A comprehensive analysis of more than 800 studies questions the idea of a universal “safe” drinking limit and found stronger evidence for alcohol-related harms than uncertain health benefits.

Research: Health effects associated with alcohol consumption: Burden of proof study. Image credit: Shchus / Shutterstock
In a comprehensive review recently published in the magazine natural healthresearchers synthesized the complex and often contradictory evidence on alcohol consumption and human health. This review analyzed data from over 800 observational studies and mapped statistically supported risk levels for 20 different diseases, including various cancers and cardiovascular diseases.
Research has found that while light drinking may slightly reduce the risk of certain metabolic and heart diseases, the habit also increases the risk of many cancers, suggesting that a universal safe drinking threshold may be difficult to justify.
Gaps in alcohol guidelines and evidence
Despite decades of research and significant public health funding, the overall health effects of alcohol consumption across a variety of diseases remain the subject of intense scientific debate. National alcohol intake (drinking) guidelines have been found to vary widely around the world, leaving consumers confused about what is a “safe” amount to consume.
Previous meta-analyses have aimed to elucidate the association between alcohol consumption and human health, but they have primarily focused on isolated diseases and have failed to account for underlying research biases. An often-cited example of this is known as the “sickness effect,” where people refrain from drinking alcohol simply because they are already sick.
As a result, the widely accepted and popular theory that low to moderate alcohol consumption (up to 20 grams of pure alcohol per day) protects against cardiovascular disease (CVD) and type 2 diabetes (T2D) is increasingly being called into question. Unfortunately, the association between different doses and a wide range of alcohol-related chronic diseases remains unresolved.
How to consider burden of proof
This review aimed to address this knowledge gap by employing the burden of proof meta-analysis framework, a highly conservative mathematical model designed to objectively quantify the strength of evidence. This study combined data from 16 systematic reviews across four major databases (PubMed, Embase, CINAHL, Web of Science) up to 2023 and included 843 cohort and case-control studies.
Studies were screened to include studies that reported relative risks for alcohol consumption and 20 specific health outcomes (‘clinical endpoints’), including 10 types of cancer, four cardiovascular diseases, and six other diseases such as tuberculosis and cirrhosis.
Statistical analyzes included a meta-regression Bayesian, regularization, and trimming (MR-BRT) model to generate relative risk (RR) functions, while systematically trimming potential data outliers and adjusting for known study design biases. A common example is past studies that failed to adjust for patient age.
This review also includes the Burden-of-Proof Risk Function (BPRF), which represents the most conservative estimate of risk. These findings were eventually translated into an easy-to-understand 0 to 5 star rating system for each specific disease.
Alcohol-related cancer and liver risks
The results of this review revealed that even low alcohol intake is associated with an increased risk of some cancers. The strongest evidence, a 5-star rating, shows an association between alcohol and other throat cancers, with this study demonstrating that 76 grams of alcohol per day was associated with an average RR of 4.24 (95% uncertainty interval (UI): 3.33 to 5.40), while conservative BPRF estimates indicate an increased risk of at least 105% across the typical exposure range.
Moderate 3-star evidence showed that alcohol is associated with a significantly increased risk of cirrhosis and chronic liver disease by at least 40% (RR 4.25, 95% UI: 1.87 to 9.66 at 61 grams per day). The data further revealed that colorectal, laryngeal, lip, and oral cancers were similarly associated with alcohol intake.

beAlcohol consumption and other pharyngeal cancers. bAlcohol consumption and cirrhosis and other chronic liver diseases. cAlcohol consumption and type 2 diabetes. dAlcohol consumption and esophageal cancer. eAlcohol consumption and ischemic heart disease. The risk curve is calculated compared to no alcohol consumption. In the left and center plots, the dark line shows the average RR (in grams per day) across alcohol consumption levels. Light and dark shading indicate the 95% UI with and without between-study heterogeneity, respectively, and the red line highlights the burden of proof function. The size of the data points corresponds to the inverse of the standard error, and data points that were cropped during the model fitting process are marked with a red x. Dashed lines represent the 15th percentile of the reference exposure and the 85th percentile of the alternative exposure. To visualize the log RR points in the left plot, we plotted each data point using the x value of the midpoint of the alternative group and the y value offset by the difference between reported and predicted log risk. In the middle plot, we raised the y values in the left plot to a power to create the RR curve. A customized funnel plot is shown on the right. The x-axis represents the residual between predicted and observed RR, and the y-axis represents uncertainty due to both measurement error and between-study heterogeneity.
Benefits of light drinking remain uncertain
Conversely, this review revealed a J-shaped or U-shaped relationship between alcohol intake and several cardiometabolic conditions. Low to moderate alcohol consumption was found to be associated with weak 2-star evidence of a 5% reduced risk of T2D and a 6.4% reduced risk of Alzheimer’s disease and other dementias (AD).
This review found that the theoretical minimum risk exposure level for T2D is 18 grams per day (RR 0.80, 95% UI: 0.67 to 0.97). However, the analysis found that at higher intake levels, the metabolically protective associations rapidly dissipated and were replaced by severe and potentially chronic health effects.
Safe alcohol thresholds remain elusive
The present review shows that the cancer risk associated with alcohol use is well documented and supported by stronger and more consistent evidence than modest and uncertain observational associations suggesting cardiometabolic benefits of light and moderate drinking. Because this data shows that even low levels of alcohol intake are associated with increased risk for some, but not all, cancer outcomes, the researchers argue against universal and arbitrary safe drinking thresholds, such as the currently commonly prescribed 20 grams per day limit of pure alcohol.
Instead, public health strategies must pivot towards clear, evidence-based messaging that communicates these growing risks. The authors also emphasized that the study results should not be interpreted as supporting alcohol consumption for health benefits, as the low-risk associations with cardiovascular disease, T2D, and dementia remain observational and uncertain. As the scientific community continues to improve its understanding of diet and disease, individuals should carefully consider these discoveries when making lifestyle choices.
Click here to download your PDF copy.

