A large prospective cohort investigation tracking hundreds of thousands of British adults over more than a decade has demonstrated that the relationship between alcohol consumption and long-term mortality risk differs substantially depending on beverage choice, total intake volume, and whether drinking takes place alongside meals. Evaluating data from 340,924 participants enrolled in the UK Biobank between 2006 and 2010 who were followed through December 2022, researchers documented 22,381 total deaths, including 4,288 deaths from cardiovascular disease, 11,063 from malignant neoplasms, and 7,030 from other conditions. The findings reveal that high total alcohol consumption was linked to a 24 percent higher risk of all-cause mortality, a 14 percent elevated risk of cardiovascular death, a 36 percent increase in cancer mortality, and a 12 percent rise in other causes of mortality when compared against participants who never or only occasionally consumed alcoholic beverages. In contrast, low to moderate intake of total alcohol exhibited no protective association against all-cause or cardiovascular mortality, while moderate total alcohol use was specifically associated with an 11 percent increase in cancer mortality.
The health impacts of alcohol use have long generated scientific debate, particularly regarding whether low to moderate alcohol intake offers cardioprotective advantages or constitutes an avoidable health hazard. Although heavy drinking has long been established as a primary contributor to worldwide morbidity and premature death, national dietary guidelines across the United States, the United Kingdom, and Canada have historically framed low to moderate alcohol consumption as relatively safe for healthy adults. However, recent international position statements have complicated public health guidance; the World Health Organization concluded there is no safe threshold of alcohol consumption, while early 2025 advisory warnings from the United States Surgeon General advised zero alcohol intake to prevent malignancies, and the American Heart Association affirmed that alcohol’s role in a healthy lifestyle remains unclear. By distinguishing beverage types, individual cancer sites, and mealtime habits, the current study sought to clarify these contradictions across a massive, well-characterized cohort.
To rigorously evaluate these associations while mitigating systematic biases that affected historical observational research, the investigators implemented strict exclusion criteria. From an initial base of 502,507 individuals aged 40 to 70 years, the research team excluded 109,452 participants who lacked complete baseline alcohol information, 90 women who were pregnant at baseline, 50,488 individuals who already possessed diagnoses of cardiovascular disease or cancer at enrolment, and 1,298 participants who lacked subsequent follow-up records. Crucially, the analysts also excluded 255 former drinkers from the reference population. In earlier epidemiological cohorts, individuals who ceased drinking due to declining health or emerging chronic illnesses were frequently grouped into non-drinking control categories, creating a well-documented reverse causation phenomenon that made moderate drinkers appear healthier by comparison. Alongside this exclusion, the analysts incorporated a comprehensive 49-deficit frailty index to adjust for underlying baseline health vulnerability.
Intake amounts were quantified through baseline alcohol-frequency questionnaires detailing the frequency and volume of specific alcoholic beverages consumed. Using standardized ethanol conversion metrics where one unit represents eight grams of pure alcohol, the researchers categorized total alcohol consumption into standardized tiers: never or occasional consumption defined as 20 grams or less per week for both sexes; low consumption defined as more than 20 grams per week up to 20 grams per day for men and up to 10 grams per day for women; moderate consumption spanning greater than 20 up to 40 grams per day for men and greater than 10 up to 20 grams per day for women; and high consumption exceeding 40 grams per day for men and 20 grams per day for women. Due to a narrower consumption range, spirits intake was evaluated across three tiers: never or occasional, low, and moderate-to-high consumption, while drinking preferences were classified by whether more than 80 percent of alcohol came from wine.
Multivariable Cox proportional hazard models evaluated mortality risks across beverage types, adjusting for an extensive panel of potential confounders. Covariates incorporated into statistical models included chronological age, sex, race, regional assessment center, cigarette smoking history, educational attainment, the Townsend Deprivation Index measuring socioeconomic circumstances, diet quality assessed via the Healthy Eating Index-2020, physical activity in metabolic equivalent task minutes per week, body mass index, systolic blood pressure, serum glucose, circulating cholesterol, family histories of diabetes, cardiovascular disease, and cancer, as well as frailty scores. When assessing specific beverages—such as beer and cider, spirits, and red, white, or fortified wine—the models mutually adjusted for the intake of other alcohol types to isolate independent associations while confirming the proportional hazards assumption through Schoenfeld residual diagnostics.
The positive associations observed between total alcohol consumption and elevated mortality appeared to be primarily driven by the consumption of beer, cider, and distilled spirits, displaying linear relationships across increasing intake levels. Compared to never or occasional drinkers, individuals reporting high intake of beer or cider experienced a 51 percent higher hazard of all-cause mortality, a 47 percent higher hazard of cardiovascular mortality, a 66 percent higher hazard of cancer mortality, and a 34 percent increase in deaths from other causes. Moderate beer and cider intake remained associated with an elevated risk of all-cause mortality by 17 percent, cancer mortality by 25 percent, and other disease mortality by 15 percent, while even low consumption was linked to a 7 percent elevation in cancer death risk. Distilled spirits exhibited even stronger linear associations: moderate-to-high spirits consumption corresponded to a 62 percent increase in all-cause mortality, a 51 percent rise in cardiovascular death, a 55 percent rise in cancer mortality, and an 83 percent elevation in other causes of mortality, with low spirits consumption linked to 7 to 9 percent higher risks across all categories.
Conversely, wine consumption displayed markedly different mortality associations across the cohort, showing J-shaped dose-response curves for both all-cause and cause-specific endpoints. Compared with never or occasional wine intake, low wine consumption was associated with an 11 percent reduction in all-cause mortality, a 14 percent lower hazard of cardiovascular mortality, an 8 percent lower hazard of cancer mortality, and a 14 percent decrease in other mortality causes. Moderate wine use likewise correlated with an 8 percent lower risk of all-cause mortality, a 21 percent lower risk of cardiovascular death, and a 13 percent lower risk of death from other causes. When wine consumption reached the high category, however, the apparent mortality advantages disappeared, and participants demonstrated a 10 percent increase in overall cancer mortality, indicating that higher consumption volumes override any favorable lifestyle or biological factors linked to lighter wine intake.
Sub-analyses examining specific wine varieties demonstrated that these favorable associations were primarily attributable to red wine rather than white or fortified alternatives. Compared with never or occasional intake, low and moderate consumption of red wine correlated with 8 to 14 percent and 8 to 16 percent reductions in all-cause, cardiovascular, and other mortality, respectively, whereas high red wine consumption was linked to a 9 percent increase in total mortality and a 25 percent increase in cancer death. White wine consumed at high levels was associated with a 16 to 25 percent higher risk across all-cause, cancer, and other disease deaths. High intake of fortified wine demonstrated dramatic increases in risk, corresponding to a 59 percent increase in all-cause mortality and a 141 percent elevation in other causes of death, with moderate fortified wine consumption associated with a 35 percent increase in other mortality.
A major focus of the investigation was characterizing site-specific cancer mortality across 20 distinct anatomical malignancies, with statistical significance established using Benjamini-Hochberg false discovery rate corrections. Across the cohort, alcohol intake showed statistically significant associations with increased mortality from 15 site-specific cancer outcomes. Participants in the high total alcohol bracket faced triple the risk of dying from head and neck cancers, alongside a 79 percent higher risk of mortality from respiratory system cancers, specifically tracheal and lung cancers, and a 48 percent increase in gastrointestinal cancer mortality, including colorectal and other digestive tract malignancies. Furthermore, moderate and high intake of beer, cider, and spirits, as well as high wine consumption, were notably linked to mortality from liver and gallbladder cancers, hematological malignancies, female breast cancer, and genital cancers in both women and men, notably including prostate cancer.
The timing of alcohol consumption relative to meals emerged as an important modifying factor in joint survival analyses. Compared to participants who consumed alcoholic beverages outside of meal times, individuals who reported consuming alcohol with meals sometimes or usually exhibited 7 to 18 percent and 12 to 22 percent lower hazards for all-cause and cause-specific mortality, respectively. In joint analytical evaluations, consuming alcoholic beverages outside of meals worsened the adverse associations linked to moderate and high consumption of beer, cider, and spirits, as well as heavy wine intake. Concurrently, drinking away from meals attenuated the lower mortality risks observed among low to moderate wine consumers, consistent with pharmacological mechanisms suggesting food intake slows ethanol absorption and reduces peak systemic acetaldehyde bioavailability.
Stratified sensitivity analyses revealed that the observed associations between total alcohol consumption and mortality differed across sociodemographic and lifestyle subgroups. Interaction models confirmed that the adverse relationships between total alcohol intake and all-cause mortality were significantly stronger among younger participants, male individuals, non-white participants, active cigarette smokers, and individuals without obesity. Baseline demographic assessments had also shown distinct lifestyle patterns among beverage consumers: high beer and cider consumers were disproportionately male with lower educational attainment and diet quality, high spirits consumers similarly exhibited poorer dietary scores, whereas high wine consumers were more frequently female, white, and maintained higher diet quality. Importantly, the main findings remained consistent in sensitivity analyses that excluded participants who died within the first two years of follow-up, lifetime non-drinkers, and when modeling competing causes of death.
Despite the study’s extensive sample size and multivariable adjustments, the authors highlighted several methodological limitations inherent to observational cohort research. First, alcohol consumption was gathered through self-reported questionnaires at baseline, making exposure estimates vulnerable to social desirability bias and recall underreporting, which typically attenuates observed risk relationships toward the null. Furthermore, drinking behaviors were not systematically re-assessed across the entire population throughout the 13.4-year follow-up, leaving temporal variations in consumption uncaptured. In addition, because the UK Biobank cohort is predominantly composed of white British residents, caution is required when attempting to generalize these risk estimates to diverse ethnic populations exhibiting divergent lifestyle characteristics or genetic metabolic traits. Finally, while investigators extensively controlled for socioeconomic, dietary, and clinical variables, residual confounding cannot be entirely eliminated in non-randomized observational designs.
Subject of Research: Prospective cohort analysis of alcohol consumption patterns, beverage types, and mortality risks in middle-aged UK adults
Article Title: Total and beverage-specific alcohol use at mid-life and all-cause and cause-specific mortality: a prospective cohort study
Article References: Li, Z., Tan, B., Desjardins, C., Wang, L., Khandpur, N., Feng, Y., Oh, H., Li, Q., Qian, F., Zong, G., Zhou, S., Voortman, T., & Chen, Z. (2026). Total and beverage-specific alcohol use at mid-life and all-cause and cause-specific mortality: a prospective cohort study. Nutrition Journal, 25(1), Article 114. https://doi.org/10.1186/s12937-026-01361-3
Image Credits: AI Generated
DOI: 10.1186/s12937-026-01361-3
Keywords: UK Biobank, Alcohol Consumption, All-Cause Mortality, Cancer Mortality, Cardiovascular Disease Mortality, Wine, Beer and Cider, Spirits, Prospective Cohort Study, Epidemiology, Dose-Response Relationship, Site-Specific Cancers, Dietary Assessment, Public Health, Drinking Patterns, Hazard Ratios
Cite Scienmag News
Phoebe Ingram. (November 20, 2026). UK Biobank Study Links Specific Alcohol Types to Differing Mortality Risks. Scienmag. https://scienmag.com/uk-biobank-study-links-specific-alcohol-types-to-differing-mortality-risks/
Phoebe Ingram. "UK Biobank Study Links Specific Alcohol Types to Differing Mortality Risks." Scienmag, 20 November 2026, https://scienmag.com/uk-biobank-study-links-specific-alcohol-types-to-differing-mortality-risks/. Accessed 20 November 2026.
Phoebe Ingram. "UK Biobank Study Links Specific Alcohol Types to Differing Mortality Risks." Scienmag. November 20, 2026. https://scienmag.com/uk-biobank-study-links-specific-alcohol-types-to-differing-mortality-risks/








