The Role of Omega-3 Polyunsaturated Fatty Acids in Heart Failure: A Meta-Analysis of Randomised Controlled Trials

Many new clinical trials about the effect of omega-3 polyunsaturated fatty acids (PUFAs) in heart failure (HF) patients have shown inconsistent results. Therefore, a meta-analysis of randomised controlled trials (RCTs) was performed to determine the benefits of omega-3 PUFAs in HF patients. Articles were obtained from PubMed, EMBASE, and the Cochrane Library. RCTs comparing omega-3 PUFAs with placebo for HF were included. Two reviewers independently extracted the data from the selected publications. The I2 statistic was used to assess heterogeneity. The pooled mean difference and associated 95% confidence intervals were calculated, and a fixed or random-effects model was used for the meta-analysis. A total of nine RCTs involving 800 patients were eligible for inclusion. Compared with patients taking placebo, HF patients who received omega-3 PUFAs experienced decreased brain natriuretic peptide levels and serum norepinephrine levels. Although the left ventricular ejection fraction (LVEF) and clinical outcomes (Tei index, peak oxygen consumption) did not improve, subgroup analysis showed that the LVEF increased in dilated cardiomyopathy (DCM) patients. Overall, omega-3 PUFA supplements might be beneficial in HF patients, especially in DCM patients, but further studies are needed to confirm these benefits.


Introduction
Heart failure (HF) is a common and growing public health problem worldwide. At present, approximately 26 million people are living with HF in the world. In the United States, there were 5.8 million patients with HF in 2012, and this is expected to increase to 8.5 million by 2030 [1]. Despite advances in pharmacological and non-pharmacological therapies, the treatment of HF remains a challenge [2,3]. Omega-3 polyunsaturated fatty acids (PUFAs), including eicosapentaenoic acid (EPA), docosahexaenoic acid (DHA), and plant-derived alpha-linolenic acid (ALA), are mainly derived from marine sources such as fish oil, as well as from nuts and vegetable oils. Omega-3 PUFAs are widely consumed as nutritional supplements. Health authorities have long recommended these supplements to provide cardiovascular protective effects in patients with coronary heart disease (CHD) [4]. Substantial evidence from experimental and epidemiological studies has suggested that the intake of omega-3 PUFAs could be beneficial for HF patients [5][6][7][8]. The GISSI-HF trial showed that HF patients receiving long-term fish oil treatment showed reduced endpoints of mortality or hospitalizations for cardiovascular reasons [9]. However, many questions still remain. Further clinical studies are warranted to clarify the optimal dose, the differential benefit of EPA and DHA, and the mechanism of action. A previous meta-analysis performed in 2012 [10] indicated that omega-3 PUFAs could improve cardiac function and functional capacity in HF patients. However, since the publication presented as weighted mean difference (WMD) with confidence intervals (CIs) for measurements of the same parameter or standardised mean difference (SMD) for various measurements in the included studies. Heterogeneity was evaluated using the I 2 test (I 2 > 50% indicating significant heterogeneity). An inverse variance fixed-effect model was used to calculate the WMD or SMD with 95% CI if there was no significant heterogeneity among the included studies; otherwise, a random-effect model was chosen. Sensitivity analysis was used to determine the stability of the statistical results by excluding each study sequentially. In addition, publication bias was evaluated with the use of funnel plots. Statistical significance was set at p < 0.05.

Eligible Studies
A total of 2509 (PubMed (1362), EMBASE (1103), and the Cochrane Library (44)) potentially relevant articles were found by searching the databases. The titles and the abstracts of the retrieved articles were examined, and 2475 were excluded. A full-text evaluation was performed for the remaining 34 articles, and 25 articles were excluded for the following reasons: reviews (6), meta-analysis (6), and did not report our endpoints (13). Therefore, nine studies were eligible for our meta-analysis. The process of the literature search and the reasons for exclusion are provided in Figure 1.
Nutrients 2017, 9,18 3 of 16 the same parameter or standardised mean difference (SMD) for various measurements in the included studies. Heterogeneity was evaluated using the I 2 test (I 2 > 50% indicating significant heterogeneity). An inverse variance fixed-effect model was used to calculate the WMD or SMD with 95% CI if there was no significant heterogeneity among the included studies; otherwise, a randomeffect model was chosen. Sensitivity analysis was used to determine the stability of the statistical results by excluding each study sequentially. In addition, publication bias was evaluated with the use of funnel plots. Statistical significance was set at p < 0.05.

Eligible Studies
A total of 2509 (PubMed (1362), EMBASE (1103), and the Cochrane Library (44)) potentially relevant articles were found by searching the databases. The titles and the abstracts of the retrieved articles were examined, and 2475 were excluded. A full-text evaluation was performed for the remaining 34 articles, and 25 articles were excluded for the following reasons: reviews (6), metaanalysis (6), and did not report our endpoints (13). Therefore, nine studies were eligible for our metaanalysis. The process of the literature search and the reasons for exclusion are provided in Figure 1.

Study Characteristics
The remaining studies were nine clinical trials [11,12,14,[17][18][19][20][21][22] and included a total of 800 patients, as described in Tables 1 and 2. All nine studies were RCT designs published from 2006 to 2016. These studies reported that the participants were clinically and stable HF patients who received optimal medical therapy based on modern HF treatment strategies. The medication treatment was maintained during the follow-up interval. The aetiology of HF was dilated cardiomyopathy (DCM) in three studies [17,18,20], ischaemic cardiomyopathy (ICM) in two studies [11,21], and both ICM and

Study Characteristics
The remaining studies were nine clinical trials [11,12,14,[17][18][19][20][21][22] and included a total of 800 patients, as described in Tables 1 and 2. All nine studies were RCT designs published from 2006 to 2016. These studies reported that the participants were clinically and stable HF patients who received optimal medical therapy based on modern HF treatment strategies. The medication treatment was maintained during the follow-up interval. The aetiology of HF was dilated cardiomyopathy (DCM) in three studies [17,18,20], ischaemic cardiomyopathy (ICM) in two studies [11,21], and both ICM and DCM in four studies [12,14,19,22]. ICM was defined as a history of myocardial infarction with significant coronary artery disease (>70% luminal stenosis in at least two major coronary arteries) [12]. The mean baseline LVEF of the participants varied from 24% to 45%, and the New York Heart Association class was from I to IV. The dose of omega-3 PUFAs (calculated as the total dose of ALA, EPA, and DHA) ranged from 1 g/day to 4 g/day, with the ratio of EPA to DHA varying from 0.60 g/day to 1.53 g/day. The duration of the investigation varied from 3 to 12 months. None of these studies reported adverse events related to omega-3 PUFA supplementation. The quality score of the seven studies ranged from 3 to 5 ( Figure 2). All of the studies were randomised and placebo controlled; only one study was performed in a single blind fashion [12].
Nutrients 2017, 9, 18 4 of 16 DCM in four studies [12,14,19,22]. ICM was defined as a history of myocardial infarction with significant coronary artery disease (>70% luminal stenosis in at least two major coronary arteries) [12]. The mean baseline LVEF of the participants varied from 24% to 45%, and the New York Heart Association class was from I to IV. The dose of omega-3 PUFAs (calculated as the total dose of ALA, EPA, and DHA) ranged from 1 g/day to 4 g/day, with the ratio of EPA to DHA varying from 0.60 g/day to 1.53 g/day. The duration of the investigation varied from 3 to 12 months. None of these studies reported adverse events related to omega-3 PUFA supplementation. The quality score of the seven studies ranged from 3 to 5 ( Figure 2). All of the studies were randomised and placebo controlled; only one study was performed in a single blind fashion [12].   Abbreviations: RCT, randomised controlled trial; HF: heart failure; NYHA, New York Heart Association; ICM, ischemic cardiomyopathy; DCM, dilated cardiomyopathy; EPA, eicosapentaenoic acid; DHA, docohexaenoic acid; ALA, plant-derived alpha-linolenic acid; TNF-α, tumor necrosis factor-α; IL-1,6, Interleukin-1,6; BNP, brain natriuretic peptide; NT-proBNP, N-terminal pro B-type natriuretic peptide; CRP, C reactive protein; MI, myocardial infarction; peak VO 2 , peak oxygen consumption.

Publication Bias and Sensitivity Analysis
The symmetry funnel plot of the LVEF indicated that there was no significant publication bias in the analysis (Figure 11). We also performed a sensitivity analysis by excluding individual studies one by one to evaluate the stability of our estimates. Based on the results of this analysis, there was no significant modification of our conclusions.

Publication Bias and Sensitivity Analysis
The symmetry funnel plot of the LVEF indicated that there was no significant publication bias in the analysis (Figure 11). We also performed a sensitivity analysis by excluding individual studies one by one to evaluate the stability of our estimates. Based on the results of this analysis, there was no significant modification of our conclusions.

Publication Bias and Sensitivity Analysis
The symmetry funnel plot of the LVEF indicated that there was no significant publication bias in the analysis (Figure 11). We also performed a sensitivity analysis by excluding individual studies one by one to evaluate the stability of our estimates. Based on the results of this analysis, there was no significant modification of our conclusions.

Discussion
Data from our meta-analysis indicated that additional omega-3 PUFA supplementation led to lower BNP and SNE levels, which would represent evidence of the benefits of omega-3 PUFAs in patients with HF. However, omega-3 PUFAs did not improve the LVEF and other clinical outcomes (Tei index, peak VO 2 ) in patients with HF overall. Although the subgroup analysis showed that the LVEF increased in DCM patients and with long-time treatment, the dose-response analysis did not reach statistical significance.
The result of our meta-analysis showed that omega-3 PUFAs could decrease BNP and NT-proBNP levels. Plasma BNP or NT-proBNP concentration is an indicator of the severity of HF and increases exponentially as the cardiac condition worsens [23][24][25][26]. Linssen et al. [27] demonstrated that the doubling of NT-proBNP level was associated with a 22% increase in all-cause mortality and a 16% increase in cardiovascular events. Although the results were obtained from limited studies, based on our meta-analysis, the lower BNP level might demonstrate that omega-3 PUFAs were beneficial to HF patients. BNP levels were correlated with increased ventricular wall stress, and a beneficial reduction in wall stress can be assumed with lower BNP levels, which is in accordance with an improvement in diastolic function. In addition, adipose tissue exhibits a high density of natriuretic peptide receptors, which has an underrated role in the regulation of the fat metabolism. HF patients who are overweight or obese have a better prognosis than HF patients with normal body weight, this phenomenon is called the "obesity paradox" [28]. It was suspected that a higher PUFAs concentration could cause a better prognosis in patients with abdominal obesity [29].
The present meta-analysis showed that omega-3 PUFAs could not increase the LVEF and improve other clinical outcomes (Tei index, peak VO 2 ), and this finding was not consistent with that of a previous meta-analysis [11]. However, our findings should be interpreted with caution, because the studies by Chrysohoou et al., 2016 [14] and Kohashi et al., 2014 [12] showed negative results and significant heterogeneity among the included studies. We assume that the differences in ages [30], underlying causes of HF, varying severity of HF, and higher daily doses and ratio of EPA-DHA may have caused the heterogeneity. Subgroup analyses suggested that there were increased effects for non-ischaemic HF than for ischaemic HF patients, which were in accordance with the findings of previous studies [10]. It is currently debated whether supplementation of omega-3 PUFAs in ICM patients is associated with a protective effect on major cardiovascular events. The Alpha Omega Trial, which included 4837 patients, showed that low-dose omega-3 PUFAs did not significantly reduce the rate of major cardiovascular events in post-myocardial infarction (MI) patients [31]. The study by Hoogeveen [13] also demonstrated that there were no effects of omega-3 PUFAs supplementation on NT-proBNP levels after MI. The potential mechanisms might be that excess fatty acid intake magnifies post-MI chemokine signalling and induces non-resolving overactive inflammation within cardiosplenic and cardiorenal network with aging [32]. The inflammatory response into the infarcted zone lead to migration of monocytes, stimulation of the sympathetic nervous system and renin-angiotensin-aldosterone system, and the release of natriuretic peptides [33].
However, the GISSI-Prevenzione investigators showed that the reduction of cardiac death due to supplementation with omega-3 PUFAs was most pronounced for patients with systolic left-ventricular dysfunction [9]. In the Acute Studies of Nesiritide in Decompensated Heart Failure (ASCEND-HF) trial, when the LVEF was evaluated as a continuous measure, it exhibited a U-shaped pattern with mortality [34], so modest improvement in the LVEF was barely in direct relationship with mortality. Further investigation is warranted to elucidate the effect of omega-3 PUFAs in the mortality of HF patients. Furthermore, in the analysis of the dose-effect and subgroup effects, we found that both dosages of 2 g/day and 1 g/day did not show any significant difference, but that there were significant differences in long-term treatment (12 months). More intervention studies should explore the effects of higher doses and long-term use of omega-3 PUFA supplementation in HF patients [35].
Our meta-analysis showed that omega-3 PUFAs could significantly decrease SNE levels. Previous studies indicated that the benefits of omega-3 PUFAs on cardiac function and remodelling may be mediated by multiple effects on multiple processes and mechanisms [36,37]. Omega-3 PUFAs could preserve cardiac mitochondrial function by stimulating the expression of proteins involved in cardiac lipid metabolism to reduce myocardial oxygen consumption, allowing improved mechanic efficiency of the ventricle [38], or by inhibiting inflammatory nuclear transcription factors such as TNF-α, IL-1, and IL-6 [31]. Sympathetic nervous system hyperactivity and increased synthesis and release of cardiac BNP are cardinal features of HF [39]. There is a well-recognised, direct correlation between enhanced norepinephrine levels and increased mortality in HF patients [40,41]. Some data have supported the finding of a modulating effect of omega-3 PUFAs on cardiac autonomic control, but the evidence is not conclusive [42]. Decreasing SNE levels might be one of several beneficial effects of omega-3 PUFAs in HF patients, but because of the small number of participants and short-term follow-up, large-scale clinical trials are required to confirm this benefit.
There are several limitations to the present meta-analysis that should be considered. Firstly, the small sample size and limited number of clinical trials might lead to imprecise results. Second, the changes in BNP levels, LVFF, and SNE levels are all surrogate endpoints. With few studies evaluating terminal outcomes, we were unable to perform analyses of the effects of omega-3 PUFAs on cardiovascular events and mortality in patients with HF. Thirdly, different lengths of intervention time, different doses, and different composition of omega-3 PUFAs in each study might cause a confounding bias and affect the treatment effect. Fourthly, significant heterogeneity in LVEF analyses might affect the interpretation of the true intervention effect. There were not enough studies to perform a meta-regression to explore the source of heterogeneity. Finally, in the subgroup analysis, only a few studies tested the effect of omega-3 PUFAs on patients with only ICM or DCM.

Conclusions
In conclusion, treatment with omega-3 PUFAs decreased the BNP and SNE levels and increased the LVEF in DCM patients. These findings suggest that omega-3 PUFAs might be beneficial to HF patients, but the evidence is not definitive. In the future, high-quality RCTs with large numbers of participants are needed to investigate the optimal dosage and components of omega-3 PUFAs to improve the prognostic outcomes of HF patients.

Supplementary Materials:
The following are available online at http://www.mdpi.com/2072-6643/9/1/18/s1. Author Contributions: Jing Hang drafted the manuscript. Chunbin Wang and Bo Xiong analysed the data, contributed reagents/materials/analysis tools, and wrote the paper. All authors contributed to the development of the selection criteria, the risk of bias assessment strategy, data extraction criteria, search strategy, and statistical expertise. All the authors read, provided feedback, and approved the final manuscript.

Conflicts of Interest:
The authors declare no conflict of interest.