Hacker News Reader

Comment by nostromo | Hacker News Reader

We now know that isn't true, especially not for heart disease. This study, as one example, separated out "never drinkers" from "former drinkers" to address this bias and found drinking to still be beneficial.

https://bmcmedicine.biomedcentral.com/articles/10.1186/s1291...

I can't help but feel bias against drinking has religious roots. This has been studied to death, and generally moderate drinking is shown to be beneficial, or at very least not harmful. But some folks are so opposed to that result they keep moving the goal posts.

View on HN

On the individual level the science is still murky, but we do know that people from cultures that have a glass of wine in the evening tend to be healthier and live longer that those who don't. That said, I don't necessarily think the magic is in the antioxidants in wine, nor do I think it has something to do with the cardiovascular effects of having alcohol in the bloodstream.

My complete layperson theory is that no amount of alcohol is "healthy" per-se: its negative effects are well known and documented, and almost assuredly outweigh the few dubious positive effects. Instead, the benefits of drinking small amounts of alcohol might be entirely tied to stress management. Stress kills. It absolutely ages people and leads to early death.

Having a wind-down ritual in the evening is probably the important bit. A glass of wine with dinner is a firm punctuation mark in a person's day. It signals that work is now over, and what's been left undone can be resumed tomorrow.

That's exactly how I appreciate a beer after work every couple of days. A personal ritual to relax, taste a different - if possible craft - beer, and unplug from work. One small bottle (35 cl) is enough to make me happy ;)

>> This study, as one example, separated out "never drinkers" from "former drinkers" to address this bias and found drinking to still be beneficial.

Careful. That result is not so clear-cut:

>> In the few studies that excluded former drinkers from the non-drinking reference group, reductions in risk among light-to-moderate drinkers were attenuated. [Section Abstract/Results]

>> Pooled analysis of estimates relative to non-current drinkers showed a reduced mortality risk for an alcohol intake up to approximately 75 g/day. However, when studies with former drinkers in the reference group were excluded, the association was considerably weakened (see Additional file 1: Figure S9). In addition, among those studies using post-event alcohol measures, the result did not change substantively; a similar trend was seen in studies with multiple measures but failed to reach statistical significance, probably because of the low number of curves (n=2) in this subgroup (see Additional file 1: Figure S10). [Section Alcohol consumption and all-cause mortality among CVD patients]

Btw, "post-event" means that participants reported their results after they had their major cardiovascular event:

>> In addition, most of the included studies asked patients to report their average consumption since the occurrence of their primary events (post-event alcohol assessment), whereas three studies used alcohol intake in the year prior to primary events (pre-event), assuming drinking habits remained stable over time, even following events [14, 44, 45]. [Section Data extraction and quality assessment].

Why is everyone so confident with their conclusions without RCTs? This article points out how deeply complicated ethanol consumption is with our biochemical systems. Without RCTs we really can't measure the overall effect of ethanol without quantifying the effects of ethanol on each of those systems.

Can't we just reserve judgement here, and use some humbleness in the face of our ignorance to drive for a real RCT funded by the public?

Can we try to avoid the trap of becoming confidence-men?

This is an impossible ask and exactly what I mean about moving the goalposts.

For example: there are zero RCTs that prove to us cigarettes cause cancer. But we're still comfortable in saying cigarettes cause cancer.

It's entirely ridiculous to suggest we do a RCT and ask someone to smoke for 30 years to remove all doubt that cigarettes cause cancer. And who would pay for it? So all we have are mountains of data showing that cigarettes shorten lifespan.

The same is true here, but the opposite result. Doing a RCT on alcohol for decades will never happen. All we have are mountains of data showing either no harm or a small benefit on lifespan. So that should be our null hypothesis: that moderate drinking is either harmless or slightly beneficial.

This is a fun one because it's a good example of how to bend the truth with the truth.

About 80 to 90 percent of people who have lung cancer were smokers.

Only about 10 to 15 percent of smokers will develop lung cancer.

So it's not a guarantee. It increases the risk.

Just like alcohol (https://www.cancer.gov/about-cancer/causes-prevention/risk/a...)

The problem with doing RCT with alcohol and with smoking is that we're talking about lifestyle studies. You can't really do a study with the goal being "see if we can give this dude cancer during his lifetime".

And let's also remember that the study that people who drink moderately had longer lifespans just really notes correlation. It does not prove a causation.

Drinking is ubiquitous. Most people drink. Drinking is also a luxury. So if you're poor, you either drink a lot or very little. Because you either have a problem or you're too broke to afford it. (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3185179/)

Guess which income bracket engages in moderate drinking the most? That's right, the income bracket that also allows you to afford better health outcomes overall. So while it may be true that moderate drinking correlates with longer lifespans, it's a knock-on effect to the fact that being poor sucks.

The issue is, the difficulty of doing an randomized study doesn’t nullify the benefits of randomization (a big one being eliminating confounding variables, for which there can be a lot for who decides to take up an addiction and why).

The control part would be particularly difficult (unless maybe you gave everyone a real or sham nicotine patch? But then you can’t exclude some other benefit from straight nicotine itself).

One could control for all the confounders they want, but you still risk missing some that are unknown or undervalued.

So your argument here is that it is ridiculous to suggest that an RCT is possible in this circumstance, so we should therefore accept observational studies which derive correlations as truth.

Not only does that conclusion not follow from your premise (a logical fallacy), but it's an absurd suggestion as a process for deriving truth. Observational studies, at best, offer an insight into possible hypotheses, and should by no means be considered persuasive unless all interacting systems have been controlled for. In the case of ethanol, there are an enormous amount of interacting systems that need to be controlled for.

Your premise is also strictly false, as it's clear that not only is an RCT for ethanol consumption possible, but it was planned and partially in progress before it was terminated.

It's truly incredible to see the kind of language that you use here, how confident you are about your fallacious argument. I kindly request that you turn down your confident language. Science is a long and careful process of pushing back the fog of ignorance, and if you are serious about the search for truth you shouldn't use such confident language.

Because only people with a high enough conviction of benefits of moderate drinking are commenting here and arguing for. However I have no idea how they do become so strongly convinced in that, given the low quality of the data that exists, the general reproducibility crisis in bio-sciences, precedent of complete compromise of a large study, things like https://slatestarcodex.com/2014/12/12/beware-the-man-of-one-..., the argument that if moderate amounts of ethanol were beneficial we would have just evolved to synthesize it, etc.

I think the most likely outcome is boring: moderate amounts of alcohol are moderately harmful, but not harmful enough to be immensely obvious and not harmful enough to outweigh the non-biological benefits some people extract from it.

I don't think it's necessarily about religion. I have nonreligious, very sciency type friends who are simply opposed to drinking. Even after I show them the studies, they do as you describe, move goalposts and desperately try to justify their beliefs in some type of science.

It's just a morality/cultural issue. They see people drinking & driving, excessively drinking and being violent, depicted IRL and in movies, and want to distance themselves from these bad types of people that drink alcohol.

It can be argued that Americas Puritan rejection of alcohol and tobacco was retained with a different justification.

Plenty of very secular countries have drink and smoke a lot (for communist countries, for example).

There was never a Puritan rejection of tobacco.

Not sure how you've reached the "religious" conclusion. With the glaring exception of Islam, alcohol plays a central role in most religions.

Obviously drunkenness is another issue altogether, but that's not what such studies examine in the first place, so that's a different discussion.

As for the benefits of wine in moderate drinkers, I remember reading a study which compared consumption of wine vs equal amounts of other types of alcohol, and concluded that the antioxidants and tannins in wine were the main carriers of that benefit, and if anything alcohol still had negative effects but simply the benefit from the other substances in wine overrode the harm caused by the moderate amounts of alcohol in wine.

Obviously since I don't remember the study to cite it, this makes it an anecdotal claim here, but ...

In the US the temperance movement that led to prohibition was motivated in no small part by evangelical Christianity.

And where did evangelicals get the idea to ban alcohol?

They didn't read in the Bible that alcohol was bad. They looked around at society and blamed a lot of its ills on problematic alcohol use.

Some background here: https://www.thegospelcoalition.org/article/9-things-you-shou...

Everybody gets this causal arrow strangely backwards. Alcohol causes people to sin; that's why Christians opposed its use. It's the negative effects of alcohol that led people to oppose its use, not "religion" per se (or even in any kind of explicit scriptural or liturgical sense).

Given that "sin" is a purely religious concept, and that one's culture around alcohol has a pretty big effect on whether alcohol is associated with misbehavior (there are lots of people and places for whom alcohol does not cause "sin"), it's totally fair to say that alcohol bans were at least partially religiously motivated.

I don’t think either of your premises are true, but it’s my fault for using the word “sin.” Alcohol is implicated in something like 50% of all homicides. There’s no culture where that behavior is tolerated and prohibition of murder isn’t a religious concept.

That’s all I meant by “sin.” I only used the religious term because I was speaking from the viewpoint of the religious of the time. Didn’t mean anything by it other than ordinary misbehavior.

Alcohol is implicated in something like 50% of all homicides.

Perhaps, but P(A|B) is drastically different from P(B|A). In other words, what percentage of consumed drinks of alcohol are associated with a homicide?

Beyond that, could alcohol use be prevalent enough that alcohol could be implicated in 50% of everything? It's sort of like how "excessive speed" is a factor in most car accidents, but that's because posted speed limits are too low and everybody speeds at least sometimes.

Hinduism/Buddhism/Sikhism has no role for alcohol as far as I am aware.

The New England Journal of Medicine study that came out in the early 2000s (I don't have a citation for you) concluded the opposite: It's the alcohol that kept the arteries clear. Their study was motivated by the observation that the cadavers med students work on are typically homeless or strongly disadvantaged alcoholics. The bodies were often disease-ravaged, but their blood vessels were in very good shape.

I think of it as an "alcohol as a plaque solvent" model.

Another way to see it is that if you starve and don't eat well and try to store as sugar in your liver (but you can't because the liver has exit the room) you don't have much remaining fat in any part of your body, including the blood vessels.

They probably considered that.

Beneficial implies casualty, which is clearly not established in the association study you linked.

This meta study you link literally cites 6 papers authored by our main villain Mukamal.

Is this "no one reads the article anymore" again?

> I can't help but feel bias against drinking has religious roots.

I don’t think it’s necessary to propose novel explanations for opposition to drinking. Alcohol has enormous negative effects on society and individuals. Nobody who has ever lived with an alcoholic is even remotely confused about this, but even those of us who have been lucky in that respect can easily observe large societal detriment. A staggering percentage of all violent crimes are committed by drunk people, most of us know someone who died in a drunk-driving related incident, and we’ve all been around drunk people who we found to be incredibly annoying (and often actually violent).

Opposition to drinking doesn’t require additional explanation.

The majority of drinking is done in a pro-social manner, and you're missing a large part of the picture when you say "Alcohol has enormous negative effects on society and individuals" as if it's a blanket negative impact. I agree that it does have a substantial negative component of course.

There was a (ineffective and violent) prohibition on drinking a century ago. Some religions also have strong views against drinking. To ignore these societal influences on western culture and call it a novel explanation is a bit of a stretch.

This is a bit of a non sequitur. The benefits of drinking are obvious! And, yes, religions and political parties have opposed alcohol...because the harmful effects are equally obvious. The causal arrow runs that direction. There was indeed a prohibition on drinking a century ago, spearheaded mostly by women who were tired of their husbands beating them when drunk.

Prohibition is widely regarded today as an obviously ridiculous blunder, but that's mostly revisionist and ahistorical.

German Lopez at Vox did a couple good pieces on what people get wrong about it:

* https://www.vox.com/2015/10/19/9566935/prohibition-myths-mis...

* https://www.vox.com/the-highlight/2019/6/5/18518005/prohibit...

There's also a bit of an asymmetry that makes weighing the pros and cons difficult. If 40 people enjoy a night at the bar, but one one of them goes home and beats their wife and another kills a family in a drunk-driving accident on the way home, do we just say, "well, 38 of the 40 peacefully enjoyed their evening, so it's mostly good." That looks like a horrific outcome, to me.

A critical fact that destroys modern narratives about the supposedly naive rationales behind Prohibition is that the 18th Amendment was only supposed to apply to what we now call hard liquor, not to beer and wine. Indeed, much of the country already had similar or identical legislation yet beer and wine was perfectly legal in most (all?) of those localities. The Federal prohibition was simply for national consistency with a public policy that was well tested.

Unfortunately, an overly pedantic Supreme Court interpreted "intoxicating liquors" very broadly, resulting in a national ban that far exceeded the state and local legislation with which most people were accustomed and comfortable. It was the narrower definition only encompassing distilled liquor that the public knowingly gave their consent.

This isn't quite accurate either. It was the Volstead Act - an act of Congress - that defined the breadth of the ban (i.e. what constitutes intoxicating liquors). The Supreme Court did uphold the legality of the Volstead Act.

Noted. Thank you. And for more context, the Volstead Act passed with supermajorities in Congress, enshrining a war-time prohibition (based on war-time powers) that was already in effect. So there were indeed some very strong political dynamics pushing for strict alcohol prohibition, but arguably not enough alone to get from inception to a constitutionally enumerated power. Perhaps the better lesson is that there is no simple narrative.

Is there any data that points to causation?

Is there any reason to believe the chemicals in alcoholic drinks would boost your health?

My intuition is it's behavioral or emotional in that maybe drinking a glass of wine a day is relaxing, maybe more likely to be drinking with friends?

I’ve heard a number of theories, here are a few:

Alcohol thins the blood, much like aspirin, which at small doses decreases all cause mortality.

Some types of alcohol, especially wine, are rich with antioxidants and flavonoids.

Alcohol may decrease stress and anxiety.

Alcohol is often consumed with friends and family, and we know socialization is correlated with better health.

Thank you. I have not heard of thinning blood that is interesting! Kind of on that topic I have read a lot about inflammation casual relationship to lots of problems. As I'm getting older in my sport I take a lot more Ibuprofen than I used too ;) maybe going daily isn't such a bad idea?

Yes 100% on socialization that was my first intuition.

Do you really think the bias against drinking (that might or might not have religious roots) is stronger than the massive bias that comes from the vast majority of people enjoying alcohol a lot?

Of course we 'know' nothing for sure - which is a truism that applies both to the original comment and to yours... but my point is that you should rarely be convinced by a single study.

That said, I have no specific criticism on the paper you refer to. It might be that the lower risk of heart disease is outweighed by an elevated risk for e.g. cancer? In that way both results can be true.

"Religious roots", hah. The Catholic church was opposed to teetotallism, as were lots of conservative denominations. Modern Teetotalism started as a secular movement out of dismay at the visible damage of alcohol use, just like socialism started out of outrage against industrial labour exploitation, and the sanitation movement started out of literal physical disgust at foul smells. The common thread wasn't religion, but progressivism: seeing society change immensely around them due to technological process, they believed (far more than we do today, ironically) that society could and should be consciously reformed.

And of course, what those "studies to death" you mention have established is that alcohol is even nastier than suspected, causing birth defects and cancer like no comparable substance we consume, in addition to all the obvious negative health effects we all can see.

The supposed dip at the low end, for only a few of the nasty effects, is a counter-theory to distract from the (pink?) elephant in the room. And even that counter-theory has been discredited by mendelian randomization studies.

Feels somewhat disingenuous to combine the data for men and women, not sure about that discrepancy between surveys either: https://i.imgur.com/BVJvRx6.png

Reply on news.ycombinator.com