© 2011 Roy Benaroch, MD
Oh noes, yet another foodstuff to worry about!
The Center for Science in the Public Interest has begun a press campaign to inform the public that commonly-used “caramel coloring” contains cancer-causing chemicals. It’s made it into newspapers and websites under cleverly alliterative headlines like “Caramel Coloring in Cola Could Cause Cancer.” The CSPI is urging the FDA to prohibit its use. The list of chemicals that are killing us and our children seems to be growing every day. But is this really something we need to worry about?
Two chemicals are being implicated as the cancer-causing contaminent: 2-methylimidazole and 4-methylimidazole (4-MI). These chemical are created during the production of caramel coloring, which apparently involves steeping sugar in ammonia and other tasty chemicals. Modern production methods don’t seem to resemble grandma stirring a pot of sugar as it browns—and I agree with the CSPI’s objection to the name “caramel coloring,” which makes the stuff sound more wholesome than it is.
But it’s not just the name that the CSPI is criticizing. Their press release directly implicates caramel color as a carcinogen:
“The ‘caramel coloring’ used in Coca-Cola, Pepsi, and other foods is contaminated with two cancer-causing chemicals and should be banned, according to a regulatory petition filed today by the Center for Science in the Public Interest.”
Their conclusions stem from research studies, such as this one by Chan and colleagues (reference #15 in the CSPI petition), which they claim shows that these chemicals cause cancer in lab animals. However, the study does not justify their conclusion; ironically, the study might show tantalizing evidence that caramel coloring protects rodents from at least on kind of cancer.
Chan’s study is typical of toxicology research. Standard lab animals (in this case, 200 rats and mice) are fed standard rodent-chow with added amounts of the chemical in question. This study followed the animals and measured daily consumption of 4-MI for 2 years. Just how much caramel coloring did they consume? There was a range of values, with some animals eating far more than others; overall consumption was designed to fall in a range of about 30-250 mg/kg day. Using the CSPI’s estimate of up to 130 micrograms of 4-MI in a 12 oz can of cola, and estimating an average teenager weighs 110 pounds, I work out that the range of exposures is similar to a teenager drinking from 11,538 – 96, 153 cans of cola a day. Every day. For two years.
I am certain that this much soda would kill a teenager, most likely by drowning. The numbers look ridiculous, so I’ll post my math at the end.
OK, let’s just say, for the sake of argument, that our hypothetical teenager really is drinking this much soda. He has a big test to study for, or something like that. What did the study show happened to the exposed rats and mice? Though in some cases, certain cancers were occurred more frequently in the more-exposed rodents (some lung cancers, for instance), in other cases the rates of cancer didn’t show a consistent dose-response relationship (thyroid lesions in female but not male mice). There was also a complex gamish of other microscopic tissue changes at a variety of inconsistent dosing levels, seen sometimes in only one kind or gender of animal. Also, 4-MI exposure seemed to protect against possibly pre-cancerous changes in breast tissue at every exposure level. The overall mortality among exposed and control animals was the same.
Basically, there was no consistent pattern of much of anything; some animals got cancer, some did not, and their tremendous exposures to 4-MI didn’t consistently correlate with health.
Don’t misunderstand me: there’s plenty wrong with drinking soda, especially in children. Loads of sugar and acid to rot teeth, and all of those unnecessary calories contribute to obesity. So I’m not saying here that soda is good for you. I agree with the CSPI that less soda consumption can improve health. What I disagree with is the breathless worry, based on barely conclusive animal studies on phenomenally large exposures, that 4-MI is a proximal cause of cancer. The science doesn’t support that, and we ought to be more truthful with people. There’s plenty enough real worry out there, and crying wolf isn’t helping anyone raise healthier kids.
The lowest range of exposure was for male rats getting 30 mg/kg/day of 4-MI. A 12 oz can of cola contains 130 micrograms = 0.13 mg of 4-MI. Divide that by an estimated teenager’s weight of 50 mg, you get an exposure of 0.0026 mg/kg/day. 30/.0026 = 11,538 – so a teenager, to reach the lowest end of 4-MI exposure in the Chan study, needs to drink 11,538 cans of cola every day for 2 years. That would be about 2060 2 liter bottles, or about 100 of those 5 ½ gallon drums, or if you were to fill up the interior of my car (total interior capacity 14.7 cubic feet), it would be drinking ten entire Honda Accord’s worth of soda every day. Isn’t math fun?