Heavy Metals in Fish Oil: What the Testing Actually Reveals

Across 1,894 commercial fish oil samples, mean mercury measured 0.0005 mg/kg. The industry ceiling is 0.1 mg/kg. That is one two-hundredth of the allowance, in the largest contaminant dataset anyone has published on this category.

Fish oil has a reputation problem it has largely outgrown. Ours is one of the brands that helped build it — an earlier post on this site leaned on "heavy metals, PCBs and contaminants" as a reason to choose algae. On the metals, that framing does not survive the data.

The quality problem in fish oil is real. It just sits somewhere else.

What the assays actually find

The largest dataset comes from Nutrasource Diagnostics, working with the omega-3 trade body GOED. They pooled five years of finished-product testing: 1,894 samples, roughly 44 brands, eight countries.

Mean cadmium was 0.0006 mg/kg. Mean lead, 0.0019 mg/kg. Mean mercury, 0.0005 mg/kg. All sit against a 0.1 mg/kg specification. Of 792 samples tested for mercury, 99.75% came in under it.

Smaller independent surveys agree. Czech researchers tested 19 fish oil capsule products and found total mercury between 0.013 and 2.03 ng/g. Methylmercury, the form that matters, was below the detection limit in every capsule sample.

A 2023 Polish survey of 36 supplements found mercury from 0.023 to 0.427 micrograms per kilogram. The highest single reading was 0.4% of the legal ceiling for supplements.

That survey also produced the finding that should end the conversation. Mean mercury in the vegetable oil supplements was 0.218 µg/kg — more than three times the shark liver oils at 0.065 µg/kg.

Why the metal barely makes it into the oil

Methylmercury binds to cysteine, an amino acid in muscle protein. It travels with protein, not with fat.

Press a fish, and the oil separates from the tissue that holds most of the mercury. Refine that oil — steam stripping, cold filtration, molecular distillation — and more comes out. The chemistry does the work before any testing regime gets involved.

This is why the 2003 Massachusetts General Hospital analysis found nothing much. Five retail brands: three below the 6 µg/L detection limit, two at 10 and 12 µg/L. Six capsules of the worst performer carried 0.072 µg of mercury, about 2% of an average adult's daily intake from food.

The GOED data adds a wrinkle. Concentrated oils above 50% EPA+DHA averaged 0.0003 mg/kg mercury. Less concentrated oils averaged three times that, at 0.0009 mg/kg. Concentration involves more processing, and processing strips contaminants.

Cheaper, weaker oil is the dirtier oil. That is the opposite of what most people assume.

The contaminant that did have a problem

Persistent organic pollutants are a different story, because they are fat-soluble. They go where the oil goes.

In 2000–02 the UK Food Standards Agency surveyed 33 fish oil supplements sold here. Dioxin levels ranged from 0.18 to 8.4 ng/kg WHO-TEQ. Dioxin-like PCBs ran from 1.1 to 41 ng/kg.

Shortly afterwards, a maximum limit of 2 ng/kg WHO-TEQ came in for dioxins in marine oils. Twelve of those 33 UK products would have failed it.

The current GB limits for fish oils intended for human consumption are 1.75 pg/g fat for dioxins, 6.0 pg/g fat for dioxins plus dioxin-like PCBs, and 200 ng/g fat for the six indicator PCBs. Against those, the GOED dataset averaged 0.55 ppt WHO-TEQ for dioxins and furans across 484 samples.

Regulation worked. A category that had a genuine dioxin problem in 2002 no longer has one at the average.

Where contamination still concentrates

Averages hide outliers, and the outliers here are predictable. Liver oils and top-predator oils.

Canadian analysis of supplements on retail sale found a shark oil sample at 10,400 ng/g total PCBs. A mixed anchovy, mackerel and sardine oil in the same set measured 0.711 ng/g. That is four orders of magnitude apart.

Two Japanese deep-sea shark liver oils came in at 290 and 340 ng/g. Cod liver oil sits between the extremes — higher than body oil, because the liver is where a fish stores what it cannot excrete.

The rule is simple. Small, short-lived, low in the food chain, body oil not liver oil. Anchovy and sardine oils are consistently the cleanest fish sources tested.

The number nobody puts on the label

Here is where fish oil does have a measurable problem, and it has nothing to do with metals.

Omega-3s have five and six double bonds. Double bonds oxidise. As an oil goes off, the EPA and DHA you paid for turn into peroxides, then into aldehydes and ketones.

In 2015 a New Zealand team tested every encapsulated fish oil on that market. Of 32 products, 83% exceeded the recommended peroxide value and 50% exceeded the TOTOX ceiling of 26. Only 8% cleared all three oxidation checks. Only three products contained the EPA and DHA their labels claimed; 69% held under two-thirds of it.

Two years later, 47 products from the same market were tested again, this time by GOED. Compliance was 72% on peroxide value, 77% on TOTOX, and 91% on label content.

Those results cannot both be typical. The second study was run by the trade association whose members were being graded, which is worth stating plainly. The first was a single academic laboratory. Sampling, storage and pooling methods differed.

Take the honest reading: somewhere between 8% and 77% of fish oils on a given shelf meet the voluntary freshness standard, and nobody can tell you which from the packaging.

Does oxidised oil actually harm you?

Less than the oxidation panic implies, on the evidence available.

A Norwegian group randomised 54 healthy adults to 8 g a day of fresh fish oil, oxidised fish oil, or high-oleic sunflower oil for seven weeks. Both fish oil arms delivered 1.6 g a day of EPA+DHA.

Read that carefully before relaxing. It is one trial, seven weeks, in adults aged 18 to 50, with 17 to 19 people per arm. It is powered to detect large effects and nothing else.

The industry oxidation limits themselves are built on palatability, not on health outcomes. There is not enough human data to set them on anything else. That is an honest gap, and it cuts both ways: no one has shown oxidised oil is harmful, and no one has shown it is fine over years.

What oxidation definitely costs you is potency. Oxidised EPA and DHA are no longer EPA and DHA. In the 2015 survey, missing label content correlated with anisidine value at r = 0.45.

If you are taking 2 g a day and a third of it has degraded, you are on 1.3 g. That is worth knowing when you work out your dose.

What to ask a brand for

Stop asking about heavy metals. The answer is always the same and the number is always tiny.

Ask instead for a batch-specific certificate of analysis carrying five figures: peroxide value, anisidine value, TOTOX, dioxin plus dioxin-like PCB TEQ, and measured EPA and DHA in mg per capsule.

Reasonable thresholds to look for: peroxide value under 5 meq/kg, anisidine under 20, TOTOX under 26. The European and British Pharmacopoeias allow looser numbers — peroxide under 10, TOTOX under 50 — so a certificate quoting those is quoting the softer standard.

Three practical checks that cost nothing:

• Match the batch code on the certificate to the code on your bottle. A generic PDF from 2023 tells you nothing about the bottle in your hand.
• Bite one capsule. Fresh marine oil tastes of almost nothing. Rancid oil is sharp, and the burp is worse. This is a crude test that catches the worst cases.
• Store it cold and dark. Room-temperature oils in clear bottles on a sunny shelf keep oxidising after you buy them. The fridge costs you nothing.

Best-before date, price, country of origin and brand exclusivity all failed as predictors of quality in the 2015 survey. Expensive did not mean fresh.

Where this leaves the choice

Refined fish oil is not a mercury risk. That claim, including the version we made ourselves, does not hold up against 1,894 samples and three independent surveys.

Dioxins and PCBs were a genuine UK problem in 2002 and are largely handled now by regulation, except in liver oils and shark oils. Avoid those two and the persistent-pollutant question mostly goes away.

Oxidation is the live issue, and it is unresolved. It costs you measurable potency, and its long-term health significance is unknown because the trials have not been run.

Algae-derived DHA sidesteps the marine food chain entirely — grown in closed steel vessels, so there is no bioaccumulation to refine out. But it oxidises on exactly the same chemistry, and it needs exactly the same certificate. The source changes which questions matter. It does not exempt anyone from being asked.

References

1. John, T. E., Rebry, R. M., Yan, K. L., Rice, H. B., & Rowe, W. J. "Measurement of Environmental Contaminants in a Globally-Representative Sample of Fish Oil Supplements." Nutrasource Diagnostics / Global Organization for EPA and DHA Omega-3s (GOED) white paper, covering finished-product testing data from 2005 to 2010.
2. Foran, S. E., Flood, J. G., & Lewandrowski, K. B. (2003). "Measurement of mercury levels in concentrated over-the-counter fish oil preparations: is fish oil healthier than fish?" Archives of Pathology & Laboratory Medicine, 127(12), 1603-1605.
3. Smutna, M., Kruzikova, K., Marsalek, P., Kopriva, V., & Svobodova, Z. (2009). "Fish oil and cod liver as safe and healthy food supplements." Neuro Endocrinology Letters, 30(Suppl 1), 156-162. PMID: 20027164.
4. Brodziak-Dopierala, B., et al. (2023). "Mercury Exposure from the Consumption of Dietary Supplements Containing Vegetable, Cod Liver, and Shark Liver Oils." International Journal of Environmental Research and Public Health, 20(3), 2129. PMID: 36767496.
5. Fernandes, A. R., Rose, M., White, S., Mortimer, D. N., & Gem, M. (2006). "Dioxins and polychlorinated biphenyls (PCBs) in fish oil dietary supplements: occurrence and human exposure in the UK." Food Additives & Contaminants, 23(9), 939-947.
6. Albert, B. B., et al. (2015). "Fish oil supplements in New Zealand are highly oxidised and do not meet label content of n-3 PUFA." Scientific Reports, 5, 7928.
7. Bannenberg, G., Mallon, C., Edwards, H., Yeadon, D., Yan, K., Johnson, H., & Ismail, A. (2017). "Omega-3 Long-Chain Polyunsaturated Fatty Acid Content and Oxidation State of Fish Oil Supplements in New Zealand." Scientific Reports, 7, 1488.
8. Ottestad, I., et al. (2012). "Oxidised fish oil does not influence established markers of oxidative stress in healthy human subjects: a randomised controlled trial." British Journal of Nutrition, 108(2), 315-326. PMID: 22136711.
9. Commission Regulation (EC) No 1881/2006 setting maximum levels for certain contaminants in foodstuffs, as retained in GB law.
10. Commission Regulation (EU) 2023/915 on maximum levels for certain contaminants in food, Table 1: marine oils.

Numbers on the certificate, not on the front of the bottle

Our DHA comes from fermented algae rather than fish, so the marine food chain never enters the picture. Every batch is tested for oxidation and contaminants.

See Omega-3 DHA →