Tuesday 20 September 2022

Snus is safe and e-cigarettes are miracle products

There were a couple of studies published last week that are worth bookmarking. They don't tell us anything we don't already know, but they contradict what a lot of people think they know.

The first looks at smokeless tobacco and oral cancer risk. It finds that the kind of smokeless tobacco widely used in Asia, such as gutkha, is strongly associated with oral cancer (partly because these products contain many other ingredients in addition to tobacco), but there is one notable exception.

Except snus, all SLT [smokeless tobacco] products sold in different WHO regions are strongly associated with OC [oral cancer] incidence.

Readers of The Art of Suppression may recall that snus was banned in the EU because it was assumed to have a similar risk profile to other smokeless products. It doesn't. Not even close.

The other study looks at a remarkable feature of vaping that I have mentioned before - its extraordinary ability to help people quit smoking even when they have no intention of doing so.

Overall, 12.7% of smokers quit smoking. Smokers not initially planning to quit within 6 months experienced higher odds of smoking cessation when they took up daily vaping (32.4%) versus no vaping (6.8%; adjusted odds ratio [AOR]=8.58, 95% confidence interval [CI]:5.06-14.54).

... Uptake of nicotine vaping appears to be strongly associated with cigarette smoking cessation among smokers with no initial plans to quit smoking.

The difference between a 6.8% quit rate and a 32.4% quit rate is enormous. The awesome power of vaping to get unmotivated smokers to quit deserves more attention, especially since only half of British smokers say they intend to quit.
Imagine being an anti-smoking campaigners and being opposed to e-cigarettes...

Friday 16 September 2022

Axe the tax, Liz

More from me on reports that Liz Truss may be ditching Boris Johnson's obesity strategy and repealing the sugar tax...

In public health circles, it is considered terribly gauche to expect policies to work. You might think, for example, that a trailblazing intervention designed to reduce obesity would be considered a failure if obesity rates rise to record highs after it has been implemented. Not so with the sugar tax. Obesity among both children and adults has gone up since it was introduced in 2018, but the health lobby does not consider it to be a failure. Contrary to the evidence of your eyes, they say, it has actually been a success. The only failure is the failure of the government to do lots of other things in addition.

This is one of a number of ways in which ‘public health’ differs from medicine. If a patient’s health gets worse after being given some pills, doctors do not claim that the treatment has been a success and double the dose.


Wednesday 14 September 2022

Dare to dream

Has the Tory party finally found a leader with sound convictions and the bottle to get things done? The early signs are encouraging. Today's Guardian front page laments that...

Liz Truss could scrap anti-obesity strategy in drive to cut red tape

Exclusive: Health officials ‘aghast’ as review launched of measures to deter people from eating junk food

... The review is so radical in scope that it may even look at whether the sugar tax, which began in 2018 and has helped make soft drinks much less unhealthy, should go too. Health experts have hailed the levy as a key initiative in the fight against dangerous obesity.


Civil servants being aghast is a good thing when they work for what used to be known as Public Health England. Everyone from the Institute for Government to the Soil Association is horrified by the idea of an ineffective, regressive tax being ditched.
I've written about this for Spiked...

The sugar tax has achieved the square root of diddly squat while extracting £1 billion from the pockets of hard-pressed consumers and ruining the taste of several cherished soft-drink brands. For a prime minister focused on deregulation, lowering taxes and the cost of living, getting rid of it is an obvious place to start.

If Truss does scrap the tax, she will inevitably face fierce resistance from ‘The Blob’. The usual state-funded pressure groups and media midwits are already up in arms at the mere suggestion, as are the activist-bureaucrats who run the Department of Health. Although the sugar tax has only existed for four years and patently has no value as a public-health measure, you would think Truss was considering a ban on penicillin from the hysteria in some quarters.

Useless and regressive, the sugar tax was – to quote Jamie Oliver – a ‘symbolic’ measure. Its repeal would be no less symbolic. It would show that after getting through four prime ministers in six years, the Conservative Party had finally landed on a leader who genuinely believes in free markets and personal freedom. It would show that the nanny-state lobby does not always have to win, that Britain is not condemned to ever-more restrictive lifestyle regulation and that there is an exit on the road to serfdom. If Truss picks this battle, she will have my sword, and I hope yours, too.


Tuesday 13 September 2022

Low risk gambling guidelines

Canada is swiftly becoming nearly as much of a lost cause as Australia, liberty-wise. That the Canadian 'public health' racket is out of control was shown recently when the Canadian Centre on Substance Use and Addiction recommended that the drinking guidelines be reduced to two drinks per week. Yes, that's per week, not per day. The Canadian government also plans to put health warnings on each individual cigarette from next year.

I recently discovered that Canada also has 'lower risk gambling guidelines'. I haven't heard of this concept anywhere else in the world. There are ways of diagnosing problem gambling, but this is the first time I've come across general advice to the population to help people avoid becoming a problem gambler.

The government boasts that its guidelines are based on "the most current scientific evidence available". I'd be interested to see this evidence because it seems to me that the guidelines have been rectally sourced. They are: 
1: Gamble no more than 1% of household income
2: Gamble no more than 4 days per month
3: Avoid regularly gambling at more than 2 types of games
Moreover, if you have "problems from alcohol, cannabis or other drug use", you should gamble even less than this and preferably not at all. 

I suppose problem gamblers break all three of these rules on a regular basis, but that doesn't mean that you or I will become a problem gambler if we break any of them. 
Since the 'harm' associated with being a problem gambler comes from losing more money than you afford, sticking to the first of these tips (Gamble no more than 1% of household income) is guaranteed to prevent all 'gambling-related harm'. They are therefore not 'low-risk' gambling guidelines. They are zero-risk gambling guidelines.

How useful and evidence-based are they? The academics who created them identified several traits of problem gamblers, such as a tendency to engage in various different forms of gambling, and are trying to steer people away from them. But isn't this the tail wagging the dog? Problem gamblers play different games because they gamble at every opportunity. They are problem gamblers, after all. Can we really reverse cause and effect and assume that if someone sticks to one or two forms of gambling, they won't become a problem gambler? 
Can we assume that a problem gambler will pay any attention to these guidelines at all? They are about as much use as telling people to drink no more than two drinks a week. As with the drinking guidelines, the only people who will abide by them are people who are never going to run into trouble in the first place.

One of the covert purposes of constantly reducing the drinking guidelines is to inflate the number of 'hazardous drinkers’ in society, thereby creating demand for action from politicians. But hazardous drinking isn’t a proper clinical concept. If your doctor thinks you’re drinking too much you might get given a questionnaire to assess ‘harmful drinking’ - which is a clinical concept - but the questions have nothing to do with 14 units a week in the UK (or whatever arbitrary figure has been picked elsewhere). 

With problem gambling rates currently at the very low level of 0.2% in Britain, the concept of 'at risk' gambling has already been adopted to come up with a scarier figure. How long before the UK introduces low risk gambling guidelines and uses estimates from it, which will presumably run into the millions, to get an even bigger number?

Monday 12 September 2022

'Problematic information' about nicotine and COVID-19

A study titled 'The influence of pro-vaping "gatewatchers" on the dissemination of COVID-19 misinformation on Twitter' has been published in something called the Journal of Medical Internet Research. It's paywalled but the pre-print version is here. It is yet another whinge about people retweeting things that 'public health' academics don't agree with. 

The authors lament the lack of gatekeepers on Twitter to control what scientific information people have access to. They then invent the meaningless term 'gatewatchers' to describe anyone who has built up a following and puts out interesting tweets. After the usual dredging of Twitter, they conclude that quite a few of the people who mentioned the evidence that smokers are less likely to get COVID-19 in 2020 had a 'pro-vaping bias'.

A review of early clinical data in Wuhan province found that smokers were less likely to be admitted to the ICU due to COVID-19 complications, compared to non-smokers [10]. Although the study itself did not provide any evidence of a causal relationship between smoking and COVID-19 infection or progression, the authors posited that the anti-inflammatory properties of nicotine might be responsible for the unexpectedly low prevalence of COVID-19 infected smokers in countries with high smoking rates [12, 13].

Indeed they did. This was sparked by evidence in an NEJM study showing surprisingly few smokers in hospital with Covid in China and a study by Dr Konstaninos Farsilinos published in May 2020.

While no subsequent evidence has been found to support a protective role of nicotine, the notion that smoking, vaping, or nicotine use would prevent COVID-19 circulated, leading researchers to document misinformation about smoking, vaping, and nicotine as protective against COVID-19 across communication channels, particularly on Twitter [14-16].

If you're going to write a study about misinformation, it's best not to put a big fat lie in your introduction. No subsequent evidence?!? I have so far found 98 studies looking at this question and the vast majority of them find that smokers are significantly less likely to get COVID-19. A less expansive meta-analysis published in Addiction last year found the same thing. Nicotine is the obvious explanation and there are biological reasons why nicotine would make a difference. It could be something else, but a study published only last week found that smokers given nicotine in hospital had better outcomes than those who weren't.
No amount of evidence will satisfy the authors (one of whom works for the USA's risible Truth Initiative), but to say that there is no evidence at all is just a lie. 

Such counterintuitive findings are not misinformation in a direct sense, in that they do not present demonstrably false information [21].

You can say that again.

Rather, the extrapolation of the study’s findings out of context or with overreaching implications exemplifies the sort of claims that are not egregiously false, but rather represent unsubstantiated and misleading implications that run counter to the best available scientific evidence [11, 22].

In other words, people on Twitter read about the research and discussed what it might mean.

The 'unsubstantiated and misleading implications' that people were tweeting about, according to the authors of this study, included articles in the Economist ('Smokers seem less likely than non smokers to fall ill with Covid-19'), the Telegraph ('Smokers 'four times less likely' to contract Covid-19, prompting nicotine patch trials on patients') and the Guardian ('French study suggests smokers at lower risk of getting coronavirus'). All were perfectly factual, even if the evidence displeased anti-vaping academics in America.

The authors quote the top 10 posts by the 'top users'. They don't give their names but typing the text into Twitter shows that they were the scientists Konstantinos Farsalinos and Eli Rubashkyn, the journalist Alex Norcia, a South African radio station, occasional commenter on this blog Klaus K., the news website News24, a lawyer who tweeted the News24 article and someone who can't be traced because the tweet has disappeared.

Several tweets by Farsalinos are in the top 10. You may recall that he wrote one of the studies. So a scientist tweeting about his own research falls under the umbrella of not-misinformation-as-such-but-we-don't-like-it as far as the authors are concerned, as does any Twitter user linking to a respectable news source.

This research demonstrates the utility of the gatewatching framework for examining the dissemination of problematic information on Twitter. More than half of our sample of top users and 22 of the 25 most prolific users producing and disseminating content about COVID-19 and tobacco in the first nine months of the pandemic were pro-vaping “harm reduction” advocates. 

Anti-vaping harm reduction opponents obviously saw nothing to be gained from mentioning it. So what? Perhaps the fault lies with them.

Building on previous research both identifying and quantifying the extent of a specific piece of misinformation that nicotine can prevent COVID-19, we show the disproportionately broad reach of this claim across the most retweeted content during this time period [16, 47, 48]. Even in May 2020, when the original study by Farsalinos et al was published, the preponderance of scientific evidence, including multiple meta-analyses, still opposed the notion that nicotine, and especially smoking, would protect people from COVID-19 [10].

This is untrue and reference 10 is Farsilinos's study so it provides zero support for the authors' claim.

Still, in our sample of the top 1000 retweets, propagation of this claim was more than 20 times more common than the five tweets trying to debunk the claim and was retweeted nearly 17 times as often.

One can only assume that Twitter users found the attempts to debunk the claim unconvincing.

This research highlights the growing challenge of addressing scientific distortions that while not themselves misinformation can nonetheless drive false beliefs. There is no reason to believe that the study finding fewer smokers than expected was falsified. In fact, this “smoker’s paradox” drove significant research interest and calls to pre-register hypotheses towards the goal of rigorously investigating the effects of nicotine on COVID-19 [53]. A substantial body of literature has provided strong evidence that smoking during a respiratory pandemic increases the risk of severe illness and death [5-8].

References 5-8, which include a hastily written junk meta-analysis from Stanton Glantz, looked at the health outcomes of smokers who have contracted COVID-19. They did not look at the relevant question of whether smokers are less likely to get COVID-19 in the first place.

Our study does not address this complicated body of literature.

There is a good reason for that.

Opinion leaders on Twitter do not have control over the content posted on the platform. However, they have outsized influence over the dissemination of certain perspectives over others.

Because people are interested in what they have to say.

Although more research is needed, we contribute strong evidence that the ideological lean of the most prolific tweeters on a given subject (pro-vape users discussing COVID-19 and nicotine) directly influenced the spread of problematic information (that nicotine could prevent COVID-19) through retweeting much of the most broadly disseminated posts.

Now do anti-vapers and see what they've been saying. You wouldn't believe the amount of 'problematic information' those guys pump out.

These findings are reminiscent of previous research showing the majority of disinformation in another context, anti-vaccination, emanated from only twelve users [56].

Anti-vaxxers are an interesting comparison, but not for the reasons the authors think. Anti-vaxxers say that there is not enough evidence that COVID-19 vaccines are safe in the long-term despite them having been in use for a considerable length of time and despite there being ample evidence from beyond observational epidemiology that they do not pose any significant long-term health risks. Anti-vaxxers focus on rare examples of harm while ignoring the major benefits they provide to the vast majority. And they claim that there is no evidence of efficacy despite the death rate from COVID-19 falling dramatically as soon as the population is immunised.

Similarly, anti-vapers claim there is no evidence of long-term safety despite ample evidence to the contrary, focus on very rare examples of harm (batteries exploding, etc.) while ignoring the broader benefits, and claim that e-cigarettes do not help smokers quit despite smoking prevalence falling noticeably wherever they have become popular.

However, in contrast to the “Disinformation Dozen,” pro-vaping gate watchers on Twitter are not producing and disseminating overtly false information.

So you keep (rightly) saying. So what exactly is the problem? 

Rather they serve as mediators between the scientific community and the broader, Twitter-using public, and privilege scientific findings that support a pro-vaping narrative while dismissing, ignoring, and countering a preponderance of evidence that does not.

The preponderance of evidence on this issue is all on their side. Even if it weren't, the US anti-vaping racket is vastly better resourced to counter it.

The broader implication of this process of dissemination is that the proverbial deck is stacked against effective public health communication on Twitter. 

If 'effective public health communication' means telling people that smokers are more likely to get COVID-19 and that vaping is very dangerous, then it would be a good thing if the proverbial deck was stacked against it. However, growing public ignorance about the risks of vaping suggests that it is not.

The observable pro-vaping bias of the most outspoken users discussing COVID-19 and nicotine inevitably meant that even attempts to debunk such information on the platform did not receive nearly the same amount of traffic. 

Boo hoo. Go ask Mike Bloomberg for some more money.

Wednesday 7 September 2022

Modern temperance

Professor David Nutt is a curious fellow. He is quite sound on vaping and drugs, but horribly puritanical about alcohol. The problem lies in his blinkered view of psychoactive substances which focuses solely on ‘harm’ and ignores both the benefits of taking the substance and the societal context in which consumption takes place.

For Nutt, ‘harm’ generally means the damage to the health of the user and those around him, but he will sometimes include barely measurable harms such as ‘loss of relationships’, ‘family adversities’ and ‘community’ to bulk up the figures. By this dubious method, he once produced a league table of substances in which ketamine, GHB and benzodiazapine were portrayed as better than alcohol. In fact, everything in the league table was better than alcohol, even tobacco.

From this, Nutt concluded that ‘the present drug classification systems have little relation to the evidence of harm’. The implication was that either alcohol should be banned or everything should be legal.

Now, it might be true in some sense that ketamine is a safer drug than alcohol, but it would hardly be an appropriate substitute for alcohol at a wedding, for instance. It isn’t even much of a substitute for alcohol in a pub.



And whilst Nutt insists that methamphetamine does not cause much harm to other people, I would argue that if everyone who drank alcohol switched to crystal meth tomorrow, society would soon fail to function.

While I agree with Nutt that MDMA and cannabis should be legal, I have long had concerns over his research, much of which seems to be blatantly agenda-driven and sloppy. His articles about alcohol, in particular, are riddled with errors. He is happy to repeat any old canard from the temperance lobby so long as it paints booze in a bad light. I have written about this again and again and again.

In January 2020, he published a book called Drink?: The New Science of Alcohol and Your Health. I haven’t read it and have no urge to do so, but I could see from his summary of it in the Daily Mail that it was packed full of half-truths, lies and exaggerations.

I had forgotten all about it until I read this post by Fergus McCullough who is very impressed by the book. The parts he quotes irritated me because they are largely untrue, but it irritates me even more to think that somebody believed them. This is not Fergus’s fault, as such. People should able to believe a book written by a professor. Nevertheless, what Nutt says is frequently wrong.

Take the health benefits of moderate drinking, for example, which seem to rile the anti-alcohol lobby more than anything. Fergus writes:

What I didn’t realise before, though, is how poorly evidenced the beneficial effects of alcohol are. Looking at the available studies, Nutt writes that the positive effect on cardiovascular health has never been definitely proven (i.e. beyond mere association), and even if there is a small positive effect, the optimal level of consumption would be around one unit a day. The benefits don’t outweigh all the other risks.

Never definitely proven ‘beyond mere association’? OK, so we don’t trust observational epidemiology. But two paragraphs earlier, Fergus quotes the following from Nutt’s book:

Alcohol use is one of the top five causes of disease and disability in almost all countries in Europe. In the UK, alcohol is now the leading cause of death in men between the ages of 16 and 54 years, accounting for over 20 per cent of the total. More than three quarters of liver cirrhosis deaths, 7 per cent of cancer deaths and 25 per cent of injury deaths in adults under 65 years of age in Europe in 2004 were estimated to be due to alcohol.

How do we know that alcohol causes these diseases and injuries? From ‘mere association’ in epidemiological studies - the same kind of epidemiological studies that have consistently shown lower rates of cardiovascular disease and lower overall mortality among moderate drinkers for decades.
Later in his post, Fergus says (presumably quoting Nutt):

Treating the results of excessive alcohol consumption is a huge burden for the NHS. In England alone, around 350,000 hospital admissions per year are mainly attributable to alcohol.1

Nobody is counting these people in hospitals. The 350,000 is an estimate based on attributable fractions which simply assume that a certain proportion of hospital admissions for a given ailment is caused by alcohol. So, for example, if two people die from drowning, one of them is assumed to be an alcohol-related death. All of this is ultimately derived from the ‘mere associations’ of observational epidemiology. (Last year, the Office for Health Improvement and Disparities (ne√© Public Health England) changed the methodology and the number of ‘alcohol-related’ hospital admissions dropped massively.)

There is no ‘definite proof’ that alcohol causes cancer. There is no ‘definite proof’ that smoking causes cancer, for that matter. It is practically imposssible (and unethical) to conduct randomised controlled trials to prove it either way. What we have instead is a wealth of observational evidence backed up with a plausible biological mechanism and no other reasonable explanation for the statistical associations. And that is what we have to show that moderate drinking is good for the heart and helps people live longer.

In fact, the evidence of health benefits from moderate drinking is stronger than the evidence for alcohol causing any form of cancer. There are more studies from more countries over a longer period of time and the hypothesis has been tested more rigorously precisely because people like Nutt don’t want to believe it.

I am on two mailing lists for alcohol research and barely a week goes by without a new study showing health benefits from moderate alcohol consumption. Occasionally I will tweet them, but generally I ignore them. For anyone familiar with the field, it is a non-story. The evidence is so overwhelming that it takes a huge amount of motivated reasoning to ignore it. The response from the likes of Nutt is exactly the same as the tobacco industry’s response to the evidence on smoking and lung cancer in the 1950s. They dismissed it as a mere statistical association and demanded an impossible burden of proof.

I have written plenty about the alcohol J-Curve elsewhere so won’t go over it again, but to give you an idea of the double standard at work, here are the results from a meta-analysis of light alcohol consumption and breast cancer risk, which Nutt believes to be conclusive.

Only a handful of the studies produced statistically significant results and the combined relative risk was a tiny 1.09 (1.06-1.12).

And here are the results from a meta-analysis of moderate drinking and coronary heart disease which Nutt thinks ‘has never been definitely proven’.

Here, most of the studies are statistically significant and the effect is larger. A relative risk is 0.75 (0.68 to 0.81) means the moderate drinkers are 25% less likely to die from a very common disease. If moderate drinking was a drug, they’d be prescribing it.

Incidentally, the authors of the first study found a statistically significant reduction in lung cancer risk among the light drinkers, a result that I suspect Nutt would not take seriously (and I wouldn’t blame him).

As for the claim that ‘The benefits don’t outweigh all the other risks’, here is what overall mortality looks like. The teetotallers have a 20% increased risk of premature death compared to moderate drinkers. Note also how these graphs refute the tired old cope from the ‘sceptics’ that teetotallers only die younger because many of them are sickly former alcoholics. These graphs separate never-drinkers, ex-drinkers and current drinkers.

We then move on to economics. Nutt writes:

…it’s been estimated that when you add in the costs of alcohol to society, there is a net loss to the Exchequer. This is undeniably a difficult argument to disentangle economically, and a complicated sum. But the costs of alcohol to society are relatively well established. These are: £3.5bn on health, especially hospital admissions and accident and emergency attendances; £6.5bn for policing drunkenness; £20bn for lost productivity through hangovers. The total is £30 billion.

So two-thirds of the ‘cost to society’ consist of lost productivity as a result of hangovers? As health economists tire of having to point out, lost productivity is not an external cost. If you are less productive, you get paid less and get passed over for promotion. The cost falls on you. It is not a cost to society and certainly not a cost to ‘the Exchequer’.

In any case, drinkers seem to be more productive than teetotallers and get paid more, probably because they increase social capital and have larger social networks.

As for the more relevant costs to government, some of these are real but when I looked at them in 2015 they amounted to £3.9 billion, which is barely a third of what the government gets from alcohol duty.

Fergus says:

It’s strange that this isn’t talked about more. Polls suggest that the NHS is the one institution that almost everyone in Britain cares about. So why not ease the burden on its workers – and the public purse – by reducing our alcohol consumption?

Personally, I don’t care about the NHS. In fact, I despise it. And there are good reasons why we don’t we ‘ease the burden on its workers’ by reducing our alcohol consumption. It’s because alcohol duty comfortably pays for alcohol-related healthcare costs and because the NHS is there to look after us, not the other way round.
Nutt seems to think that people in the UK are exceptionally heavy drinkers and Fergus has taken this on board, saying:

Notably, that there are enormous differences in the levels of alcohol consumption around the world. Some of this is because of religious prohibitions on alcohol in Muslim-majority countries, or because many people carry genes that mean they feel flushed and nauseous after consuming alcohol (commonly known as Asian flush). But even among countries with no religious or biological hindrances to alcohol, consumption varies a lot, and Britain and Ireland do unusually poorly on this front. 

Doing poorly means drinking a lot, as far as Fergus is concerned. But we don’t actually drink particularly heavily in the UK, as the chart below shows. Consumption in Ireland is higher but has fallen a lot in recent years.

One might assume that the British (and Irish) have always been habitual drunks. In fact, there has been significant variation in the amount of alcohol British people have consumed over the past few centuries. We’re at (or just after) the peak of a decades-long trend of more alcohol being consumed in Britain.

We’re definitely after the peak, which was in 2004, and we’re now back to the levels of the 1980s. The world wars clobbered alcohol consumption and the last decades of the 20th century saw drinking climb back to pre-WWI levels, although it has never returned to Edwardian levels.
Still, what is to be done? Quite a bit, according to Nutt.

Nutt suggests his own set of policy solutions: taxing drinks by the amount of alcohol in them and increase that tax back to 1950s levels (i.e. triple it); stop selling strong alcohol in supermarkets; make it a law that all alcohol outlets must sell non-alcoholic drinks; install breathalysers in pubs and stop drunk people from buying more alcohol; banning all alcohol advertising; and many more.

Yikes! I much prefer his work on Ecstasy, to be honest.

He focuses most on minimum unit pricing (MUP), i.e. a floor on the price at which a unit of alcohol can be sold. As the government has not raised its duty on alcohol, it now costs a third of what it did in 1970 in real terms.

It does not cost less in real terms. It costs more in real terms. The price of alcohol has gone up by more than the cost of a basket of goods, i.e. above the general rate of inflation. It has come down in relation to average incomes which means that it is more affordable - making things more affordable is the whole point of raising incomes - but the price has not fallen in real terms. This is a common misunderstanding.
As for minimum pricing, the jury was still out when Nutt was writing his book in 2019, but his high hopes for the policy have not aged well.

Scotland introduced MUP in 2018 – with initially positive results:

…the amount of alcohol bought in shops and supermarkets per person per week fell by 1.2 units (just over half a pint of beer or a measure of spirits) compared with what would have been drunk without MUP. In England over the same time, consumption increased. The biggest drop – two units a week – was in the heaviest fifth of drinkers. 

Targeting the heaviest drinkers is important because they are the worst affected; they account for the vast majority of the health costs. 

Alas, the heaviest drinkers did not play ball and the modest reduction in alcohol consumption did not yield any health benefits. As I discussed in a recent post, minimum pricing has backfired horribly. The official evaluation concluded that…

There is no clear evidence that MUP led to an overall reduction in alcohol consumption among people drinking at harmful levels or those with alcohol dependence, although some individuals did report reducing their consumption.

People drinking at harmful levels who struggled to afford the higher prices arising from MUP coped by using, and often intensifying, strategies they were familiar with from previous periods when alcohol was unaffordable for them. These strategies typically included obtaining extra money, while reducing alcohol consumption was a last resort.

And a recent study found that among the heaviest drinking men, consumption actually increased

For the 95th percentile the introduction of MUP was associated with an increase in consumption for men of 13.8 g (95% CI 5.8 to 21.5), but not for women (4.8 g, 95% CI −4.0 to 13.7).

Nevertheless, Fergus is so impressed with this book that he thinks Nutt should go even further.

The evidence here is so damning that I wonder how Nutt can still suggest, in good conscience, that we should still drink (within limits). Some of the arguments he makes in favour of alcohol are ridiculous:

A glass to hold gives us something to do with our hands in awkward situations, particularly now smoking has become so vilified.

I agree that is ridiculous. The only justification people need to drink is that they enjoy it. Same with vaping. Same with cannabis. Same with ecstasy. So long as you pay your way and don’t hurt anyone, do what you like.

Alcohol. Where would we be without it? Qatar.

And then comes the big reveal…

I don’t enjoy drinking alcohol, frankly.

Knock me down with a feather! 

I drink a little bit on some social occasions, but, following my arguments here, I’m not sure that I should. Abstention from alcohol is more or less costly for different individuals, and I am one of those fortunate enough to manage it with ease.

Whatever works for you. 

It’s possible to consume alcohol in a safe manner, with risks to your long-term health that you might consider acceptable, if you enjoy drinking it. But, if you do so, then, at the margin, you are contributing to the idea that it’s normal to drink it.

It is normal to drink it. 83% of adults drink alcohol in Britain and we’re not going to stop just to make you feel less weird.

And that norm should be broken down. Drinking alcohol should always be a choice, not a default. 

What does this even mean? When is drinking alcohol the default? In a pub, I suppose, but even there you still have a choice. If you really don’t like drinking, pubs might not be the best place for you, especially if you’re going to spend the whole time citing dodgy factoids and demanding neo-prohibitionist legislation.

I don’t mind people not drinking, I just wish they wouldn’t do it around me.

First published on the Snowdon Substack.

Monday 5 September 2022

Drinking is not smoking

From the Daily Express...

Cancer warning: Popular drink equivalent to smoking five to 10 cigarettes UK study finds 

CANCER is a terrifying prospect but one that can be mitigated against to some extent. For the first time, a UK study has found knocking back a popular drink is the equivalent of five to 10 cigarettes.

The popular drink is wine. I'm not sure why they don't just say that in the headline.

Drinking a bottle of wine per week may be like smoking five to 10 cigarettes in the same time period, in terms of cancer risk, according to a study from the United Kingdom.

The research, published in the journal BMC Public Health, should serve as a clarion call for the general public to drink in moderation.

The study was published three years ago. It was garbage then and I have no idea why it's being revived now, but since it has been, here's what I said about it at the time on the long gone Spectator Health website...

If you’ve been watching closely, you'll have seen the signs. The campaign to turn alcohol into The New Tobacco happened gradually at first but is now approaching full speed. It accelerated last year when two widely reported studies in The Lancet claimed that there is no safe level of drinking. The evidence in these studies, such as it was, didn't actually support that claim but along with the equally dubious assertion that alcohol is a major cause of cancer, it forms the bedrock of the drive to treat drinkers like smokers.

If there was a turning point, it was in January 2016 when the Chief Medical Officer, Sally Davies, pushed the “no safe level” line to the public whilst announcing the new drinking guidelines. The guidelines didn’t actually say that there was no safe level, but the small band of anti-alcohol activists who created them managed to downplay the health benefits of moderate drinking to such an extent that it only required a little more statistical chicanery to make the benefits disappear altogether.
The Lancet studies made the final push.

By August 2018, decades of evidence showing that moderate drinkers live longer than teetotallers had been debunked, at least as far as the BBC was concerned [the same researchers who made this claim produced a study this year showing that moderate drinkers do live longer after all, but the Beeb didn't cover that one - CJS 2022]. If alcoholic drinks, like cigarettes, are carcinogens that cannot be consumed safely, then the anti-tobacco blueprint of cancer warnings, advertising bans and the endgame of total eradication can be rolled out.

The most shameless example yet of the deliberate conflation of smoking and drinking was published today. A study in BMC Public Health claims that drinking one bottle of wine a week increases the risk of cancer by the same amount as smoking ten cigarettes a week (if you’re a woman) or by five cigarettes (if you’re a man). Its authors say that their intention is to provide ‘a useful measure for communicating possible cancer risks that exploits successful historical messaging on smoking’. They insist that they are ‘not saying that drinking alcohol in moderation is in any way equivalent to smoking’.

Yeah, right. The only reason this study was conducted was to generate headlines to the effect of 'drinking is as bad as smoking' - and it has worked. The 'study' might boost the impact factor of the journal that published it, but it has no academic merit. There is no reason for it to exist other than as propaganda.

Do you know what the cancer risk of smoking ten cigarettes a week is? No, because hardly anyone does it. The authors admit that ‘the risk of smoking approximately five cigarettes per day (35 cigarettes per week), [is] generally the lowest level of risk detailed in [epidemiological] studies’. There is a study that estimates the cancer risk of smoking very few cigarettes, albeit in a low quality journal, but the authors ignore it and choose instead to extrapolate from other data. To put it in plain English, they had a guess.

So much for it being a 'useful measure for communicating possible cancer risks'. They are comparing the risk of drinking one bottle of wine to a risk that most people have never thought about and which even experts cannot reliably quantify. It throws shade, not light. If this is an attempt to 'make advice simpler', as The Times says, then it has failed. Unless, of course, the advice is to not drink at all.

The authors' real aim is laid bare in the study's introduction when they assert, with breathtaking chutzpah, that: 'There is now robust evidence that low levels of alcohol intake do not provide any protective health benefits.' They cite one of the ridiculous Lancet studies as evidence for this, but it could not be further from the truth. Decades of epidemiological studies have shown a strong and consistent association between moderate alcohol consumption and lower mortality, mostly thanks to lower rates of cardiovascular disease. This finding has been tested and retested dozens, if not hundreds, of times by researchers all over the world and has always come up smiling. To claim that there are no health benefits from moderate consumption would make you odd, but to claim that there is 'robust evidence' disproving those benefits puts you in Flat Earth territory.

By contrast, the claim that there is no safe level of alcohol consumption rests on a much smaller, thinner and shakier body of evidence, mostly concerned with the effect of light drinking on breast cancer (which is why the authors' cigarette equivalents for men are half those of women's).

So, on the one hand, we have decades of epidemiological evidence backed up by biological experiments which show that moderate alcohol consumption reduces the risk of death from heart disease by 15-30 per cent, and, on the other hand, we have a slimmer body of epidemiological evidence which suggests that moderate alcohol consumption might have a small effect on breast cancer risk. It speaks volumes about the 'public health' lobby that they have spent years denying, doubting and ultimately dismissing the former finding while treating the latter as gospel.

Let us take a moment to look at the evidence. A meta-analysis of prospective cohort studies (which track people’s drinking habits and health status over a number of years and are the most reliable studies in observational epidemiology) found that drinkers were 25 per cent less likely to die from coronary heart disease than teetotallers. The 31 studies reviewed are shown below. Note that most of them show a statistically significant reduction in risk.

By contrast, here are the results from a similar meta-analysis of cohort studies looking at moderate/light drinking and breast cancer risk. Of the 25 studies, only six produced statistically significant results and the overall estimate was an increased risk of just nine per cent.

That’s breast cancer incidence. If we look at breast cancer mortality, the evidence is even less scary. For 'very light drinking' - defined as consuming up to five units a week - the risk of dying from breast cancer actually decreases, as does the risk of dying from lung cancer, as the authors of the meta-analysis note:

Very light drinking reduced the mortality of both female and male lung cancer (RR, 0.81; 95% CI, 0.69 to 0.94; I2=0.0%; n=2), female lung cancer (RR, 0.70; 95% CI, 0.56 to 0.89; n=1) and female breast cancer (RR, 0.79; 95% CI, 0.64 to 0.97; I2=0.0%; n=2). There was no significant association between very light drinking and the mortality of colorectal cancer, gallbladder cancer, prostate cancer, and hematologic malignancy.
'Light drinking' - defined as drinking up to eleven units a week, ie. a bottle of wine - is also associated with lower cancer rates in some sites and is not associated with breast cancer mortality...
.. light drinking reduced the mortality of female stomach cancer (RR, 0.65; 95% CI, 0.44 to 0.98; n=1) and male lung cancer (RR, 0.79; 95% CI, 0.70 to 0.87; I2=0.0%; n=5). There was no significant association between light drinking and the mortality of oropharyngeal cancer, esophageal cancer, larynx cancer, colorectal cancer, liver cancer, female gallbladder cancer, pancreatic cancer, breast cancer, cervical cancer, prostate cancer, and hematologic malignancy.
Only when we get to 'moderate drinking' - consuming between 11 and 22 units a week - is there any increased risk of dying from any form of cancer, although it was associated with a lower risk of dying from kidney cancer.
...overall moderate drinking was not associated with the mortality of most cancers. However, it increased the mortality of female colorectal cancer (RR, 2.51; 95% CI, 1.31 to 4.82; n=1) and female breast cancer (RR, 1.04; 95% CI, 1.01 to 1.07; I2=0.0%; n=2), while it reduced the mortality of male kidney cancer (RR, 0.46; 95% CI, 0.23 to 0.93; n=1).

The claim that there is no safe level of drinking makes no sense unless there are risks from very light drinking. We know that very light drinking confers significant health benefits to the heart, and the strongest epidemiological evidence suggests that very light drinking makes breast cancer slightly more likely to occur (+4 per cent) while making fatal breast cancer considerably less likely to occur (-21 per cent).

In the face of seemingly contradictory findings like this, you might be inclined to dismiss observational epidemiology as quackery with mathematics. You might say that meta-analyses are a way of spinning gold out of garbage. If so, I can't wholeheartedly disagree, especially when dealing with ultra-low relative risks of this order, but whatever your view of the science, you can’t have it both ways. You can’t cling to a handful of studies which show a tiny relationship between light/moderate drinking and a few (mostly rare) cancers while dismissing a much larger body of evidence showing a strong, large and clinically relevant protective association between light/moderate drinking and coronary heart disease, nor can you ignore the protective effect of alcohol on the most important health metric of them all: mortality.
The fact remains that the positive effects of moderate drinking on heart disease and other conditions exceed and outweigh the negative effects on cancer risk. To focus on the risks while ignoring the benefits is to lie by omission. 

Given that coronary heart disease is the most common cause of death in the UK, and that the risks of not drinking are more serious than the risks of drinking a little, perhaps we should put a cigarette equivalent on the heart disease risk of teetotalism. Five minutes with PubMed and a pocket calculator tell me that abstaining from alcohol is as risky as smoking five cigarettes a week. Stay safe. Cheers!

A swift half with Angela Knight

In the latest episode of The Swift Half I spoke to the former CEO of Energy UK, Angela Knight, who explained what is going on with energy prices and what can be done to bring them down. Angela is knowledgeable and non-partisan. Several people have said that they found the interview to be very informative. I hope you will too.