March 20, 2012

Stanton A. Glantz, PhD

Philip Morris USA joins Philip Morris International in responding to our paper on Project Mix

We must have hit a really raw nerve at Philip Morris with our December 2011 paper, "The toxic effects of cigarette additives. Philip Morris' Project Mix Reconsidered: An Analysis of Documents Released through Litigation," that showed how Philip Morris used tricky normalizations and underpowered studies to try an argue that additives did not increase cigarette toxicity when, as we said in the paper, "The case study of Project MIX shows tobacco industry scientific research on the use of cigarette additives cannot be taken at face value. The results demonstrate that toxins in cigarette smoke increase substantially when additives are put in cigarettes, including the level of [total particulate matter]."

Philip Morris International responded in January and Philip Morris USA responded in March, both trying to talk their way around the central issues we raised.

You can read their comments on the PLoS Medicine  website (links above) together with our responses.  Here is the latest response, to Philip Morris USA:

Like Philip Morris International’s  response to our paper on Project Mix, Philip Morris’ USA’s response sidesteps the two most important findings in our paper: (1) After it found that the levels of many toxins in the smoke from cigarettes with additives increased (and TPM, generally by even more), Philip Morris changed the Project MIX protocol to normalize by total particulate matter (TPM), thereby obscuring the absolute increases; and (2) The animal toxicology studies were seriously underpowered, making it unlikely that it would detect statistically significant changes in biological effects due to the additives.

Philip Morris USA simply ignores the first finding, that the use of a post-hoc normalization by TPM obscures the absolute increases in many toxins associated with putting the additives in the cigarettes.

In response to the fact that the Project Mix studies were seriously underpowered, Philip Morris USA says that “We used classical toxicological evaluation assays and internationally recognized methods (including group size recommendations) [emphasis added].”  They go on to state that “we recently published on the power of these assays [used in Project Mix] to discriminate effects of ingredients added to cigarettes.”  As Philip Morris notes in reference 12 (a paper by Oldham himself) in its comment, the power of a study depends on the size of the effect one seeks to detect, the confidence with which one wishes to be able to detect that effect, and the underlying variation in the response.  These are all factors which depend on the specific situation, so one cannot rely on “group size recommendations” developed in general.  Indeed, the power calculations presented  in reference 12  demonstrate a very wide range of minimum detectable effects (reflecting corresponding wide variations in power) for the Project Mix assays.  More important, reference 12 shows that many of the assays had very low power to detect the kind of small changes in toxicity that would be important when exposing tens of millions of people to a toxin.  The minimum detectable differences (with 80% power) for many of the endpoints reported in reference 12 exceeded 20%, with many exceeding 50% or even 100%.  Even a 5% increase in toxicity would have devastating population impacts on the 40 million smokers in the United States.

Indeed, given the need to demonstrate a negative result (i.e., no increase in toxicity due to the additives), one could argue that studies should be large enough to achieve 95% power to detect modest effects (similar to the 95% confidence often used when drawing positive conclusions).

Philip Morris USA cites several studies that “reached the same conclusions” (references 13-32 in their comment) as Project Mix.  Every one of these studies was conducted by people working for a tobacco company (Philip Morris/Altria: references 13, 21-37; Lorillard Tobacco: 14, 15; RJ Reynolds: 16-18; British American Tobacco: 19; Japan Tobacco International: 20).

Philip Morris USA cites a paper by long-time industry consultant Peter Lee (reference 37) and  coauthored with EB Sanders from Philip Morris and RA Carchmann, Vice President, Scientific Affairs at Philip Morris Incorporated from 1997-1999,  as evidence that “recent human epidemiological evidence … indicates there is no direct evidence suggesting that smoke from cigarettes with ingredients is any more hazardous than smoke from cigarettes without.”  (This paper did not indicate any funding source.)  The first problem with this statement is that reference 37 is not a study of additives, but rather a study of flue-cured versus blended cigarettes on chronic obstructive pulmonary disease.  While reference 37 states that flue-cured cigarettes use fewer additives than blended cigarettes (page 405, the flue-cured cigarettes are not additive free. In addition, there are many other differences between these two types of cigarettes than just the additives that are used.  Moreover, as demonstrated by our analysis of the Project Mix toxicological studies shows, the additives as associated with many toxicological changes beyond those that lead to chronic obstructive pulmonary disease.

It is also important to consider the source of reference 37.  Peter Lee has an established history of reaching conclusions that tobacco products and secondhand smoke do not increase disease risk.  In one particularly well-documented case, he worked under the direction of industry lawyers and executives to publish a paper designed to contest the first large Japanese prospective epidemiological study (Hirayama, 1981) linking secondhand smoke with lung cancer (Hong and Bero, 2002; Yano, 2005).  The tobacco companies, working with the Covington and Burling law firm, wanted to contest the accuracy of the measure of exposure to secondhand smoke in the Hirayama study as a way to undermine its conclusion that secondhand smoke caused lung cancer in nonsmoking women married to smokers.  In the original plan,  the paper was to be authored by Japanese investigators, Eiji Yano and Jun Kagawa, with Lee serving as a consultant rather than an author.  When the Japanese investigators refused to put their names on the paper the tobacco industry drafted, they were replaced by Lee.  Later, when learning that Lee had published the results, Yano (2005) published an analysis showing that the measures of exposure to secondhand smoke were accurate.  Given this history, one cannot rely on reference 37 as independent evidence that additives do not increase the toxicity of cigarettes.

Philip Morris USA noted that the experiments were conducted in accordance with good laboratory practices.  We did not identify any problems with the actual conduct of the experiments.  Quite the contrary, we assumed that their measurements were unbiased when we concluded in our paper that the appropriate interpretation of the Project MIX results is that “many of the toxins in cigarette smoke increase substantially when additives are put in cigarettes, including the level of TPM, and that, assuming that the toxicological results from Project MIX represent unbiased estimates of the true biological effects, these data show many adverse biological consequences (and that the failure to reach statistical significance was the result of underpowered studies rather than lack of an effect).”

Nothing in  Philip Morris’ USA’s letter changes our conclusion that, “In particular, regulatory authorities, including the FDA and similar agencies elsewhere who are implementing FCTC articles 9–11, could use the Project MIX data to eliminate the use of these 333 additives (including menthol, which is the major component of ingredient group 3) in cigarettes.”

Indeed, like the exchange with Philip Morris International, Philip Morris’ USA’s refusal to forthrightly engage the substantive issues raised in our paper reinforces our recommendation that scientists and regulators cannot take at face value statements regarding the safety of its products offered by Philip Morris (or other tobacco companies).

Marcia Wertz, PhD RN
Thomas Kyriss, MD
Suman Paranjape, PhD
Stanton A. Glantz, PhD


Hirayama T (1981) Non-smoking wives of heavy smokers have a higher risk of lung cancer:  A study from Japan.  BMJ 282:183-5.

Hong MK, Bero LA (2002) How the tobacco industry responded to an influential study of the health effects of secondhand smoke. BMJ 325: 1413-1416.

Yano E (2005) Japanese spousal smoking study revisited: how a tobacco industry funded paper reached erroneous conclusions. Tob Control 14: 227-233; discussion 233-225.

Add new comment

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.