Acute Alcohol Use Raises Suicide Risk More Strongly in Women, Meta-Analysis Finds

Drinking shortly before a suicidal act appears to raise the odds of that act occurring, and the effect is not the same for men and women. That is the central conclusion of a large new meta-analysis published in BMC Psychiatry, in which researchers pooled data from eleven observational studies covering 352,634 individuals. The team, led by Yangyang Wang and colleagues based at the Second Affiliated Hospital of Zhejiang University School of Medicine and Hangzhou Normal University, set out to answer a question that has long been difficult to pin down with precision: does acute alcohol use, meaning drinking in the hours immediately preceding a suicidal event, carry a different weight of risk for females than for males, and does the association vary depending on the method of suicide involved?

The scale of the underlying evidence base is one of the study’s strengths. The investigators systematically searched PubMed, Embase, and the Cochrane Library from their inception through May 2025, following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses, the internationally recognized reporting framework known as PRISMA. They included observational studies that examined suicidal behaviors occurring after acute alcohol use and that reported odds ratios with 95 percent confidence intervals for overall suicide risk and for specific suicide methods. The protocol was pre-registered on PROSPERO, the international prospective register of systematic reviews, under registration number CRD420251054210, a step that helps guard against selective reporting after results are known.

Statistically, the pooled picture was clear even if the certainty of the evidence was not high. For females, acute alcohol use was associated with roughly a doubling of the odds of suicide risk, with a pooled odds ratio of 2.04 and a 95 percent confidence interval spanning 1.24 to 3.36. For males, the pooled odds ratio was 1.62, with a confidence interval of 1.24 to 2.11. In plain terms, both figures indicate that drinking acutely before a suicidal event is associated with increased odds of that event, but the numerical estimate was higher for women than for men. The researchers combined estimates using fixed-effects models when heterogeneity statistics met pre-specified thresholds, specifically when the heterogeneity P value was at least 0.1 and the I-squared statistic was 50 percent or below, and random-effects models otherwise, which is the standard approach for deciding how to pool studies whose results vary in magnitude.

That variation, in fact, was extreme. The I-squared statistic, which describes the proportion of variability in observed effects that cannot be explained by chance alone, reached 97.8 percent for the female estimate and 98.7 percent for the male estimate, with heterogeneity P values below 0.001. Values that high signal that the individual studies differed enormously from one another, whether because of differing populations, settings, definitions of acute alcohol use, or measurement approaches. Consistent with that, the research team applied the Grading of Recommendations Assessment, Development, and Evaluation system, known as GRADE, and rated the overall quality of the evidence as very low for these headline estimates. The GRADE framework judges evidence across domains including risk of bias, inconsistency, indirectness, and imprecision, and very low ratings mean the true effect could plausibly be substantially different from the pooled point estimate.

Subgroup analyses added texture to the headline numbers. When the researchers separated suicide attempts from suicide deaths, the association with acute alcohol use was markedly stronger for attempts. Among males, the odds ratio for suicide attempts was 6.82, with a confidence interval of 1.71 to 27.1, and among females it was 3.62, with a confidence interval of 1.67 to 7.86, both rated very low quality. Geographic patterns also emerged. The strongest associations were observed in North America, where the pooled odds ratio was 2.44 for males, with a confidence interval of 1.49 to 4.00, and 2.62 for females, with a confidence interval of 1.54 to 4.46, again with very high heterogeneity and very low GRADE ratings. The authors caution, implicitly through these ratings, that such subgroup findings should be read as signals for further study rather than definitive effect sizes.

Perhaps the most distinctive contribution of the analysis lies in its examination of suicide methods. Acute alcohol use was significantly associated with an increased risk of suicide by drowning, with a pooled odds ratio of 1.28 and a confidence interval of 1.04 to 1.58, and notably this estimate showed comparatively modest heterogeneity, with an I-squared of 46.2 percent, and was rated low rather than very low quality. By contrast, no significant association was found between acute alcohol use and suicide by jumping, hanging, or firearm use. This method-specific pattern is biologically and behaviorally interesting. Alcohol impairs judgment, motor coordination, and swimming ability, and intoxication near water could plausibly increase lethality in drowning scenarios, whereas methods such as hanging and firearms are highly lethal regardless of intoxication status, leaving less room for alcohol to shift the odds.

The robustness of the findings was assessed with the standard toolkit of meta-analytic quality control. Sensitivity analyses, which re-run the pooled estimates while excluding individual studies or varying modeling assumptions, indicated that the results were robust. Funnel plots and Egger’s test, a statistical regression that detects asymmetry in the distribution of study estimates, found no significant publication bias, meaning there was no strong statistical signal that small studies with dramatic results were being preferentially published while null findings languished unpublished. These checks matter because meta-analyses inherit the weaknesses of their component studies, and observational designs in particular cannot establish that alcohol causes the suicidal act, only that the two co-occur more often than expected.

That causal caveat deserves emphasis. Observational studies of acute alcohol use and suicide typically rely on postmortem blood alcohol concentration measurements, toxicology screens, or proxy reports, each of which carries measurement limitations. Blood alcohol concentration degrades after death, and the absence of detectable alcohol does not prove abstinence in the preceding hours. Reverse causation is not the concern here, since drinking preceded the event by definition, but confounding is: people who drink acutely in a crisis may differ systematically from those who do not in ways that independently affect risk, including psychiatric history, substance use disorders, and access to lethal means. The very low GRADE ratings reflect exactly these structural uncertainties, and the authors’ own framing of the evidence quality is a model of appropriate restraint.

Why might the association be numerically stronger in women? The study itself does not establish a mechanism, but several hypotheses circulate in the literature on sex differences in alcohol and suicide. Women generally have lower average alcohol consumption and lower rates of alcohol use disorders, so acute intoxication may represent a more marked deviation from their baseline state, potentially amplifying disinhibition and impulsivity effects. Pharmacokinetic differences, including differences in body water composition and alcohol metabolism, mean equivalent doses produce higher blood alcohol concentrations in women on average. Social and psychiatric factors, including higher rates of depression diagnoses among women, could also interact with intoxication differently. None of these explanations is tested directly by the pooled data, and the authors appropriately frame the finding as underscoring the importance of sex-specific interventions rather than claiming a biological cause.

For suicide prevention, the practical implications are nonetheless concrete. If acute alcohol use roughly doubles the odds of suicidal events, and if the association is strongest for attempts and in North American settings, then clinical and public health interventions that address drinking in moments of crisis become a plausible target. Crisis services, emergency departments, and alcohol treatment programs could incorporate screening for suicidal ideation when acute intoxication is present, and means-restriction counseling could take account of the drowning-specific signal. The authors conclude that these findings underscore the importance of sex-specific interventions targeting acute alcohol use in suicide prevention efforts. Given that the evidence is rated very low quality for most estimates, the appropriate response is not alarm but investment: better-designed prospective studies with standardized definitions of acute alcohol use, consistent measurement of blood alcohol concentration, and sex-stratified reporting would allow future syntheses to convert these suggestive pooled odds ratios into findings solid enough to guide policy with confidence.

Subject of Research: Sex-specific associations between acute alcohol use and suicide risk, including suicide methods, in observational studies

Article Title: Sex-specific association of acute alcohol use with suicide risk: a meta-analysis of observational studies

Article References: Wang, Y., Yin, N., Li, J., Yao, X., Zhang, J., & Yang, M. (2026). Sex-specific association of acute alcohol use with suicide risk: a meta-analysis of observational studies. BMC Psychiatry. https://doi.org/10.1186/s12888-026-08531-x

Image Credits: AI Generated

DOI: 10.1186/s12888-026-08531-x

Keywords: acute alcohol use, suicide risk, meta-analysis, sex differences, suicide attempts, suicide methods, drowning, observational studies, GRADE evidence quality, BMC Psychiatry, suicide prevention, odds ratio