Wednesday 27 May 2015

Under-reporting alcohol

From the Telegraph...

Just one alcoholic drink a day damages hearts of elderly women, Harvard study warns
Only one alcoholic drink a day damages the hearts of elderly women, a new study has warned.

Despite previous research that suggested a drink a day might protect against some cardiovascular disease, the findings suggested otherwise, at least in the elderly.

Women who have just one alcoholic drink a day are defined as light to moderate drinkers. Yet new research found women with an average age of 76 who drink moderately had small reductions in heart function because of an enlargement of the wall of the heart's main pumping chamber - the left ventricular mass.

I'll take a small reduction in heart function at the age of 76 if it means I have less chance of getting heart disease. And make no mistake, heart disease is significantly less prevalent amongst drinkers than it is amongst teetotallers.

The temperance lobby hates this fact, which is why there is a concerted and growing effort to persuade people that there is no safe level of drinking. I don't know whether this particular study was designed to serve that end but the merchants of doubt will be pleased to see it in the media.

Also in the media recently was a story about drinkers under-reporting what they drink. This comes up from time to time (eg. here in 2013) and could give the unwary reader the impression that Britons drink much more than was previously thought. This would be true if we estimated how much people drink by asking them, but we don't. We work out national consumption by looking at how much is sold. This does not give a perfect picture because some alcohol is never drunk and some alcohol is bought illegally (and therefore doesn't show up on HMRC tax receipts), but it is fairly accurate and it is certainly more accurate than asking drinkers.

Under-reporting was in the news last week because some researchers decided to ask people how much they drink on holiday and on special occasions. Surprise, surprise, they drank more than in the average working week.

This is much as might be expected. Under-reporting in endemic in the alcohol field and researchers are rightly sceptical about self-reported evidence. But the important point is that virtually all alcohol epidemiology is based on self-reported evidence. If people under-report their consumption by 40-60 per cent, then observational studies which show a risk from some disease or other at (for the sake of argument) 50 units a week actually show risk kicking in at 75 units a week.

Alcohol researchers occasionally mention this. Here is Sadie Boniface in 2013, for example...

Much of what is known about the relationship between drinking and harm is based on self-reported data, where consumption was under-reported. This means that the relationship between drinking and harm may have been over-estimated, and that drinking is effectively ‘safer’ than the Government’s drinking guidelines suggest. This raises the question of whether the guidelines should be raised to reflect the actual relationship between drinking and harm, creating perverse incentives for alcohol policy (not just in the UK but worldwide).

So it's not that researchers are unaware of the problem. They just don't seem to be as interested in it as they could be considering the implications (tellingly, Boniface says that "there wasn’t space to discuss [this issue] in detail in the paper [that she had just published]"). Certainly, there is no urgency to raise the guidelines. Quite the reverse.

Last week, the subject got an airing in The Parliament Magazine. Referring to research conducted in the US, it concluded...

In summary, scientific data clearly show a relation between heavy drinking and certain types of cancer. For even light-to-moderate drinking, many studies have shown a slight increase in the risk of certain cancers, especially breast cancer.

However, when subjects likely to be underreporting their intake are removed from that pool, new data indicate that light to moderate drinking is not associated with an increased risk for breast cancer or for other types of cancer.

The question of alcohol and cancer boils down to dose: heavy consumption increases risk but light to moderate consumption is unlikely to increase risk.

This is the exact opposite of what the 'no safe level' faction of the neo-temperance lobby want you to believe so you can expect the message to be confined to niche publications and blogs while claims about the putative health risks of drinking tiny amounts of alcohol - which are not, in all probability, so tiny - proliferate.

No comments: