CATCH-17

Rich White again

This is likely to be my last post, for the time being at least. The next few weeks are going to be extremely hectic for me and I won’t have time to submit any more posts, except possibly the occasional comment. Although if Frank is putting this on hiatus over Christmas then it won’t matter anyway.

This debate has become somewhat circular, with continued discussion on the merits of epidemiology and the relevance of animal studies. From the looks of it, we’re never going to get anywhere if we continue with this. Some of the most interesting discussions and points have been raised by Frank’s deterring from the main body of points and putting forth different theories, like actual dosage from cigarettes smoked etc. With that in mind, I am taking a slightly different path with this post and will be looking beyond the correlations and statistics and into areas that are new for CATCH and hopefully readers will find it interesting.

Smoking is a funny topic. Even the most rational, level-headed and sceptical person can suddenly turn into something of a wreck with emotions running high. While we can question the evidence of such about anything – even if it just serves to reinforce the belief it perpetuates – with smoking, challenging the evidence makes us crazy, denialists, cranks; we’re ignored, ridiculed, met with derision – by apparently people on our own side.

But what is the basis for all this? It’s flimsy indeed. Chris bases his entire faith on the smoking/cancer premise on the statistics. It’s interesting to note than despite tobacco being so intensively studied for over 60 years, all we still have is statistics. To some, that isn’t a problem. To others, it’s a big red flag. Statistically, we could show a dose-dependent relationship between almost anything. People with headaches will take paracetamol, and the worse/more frequent the headaches, the more paracetamol they take. Statistically then, paracetamol causes headaches. Or respiration equipment is linked with difficulty breathing, in a dose-dependent way (the worse the breathing, the more respiration equipment), and so statistically respiration equipment causes difficulty breathing.

Of course, causation is always linked with correlation, but the opposite isn’t true at all, and as shown above, correlation alone can be hugely misleading. The soft science of epidemiology and statistics provide clues for hard science to investigate. The hard science would be animal studies and randomized clinical trials. The latter cannot really be done in smoking studies because it would be deemed unethical to make people start smoking (although intervention trials have been conducted, with underwhelming results for the investigators – look for Whitehall and MRFIT), but the former has been done.

So what do we find from animal studies? In addition to the fact that tobacco smoke has not induced lung cancer in any animal (Chris still refutes this, but we’ll stick to the legal and medical professionals who quite readily admit they failed. And of course, if the animals were dying from smoke, they would never have needed to use the F344 rats and A/J mice, genetically bred to develop vast amounts of tumours), there are a great many study where smoke-exposed animals outlived the non-smoking ones, whose maximum lifespans were actually increased, and in other studies where animals exposed to radon or radioactive particules all died, those exposed to tobacco smoke had a massively decreased mortality rate. In a critique of the 1982 SG Report we are told various things about animal studies.

“Chronic Inhalation of Cigarette Smoke by F344 Rats”: ‘smoke exposure did not change the total number of tumor-bearing animals relative to controls; however exposed rats had significantly fewer tumors in the hyphphyses, hematopietic-lymphoid systems, uteri, and ovaries but an increased number of tumors in the respiratory tracts and dermes.’

‘These animal data fit in with the concepts of Prof. Oeser in Berlin and Dr. Lock in Hamburg, that, if properly assessed, the epidemiological cancer rates have not changed, and that the only thing which has changed is that the increase in one type of cancer is compensated for by a decrease in other organ cancer’ [For those who missed it, I explained in a comment in CATCH-14 how one plausible explanation for this is that nicotine stimulates vascular growth, i.e. rich blood supply in the lungs and aerodigestive tract. Cancer can, and will, migrate to the site that will benefit it most. This would mean smokers do get more lung cancer, but no more actual cancer than had they not smoked – the cancer just moves to the lungs. I then said it would be interesting to see the percentage of primary and secondary lung cancers in smokers.)

“A.P. Wenner et al”

‘The results show that the highest number of tumors occurred in the untreated control rats. The next highest number of tumors occurred in rats subjected to sham smoking, i.e. rats which were placed in the smoking machine without smoke exposure, and the lowest number of tumors occurred in the smoke-exposed rats…The bottom line of this research is that again, in spite of tremendous efforts and millions of dollars of expenditures, all efforts have failed to induce tumors having any resemblance to human tumors, by exposing animals to cigarette smoke.’

The notorious Nightlight posted in a forum:

“The most interesting tidbit though, buried deep in the dogs experiments section (experiments by Cross et al, 1982), at the end of a sentence (p. 555), are the results of combined effects of radon (via uranium dusts) inhalation and tobacco smoke. While the radon alone had no problems in inducing lung cancers in dogs (or any other animals), the combined effect of tobacco smoke and radon resulted in significant protection against lung cancers in smoking dogs — while the nonsmoking dogs exposed to radon had 37% incidence of lung tumors, the smoking dogs exposed to radon had only 5% lung tumor incidence (from the original Cross et al. 1982 paper, p. 48):

“Results of this experiment indicate that cigarette-smoke exposure, under the conditions of the experiment, had a mitigating effect of radon-daughter induced respiratory tract cancer in dogs… This difference is significant (p=0.03 Fischer’s Exact test).”

(source)

Further still:

“Inhalation Studies with Syrian Golden Hamsters”

‘The authors confirmed their previous data of experiments financed by the National Cancer Institute, of a “significantly longer lifespan of the smoke exposed hamsters”, as compared to non-smoking hamsters.

(The increased lifespan of animals became so common, and clearly problematic to the researchers, that animals were ‘terminated’ before they died naturally. This solved the little indiscrepancy.)

Bear in mind that in all these studies, the researchers tried very hard to cause cancer, exposing them to so much smoke that they were borderline asphyxiated (the researchers called it Maximum Dosage Tolerance, MDT), and yet they still couldn’t get the desired result. These animals sometimes smoked more than is humanly possible for us to smoke without vomiting, and they still outlived the non-smoking animals.

So if animals aren’t dying from smoke, and are outliving the non-smoking animals, why are we dying from smoke? We’re either the only single living animal to die from smoking (albeit after 50 or 60 years of doing it, which to many is rather peculiar, myself included), or we’ve made an error somewhere along the way. After 60 years, we have no biological evidence of how smoking causes cancer (Chris stated the animal studies are biological evidence, momentarily forgetting the animal studies have failed in inducing lung cancer in animals.) Given that the world’s oldest people were all smokers, and Dr Denson found that smokers who eat well and exercise outlive non-smokers, it does warrant investigation.

We’ve already discussed the initial epidemiology (which Doll conducted while under the pay of an asbestos company, and subsequent studies didn’t advance much on, as Fisher noted they were just repetition. Chris produced the Kaiser report in CATCH 14, a study which took account of absolutely no confounders. Chris said it didn’t matter as they wouldn’t have any effect on the results!), so what else do we know? In his study Comparative Epidemiology of Cancer Between the United States and Japan, Wynder stated the Japanese smoked less than the Americans. Chris has said the researchers of the 1950s were honest. But really? Wynder said: “The age adjusted mortality rates for laryngeal cancer during 1955 are higher in US White[s] than in the Japanese. These differences can be partially explained by the higher levels of cigarette and alcohol consumption in the US” even though his own tables demonstrated the Japanese smoked more. And if the research in the ‘50s was really objective and not affected by bias, why was smoking singled out and not looked at collectively with other factors? The very fact that Doll and Hill decided to straight away look at tobacco usage shows an element of bias. It’s odd that Chris thinks researchers were more honest then, because his own book is a demonstration of how far back negative attitudes towards tobacco go.

So what’s causing the deaths in smokers? It’s easy to say smoking is killing them, but this is too simplified. If tobacco is really ‘unique’ and such a huge killer, we would expect more deaths. Chris asserts 10% is a big risk, and perhaps in risk terms it is; but in real terms, a 90% chance of not getting ill is pretty good. If I had a 90% chance of winning the lottery, i’d play it. One of Chris’s champion reasons for accepting smoking=cancer is the statistics of females. Yet Chris also argues that different races are less likely to get lung cancer, so who’s to say women simply don’t have more resistance than men? Or, going back to the quote above about total cancer being constant but cancers of varying organs changing in regularity, perhaps women have the same total cancer as they did previously but more in the lungs. In which case, it could be explained by my aforementioned theory on cancer migrating i.e. secondary cancer. Let’s not forget, breast cancer affects men as well as women, but the rates are very low in men while seemingly increasing in women. There’s a strong statistical correlation between wearing a bra and developing breast cancer. But then again, maybe women have less lung cancer because they have more breast cancer – and their lung cancer rates are caused by secondary cancer, migrating to the lungs.

The negative placebo, or ‘nocebo’ effect, is very real with much research conducted into it. Nightlight posted on a forum the following:

There was a study in Heidelberg, described by Professor Eysenck in Psychological Reports (1989) in which 528 men were asked whether they, as smokers, were convinced that they would be very likely to develop lung cancer, heart disease, or other ‘smoking related diseases’. The 72 who answered ‘yes’, while admitting that their views were taken from information in the media, had an almost three times higher death rate at the end of 13 years than those who were not so influenced.

Fear can kill. This has been known since disease was first studied. We are entitled to wonder how many people have been killed more by the fear of ‘smoking related diseases’ than by any actual disease itself.

This is relevant nowadays, but what of disease in the 1950s when people weren’t worried about smoking? The self medication hypothesis could go some way to explaining it. We know smokers smoke for various reasons, and people note that it wakes them up in the morning, perks them up when they’re tired, suppresses hunger, and various other things. And yet fatigue is just one sign of various diseases, so how can we know that people don’t have an underlying or developing disease and smoke more to mask the symptoms? Indeed, this would perfectly correlate with the dose-dependency shown in the statistics of ‘the more you smoke the more you die’. Certainly none of Chris’s beloved epidemiology looks into it. The fact is, the epidemiology simply makes no attempt to learn why people smoke. There is a lot of research into the possibility that people are genetically predisposed to want to smoke, and independently develop lung cancer also. Does the epidemiology consider this? No, it doesn’t. If true, we realise that the smoking is just a correlation – the people have alleles in certain genes that predispose them to lung cancer regardless, but the fact they smoke confuses the results. I have repeatedly made the point that as more non-smokers and never smokers are now seemingly getting cancer than before, in lieu of the fact there are simply more non/never smokers now than ever before, it is entirely possible that the same overall rates exist but before they were in smokers and so hidden under ‘smoking related deaths’. This theory perfectly correlates with the above studies.

Back to self-medication, there is some validation of this in schizophrenia sufferers, as over 90% of them smoke and smoking helps alleviate the symptoms. Statistically, smoking could be shown to cause schizophrenia, but in reality the victims use smoking to help them – just like paracetamol and headaches. There is a lot of research into self-medication and it isn’t a crackpot theory designed to make us feel better about our filthy habit.

Nicotine is just one of many compounds of tobacco smoke; we are aware of the MAO B inhibitors in the smoke, which we can use to ‘self medicate’ to feel happier for instance. But what of physical ailments? Smoking has anti-inflammatory effects too, so we can ‘self medicate’ for various other problems. There’s more too, and I will quote again from Nightlight:

Specifically, some still unknown components of tobacco smoke upregulate substantially the key internal antioxidants and detox enzymes (glutathione [or more recent ref cf. Fig 4, p. L1076], catalase and SOD, all nearly doubled). These (families of) enzymes are responsible for neutralizing and excreting virtually every environmental or industraial toxin and carcinogen we are exposed to (WIKI: glutathione, SOD, catalase).

I stated some time ago about people smoking for different reasons and used the single example of stress. Dr Ian Dunbar spent much of his career looking at psychopharmacology and noticed the effects stress or insecurity can have on physical wellbeing. He explained it quite succinctly for this debate, but this brief quote is by no means exhaustive of the whole issue:

I am afraid that all the statistics about tobacco are worthless because they omit a common variable from their compilation.

I have gone on before about the importance of the mind. It is still generally dismissed by those who profess to be scientists as a figment of the imagination and ignored. However, it must be regarded as a physical entity with its own physiology and biochemistry like the kidneys or lungs.

Had scientists been more scientific they would be able to confirm that the mind is linked to the body by three hormone/neurotransmitter systems. The chemistry of those systems varies according to whether the individual sees itself as being secure in the environment or insecure in the environment.

That perception can be physical, intellectual or emotional and the security or insecurity acute or chronic.

If the individual is chronically insecure, physically, intellectually or emotionally then the body can become poisoned by the associated biochemistry just as cirrhosis of the liver is caused by chronic consumption of alcohol. I would submit that it is that poisoning that lies at the root not only of cancer but of ageing and of illness generally. People who are chronically insecure seek respite with sedative activities such as food, drink, drugs and sex all of which can be hazardous to health.

They also tend to smoke tobacco and inhale it deeply but that is a trivial side issue distracting people from facing up to the real problem.

So where does that take us? I’m not sure Chris realises the existence of psychopharmacology, but it is very real and recognised (stress produces cortisol for instance, while other ‘psychological’ distress causes a surge in adrenaline, and so on. Before Chris turns this into accusing me of making the claim that adrenaline is causing cancer, I’m really just making a couple of brief examples of the very real existence of ‘psychopharmacology’). Chris asked for some other reasons why smokers get more disease than non-smokers, and these are just a few very real, plausible, possible and recognised suggestions. Beyond that, what could it be? Why was there a surge in lung cancer in the 1930s? I already explained about advanced diagnostic tools that became used in the 1930s which caused an immediate surge in LC – the surge itself wasn’t real in terms of cases, but those now correctly diagnosed showed an increase. But around the 1880s, a case of lung cancer was found and it caused a stir in the medical community because it was believed they would never see it again. What does this mean? People were smoking long before this, the first cigarette maker was patented in 1881, and the first cigarette shop opened in, I believe, 1860 (correct date mentioned in an earlier CATCH, but it was around this time). So we know people were smoking, we know lung cancer existed but apparently not particularly common, even making allowances for misdiagnosis to consumption and TB, and we know this changed in the 20th century. Is it very likely that tobacco suddenly became dangerous? Not really. To go back to Nightlight once again, he briefly explains the inherent problems with secular graphs, but the second paragraph he gives some possible explanations for the rise in LC:

The secular trends are by far the weakest and the lowest resolution tool, among already low-res statistical methods on self-selected subjects, for reconstructing the complex graphs of causes and effects leading to diseases. They lump up far too many factors, quantified in extremely coarse grained, fuzzy paramatrizations or largely ignored altogether, to get close to even pretending to resolve between the models (a), (b) or ©. There are so many other secular trends going up or down at every scale, intertwined in myriad ways, that to pin anything on anything is a pure act of faith. It’s fine if one looks at it as a heuristic, providing coarse grained hints for subsequent use of more precise epidemiological tools, which in turn will provide hints for hard science to actually resolve the complex cause-effect graphs.

Regarding smoking & LC, how do you account for and adjust for the effect of nuclear fallouts, which also rose rapidly with atmospheric nuclear tests in early 1950s, coinciding in time and often in space, with rapid rise in lung cancers and melanomas (lungs & skin are the frontline interface of our bodies to the external air). Or simultaneous rise in EMF pollution, especially in high frequency ranges (whose quanta energies increasingly overlap in spectrum with the sectra of cellular biochemical processes)? Or trends in traffic pollution, fuel & engines trends,…? Or about myriad of other environmental, nutritional, medical, social,… trends, be it known, unknown, unquantified or unquantifiable, unfolding all in parallel on the same populations? It just can’t be done through such coarse grained methods. At most you can honestly get out of it is a hunch about a rough domain where to look for a better hint (e.g. using case control epidemiological studies), before focusing the finer tools of hard science to actually resolve between the alternative models (a), (b) or ©.

Interjecting this trail, Chris says that Asians have a resistance to lung cancer than westerners don’t have and thus have lower lung cancer rates. Perhaps that’s true, let’s run with it. Japan has the top life expectancy in the world right now, which depends on more than lung cancer. Chris asserted that lung cancer isn’t higher in smokers because of emphysema and heart disease carrying them off beforehand. So are we now to accept than Asians have resistance to all smoking related diseases, allowing them to puff away and still have the top life expectancy? The USA ranks a lowly 36 in the list despite having low smoking rates in terms of its own population, and the UK sits at number 20, while France and Spain, heavy smokers without the lung cancer resistance, are above us. In fact the Spanish smoke far, far more than nearby Portugal, share similar diets and culture, and outlive the Portuguese by some margin. You’d think really that if smoking was wiping them out in their droves, that would seriously hamper the life expectancy of these countries.

So basically, we have: the fact that statistics are misleading; the fact that epidemiology cannot prove anything and has missed (intentionally or accidentally) confounding factors and made no attempt to look at why people smoke; psychopharmacology providing some insight into reasons why people smoke and the effect these underlying reasons can have on physical health; scientific studies showing that some people have a genetic code that prompts them to smoke, and independently causes them to have lung cancer later in life; and a self-medication hypothesis that fits perfectly with the statistics, fits perfectly with other ‘medication/illness’ statistics, has a multitude of studies and is backed up by the rest of this paragraph.

At the end of the day, only epidemiology and statistics exist to link smoking to cancer. And when you only have epidemiology and statistics, you’re asking for trouble. If, after 60 years of intensive research, nothing of more substance exists, you’re asking for trouble. Especially when there are theories and evidence that explain away those statistics.