Post by Bruce Watsonit could not explain why the retrospective studies to date were revealing a disproportionately high incidence of the disease among smokers; if cars were at fault,smokers and nonsmokers ought to be affected equally
That's like asking why disproportionate number of people using blood pressure meds end up with more heart attacks and strokes than people who don't use these meds (because they don't need them). Or why people who wear sunglasses will statistically end up with more sunburns than who don't wear them (because they don't need them e.g. indoors/nightly folks or those living in less sunny climate).
The positive statistical correlation (on _non-randomized_ samples) between some disease D and consumption of X exists and can be equally strong, whether X causes D or whether it is therapeutic/protective against D or its precursors.
They key term in this "paradox" is in the condition "on _non-randomized_ samples" i.e. that subjects in such study are not randomly selected by researchers 'to consume X' or 'not to consume X', but consume X for reasons of their own (e.g. people exposed to sun will more often decide to wear sunglasses) which are outside the control of the researchers.
In normal science (unlike antismoking "science"), whenever such correlations are observed on non-randomize samples, the real _hard science_ is done in the followup (randomized study or experiment). Only hard science can disentangle the often complex web of causes and effects and then uncover the nature of the links (causal/protective) connecting X and D, which were only hinted at by the non-randomized study.
Strangely, as even your own above post illustrates, with antismoking "science" all you will hear about, for over 60 years since it started in the USA & UK, are the non-randomized studies, discovering that smokers have more of this or that health problem than non-smokers. Where was the normally immediate hard science followup to establish the nature of such statistical links observed on non-randomized samples? Why trot out only the weakest and most ambiguous kind of evidence?
Interestingly, already in 1958 the father of scientific statistical techniques, famous UK mathematician R.A. Fisher noticed this peculiarity of the antismoking "science" and called them out on it (pdf: http://www.york.ac.uk/depts/maths/histstat/fisher274.pdf ):
"<b>But the time has passed</b>, and although further investigation, in a sense, has taken place, it has consisted largely of the <b>repetition of observations of the same kind</b> as those which Hill and his colleagues called attention several years ago. I read a recent article to the effect that nineteen different investigations in different parts of the world had all concurred in confirming Dr. Hill's findings. I think they had concurred, but I think <b>they were mere repetitions of evidence of the same kind.</b>"
Yet, over half century later, they are still stuck spinning in the same hint & wink loop, parroting only the non-randomized associations as their science.
It's not that many experiments on this question were not done in these 6 decades. There were hundreds. The reason you won't hear about them is that they all went the "wrong" way -- smoking animals end up living longer, staying healthier and sharper into the old age (e.g. in radon exposure experiments, smoking dogs get 7 times less often lung cancer than non-smoking dogs). Similarly, in _randomized_ quitting trials on humans (group of smokers randomly divided in 2 halves, then one half pressured into quitting, the other half left alone to smoke as before), the quit group ends up within several years or decades with more lung cancers (or heart attacks) than smoking group.
How can that be? Just like that example with use of blood pressure meds correlating with more strokes and heart attacks, tobacco is an ancient panacea with therapeutic or protective effects for numerous health problems. Hence its use will correlate (on _non-randomized_ samples) precisely with health conditions/diseases for which it is therapeutic or protective.
For example, smokers are twice or more as likely to get rheumatoid arthritis (R.A.) than non-smokers. As result doctors will strongly urge their R.A. patients to quit smoking. Yet, when animal experiments are done, both nicotine alone, but more so the full tobacco smoke, are strongly protective against R.A., delaying its onset and reducing its severity in existent cases.
Hence, it is the well established potent anti-inflammatory effects of tobacco smoke (and weaker ones of nicotine alone) that are the underlying mechanism for observed positive statistical correlation between R.A. and smoking, exactly like that between use of blood pressure meds & future strokes or heart attacks.
For references & discussion of the above R.A. experiments, as well as many other animal experiments, see this post: http://www.longecity.org/forum/topic/61248-the-intelligent-smoker-what-should-a-smoker-take-to-nullify-harm/#entry564686