Guest Blog – Fishing for Toxins (part 2 of 2)

Today’s guest blog is Erik Davis’ followup to his post from Sunday. Part 1 was published on Sunday, December 6. Anyone interested in contributing to Skeptic North can email the editors, skepticnorth [at] gmail [dot] com.

As we discussed in Part 1 of this article last Sunday, there’s a significant disconnect between the way the public has reacted to the assurances of safe levels of mercury in vaccines and the warnings of higher levels of mercury in fish — which seems to be a case of our hard-wired mechanisms for evaluating risk butting heads with science.  Unfortunately for those of us who aspire to think critically, our brains are the product of 250,000 years of human evolution — and millions of years of pre-human evolution — while modern science has only been around for the last 150 or so.  And so we continue to assess the risks around us through a series of heuristics that worked well on the pre-historic plains of Africa, but less so on the modern streets of Toronto.

It’s pretty clear this is what’s causing many people to overstate the risk of mercury in vaccines, a topic that’s been well-addressed by others on this site. In today’s discussion, I’ll compare the actual levels of mercury in fish with the public health advisories, to see whether the same is true there as well.

Evaluating the risk

Two things we know for sure are that mercury poisoning is serious and that most of our mercury exposure comes from eating fish.  But that doesn’t, in and of itself, mean that eating fish is risky.  To know that, we need to understand how much mercury (specifically, a form called methyl-mercury) is too much, and how much we get from typical consumption.

In 1972, the World Health Organization (WHO) first set the Tolerable Daily Amount for adults at 0.47 micrograms of methyl-mercury per kilogram of body weight, and it was adopted by Health Canada as the national standard.  That advice has remained remarkably consistent for 37 years, with one carve out — the WHO now recommends that pregnant woman, women of childbearing age, and children under 12, stay below 0.23 micrograms per kilogram of body weight per day.  Canada’s standard is slightly more conservative, at 0.20 micrograms.  These recommendations result in the following tolerable daily amounts of mercury (in micrograms):

Staying below the tolerable level is important, because methyl-mercury is excreted (in feces and urine) quite slowly.  After stopping exposure, it can take 45-70 days for half of the accumulated methyl-mercury to leave the blood, so it’s not like you can just skip sushi night this week and take care of it.  Overexposure has serious consequences and is hard to correct, so staying within the tolerable levels is key.

We’re fortunate to have really good data on mercury levels in a wide variety of fish, in the form of the 2007 Human Health Risk Assessment of Mercury in Fish and Health Benefits of Fish Consumption issued by Health Canada.  Health Canada has been setting standards for mercury levels in fish since the late 1960′s, and this risk assessment was an update to the previous one 5 years earlier.  We all love to slag our government, but this is an example of them really doing their job well — weighing risks and benefits based on solid evidence and putting forth (in the companion risk management strategy) reasoned recommendations for consumers.

The good news is that the vast majority of fish are well within safe territory — defined as under 20 micrograms of mercury per 100g serving.  As you can see from Table 1, this is well within tolerable limits for daily consumption by the general population, though should be eaten no more than every 2 or 3 days by children and woman who are pregnant or planning to be so.  There is a massive list of fish that fits within this classification, and the most popular ones (canned tuna, salmon, shrimp, cod) are all quite a bit lower than the threshold (see Table 2 for just how low).  Here are the low-mercury fish you shouldn’t worry about:

That said, there are indeed a handful of higher mercury fish that consumers should take note of, as outlined on the following table (note that the shaded area are common low-mercury fish included for comparison).

This means that an average non-pregnant Canadian adult, weighing in about 65 kg, can eat all the shrimp and canned light tuna he wants — and shouldn’t worry about his yellowfin or halibut sashimi habit so long as it’s not daily. He needs to keep albacore, grouper, and sea bass on a less regular schedule, and shark, swordfish, and marlin should be reduced considerably, esp. since many of us easily go over the 100g serving size this table is based on.

And that’s more or less what the Health Canada recommended:

  • For fresh / frozen tuna, shark, swordfish, escolar, marlin and orange roughy, the general population can eat up to 150 g per week combined; pregnant/child-bearing-age woman get 150g/mo; children 5-11 get 125g/mo; and children 1-4 get 75g/mo
  • Canned albacore tuna is OK for the general population, but limited for pregnant/child-bearing-age women and children 1-4 to 75g/week; children 5-11 get 150g/week

Note that no recommendation was made with respect to grouper, sauger, sea-bass or barracuda, all of which are above fresh / frozen tuna and canned albacore on Table 2. This is based on popularity, i.e. fresh/frozen tuna and albacore wouldn’t have been included either but for the fact that we eat a lot of them in Canada. To my mind, that argument works for everything except sea bass, which is both higher in mercury and surely more popular than escolar & roughy (eww, slimehead). Many of us have already curtailed our sea bass consumption due to overfishing, and many chefs have stopped offering it, but it certainly wouldn’t be unreasonable to add it to the watch list.

Or don’t. After all, it’s a third of swordfish’s mercury count and half of shark’s, and you could eat 100g every other day and still stay inside your tolerable limit. The fact is that most of us don’t eat enough sea bass or any other bad fish to run afoul of the guidelines, even if we weren’t aware of them. The average daily consumption among Canadian’s who eat fish is only 26g, a safe level for everything except swordfish. That’s an average of course, and it’s not that hard to envision the extreme. Three restaurant-sized (150g+) portions of sea-bass a week could cause trouble for the average Canadian. Add in the tuna sushi pizza for lunch and the Szechuan grouper hot pot after last call, and mercury toxicity is certainly not out of the realm of possibility.

Yet still, it’s just not a big risk. It’s listed as a “rare disease” by the US National Institute of Health, a classification reserved for diseases that affect fewer than 200,000 Americans (about 6.5 people in 10,000), which puts an upper limit on incidence rates. Yet I couldn’t find actual incidence rate anywhere, which usually happens when something is too rare to track. This supposition is supported by the Canadian Health Measures Survey, self-billed as “the most comprehensive direct health measures survey ever undertaken on a national scale in Canada”, which found that fewer than 1% of Canadians 20-79 have mercury concentrations above the Health Canada guidance level of 20 micrograms per litre of blood. The authors admit this figure is the conservative upper limit of a data set with a lot of variability, and of course, the guidance levels themselves are highly conservative, so violating them certainly does not mean immediate mercury poisoning.

And the conservativism doesn’t stop there. The Health Canada research, like most such studies, does not break out the different types of mercury in the fish, primarily to due to cost & complexity. Instead, they count the total mercury and treat it as if it’s all methyl-mercury — which overstates the risk, since methyl-mercury may make up as little as 30% or as much as 95% of the mercury in fish. As well, samples are generally taken at the processing plants, not at the retail level, so the data may not capture mercury that’s eliminated in processing. All this to say that the recommendations have even more of a safety buffer than is usual for government advisories, so occasionally overdoing the sushi shouldn’t throw you into full blown panic attack.

Lizard Brain Gets It Right
And as we saw in Part 1, this is consistent with how we’re actually responding to the government advisories & media coverage, weighting the health benefits of fish higher than the mercury risks, and continuing to eat it despite some cognitive dissonance. But it’s important to point out that this response is just as arbitrary as the overreaction to mercury risk in vaccines, since it’s the result not of a logical process that evaluates the science, but of hard-wired heuristics that lend more credence to visceral, human anecdotes than cold statistics. It may only take a few high profile cases of mercury poisoning to change the tide of public opinion.

Which could be a bad thing for our hearts and life expectancy, as a recent analysis by the Harvard Center for Risk Analysis shows. The study looked at the effects on coronary heart disease, stroke, and IQ under 3 scenarios representing potential consumer responses to repeated public health warnings about mercury levels in fish. In the optimistic scenario, pregnant/child-bearing-age women ate the same amount of fish but switched to lower mercury species. In the middle scenario, pregnant/child-bearing-age women reduced their fish consumption by 17% without changing the mix. And in the pessimistic scenario, the entire adult population reduced their consumption by 17% without changing the mix. No surprises here, but the pessimistic scenario — the one most likely if media activity starts to coalesce around the type of heartbreaking personal anecdotes that ring our hard-wired risk bells — results in an additional 9400 deaths a year in a US population of 304M. Granted, that’s not a lot — only about 3 in 100,000 — but it is an entirely preventable side-effect of bad risk assessment.


Erik Davis thinks we really are living in the best of all possible worlds, and is astounded at how many people don’t see it. Accordingly, he’s agreed to be an occasional guest contributor on Skeptic North on the topic of risk perception. A long-time opponent of all things Web 2.0, he was finally induced to blog by PharmacistScott, who threatened to withhold his OxyContin.

Comments are closed.

  • Erik Davis

    Erik is a technology professional based in Toronto, focused on the intersection of the internet and the traditional media and telecommunications sectors. A reluctant blogger, he was inspired by the great work Skeptic North has done to combat misinformation and shoddy science reporting in the Canadian media, and in the public at large. Erik has a particular interest in critical reasoning, and in understanding why there’s so little of it in the public discourse. You can follow Erik's occasional 140 character musings @erikjdavis