12.9 Specific carcinogens and cardiovascular toxicants in Australian cigarettes

As was noted in section 12.3, it has long been assumed for practical purposes that the composition of "tar" does not differ between brands in ways that produce significant differences in smokers' disease risks. In other words, it has been assumed that probable variations in the constituents of "tar" amount to a zero sum game. However, in recent years there has been increased interest in determining the extent to which different cigarette brands are likely to produce differing exposures to specific carcinogens and cardiovascular/ respiratory toxicants. Following findings that emissions of some harmful smoke constituents vary markedly between brands,1,2 even after controlling for variation in tar and nicotine yields, there has been renewed interest in the possibility that cigarettes could be made less harmful through selective removal of specific known carcinogens and toxicants.2,3 One possible mechanism for achieving this would be to set emission limits for specific harmful smoke constituents, using emissions measures that take compensatory smoking into account.4

In 1999, the Australian Government planned to require annual disclosures of the emissions of a group of 40 carcinogens and cardiovascular toxicants for a representative sample of Australian brands, following the lead set by the Province of British Columbia in Canada. The Voluntary Agreement for Disclosure of Ingredients, which eventuated in 2000, did not include recurrent emissions disclosures.5 However, the industry did agree to a once-off disclosure for a sample of top-selling brands. This occurred in 2001 and included fifteen brands: six brands each for Philip Morris and British American Tobacco Australia and three brands for Imperial Tobacco Australia. This once-off disclosure provides most of what is known about the specific emissions of Australian cigarettes. An analysis of the Australian emissions data in tandem with the British Columbia emissions data from the same year reveals a number of by-manufacturer and by-country differences in emissions.4

During lobbying against the proposed annual emissions disclosures, the Managing Director of Imperial Tobacco Australia, Nick Cannar, claimed:

There is considerable evidence in the published scientific literature that the composition of smoke is essentially consistent, i.e. the proportion of each emission per milligram of tar is essentially fixed. In other words, individual brand differences between products of the same tobacco style, (i.e. flue-cured, as in Australia) would have a minimal, if not undetectable, impact on emission data.6

As Cannar would have it, flue-cured or Virginia cigarettes may have different emissions patterns to, say, US-blended cigarettes but, within any group of flue-cured cigarettes, there will be no differences in emissions that do not relate directly to tar yields. It would then follow that either tar and nicotine yields are consequential for smokers' exposures after all or, alternatively, all Virginia cigarette smokers with comparable nicotine intakes gain essentially the same exposures to specific harmful smoke constituents, regardless of the specific brand they smoke.

King, Borland and Fowles 4 attempted to test the veracity of Cannar's claims. In order to provide the best correction of the data for compensatory smoking and thus the most realistic comparisons possible between the emissions of "regular", "mild" and "ultra mild" brands, the analyses were conducted using emissions per milligram of nicotine, rather than emissions per cigarette (as the data was presented in the disclosures). King, Borland and Fowles 4 also focussed on the emissions data determined using the ISO intensive condition yield test. This test uses 55ml puffs, taken once every 30 seconds, with the filter ventilation taped over. While this test has the same pitfall as the standard ISO yield test in that it does not take compensatory smoking into account, the size of the puff at the burning cone and the interval between puffs is closer to the actual smoking behaviours of most smokers than the parameters of the standard ISO yield test. By using this test and making the adjustment for nicotine delivery, a reasonable approximation of relative deliveries of specific smoke constituents can be obtained.

Table 12.9.1 presents Australia–Canada comparisons for 13 selected smoke constituents under the ISO intensive testing condition. The Australia–Canada differences in mean levels were statistically significant for 10 of these 13 agents. On these figures, Cannar's claim6 would appear to be seriously wide of the mark. Further, while there were no marked by-manufacturer differences for the Canadian brands (all of which used tobacco grown in southern Ontario), there were a number of by-manufacturer differences for the Australian brands. For instance, the British American Tobacco brands had higher adjusted emissions of tobacco specific nitrosamines (NNK and NNN) than the Philip Morris and Imperial brands. However, the Philip Morris and Imperial brands had higher emissions of benzo{a}pyrene (BaP) than the British American Tobacco brands. This may be explained by British American Tobacco brands having higher levels of nitrates in the tobacco, as nitrates "scrub" BaP but react with nicotine and related substances (collectively know as nicotine alkaloids) under combustion to produce tobacco specific nitrosamines.

Table 12.9.1
Mean adjusted emissions of 13 selected agents under intensive ISO testing conditions for Australian brands (n=15) and Canadian brands (n=21) tested in 2001




4-aminobiphenyl ng/mg



Benzoapyrene ng/mg



Acetaldehyde ug/mg



Acrolein ug/mg



Hydrogen cyanide ug/mg



Cadmium ng/mg



NNN ng/mg



NNK ng/mg



Quinolene ng/mg



M+p-cresol ug/mg



1,3-butadiene ug/mg



Acrylonitrile ug/mg



Benzene ug/mg



NNN = n-nitrosonornicotine

NNK = 4-(N-nitrosomethylamino)-1-(3-pyridyl)-1-butanone

* Difference in means between Australian and Canadian brands statistically significant.

Differences in the way manufacturers produce their cigarettes thus appear to lead to higher emissions of some agents and lower emissions of others. However, King Borland and Fowles did not find evidence of cigarettes that were likely to have low levels of harmful smoke constituents across the board. King, Borland and Fowles found no brands with below average emissions for all 13 smoke constituents. Thus no brands could be singled out as probably less harmful, even assuming that the measured smoke constituents were the most important ones.4 Further, given the once-off nature of the Australian disclosure, it is not possible to know the degree to which the emissions of Australian brands vary from year to year.

Future research should make clearer the extent to which levels of toxins/ carcinogens in cigarette smoke vary independently of each other and the extent to which they are interrelated (so that reducing the level of one toxin/ carcinogen will necessarily result in another increasing). That will strengthen the evidence base for regulatory proposals which aim to reduce the harmful emissions of cigarettes.7,8

12.9.1 Proposal for testing/regulation

[content in development] 

Relevant news and research

For recent news items and research on this topic, click here. ( Last updated May 2021)


1. Gray N, Zaridze D, Robertson C, Krivosheeva L, Sigacheva N, Boyle P, et al. Variation within global cigarette brands in tar, nicotine, and certain nitrosamines: analytic study. Tobacco Control 2000;9(3):351. Available from: http://tobaccocontrol.bmj.com/cgi/content/full/9/3/351

2. Ashley D, Johnson DR, McCraw JM, Richter P, Pirkle JL, Pechacek TF, et al. Tobacco-specific nitrosamines in U.S. brand and non-U.S. brand cigarettes. Nicotine and Tobacco Research 2003;5(3):323-32.

3. Gray N and Boyle P. The case of the disappearing nitrosamines: a potentially global phenomenon. Tobacco Control 2004;13(1):13-16. Available from: http://tobaccocontrol.bmj.com/content/13/1/13.full.pdf

4. King B, Borland R and Fowles J. Mainstream smoke emissions of Australian and Canadian cigarettes. Nicotine & Tobacco Research 2007;9(8):835–44. Available from: http://www.informaworld.com/smpp/content~db=all?content=10.1080/14622200701485109

5. Commonwealth Department of Health and Ageing. Voluntary agreement for the disclosure of the ingredients of cigarettes. 16 June 2004, Commonwealth Department of Health and Ageing, 2000. Available from: http://www.health.gov.au/pubhlth/strateg/drugs/tobacco/agreement.pdf

6. Cannar N. Draft voluntary agreement for ingredient disclosure Letter to S. Kerr, Dept. of Health. Sydney: Imperial Tobacco Australia Limited 2000, viewed 7 Dec 2007. Available from: http://legacy.library.ucsf.edu/tid/mjq56c00

7. Gray N. The modern cigarette, an unregulated disaster. Medical Journal Australia 2007;187(9):502-3. Available from: http://www.ncbi.nlm.nih.gov/pubmed/17979612

8. World Health Organization. The scientific basis for product regulation. Technical Report Series, 945. Geneva: World Health Organisation, 2007. Available from: http://www.who.int/tobacco/global_interaction/tobreg/who_tsr.pdf