The Alcohol J-Curve Did Not Survive Better Methods
Studies that separated lifelong abstainers from former drinkers found the protective dip largely disappeared.
Dr. Elena Vance
Chief of Research
Sep 27, 2026 • 10 min
What the data says
Abstainer bias
Grouping former drinkers with lifelong abstainers made the abstainer reference group look unhealthier than it is.
Source: Reanalyses of alcohol cohort data
The famous curve suggesting moderate drinking beat abstaining came from cohorts whose non-drinking reference group included people who had quit because they were unwell.
Reanalyses that isolated lifelong abstainers, and Mendelian randomisation studies using genetic variants affecting alcohol metabolism, found the protective dip shrink or vanish.
The remaining picture is a dose-dependent increase in several cancers with no clearly protective range, alongside real social and psychological effects that cohorts cannot capture.
This is not a temperance argument. It is that anyone drinking for health reasons is acting on a finding that no longer stands.
Editorial note: this article summarises published research for general information. It is not medical advice. Discuss any intervention with a qualified clinician.
