Regular alcohol consumption is commonly assumed to promote uneven truncal fat accumulation, colloquially referred to as a “beer belly”. Enhanced imaging techniques (e.g., MRI, CT, or DXA) are required to verify whether this abdominal fat is superficial or intra-abdominal/visceral fat mass (VFM).

Alcohol is the second-most energy-dense ‘macronutrient’ (after fat) [1], contributing substantially to daily caloric intake in heavy drinkers. Strong associations between alcohol consumption and measures of total fat have been reported [2,3,4,5], yet its effect on specific fat depots remains unclear, limiting explanations for the regional fat accumulation attributed to heavy drinking.

Ethanol, or its metabolites, may directly influence adipose tissue by both inhibiting lipolysis, and providing substrates for de novo lipogenesis [1, 6]. Acetaldehyde, the primary metabolite of ethanol, may also stimulate the hypothalamic-pituitary-adrenal (HPA) axis, promoting a pseudo-Cushing’s syndrome, that evokes truncal adiposity [7, 8]. In extreme cases of heavy drinking, multiple symmetric lipomatosis (Madelung’s disease) can be seen, supporting an alcohol-specific impact on regional fat distribution [9]. Finally, recent evidence suggests that acetate, generated by hepatic ethanol oxidation, may induce histone acetylation in distant tissues, profoundly altering global transcriptional control [10], though this mechanism has not been explored in adipose tissue.

Several studies examining alcohol and body composition have used imprecise measures of VFM. Large-scale observational studies often rely on conventional anthropometrics (e.g., waist circumference, waist-to-hip ratio) [2,3,4], which cannot accurately distinguish between visceral and subcutaneous fat. Furthermore, the strong trend of increasing BMI with greater alcohol consumption tends to weaken associations after BMI adjustment [4].

Most studies employing more precise imaging methods (CT, MRI or DXA) have been small [11], or focused on heavy drinkers/alcoholics [12, 13]. The larger studies commonly faced similar limitations, such as exclusively studying one sex, limiting their generalisability to the wider population [5, 14].

To overcome the issues of sample size, population representation, and precise VFM quantification, we assessed the relationship between alcohol consumption and VFM in the Oxford BioBank, which is a population-based cohort (n = 5761) with DXA-measured VFM.