In approaching the subject of the relationship between beverage alcohol and cancer I am conscious of the fact that I am not a medical scientist, nor an epidemiologist. But I have spent over a decade studying, writing and speaking about the relationship between alcohol and society in all its aspects. I’ve published over 250 articles on the subject and one book – Culture Wars and Moral Panic, the Story of Alcohol and Society.
What I have learned is not to simply trust “the experts”; not to accept appeals to authority, because science can be abused and corrupted by a combination of good intentions and moral certainty and this is particularly true of alcohol science. What follows is a synopsis of the understanding my reading and study has afforded me in respect of alcohol as a cause of cancer.
Initially my writing and research on alcohol concentrated on alcohol as a driver of crime and disorder, but over the past seven years alcohol policy and health has been my main focus because, overwhelmingly, it has been “public health” that has led the attack on beverage alcohol. The three strategic thrusts that underpin the “public health” attack on alcohol haven’t changed; it is still the “three As” – availability, affordability and advertising – and it is clear that “healthists” would like all three to be reduced if not eliminated!
However, availability has a demand side as well as a supply side, and attempts to persuade people to drink less (reduce demand) have come to the fore in recent times. In particular, the publication in March 2016 of the Chief Medical Officers’ of Health new ‘low-risk’ drinking guidelines of 14 units a week for both men and women is intended to reduce demand for alcohol by persuading the public to drink less. What has been key to this effort has been a concerted attempt by medical authorities to link alcohol consumption, even at very low levels, with cancer. This in turn has fed into the Chief Medical Officer’s narrative: “There is no safe level of drinking.”
The purpose of this paper is to examine the links between alcohol and cancer, and to unpick and demystify what is being said. What is essential to my task is to explain the difference between ‘correlation’ and ‘causation’. Most of the studies that link drinking alcohol with various cancers are epidemiology studies. ‘Epidemiology’ is about establishing the risk factors for various diseases in a given population. In order to do so epidemiologists look for patterns that associate one thing with another; in other words, they look for statistical correlations. A key problem with epidemiology studies is that they often show correlations that are not causal. For example: smoking tobacco is highly correlated with getting lung cancer (and the correlation is causal); drinking beer is highly correlated with smoking tobacco (a pint and a fag), and therefore, statistically, drinking beer is also highly correlated with lung cancer (but doesn’t cause it).
A more studious example of this confusion between correlation and causation appeared in a recent article published in the journal ‘Addiction’ by an academic, Jennie Connor, from Otago University in New Zealand. She wrote about alcohol as a cause of cancer at seven different sites in the human body. There was nothing new in this article, but the media picked up on it and declared it was a “new study”. The headline read: “Finally, proof that alcohol causes cancer.” Commenting on this opinion piece Hank Campbell from the American Council on Science and Health said “While overuse of alcohol can certainly cause something like fatty liver disease and then cirrhosis, so can too many cheeseburgers. Lots of things can be harmful when misused, that is why the American Council on Science and Health talks about dose-response when environmentalists only want to talk about hazard.”
So, it appears that the purpose of Connor’s opinion piece, and others like it, is to promote the idea that it is plausible to represent the (weak) correlations between drinking alcohol and developing certain cancers that have been observed in population studies, as constituting evidence that drinking alcohol increases your risk of cancers because it causes them. The conclusion that it does is based almost entirely on epidemiology studies that show that drinkers have slightly higher rates of some specific cancers. Commenting on this methodology Professor Byron Sharp of the University of South Australia wrote:
“A more sober analysis would adopt the epidemiologists’ rule of not considering risk assessments of less than a factor of 3 as indicating any causality (see Taubes, G. 1995. ‘Epidemiology faces its limits.’ Science, 269:5221, 164-69). Unlike smoking, alcohol and cancer studies seldom report risk estimates anywhere near this level.”
My understanding is that there are hundreds of different cancers, but most epidemiologists use the medical convention of labelling a cancer according to where the tumour occurs. So, for the purpose of this discussion we might categorise cancers as follows:
- Oral/throat/oesophagus (aerodigestive cancers);
- Liver cancer;
- Colorectal; and
- Breast cancer.
What is often lacking in epidemiology studies that find statistical correlations between alcohol drinking and the cancers above, is any reference to the possible biological mechanisms by means of which alcohol might cause these cancers. Alcohol has been shown not to be carcinogenic in animal tests, and the evidence that ethyl alcohol itself is carcinogenic in relation to human cells is weak. And the epidemiology shows associations with only some cancers not others (indeed drinking seems to be associated with less of some cancers).
So, what can we say about the relationship between alcohol and these cancers?
A possible biological mechanism in relation to aerodigestive cancers, is that bacteria in the mouth and digestive tract convert alcohol to acetaldehyde, which is a carcinogen. This may provide a plausible explanation where acetaldehyde comes into direct contact with tissue, as in oral/throat cancers, which are very rare. So for oral/ throat cancers we have the highest correlations in epidemiology data (though nothing like the magnitude for smoking), and we have a plausible mechanism (the culprit being acetaldehyde). But these cancers are extremely rare, and rarely fatal. As Professor Byron Sharp notes:
“This is by far the best case for making a causal connection between drinking alcohol and a cancer. Though it must be noted that aerodigestive cancers are rare (e.g., a tiny 0.3 percent of US deaths are from oral cancers, and most are caused by smoking, age and genes) even though most of the population drinks alcohol.”
In relation to liver cancer, it is highly unlikely that acetaldehyde is the biological mechanism except in cases where the drinker is consuming very large quantities of alcohol. This is because the liver is very efficient at breaking down acetaldehyde.
Most of the deaths that arise from very heavy drinking in the UK occur in relation to alcoholic liver disease or cirrhosis of the liver, where the average consumption of sufferers is 200 units of alcohol a week (the equivalent of a 75cl bottle of Scotch a day). Alcoholic liver disease is often confused or conflated with liver cancer in the popular press. Liver cancer is a rare, though deadly disease. It may be that alcoholic liver disease increases the risk of developing a liver cancer, but actually most cases of it are caused by viral hepatitis, obesity, diabetes or genetic predisposition. Very few drinkers develop liver cancer, so the degree of absolute risk is tiny and probably only for long-term alcoholics or for those with liver damage from hepatitis or other diseases.
In relation to colorectal cancer and potential biological mechanisms: from my reading of the research it appears that the flora of the human gut is highly complex, and whilst some faecal bacteria has been shown to convert alcohol into acetaldehyde, other bacteria has been shown to break it down. There is no settled science around the issue of a biological mechanism linking alcohol with colorectal cancer. Further research is needed. In addition, alcohol is absorbed in the stomach and small intestine which makes it harder to explain how alcohol might reach the colon where it could be converted to acetaldehyde by bacteria.
In relation to breast cancer, which is what drives public health claims that “there is no safe level of drinking”, the lifetime risk of a female ‘never drinker’ developing breast cancer is just under 10 percent. For women who regularly drink moderate amounts of alcohol, it elevates to just over 10 percent. But 90 percent of breast cancer is not fatal and meta-analysis shows no link between moderate consumption of alcohol and breast cancer mortality, nor its recurrence (Gou et al 2013).
A large study which followed up women with breast cancer correlated with better chances of survival if they were regular drinkers before diagnosis. If they altered their drinking after diagnosis this had no link to their chance of dying from breast cancer, but an increase in drinking was associated with an overall increase in life expectancy, largely due to substantially fewer heart disease deaths among those who increased their alcohol consumption (Newcomb et al 2013). This is strong causal evidence that alcohol prevents heart disease, and seriously undermines the proposition that alcohol causes breast cancer.
The modest degree of increased risk for alcohol drinkers may simply be down to a confounding factor: namely drinkers are more likely to screen for breast cancers, which results in more diagnoses (Mu and Mukamal 2016).
For the categories of cancer referenced above the evidence that light drinking is a causal factor is extremely poor, whereas the evidence is much more convincing that the regular consumption of moderate amounts of alcohol reduces heart disease, diabetes, dementia and overall mortality.
There are numerous meta-analyses of population studies, supported by plausible biological mechanisms, that show the protective effects of moderate drinking in relation to heart disease.
Finally, there is a paradox that is worthy of consideration: to the extent that regular, moderate consumption of alcohol is correlated with longer life, as compared with the lower life expectancy of ‘never drinkers’, this does, by definition, increase their lifetime risk of cancer. Why? Because for most cancers age is the dominant risk factor. Half of all colorectal cancers occur in people aged over 70 years old; half of breast cancer is for women aged over 62. Moderate drinkers live longer, largely due to reduced levels of heart disease, diabetes and dementia and will therefore have to get more cancer. Everyone has to die of something.
In conclusion, I would say that the scientific evidence suggests alcohol might cause a few cancers, but at low levels of consumption the evidence is extremely weak; the cancers are rare and the elevated risk of dying is tiny. It is wrong to misrepresent this evidence and unethical to use it to scare the public.
Paul Chase is director of CPL Training and a leading commentator on alcohol and health policy