A systematic review published in Public Health in 2026 examined all studies on alcohol and cardiovascular outcomes released over 2.5 years—from January 2023 through July 2025—to assess their methodological rigor.

What they did

Eight publications met the criteria. For each, researchers evaluated how the control group was constructed, whether known sources of bias were accounted for, and whether authors discussed limitations in their own study design.

What they found

Of the 8 studies:

  • 5 reported protective effects of low to moderate alcohol consumption
  • 3 found no protective effects

The key difference lay not in the datasets but in methodology.

The five "positive" studies compared drinkers against non-drinkers or lifetime abstainers. This introduces bias: among people who don't drink at all, there's a disproportionate number who quit because of illness or poor health. When you compare a moderate drinker to such a group, the moderate drinker automatically appears "healthier"—not because alcohol helps, but because the control group is already sicker on average.

Only one of the eight studies even mentioned this problem. None applied methods to assess or correct for it.

The three studies with neutral or negative findings took a different approach:

  • they excluded non-drinkers from comparison, working only with drinkers of varying intensity
  • or they used Mendelian randomization (MR)—a method that substitutes genetic variants for self-reported consumption

Mendelian randomization is one of the few ways to approach causal inference in observational research: a person doesn't choose their genes, so genetic predisposition to lower consumption isn't confounded with other behavioral factors. These three studies found no protective effects.

One claimed benefit the review authors call outright "implausible": a 40% reduction in cardiovascular events from consuming less than one glass per week compared with non-drinkers. No known physiological mechanism could explain this.

The authors also noted another issue: some studies could present their data as either positive or negative—researchers chose which angle to emphasize in their main conclusions.

What this means

This isn't an indictment of specific scientists. Alcohol epidemiology has several well-known methodological traps, and most studies don't account for them. Or they do, but don't mention it in the text.

So I'd propose two questions to ask of any such study:

  • Who are they comparing? If the control group is lifetime abstainers or simply "non-drinkers" without distinction, the result is questionable.
  • Were confounding correction methods applied? Mendelian randomization, exclusion of former drinkers who quit due to illness—if not, confidence in a protective effect should be low.

Or put simply: who were they actually comparing against?

Caveats

Eight studies over 2.5 years is a small sample for broad conclusions. The review authors didn't assess conflicts of interest or funding sources for the original studies. There may be the authors' own bias toward rigorous methods. Still, the claim of an "implausible" 40% benefit at less than one glass per week is hard to dispute without explaining the underlying physiology.

Source: Public Health, doi:10.1016/j.puhe.2026.106331