That new alcohol study…

Posted on August 27, 2018

(Last Updated On: August 27, 2018)

Last week, in my routine meandering the internet, I repeatedly came across headlines of,


At least that’s how it felt. It was more like “No amount of alcohol is safe.”

The study,

Alcohol use and burden for 195 countries and territories, 1990–2016: a systematic analysis for the Global Burden of Disease Study 2016

I’ve written I don’t like full denouncements of alcohol. My main experience with this in seeing personal trainers tell their clients they HAVE to stop drinking. But you need to weigh the benefits against the negatives. Such as how drinkers may be more social, deal with depression better, or have a healthier bodyweight.

Not to mention life can be a lot more interesting. As the song goes, “Everything that kills me makes me feel alive.”


“we found that consuming zero standard drinks daily minimised the overall risk of all health loss (figure 5).


This weighted relative risk curve took into account the protective effects of alcohol use associated with ischaemic heart disease and diabetes in females. However, these protective effects were offset by the risks associated with cancers


the curve changed significantly only in settings where diabetes and ischaemic heart disease comprised more than 60% of total deaths in a population.”

So that’s interesting. Granted, I’m not sure there is any big, developed country, where 60% of the deaths are from diabetes and ischaemic circulatory issues. (Best I can tell America is at 30%.) But, it is a lot, and a growing percentage.

Below, if you’re above 1, that means you have greater risk of heart disease. Below 1, lesser risk. Drinkers have less risk, until they get to 6 (!) drinks per day.

Heart disease deaths, up to a point, go *down* as people drink more. (Specific chart is for females but same trend happens for males; bottom ticks each represent one standard drink per day.)

And that may really matter considering how much those diseases kill people, and again, how much the trend is them killing more. Some, Bill Gates included, think we’ll cure cancer soon, while we’re nowhere near curing heart disease or death from the sugars. Twenty years from now, will we want people drinking???

Next, this can vary dramatically by country. After all, if you’re in a country where you’re not living long enough for heart disease to kill you, you’re not as likely to get a benefit from drinking.

SDI = Sociodemographic Index. Lower = poorer. DALY = Disability Adjusted Life Years.

Notice the mountains only go up. That is, drinkers in middle to low sociodemographic countries only have greater risks due to all those ailments.

But for a rich country:

Over a lifespan, there is a lot of benefit in drinking -that purple bar got big!- because deaths from heart disease, diabetes, and ischaemic stroke go down. It’s not quite enough to fully offset the other diseases, but it cuts a dramatic amount.


Let’s think about the drinkers those headlines were aimed at. Obviously, on balance, many grasp alcohol increases risks of all kinds. Car accident, self harm, fighting, etc.

Similarly, nobody advocates binge drinking.

Where this topic routinely ends up is in the the realm of everyday people, in rich countries, 30 years or older, who like to have some drinks on a fairly regular basis. Is that ok? Or is it analogous to smoking?

For the sake of those people, and I get this may be too optimistic for some, let’s cut out the deaths from,

  • Transport injuries
  • Unintentional injuries
  • Self harm
  • Interpersonal violence

Which causes our graph to look more like (it’s not going to look perfect but you get the idea),.

From left to right, the top mountain got lower,

Over a full lifespan, we’re pretty damn close to a perfect canceling out effect now. I bet if you took out the heavy drinkers, likely lessening cirrhosis, accounted for BMI (similar to heart disease risk, at some point, the more you drink, the heavier you get), factor in how many people also indulge in cigarettes the moment they pop some bottles, and you’re getting to a much less headline worthy finding.

-> An analysis showing eliminating heavy drinkers and smoking and you eliminate nearly all alcohol caused cancers

Of course, and again, on a population level, especially a global one, that’s not how it works. But I pretty much guarantee the authors of this study ain’t exactly trying to influence policy decisions in the third world. This was a paper done on the global population, with press releases done almost assuredly damn near exclusively in rich countries, whose average finding is not going to apply to a great deal of those rich countries populations.

-> There is also room to adjust assumptions the authors use. An enormous factor in this study is trying to parse out how much of a death is attributable to alcohol. Person X gets Y cancer, how much of that was from alcohol? Not easy. I can’t imagine it’s not highly debatable. A little adjustment one direction or the other could dramatically change the numbers.

-> Note, this is over the lifespan. As the graph shows, if you’re 30 years old, drinking has a host of increased risks, but again that’s not really what people think of what they talk about this topic. I mean, it’s not news 25 year olds who drink have riskier lives than those who don’t.

Sitting at home every weekend and never having sex is certainly safer than going out and gettin’ down…who wants that life?


Alcohol has been found to…decrease cancer too


Moderate Alcohol Intake and Cancer Incidence in Women

Notice how thyroid, non-hodgkin lymphoma, renal cell, stomach, bladder, multiple myeloma, and maybe a couple others have a lesser risk when women drink.

As far as I can tell, none of these cancers were included in original study we were talking about…


Two broader points

There has been a surprising amount of contemporary research contradicting what we’ve long believed about alcohol- that moderate consumption is fine, if not healthy. You’ve probably seen this with lots of topics. Hell, I’m even seeing salt and sugar being rethought now. When I was in college, fruit (!) of all things was being debated.

I would not at all be surprised if part of this is because researchers need to “find” something. This isn’t scientific, but if you read enough papers, sometimes you can feel when a paper is trying to justify a point of view. This paper gave me it- that the researchers wanted to find something to refute a long held belief.

One quick example. I previously quoted this,

“This weighted relative risk curve took into account the protective effects of alcohol use associated with ischaemic heart disease and diabetes in females.”

What’s weird about that sentence is it makes it seem as if the study only found a benefit for females, but if you look at their data, they also found the might-as-well-be-indentical benefit for males. A is females; B is males.

There were definitely differences in males vs females. The SDI charts being significant, but it was one of those sentences which could make you go “hm.”

Males didn’t get as much benefit from drinking as females. In large part because males are more violent. Shocking. (I used females in SDI charts earlier.)

One more quick example. The authors early on state (bolding mine),

“we did a new meta-analysis of relative risks for 23 health outcomes associated with alcohol use”

Yet from after that point, it’s this,

“The level of alcohol consumption that minimised harm across health outcomes was zero.”

Which comes across as if all causes of death were looked at. Again, they weren’t. Only 23 were. That statement should say,

“The level of alcohol consumption that minimised harm across the 23 health outcomes looked at…

and, so far as I can tell -there are thousands of pages in the appendices so I admit I could have missed it- there is no justification why they only looked at those 23…or why the cancers they looked at happen to be the ones most known to be negatively influenced by alcohol…

Last example. Remember, the authors, and the press releases, say “no,” “zero,” “nada” amount was safe, but do you see any difference between 0 and 1 drink per day in this graph?

That bar doesn’t clearly start going up until two drinks per day. Even if there is a difference there, one you need a microscope to see, is that really clinically relevant? (No.)

-> Hat tip

There is a well known reproducibility issue in research, most prevalent in anything having to do with studying humans. There are likely too many PhDs chasing too little research. While one group of researchers look at overall mortality and drinking, another group decides “Hey, we’ll only look at 23 health outcomes. That’s technically different, so it’s publishable.” Meanwhile the rest of us are left thinking “Why is it every six months I’m told I need to flip my behavior back the other way?”

When there is debate between old and new, with much of it -definitely not all- rather than lean towards the the newer research because it’s new, I lean towards the older. The newer has such esoteric statistical modeling and is often so complicated it’s hard to believe it’s not so much better as it is purposely convoluted in the aim of finding something worth printing. (The more complicated it is, the more easily the reader, and the researcher(s), are tricked.)


Again, this is not rigorous, in fact I know the below is flawed, but it’s hard to imagine there is not SOMETHING to be said for looking at this much more broadly. This was from the press release,



Highest prevalence:

  1. Denmark: 97.1
  2. Norway: 94.3
  3. Argentina: 94.3
  4. Germany: 94.3
  5. Poland: 93.8
  6. France: 93.1
  7. South Korea: 91.3
  8. Switzerland: 91.2
  9. Greece: 90.8
  10. Iceland: 90.3

Lowest prevalence:

  1. Pakistan: 0.9
  2. Bangladesh: 1.0
  3. Egypt: 1.1
  4. Mali: 2.5
  5. Morocco: 3.0
  6. Senegal: 3.2
  7. Mauritania: 3.2
  8. Syria: 5.0
  9. Indonesia: 7.2
  10. Palestine: 7.9

Is anybody concerned with Denmark’s life expectancy? Norway’s?

But, from a life expectancy standpoint, I bet nobody’s ass is yearning to live in the bottom 10.


Random questions / points

  • It’s tough to say someone should start drinking if they simply don’t want to be a drinker. However, if you have a much more serious family history of heart disease compared to cancer, could that be prescribed at all?
  • It’s tough to say moderate drinkers should stop drinking.
  • Why do college athletes, who binge drink more than regular students, live longer? If you drink and exercise does that benefit clearly outweigh the risks?
  • Anecdotally, many drinkers are heavy coffee drinkers. Coffee has over and over again been found with increasing life expectancy. Should we be controlling for coffee consumption too?
  • Portugal is one of the highest drinking countries per capita, yet has a very good life expectancy.
  • Why do so many other countries start drinking (legally) earlier than the U.S. yet live longer?
  • Think about the differences in cultures or personalities which endorse drinking vs those who banish it, who seems healthier psychologically?
    • Does trying to be too strict in your life backfire to some degree?
  • Moderate drinkers either live longer or the same as non drinkers
    • Notice how one of the authors says it’s not so much the drinking, it’s how social the moderate drinkers are. Wellllll, can you really separate the two? A lot of people aren’t social without alcohol around. Do the people who go out for a drink after work become more affluent because of the going out for the drink?
  • How many suicides / self harm injuries are avoided because of alcohol?

The common sense argument, or how I tend to view alcohol:

It’s like Vitamin D and skin cancer.

We’ve gone from all sun is terrible to everyone is Vitamin D deficient.

Common sense says some sun is good because it feels good. Too much is bad because it feels bad. Different people; different tolerance. That’s more or less what we now know. No sun is bad; too much sun also bad.

Is alcohol any different?

What I tell clients (usually in the vein of weight-loss but general health applies too),

Client “I like to have one or two drinks on a regular basis.”

Me “That’s fine. Nothing worry about.”


Client “I drink a bottle or two of wine a night”

Me “See a therapist.”


-> You could easily make an argument this line of thought does not work with smoking. I love a random cigar as much as anybody, but you’re unlikely to find anything remotely showing a benefit for even very sporadic smoking.

-> It’s fair to say this article may read like it was written by someone who wants to justify drinking, because they drink. While I used to enjoy Saturday nights quite a bit, I get out about once every six months these days -having kids does that to you- I didn’t drink until the end of my senior year in college, and have never been much of an at-home drinker.

Subscribe to Blog via Email

Enter your email address to subscribe to this blog and receive notifications of new posts by email.