Log in

No account? Create an account


Frank Davis

Banging on about the Smoking Ban

Previous Entry Share Next Entry
Chris Snowdon returns. (I've made 3 or 4 minor corrections, Chris. Let me know if anything needs to be changed.)

To try and maintain some sort of structure to this debate, I’m going to use each post to respond to one major issue raised by the defence, then present a piece of evidence for the prosecution, and then respond to various questions, factual errors and logical fallacies at the bottom. To give you notice, in the next three posts, I plan to write about:

  1. ‘Only 10% of smokers get lung cancer’

  2. ‘Confounding factors could explain the association’

  3. ‘The evidence is mere statistics and proves nothing’

So, first up is the idea that because “only” around 10% of smokers develop lung cancer, smoking does not really ‘cause’ lung cancer at all. Firstly, I’ll address the point that some people have raised that goes along the lines of “even if smoking does cause lung cancer, 10% isn’t much of a risk and is therefore not worth worrying about”. Even the normally rational Leg-Iron said something similar on his blog the other day:

“It seems sensible to assume that anything unnatural taken into the body will carry some risk. Then again, the risk is demonstrably small, because 90% of smokers don't get lung cancer. That's better odds than crossing the road.”

I don’t know what the roads are like 'round Leg-Iron’s way, but statistically a smoker is much more likely to be killed by lung cancer than he is crossing the road. Clearly 1 in 10 people do not get killed crossing the road.

In fact, compared to almost any form of mortality, 1 in 10 is a high risk (and it could be 1 in 8 or 1 in 5 if you’re a very heavy smoker). Breast cancer incidence for women is a little higher (13% or so) but mortality is a lot lower. The closest thing would be heart/circulatory disease, a term that covers a multitude of sins, including what many of us would call ‘dying of old age’ and CVD risk is, of course, also increased by smoking. One of the reasons the lung cancer rate isn’t higher is that so many smokers get taken off by heart disease, cancer of the larynx, emphysema, etc.

However, if that’s the risk you want to take, I’ll be the last person to stop you. Just don’t take a 10% mortality rate as evidence that smoking doesn’t cause lung cancer. No one ever said it always causes lung cancer, just as no one ever said that lung cancer is always caused by smoking. Both human beings and human diseases are more complex that that. It comes back to the defence demanding an impossible level of proof—a level they never ask for when saying that other things cause lung cancer.

There are very few risk factors (ie. causes) that kill everybody who comes into contact with them, and there are few risk factors that are the exclusive cause of one type of disease, or death.

We can say, for example, that drink-driving is a cause of excess mortality. We can say, quite literally, that drink-driving kills. Reasonable people do not take this to mean that it always kills or that nothing else kills. Fewer than 10% of drink-drivers get killed by the act, fewer than 10% of deaths are caused by drink-driving and most people who get killed on the roads don't get killed by drink-drivers. Nonetheless, we recognise that it kills because of the aggregate statistics that the defence is so dismissive of.

Or to take the famous case of the Bow Street Pump in which Dr Snow identified a cholera infected water pump as the cause of a cholera outbreak by looking at where the victims lived. In what is often cited as an early piece of epidemiology, the close correlation between proximity to the pump and developing cholera helped Snow work out that the water was the cause (read here if you’re not familiar). Not everyone who drank from the water got cholera and people got cholera without drinking from that pump. People did not understand the biological reasons why they were getting sick from the water, but Snow’s observation was proof enough to identify the problem and get the pump shut down.

Just as the Victorians did not understand the exact workings of cholera, so we do not understand the exact workings of cancer. Until we do, who gets it and who avoids it will seem partly the result of luck. Factors currently unknown to us are at work for sure, and Rich urges us to keep an open mind. And of course, scientists do keep an open mind, but unlike Rich they don’t reject what we do know in the mean time, just because our knowledge is not perfect in the here and now. If John Snow kept the pump open in 1854 because ‘it might be something else’ (which was true) or ‘we don’t know everything’ (which is always true), he would have been rejecting the known, embracing ignorance and endorsing the Nirvana fallacy which I mentioned last time (and which Rich misunderstood from the very first paragraph. Perhaps I didn’t make it clear enough that I wasn’t talking about ETS.)

There’s no justification or precedent in science to demand that everyone, or a majority, or even 2% of people must develop a disease as a result of exposure before we cry causation. We do not expect every fat person to develop diabetes, expect every alcoholic to get cirrhosis or expect every Victorian street-dweller to get cholera before we say that we have proof of causation. The statistic that really matters is the chance a nonsmoker has of getting lung cancer, which is well under 1%. And it is the difference between the incidence within the case and control group that tells us what is going on. These are, let's remind ourselves, not 'mere statistics'. They are real people dying of real diseases for real reasons.

The Kaiser Permanente Experiment

As much as I’d love to talk about Doll and Hill’s 1950 study until the cows come home, I thought it would be nice to look at some of the hundreds of subsequent studies that found the same thing; most of which addressed the criticisms directed at Doll/Hill (let’s say it again - Doll/Hill was preliminary study). To make this a fair fight, I haven’t bothered to do any more than skim this, so if it’s terribly flawed, we’ll all find out together.

But the gist of it is: Cohort study. 60,000 subjects aged 35 of older monitored between 1979 and 1987. Relative risks varied a lot due to race (as I was saying the other day), but the female RR for smoking and lung cancer was 8.5 (less than a pack a day) and 13.9 (pack or more a day). For men, it was 4.7 and 10.4. Ever, never and former smokers were all categorised, dose-response was evident and tables are shown for all. This is the one for male current smokers:

Kaiser (click on picture for larger image)

Questions, errors, fallacies etc.

Air pollution

Chris touts the global statistics proudly, but repeatedly omits to mention that 90% of smokers never get lung cancer. As Frank pointed out previously, what if the cause of lung cancer was often something in the external air, that housewives weren't so exposed to?

Then we would expect smokers and nonsmokers alike to be dropping dead of lung cancer at the same rate. We would also expect some canny scientist to have spotted this massive global outdoor carcinogen.

Mouse painting

Was it a little speck, a generous helping, or did they cake the entire inside of their lungs with the stuff? And are you seriously suggesting that dried and 'painted' tar is the same as inhaling and exhaling smoke?

They used between 25mg and 75mg of tar. It was painted on their backs, as I would have thought Rich would know, really. Certainly they didn’t “cake the entire inside of their lungs with the stuff” (!). No matter what the doses were, the critics would say they were too high and that the experiment was unrealistic, which it was, if we are to compare it to human beings' smoking. The object of the painting of cigarette tar on mice was only to show that there were carcinogens in cigarettes. And the scientists did what was asked of them by the critics—ie. produce tumours in animals. The critics then demanded they produce the same results from smoke. Which brings us to...

The beagle experiment (revisited)

On the subject of the tobacco industry secretly accepting the Auerbach study as conclusive, Rich says “because the industry said so, they must be correct?” No, but when the industry that has spent millions investigating the issue, and has every reason to want the study to be flawed, admits that the experiment is valid, then it’s a pretty good bet that it is valid. If their legion of well-paid expert scientists from around the world didn’t see a problem with it then maybe we shouldn’t be so quick to call it garbage at a 40 year remove from behind a computer.

“Also, the dogs didn't develop lung cancer.”

No, as I said in the last post they delivered invasive tumours which is the step before lung cancer, and they developed lots of them. The nonsmoking dogs didn’t develop any.

On the criticism that there weren’t enough control dogs, same answer again, really. Since every one of the controls were absolutely fine and all the cases were dead, dying or in a bad way, we don’t need hundreds of controls to tell us what we already know - that healthy beagles don’t just suddenly get a range of smoking-related diseases for no reason.

Rich then denies ever having said that no animal experiment succeeded in getting animals to develop lung cancer through tobacco smoke, but in his very first post he said:

“To this day, scientists have not managed to induce lung cancer in any animal through tobacco smoke (excluding the F334 rats and A/J mice, specially bred to develop cancer).”

Then Rich denies having complained that “people do not smoke like the beagles in this experiment.” He said:

“I didn't, and explained this in the very first comment.”

But he did. He said: “The methods of this study are not remotely natural, given that no person smokes in those manners.”

Then he weirdly criticises me for saying, as an aside, that “by all accounts, the dogs quite enjoyed the smoking and would wag their tails!”

“By all accounts is a pretty big stretch of the truth - in a report in a medical journal by the experimenters, it was claimed some dogs were so addicted to the nicotine they jumped into the box and wagged their tails. Not all the dogs, and not by all accounts.”

“By all accounts” is a turn of phrase for Christ’s sake! The tail-wagging was in the study and Kluger reports it in his account of the story, as have other people who have written about it. Apparently the tail-wagging element being written “in a report in a medical journal by the experimenters” is not sufficient proof that it happened.

The reason for this focus of trivial linguistic tics is to avoid discussing whether the dogs were stressed. Stress is the perfect excuse for the critics because it can’t be measured very reliably, it’s vague in meaning and whatever effect it has on health is impossible to quantify. Rich tries to have his cake and eat it by suggesting that stress causes lung cancer (“Who's to say that the cancer was not caused by post-operative complications? Or indeed the stress.”) while insisting that he’s not putting forward an alternative explanation (“I didn’t assert the real cause was stress”) because it would be ripped to shreds. But, he says, “it could certainly be a contributing factor”. I’ll come to confounders in my next post, but I didn’t understand why Rich felt so strongly that stress was the explanation in the beagle experiment. But now I understand. He thinks the stress was caused by the tracheotomies:

“Oh, because the non-smoking dogs didn't get sick? True, but then the non-smoking dogs didn't endure tracheoctomies either.”

Oh dear. It’s statements like this that makes me wonder whether Rich has (a) read the study, and (b) understand what a control is. The control dogs absolutely did have tracheotomies and were kept in identical environments and fed the same food as the other dogs. The only difference between the controls and the subjects was that the subjects smoked cigarettes. That’s what makes them controls! Otherwise they’d just be some dogs in a kennel.

From the study: “Ninety-seven pedigreed male beagles were purchased from breeders who stated that the dogs were in good health. They were kept under observation in our laboratory for several weeks to make sure that they were in good health. Tracheostomy was then performed on each dog.... The ninety-seven dogs were assigned to cages by computer-generated random numbers. Eight of the dogs were then selected by random numbers and kept as controls.”

Rich also gets the number of dogs wrong, which also makes me think he hasn’t read the study ( not for the first time).

The Semai miracle

Speaking of which, I’m quite surprised he dared come back to this doozie:

Chris isn't convinced by the Semai evidence, but this is a pattern now. When smoking rates tally lung cancer rates, that's definitive proof; when they show the opposite, it's wrong.

No, Rich, that’s not what it is. The problem with this Semai ‘evidence’ is that there isn’t any. There is no study, no report, no statistics, no evidence. The whole thing comes from a unverified letter written by some Singaporean GP 33 years ago, which you only know about because you read about it in a book written by a notorious crank. The data showing how many people smoke and how many people get lung cancer, on the other hand, is verified, serious evidence taken from around the world over many decades. I find it really quite worrying that anybody could fail to spot the difference between these two sources, let alone treat them as if they had equal weight.


What about the Asians, or the Native Americans, or, hell, the Causacasians that hold the longevity records and all smoked?

What about them? See p. 482 of the Kaiser study to see lung cancer data stratified by race in the same country. Surely this isn’t one of those ‘my granny smoked and she lived to be 102’ arguments? I rather think it is, because it leads onto things like this:

“10% of smokers get lung cancer, that's all we really know.”

And we know that the percentage of nonsmokers who get lung cancer is much, much lower.

“No biological evidence exists, whatsoever, and animals studied have failed.”

Both these statements are, quite simply, lies, and transparent ones at that, considering Rich has just admitted that at least one animal study succeeded.

Rich claimed in a previous post that smoking rates fell to zero during the war. He then said this was in South Africa and gave a source. Since then, he’s admitted that he hasn’t read the source (it’s another cut 'n' paste from Whitby’s book) and it was Holland, not South Africa. Seeing as the smoking rate didn’t fall to zero during the siege of Leningrad, I doubt it fell to zero anywhere else and Rich’s ‘source’ doesn’t make me think otherwise. His point was that smoking fell to zero but the rate of smoking-related diseases didn’t. There seems to be no proof of that, but there is proof (in Proctor’s Nazi War on Smoking, figures given here) that the decline in cigarette consumption in Germany in the 1940s had a clear effect on lung cancer rates compared to other countries.

Rich complains that graphs showing cigarette consumption against lung cancer don’t “take into account such facts as, for instance, US tobacco being more radiaoctive than foreign tobacco, which should be explored.” He has since admitted that there is no evidence for this claim and the source he gave me is comically inept.

When referring to the rates of lung cancer in women going up in line with rates of smoking, some 40 years after the rise in male lung cancer, I said that “women started smoking cigarettes in a big way from the 1920s”. Rich decided to pick up on the turn of phrase and avoided the point:

In a big way, sure. But women were smoking before the 1920s.

So let me be more explicit: Very few women smoked until the 1920s and very few women got lung cancer. By the 1970s, very many women were smoking and very many women were getting lung cancer.

On the wacky idea that the chemicals and toxins in cigarettes “are tiny, less than we inhale in normal air”, I pointed out that all of Rich’s examples involved secondhand smoke, not directly inhaled smoke.

My figures were how much a cigarette contains, yes. Higher, yes, but how much?

It varies greatly depending on the chemical, but 1,000 times higher is the frequently given estimate. Whatever the dilution, it is completely fatuous to equate secondhand smoke with directly inhaled smoke.

On the same topic [chemical constituents of cigarettes], he asks:

Again, as I said at the outset, where are the details?

Find them yourself and get back to us. You’re the one who made the argument. It’s not in the nature of a debate for one person to demand his opponent find all his data for him.

Clearly the CO can't be too bad, because we'd all have carbon monoxide poisoning for instance.

You’re confusing acute health problems with chronic health problems. CO converts haemoglobin to carboxyhaemoglobin. This is not great in the short-term but can be very bad in the long-term as it starves your organs of oxygen and leads to chronic disease.

Logical fallacies

“Chris, your statement seems to read that because each study has some flaws, we must accept them as accurate?”

A massive straw man, there. My point was—and I think this is crystal clear as it was the point of the whole post—that all studies have flaws and we must decide whether the flaws are big enough to invalidate the study’s finding. Since the flaws raised by the defence are petty issues like ‘they didn’t have enough control dogs’ and ‘my mate smoked 105 cigarettes but he calls himself a nonsmoker’, the answer in this case is no.

And speaking of straw men...

“Perhaps the 'smoking causes lung cancer' statement is as true/flawed as 'eating causes heart disease'”

Show me one person who says ‘eating causes heart disease’.

“Phil's point on the expensive brands, well, is it not possible that the cheaper brands use crappier ingredients which are thus more harmful?”

Not really, no, but if it was - and if doctors smoked more expensive, ‘safer’ brands then the doctors should have a lower cancer rate than the rest of society. Therefore, Doll and Hill underestimated how dangerous smoking was. It’s not an argument that helps the defence much though.

And this was in response to me saying that the Auerbach study proved that tobacco smoke can cause malignant tumours in the lungs of dogs:

No more than it was 'proved' red hair 'causes' heart disease.

This being a reference to Scottish people having higher rates of heart disease and also being more likely to have red hair. It’s a nice Epidemiology 101 example of correlation and causation. Doesn’t really work here, partly because the Auerbach study was one of Rich’s beloved lab experiments where correlation and causation are easier to separate, and because the increased risk of being Scottish and having red hair is roughly equivalent to the risk of being Scottish and having heart disease. What you need to explain the lung cancer risk (and I’ve said this about 8 times now) is a factor that is not just correlated with smoking but is as strong or stronger than the association found between lung cancer and smoking. But that’s epidemiology, whereas this is an animal experiment in which no controls got a disease, so Rich picked the worst possible moment to re-use Phil Button’s metaphor.

And the logical fallacies keep on coming:

It was once considered a given that masturbating caused blindness....

Not by doctors and scientists, I think you’ll find, and certainly not by the entire medical establishment.

And here’s another, in reply to me saying that Japanese men have much higher lung cancer rates than Japanese women...

Which would surely suggest the cigarettes are responsible in your argument, correct? In which case, the aforementioned resistance isn't quite so prominent as is being made out.

See what he did there? I said that Asian people had a resistance to lung cancer and he responded as if I said Asian people were immune from lung cancer. Even though I’d already corrected him on this in a previous post. Would you like a light for your straw man, Rich?

Hospital study again

Back to Richard Doll (for a change), I said: "I’m happy to accept the possibility that the [smoking] rate could have been even higher [than 80%] in London, but common sense tells us that it wasn’t 98%"

Why would it need to be?

Because if it was less than 98% then there were a disproportionately large number of smokers in hospital, which ‘anti-smoking fanatics’ like me would suggest is evidence that smoking isn’t too good for you. And since 99.7% of lung cancer patients were smokers, we’d need 99.7% of Londoners to be smokers for there not to be a positive association when measured against the general population.

“Do 40 a day smokers start on 40 a day? Absolutely not. 4 years ago I smoked perhaps 4 a day. The year before that I smoked maybe one a day everyday and half a pack on the weekend.”

Rich, that’s because you’re a young lad who started smoking a few years ago. The people dying of lung cancer in the hospital were not. And, yes, smoking habits do change and that’s why Doll and Hill asked for the average amount smoked over 20 years. As I said in the post, a bit of inaccuracy here or there on this sort of question really is a trivial issue.

I don't see what's difficult about what I said to be honest. I think you're referring to the Men of 25 States study, so i'll repeat what i said in CATCH-6:

No, God, don’t repeat things. I read it. Just give me a source for your claim that the study didn’t take note of smoking status and explain what you mean about comparing people in their 30s to people in their 80s.

P.S. Requested corrections made 1:26am 11 Dec 2010.
Tags: ,

  • 1
At 1.48am i don't have time to respond to all that yet, but I want to clear a few things up.

"On the criticism that there weren’t enough control dogs, same answer again, really. Since every one of the controls were absolutely fine and all the cases were dead, dying or in a bad way, we don’t need hundreds of controls to tell us what we already know - that healthy beagles don’t just suddenly get a range of smoking-related diseases for no reason."

All i said was an equal number. Not too much to ask is it? You did say we can't have 'my nan smoked and lived to 102' so why should a measly 8 dogs count? 8 dogs in the smoking sample were healthy too.

Rich then denies ever having said that no animal experiment succeeded in getting animals to develop lung cancer through tobacco smoke, but in his very first post he said:

“To this day, scientists have not managed to induce lung cancer in any animal through tobacco smoke (excluding the F334 rats and A/J mice, specially bred to develop cancer).”

Did i deny it? Where? I can't remember that, because i still absolutely maintain that it hasn't been done, and that statement, far from being transparent, has been reaffirmed in multiple court cases by both sides. You'd think really if it was such a massive success it would have been repeated to really bring the hammer down, no?

Then Rich denies having complained that “people do not smoke like the beagles in this experiment.” He said:

“I didn't, and explained this in the very first comment.”

But he did. He said: “The methods of this study are not remotely natural, given that no person smokes in those manners.”

Get a grip Chris, i deliberately clarified this in my comments. I said that that particular excerpt was from my book, as a direct response to the SG saying it was 'natural'. I needed to clarify it because that excerpt was not included in my post. I didn't deny the statement, i denied that i was using it as a critique of the study.

Which would surely suggest the cigarettes are responsible in your argument, correct? In which case, the aforementioned resistance isn't quite so prominent as is being made out.

See what he did there? I said that Asian people had a resistance to lung cancer and he responded as if I said Asian people were immune from lung cancer. Even though I’d already corrected him on this in a previous post. Would you like a light for your straw man, Rich?

Chris, wtf? The very quote of mine you used says 'resistance' and no mention of 'immune'. It says, and you copied it, 'resistance isn't quite so prominent as is being made out'.

As for the rest, I'll try to get to it over the weekend. But Chris, seriously, give it a break on the digs and insults because it's completely unnecessary. Either debate sensibly or don't bother. Frank has quite explicitly asked this to be calmed down, and you've amped it up.

Frank has quite explicitly asked this to be calmed down, and you've amped it up.

That's not my view. I don't think Chris has 'amped it up'. I think that he's doing exactly what I asked him to do, and toned it down.

What was bothering me a few days back was that it looked like it was all getting much too personal. And I simply didn't see any call for that. Which is why I intervened to call for restraint.

We're dealing with a super-hot issue here, with smoking and lung cancer. Rather to my surprise, the commentariat here seem to have been coming down on the side of Rich and myself. Which makes for an uphill battle for Chris. Everybody knows what it's like to argue your corner when more or less everybody else is against you. And I think that Chris has done a sterling job with his latest post, without getting seriously nasty. And that's very creditable. And it was all I was asking for.

I think that maybe I'll reply tomorrow (if I can find time from sorting and packing), to try to widen the debate slightly.


(no subject) (Anonymous) Expand
(no subject) (Anonymous) Expand
(no subject) (Anonymous) Expand
(no subject) (Anonymous) Expand
(no subject) (Anonymous) Expand
Smokescreens (Anonymous) Expand
(no subject) (Anonymous) Expand
Just some observations for a non-scientific, non-mathematical reader .....

One of the reasons the lung cancer rate isn’t higher is that so many smokers get taken off by heart disease, cancer of the larynx, emphysema, etc.

Shouldn’t that statement be “One of the reasons the lung cancer rate isn’t higher may be that so many smokers …….” There’s no way of knowing, for example, that a smoker who hadn’t died of one of those illnesses wouldn’t have succumbed instead to something completely unrelated to smoking. There’s certainly no guarantee that if they hadn’t developed heart disease, laryngeal cancer etc that they would automatically have got lung cancer instead.

No one ever said it always causes lung cancer, just as no one ever said that lung cancer is always caused by smoking

But isn’t that the whole point? That when scientists state, categorically, that they have discovered a “cause” for some illness it is because that cause, when applied to a hitherto healthy subject, results in the stated ailment? And, perhaps more importantly, absence of that cause never does? In scientific terms, doesn’t anything which only results in 10% of the subjects falling prey to said ailment indicate something rather more along the lines of an “association” or a “risk factor?” I don't know what the standard is for indicating that a cause is just that - I'm sure it wouldn't be as high as 100%, but I'd be very surprised if it was as low as 10%.

Maybe some scientists out there can confirm this, but I always believed that for scientific “proof” to be claimed, a new and separate group of healthy subjects must be deliberately “made ill” by application of the “cause” after the initial experiments – with control group and the like – had been completed. In fact, wasn’t that one of the stumbling blocks in being able to declare the gonorrhea bacterium the cause of the STD Gonorrhea many years ago? Again, I would have expected that the number which must be "made ill" would need to be pretty high to be regarded as "proof."

We can say, quite literally, that drink-driving kills. Reasonable people do not take this to mean that it always kills or that nothing else kills

Au contraire. The statement that “drink-driving kills” is in a way similar to the “smoking kills” statement in that, as Chris says, drink-driving can kill, but doesn’t always and isn’t the only cause of death on the roads. But, as the anti-smoking movement know full well, “reasonable” people are far outnumbered in the general population than “less reasonable” ones (I am assuming here that Chris is using the word “reasonable” to refer to people who use their reason, rather than people who are fair in their dealings with others). It’s a campaigners’ classic sleight of hand – make a statement which you know will be over-interpreted by the majority of the population, make it often enough, and it’ll enter the public consciousness wholesale and without question and will become an unchallengeable, known "fact." In fact, this is a good example to watch in the future, because I strongly suspect that the "drink-driving kills" will become the new "smoking kills" over the next few years. Watch this space, maybe.

and there's more .....

And of course, scientists do keep an open mind

The trouble is, where smoking is concerned, they no longer do. This is, no doubt, partly due to the sheer power and insidious nature of the anti-smoking movement, which has successfully inveigled itself into every area of life which might have any connection with smoking or smoking research, and wields enough power to prevent any curiously-minded scientist from ever undertaking research which might yield results that run counter to the “smoking causes cancer” mantra, and have sufficient clout to ensure that any such research would in any case never see the light of day in any publication – certainly not without robust counter-allegations by wheeled-in big-guns from the profession and at severe risk to the researcher’s own career.

Closing all avenues against further research into the causes of lung cancer – forcibly, if you like, closing scientists’ usually open minds - and pinning the blame squarely on smoking, and nothing but smoking, thus closing off avenues into research into other causes of lung cancer (because to do so you have to start by admitting that there may be other causes, which gives the anti-smoking movement an immediate problem), may explain much as to why now, some 40-odd years on from the start of the sharp decline in the number of people smoking, we have yet to see a corresponding sharp decline in lung cancer rates. Of course, we’ll all have to wait another ten years to be certain, but early indications so far don’t look great for the “pin all the blame on smoking” theory.

The tail-wagging was in the study

Just for the record, tail-wagging isn’t automatically a sign of joyous anticipation as it’s commonly assumed. It’s a much more general indication of a state of enhanced feeling. Dogs will sometimes wag their tails if they are protecting their territory, jostling for a bite of a kill (certainly not the friendliest of environments within a dog pack), or as a sign of submission and thus appeasement – badly neglected or physically beaten dogs will often wag their tails when violent owners come home in the hope that they can avoid a thrashing. I wouldn’t necessarily expect the Beagle researchers to be experts on dog behaviour, but because they probably weren't, it’s highly likely that the dogs were in fact trying – rather hopelessly – to stop the researchers doing what they habitually did whenever they walked into the room. And, quite frankly, I can't blame them. It sounds horrific!

Stress is the perfect excuse for the critics because it can’t be measured very reliably, it’s vague in meaning and whatever effect it has on health is impossible to quantify

Does this therefore mean that it should be discounted? Research into the effects of stress are in their infancy and, as Chris says, it is hard to measure accurately. But studies do seem to be increasingly showing that again, whilst not in and of itself a hard-and-fast cause of many illnesses, it would seem to be a contributory factor. Probably not dissimilar to smoking itself ……

Just some thoughts.

There are a number of, essentially, side-issues. One of them is animal studies. Although these could be debated no end, they really add nothing to the issue of smoking and LC in humans. It should be noted that there is great difficulty in extrapolating animal studies to human physiology/bio-chemistry: The applicability of animal experimental findings to the human condition is highly arguable.

The combination of animal and human experimentation/epidemiology in both “carcinogenicity” and pharmaceutical analyses is a long-standing protocol that has had its critics over many, many decades. This combination or protocol gives the appearance of a formal, exhaustive study of an issue. However, if animal studies have poor/no transfer value to the human condition, then subjecting animals to such experimentation is only a venture into cruelty.

Chris, just a few points on your choice of analogies.
“We can say, quite literally, that drink-driving kills.” In the instance of drink-driving and accidents producing injury/fatality, the causation of trauma to physiology and vital organs, and the cutting of blood supply in injury and death is already known independently of driving producing these sorts of circumstances. In the instance of cholera, there is a bacterial micro-organism at play, i.e., an entirely different phenomenon. Concerning smoking and LC, the causal “argument” is based entirely on the differential in LC between smokers/nonsmokers: The differential is the entire basis for the causal attribution.

For the purposes of this discussion, this differential in LC is about 10% (for heavy smokers). The entire “smoking causes LC” claim rests on the size of this differential. Yet, the point that has been raised in this discussion is that the 10% figure is highly questionable. Two sources of error that can influence this 10% have been raised – misclassification error and, particularly, detection/diagnostic bias. For example, properly accounting for these two factors could elevate the accurate detection of LC for nonsmokers from 2% (in absolute terms) to say 4-5% and reduce LC for smokers from 12% to say 7-8%. The differential would then be 2-4%, not 10%. These factors have the potential of wiping the differential altogether. The problem is that, particularly with detection bias, the little relevant research indicating a substantial effect, i.e., an important factor, was never followed through. So we have constantly been fed this 10% differential.

Chris, you indicated that you believed that research has been exhaustive, that there can be no other vital factors. It should not be a requirement that unless someone can account for the [questionable] differential in LC, then smoking will be blamed in causal terms. Yet even in this circumstance, there is evidence that HPV might be relevant in a very significant percentage of LC in both smokers/nonsmokers. Yet even this possibility has been steered away from in mainstream epidemiology.

Hope you address these issues in your next post.



From 1900 onwards there is an observation repeated over and over again and up to the present day that LC is a disease of industrial cities and not the rural areas.

Do you have any ideas on this anomaly?


From the Mc Tear case:

"[5.358] Dr Kerr had not been familiar with Spain et al. 1970 before coming to court to give evidence. According to the abstract of that paper, in 500 apparently "healthy" adults who died suddenly and unexpectedly and who lived and worked within the environs of a large urban area noted for its air pollution, autopsy studies revealed that 50% of all adult non-smokers had metaplasia of their tracheobronchial epithelium. Metaplasia was present in 70% of male cigarette smokers. Women in each category revealed half the frequency of metaplasia present in the men. Metaplasia was present in less than 20% of men under the age of 30 years, rose to a peak of 80% at age 50 years, and gradually declined in frequency after the age of 60 years."


(no subject) (Anonymous) Expand
First of all, I apologize for my bad English. I think Chris wrote a very good post. Well, it seems that everyone agrees that smoking is not a necessary condition for lung cancer. Why? Because there are people with cancer who are not (or are not former) smokers. Is smoking is a sufficient condition? Again, no. Because there are people who smoke and do not get lung cancer. So we must deal with statistics, with probabilities. And Rich is right: epidemiology IS statistics (perhaps bad statistics). But what "probability" means in this context? Of course, there is considerable disagreement about the interpretation of a probability model in other sciences (frequentist versus Bayesian interpretations). This disagreement is important. It influences how we use probability analysis, and what we think this analysis means and can prove. When Chris says "in the end, it becomes obvious that there is no evidence that will satisfy the critics" he is right. But this is inherent in science especially with regard to empirical and statistical studies (We need not be Kuhnian to recognize the problematic nature of empirical evidence in scientific controversies).
Anyway, we should be investigating whether smoking and other factors (A1, A2, ...) are sufficient conditions for cancer. Obviously, smoking AND Ai cannot be necessary conditions for cancer because any properties which are absent when the effect is present cannot be necessary conditions for the effect (and smoking can be absent). But some Ai may be a necessary condition. It is perfectly possible that we would never identify a necessary condition Ai. But research should focus on these factors, because: (1) they may be necessary conditions, (2) they may increase the predictive power of the statistical model. The emphasis on studies of tobacco can be at the service of public health (a complicated concept) but might not be contributing to science, to the growth of knowledge.


You say that tobacco industry scientists agreed that Auerbach had succeeded in producing lung cancer and saw no problem with the research. Do you have a link to any documentary evidence?
This is really a follow up to my earlier question in CATCH-9:

Your analogy with the Broad street pump handle seems to me to miss an important point. In the case of the pump, it was significantly associated with Cholera but scientists did not blame the pump itself. Instead they took it as a clue and went on to discover the real cause; a bacterium which turned out to be a uniquely responsible for Cholera. In other words you cannot contract Cholera unless you have the bacterium in your system.



There's this Philip Morris memo from several weeks after the study was published. He interviewed Auerbach personally.

“I would say that the experiment is a crude one but effective in that carcinoma in dogs has been produced ..The crux of the situation is whether there is general agreement by qualified pathologists that carcinoma, squamous-cell carcinoma has indeed been produced. And even if the cancer production is invalidated, the obvious emphysema
produced cannot be denied”


And this from the research manager at Gallaher (p. 38) from 2 months later:

"the Auerbach work proves beyond all reasonable doubt that fresh whole cigarette smoke is carcinogenic to dog lungs and therefore it is highly likely that it is carcinogenic t o human lungs … the results of the research would appear to us to remove the controversy regarding the causation of the majority of human lung cancer … to sum up we are of the opinion that the Auerbach’s work proves beyond reasonable doubt the causation of lung cancer by smoke"



(no subject) (Anonymous) Expand
(no subject) (Anonymous) Expand
(no subject) (Anonymous) Expand
(no subject) (Anonymous) Expand

The urban-rural divide.

Hueper - 1956

"In fact the first observations on an appreciable rise in the frequency of lung cancer were reported from the highly industrialized cities of densely populated Saxony during the first two decades of this century.

Some years later it was found that high lung cancer rates existed for the population of the industrialized territory of the Ruhr valley, while they were below average for the agricultural region of the Main valley."

Hueper - Court Case 1957

Q Actually Dr. when was an increase in lung cancer first noted?

A The pathologists of central Europe, particularly in Germany where autopsies were performed for many years in hospital institutes of pathology, the University institutes of pathology on the majority of patients who died in these hospitals, the percentage rate ran between 90% and 95% of the deaths which occurred, and showed around 1920 that there had been a gradual, and progressive increase in lung cancer deaths observed in autopsies since the turn of the Century.

Q Since about 1900?

A Around 1900

Q And then it gradually started to increase,

A the number of cases increased.

Having in mind that you were living in Central Europe, at that time, was there any increase in smoking during that period of time that parallelled the increase in lung cancer from 1900 to 1920?

A It is pretty well established from statements in the literature that a wide spread and general use of the cigarette among the people of industrialized countries started with the first World War.
That means there was an increase in the lung cancer incidence before the custom of smoking cigarettes became a widespread habit.
m=750&ocr_position=above_foramatted&start_page=1 page 7

"Bessy Braddock, Labour MP for Liverpool Exchange, favoured an environmental explanation, and therefore found the urban–rural divide a barrier to acceptance of the smoking–lung cancer connection.
‘In view of the fact that cigarette and pipe smoking goes on all over the country, it is folly to say that it is the main cause of lung cancer.’

"The text of a TV broadcast on the subject in 1953 after publication of the Doll–Hill research on smoking and lung cancer in 1950 gives a sense of the focus on both individual and environment.
Introduced by Charles Fletcher, later famous for his pioneering Your Life in Their Hands, the programme was called Matters of Medicine.

Dr Guy Scadding, taking part, expressed the views clearly:smoking cannot be called the cause of lung cancer, since non-smokers also get the disease, and moreover the increase in cigarette smoking is not likely to be the only cause of the increase in the lung cancer death rate.

"The effect of smoking cannot explain the difference in mortality between town and country dwellers."
http://fds.oup.com/www.oup.co.uk/pdf/0-19-926030-3.pdf Page 17

Kitty Little

"The increase in lung cancer was primarily an urban phenomenon, and it was not observed in genuinely rural communities."


"Numerous studies have shown that lung cancer rates are higher in urban than rural areas, controlling for differences in age and sex profiles. One explanation is that smoking rates are higher in urban areas, although it is not clear whether the variations in smoking behaviour fully account for the observed urban/rural gradient in lung cancer incidence.

Indeed, some studies have demonstrated an excess of cases in urban areas, even controlling for smoking behaviour."

Of course we have trucks and tractors all over the place now, but the urban-rural divide still shows up.


I think the last of those sources puts it well:

"Smoking behaviour did account for much of this urban excess in lung cancer, although it did not explain the entire effect. These results suggest that there are urban effects that influence the incidence of lung cancer that are not explained entirely by smoking behaviour. Possible explanations include the variations in exposure to air pollution, occupational differences and the legacy of selective migration between urban and rural areas."

In mid-century USA, the larger gap between rural and urban lung cancer would be largely explained by rural preference for chewing tobacco. You see the same effect in Sweden where the (rural) North has the highest consumption of snus and the lowest lung cancer rates. I'd be interested to see more recent figures for the USA.

A comment

As a fifty year old smoker I can only in this instance give my personal opinion.
I do believe two pint's of beer a night has no detrinental effect on health.
I do believe two bottles of vhodka a day does.
I do believe ten roll up's a day has no detrimental effect on health.
I do believe 60 cigarettes a day does.
I do believe being slim has no detrimental effect on health.
I do believe being even 1 stone overweight does.

I do believe people who have a roll up instead of unneccesary nourishment are doing themselves far less harm.

I do believe that the life expectancy average will go down in a few decades because too much food is far worse than the odd ciggy or beer.
I see many slim people in their 70's 80's 90's some smoke ,some do not.
However the smokers are far more likely to be be slim.

‘Only 10% of smokers get lung cancer’

"‘Only 10% of smokers get lung cancer’"

Actually, that number is way TOO high.

In the USA, in any given year, 99.93% of the current adult smokers will NOT die from lung cancer.

Table 2
We see that current smokers account for 20.9% of the lung cancers.

There are 157,000(USA) lung cancer deaths per year.

Current smokers are 32,813 of those lung cancer deaths.

There are about 46 million current adult smokers.

That gives us a current smoker lung cancer death rate of about 7/10,000 current smokers per year.

So; in any given year, 9,993 out 10,000 current smokers will NOT die from lung cancer.

That is 99.93%.

In any given year, only 7/100ths of 1% of the current smoker adults will die from lung cancer.

They average age of lung cancer death is 72 and there are NO lung cancer deaths below the age of 35. So, current smokers have, on average, 37 years of that 7/10,000 death rate.
37 x 7 = a 259/10,000 current smoker lifetime lung cancer death rate.

Over the course of a lifetime of smoking, 9,741 out of 10,000 current smokers will NOT die from lung cancer and that is 97.41%.

Gary K.

Re: ‘Only 10% of smokers get lung cancer’


I tend to agree that the figure of 10% is way too high. I have done a similar calculation on lung cancer deaths in Australia and have come up with a figure of around 3%-4%.

I get the feeling that, if lung cancer death rates were that high, we'd all know many, many people who'd succumbed to lung cancer. I am now 42 years old, and I can't name one family member/family friend who has. That's purely anecdotal, of course, but there are lots of very heavy current and ex-smokers amongst them.

Of course, it may well be that something else took them away before the lung cancer got to them, as Chris surmises, however I think that it's very important to verify whether or not this 10% number is correct. 10% is a very different kettle of fish to 3%.

Lung cancer and never-smokers


Antis prefer to babble on about the current smokers increased risk of lung cancer as compared to never-smokers.

Here is something that we SHOULD talk about and insist is important.

Current smokers have 99.95% of never-smokers chances of NOT dying from lung cancer, in any given year.

Using the link above, we find that the never-smokers lung cancer death rate is 2/10,000 in any given year.

Thus, 99.98% of the never-smokers will not die from lung cancer in any given year.

99.93% of the current smokers will not die from lung cancer in any given year.

99.93 is 99.95% of 99.98.

Current smokers have 99.95% of never-smokers chances of NOT dying from lung cancer, in any given year.

Gary K.



Over the course of a life-time, current smokers have 98.14% of the never-smokers chances of NOT dying from lung cancer.

Mulberry Accessories

[url=http://mulberrybrynmorebag.webs.com/]mulberry brynmore bag[/url]
[url=http://mulberryalexareview.webs.com/]mulberry zuccarin side effects[/url]
[url=http://mulberryfruit.webs.com/]mulberry recipes[/url]
[url=http://mulberrysales1.webs.com/]mulberry sales[/url]
[url=http://handbagmulberry.webs.com/]hand bag mulberry[/url]
[url=http://ebaymulberrybag.webs.com/]mulberry bags ebay[/url]
[url=http://mulberryhandbagsoutletyork.webs.com/]mulberry factory shop york review[/url]



  • 1