The Physics Police

The Physics Police

Monday, December 2, 2013

The Maillard Overreaction

The Millard Reaction is a chemical reaction between amino acids and sugar. It's responsible for the "golden brown and delicious" quality of baked foods Alton Brown so admires. It creates a wonderful panoply of chemicals, many of which are very good eats. Some less so, as evidenced by the acrid taste of charred meat.

One of these molecules, called 4-Methylimidazole (4-MeI), was studied in 2007 by the National Toxicity Program. In this study, rats and mice were exposed to very high doses in their food.
There was clear evidence of carcinogenic activity of 4-methylimidazole in male and female B6C3F1 mice based on increased incidences of alveolar/bronchiolar neoplasms.
These results were only seen for mice in the highest dosage groups, 625 μg/g and above.

In 2011, the California Office of Environmental Health Hazard Assessment (OEHHA) set the "No Significant Risk Level" (NSRL) of 4-MeI 29 μg per day. In accordance with Proposition 65, a big scary cancer warning must be printed on consumables that exceed the NSRL.

This posed a problem for Coke and Pepsi, which both use caramel color. These sodas contains more than the NSRL level of 4-MeI per bottle. Both companies had to change their ingredients or risk having a cancer warning slapped on the label.

It isn't clear to me how OEHHA arrived at this ridiculously conservative NSRL for 4-MeI. Thanks to Dr. James Coughlan via Chris Hamble for this calculation demonstrating its absurdity:
... the amount of 12-ounce cans a woman would have to drink, a day, for the rest of her life, in order to get to the carcinogenic levels of 4-Mel would top 37,000 cans. And for men, a whopping 95,000 cans.
That's an extrapolation based on the National Toxicity Program study, accounting for concentration and body weight. A 12-ounce can of soda is 335 grams. If that can contains 29 μg of 4-MeI, than the NSRL concentration for a beverage is 0.0866 μg/g. That's 86.6 parts per billion, if you like.

In reality, the FDA, European Food Safety Authority, and Health Canada all consider caramel color safe. This is because the dose of 4-MeI you get from the food additive equivocal to that you ingest from roasted foods, baked goods, grilled meats, etc.

It's even in coffee! Lojková et al. looked for 4-MeI in several off-the-shelf coffee products.

4-MeI ranged from 0.35 to 0.77 μg/g in roasted coffee from around the world.

That's between 4.04 and 9.98 times the NSRL concentration! So, does that mean...?

Yes, coffee is toxic sludge.

Anyway, Coke and Pepsi are both transitioning to a new, low 4-MeI formulation of caramel coloring.

In July, a watchdog group called the Center for Environmental Health conducted tests on Pepsi and Coke from all over the country. They found bottles sold outside California still contained traditional caramel color. These results were published July 3rd, accompanied by an hysterical press release and online petition demanding that Pepsi "clean up" the ingredients ahead of schedule.

The Associated Press ran a story based on this press release, which was quickly picked up by junk news outlets. I call them junk because they distort their readers impression of a story with misleading headlines. One such headline reads US group finds ‘worrying’ carcinogen levels in Pepsi. Nobody ever said the word "worrying" except the anonymous poster of this story. Nevertheless, this misleading headline appears on 1,720 web pages!

Junk syndication like this is how several people I know get their news. If you ask me, this does far more damage than drinking a goddamn Pepsi.

If you take away one lesson from all this, let it be a healthy distrust of headlines, especially from junk news outlets. They're incessantly deceptive.

And as always, be wary of black and white health claims made in absence of dosage. Some things give rats cancer in high doses. That doesn't mean they will necessarily harm humans in low doses.

We can be glad OEHHA is doing its best to protect our health without believing alarmist nonsense.

1 comment:

  1. In case anyone doubts Dr. James Coughlan's math, I've replicated it below.

    The average daily dose fed to mice showing statistically significant increased risk for cancer was 80 mg 4-MeI per kg body weight.

    I will use the average 0.504 mg 4-MeI per kg coffee from the table above by Lojková et al. and an assumed human weighing 75kg.

    The equivalent dose of coffee, PER DAY, is (75 kg) * (80 mg/kg) / (0.505 mg/kg) = 11,881 kg or 50,218 US cups.

    I dare you to drink that in a lifetime.