No Safe Level: What the Suppressed Federal Alcohol Study Actually Found

The study that landed in the Journal of Studies on Alcohol and Drugs on June 9 is not reading the tea leaves or extrapolating from imperfect data. It is a systematic review commissioned by the Substance Abuse and Mental Health Services Administration, conducted over roughly two years by a federally-convened expert panel, and peer-reviewed by the journal before publication. Peer review, not suppression, is the appropriate quality control for science. That it passed is the relevant fact.

Here is what it found.

There is no level of alcohol consumption associated with net mortality benefit

The study examined all-cause mortality across levels of alcohol consumption and found that the J-curve — the shape of evidence that appeared to suggest light drinking was cardioprotective — does not survive rigorous methodological control. Earlier studies showing cardiovascular benefit from light drinking suffered from several documented confounders: the “sick quitter” problem (abstainers include former heavy drinkers who quit due to illness, making them look sicker than moderate drinkers); failure to disaggregate lifetime abstainers from former drinkers; and inadequate control for socioeconomic variables correlated with both drinking patterns and health outcomes.

When those confounders are controlled, the protective effect disappears. What remains is a dose-response relationship that runs consistently upward from zero: any consumption raises certain disease risks, and that risk increases with volume.

One drink per day shifts your risk profile

At a single standard drink per day — the level commonly marketed as “moderate” and previously described in dietary guidance as acceptable — the study found elevated risk of death from liver cirrhosis, esophageal cancer, oral cancer, and injury-related causes. The absolute risk increase at one drink per day is small. But it is present, and it is measurable, and there is no threshold below which it disappears.

This matters for clinical communication. For years, the framing used in provider-patient conversations has been: “Light to moderate drinking is probably fine, and may even be protective for your heart.” That framing is not supported by the current evidence base. Providers counseling patients who drink at low levels should now communicate that the evidence does not support a safe level, even as they acknowledge that absolute risk at low consumption is modest.

The risk curves for women are steeper

The study’s analysis of sex-stratified risk found that women reach elevated all-cause mortality risk at lower consumption thresholds than men. At fourteen drinks per week for men, the lifetime mortality risk from alcohol-related causes is approximately one in twenty-five. Women face comparable or steeper risk curves at lower weekly volumes. This is consistent with established pharmacokinetics: women have lower body water percentage and different alcohol dehydrogenase enzyme concentrations, producing higher blood alcohol concentrations per drink and longer clearance times.

The clinical implication is that drinking guidance that treats all adults the same — a standard that previous guidelines sometimes implied — understates the risk for women.

What the methodology does and doesn’t prove

A systematic review is not a randomized controlled trial. It synthesizes existing evidence, and the evidence it synthesizes is observational. The study cannot prove that alcohol causes any particular death — only that population-level risk is elevated at any consumption level when confounders are properly controlled. Critics of the study’s findings will note this limitation, and they are correct to do so. But the limitation applies in both directions: the absence of randomized controlled trial evidence does not prove alcohol is safe. The precautionary logic runs toward the conclusion the study reached.

The suppression of this research did not improve the scientific evidence base. It removed rigorous findings from official channels and replaced them with guidelines that said “less alcohol for better overall health” — language specifically chosen to avoid triggering clinical alarm or labeling changes that would have followed from a direct statement of what the research showed.

The suppression of this research did not improve the scientific evidence base.

For providers and recovery navigators: the evidence is now in the peer-reviewed record. The Dietary Guidelines for Americans are a policy document, not a scientific one. Base your counseling on the science. The science says there is no level below which alcohol is without risk, and that the risk is real even at levels widely considered safe.

That’s what was in the study. That’s what Washington didn’t want in the guidelines. And that’s what the House Appropriations Committee, this Thursday, voted to prevent researchers from ever studying again.