Friday 29 July 2016


Cancer = Death. And sooner rather than later, or at least that’s the public’s perception. So any link between alcohol and cancer implies precisely that equation and is therefore guaranteed to generate headlines. So last week headlines in the I-online and Guardian newspapers that screamed “Alcohol directly causes seven kinds of cancer, says study”, and “Proof: drinking alcohol causes cancer” were designed to alarm public opinion and to feed into the ongoing ‘public health’ narrative that “there is no safe level of alcohol consumption”.

So, presumably, in order to justify these headlines there must be some new weighty piece of research that’s been published, or perhaps a meta-analysis of numerous other peoples’ research that give us, finally, irrefutable proof of this claim. But actually nothing of the sort! What the newspapers were referring to was an opinion piece by an academic named Jennie Connor, who works at the University of Otago, New Zealand, and that was published in the Journal ‘Addiction’. It consists of just over 3,700 words plus references.

Without wishing to be overly pedantic, it is not ‘alcohol’ but the drinking of alcohol, mostly in copious quantities, that causes a number of different cancers and there is nothing new in this claim nor is it, for the most part, disputed. The issue here is about which alcohol-related cancers require the drinker to really work at it by drinking to excess and which can be triggered by relatively low levels of consumption. The alcohol mortality statistics for England show that over 80 per cent of deaths from diseases caused by alcohol consumption are deaths from alcoholic liver disease (4,333 in 2014) and liver fibrosis/cirrhosis (1,520 in 2014). Bear in mind that some 28 million people in England drink alcohol on a weekly basis, so these drinkers are at the extreme end of consumption – often drinking in excess of a bottle of Scotch a day for many years.

The claim made in the article by Jennie Connor is that alcohol-attributable cancers make up 5.8 per cent of all cancer deaths worldwide. But if you look at the alcohol mortality statistics for England you will not find any recorded deaths from lip, oral cavity and pharynx cancer, oesophageal cancer, colorectal cancer or cancer of the larynx. As I said above, alcohol-related deaths are mostly caused by alcoholic liver disease or liver fibrosis/cirrhosis.  Curiously, there is no reference to deaths from alcohol-related breast cancer in the official figures, but studies estimate that about 94% of breast cancers are linked to causes other than alcohol, so separating out alcohol-related breast cancer let alone those cases that result in death is extremely speculative.

And this is the crux of the matter: at what point can we say that a statistical association between alcohol consumption and a disease becomes sufficiently well established to conclude that the relationship is causal? For what it’s worth, I think that Jennie Connor’s claim that the statistical association is strong enough for us to conclude that alcohol consumption causes cancers at seven sites is probably correct (oropharynx, larynx, oesophagus, liver, colon, rectum and female breast). But with the exception of the liver and a tiny minority of breast cancers these cancers are rare and rarely result in death.

The link between alcohol consumption and some rare cancers needs to be balanced by an appreciation of the clear proof that there are overall health protective effects from light to moderate drinking, particularly in relation to heart disease and stroke. I am of course talking here of the famous j-shaped curve that shows that regular, moderate drinkers live on average longer lives than either heavy drinkers or ‘never drinkers’, and are less likely than never drinkers to die of heart disease or a cancer. Jennie Connor, in her article, casts doubt on that hypothesis and space prevents me from going into detail on that, although I have done so in previous articles. Her scepticism of the j-curve hypothesis rather spoils what is otherwise a rational article that largely avoids bias.

My main criticism is of the media. Even supposedly ‘high-brow’ newspapers like the Independent and the Guardian cannot resist the temptation to get a wowser headline  by bigging-up an article that contains nothing new into a ‘study’, when it isn’t. This is how moral panics are started and sustained by a public media that creates and perpetuates myths which get repeated so often they become “well-known facts”. Are the media behaving cynically? Do they know they’re doing this? I am left with the thought that a man who tells a lie is merely hiding the truth; a man who tells a lie so often he’s convinced himself has forgotten where he’s put it.

No comments:

Post a Comment