The following article was published in Food
Focus
(a publication of the New Zealand Food Safety Authority)
May 2005
Jay D Mann
Please feel free to print this but be sure to
acknowledge Food Focus as the source. It has been slightly modified
for reading as a web page. You can download a complete .pdf version
of the original from http://www.nzfsa.govt.nz/publications/food-
focus/2005-02/ffmay05.pdf
Imagine that some madman has injected cyanide into just one
orange in a bin containing ten thousand oranges. Anyone eating
these fruits will have one chance in ten thousand of receiving
serious damage or death. This is clearly an unacceptable
situation.
Suppose, however, that the oranges were turned into juice,
enough for ten thousand glasses. Each person drinking that juice
will receive one ten-thousandth of a lethal dose of cyanide.
Newspaper headlines and TV bulletins proclaim “Poison Found
on NZ Breakfast Tables”.
This imaginary tale illustrates how easy it is to confuse
distributed risk with all-or-nothing risk. No matter how many
people drink the cyanide-containing juice, none will die. At worst,
each person will be fractionally less healthy. Using
straightforward mathematics of dubious validity, we can say that
death is one hundred per cent ill health. So each person getting a
ten-thousandth share of the cyanide-containing juice will be
perhaps 0.01% less healthy. (In fact, I'll show later in this
article that no one at all will be injured!)
Language and logic are not always good partners. We don't have
words that readily distinguish between risks that are distributed
as compared to risks that are all-or-nothing. It is all too easy
for some misguided person to complain that “the government is
allowing the continued sale of orange juice that could result in
the death of ten New Zealanders per every hundred
thousand.”
Our language also lets us speak or write about “a low dose
of a toxin”, even though this is really a meaningless phrase.
At a low enough dose a supposed 'toxin' is simply not toxic. It is
time we stopped worrying about imaginary risks.
Descriptive adjectives such as 'toxic', 'sweet', 'bitter', and
'yellow' are designed for application to the ordinary effects of
chemicals (both natural and man-made). Grammar makes it easy to
turn these characteristics into nouns: 'toxin', 'sweetness',
'bitterness', and 'yellowness'. It sounds, on the face of it, as
though a 'poison' is inevitably 'poisonous'.
It is difficult to think logically about 'toxins' and 'poisons'.
Think, instead, about 'sweetness'. You can run a real or a 'thought
experiment'. Pick a sweet chemical, such as sucrose (cane sugar),
saccharine, or aspartame, and dissolve enough in water to make an
obviously sweet solution. Then make a series of serial dilutions,
taking something like a teaspoon of solution mixed into a fresh cup
of water. In my tests, by the third or fourth dilution, few people
can detect any difference between my test solution versus tap
water. With further dilution, no one at all will be able to detect
a difference: you have reached a biological threshold where
biochemical 'noise' in the taste buds masks any sweet-tasting
action.
Some people may be much more sensitive than others. Researchers
have reported that for certain taints, there can be a hundred-
thousand-fold range between least and most sensitive tasters.
Nevertheless, even the most sensitive individual will find there
are sucrose/saccharine/aspartame levels below which he can detect
nothing.
Even in our most dilute solutions, there is a good chance that a
skilled chemist with sufficiently expensive equipment, could detect
low levels of erstwhile sweetening agent. In any case, having made
the dilutions ourselves, we know that our sample solutions contain
calculable concentrations of 'sweetener'. Yet in these low
concentrations, the chemicals, which new media might describe as
“linked to sweetness” are no longer sweet! 'Sweetness'
is, therefore, a biological action that occurs only in response to
sufficiently high concentrations. (I feel ashamed to be reminding
my audience of this elementary fact. My excuse is that if we
replace 'sweet' and 'sweetness' with 'toxic' and 'toxin', exactly
the same situation holds, but our emotive reaction to the thought
of 'poison' tends to block rational thought.)
Our bodies are crude, rugged, and adapted to imperfection.
Is it possible that our metabolism is so magnificently balanced,
so precisely engineered, that long-term exposure to even a trace of
a harmful chemical, no matter how low a dose, can mess up the
system? This concept of such Swiss-watch perfection has evaporated
under the scrutiny of modern biochemistry. We might have guessed
that a 7physical design such that people continue to function
despite the loss of half their lungs and kidneys, most of their
liver, and major portions of their brain, is rugged. Complexity is
not an accident: the futurologist Freeman Dyson commented
“Life by its very nature is resistant to simplification,
whether on the level of single cells or ecological systems or human
societies.” (“Infinite in all Directions”, 1988,
Pelican Books) Simplicity and efficiency might appeal to management
consultants, but man-made organisations designed for simplicity,
clarity and efficiency tend to be brittle and unable to survive new
challenges.
My training, a long time ago, was as a biochemist. I'm
constantly amazed by the newest discoveries in my erstwhile field.
My instructors emphasised the high degree of specificity in
enzymes, their lack of error, but nowadays, it is quite clear that
most enzymes are only moderately specific, and mistakes can happen.
In fact, by the time you multiply the number of cells in our body
by the number of enzyme pathways in each cell, mistakes are
inevitable. What keeps us going (most of the time), is “a
complicated web of molecular structure” (Dyson) with
different routes for obtaining the same results and with numerous
error-checking procedures. For instance, the DNA that carries our
vital genetic information is constantly under attack by random
chemical reactions. Dr Bruce Ames and his coworkers have shown that
every day, in every cell of our body, there are about thirty
thousand 'hits' on DNA, mostly from free radicals that are formed
in normal metabolism. (Not surprisingly, some of the free radical
formation is part of a mechanism for killing invading bacteria, but
much is just slipshod errors by oxidative enzymes.) There are,
however, at least four different quality- control inspection
systems, acting independently, that try to restore the damaged DNA,
or failing that, to prevent it being used. DNA replication in the
test tube has one error per hundred, but double- checking and
error-correcting within the living cell bring it down to, say one
error per 100 million.
Another example of overlapping functions: In the lab, individual
genes can be inactivated by 'anti-sense' genes. To the surprise of
the scientists who did this, in about one third of cases there was
no effect of a missing gene product. This doesn't mean the gene's
function wasn't important, but rather it confirms the existence of
other pathways ready to step in to accomplish the same result.
I have gone into a little bit of detail to show how metabolic
complexity is built in to our cells, making them surprisingly tough
and rugged. If a gene can be totally eliminated without obvious
damage, why should we expect any damage from a hypothetical
inhibition of an enzyme by a fraction of a percent? To quote
Freeman Dyson again, “error tolerance ... is inherent in life
from its earliest beginnings” .
I promised early in this editorial to reveal the truth about
cyanide. There is no doubt that this chemical is, even in legal
terms, a 'poison'. Doses of about 50 mg can be lethal, because high
doses of cyanide block the transport of oxygen within the body. On
the other hand, cyanide is also a natural product, found as a
protective chemical in apple pips, seeds and leaves of apricots and
cherries, cassava roots, cabbage, mustard, and Lima beans, to name
just a few examples. (Tobacco smoke has cyanide too.) Because
cyanide is so wide spread, we should not be surprised to learn that
detoxifying enzymes such as rhodanase are found in our tissues.
Rhodanase converts cyanide to the less toxic chemical
thiocyanate, which is excreted in the urine. However thiocyanate
antagonises the uptake and utilisation of iodine, which is
essential to the proper functioning of the thyroid gland. People
with inadequate levels of iodine, for any reason, develop goitre
(the bulging necks seen in paintings of medieval peasants).
Children born to iodine- deficient mothers are at risk of brain
damage, ranging from a few per cent all the way to full-blown
cretinism. In parts of Africa where cyanide-containing cassava is
consumed every day, goitre and cretinism are major health
problems.
What happened to our "dangerous" orange juice?
Our hypothetical cyanide-containing orange juice, on the other
hand, will be ingested by New Zealanders with reasonably adequate
levels of iodine intake. Since cyanide-wielding madmen are not
everyday occurrences here, there is little chance that
cyanide-enriched orange juice will become a regular component of
our diet. After all, New Zealanders are already consuming modest
levels of naturally occurring cyanide from plant-based foods, and
minor sources like the (imaginary) orange juice are
unimportant.
What started out as a fable about a seemingly dangerous risk of
cyanide poisoning, has now turned out to be a trivial intake of
biologically unimportant levels of a chemical by a large number of
people. Although cyanide is a natural chemical, that is not
important. Low enough concentrations of any chemical, no matter
what its origin, will have zero effect. We have more important
things to worry about.
A final comment
If there is one take home lesson for readers, it is that any
time you read the phrase “a low level of toxin” or
“linked to cancer/death/illness”, don't panic. Use of
such phrases is evidence that the speaker or writer simply doesn't
understand what they are talking about.
