Booze is a leading cause of death and disease worldwide, and no amount is devoid of risks, according to a massive meta-analysis on global alcohol use.
The study, published late last week in The Lancet, appears to contradict some health advice that suggests low-to-moderate drinking is fine and may even provide health benefits. But its important to keep in mind that the study is focused on risk at the population level—individual, absolute risks can still be quite small, even nearly negligible, depending on how much you drink.
For the study, hundreds of researchers collaborated to lump together 3,992 estimates for relative risks of alcohol drinking. That is to say, they combined estimates of how drinking increases a persons risk of a particular potential harm—such as being injured in a drunken accident or developing throat cancer—relative to someone who does not drink or drinks less. Those estimates were distilled from 592 different studies. The researchers also amassed data on alcohol exposure from 694 different sources, extracting 121,029 data points. This helped the researchers estimate how much men and women in 195 countries might actually be drinking, broken out by age.
Stiff drinks
This massive analysis had some notable advantages over earlier work that tried to glimpse the totality of alcohols effect. For one thing, the researchers tried to minimize the complicating factor of the so-called “sick quitter” hypothesis. This is the concern that studies assessing the risks of alcohol by comparing the health of drinkers to non-drinkers may inadvertently include non-drinkers who quit because of health problems. These ailing former imbibers could collectively drag down the health of the non-drinking group, making drinkers and drinking appear healthier, or so the hypothesis goes. To try to avoid this or at least account for the problem, the new analysis included more specific reference-group data and age categories.
For assessing how much people drink in different places and age groups, the researchers turned to alcohol-sales data as well as drinking-survey data. One unique feature of their analysis is that the researchers went the extra step to try to estimate how much alcohol in a particular place went to boozy tourists, which could otherwise lead to an overestimate of how much residents actually drink. They did this by including data on tourism and travel. On the flip side, the researchers also used some studies to try to estimate otherwise “unrecorded” consumption—drinking from illicit alcohol production or home brewing. Without these sources factored in, the analysis could underestimate drinking in a given area.
The researchers then re-scaled the data to fit estimates of drinking by age group in various areas and among the sexes. They then calculated dose-response curves for the risks of 23 alcohol-associated health problems, including cancers, injuries, and less-direct issues, such as tuberculosis.
While the analysis includes an impressively large set of data and rigorous attempts to control for complicating factors, at the end of the day, it is still a meta-analysis. These types of analyses rely on repurposing published data from disparate studies, which are conducted by different researchers at different times and places. However exhaustive the controls, they are always vulnerable to the pitfalls of publication biases and poor data quality, and they always have caveats.
For instance, in this study, when the researchers estimated the risks of motor vehicle injuries and deaths related to alcohol use, they did so using data only from the US—it was just the best data available. The researchers admit they assumed that the patterns of risks seen in the US are the same everywhere else. But thats just their assumption.
Also, one of the 23 alcohol-associated harms included in the study is death from tuberculosis. This is an association that has been studied extensively—it has its own meta-analyses even. But, the absolute risks an individual has of dying of tuberculosis obviously varies a lot by location, regardless of alcohol. More than 95 percent of tuberculosis deaths occur in developing countries, according to the World Health Organization.
Drowning our sorrows
With the data, limitations, and caveats, the researchers came to the estimation that 32.5 percent of people worldwide were drinkers in 2016. That corresponds to about 2.4 billion people, aged between 15 and 95, broken down to 1.5 billion men and 0.9 billion women. But prevalence dramatically varied by location. Denmark poured in with the highest prevalence, with 97 percent of males and 95 percent of women drinking. On the other end of the spectrum was Pakistan, where only 0.9 percent of men and 1.5 percent of women knocked back drinks.
Meanwhile, Romania took the lead for the highest average drinks per day by men—with a mind-numbing 8.2. Ukrainian women came in first in this category, with 4.2 average drinks per day.
For this study, the researchers considered a drink to be 10 grams of pure alcohol, standardized regardless of what type of alcohol people drank. That corresponds to about 3.4 ounces of 13-percent wine, 12 ounces of 3.5-percent beer, or one ounce of 40-percent liquor.
All that drinking led to 2.8 million deaths in 2016. That places alcohol use as the seventh leading risk factor for death and disability in 2016. Among those aged 15 to 49, alcohol use was the leading cause of risk-attributable disease burden globally. For drinkers aged 50 and over, the biggest concern was cancers. Alcohol-related cancers explained 27 percent of alcohol-related deaths in women and about 19 percent of alcohol-related deaths in men.
Based on dose-response curves, the researchers found that to minimize the risk of all the potential harms from alcohol, a person should have “zero standard drinks daily.”
The researchers concluded:
Taken together, these findings emphasize that alcohol use, regardless of amount, leads to health loss across populations. Although we found some protective effects for ischaemic heart disease and diabetes among women, these effects were offset when overall health risks were considered—especially because of the strong association between alcohol consumption and the risk of cancer, injuries, and communicable disease.
Were all gonna die
The researchers behind the analysis argue that any notion that alcohol drinking has benefits “needs revising” and that to reduce alcohols burdens, policies and health guidelines should aim to lower population-level consumption.
But, while the analysis spills a grim mess for our health collectively, some argue that it doesnt necessarily make a strong case for teetotalism as individuals.
In a statement, David Spiegelhalter, Winton Professor for the Public Understanding of Risk at the University of Cambridge, noted that:
According to data provided by the authors but not published in the paper, to suffer one extra alcohol-related health problem, around 1,600 people would need to drink two drinks totaling 20g (2.5 units) of alcohol a day for a year. This is equivalent to around 32 standard 70cl [700 milliliter] bottles of gin over a year, so a total of 50,000 bottles of gin among these 1,600 people is associated with one extra health problem. This indicates a very low level of harm in moderate drinkers and suggests UK guidelines of an average of 16g a day (2 units) are very low-risk indeed.
The lead researcher on the meta-analysis is Emmanuela Gakidou, a professor of global health at the University of Washington. In an email to The Washington Post, she put things in another way for individuals thinking about their own risks. At one drink a day, she notes, a persons risk of developing one of the 23 alcohol-related health problems increases by 0.5 percent. She called this “a small increase in risk.”
That risk wasnt lost on Spiegelhalter, who went on to say:
Given the pleasure presumably associated with moderate drinking, claiming there is no safe level does not seem an argument for abstention. There is no safe level of driving, but governments do not recommend that people avoid driving. Come to think of it, there is no safe level of living, but nobody would recommend abstention.
The Lancet, 2018. DOI: https: 10.1016/S0140-6736(18)31310-2 (About DOIs).
[contf] [contfnew] 
Ars Technica
[contfnewc] [contfnewc]
