Causal role of EPA on ischemic heart disease, triglyceride rich lipoproteins and related traits: A two-sample Mendelian randomization analysis

This two-sample Mendelian randomization study suggests that while eicosapentaenoic acid (EPA) remodels triglyceride-rich lipoproteins by shifting particle sizes, it does not demonstrate a protective causal effect against ischemic heart disease in the general European population and may even be associated with a slight increase in risk.

The Gist

The Big Question: Is the "Fish Oil" Miracle Real?

For a long time, people have believed that eating fish or taking fish oil supplements (specifically a type of fat called EPA) is like a magic shield against heart attacks. Some studies on people with very high heart disease risk suggested this was true. But other studies on the general public showed little to no benefit.

This paper asks a simple question: Does EPA actually protect the average person's heart, or is it just a myth?

To find the answer, the researchers didn't just ask people what they ate (which is unreliable). Instead, they used a scientific "time machine" called Mendelian Randomization.

The Time Machine Analogy: The Genetic Lottery

Imagine that at the moment of conception, every person is dealt a hand of genetic cards. Some cards make your body naturally produce more EPA (like having a factory that makes more of a specific chemical), while others make you produce less.

Because these cards are dealt randomly (like a lottery), they act like a natural experiment. If people with "high EPA cards" get fewer heart attacks than people with "low EPA cards," we can be sure it's the EPA causing the protection, not their diet or lifestyle.

The researchers looked at the genetic "cards" of over 14,000 people to see how their natural EPA levels affected their heart health and blood fats.

The Findings: A Mixed Bag of Results

1. The Heart Attack Result: No Clear Shield

The study found that having higher natural levels of EPA did not lower the risk of ischemic heart disease (heart attacks caused by blocked arteries) in the general population.

• The Analogy: Imagine you are trying to stop a leak in a boat by plugging one hole. You might think, "If I plug this hole, the boat won't sink." But the study suggests that for the average person, plugging this specific hole (raising EPA) doesn't actually stop the boat from sinking. In fact, the data hinted slightly that it might even make things a tiny bit worse, though the evidence wasn't strong enough to be certain.

2. The Blood Fat Result: The "Traffic Jam" Remodeling

This is where the story gets interesting. While EPA didn't stop heart attacks, it did drastically change the traffic pattern of fats in the blood.

• The Big Trucks vs. The Small Cars: Think of your blood fats as vehicles on a highway.
  • Large VLDL particles are like big, slow-moving trucks carrying heavy loads of triglycerides (fats).
  • Small VLDL particles are like tiny, fast cars.
• What EPA Did: The study found that EPA acts like a demolition crew that breaks down the big trucks into smaller cars.
  • It successfully reduced the number of big trucks (lowering triglycerides).
  • However, it didn't clear the smaller cars off the road. Instead, the traffic shifted from big trucks to a swarm of small cars.
  • It also increased the number of "remnant" particles (the leftover pieces of the trucks) and slightly increased total cholesterol.

3. The "Cleanup Crew" Problem

The researchers suspect that EPA tells the body to break down the big fat trucks (lipolysis), but it doesn't tell the body to remove the resulting small pieces from the bloodstream efficiently.

• The Analogy: It's like a construction crew that demolishes a large building (the big fat truck) but leaves all the rubble (the small fat particles) sitting on the street. The street looks different, but it's still cluttered with debris.

The Conclusion: Why the Confusion?

The paper suggests that EPA might be a "double-edged sword" for the average person:

1. Good: It clears out the big, heavy fat trucks.
2. Bad: It leaves behind a swarm of smaller, potentially dangerous fat particles that stay in the blood longer.

Because these two effects cancel each other out, the net result for the general population is no protection against heart disease.

What This Means (According to the Paper)

The authors are careful to say this study does not prove EPA is useless for everyone. They note that:

• Previous trials showed EPA did help people who were already at very high risk for heart disease.
• This study only looked at the general population.
• The "demolition without cleanup" mechanism might explain why EPA works for some groups but not others.

The Bottom Line:
Based on this genetic study, if you are a generally healthy person, simply raising your EPA levels might not prevent a heart attack. It changes the type of fat in your blood (breaking big trucks into small cars), but it doesn't necessarily clear the road enough to save your heart. The authors suggest we need a new, large-scale trial specifically testing EPA on the general public to see if this "traffic jam" theory holds up in real life.

Technical Summary

1. Problem Statement

While randomized controlled trials (RCTs) generally show that combined omega-3 fatty acids (EPA and DHA) have little effect on cardiovascular disease (CVD) risk, specific trials (e.g., JELIS, REDUCE-IT) suggest that Eicosapentaenoic Acid (EPA) monotherapy may reduce CVD risk, particularly in high-risk populations with high triglycerides. However, it remains unclear whether EPA is beneficial for the general population and how it mechanistically affects triglyceride-rich lipoproteins (TRL) and related lipid traits. Previous Mendelian Randomization (MR) studies on omega-3s have yielded mixed results, often due to limited genetic instruments or violations of MR assumptions (specifically horizontal pleiotropy). This study aims to resolve these uncertainties by evaluating the causal effect of genetically predicted plasma EPA on Ischemic Heart Disease (IHD) and a comprehensive panel of lipoprotein subfractions in a general European population.

2. Methodology

The study employed a Two-Sample Mendelian Randomization (MR) design using summary statistics from Genome-Wide Association Studies (GWAS).

• Data Sources:
  • Exposure (Plasma EPA): Meta-analysis of GWAS from EPIC-Norfolk and INTERVAL cohorts (N=14,267, European ancestry).
  • Selection Datasets (Instrument Identification): Meta-analysis of CHARGE consortium and Canadian Longitudinal Study of Aging (CLSA) to identify candidate variants.
  • Outcomes:
    • IHD: Meta-analysis from UK Biobank, CARDIoGRAMplusC4D, and FinnGen (Cases/N ≈ 213k/1.38M).
    • TRL & Related Traits: GWAS from 26 cohorts (N=68,559) using NMR metabolomics to measure lipoprotein subclasses, apolipoproteins, and cholesterol.
• Instrument Selection:
  • Variants associated with plasma EPA at $P < 5 \times 10^{-5}$ in both selection and exposure datasets were considered.
  • Exclusion Criteria: Variants were removed if they were palindromic, had low MAF (<0.01), or were located on/linked to FADS and ELOVL gene clusters (to avoid horizontal pleiotropy via omega-6 synthesis). Variants associated with confounders (smoking, alcohol, education) or DHA were also excluded.
  • Final Instruments: 8 independent SNPs (F-statistics range: 26.16–177.64) were retained.
• Statistical Analysis:
  • Primary Estimator: Inverse-Variance Weighted (IVW) with multiplicative random effects.
  • Sensitivity Analyses: MR-Egger (to test for directional pleiotropy), Weighted Median (WM), and leave-one-out analyses.
  • Controls: Positive controls (TG, VLDL) to verify instrument validity; Negative control (Age at recruitment) to assess selection bias.
  • Correction: Bonferroni correction applied for multiple testing ($P < 0.002$).

3. Key Contributions

• Expanded Lipid Profiling: Unlike previous MR studies focusing only on total lipids, this study utilized high-resolution NMR data to assess specific lipoprotein subfractions (e.g., large vs. small VLDL, IDL, LDL, HDL) and apolipoproteins (ApoB, ApoA-I).
• Refined Instrument Selection: The study rigorously excluded variants in the FADS and ELOVL regions, which are known to be pleiotropic (affecting both omega-3 and omega-6 pathways), thereby strengthening the validity of the causal inference.
• General Population Focus: It specifically addresses the gap in knowledge regarding EPA's effects in the general population, contrasting with high-risk cohorts used in previous RCTs.

4. Results

A. Ischemic Heart Disease (IHD)

• Main Finding: Genetically predicted higher plasma EPA was associated with a non-significant increase in the odds of IHD.
  • IVW Estimate: OR = 1.05 (95% CI: 1.00, 1.10).
  • MR-Egger and WM estimates were similar or closer to the null, with no evidence of directional horizontal pleiotropy (MR-Egger intercept $P=0.355$).
• Conclusion: There is no evidence of a protective effect of EPA on IHD in the general European population; if anything, a weak harmful trend was observed, though not statistically significant after strict correction.

B. Triglyceride-Rich Lipoproteins (TRL) and Lipid Traits

EPA induced a distinct remodeling of the lipoprotein profile:

• Triglycerides (TG): Significant reduction ($B = -0.16$).
• VLDL:
  • Decrease in large VLDL particles.
  • Increase in very small VLDL particles.
  • Overall reduction in mean VLDL diameter.
• IDL and LDL:
  • Significant increases in Intermediate Density Lipoproteins (IDL) and both large and small LDL particles.
  • Increase in mean LDL diameter.
• Apolipoproteins:
  • No change in ApoB (indicating the total number of TRL particles remained stable).
  • Increase in ApoA-I (consistent with larger HDL).
• HDL: Increase in very large and medium HDL particles; no change in small HDL.
• Cholesterol: Increases in both Remnant Cholesterol and Total Serum Cholesterol.

5. Significance and Discussion

• Mechanistic Insight: The results suggest EPA promotes lipolysis (breaking down large VLDL) but fails to promote the subsequent clearance of the resulting smaller particles. This leads to a shift from large TRL to smaller, potentially more atherogenic remnant particles (IDL and small LDL) that remain in circulation longer.
• Clinical Implications:
  • The lack of IHD protection in the general population may be due to this "remodeling without clearance" mechanism, where the reduction in TG is offset by an increase in atherogenic remnants.
  • This contrasts with EPA monotherapy trials in high-risk populations (e.g., REDUCE-IT), suggesting effect heterogeneity based on baseline CVD risk or ethnicity (e.g., Japanese vs. Western populations).
  • The findings challenge the assumption that lowering TG via EPA automatically translates to reduced CVD risk in the general population.
• Future Directions: The authors call for a cardiovascular outcome trial of EPA monotherapy in a general population that specifically collects lipoprotein subfraction data to confirm whether the observed remodeling is detrimental or if the high-risk population benefits from a different mechanism.

Conclusion: This study provides causal evidence that while EPA effectively lowers triglycerides and large VLDL, it remodels the lipoprotein profile toward smaller, potentially more atherogenic particles without reducing total particle number (ApoB). Consequently, EPA does not appear to protect against IHD in the general European population, highlighting the complexity of lipid metabolism and the limitations of using TG reduction alone as a surrogate for CVD risk reduction.