheartj/ehz455CrossrefMedlineGoogle Scholar
  • 6. Nicholls SJ, Lincoff AM, Garcia M, Bash D, Ballantyne CM, Barter PJ, Davidson MH, Kastelein JJP, Koenig W, McGuire DK, et al. Effect of high-dose omega-3 fatty acids vs corn oil on major adverse cardiovascular events in patients at high cardiovascular risk: the STRENGTH randomized clinical trial.JAMA. 2020; 324:2268–2280. doi: 10.1001/jama.2020.22258CrossrefMedlineGoogle Scholar
  • 7. Kalstad AA, Myhre PL, Laake K, Tveit SH, Schmidt EB, Smith P, Nilsen DWT, Tveit A, Fagerland MW, Solheim S, et al; OMEMI Investigators. Effects of n-3 fatty acid supplements in elderly patients after myocardial infarction: a randomized, controlled trial.Circulation. 2021; 143:528–539. doi: 10.1161/CIRCULATIONAHA.120.052209LinkGoogle Scholar
  • 8. Albert CM, Cook NR, Pester J, Moorthy MV, Ridge C, Danik JS, Gencer B, Siddiqi HK, Ng C, Gibson H, et al. Effect of marine omega-3 fatty acid and vitamin D supplementation on incident atrial fibrillation: a randomized clinical trial.JAMA. 2021; 325:1061–1073. doi: 10.1001/jama.2021.1489CrossrefMedlineGoogle Scholar
  • 9. Moher D, Liberati A, Tetzlaff J, Altman DG; PRISMA Group. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement.Ann Intern Med. 2009; 151:264–9, W64. doi: 10.7326/0003-4819-151-4-200908180-00135CrossrefMedlineGoogle Scholar
  • 10. Aung T, Halsey J, Kromhout D, Gerstein HC, Marchioli R, Tavazzi L, Geleijnse JM, Rauch B, Ness A, Galan P, et al; Omega-3 Treatment Trialists’ Collaboration. Associations of omega-3 fatty acid supplement use with cardiovascular disease risks: meta-analysis of 10 trials involving 77 917 individuals.JAMA Cardiol. 2018; 3:225–234. doi: 10.1001/jamacardio.2017.5205CrossrefMedlineGoogle Scholar
  • 11. Mariani J, Doval HC, Nul D, Varini S, Grancelli H, Ferrante D, Tognoni G, Macchia A. N-3 polyunsaturated fatty acids to prevent atrial fibrillation: updated systematic review and meta-analysis of randomized controlled trials.J Am Heart Assoc. 2013; 2:e005033. doi: 10.1161/JAHA.112.005033LinkGoogle Scholar
  • 12. Mozaffarian D, Wu JH, de Oliveira Otto MC, Sandesara CM, Metcalf RG, Latini R, Libby P, Lombardi F, O’Gara PT, Page RL, et al. Fish oil and post-operative atrial fibrillation: a meta-analysis of randomized controlled trials.J Am Coll Cardiol. 2013; 61:2194–2196. doi: 10.1016/j.jacc.2013.02.045CrossrefMedlineGoogle Scholar
  • 13. Grundy SM, Stone NJ, Bailey AL, Beam C, Birtcher KK, Blumenthal RS, Braun LT, de Ferranti S, Faiella-Tommasino J, Forman DE, et al. 2018 AHA/ACC/AACVPR/AAPA/ABC/ACPM/ADA/AGS/APhA/ASPC/NLA/PCNA guideline on the management of blood cholesterol: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines.Circulation. 2019; 139:e1082–e1143. doi: 10.1161/CIR.0000000000000625LinkGoogle Scholar
  • 14. Gencer B, Marston NA, Im K, Cannon CP, Sever P, Keech A, Braunwald E, Giugliano RP, Sabatine MS. Efficacy and safety of lowering LDL cholesterol in older patients: a systematic review and meta-analysis of randomised controlled trials.Lancet. 2020; 396:1637–1643. doi: 10.1016/S0140-6736(20)32332-1CrossrefMedlineGoogle Scholar
  • 15. Hartung J, Knapp G. A refined method for the meta-analysis of controlled clinical trials with binary outcome.Stat Med. 2001; 20:3875–3889. doi: 10.1002/sim.1009CrossrefMedlineGoogle Scholar
  • 16. Bender R, Friede T, Koch A, Kuss O, Schlattmann P, Schwarzer G, Skipka G. Methods for evidence synthesis in the case of very few studies.Res Synth Methods. 2018; 9:382–392. doi: 10.1002/jrsm.1297CrossrefMedlineGoogle Scholar
  • 17. Jackson D, Law M, Rücker G, Schwarzer G. The Hartung-Knapp modification for random-effects meta-analysis: a useful refinement but are there any residual concerns?Stat Med. 2017; 36:3923–3934. doi: 10.1002/sim.7411CrossrefMedlineGoogle Scholar
  • 18. Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test.BMJ. 1997; 315:629–634. doi: 10.1136/bmj.315.7109.629CrossrefMedlineGoogle Scholar
  • 19. Duval S, Tweedie R. Trim and fill: a simple funnel-plot-based method of testing and adjusting for publication bias in meta-analysis.Biometrics. 2000; 56:455–463. doi: 10.1111/j.0006-341x.2000.00455.xCrossrefMedlineGoogle Scholar
  • 20. Sterne JA, Sutton AJ, Ioannidis JP, Terrin N, Jones DR, Lau J, Carpenter J, Rücker G, Harbord RM, Schmid CH, et al. Recommendations for examining and interpreting funnel plot asymmetry in meta-analyses of randomised controlled trials.BMJ. 2011; 343:d4002. doi: 10.1136/bmj.d4002CrossrefMedlineGoogle Scholar
  • 21. Marston NA, Giugliano RP, Im K, Silverman MG, O’Donoghue ML, Wiviott SD, Ference BA, Sabatine MS. Association between triglyceride lowering and reduction of cardiovascular risk across multiple lipid-lowering therapeutic classes: a systematic review and meta-regression analysis of randomized controlled trials.Circulation. 2019; 140:1308–1317. doi: 10.1161/CIRCULATIONAHA.119.041998LinkGoogle Scholar
  • 22. Bowman L, Mafham M, Wallendszus K, Stevens W, Buck G, Barton J, Murphy K, Aung T, Haynes R, Cox J, et al; ASCEND Study Collaborative Group. Effects of n-3 fatty acid supplements in diabetes mellitus.N Engl J Med. 2018; 379:1540–1550. doi: 10.1056/NEJMoa1804989CrossrefMedlineGoogle Scholar
  • 23. Parish S, Mafham M, Offer A, Barton J, Wallendszus K, Stevens W, Buck G, Haynes R, Collins R, Bowman L, et al; ASCEND Study Collaborative Group. Effects of omega-3 fatty acid supplements on arrhythmias.Circulation. 2020; 141:331–333. doi: 10.1161/CIRCULATIONAHA.119.044165LinkGoogle Scholar
  • 24. Aleksova A, Masson S, Maggioni AP, Lucci D, Fabbri G, Beretta L, Mos L, Paino AM, Nicolosi GL, Marchioli R, et al; GISSI-Heart Failure Investigators. n-3 polyunsaturated fatty acids and atrial fibrillation in patients with chronic heart failure: the GISSI-HF trial.Eur J Heart Fail. 2013; 15:1289–1295. doi: 10.1093/eurjhf/hft103CrossrefMedlineGoogle Scholar
  • 25. Roncaglioni MC, Tombesi M, Avanzini F, Barlera S, Caimi V, Longoni P, Marzona I, Milani V, Silletta MG, Tognoni G, et al; Risk and Prevention Study Collaborative Group. n-3 fatty acids in patients with multiple cardiovascular risk factors.N Engl J Med. 2013; 368:1800–1808. doi: 10.1056/NEJMoa1205409CrossrefMedlineGoogle Scholar

eLetters(0)

eLetters should relate to an article recently published in the journal and are not a forum for providing unpublished data. Comments are reviewed for appropriate use of tone and language. Comments are not peer-reviewed. Acceptable comments are posted to the journal website only. Comments are not published in an issue and are not indexed in PubMed. Comments should be no longer than 500 words and will only be posted online. References are limited to 10. Authors of the article cited in the comment will be invited to reply, as appropriate.

Comments and feedback on AHA/ASA Scientific Statements and Guidelines should be directed to the AHA/ASA Manuscript Oversight Committee via its Correspondence page.