Secondhand smoke and cancer: Where's the proof"

CMAJ 1998;159:441-2 See response from: C. Gray In her desire to refute a Daily Telegraph article that claimed secondhand smoke was noncarcinogenic, Charlotte Gray resorted to hyperbole in her own article, "Secondhand-smoke story goes up in flames" (CMAJ 1998;158[9]:1178-80). Some of the inferences that Victoria MacDonald drew from a World Health Organization (WHO) study were unjustified, but to say that she has "no understanding of scientific practices" appears manifestly untrue, if the subsequent rebuttal that appeared in the Mar. 15, 1998, issue of the Sunday Telegraph is any indication.

Gray describes the WHO study as a small, run-of-the-mill study involving exposure to secondhand smoke that consisted mostly of tobacco-lobby spin and a lot of egregious mistakes. To our knowledge, the study has not been published. How does Gray know about the egregious mistakes" Has she checked the statistics" [On Mar. 9, the WHO released the following statement: "In February 1998, in accordance with usual scientific practice, a paper reporting the main study results was sent to a reputable scientific journal for consideration and peer review. That is why the full report is not yet publicly available. Under the circumstances, however, the authors have agreed to make an abstract available to the media." WHO also said that the media — the Daily Telegraph — had "completely misrepresented" the study and its results.—Ed.]

Our impression of the WHO and its published studies and statements is that they are carefully considered and contain sound science, although they sometimes lapse into "bureaucratese." In the study, the relative risk (RR) of a nonsmoker who lives in a house with a smoker was given as 1.16, with confidence intervals of 0.093 to 1.44. Thus, it is entirely possible that the RR would be less than that expected — below 1.0 — but it also could be appreciably higher than 1.16. This suggests that the figure is not statistically significant, but the statistical and epidemiologic fraternities prefer to give their results with the appropriate confidence intervals rather than tests of significance.

Gray also quotes Neil Collishaw of WHO's Tobacco or Health Unit, who points out that a major meta-analysis of 40 studies of passive smoking in lung cancer was published in the British Medical Journal in 1997. Unfortunately, positive studies are much more likely to be included in meta-analyses than negative ones. Meta-analyses need to concern themselves not only with published studies but also with other studies that for one reason or another have not been published. Much more importantly, meta-analyses need to review and check the raw data of all published investigations to ascertain whether the data have been analysed appropriately or manipulated to support a particular point of view.

In the hospital where we work, we see 5 or 6 new lung cancer patients each week, or about 250 a year. Yet over the past 20 years or more we have seen only 3 definite cases of primary lung cancer in life-long nonsmokers. It also needs to be emphasized that many smokers, especially those seeking compensation for work-related conditions, are economical with the truth when it comes to their smoking habits. One study indicated that 25% of the smoking histories obtained from subjects exposed to asbestos who were dying of lung cancer were completely incorrect.1 Many of the men denied smoking when applying for benefits, but an about-turn took place once histories were taken from relatives after the subjects died. It is highly probable that some such alleged nonsmokers are included in most epidemiologic studies.

We loathe and detest tobacco companies for their evasion, lies and attempts to trick adolescents and others into taking up smoking. However, the rejection of truth and the acceptance of unproven hypotheses to further one's concept of ethics or social justice is wrong too. Many studies involving secondhand smoke are not convincing, and answers about whether it causes lung cancer are far from established. Unfortunately, it has become customary to torture the data until they confess. We need more science, less hyperbole and less enthusiasm for unproven points of view. We support regulations banning smoking in airplanes, hospitals and public places, not because secondhand smoke causes lung cancer but because many nonsmokers suffer discomfort as a result of the habit.

Dildar Ahmad, MD
W. Keith Morgan, MD
Chest Diseases Unit
London Health Sciences Centre
London, Ont.

Send a letter to the editor Envoyez une lettre à la rédaction

Reference

1. Berry G, Newhouse ML, Antonis P. Combined effects of asbestos and smoking on mortality from lung cancer and mesothelioma in factory workers. Br J Ind Med 1985;42:12-8.