Thursday, 27 April 2023 14:45

Meta analysis on total mortality and drinking- Evidence-based science or ideology?

Researchers at the Canadian Institute for Substance Use Research (*) published a new meta analysis about the effects of various levels of alcohol consumption on all-cause mortality and concluded that drinking less than two standard drinks per day (<25g/d) was not associated with protection against heart disease or contributed to a longer life expectancy. Thus, disputing the well-established J curve and that wine is protective against heart disease.


Main findings


  • The current meta-analysis of 107 cohort studies involving more than 4.8 million participants did not find a significant reduction in risk of all-cause mortality for “low-volume” drinkers consuming less than 25 g of ethanol per day compared to lifetime abstainers (after adjusting for key characteristics such as age and gender).
  • These results seem to be inconsistent with the data presented in the publication’s
  • When they separated the underlying (or original) studies into older participants and younger ones, this meta-analysis found no risk associated with any drinking up to 45g/day and
  • an increase in risk among heavier drinkers across both age groups compared to abstainers and to occasional drinkers.
  • The meta-analysis found that the mortality risk appears to be different for men and women, with an increase for women at drinking levels above 24g/day and for men above 44g/day. But the risk did differ by age (even though the authors did not do an analysis to test the significance of these differences).
  • These results as well as larger increases in risk for women compared to men for certain drinking levels are consistent with the findings of previous studies.


What does it mean?

  • The current publication is an update of the authors’ 2016 meta-analysis, where 20 new studies (from 2016 to July 2021) were included. Their findings are inconsistent with most other meta-analyses conducted during past two decades.
  • The study assumptions appear to be subjective and based on the authors’ previous publications.
  • The authors’ conclusion that ”daily low or moderate alcohol intake was not significantly associated with all-cause mortality risk” does not fully reflect the study’s results. There are several limitations to this study’s design, analysis, and interpretation that warrant caution when considering its results and conclusions.



  • The authors adequately acknowledge the study’s main strengths but understate some of its limitations:
  • In many studies, the measurement of alcohol consumption is imprecise and was assessed only once
  • Underreporting is a problem: individuals report less than they actually consume, which distorts the research results and the associated health risk of alcoholic beverages (ie. if they report 1 glass of wine but they actually consume 2 glasses, the risk appears to be higher than if they would actually only consume 1 glass),
  • Few underlying studies included sufficient data on drinking patterns to fully capture their likely effect. For example, a binge drinking pattern affects the risk quite differently, even though the averaged daily drinking volume is low.
  • It is important to know that many of the studies that went into the meta-analysis do not assess certain confounding factors such as drinking patterns (only average alcohol consumption), type of alcoholic beverages and lifestyle factors such as diet, smoking, body weight, exercise)
  • Despite these limitations, the authors conclude that moderate, “low-volume” alcohol drinking was not associated with protection against death from all cause.
  • However, the figures presented in the publication and the J curve in the appendix (eFigure 4) do not support these conclusions. The data clearly show that both the non-adjusted as well as the adjusted data support a J-shaped relation between alcohol consumption and overall mortality. This means that low volume drinking (defined as 1.3 – 24 g alcohol per day, or up to two drinks) is associated with a significantly reduced risk of dying.
  • The paper does unfortunately not present any new ideas to support their data interpretation; it is just an extension of previous papers by the same group. The hypothesis put forward by the authors a long time ago (the “sick quitters” hypothesis) has been disproven decades ago, and all studies since the 1990s have excluded those who stop drinking from the non-drinker’s pool. This appears to be just be another attempt to revive their misconceptions.
  • Furthermore, the authors deliberately excluded numerous strong epidemiological studies showing a J-shaped association between drinking and mortality. They also did not acknowledge the well-described mechanisms that explain the causal relation between regular light to moderate consumption of alcoholic beverages and reduced risk of coronary heart disease and, therefore, reduced risk of death.


(*) the authors disclose funding from Canadian Centre on Substance Use and Addiction under a Health Canada grant related to the recent review of Canada’s Low-risk drinking guidelines from 2011.


The authors have taken reasonable care in ensuring the accuracy of the information herein at the time of publication and are not responsible for any errors or omissions. Read more on our disclaimer and Privacy Policy.