Tuesday
Apr072009
The politics of secondhand smoke
Tuesday, April 7, 2009 
I have been reading the comments (101 so far) on last Friday's post. I have nothing to add except to refer readers, especially our friend Rollo Tommasi, to the 2005 Forest report Prejudice and Propaganda: The Truth About Passive Smoking.
The 50pp document, which features a lively introduction by Lord Harris of High Cross, our late chairman, is divided into three sections: Questions and Answers about ETS, Epidemiology and ETS, and Examining the Evidence on ETS.
Nothing that has been said or written since prompts me to change my mind that the dangers of secondhand smoke have been greatly exaggerated for reasons that have far more to do with politics than health.
You can download the full report HERE.
Reader Comments (156)
Rollo,
"I’m not sure whether your question refers to SCOTH looking again at evidence into the risks of active smoking or passive smoking." - Rollo.
I was speculating about whether SCOTH has sufficient objectivity to evaluate the oncogenic effects of HPV on active smokers and passive smokers in the context of cervical cancer and lung cancer.
"On active smoking, the Klein et al study is very interesting. I think the killer message is the last sentence of the abstract: “The data suggest that HPV is the second most important cause of lung cancer after cigarette smoking and strongly argues for additional research on this issue.”" - Rollo.
I agree that the Klein et al study is very interesting. However, I think the "killer message" is that prior to this very recent research, HPV was not revealed to be a highly plausible oncogenic factor in lung cancers.
In the case of cervical cancer there is already evidence to suggest that an estimated RR of ~1.19 (for passive smoking => lung cancer) is swamped by at least one confounder, that is
explained by differences in human behaviour and not explained by the theory that tobacco smoke has oncogenic properties. This study (J.Peto) "(and cigarette smoking, with strong evidence for a dose- response (OR for current smoking habit 20+ per day = 2.57; 95% Cl = 1.49-4.45)" is well above stated risks of passive smoking (lung cancer) and easily with in the arms of
of active smoking (R.Peto). What I am saying is, that cervical cancer data shows that people who smoke are at higher risk of (cervical) cancer but that smoking does not cause this cancer (HPV does, 99.7% of cases). This confounder could be one of many confounders that may equally apply
to lung cancer. Given that HPV is common to both cancers - it would be a really good idea if SCOTH would point out these facts, as a matter of public interest, now, this year, not later.
http://www.springerlink.com/content/q311672081186pk1/
Contrary to what Rollo has been saying, a RR of 1.38 in a meta-analysis is still considered insignificant and not high enough to determine a causal effect. And 1.38 is higher than the SHS RR. Looks like Rollo was wrong that a low RR is actually high in a meta-analysis.
And:
Strength of association and dose-response
There is an important distinction to be made between improving the precision of the relative risk estimate—which meta-analysis offers—and interpreting the causal relevance of the absolute value of the summarized estimate—which it does not.11,18,20,38 Simply put, the practitioner of causal inference, having completed a meta-analysis, remains faced with the problem of judging whether the summarized estimate—small (i.e. weak) or large (i.e. strong)—can be explained by confounding or bias. Thus it is reasonable to accept a relative risk estimate of 2.0 emerging from a meta-analysis as a better estimate than that which we may have opined through a judicious application of our ‘judgement’, somehow estimating a summary average value from a long list of single study estimates. Nevertheless, that same conclusion, emerging as it does from a meta-analysis, provides no additional warrant for a causal claim. Put another way, a summary estimate of 2.0—whether it emerges from a meta-analysis or not—remains on the borderline of what is typically called a ‘weak’ association. The same argument applies to dose-response curves,39 which can also be made more precise by judicious application of meta-analysis. Nevertheless, whatever the form of those curves—increasing, decreasing, ‘S’ or ‘J’, shaped—the practitioner of causal inference is left with the task of explaining them in terms of the biological sense of the revealed pattern, potential biases, and other concerns.
http://ije.oxfordjournals.org/cgi/content/full/29/3/387
From the same source:
"These conclusions imply that practitioners of meta-analysis should refrain from making causal claims without first considering the issues brought out by applying the largely qualitative causal inference methods"
and
"Indeed, the scope of the problem of making causal claims from scientific evidence is widening. Recently, some have argued that an understanding of the determinants and distribution of disease will be improved by examining not only epidemiological and biological evidence but also social evidence for causation.40–42"
I have long said that social factors are hugely important when looking at smoking studies, because smoking has never been an isolated factor. The majority of cigarette smokers are in the lower classes, thus tend to eat worse, exercise less, have less healthcare, be more stressed and so forth. Without properly accounting for such variables, smoking studies are at best to be treated with caution and at worst to be thrown away.
Bottom line: Rollo, a RR of 1.2-1.3 in a meta-analysis, no matter how much you squirm and protest, is still statistically insignificant and not the proof you tout it as.
My god Rollo, what a sad person you must be. We should feel sorry for you because obviously you don't have a life do you? I'd like to know what you actually do to enjoy yourself. For God's sake get a life.
Fredrik
You seem to be saying that, because almost all women with cervical cancer show at least one form of HPV, that is the only cause of “cervical cancer”. In particular, you say “cervical cancer data shows that people who smoke are at higher risk of (cervical) cancer but that smoking does not cause this cancer”. I totally disagree. While almost all women with cervical cancer are HPV+, most women who are HPV+ do not contract cervical cancer. There are numerous risk factors for cervical cancer – e.g. diet, family history, multiple pregnancies – all of which can be a cause of cervical cancer for women infected with HPV. Smoking is recognised as a cause of cervical cancer – smokers infected with HPV are much more likely than non-smokers with HPV to contract cervical cancer and scientists believe they know why (by significantly increasing the likelihood of dysplasia (abnormalities on tissue cells)).
There is nothing in the link between lung cancer and HPV which undermines the status of other risk factors for lung cancer (whether it is smoking, passive smoking, exposure to radon gas, air pollution, family history, etc). Like I said before, I would be interested in seeing research into the relationship between HPV infection, exposure to SHS and lung cancer, just to be sure. But I see nothing in the link with HPV which gives any reason to question the role that exposure to SHS plays in contracting lung cancer.
And of course, HPV does not even begin to explain all the different risks to health which are increased by active and/or passive smoking (e.g. certain other cancers, heart disease, stroke, bronchitis, emphysema).
Rich: You seem to understand neither my comments nor the commentary you cite.
I did not say “that a low RR is actually high in a meta-analysis”. I was saying that something can be a major public health risk, even if the risk to an individual is low, where many people can potentially be affected. In fact, one of the links referred to in the second article you cite says my point exactly:
“In these situations the risks involved are generally small, but once a large proportion of the population is exposed, the potential public health implications of these associations—if they are causal—can be striking.” (Egger ,Schneider & Davey Smith).
If you look properly at the EMF meta-analysis, you will find that they would have been happy to reach a firm conclusion on the basis on the relative risk they reached. However, scientific findings require more than a RR. In this case, the researchers could not reach a consistent conclusion about whether and how EMF fields risk health. Their conclusion? Not that ENF fields present no risks. Just that the issue requires further exploration. In other words, absolutely NOT the conclusion you claim the meta-analysis reached.
Nor does the second article you cite say anything which states that a meta-analysis’ results are meaningless (and “statistically significant”? – Where did you get that from).
Looks like your bottom line has been drawn on shifting sands. Because I can’t find it.
Sylvia - Perhaps you'd like to direct your remarks to those people on this board who spend a LOT more of their time participating in these sorts of boards than I do???