Booze – another gulp.

Whether a small amount of booze is good for our physical health was back with a vengeance in The Lancet this week, and the evidence is intriguing. 

I posted recently about a piece I wrote for Prospect magazine discussing that evidence – in particular what we’re learning from the brilliant, new, genetically-based technique of mendelian randomisation (there’s a two-minute, YouTube primer on MR here). This claims to offer a more robust answer about the direction of the effect, and it was this that made the week’s headlines.

But it’s not the answer we’d like. And it’s not the answer we once thought – that a small amont of alcohol could be on balance good for health. The new bottom line? ‘It’s bad for ya!’ – as the comedian George Carln liked to say. 

Yet again, supposedly robust knowledge confirmed through repeated and careful study has been challenged by new evidence. 

In this case, the old evidence came from observational studies of associations between reported levels of drinking and measures of ill-health. These studies had suggested that booze presented a risk of cancer at all levels of consumption, but for causes of heart disease there was a j-shaped curve, seen below in the data from these older studies published in The Lancet: the lowest risk was not – or so it seemed – at zero consumption, but slightly further along the curve at a glass or two (before rising again with higher levels of consumption.) That is, a glass or two was better than nothing – best of all, in fact – implying a protective effect.

But the mendelian randomisation evidence (see the Prospect article for an explanation of how this works and why it’s persuasive) removes the J-shape to produce a more-or-less straight line (see charts below). More booze is most probably worse for the health of your heart all the way along the line. Those old studies were quite likely confounded.

Here’s the new data as summarised in the Lancet.

“Enquiry is not standing upon the bedrock of fact. It is walking upon a bog, and can only say, this ground seems to hold for the present. Here I will stay until it begins to give way.”

Simon Blackburn, Truth, 2017 (thanks to Daniel Thornton).

The finding itself is striking enough. More striking still (to me) is the fact that we bust a gut to test the old evidence to destruction and for decades it seeemed to stand up. Then along came a new technique and the whole edifice of that understanding suddenly looks shaky.

So, is that it, settled at last? Is ‘it’s bad for ya!’ all you need to know? 

Not quite. Even if this new data is nailed-on certain and tells the true story – and I strongly suspect it’s a lot more likely to be right than the old evidence – the uncertainty hasn’t gone away. It’s still only nailed to the jelly of personal probability.

And the probability that you personally will be harmed by very low levels of drinking is still small. It was low before, and it’s probably low now (although when it comes to risk, “low” is always a matter of opinion). For a long time, we’ve known there was a chance that low levels of drinking could harm us – because of the effect on cancer. Now, the chance of harm has gone up a bit because it might include an additional risk – instead of a benefit – to your heart. But at these low levels it’s still a low probability, in my view. In that sense, things haven’t changed all that much. 

In other words, even if we now have strong evidence about the direction of the likely health effect (bad, not good), we’re still in the dark as individuals, because we’re still at the mercy of weak probabilities (though bear in mind that the probabilities rise rapidly with heavier drinking).

The evidence from Mendelian Randomisation is compelling, the technique itself a piece of genius, the lesson for population-level health important – but you and I are not populations, we’re individuals for whom almost every decision was and always will be a roll of the dice.

And that’s before you even begin to ask whether a drink or two – and the risk – is worth it for the pleasure it gives. Just don’t kid yourself any more that it’s doing your physical health any good. And so if you do decide to raise a glass, raise it not to good health, but to uncertainty.


Posted in: Uncategorized