Felicity Carter, in her usual brilliant way, has uncovered a neo-prohibitionist rat’s nest under the WHO.
I went digging after seeing her article and found a wee bit more regarding some of the other organizations involved. I’ve recapped and thrown in some new info in the Wine News in 5 video and article on JR.
@Adam_Frisch your post last year was on point.
As a practicing physician with training in public health, I think this article is somewhat disingenuous and dismissive of what is actually a well-studied issue.
The issue with large-scale epidemiological association studies is that it is impossible to control for all potential confounders. The paper that she herself cites as the basis for her argument itself acknowledges that -
In conclusion, while some studies have suggested a J-shaped relationship, there are limitations in the study design, confounding factors, and individual variability that challenge the generalizability and interpretation of these findings. More recent studies with advanced methodological designs have challenged the J-shape association.
Similarly, “safe” levels of alcohol are defined arbitrarily based on cutoffs that produce statistical significance when compared categorically, whereas the relationship is certainly continuous and to some extent exponential. Simply put, the more the worse, but when you go from 1 to 2 drinks a day for instance, the incremental risk is minimal. When you go from 4 to 5 drinks, the incremental risk is relatively higher. So therefore if you draw a line at a certain value and compare all the people below and above the line for differences in mortality, you are likely to see a bigger difference if the line is at a higher threshold than at a lower.
Beyond just a simple sticker outcome of overall mortality, the association with developing cancer, liver disease and so on is also relatively uncontroversial and indisputable.
From the WHO perspective, problem drinking is an issue in many low-to-middle income countries, and so I can see why they might choose to sacrifice a more nuanced message for something clearer and aimed at overall drinking reduction to address this issue. But that’s a side point.
Rather than cling to the falsely-reassuring belief that there must be a protective effect of “low” levels of alcohol consumption, it is far better for the average person to recognize that while alcohol consumption itself is directly harmful, there are many other associated benefits such as better mental health, wellbeing, cultural identity, family togetherness and so on that are not accounted for in a simple mortality outcome.
As with most things in modern life, it’s simply a trade-off. I drink wine despite knowing that it may increase my risk of cancer in the long-run, but choose to do so at a level that the incremental risk is (hopefully) low, and trade that increased risk for the happiness it brings me. At the same time, I try my best to mitigate this with other healthy behaviors such as regular exercise and a relatively healthy diet (to the extent that us wine lovers also tend to be food lovers).
The same could be said for other less controversial food exposures like sugar, salt etc which definitely increase overall risks of heart disease, mortality in the long run - but we still reach for the salt when eating because under seasoned food is no joy at all. Ignoring the issue doesn’t make it go away - but understanding it helps us continue to drink responsibly and stay happy!
Another recent controversy is that of the link between artificial sweeteners and cancer and the similar advice against consumption by the WHO last year. But these studies don’t account for the alternative - sugar - which is also linked to cancer and a whole host of other metabolic complications besides. I tell my patients to avoid both if they can, but if not, then the artificial sweeteners are still probably the lesser of two evils, but acknowledging that they still do come with increased risks overall. That’s life!
Gerald,
Thanks for your thoughtful response to a complex question. Well said.
Gerald, appreciate the insight. Have you heard or followed anything put out by Christopher Snowdon on this subject matter? He seems to be pretty confident that the J-curve is real and has written some pieces on it and the WHO as well. From what I can tell he seems to be a pretty well respected journalist too. This is a pretty complex and nuanced subject matter and so I feel like it’s pretty difficult to ever feel like I have a solid grasp on what’s actually true.
I’m not familiar with Christopher Snowdon, but have just had a quick read of the articles you’ve linked.
I don’t claim to be an expert in this field, but I would say that there is a difference between association and causation.
It’s one thing to say that perhaps having 1 drink a week is associated with lower mortality. The validity of this association is clearly disputed in the literature, and I would say that despite his protestations it is still impossible to measure for all confounders in any retrospective observational study (for instance you might control for age, gender, BMI, medical conditions, previous drinking habits. But can you control for diet? Exercise levels? Supplement use? Etc). But let’s say we assume that this J curve for mortality actually exists for purposes of discussion (which is still plausible to me and I would not strongly argue against its existence).
This still does not mean that the alcohol intake is the cause of the reduced mortality. Rather, I would sooner believe that the alcohol intake is a bystander that is associated with the true reason for the reduced mortality. Why is that?
It is a well-founded maxim (Hill’s postulates) that causality needs be demonstrated by plausibility (meaning that there must be a reasonable biological mechanism or at least a theory of one). We know how the human body metabolizes alcohol well (it is in most high school or college level biology textbooks). We know from detailed studies in animal models and humans that the metabolism of alcohol does produce harmful by-products like reactive oxidative species that are known to cause tissue-level damage that may lead to outcomes like cancer or liver cirrhosis in the long run. With the caveat that this isn’t really my field, I haven’t come across any studies that show how alcohol is beneficial to cellular functions or biological processes. There is at best some evidence that alcohol does regulate our lipid levels somewhat, which may be a mediator of purported cardiovascular protection. However, this is just an intermediate step, and also the studies are themselves somewhat confounded. It is perhaps telling that most of popular science will point not to alcohol but instead other associated compounds in wine - for instance resveratrol - as the mediator of purported benefits, even though there are unfortunately few studies that demonstrate such a fact in a convincing manner.
Conversely, moderate alcohol intake is itself associated with many factors that are associated with increased life expectancy, such as higher income levels, adequate social interactions, a lack of comorbid disease, good overall health and so on. Yes, I acknowledge that the biological mechanism of these interactions are hard to prove, but I would sooner believe that they are more likely the cause of reduced mortality, than the ingestion of a compound that is proven to be harmful on a cellular level.
All this talk about association studies also only focuses on the single outcome of mortality. A few studies also look at risk of cardiovascular disease. But the strong associations with cancer (dismissed as trivial by the author, which anyone who has had a family member with cancer knows is certainly anything but true), or liver cirrhosis are conveniently excluded. As are studies on mental health outcomes and other measures.
Putting it all together, I do get a little peeved by inflammatory language like this and additionally the assertion that moderate drinking is protective, when that statement implies a level of causality for which evidence is sorely lacking.
There are some potential real-world implications of such messaging. I for one know that it is easy to underestimate how much one will drink before one starts drinking (the author alludes to that too). It is easy for an average person to come away from reading such an article and simply take home that moderate drinking is “beneficial”, thereby enabling a higher level of intake. I would much prefer to talk about the issue in the terms of what I mentioned in my previous post - of balancing potential harms against potential benefits - instead of simply concluding that there are no harms below a certain arbitrary threshold.