Omega-3 supplementation for CVD

Last update: 24 August 2021

Omega-3 fatty acids are a biomarker strongly negatively associated with cardiovascular disease1. However, when systematically reviewed, the RCT evidence shows at best little benefit2. This may or may not be due to failure of the intervention (usually – taking pills) to affect body lipids3. To my knowledge, nobody has done an analysis of trials that did measure effectiveness of the intervention by how it affected n-3 biomarkers.

Methods

I went through the Cochrane systematic review’s references section2 and selected the trials that, according to the reviewers, collected biomarker information (as per the trial summaries, mainly the “compliance” sections).

Where there were multiple markers available, I focused on those better reflecting recent intake of n-3 PUFA (f.ex. plasma over adipose). Where I had no access to the article, I used whatever information was provided in the abstract or in the Cochrane review.

I skipped studies, labelling them irrelevant, that did not report cardiovascular outcomes. I focused on major outcomes (ACM, CVM, CVE, CHDM, CHDE, Stroke, MI, Arrythmia, SCD – as reported in the review) and remarked where the trial only employed risk factors (or other metrics) for outcomes.

Results

46 trials were identified, of which 30 were relevant to cardiovascular disease.

TrialMarkersOutcomes
AFFORD 2013n-3 index
baseline: ~4.6%
at 6mo: ~7.7%
ACM: null
CVE: null
Stroke: null
Arrythmia: null
AlphaOmega 2010
ALA
serum CE
at 40mo:
ALA +68.8%
EPA +41.6%
ACM: null
CVM: null
CVE: null
CHDM: null
CHDE: null
Stroke: null
MI: null
AlphaOmega 2010
EPA+DHA
serum CE
at 40mo:
EPA +61.6%
DHA +30.0%
ACM: null
CVM: null
CVE: null
CHDM: null
CHDE: null
Stroke: null
Arrythmia: null
MI: null
Baldassarre 2006plasma FA
baseline:
EPA 0.65%
DHA 1.55%
at 24mo:
EPA 3.58%
DHA 4.06%
CVE: null
CHDE: null
MI: null
Bates 1989irrelevantirrelevant
Berson 2004irrelevantirrelevant
Brox 2001serum FA (sum n-3)
cod liver oil:
baseline: 1.04%
14mo: 2.05%
seal oil:
baseline: 1.01%
14mo: 2.38%
ACM: null
CVM: null
CVE: null
CHDM: null
CHDE: null
MI: null
SCD: unestimable
Caldwell 2011irrelevantirrelevant
Deslypere 1992RBC lipid EPA
baseline: 0.7-0.78%
12mo:
3g: 2.82%
6g: 4.31%
9g: 5.86%
biomarkers only
DIPP 2015irrelevantirrelevant
DISAF 2003RBC FA
increased n-3
ACM: null
Arrythmia: null
DO IT 2010plasma EPA and DHA
increased
ACM: null
CVE: null
CVM: null
CHDE: null
CHDM: null
Stroke: null
MI: null
SCD: null
DREAM Asbell 2018irrelevantirrelevant
EPE-A 2014irrelevantirrelevant
EPOCH 2014irrelevantirrelevant
FAAT 2005RBC PL EPA+DHA
baseline: 3.4%
12mo: 7.6%
ACM: null
CVM: null
Arrythmia: null
CVE: null
CHDM: null
Arrythmia: null
FLAX-PAD 2013plasma ALA
baseline: 0.024 mg/ml
6mo: 0.067 mg/ml
ACM: null
CVM: null
CVE: null
CHDM: null
CHDE: null
Stroke: null
MI: null
FOSTAR 2016serum PL n-3 (high dose)
baseline: 6.3%
24mo: 14.5%
Arrythmia: null
SCD: null
Franzen 1993serum lipid FA
baseline: 3.6 mg/dl
4mo: 16.2 mg/dl
ACM: no deaths
CVE: no events
otherwise biomarkers only
HARP 1995adipose FA (EPA, DPA, DHA)
intervention: 2.37%
control: 0.76%
ACM: null
CVM: null
CVE: null
CHDM: null
CHDE: null
Stroke: null
MI: null
HERO 2009RBC FA (ALA)
baseline:
control: 0.10%
intervention: 0.12%
12mo:
control: 0.17%
intervention: 0.18%
ACM: no deaths
otherwise metrics and biomarkers only
JELIS 2007plasma FA
baseline:
EPA/AA: 0.6
treatment:
EPA/AA: 1.3
ACM: null
CVE: benefit
CHDM: null
CHDE: null
Stroke: null
MI: null
SCD: null
Kumar 2012serum FA
EPA: 2.1x vs control
DHA: 1.8x vs control
Arrythmia: benefit
Kumar 2013plasma PL
baseline:
EPA: 1.5%
DHA: 4.2%
6mo:
EPA: 4.3%
DHA: 7.7%
ACM: null
CVM: null
CVE: null
MAPT 2017RBC lipids
EPA +3.52pp
DHA +3.29pp
ACM: null
CVE: null
Stroke: null
MARGARIN 2002serum ALA
baseline: 0.52%
intervention 1: +0.36pp
intervention 2: +0.47pp
ACM: null
CVM: null
CVE: null
Stroke null
MARINA 2011RBC lipids (1.8g group)
baseline:
EPA: 1.3%
DHA: 6.5%
12mo:
EPA: ~+2.5pp
DHA: ~+2.3pp
biomarkers only
MENU 2016RBC ALA
baseline: 0.12%
6mo: 0.17%
biomarkers only
NAT2 2013irrelevantirrelevant
Nodari 2011 HFcirculating EPA + DHA
baseline: 1.76%
12mo: 4.31%
CVE: benefit
Stroke: null
Arrythmia: null
OPAL 2010irrelevantirrelevant
Raitt 2005n-3 index
baseline: 4.7%
24mo: 8.3%
ACM: null
CVM: null
CVE: null
CHDM: null
CHDE: null
Arrythmia: null
MI: null
SCD: null
Ramirez-Ramirez 2013irrelevantirrelevant
REDUCE-IT 2019median serum EPA
baseline: 26.1 ug/ml
12mo: 144 ug/ml
ACM: null
CVM: benefit
CVE: benefit
CHDE: benefit
Stroke: benefit
Arrythmia: null
MI: benefit
SCD: benefit
Rossing 1996PLT lipids
baseline:
EPA: 0.42%
DHA: 1.66%
12mo:
EPA: 2.72%
DHA: 3.59%
ACM: no deaths
otherwise biomarkers only
Sandhu 2016irrelevantirrelevant
SCIMO 1999RBC PL
EPA +2.89pp
DHA +6pp
ACM: null
CVM: null
CVE: null
CHDM: null
CHDE: null
Stroke: null
MI: null
seAFOod Hull 2018irrelevantirrelevant
Shinto 2014irrelevantirrelevant
SHOT 1996serum EPA + DHA
baseline: 177.1 mg/L
6mo: 256.8 mg/L
ACM: null
CVM: null
CVE: null
CHDM: null
CHDE: null
Stroke: null
MI: null
SMART 2013n-3 index
baseline:
fish: 3.5%
fish + supplement: 3.6%
12mo:
fish: 3.4%
fish + supplement: 4.4%
ACM: no deaths
CVE: no events
otherwise metrics and biomarkers only
SOFA 2006serum CE EPA
baseline: 1.1%
12mo: 2.3%
ACM: null
CVM: null
CVE: null
CHDM: null
CHDE: null
Arrythmia: null
MI: null
SU.FOL.OM3 2010plasma n-3
+37% vs placebo
ACM: null
CVM: null
CVE: null
CHDM: null
CHDE: null
Stroke: null
Arrythmia: null
MI: null
SCD: null
WAHA 2016irrelevantirrelevant
Weinstock-Guttman 2005irrelevantirrelevant
WELCOME 2015irrelevantirrelevant
Zhang 2017irrelevantirrelevant
ACM: all-cause mortality, CVE: cardiovascular events, CVM: cardiovascular mortality, CHDE: coronary heart disease events, CHDM: coronary heart disease mortality, MI: myocardial infarction, SCD: sudden cardiac death, CE: cholesteryl esters, FA: fatty acids, ALA: alpha-linolenic acid, EPA: eicosapentaenoic acid, DHA: docosahexaenoic acid, RBC: red blood cell, PE: phosphatidylethanolamine, PL: phospholipids, PLT: platelets, AA: arachidonic acid

Conclusion

I substantially concur with the Cochrane review that any benefits of n-3 supplementation are likely small. This does not seem to be driven by failure to induce body lipid recomposition, but it is still possible that targeting fivefold improvements (such as in the REDUCE-IT trial) is the level at which you can more reliably detect hard outcome benefits. In any case, such high-dose supplementation isn’t harmful, so it’s a definite avenue for further studies.

Note: I have noticed some inconsistencies in some writeups of the studies in the review (regarding the n-3 biomarkers), compared what the studies reported, but it’s unclear to me if I’m in error or if the Cochrane authors made typos or thinkos sometimes.

References

  1. William S Harris, The omega-3 index as a risk factor for coronary heart disease, The American Journal of Clinical Nutrition, Volume 87, Issue 6, June 2008, Pages 1997S–2002S, https://doi.org/10.1093/ajcn/87.6.1997S
  2. Abdelhamid  AS, Brown  TJ, Brainard  JS, Biswas  P, Thorpe  GC, Moore  HJ, Deane  KHO, Summerbell  CD, Worthington  HV, Song  F, Hooper  L. Omega‐3 fatty acids for the primary and secondary prevention of cardiovascular disease. Cochrane Database of Systematic Reviews 2020, Issue 3. Art. No.: CD003177. DOI: 10.1002/14651858.CD003177.pub5. Accessed 22 August 2021.
  3. von Schacky C. Omega-3 index in 2018/19. Proc Nutr Soc. 2020 May 11:1-7. doi: 10.1017/S0029665120006989. Epub ahead of print. PMID: 32389149.