Wednesday, December 9, 2009

Polycyclic Aromatic Hydrocarbons in Smokeless Tobacco: Much Ado About Snuffing


In August, I wrote about a report by a prohibitionist University of Minnesota group claiming that one pinch of smokeless tobacco is five times more dangerous than smoking. They reported that smokeless tobacco products contain PAHs, a complex family of chemicals that are formed during the incomplete burning of almost any product, including coal, oil, gas, wood or other organic matter.

The researchers, led by Irina Stepanov and Stephen Hecht, pitched these claims to national media at a press conference sponsored by the American Chemical Society. But they refused to share their results with other scientists, pleading that “we are in process of publishing a full paper on this study.”

That paper has just been published, and it has been pitched again to national media. Again, the headlines scream “high levels of toxins,” and Stepanov et al. portray the results in almost apocalyptic terms: “Our findings render PAHs one of the most prevalent groups of carcinogens in smokeless tobacco. Urgent measures are required from the U.S. tobacco industry to modify manufacturing processes so that the levels of these toxicants and carcinogens in U.S. moist snuff are greatly reduced.”

Are the PAHs really toxins? Are the levels really “high”? Are emergency measures really indicated?

Unproven toxins

Stepanov analyzed smokeless tobacco samples from the U.S. market for 23 PAHs, and the manuscript implies that all of these are proven cancer-causing chemicals (carcinogens). Not even close.

I compared Stepanov’s list with a database maintained by the authoritative International Agency for Research on Cancer. IARC has reviewed several hundred chemicals for cancer-causation, and it has classified them on the strength of the scientific evidence as (Group 1) Carcinogenic to humans, (Group 2a) Probably carcinogenic, (Group 2b) Possibly carcinogenic, and (Group 3) Not classifiable. Groups 2a and 2b are largely based on animal or laboratory studies with no direct human evidence. IARC is a recognized authority, but it is inclined to label chemicals as human carcinogens if there is even a shred of evidence.

Readers will be amazed at how many IARC Group 1 carcinogens are on Stepanov’s list: Exactly ONE. Her list included one chemical IARC calls “probable”, while eight agents are in the still less threatening category of “possible”. More importantly, twelve (over half!) are NOT EVEN CLASSIFIED. (I couldn't find the remaining chemical in any IARC list.)

Low, Low Levels

Stepanov not only misrepresented the number of carcinogens. She implied that levels were dangerously high, stating that “moist snuff can be considered an important source of human exposure to PAHs.” Her data contains virtually no evidence for this statement.

In order to discuss Stepanov’s data, I’ll have to give a short primer or review of the metric system weights. Let’s start with a reference weight 1 kilogram, which is about 2.2 pounds.

1 kilogram = 1000 grams
1 gram = 1000 milligrams
1 milligram = 1000 micrograms
1 microgram = 1000 nanograms

Stepanov reported PAH concentrations in nanograms per gram of (dry) tobacco. A nanogram is one trillionth of a kilogram, or one billionth of a gram, so we are talking about vanishingly small amounts.

A can of moist snuff contains about 30 grams of tobacco, but about 50% of this weight is water, so the effective tobacco content is 15 grams.

The most important PAH on Stepanov’s list is benzo(a)pyrene (BaP), the only IARC Group 1 carcinogen. In Stepanov’s ideal world, no one would consume BaP; but we live in the real world, in which a fundamental tenet of toxicology is: the dose determines the poison. Despite Stepanov’s rhetoric, the level of BaP is very low in her tests, at an average of 56 nanograms per gram. That means that a one-can per day user is potentially exposed to 840 nanograms of BaP per day.

What does this level mean? Fortunately, the federal Environmental Protection Agency (EPA) has information that provides perspective. It says that “BaP is commonly found … in cigarette smoke, in grilled and broiled foods, and … in some water sources.” According to a study published in the Journal of Agricultural and Food Chemistry, a person eating a char-broiled quarter-pound hamburger consumes about 400 nanograms of BaP. The EPA says that water can have 780 nanograms of BaP per gallon and still be acceptable. This means that someone who eats a quarter-pound burger and drinks a half-gallon of water may consume about as much BaP as a can-per-day moist snuff user.

The EPA gives guidance for consumption of 5 other PAHs on Stepanov’s list, which are available in this document from the CDC. In fact, it gives specific consumption levels of these chemicals which are “not likely to cause any harmful health effects.” Using this information, I have estimated the number of cans of moist snuff that can be safely consumed daily by a 110-pound man with respect to these 5 PAHs:





















PAHStepanov’s Concentration (nanograms per gram)Number of Cans Per Day That Can Be Safely Consumed
Anthracene8441185
Acenaphthene1051905
Fluoranthene140095
Fluorene827161
Pyrene129078


I have been very conservative in developing these estimates. First, Stepanov used special solvents to get maximal extraction of PAHs from the samples; I have assumed that smokeless users consume those amounts, which is probably a serious overestimate. It is well known that PAHs are not very soluble in water, so it is likely that actual PAH consumption by smokeless users is much lower than the level reported by Stepanov.

One final comment: the EPA has estimated that on average, a person in the U.S. is exposed to about 3 milligrams of all PAHs every day. That’s 3 million nanograms. A person consuming a can of moist snuff per day is exposed to 174,000 nanograms, which is less than 6% of normal daily exposure.

No tobacco product is perfectly safe, but there is very little epidemiologic evidence that using smokeless tobacco, including American moist snuff, is associated with any cancer. Despite this, Stepanov and colleagues chemically characterize these products as major causes of cancer. Their rhetoric is not sustained by their data.

1 comment:

Carl V Phillips said...

Excellent post, Brad. We are including a reference to it in our upcoming paper that is partially about how the University of Minnesota researchers' unethical behavior tricked the mainstream press (which shares the blame) into reporting grossly false claims about ST.

In case any readers might have missed the understatement, I would like to reiterate your point about extraction. Though they pretend that their results are about human health, the Minnesota chemists extracted most all of the chemicals in question that are in the ST, while actual human consumption of the product would not do so -- probably not even close. Indeed, this brings up the point that if they were actually interested in providing useful health information, rather than just advancing their political agenda, it would have been much more useful to assess how much was actually absorbed. This would be valuable in informing the world about whether swallowable products, like the new innovations by RJR, might have health risks that standard snuff use does not.

Even though they chose to not do the most useful study, it turns out that they accidentally still provided a (partial) useful answer to that with respect to BaP etc.: Since the total amount of the chemical in the product is so low as to not be a substantial health threat, then even swallowing it and completely absorbing it all does not pose a substantial health threat from this class of chemicals. This is good news about novel THR products.

Finally, I think it is worth reiterating that this analysis makes clear that the unethical nature of this group's message to the public has redoubled: Not only were their health and policy claims grossly inappropriate, since their study had no human health science or policy analysis content and they apparently have little understanding of either, but their claims actually flatly contradict their study results.

I wish I could honestly say that I would have expected better behavior.

--Carl Phillips