Wednesday 2 March 2011

Passive smoking and breast cancer

From the Beeb:

Passive smoking 'raises breast cancer risk'

Er, no. The study does not say that. This is what it says:

Our data suggest that extensive exposure to passive smoking may increase breast cancer risk. However, since risk of breast cancer was restricted to the most extensive passive smoking category with no clear dose response, the association with passive smoking should be considered suggestive only and needs confirmation from other studies.

The statistical association between passive smoking and breast cancer in this study is actually about as suggestive as a burka. The associations are all over the place and none of them achieve statistical significance. The main finding, based on a fairly substantial 1,515 cases, show that nonsmokers exposed to secondhand smoke have a nonsignificant relative risk of 1.08. Has epidemiology really got to the point where anything other than a relative risk of exactly 1.00 is considered "suggestive"? (Chance alone dictates that RRs of 1.00 are rare, but epidemiologists don't believe in chance when they want to make a name for themselves.)

The fact of the matter is that there has never been any good evidence that active smoking causes breast cancer, which makes an association with passive smoking somewhat unlikely, to say the least. Nothing has been more closely studied in the last sixty years than smoking and disease. Since breast cancer is the most prevalent form of cancer in the world, if there was a real association with smoking, it would have been spotted long before now.

It hasn't been for want of trying. Anti-smoking campaigners have always been very keen to find an association because the high incidence of breast cancer would allow them to create even more stratospheric estimates of how many lives are lost to smoking/passive smoking than already exist. But, as I wrote in Velvet Glove, Iron Fist, they have always failed to do so.

A 1994 paper published in the British Journal of Cancer found no link between breast cancer and first or secondhand smoke, nor did a massive assessment of 53 studies that encompassed 55,515 breast cancer patients in the British Journal of Cancer. The Centers for Disease Control, the American Cancer Society, the IARC, the Australian Medical Journal, the British Medical Journal and the US Surgeon General all agreed that there was no link. Geographical and historical spread of cigarette consumption showed no correlation between breast cancer prevalence and smoking, and while lung cancer rates in women began rising in the US from the mid-1960s, breast cancer rates were unaffected by the post-war surge in female smoking.

Velvet Glove, Iron Fist; p. 237

The authors of this new study even acknowledge the lack of evidence for the smoking/breast cancer hypothesis:

...systematic reviews of epidemiological studies published as of 2002 concluded that there was no overall association between active smoking and breast cancer risk, and attributed conflicting results of individual studies in part to the confounding effects of alcohol.

But add that:

However, recent reappraisals of evidence from recent cohort studies have suggested an increased risk of breast cancer that is independent of the effects of alcohol

In other words, decades of research failed to show anything until a few studies were "reappraised" in the era of anything-goes epidemiology in which eager epidemiologists and fevered advocacy groups leap on every finding, no matter how weak. We are now invited to ignore a mountain of evidence built up over decades in favour of single digit relative risks which hover around 1.00, fail to reach significance and fail to display a dose-response relationship. What a state this field of science is in.

If, fifty years ago, you told the scientists who developed the techniques of epidemiology that you thought "passive smoking raises breast cancer risk" because you'd found a statistically insignificant elevated risk of 8% they'd punch you in the face.

That is all.

8 comments:

Brian Bond said...

If, fifty years ago, you told the scientists who developed the techniques of epidemiology that you thought "passive smoking raises breast cancer risk" because you'd found a statistically insignificant elevated risk of 8% they'd punch you in the face."

Quite so! In fact I think we should take a lead on this, on behalf of those founding fathers of epidemiology. We should seek out the authors of fantasies such as this 'study', and punch them in the face!

I would be more than happy to take the first shot.

You are right about the statistics in this 'study'. Practically none of their confidence intervals managed to creep even into the outer fringe of the 95% statistically significant range.

On this evidence there is just one, and only one valid conclusion that can be drawn:

"The null hypothesis stands."

or, in terms that the thickies at the beeb can understand..

"There is no relationship between passive smoking and breast cancer"

Any other conclusion is a lie.

Just add it to the list!

Anonymous said...

From Dave Atherton

This Meta analysis by Sir Richard Peto et al finds no correlation between passive smoking and breast cancer. It even goes on suggest that misclassification and exaggeration of exposure to cigarette smoke was the main reason other studies may have found a correlation. The highlights include:

"Methods In the large prospective study, 224 917 never smokers who completed a questionnaire that asked women whether their parents had smoked and if their current partner smoked were followed up for an average of 3.5 years for incident breast cancer

"Results In the prospective study, 2518 incident invasive breast cancers occurred during follow-up and the adjusted relative risk of breast cancer for passive exposure either as a child or as an adult vs neither exposure was 0.98 (95% CI 0.88–1.09); results were similarly null for childhood exposure (0.98, 0.88–1.08) and adult exposure (1.02, 0.89–1.16) separately. We identified seven other studies with prospectively recorded exposure data; when results of all eight studies were combined (including 5743 never smokers with breast cancer), the aggregate relative risk was 0.99 (0.93–1.05) for any passive exposure. The aggregate findings differed substantially (P = 0.0002) between these 8 studies and 17 other studies with retrospectively recorded information (including 5696 never smokers with breast cancer).

Conclusions Aggregate results from studies with prospectively reported information show that the incidence of breast cancer is similar in women who did and did not report passive exposure to tobacco smoke either as a child or as an adult. The aggregate findings from the retrospective studies may have been distorted by some women becoming more likely to report past exposures because they knew that they had breast cancer."

http://ije.oxfordjournals.org/content/37/5/1069.abstract

Ann W. said...

I was wondering if they would use the Canadian study to back up their own work and there it was...

"However, a Canadian panel of experts reviewed the extensive new research in this area and concluded that the relations between active smoking and breast cancer, both premenopausal and postmenopausal, are consistent with causality, based on the weight of evidence from epidemiological and toxicological studies and on an understanding of biological mechanisms.

Expert Panel on Tobacco Smoke and Breast Cancer Risk
"Based primarily on the evidence presented by the CalEPA and the Surgeon General, and strong recent evidence of an active smoking-breast cancer risk, the Panel concluded that the relationship between SHS and breast cancer in younger, primarily premenopausal women is consistent with causality. The evidence was considered insufficient to pass judgement on SHS and postmenopausal breast cancer."
http://www.otru.org/pdf/special/expert_panel_tobacco_breast_cancer.pdf

Anonymous said...

There must be lots of retired epidemiologists. I wonder why they are not fuming mad about the distortion of their 'science'?

On the other hand, perhaps they are, but have no voice. Even if they had a 'letter to the Times' published, the probability is that the propagandists would simply ignore it.

The probability is that the unauthorised import of tobacco will become greater and greater until a point is reached where the Gov receive very little income from tobacco duties. I have two daughters who do not smoke. They go on holiday abroad regularly. The potential for them to legally import tobacco has net even been scratched yet. Having said that, I do expect that Tobacco Control in the EU will find a way to reverse the free market in respect of tobacco - to be followed in due course by a similar reversal in respect of alcohol.

But, is it not true that 'make your own' is gaining strength? It must surely be so since going to the pub is no longer a cheerful option. If you are forced to stay at home, you might as well brew your own booze and grow your own tobacco.

It is a matter of simple economics. What idiotic, short term wankers our politicians are?

Anonymous said...

Junican, in the case of Peto, when asked for quotes at te time of the smoking ban, all he said was that he hoped it would result in fewer people smoking - nothing about passive smoking, which, in his evidence to the H of Lords, he did not consider much of a risk. Peto is actually Sir Richard. What are knighthoods for?

Anonymous said...

The study does not adequately control for alcohol consumption, which other studies have found has a dose related relationship with breast cancer. 7 or more drinks a week is not further differentiated.

Anonymous said...

"" Studies have shown that eaters of peanuts have a 0.97 relative risk of cancer. Conclusion: peanuts protect against cancer""

Can we see the logical fault in the above (hypothetical) scenario?

We can, but the fault is not always clear. The point is that, if one keeps looking for connections, one will sooner or later find them. The above argument could equally apply to people who live at the seaside, or people who jog every day. But then the 'scientists' will say that these studies have been repeated in different places and the same results have been found. The reality, however, that the same results have not been found. The SHS studies are a case in point, as these breast cancer studies seem to be also.

Chris Oakley said...

Note the author of the BBC article.