Background
N-3 polyunsaturated fatty acids (PUFAs) include plant-derived α-linolenic acid (ALA) and marine-derived docosahexaenoic acid (DHA), docosapentaenoic acid (DPA), and eicosapentaenoic acid (EPA). DHA, DPA, and EPA are long-chain n-3 fatty acids. Evidence from observational, experimental studies, and randomized controlled trials (RCT) show that n-3 PUFAs from diet or supplements confer protection against cardiovascular disease (CVD) and relevant risk factors, including cardiac death, ischemic heart disease (IHD), ischemic stroke, heart failure, and blood pressure [
1,
2]. However, integrated analyses of these studies have found null, little, or inconsistent results, no matter whether in primary or in secondary prevention of CVD [
3‐
5]. Recent evidence from high-quality large RCTs also suggests that n-3 PUFA intake probably makes little or no difference for coronary heart mortality or events [
6‐
9]. The nutrition recommendations for n-3 PUFA supplements or seafood for cardiovascular benefits have been debated in recent years, but without consensus being reached [
10]. These controversies may be confounded by background dietary consumption of fish, health status, medical treatment of IHD, socioeconomic position, lifestyle, different study populations, and different definitions of CVD and study endpoints [
1,
10].
Mendelian randomization (MR) studies use germline genetic variants as intermediate instrumental variables (IVs) to assess causal relationships in a non-experimental setting. As genetic variants are determined at conception, MR studies are less susceptible to confounders than observational studies and are not affected by disease status, thereby avoiding reverse causation bias. MR studies can be regarded as “natural” RCTs and have been applied to examine genetic predisposition conferred by several genes on IHD [
11]. In the present study, we conducted a two-sample MR study to assess the effect of genetically predicted n-3 PUFAs on IHD, using genetically instrumented n-3 PUFAs from previous studies and a very large case–control dataset of IHD from public consortia. In addition, cardiometabolic risk factors (CRFs) of IHD, including type 2 diabetes (T2D), hyperlipidemia, hypertension, and abdominal obesity were similarly assessed.
Discussion
This MR study showed that a genetic predisposition toward higher plasma ALA level is associated with a lower risk of IHD, but not MI. The effect size (beta coefficient) per 0.05-unit increase (about 1 SD) in plasma ALA level was − 1.173 (95% confidence interval − 2.214 to − 0.133) for IHD. In contrast, genetically-predicted levels of marine-derived n-3 PUFAs (DHA, DPA, and EPA) had no association with IHD or MI.
On primary prevention of CVD, Abdelhamid et al
. [
3] found increased ALA may slightly reduce the risk of cardiovascular events, coronary heart disease (CHD) mortality, and arrhythmia, and Pan et al
. [
29] found dietary ALA is associated with a moderately lower risk of fatal CHD, with each 1 g/d increment of ALA intake being associated with a 10% lower risk of CHD death. By assessing the primary incidence of CHD in generally healthy, free-living populations around the world, Del Gobbo et al
. [
30] also found ALA to be associated with a 9% lower risk of fatal CHD. From the data extracted from the UK Biobank SOFT CAD GWAS and the CARDIoGRAMplusC4D 1000 Genomes-based GWAS consortia, our results also demonstrate beneficial primary health outcomes for ALA. However, for secondary prevention of CVD, there is little or no effect of ALA, as previously suggested [
31], or the evidence is scarce.
Our results, combined with previous findings, support the favorable effects of ALA specifically for the primary prevention of IHD. Mechanistically, these findings are supported by the effects of ALA on improving lipid profile (TC, TG, LDL) [
32] and cholesterol homeostasis [
33], ameliorating sympathetic heart activity and denervation [
34,
35], decreasing fasting free fatty acid and inhibiting inflammation and platelet activation [
36]. A meta-analysis of 18 observational studies in generally healthy populations found that ALA may be associated with modestly lower risk T2D [
37]. Our results also show that genetically-predicted higher plasma ALA is associated with a lower risk of T2D and lower LDL, HDL, TG, and TC. It is well known that LDL is the initiator of IHD [
38], hypertriglyceridemia is the residual risk of IHD [
39,
40], while diabetes could negatively affect clinical outcomes of IHD, in patients admitted for ST-elevation myocardial infarction (STEMI) [
41‐
43], non-STEMI [
44] or stable IHD [
45,
46]. From this perspective, ALA can reduce the risk of IHD in many ways. N-3 PUFAs do not affect atherosclerotic progression, plaque stability, plaque rupture, or thrombosis [
10]. This may be related to the ineffectiveness of ALA on MI. Clinically, the phenotypes of MI are not equal to the presence of coronary atherosclerosis. Coronary atherosclerosis may progress as acute coronary thrombotic occlusion or MI, often due to the rupture of an unstable plaque [
47‐
49], usually occurring in plaques with a thin, eroded fibrous cap, regardless of the degree of stenosis [
47,
50]. Many patients live to advanced age with stable, significant IHD and never suffer an MI.
The clinical research on marine-derived n-3 PUFAs (DHA, DPA, and EPA) has been full of twists and turns. Before the use of statins, most of the studies on the cardioprotection of marine-derived n-3 PUFAs were positive [
51]. However, after statins became widely used, most studies reported neutral effects [
52,
53]. In recent years, the cardioprotective role of marine-derived n-3 PUFAs, especially EPA and DHA, has become increasingly disputed [
54,
55]. Dietary recommendations of EPA and DHA have also been downgraded from Class I to Class II [
53], is the reason being that a large number of more recent RCTs [
6,
7] and integration analyses have found little or no effect of EPA or DHA on cardioprotection [
4,
5], particularly for primary prevention of CVD [
3,
8,
9]. Even when there is an effect, the effect is only seen in studies with a moderate to high risk of bias [
56]. Some researchers even think fish oil has disappointing therapeutic benefits [
5].
In contrast, other researchers still have hope for marine-derived n-3 PUFAs, especially EPA. They acknowledge that, in over-the-counter formulations (EPA + DHA or fish oil) at common dosages, primary prevention of CVD by marine-derived n-3 PUFAs, is ineffective and that secondary prevention is controversial [
10]. They attribute the failure of the previous trials to the low dose and impure formulation of marine-derived n-3 PUFAs, short intervention duration, high background of fish intake, and inappropriate participants [
51,
52,
55]. With the large success of the REDUCE-IT trial [
57], proponents put their hopes on highly purified EPA (icosapent ethyl), which will lower plasma TG levels, and have given some constructive suggestions for future clinical trials [
55]. At this time, the AHA also has given more affirmative recommendations to support the use of marine-derived n-3 PUFAs for reducing the residual risk of CVD that remains after statin therapy [
58]. However, as fibrates [
59] and PCSK9 inhibitors [
60] not only reduce TG, similar to EPA, but also increase HDL and reduce LDL, a new debate arises as to whether we should use fibrates instead of EPA or PCSK9 inhibitors instead of a statin/EPA combination [
61]. So, it seems that the debate on the cardioprotection of marine-derived n-3 PUFAs will continue.
Our research explores the role of marine-derived n-3 PUFAs from another perspective and finds individual marine-derived n-3 PUFAs have no association with IHD or MI in generally healthy populations. As for CRFs, DPA is associated with a higher risk of T2D and higher HDL, LDL, TC, and WHR; EPA is associated with higher WHR, and DHA does not affect CRFs. Most prior studies showed that, except for reducing TG [
57,
61], marine-derived n-3 PUFAs do not affect most CRFs or intermediate outcomes [
3,
9], including high CAD risk factors,
i.e. LDL and T2D [
9,
37]. Only a few studies have shown that marine-derived n-3 PUFAs significantly reduce blood pressure [
62], with the greatest reductions in untreated hypertension. Some studies even suggest that marine-derived n-3 PUFAs may increase LDL [
52,
58], which may negate any cardiovascular benefits [
61]. Recently, a large general-practice RCT show that for patients with multiple cardiovascular risk factors (the criterion was defined as at least four of the following, or for patients with diabetes, at least one of the following: age of 65 years or older, male sex, hypertension, hypercholesterolemia, current smoker, obesity, family history of premature cardiovascular disease), treatment with n-3 PUFAs (1 g DHA + EPA daily, with a median of 5 years of follow-up) did not reduce cardiovascular mortality and morbidity [
63].
There are some limitations to our study. First, in our study, both IVs and outcomes come from Europe. This avoids population stratification and conforms to the homogenous principle of an MR study [
11]. However, as few n-3 PUFA GWAS are available for African Americans, Chinese, or other races, the potential effects of n-3 PUFAs by race remains to be explored. Similarly, as there are ethnic differences in the risk of cardiovascular disease [
64], we need to be cautious in applying our findings to other populations. Second, n-3 PUFAs can be detected in many components of the body (e.g., serum, plasma, phospholipids, cholesterol esters, and adipose tissue) and affect many CVD subtypes (e.g., sudden cardiac death, congestive heart failure, arrhythmia, acute coronary syndrome, and stroke), but our study did not analyze this one by one. A comprehensive analysis of n-3 PUFAs in different components and their association with different disease subtypes may help to reduce potential bias and provide a better understanding of the effect of n-3 PUFAs on cardiovascular health. Third, most instrumental SNPs explain a small proportion variance of n-3 PUFAs. This may reduce the power to detect small effects of n-3 PUFAs on IHD risk in our MR framework. Fourth, the effect of n-3 PUFAs on IHD and CRFs found in this study represent a lifelong cumulative effect and are not directly comparable to those derived from conventional observational or clinical studies. Finally, we could not assess whether the effect of n-3 PUFAs on IHD and CRFs varied by sex, age, or the baseline level of n-3 PUFAs as these data are not freely available.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.