Two Servings of Fish and a Daily Capsule Do the Same Thing to Your Blood

The chilled cabinet is where this gets decided. A fillet of salmon, priced by weight, with a use-by date three days out. Twenty feet away, two hundred softgels in a plastic jar for less than the fillet costs, no use-by date worth worrying about, no cooking, no smell in the flat afterwards. Both claim the same molecule.

I went looking for whether anyone had put the two side by side and measured what came out. Four groups had. They split two against two, and the reason they disagree turns out to be more useful than a clean answer would have been.

Two trials, two industry funders, one answer

Healthy women were randomised to get an average of 485 mg of EPA and DHA a day — either from two servings of oily fish a week, salmon and albacore tuna, or from one to two capsules daily. Sixteen weeks, red blood cell fatty acid content as the measure.

The fish group went from 4.0 ± 0.6% of total fatty acids to 6.2 ± 1.4%. The capsule group went from 4.3 ± 1.0% to 6.2 ± 1.4%. Both p < 0.0001. No significant difference between groups (Harris et al., Am J Clin Nutr 2007). Fish pulled ahead briefly — EPA rose faster in the fish group over the first four weeks (p = 0.01) — then the rates converged and the endpoint was identical.

The funding line belongs here rather than in a footnote: the work was supported in part by a Fisheries Scholarship grant from the National Fisheries Institute, a seafood industry trade association, awarded to the first author.

Four years later, a New Zealand group ran a different version. Forty-four healthy volunteers, eight weeks, randomised to either two 120 g servings of salmon a week or to two, four or six salmon oil capsules daily — a design that shows the dose-response rather than a single comparison. Both raised long-chain n-3 status. Salmon was not more effective (Stonehouse et al., Br J Nutr 2011).

That study was funded in part by the New Zealand King Salmon Company, which supplied the fish.

Two trials, two continents, each part-funded by people who sell fish, each testing whether fish beats capsules for omega-3, each reporting no advantage. Conflicts of interest matter most when a result runs with them. Both of these run against, which is why I find them more persuasive than I would a null from a neutral lab.

The samples were small — eleven completed the fish arm in 2007, forty-four across four arms in 2011 — and the first enrolled only premenopausal women. Neither is definitive. Together they are the best evidence available.

The literature is not unanimous, and the split is worth being exact about. The 2011 authors summarise it themselves: two studies found fish more effective at raising plasma EPA and DHA, two found no difference. In one of the studies favouring fish, matching the plasma EPA increase produced by 0.38 g/day of EPA from salmon reportedly required 0.80 g/day from capsules — more than double. I could not read that paper directly and am reporting its result as described in the 2011 study, which also notes its dose-response design was less than optimal.

The 2011 authors offer a specific explanation for why their own result went the other way. Their participants took the capsules with food, and the omega-3 in both their salmon and their capsules was in triglyceride form, whereas the study finding fish superior used ethyl esters. If that holds, the practical instruction isn’t “buy fish.” It’s “don’t swallow the capsule on an empty stomach, and check which chemical form you bought.” That’s a testable claim rather than a settled one.

The number the whole category is priced against

Here the arithmetic runs opposite to how the aisle is organised.

A JAMA review pooled the evidence on fish intake and coronary heart disease mortality and found risk falling by 14.6% for every additional 100 mg per day of EPA plus DHA — up to a point. Above roughly 250 mg per day, the additional benefit was reported as 0.0% per further 100 mg (Mozaffarian and Rimm, JAMA 2006, DOI: 10.1001/jama.296.15.1885). The curve flattens.

The American Heart Association’s 2018 advisory, chaired by one of that review’s two authors, recommends one to two 3.5-ounce servings of non-fried seafood weekly, preferably oily species, and notes that one serving of fatty fish a week supplies approximately the 250 mg per day figure. Multiple servings of lean fish like cod are needed to get there.

One serving. The second isn’t wrong, but it isn’t where the evidence concentrates, and almost every version of this advice treats 250 mg as a floor to clear rather than a plateau to reach.

For scale, from the USDA nutrient database: a 143 g cooked Atlantic herring fillet carries about 1.58 g of DHA. A 3 oz portion of cooked farmed Atlantic salmon, about 1.24 g. A tin of sardines in tomato sauce with the bones in, roughly 0.77 g of DHA per cup, alongside 21 g of protein, 240 mg of calcium and a very large amount of B12.

Where the food numbers stop agreeing with each other

Go looking for “omega-3 per 100 g” and you’ll find confident tables that contradict one another. Cooked salmon appears at 2,150 mg of combined EPA and DHA in one compilation and 1,500 mg in another. Herring at 2,150, 1,800 and 1,730 depending where you look. Mackerel from 2,600 to 2,670.

This isn’t sloppiness so much as an unstated variable. Wild and farmed differ. Raw and cooked differ. Species inside a category differ — Atlantic mackerel is not king mackerel. Most compilations don’t say which they used. The USDA database does, which is why the figures above come from it.

The practical version: treat any single number for “fish” as a rough band, not a value. And note which species aren’t in the conversation at all. Cod, haddock, tilapia, shrimp, scallops and yellowfin tuna all sit low. Eating fish twice a week and eating oily fish twice a week are different interventions.

The label on the jar is not the number you want

The most common error in this category costs nothing to fix.

Capsules are marketed by total fish oil. A product sold as “1200 mg Fish Oil” is telling you the weight of the oil, not the weight of the EPA and DHA in it, and those can differ by a factor of three or four. The number you need is on the supplement facts panel, usually two lines down, listing EPA and DHA separately.

Do that arithmetic before comparing prices. A jar that looks cheap per softgel may not be cheap per 250 mg of the thing you’re buying it for.

The part where both options run into the same wall

This is the finding I didn’t expect, and it needs its species stated in every sentence, because it has not been shown in humans.

DHA doesn’t diffuse into the brain. It crosses the blood-brain barrier through a specific transporter, Mfsd2a, identified through gene-deletion experiments in mice — animals lacking it developed microcephaly and brain DHA deficiency. The transporter accepts DHA only when esterified to lysophosphatidylcholine. It does not transport unesterified DHA.

In a 2017 mouse study, oral LPC-DHA at 40 mg/kg for 30 days more than doubled brain DHA content and improved performance on a spatial memory task. The same quantity of free DHA raised DHA in adipose tissue and heart but not in brain, and had no effect on the maze (Sugasini et al., Scientific Reports 2017). Mice throughout.

In that paper’s introduction, the authors state that the available preparations — fish oil, krill oil, algal DHA, DHA-enriched egg phospholipids, ethyl esters, and sardines — do not appreciably raise brain DHA in adult mammals, though peripheral tissues enrich under the same conditions.

Sardines are on that list. The bottleneck isn’t a supplement problem that whole food solves. If it applies to humans, it applies to the fillet as much as to the capsule.

Whether it applies to humans is unknown. I found no trial measuring adult human brain DHA in response to any oral form. What can be said honestly: the mechanism is well characterised in rodents, it makes a specific prediction, and nobody has tested the prediction in people.

What the cognitive trials say, and what they were measuring

A 2020 meta-analysis of 38 randomised trials, restricted to studies lasting at least 24 weeks, is described by papers citing it as finding a statistically significant but clinically unimportant benefit on the MMSE — a difference of less than 1% of baseline — with no effect on new neurocognitive illness or on cognitive impairment in other domains. I’m reporting that at one remove: I could not read the paper itself, only how several independent sources summarise it, and they agree with each other. It still deserves a direct reading before anyone leans on it hard.

What I could read: a 2025 dose-response meta-analysis of 58 trials found each 2,000 mg per day increment associated with improved attention (SMD 0.98, 95% CI 0.41–1.54) and perceptual speed (SMD 0.50, 95% CI 0.05–0.95). Those look substantial until you read the certainty ratings the authors attached — low for attention, moderate for perceptual speed. The pooled population included people with dementia, Alzheimer’s disease and mild cognitive impairment alongside healthy adults. The dose-response curves were non-linear in a way that should make anyone cautious: episodic memory fell as dose rose before turning upward (p for non-linearity = 0.01).

A separate 2025 overview of systematic reviews in non-demented and MCI adults reports that many individual trials were null, and that one meta-analysis found DHA supplementation below 580 mg per day ineffective.

Sit that next to the 485 mg per day that two servings of oily fish averaged out to in the 2007 trial.

And a 2025 meta-analysis across nine systematic reviews, 14 trials and 26,881 participants aged 40 and over found a statistically significant but modest improvement — measured on the Mini-Mental State Examination, a dementia screening instrument. Whatever the MMSE detects, it is not whether a thirty-four-year-old writes better code on a Tuesday.

No trial has tested omega-3 intake in healthy adults in their twenties, thirties and forties with same-week cognitive performance as the endpoint. No direct evidence exists for the question most people are actually asking in the aisle.

Safety, in both directions

Mercury runs opposite to most people’s assumption. Five over-the-counter fish oil brands were analysed by cold vapour atomic absorption spectroscopy; mercury ranged from undetectable, below the 6 µg/L limit of detection, to a negligible 10–12 µg/L — comparable to the concentration normally present in human blood (Foran et al., Arch Pathol Lab Med 2003, DOI: 10.5858/2003-127-1603-MOMLIC). Five brands, over twenty years ago. Later speciation work put mercury impurities in fish oil one to two orders of magnitude below industrial safety standards. On this specific measure, the capsule is the lower-exposure option.

For fish, the JAMA review concluded that for adults the benefits of one to two weekly servings exceed the contaminant risks, estimating the coronary benefit from salmon as hundreds to a thousand times greater than the estimated lifetime cancer risk from its contaminants. That conclusion was formally contested: a paper in Environmental Health argued the review gave inadequate weight to methylmercury’s cardiovascular risks and that its broad conclusion risked misleading public health guidance, while conceding the evidence for mercury’s cardiac toxicity is not conclusive. Two public documents disagreeing about the same literature. The practical mitigation both accommodate: vary the species, avoid the large predators.

The atrial fibrillation question is real and dose-dependent. Several large cardiovascular outcome trials reported more atrial fibrillation in the omega-3 arms. A meta-analysis of those trials found supplementation associated with increased risk — HR 1.25 (95% CI 1.07–1.46, p = 0.013) — with a clear dose gradient: HR 1.49 (95% CI 1.04–2.15) in trials testing more than 1 g/day against HR 1.12 (95% CI 1.03–1.22) at 1 g/day or less, p for interaction < 0.001. Meta-regression put the increase at HR 1.11 per additional gram (Gencer et al., Circulation 2021, DOI: 10.1161/CIRCULATIONAHA.121.055654).

Read the doses before you read the hazard ratios. Those trials tested 1 to 4 grams a day in patients with a mean age of 65 and established cardiovascular risk. The intake this article is about — around 250 mg a day, from two tins of sardines or a modest capsule — sits four to sixteen times below the lowest of them. It’s a reason not to take high-dose omega-3 casually. It isn’t a reason to skip the fish.

A 2025 meta-analysis expanding the pool to 34 trials and 114,326 individuals concludes the risk concentrates in high-risk cardiovascular patients on high doses. Two caveats in the same breath: it’s a preprint that has not completed peer review, and one of its authors holds stock in a company selling blood omega-3 testing. That conflict points toward reassurance, which is the direction warranting more scrutiny, not less.

Bleeding. A review of the literature from 1980 to 2017 concluded that fish oil alone does not produce a clinically significant coagulopathy capable of causing surgical bleeding, though the effect becomes noticeable in patients also taking antiplatelet therapy and, to a lesser extent, factor Xa inhibitors or warfarin (Annals of Thoracic Surgery 2018). A retrospective analysis of 573 long-term warfarin patients, 145 of them taking fish or krill oil, found no significant difference in time in therapeutic range or in bleeding incidence between groups.

If you take an anticoagulant or antiplatelet drug, that’s a conversation with your prescriber before you change anything — including a conversation about the fact that measured EPA and DHA content varies widely between products, so “the same dose” may not be.

Capsule oxidation is a real measurement problem with genuinely conflicting results. A Canadian analysis of 171 supplements found half exceeding voluntary limits on at least one oxidation marker. A US analysis of 16 top-selling liquid products found over half missing their label claims for EPA and DHA and a quarter over the recommended peroxide value. But a multi-laboratory analysis of 47 products on the New Zealand market found 72%, 86% and 77% compliant with the voluntary maxima for peroxide value, anisidine value and total oxidation, with 91% meeting content claims — a result its authors describe as in stark contrast to an earlier assessment of the same market. Two studies, one market, opposite conclusions. I can’t resolve that and won’t pretend to.

What oxidised oil does in people has not been established. A commentary in the Journal of Nutritional Science states plainly that evidence linking oxidation markers to toxicological effects in humans is lacking — and raises the more uncomfortable possibility that supplements used in past clinical trials may themselves have been oxidised, potentially confounding the trial literature.

How to check any of this yourself

Three habits, all transferable beyond this topic.

Read the supplement facts panel, not the front of the jar. Front-of-pack numbers describe the vehicle; the panel describes the cargo.

When a paper reports an effect size, find the certainty rating the authors attached to it. A standardised mean difference of 0.98 rated low certainty isn’t a stronger finding than one of 0.30 rated high — it’s a weaker one wearing a bigger number.

And if you ever measure your own omega-3 index, wait four months. Plasma phospholipid levels plateau at about four weeks, but red blood cell content kept climbing through week sixteen in the 2007 trial. Test at six weeks and you’ll be reading a number that hasn’t finished moving.

The bet

For the omega-3 specifically, the fish and the capsule are interchangeable, and both trials that tested it had every incentive to find otherwise.

That makes this an asymmetric bet with an unusual shape, because the asymmetry isn’t in the omega-3 at all.

If you’ll actually eat it, eat it. You reach the plateau on one serving, and the New Zealand trial found something the capsule measurably did not deliver: selenium status rose in the salmon group and not in the capsule groups. That’s a demonstrated difference, not an argument from composition, and it’s the honest core of the food-first case here. The broader claim that whole fish outperforms isolated oil across an entire matrix of nutrients is made in a 2023 review in Frontiers in Nutrition — that one remains plausible reasoning rather than a tested outcome.

If you won’t eat it — and the average American manages 1.3 servings a week against a recommendation of two — the capsule puts the same EPA and DHA in your blood for less money, and buying it is not a failure of discipline. Take it with a meal, and check whether you bought triglyceride or ethyl ester form, because the one trial where fish clearly won used the ester.

What neither one is, on current evidence, is a cognitive intervention. The best-supported reason to care about either is your heart. The claim that it will sharpen your thinking rests on trials in older and cognitively impaired populations, measured on dementia screening instruments, rated low certainty by the people who pooled them — and on a transport mechanism that, in mice, neither option appears to satisfy.

About this article

Written by Leah Elish. Leah covers nutrition claims, dose arithmetic, and the distance between a research protocol and a shopping list. Not a dietitian or clinician — traces nutritional claims back to the studies that supposedly support them.

Medical review: None. NeuriFuel does not currently have a licensed clinician on the editorial team, and this article has not been medically reviewed. We state this rather than implying an authority we do not have. See our About page for our full methodology.

Sources: Twenty-three, comprising three randomised controlled trials, eight meta-analyses or systematic reviews, four laboratory product analyses, two animal studies labelled as such wherever they appear, one narrative review, one guideline advisory, one published critique, one commentary, one retrospective cohort analysis, and the USDA food composition database. Funding disclosures were retrievable for four studies and are stated inline, including for the two whose conflicts point toward this article’s own conclusions; for the remainder they were not available in the sources I could reach, and I have marked that absence rather than assume it was clean. Two studies are cited as described by papers citing them rather than read directly, and are identified as such in the text. Values that could not be established are marked, and where sources conflict, both are given rather than one chosen quietly.

Corrections: Found an error? Write to hello@neurifuel.com with a source and we will fix it and log the correction.

Last updated: 31 July 2026

References

1. Harris WS, Pottala JV, Sands SA, Jones PG (2007). Comparison of the effects of fish and fish-oil capsules on the n–3 fatty acid content of blood cells and plasma phospholipids. American Journal of Clinical Nutrition 86(6):1621–1625. Supported in part by a Fisheries Scholarship grant from the National Fisheries Institute, a seafood industry trade association.
2. Stonehouse W, Pauga MR, Kruger R, Thomson CD, Wong M (2011). Consumption of salmon v. salmon oil capsules: effects on n-3 PUFA and selenium status. British Journal of Nutrition. Funded in part by the Institute of Food, Nutrition and Human Health at Massey University, the Foundation for Research Science and Technology, and the New Zealand King Salmon Company Limited, which supplied the salmon.
3. Mozaffarian D, Rimm EB (2006). Fish Intake, Contaminants, and Human Health: Evaluating the Risks and the Benefits. JAMA 296(15):1885–1899. DOI: 10.1001/jama.296.15.1885. Supported by the National Institutes of Health.
4. Stern AH. Public health guidance on cardiovascular benefits and risks related to fish consumption. Environmental Health. PMC2164937. Published critique of reference 3.
5. Rimm EB et al. (2018). Seafood Long-Chain n-3 Polyunsaturated Fatty Acids and Cardiovascular Disease: A Science Advisory From the American Heart Association. Circulation. DOI: 10.1161/CIR.0000000000000574.
6. Sugasini D et al. (2017). Dietary docosahexaenoic acid (DHA) as lysophosphatidylcholine, but not as free acid, enriches brain DHA and improves memory in adult mice. Scientific Reports 7:11263. Animal study.
7. The lysolipid transporter Mfsd2a regulates lipogenesis in the developing brain. PLOS Biology. PMC6093704. Animal and cell studies.
8. Brainard JS et al. (2020). Omega-3, omega-6, and polyunsaturated fat for cognition: systematic review and meta-analysis of randomized trials. J Am Med Dir Assoc 21(10):1439–1450. DOI: 10.1016/j.jamda.2020.02.022. Result reported here as described by citing papers; primary not read.
9. A systematic review and dose response meta-analysis of omega-3 supplementation on cognitive function (2025). Scientific Reports. DOI: 10.1038/s41598-025-16129-8.
10. Omega-3 Polyunsaturated Fatty Acids and Cognitive Decline in Adults with Non-Dementia or Mild Cognitive Impairment: An Overview of Systematic Reviews (2025). Nutrients 17(18):3002.
11. Meta-analysis of nine systematic reviews of omega-3 supplementation and MMSE outcomes (2025), 14 RCTs, 26,881 participants.
12. Comparison of the effect of omega-3 supplements and fresh fish on lipid profile: a randomized, open-labeled trial. PMC5865539.
13. Santos HO, May TL, Bueno AA (2023). Eating more sardines instead of fish oil supplementation: beyond omega-3 polyunsaturated fatty acids, a matrix of nutrients with cardiovascular benefits. Frontiers in Nutrition 10:1107475. DOI: 10.3389/fnut.2023.1107475. Narrative review.
14. Gencer B, Djousse L, Al-Ramady OT, Cook NR, Manson JE, Albert CM (2021). Effect of Long-Term Marine ω-3 Fatty Acids Supplementation on the Risk of Atrial Fibrillation in Randomized Controlled Trials of Cardiovascular Outcomes: A Systematic Review and Meta-Analysis. Circulation 144:1981–1990. DOI: 10.1161/CIRCULATIONAHA.121.055654.
15. Effects of Omega-3 Fatty Acid Treatment on Risk for Atrial Fibrillation: An Updated Meta-Analysis of 34 Trials including 114,326 Individuals (2025). Preprint; not peer-reviewed. One author reports holding stock in a laboratory offering blood fatty acid testing.
16. Role of Fish Oil in Post-Cardiotomy Bleeding: A Summary of the Basic Science and Clinical Trials (2018). Annals of Thoracic Surgery.
17. Pryce R, Bernaitis N, Davey AK et al. (2016). The Use of Fish Oil with Warfarin Does Not Significantly Affect either the International Normalised Ratio or Incidence of Adverse Events in Patients with Atrial Fibrillation and Deep Vein Thrombosis: A Retrospective Study. Nutrients 8:578. PMC5037562.
18. Foran SE, Flood JG, Lewandrowski KB (2003). Measurement of Mercury Levels in Concentrated Over-the-Counter Fish Oil Preparations: Is Fish Oil Healthier Than Fish? Arch Pathol Lab Med 127(12):1603–1605. DOI: 10.5858/2003-127-1603-MOMLIC.
19. Albert BB et al. (2017). Omega-3 Long-Chain Polyunsaturated Fatty Acid Content and Oxidation State of Fish Oil Supplements in New Zealand. PMC5431122.
20. Quality analysis of commercial fish oil preparations. PMID: 23255124.
21. Fish oil supplements, oxidative status, and compliance behaviour: regulatory challenges and opportunities (2020). PLOS ONE. PMC7774961.
22. Fishing for answers: is oxidation of fish oil supplements a problem? Journal of Nutritional Science. PMC4681158. Commentary describing the Canadian analysis of 171 supplements.
23. USDA National Nutrient Database for Standard Reference, Release 28 — EPA and DHA content tables.