Does smoking cigarettes cause lung cancer?

Question

My friend is a smoker. His dad is a doctor. Both claim there is no firm proof that cigarettes cause lung cancer. This article makes a similar claim -- in particular that smoking cigarettes is not proven to cause cancer, but that it is linked to an increased risk of dying from lung cancer:

Yes, a US white male (USWM) cigarette smoker has an 8% lifetime chance of dying from lung cancer but the USWM nonsmoker also has a 1% chance of dying from lung cancer.

Is there any firm science behind the argument that smoking does or does not cause lung cancer?

Answer 1

I don't have much to say here, because Cancer Research UK have already written a winning answer. (Sometimes the giants are tall enough that there's no need to stand on their shoulders.)

If you visit Tobacco, smoking and cancer: the evidence you will find explanations of all these facts:

• Smoking is the single biggest cause of cancer in the world
• Smoking greatly increases the risk of lung cancer
• The people with the highest lung cancer risks are those who:
  • smoke the most cigarettes per day
  • smoke over long periods of time, and
  • start smoking young
• Smoking is a major cause of several types of cancer
• Stopping smoking can reduce your risk
• Tobacco smoke contains many dangerous chemicals
• Tobacco smoke contains significant amounts of dangerous chemicals
• Chemicals in tobacco smoke can build up to harmful amounts
• The chemicals in smoke are more dangerous in combination than individually
• The poisons in cigarettes can affect almost every organ in the body
• Nicotine is a very addictive drug
• Smokers are still exposed to dangerous chemicals if they smoke filtered or ‘low-tar’ cigarettes
• Alcohol and other substances worsen the effect of smoking
• Second-hand smoking also causes cancer and kills thousands of people every year
• Children are especially at risk from second-hand smoking.
• Smoking while pregnant can harm your baby
• Smokeless tobacco can also cause cancer

They explain it all clearer than I could hope to, and include references to 93 different journal articles. You've got to admit they've certainly done their research on cancer, those guys at err.. Cancer Research UK.

Oh look, a donate button. Maybe you can persuade your mistaken friend and his father to donate to make up for all the people they have misled with their ignorance and semantic word games?

Answer 2

We know that smoking causes cancer due to research carried out a long time ago. In the original, and famous, paper Smoking and Carcinoma of the Lung by Doll and Hill, the authors attempt to answer the question of why there has been such a large increase in lung cancer. They consider different possibilities, but the data shows that smokers are more likely to get lung cancer than non-smokers (the paper contains figures by age and sex) and that heavy smokers are more likely to get lung cancer than light smokers. Other cancers are considered, to ensure that it is specifically lung cancer that is being caused. Overall, this paper is quite good evidence by itself (especially as it comes from a time when smoking was not seen as dangerous, so patients had no incentive to conceal the extent of their smoking), but of course there have been many more published since then.

Answer 3

Smoking is considered a cause of lung cancer because it significantly increases the risk of developing the disease. Most diseases, including lung cancer, are multifactorial, meaning they result from the interaction of various factors including genetic, environmental, and lifestyle factors. A cause in this context increases the probability of the disease occurring. It is not always a singular, direct cause but often one of many contributing factors. There are many "types of causes":

• Necessary Cause: A factor that must be present for a disease to occur. Without it, the disease cannot develop. For lung cancer, smoking is not a necessary cause because lung cancer can occur in non-smokers.
• Sufficient Cause: A factor that alone can cause the disease. Smoking is not a sufficient cause because not all smokers develop lung cancer.
• Component Cause: A factor that contributes to the development of a disease. Smoking is a component cause because it significantly increases the risk of lung cancer.

Causation is often understood through counterfactuals: "If A had not occurred, B would not have occurred." In the context of smoking and lung cancer, if we imagine a world where no one smokes, the incidence of lung cancer would be significantly lower. This establishes smoking as a cause of lung cancer.

Answer 4

People generally use the term "cause" as being similar to being "risk factor" and even I fall into that same people category so in that sense smoking causes cancer and I would agree with OP. However if we look at things strictly then maybe no.

Semantics

The definition of cause can vary from context to context.

First definition: This is originally from the article that cause generally means A implies B in a sense it occurs 100% or even majority (>50% of time). Smoking doesn't cause lung cancer 100% of time (Do 80 to 90 percent of lifelong smokers never develop lung cancer?).

Second definition: By definition of cause and effect, the disease-causing agent must always be present if the effect is to occur. You can get lung cancer even if you don't smoke, smoking doesn't need to be present.

That doesn't mean a diseases can have multiple causes though. C. albicans causes candidasis. Lets say even if you dont have C. albicans but have C. glabrata though candidasis can occur. But as whole candida. spp must always be present for candidasis to occur. This definition is true for microorganism caused diseases yet maynt be true for NCDs. HIV is a cause for AIDS. Infectious diseases have a cause. Non communicable diseases generally doesn't have cause but rather have risk factors.

Hence the article is basically arguing smoking is not the cause but rather risk factor. Its not arguing that smoking doesn't increase your likelihood of getting lung cancer but its arguing semantically its wrong to call it a "cause". Its also arguing that the risk factor of smoking is less than what media & organizations portrays and some of the evidence provided is pretty strong like this one. But so what, even if its not as carcinogenic as other risk factors for lung cancer, it still enhances mortality (non cancer deaths for example CVD deaths etc) even if you are lifelong smoker and didn't got cancer which is why you should avoid.

Misinterpretation of article

Even the OP misinterpreted the author semantics. This article makes a similar claim -- in particular that smoking cigarettes is not proven to cause cancer, but that it is linked to an increased risk of dying from lung cancer:

When he wrote this, ".. in particular that smoking cigarettes is not proven to cause cancer, but that it is linked to an increased risk of dying from lung cancer:" It implies a different meaning than if he would have wrote ".. in particular that smoking cigarettes is not proven to cause cancer, but that it is linked to an increased risk of cancer" .

In the first and latter pargraphs you can see:

"Yes, it is true, smoking does not cause lung cancer. It is only one of many risk factors for lung cancer".

"If they would say that smoking increases the incidence of lung cancer or that smoking is a risk factor in the development of lung cancer, then I would agree. The purpose of this article is to emphasize the need to use language appropriately in both the medical and scientific literature (the media, as a whole, may be a lost cause). "

Misc

I am writing this cause it seems to be people lack the fundamental understanding of point article is trying to prove. Most answers are supposedly trying to "disprove" the article by citing sources such as research websites but thing is you can't literally disprove semantics or definition differences cause you literally can't. Literally none of the answer's are in anyway disproving the article titled "smoking causes cancer" by providing refs to articles because those articles are simply stating what the author is stating. He does believe smoking increases risk of cancer, he doesn't simply believe it causes it because of differences in semantic differences. No matter how many references an answer has, it doesn't add any value to answer because argument itself is about semantics, not that smoking doesn't raise cancer risk. If among those references, any of them shows that smoking causes cancer majority of time (>50%) then maybe it would be disproving the article, which in that case they could simply linked that one article that disproves it all.

I have seen countless examples and hundreds of time people arguing because they have different definitions. Even if we go by lets say random definition I found in internet like this one: By definition of cause and effect, the disease-causing agent must always be present if the effect is to occur

which is specifically for diseases and cancer is a disease after all. Smoking will fail for it hence it can't be declared as a cause for this definition. While this definition might not hold true for other cases because we aren't really taught definition of cause in a concise way but rather in a broad way so in non disease context, we can use cause for things that doesn't happen 100% of time or cause must be present 100% of time but at least in disease context, it is generally believed that NCD (Non Communicable Diseases) doesn't really have a definitive cause but rather have risk factor, at least that what I was taught in school. So I would still lie more on it being a risk factor than a cause and that what article is trying to say even if you just read even first paragraph fully or even the latter one where it does claim smoking increases likelihood of getting cancer. Don't get me wrong, its another thing we humans use cause (in terms of NCD) interchangeably with increases likelihood of increasing cancer which even I would agree, and me myself would use smoking causes cancer interchangeably with smoking increases risk of cancer in real world cause you don't have to be autistic about it. Listener would understand that when you say smoking causes cancer they would mean that in sense it would increase risk of cancer.

I see people using "dependence", "habit forming" interchangeably with "addiction" which arguably could be debated on whether they are interchangeable or not but people still use it anyways. You will see articles claiming constipation drugs(laxatives not opioids) causes addiction or antihistamines cause addiction. Lets say someone makes a articles with polarizing titles like "No, antihistamines aren't addictive" or "Laxatives aren't addictive" which would be highly polarizing for sure. The article basically states that laxatives, etc doesn't cause addiction because they inherently don't have rewarding mechanism hence they don't cause cravings. You simply relapse on it simply because you will try to prevent withdrawals from happening. Antihistamine withdrawal might make you severely itch to get that another dose again yet they aren't addictive. You might even want to ask in site like skeptics whether they are addictive or not. People might even try to argue psychological dependence can be seen as addiction and use different definitions on how that article is wrong and it is in fact addictive and still depending upon definition even those website would be right. However the best answer would be a answer based on semantics on how both can be right depending on definition/semantics.

For cause, there's no objective definition but the original in the articles state cause must produce an effect 100% of time or majority of time, and this definition states cause must be present 100% of time. Either way even those two definitions might not always apply everywhere, there might be more definition and there's no universal objective definition and thats why you can't literally disprove articles claiming smoking doesn't cause cancer, antihistamines aren't addictive and amphetamine isn't nootropic. You can, however, acknowledge though, that both sides are valid and there are different ways to look at things.

Summary

The main points of articles are

i) If you can understand different POV, you will clearly understand point article is trying to make and indeed it makes a good one ii) It does seem like OP misintepreted the article, while answers were mainly focused on risk part rather than semantics part

Imo, mine answer should be rather most upvoted because it only also agrees with OP but also disagrees with OP and acknowledges because of semantic differences both can be right, the article isn't wrong and indeed one can technically call smoking not a "cause" of cancer

Edit: While I do see new answers popping up talking about causes right now, its just a matter of POV whether you view smoking as a cause or not!

Arguments against my answer:

i) What if person with HIV dies before getting AIDS:

"HIV virus injected into humans causes AIDS >99% of time." This isn't true at all. Firstly, because of retrovirals, but also because AIDS takes 10-15 years to develop - many people don't live long enough to find out if they would have developed AIDS")

I mean with that argument, any microorganism of infectious diseases wouldn't be a cause of that specific disease. What if you died or not lived long enough before incubation period is completed. Wouldn't that make microorganism not the cause of diseases it causes? Obviously we are talking about theoretical development of AIDS from HIV which would be >99% of case as long they complete the incubation period. Literally just random google searches brings this up. In all but a few rare cases, if left untreated, HIV will progress to a stage of infection called AIDS. Similarly exceptions doesn't change the rule..

ii) Arguments regarding stochastic process:

It could be argued HIV or other communicable is a stochastic process because there is very minor percentage of people that wont progress to AIDS (1-5%) usually due to genetic immunity etc. However, that's still exception rather than rule. Regardless, people are quick to criticize my definitions of cause and randomness yet none of the answers seems to be even talking about semantics.

Even if communicable diseases were a stochastic processes, one could argue its inherently more casual mainly because it requires presence of microorganism and it can't simply occur without it. NCDs will occur regardless whether you smoke or not. Whats crazy is that according to some models, cancer risk drops to almost non smokers are 10-20 years of not smoking. The subjects who had smoked for one to 19 years returned to the risk level of lifetime non-smokers within five to 10 years after stopping ... This doesn't suggest it acts in any way communicable diseases do. This lines in perfectly with the original article stating no evidence has been found for smokers who smoke less than 1 pack of cigrettes ("Occasional cigarette use (<1 pk/wk) has never been shown to be a risk factor in lung cancer."). Though not all models agrees, some do find elevated risk for former smokers and some do find occasional smoking to be risk.

Regardless, I would still stand on my original argument regarding semantics that smoking can be definitely viewed as risk factor and not cause, and also as a cause if you use different definition. Semantic differences always leads to endless argument's as I already stated before, even if you don't agree with me but based on evidences you could definitely consider smoking to be not a cause but rather a risk factor and it would still be valid. It all boils down to semantics.

iii) Weird Analogies 1:

"If this is the route you want to go down couldn't you also say something along the lines that gunshots to the head don't kill people because there are rare instances where the person survives?":

This analogy inherently invalid because it assumes if a person survives "gun shots to head are not cause of death". Think about it carefully, you are basically stating that in exception case if person survived it would be invalid to call it cause. But thing is I didn't use any exception as the rule in most of my examples, in fact rather opposite in case of HIV/AIDS. I dont see how this analogy even applies here, is there any place where I said or stated anything that would imply exception as rule logic? NCDs are uncertain in sense basically anything can be risk factor (or as people like to call it "cause"). Sleep, diet, literally anything. It has no certainity when it will occur. NCDs especially cancer are nothing like infectious disease. Communicable diseases are certain. "In these instances, medical science has demonstrated with a high degree of certainty that a disease will occur if the agent known to be causative is present." If you think thats not the case, you aren't basically arguing against me but arguing against this article which state causative agents must be present (in CDs causative agents are present) and they have effect that have high certainity as basically just stated by quote.

I am trying to understand what writer meant when he tried to write this analogy. Best I can think of is, smoking is lot risker than lets say not having a good diet in that (higher OR (odds ratio) from smoking than having bad diet). But thing is cancer is literally caused by anything (old age, lack of exercise, diet, sleep etc), and thus removing one of ten's of causes doesn't imply "its removing the rule to justify the exception". Would you even consider them as "cause" at that point? Hypothetically speaking lets say nicotine weren't addictive and X-rays were addictive and suddenly all people started using X-rays (1 billion out of 8 billion as its rate of smoking) then surely it would be greatest thing that increases risk of cancer among other things. But removing it would still not make it "removing the rule to justify the exception" or "exception as rule" thing. Rather the opposite, I essentially wrote if someone has evidence that smoking causes cancer majority of time and smoking must be present (as cause) when cancer as effect is seen. Both statement implies the rule rather than exception case.

iii) Weird Analogies 2: I don't have time to reply to all weird analogies but I am writing this because I dont want to check repeatedly about people bringing up weird analogies, and this analogy could be possibly brought up. I saw this in comment which supposedly has lot of upvotes butit doesn't make any sense to me.

"let alone about whether the smoking itself causes something or the chemicals deposited in the lungs as a side effect of that smoking :) Falling out of a 100m crane doesn't cause you to break your neck after all, it's the falling on the ground afterwards that does that :)"

First of all thats not what article is claiming to be. Suppose A is component of B and A causes X". Your analogy basically states its not the A causes X but rather than B causes X. This analogy would apply in situation if author wrote "Smoking doesn't cause cancer: Its the TSNAs and PAHs in tobacco that do". That would be more fitting. Clearly the distinction is about definition on how you define A is related to X, is A a cause or risk.

I can't think of other weird analogies. Either way, that just I how feel that those analogies doesn't really debate anything or present any good argument given I cant think how I used "exception as rule" etc. And, yes I did use word cause not risk factor in the post but if it was not obvious enough, I don't mind using cause or risk interchangeably as long meaning is understood. Will I criticize someone if they might not want to use word cause and insist on using risk. I guess not, that in some sense can be totally valid way of thinking

iv) 100% of time: Also I missed this one but just to clarify. We are talking about theoretical and practical expectations. Just like how theoretical things doesn't is always practical. A 100% effect all the time might not hold practically! HIV is theoretically expected to cause AIDS 100% of time but it causes in around >90% of time which is a majority. Car engine might have old components etc but theoretically its expected to start 100% of time. Machines in physics work all the time. It wouldn't make sense in physics classroom if student somehow asked "What if simple machine couldn't function 100% of time", "What if car engine doesn't start if engine is faulty, it got rust" and so on. Regardless, I also brought up second definition that "cause" doesnt require "effect" to be produced 100% of time, but "cause" must be present 100% of time whenever "effect" occurs. Cancer on other hand is inherently like probablity, lottery or russian roulette. I hope that clarifies things

A side note: Its best if you are editing or suggest an edit to the article, do not trim it down all and keep most of the part as it is. Maybe if you feel like something is unnecessary do trim it down but keep the general gist of it. The answer is so long to make you understand the point of article, and why question itself could be deemed off mainly because when OP asked this "Is there any firm science behind the argument that smoking does or does not cause lung cancer?", in a scientific sense some definition that's highly relevant in medical science mayn't fit for NCDs but fit for CDs, hence correct answer could be somewhere in sense "According to some scientific definitions, cancer might not be called a cause but maybe for others it could be".