High Alcohol Intake Linked to Higher Mortality in UK Biobank

Key Takeaways
- In the UK Biobank cohort, high total alcohol intake was associated with higher all-cause, cardiovascular, cancer, and other-cause mortality.
- Moderate total alcohol intake was associated with higher cancer mortality, without an increase in the other broad mortality categories, and no protective association was observed at any dose of total alcohol.
- Low-to-moderate spirits and beer or cider intake was associated with higher mortality for some outcomes, whereas low-to-moderate wine intake showed inverse associations for some outcomes, a pattern that observational data cannot attribute to wine itself.
- Higher total intake and high wine intake were associated with mortality from multiple site-specific cancers.
Investigators analyzed 340,924 UK Biobank participants aged 40 to 70 at enrollment (2006–2010), with mortality follow-up through December 2022. After exclusions (including 109,452 without complete alcohol data and 50,488 with cardiovascular disease or cancer at baseline), the analysis also excluded former drinkers (255), pregnant women, and people without follow-up data. Cox proportional hazards models related total and beverage-specific consumption, reported on a baseline questionnaire, to all-cause, cardiovascular disease (CVD), cancer, and other-cause mortality; 20 site-specific cancer outcomes were also examined, with false discovery rate correction. Models adjusted for age, sex, race, assessment region, smoking, education, deprivation, diet quality, physical activity, blood pressure, glucose, cholesterol, body mass index, a frailty index, family history, beverage preference, and drinking with meals; wine, beer or cider, and spirits were mutually adjusted. Sensitivity analyses excluded deaths in the first two years and lifetime abstainers.
Never or occasional total consumption (20 g/week or less) served as the reference for total-intake comparisons; for each beverage, the reference was never or occasional intake of that beverage. Low intake meant more than 20 g/week up to 20 g/day for men and up to 10 g/day for women; moderate meant more than 20 to 40 g/day for men and more than 10 to 20 g/day for women; high meant more than 40 g/day for men and more than 20 g/day for women. For spirits, the categories were never or occasional, low, and moderate-to-high. These are exposure definitions, not drinking recommendations. Participants were observed rather than assigned to drinking patterns.
Over a mean follow-up of 13.4 years, 22,381 deaths were recorded (4,288 CVD, 11,063 cancer, and 7,030 other-cause). Compared with never or occasional total intake, high total alcohol intake was associated with higher all-cause mortality (hazard ratio
1.24, 95% confidence interval [CI] 1.17–1.31), CVD mortality (HR 1.14, 95% CI 1.01–1.28), cancer mortality (HR 1.36, 95% CI 1.26–1.47), and other-cause mortality (HR 1.12, 95% CI 1.02–1.22). Moderate total intake was associated with higher cancer mortality (HR 1.11, 95% CI 1.03–1.20), but not with all-cause, CVD, or other-cause mortality.
The positive associations were mainly driven by beer or cider and spirits, in roughly linear dose-response patterns. High beer or cider intake was associated with higher all-cause (HR 1.51) and cancer (HR 1.66) mortality, moderate-to-high spirits intake with higher all-cause mortality (HR 1.62), and even low spirits intake with modestly higher all-cause mortality (HR 1.08). In contrast, low wine intake was associated with lower all-cause (HR 0.89), CVD (HR 0.86), cancer (HR 0.92), and other-cause (HR 0.86) mortality, and moderate wine intake with lower all-cause, CVD, and other-cause mortality but not cancer mortality; wine showed J-shaped dose-response curves, and high wine intake was associated with higher cancer mortality (HR 1.10). The inverse associations were mainly driven by red wine, while high red, white, and fortified wine intake were each associated with higher mortality for some outcomes.
Of 20 site-specific cancer outcomes, 15 showed significant associations with total or beverage-specific intake after false discovery rate correction. High total intake was associated with higher mortality from head and neck (HR 3.01), respiratory, gastrointestinal, and other cancers, and moderate-to-high beer or cider and spirits and high wine intake were associated with mortality from head and neck, respiratory, gastrointestinal, liver and gallbladder, hematological, female breast, and genital cancers. Drinking with meals was associated with 7% to 22% lower mortality than drinking outside meals, and drinking outside meals worsened the harmful associations at higher intakes. Associations between total intake and all-cause mortality were stronger in men, younger participants, non-White participants, current smokers, and participants without obesity.
These cohort associations cannot establish that alcohol intake or beverage type caused the observed differences in mortality. The authors excluded former drinkers and adjusted for frailty to reduce reverse-causation bias, but alcohol intake was self-reported at baseline only, so changes over time were not captured, and the participants were mainly White UK volunteers who may not represent other populations. Wine drinkers in this cohort had higher diet quality and were more often women and White, so the inverse wine associations may reflect residual confounding by lifestyle and do not establish a protective effect or show that switching beverages would change outcomes.
The investigators observed higher mortality across broad outcome categories with high total intake and a cancer-mortality association with moderate total intake, with no protective association for total alcohol at any dose. They concluded that alcohol from beer or cider and spirits, even at low to moderate levels, may be harmful for premature death, a conclusion that goes beyond what an observational study can establish.
Clinician Questions
What counted as low alcohol intake for men and women in the UK Biobank mortality analysis?
Low total ethanol intake meant more than 20 g/week up to and including 20 g/day for men, or more than 20 g/week up to and including 10 g/day for women. These are exposure definitions, not drinking recommendations.
How was moderate total alcohol intake defined for men and women in the UK Biobank mortality analysis?
Moderate total intake meant more than 20 and up to and including 40 g/day for men, or more than 10 and up to and including 20 g/day for women. This defines exposure, not an observed mortality effect or an intake recommendation.
Did the UK Biobank alcohol-mortality analysis use the same comparison group for total intake and individual beverages?
No. High, moderate, and low total-alcohol comparisons used never or occasional total intake (20 g/week or less) as the reference. For each beverage, the reference was never or occasional intake of that beverage, with wine, beer or cider, and spirits mutually adjusted.
This content is intended for healthcare professionals for educational purposes. The article and its image were developed with AI assistance and reviewed by members of the ReachMD team prior to publication. Please refer to the original source linked for the complete and most up-to-date information.