Several large meta-analyses show that for mouth, throat, oesophageal and breast cancer, risk already rises at very low alcohol intake, with no demonstrable safe threshold. With heavy drinking (4 or more drinks per day) the evidence for causality is strong and risk increases reach a factor of 5. For overall cancer risk the effect with light drinking is small and statistically uncertain, but that average masks serious elevations for specific cancer types.
There is no safe threshold for alcohol when it comes to cancer risk. That is the central finding from several large meta-analyses. The risk increases with the amount you drink, but for certain types of cancer that increase begins at very small amounts. The overall cancer risk with light drinking (up to 1 drink per day) is not statistically significantly elevated (relative risk 1.02), but that average masks large differences between cancer types.
For mouth and throat cancer and oesophageal cancer, risk is already elevated at a maximum of 1 drink per day: the risk of mouth and throat cancer is approximately 20% higher, and of oesophageal cancer (squamous cell type) approximately 30% higher compared with people who do not drink at all. Even with light drinking there is already a significantly elevated risk of colorectal cancer and breast cancer. For breast cancer specifically, risk appears to start rising from 3 to 6 drinks per week; below 3 drinks per week no elevated risk was found.
With moderate to heavy drinking (1-3 drinks per day and more), the risk for multiple cancer types rises further. At 4 or more drinks per day the increases are substantial: the risk of mouth and throat cancer is up to 5 times higher, of oesophageal cancer likewise approximately 5 times higher, of laryngeal cancer 2.5 to 2.65 times higher, of liver cancer approximately 2 times higher, of breast cancer 1.5 to 1.6 times higher, and of colorectal cancer 1.4 to 1.5 times higher. The associations at this level of heavy drinking are judged to be causal, not merely statistical correlations.
For liver cancer there is a particularly clear dose-response relationship: even at 1 drink per day the risk is slightly elevated (approximately 1.08 times), at 50 grams of alcohol per day (around 4 drinks) it rises to 1.54 times, and at 100 grams per day to more than 3 times higher. In people with hepatitis or diabetes the risk is considerably greater still.
The association with prostate cancer is less clear-cut. Spirits appear to slightly increase risk even at low doses, but results vary considerably by type of drink and study. For beer, a possibly reduced risk at low doses was even found. On the basis of available research, no clear conclusion can be drawn here.
In practical terms this means: anyone who wants to limit their risk of cancer of the mouth, throat, oesophagus or breast already faces a measurably higher chance even with very little drinking. The risk is greatest for mouth, throat and oesophageal cancer, and rises steeply as you drink more. There is no amount of alcohol at which cancer risk is entirely absent.
Based on several large meta-analyses (PMID 37905315, 25422909, 21864055, 26134046, 33412155, 32717903). The associations are predominantly correlational in nature, but for heavy drinking and multiple cancer types, causality is assumed on the basis of consistency and biological plausibility.