Alcohol intake → Depression
Higher Alcohol intake moves Depression only in some conditions. Read the mechanism.
Three Mendelian randomization files of alcohol on depression disagree: two ADH variants in 68,486 Copenhagen adults are null for antidepressant use, UK Biobank beer and spirits instruments raise major depression until insomnia and smoking enter the model, and ALDH2 plus ADH1B in 476 Chinese adults point the other way.
Mendelian randomization · Contested · N=68486
Mendelian randomization. It's a great tool for ruling out environmental factors, though it has its own genetic limitations.
Mechanism
Observational papers have long sold a J-curve: people who drink a little look less depressed than people who drink nothing, so a glass looks protective. Sick quitters, money, sleep, and who still drinks sit on that backdoor. Researchers use a genetic shortcut called Mendelian randomization. It is good at ruling out diet, income, and the decision to stop drinking. It can still be fooled when one gene affects two traits. Wium-Andersen, Orsted, Tolstrup, and Nordestgaard genotyped 68,486 Copenhagen adults aged 20 to 100 on ADH1B rs1229984 and ADH1C rs698. The combined genotype raised drinks per day. For hospitalization or death with alcoholism, the causal odds ratio for one extra drink a day was 28.6 (6.47 to 126). That is a different pair. For prescription antidepressants the causal odds ratio was 1.11 (0.67 to 1.83). For hospitalization or death with depression it was 4.52 (0.99 to 20.5). The authors do not claim similar evidence for depression. Zhu, Chen, Si, Li, Chen, and Zhao used ALDH2 rs671 and ADH1B rs1229984 in 476 adults from fifty villages in seven Chinese provinces (mean age 49.4, 74 percent men). Depression was a CESD-10 of 12 or more (9 percent). Two-stage least squares found less depression with more drinking: drinker versus not, b = -0.127 (-0.253 to -0.001), first-stage F 44.5. Occasions in 30 days and grams a week were also negative. Sargan tests did not reject. Cutting heavy or former drinkers did not flip the sign. Qiu, Song, Cao, Tian, Cheng, Sun, Cai, and Wang ran two-sample Mendelian randomization of UK Biobank beverage intake on a major-depression GWAS of 59,851 cases and 113,154 controls. Inverse-variance weighted odds ratios were 1.284 (1.026 to 1.608) for beer and 3.837 (1.993 to 7.387) for spirits. Red wine and white wine were null. When insomnia, smoking, and contraceptive use entered a multivariable model, beer and spirits were no longer independent. Reverse Mendelian randomization did not find major depression moving beer or spirits.
Caveats
This page is adult alcohol intake and later depression or major depressive disorder. It is not alcoholism. Wium-Andersen's strong causal file is hospitalization or death with alcoholism. That pair is not published here. Wium-Andersen used two ADH variants in Copenhagen. The depression hospitalization interval includes 1. Antidepressant use is a prescription, not a structured interview. The paper is Mendelian randomization and is available here only as an abstract. Zhu genotyped 476 people. CESD-10 with a cut of 12 is a screen, not a clinic diagnosis. Women drink little in that sample. The authors name stress relief as a candidate path. That path is not measured. A flushing-response check (n=2,216) was not genotyped. Qiu's alcohol finding is univariable and beverage-specific. The threshold was p below 5 times 10 to the minus 6, not 5 times 10 to the minus 8. Beer and spirits lost the multivariable test. Cheese is a different pair and is not on this page. The European two-sample files do not travel to East Asian ALDH2. Provincial beer and spirit sales against a depression yearbook are not Mendelian randomization. Qiu and Zhu are Mendelian randomization. Sample sizes are taken from the papers. N=68,486, N=476, and the 173,005-person depression GWAS are from the papers. This is not medical advice and not a recommendation to drink, to abstain, or to treat a score.
Effect
qualitative. Wium-Andersen ADH OR 1.11 (0.67-1.83) antidepressants; 4.52 (0.99-20.5) depression hosp/death. Qiu beer IVW OR 1.28, spirits 3.84, MVMR ns. Zhu 2SLS drinker b=-0.127 (n=476).
Nodes
Named confounders
Cite this page
What Causes What. “Alcohol intake → Depression.” https://whatcauseswhat.org/edges/e-alcohol-depression (atlas updated 2026-09-02).
Papers
- Increased alcohol consumption as a cause of alcoholism, without similar evidence for depression: a Mendelian randomization study.
MK Wium-Andersen, DD Ørsted, JS Tolstrup, BG Nordestgaard · 2016 · International journal of epidemiology
- Alcohol Use and Depression: A Mendelian Randomization Study From China.
C Zhu, Q Chen, W Si, Y Li, G Chen, Q Zhao · 2020 · Frontiers in genetics
- Associations between dairy and alcohol consumption and major depressive disorder in a mendelian randomization study.
D Qiu, R Song, X Cao, Y Tian, X Cheng, D Sun, S Cai, Z Wang · 2024 · Scientific reports