Dan's Data letters #160Publication date: February 2006.
Last modified 03-Dec-2011.
Re the last column, this Chicago Tribune piece is an interesting little read. Basically anecdotal evidence (minimal numbers of fish tested) but apparently more testing than has been done by anyone else.
If they're right, then this could indeed be significant, assuming it's repeatable. But the significance is mainly in showing up the inadequacy of the government monitoring, not in finding a Terrible Lurking Danger. The FDA consumption limit for 0.5-PPM-mercury fish is, itself, very probably much lower than it needs to be.
As you say, the Tribune didn't appear to bust themselves in the methodological rigour department, and all sorts of alarming things can scuttle out if you turn over the rock of newspaper-funded science. But if the Tribune really got actual scientists at Rutger's to do the testing, though (not some guy in a garden shed), then the results should be kosher. The numbers aren't outrageous, either.
The worst they found was swordfish (which you expect to be bad, since it's a big predatory fish; long-lived, eats other fish and concentrates their mercury in its flesh, a process known as "biomagnification"), with a level of 1.41 parts per million. Tuna is also a big carnivorous fish, though. So it's not entirely surprising that they found tuna, among other commonly eaten fish, to be around 0.5 PPM rather than the very-probably-harmless 0.2 PPM or less mentioned on the old FDA page and in their current advisory.
That page classes 0.5 PPM fish as being in the not-more-than-14-ounces-per-week category, or less for women who are, or are expecting to become, pregnant.
It's scaremongering, though, to say that "millions of American women [by which they mean the pregnant or likely-to-be ones] would exceed the U.S. mercury exposure limit by eating just one 6-ounce meal [of the worse samples] in a week", without making clear that they'd only exceed the exposure limit for that week. Any level of mercury contamination that didn't kill the fish in the first place ought to be no big deal to eat, once. Mercury-in-food issues are all about repeated consumption and accumulation of the poison.
How do I know that?
Well, pure liquid dimethylmercury is probably the stuff that people who say "mercury is the second most toxic substance in the world" are thinking of, to the extent that they're thinking at all. Even by the standards of hard-core organic chemists, dimethylmercury is alarmingly nasty. It's what killed the unfortunate Karen Wetterhahn, among others. Its oral lethal dose in humans is unknown, but certainly in the fractions-of-a-milligram range. 100 micrograms (ug) orally would probably kill you very thoroughly. I wouldn't be surprised if 10ug would do the job, eventually.
This is right up there in the pantheon of poisons. Dimethylmercury is roughly level with botulin toxin. The lethal dose for botulinum is about 300 picograms per kilogram - so only 30ug is needed to speedily kill a husky 100-kilogram adult. Even ricin is probably less toxic than dimethylmercury - you need about a milligram of ricin to kill a human.
The movie-plot poisons are rather less toxic. Good old arsenic needs 250mg to probably do in an adult. Potassium or sodium cyanide's about the same as arsenic trioxide, but cyanic acid is only about 50mg. Strychnine? About 350mg, for an adult. Curare's pretty hot stuff, though - well under one milligram per kilogram'll do you.
Getting vaguely back towards the subject, though - if you're unlucky enough to be eating fish that's loaded up with 0.5 parts per million of pure dimethylmercury, only 200 grams of the stuff will give you a one-milligram dose, and you'll be in big trouble.
Fortunately, the "mercury" in fish is likely to be mostly, if not entirely, methylmercury, which lacks one of the methyl groups of dimethylmercury and, while still bad, is far less toxic.
The oral guinea-pig LD50 for methylmercury chloride is 21 milligrams per kilogram. Let's fudge some numbers regarding comparative toxicity, and concentration by weight of nasty-methylmercury-ion versus harmless-chloride-ion, and assume that 20mg/kg of the stuff is a fatal, or at least you're-gonna-feel-real-bad, dose for humans.
That adds up to 1.5 grams of the stuff for an unremarkable-sized 75-kilogram human. Which is 300 kilograms of 0.5 PPM fish.
To get down to the point where one big serving of fishier-than-usual tuna casserole will do you a lot of harm, you'd need something like five-hundred-PPM methylmercury concentration.
So a single big meal of 1.5-PPM swordfish should not, in fact, endanger anybody, pregnant or not. Perhaps a single dose of methylmercury around 1% that needed to kill you, taken during pregnancy, increases the risk that the baby will be autistic, or something. But that's still a very big fat "perhaps" at the moment, despite the convictions of the anti-mercury fanatics.
These people are often opposed to childhood vaccinations, because they're concerned about thimerosal, an ethylmercury-salt vaccine preservative. They like the "second most poisonous substance" line a lot, but they have some dosage problems to get around.
Nobody actually makes vaccines with significant amounts of thimerosal in them any more (and no, autism rates didn't suddenly drop when they took the thimerosal out, thank you for asking). When they did, a normal 0.5ml dose of such a vaccine would deliver only 50 micrograms of thimerosal, and a much smaller dose of ethylmercury and the other compounds in thimerosal. Those other compounds include sodium hydroxide, another Nasty Substance which the thimerosal-haters overlook because it's obvious that, um, there's not enough of it in the vaccination to do any harm.
Animal tests found thimerosal most toxic when give to mice intravenously; then the LD50 was 30mg/kg. Direct dumb scaling up to human doses is close enough for government work. If you do that and consider a 3.5 kilogram baby (straight out of the birth canal and bang, give 'em the needle!), you get 105 milligrams for the same 50% chance of death.
50 micrograms from a vaccine is 1/2100th of that dose.
Modern vaccines that have a "trace amount" of thimerosal in them may contain as much as 1 microgram per dose. Now we're looking at 1/105,000th of a killing dose.
People are exposed to doses this tiny of various deadly poisons all the time, and survive. Lots of ordinary substances, like table salt, become lethal if you somehow manage to get a couple of thousand times the usually-consumed amount into your body over a short period of time. It's possible that one two-thousandth of the likely-lethal dose of something will affect you in some subtle way, but it's far from proven - and, actually, many of the symptoms of ordinary high-level organic-mercury poisoning go away when the victim stops ingesting the stuff.
Anti-mercury activists also typically get exercised about amalgam fillings, which outgas small amounts of mercury all the time, and more when you're chewing something.
This, though, is metallic mercury, which is really not that big a deal if you don't happen to be made of aluminium.
Mercury is a poisonous heavy metal that accumulates in the body, and because it's liquid at room temperature it's much more dangerous than other heavy metals, because it emits vapour all the time.
So you don't want to screw with it.
But lead is a poisonous heavy metal that accumulates in the body, and people who do electronics soldering in poorly ventilated areas breathe in plenty of lead vapour. And yet we don't all suffer from brain rot after a year.
The dose, once again, makes the poison, and the dose of mercury from fillings, or lead from the occasional snort of solder smoke, is tiny. If some voodoo priest hasn't been sprinkling the stuff all over your house, metallic mercury poisoning is very rarely found in people who don't, for some reason, work or play with the stuff all the time.
If you're pregnant and want to apply the Precautionary Principle and studiously avoid any foodstuff that might have more mercury in it than is usual then sure, be my guest. But I strongly suspect you'd protect your foetus better by avoiding more definitely risky behaviour - like making sure you don't drive anywhere unless it's strictly necessary.
It should also be noted that the Chicago Tribune's accurate-enough description of low-level mercury poisoning symptoms as including "headaches, fatigue, numbness in the hands and feet, and a lack of concentration" is, of course, a recipe for hypochondria.
I get the occasional headache. I feel fatigued and have trouble concentrating sometimes (usually when I haven't been getting any exercise).
Oh no! I'm poisoned!
"So there I was, thallium dust all over my face..."
Oh, now, come on. You can't leave it hanging like that!
I'm putting together a small element collection. I'm not interested in building a comprehensive collection, because it's time-consuming and expensive and a lot of elements aren't actually very interesting. But I've got a lovely tungsten cylinder (and may be buying another cylinder or two, now that they've got other elements on offer), a decorative bismuth egg, a good-sized copper bar for Lenz's-Law magnetic-braking demos, a silicon boule top, some unusual chunks of magnesium... and some mercury.
Mercury used to be pretty easy to buy, but it ain't no more, at least not here in Australia. You can still get it from scientific-supply places, but it seems that even "technical-grade" mercury, not the super-pure reagent-grade stuff, is quite excitingly expensive, so that's out.
Since there are no strange houses nearby from which I can siphon the DSL-conducting life-blood, my Hg source thus far has been old sphygmomanometers. I started buying 'em, on eBay, because Theodore Gray said you could get "at least a cubic inch" of the stuff out of one of them.
Well, that's a load of bollocks.
I've cannibalised three meters now, with another one waiting for my next urge to do strange mad-scientist things on the deck out the back, and have collected a total of a bit less than a cubic inch, thank you very much Mister Expert.
(If any of you readers out there have a more satisfying quantity of mercury, more than a thermometer but less than a bucket, and are wondering what to do with it, do e-mail me and we'll work something out. If you're not in Australia, though, you're probably going to have to swim here with the stuff in a little barrel around your neck. Shipping mercury internationally is not easy.)
The first meter I bought was, however, not very well packaged.
Part of the fun of mercury is the freaky way it behaves. Its density is 13.5 grams per cubic centimetre - 1.19 times that of lead. It's also very "non-wetting" - it sticks to itself much better than it sticks to most surfaces. So when mercury gets moving, it tends to keep moving. It's got lots of momentum, and it forms droplets rather than sheeting out flat and using up its energy in friction and turbulence.
When mercury hits a hard surface, it doesn't so much splash as (apparently) shatter into droplets of various sizes. The droplets, as the cleanup guides say, can travel "surprising distances", and the smallest ones are really, really small.
Mercury sphygmomanometers aren't very well sealed, so blithely sending one through the post is not a good idea unless you put it in a Ziploc bag or similar airtight packaging. I didn't know to ask for that the first time, and I didn't think to check when I put the thing on the kitchen counter, where it stayed amid a Pile O' Stuff for a while.
When I picked it up later, tiny droplets went everywhere.
Hence, the cleanup job. Only about a large-thermometer-worth of mercury escaped, but that's enough to create tiny silvery jewels (a flashlight is an essential tool when cleaning up mercury) all over the darn place.
So I squeegee-d, and I eyedroppered, and then I got down to the really fun part.
You can, you see, pick up really little mercury droplets with sticky tape.
So I did that, for a while.
You can also render mercury harmless by mixing it with powdered sulfur to re-ore-ify it. I happen to have a little jar of sulfur (I suppose it counts as part of the element collection too), so I used some of it to get the twinkly bits hiding in the floorboards. The reaction isn't what you'd call a whiz-bang one, but if there's sulfur powder on top of the mercury it ought to catch all of the vapour quite nicely.
You can never get all of the droplets if you spill mercury in any ordinary home or lab, but that doesn't matter unless you're one of the people that're terrified of super-small doses of the stuff. Mercury's dangerous because it vaporises so easily, but that also means it'll clean itself up, given time. Tiny droplets evaporate all by themselves, and most houses are leaky enough that a scattering of teeny mercury droplets won't do anything significant to the air mercury concentration.
It doesn't hurt to open a few windows as well, of course.
I'm wondering if you have encountered the BioMeridian device, FDA Regulation Number 882,1540, manufactured in Draper, Utah? A number of "practitioners" are using it to identify food allergies and vitamin/nutritional deficiencies and making great claims. I can only find the FDA registration, which is for a "Galvanic Skin Response Measurement Device", but I can't find anything explaining its accepted use (other than galvanic response, which is good for what? Detecting lies?).
I hate to see folks bilked out of a lot of money to have this test done and directed to take a slew of vitamins that may not be necessary, or healthy, for them.
I haven't seen a BioMeridian device myself, but the QuackWatch page on these things notes that BioMeridian are the company that, some years ago, sold the Listen System by which I had the pleasure of being tested, also some years ago.
BioMeridian have polished their product line up since then, though. Very Star Trek.
They're proud of their FDA 510(k) approval, which they have of course gained on the same basis that various other electrodiagnostic quack companies have gained it, because their devices are substantially equivalent to previous devices which also have never been proven to work, but are safe.
They say they adhere to FDA Approved Good Manufacturing Procedures. You can, of course, do that and still make a device that does not actually work, as long as it's not working as designed. But as we'll see, that doesn't stop them using these sorts of claims as... well, as pretty much all of the evidence in their favour.
It's like ISO 9001 certification. That doesn't prove that your procedures work, only that you've got a snowstorm of documentation that describes what you do (or claim to do) to make them and otherwise run your operation, and proves (for suitably small values of "prove") that you're doing it.
Ooh, but that page also says that BioMeridian products have been tested by a third party, allegedly at great expense!
These tests have established that samples of their devices complied with Annex IV of the Medical Devices Directive 93/42/EEC in 1999.
I wonder what that might mean (he asked, batting his eyelashes adorably).
You can get that Directive here: PDF. If you can be bothered doing so, you'll see that Annex IV is all about making sure devices "conform to the type described in the EC type-examination certificate". Look at BioMeridian's "testing Documentation" (PDF), and you'll see that this certification is all about... safety. Primarily electrical.
OK, guys, we get it. It doesn't electrocute people, screw up TV reception or catch fire, and it comes with an instruction manual. Great.
And they've got the CE Mark, too! Wow! Does that mean the things work? Absolutely not - it's another electrical and safety certification!
Next up in the certificate roll call is, you guessed it, ISO 9001. Are we starting to see a trend, here?
And they've got a Quality Assurance manual! Well, of course they have; it's hard to get ISO-9001-certified without one. Still doesn't say a thing about whether the things you're making so allegedly consistently actually work, of course.
And there's a Customer Service department, too! Since not many people who don't already believe the products work are likely to be calling that line, I think BioMeridian aren't going to be hitting a lot of feedback that challenges their world view.
But wait, I hear you say, what about their Allergy Study?
That was actually published (in the august, though now defunct, American Journal of Acupuncture; this study doesn't seem to be in Medline, though it indexes that journal), but although this study's getting on for 22 years old now, its discovery that these gadgets are good for diagnosing food allergies has, amazingly, not been replicated. Regrettably, many other great discoveries have languished in obscurity as a result of the same terrible problem.
Only a churl would suggest that the experimenters' belief in the technology, imperfect blinding (kinda hard to stop the tested people from knowing what they're allergic to...), and/or the teeny-tiny 30-subject test group, or perhaps plain old incompetence, might have something to do with the production of a spurious result. Not to mention the significant disagreement between the two conventional test techniques that're in the study along with the electrodiagnostic doodad.
The absence of any form of statistical analysis of the results is a daringly unusual feature of this test. It would be most discourteous to suggest that the experimenters could not, in fact, tell a significance level from a hole in the wall.
Moving on - what about the Icon Study? An unblinded, unrandomised test (and proud of it!) showed that something made most of the employees of a company happier and healthier over almost two years of electrodiagnostification, together with dietary supplements and... well, who knows what else?
That "who knows" part, and the detection of simple fraud, are why people do proper tests. This "study" could be used as a jumping-off point to look for real evidence - that sort of thing happens a lot, when someone says "that's funny..." and then sets out to see whether what appears to be happening actually is.
In itself, though, a funny-looking situation without any experimental method probably doesn't tell you anything at all.
Of course, the well-known fact that the management of a large company has never been known to go off on weird tangents as a result of peculiar beliefs about the world does give the Icon "study" more weight.
On to the "CAM Study". Which is not really a study, but more of a list of referenced factoids starting with a reference to a fairly famous JAMA study from 1998, but who's counting. What it's telling us is that Americans spend a lot of money on Complementary and Alternative Medicine! So, obviously, they're spending it well!
(Incidentally, the JAMA study, which has been very widely reported, categorises things like "massage" and "self-help groups" as "CAM". Several other studies have done the same. So things aren't as bad, I mean wonderful, as they seem; some large percentage of people may be using one or more of the listed modalities, but there are more people sitting in plastic chairs and talking about nutrition, or getting their sore back rubbed, than there are enjoying the attentions of iridologists and colonic irrigators.)
BioMeridian have plenty of other references, some more relevant than the CAM Study, on their Education page; dig this PDF for an example of the kind of gibberish (with underlining and stuff in the margins!) that passes for "education" in the world of "Electrodermal Screening".
Feel free to check out the similarly edifying BioMeridian Technology page as well, and alert me if they come up with any repeatable results from studies published in real journals.
I do like how, on this page, they point out that whatever it is that their stuff does, it doesn't provide a "medical diagnosis". No, it just recommends "herbal, homeopathic, and nutritional"... "products". Not medicines! No! "Products"!
And Biomeridian gear can do tests for "allergies and chemical exposure". No matter what you may previously have been told, that too is utterly unrelated to medicine.
Bear in mind that simply establishing that there's something, anything special about those "meridian points" that BioMeridian keep talking about would be enough to win BioMeridian, or any single acupuncturist, James Randi's million dollars. There are many schools of acupuncture and electroacupuncture, and they disagree about where the "meridians" are supposed to be, and how many of them there are. Similar disagreement in conventional medicine over the location and number of, say, kidneys, or toe bones, is difficult to find.
I see no reason not to presume that anybody using BioMeridian machines for their intended purpose is either actionably negligent in establishing whether they are actually doing what they think they're doing, or engaged in simple fraud.
Good luck getting the sons of bitches convicted, though.
Usually, just one corpse is not nearly enough.