Five cups of coffee a day tracked lower cirrhosis risk

Coffee has repeatedly been linked to lower rates of liver disease, but most of that literature rests on self-reported drinking and hard clinical endpoints, with little in between to show what is happening inside the liver. Questions have also lingered about whether the benefit is really caffeine, whether decaffeinated coffee counts, and whether adding sugar or sweeteners cancels the effect. Because coffee is a near-universal habit, even a modest and genuinely causal association would matter at population scale. A study that links intake patterns to imaging and blood-protein measurements as well as disease outcomes can at least test whether the story is internally consistent.

The authors analysed 354,957 UK Biobank participants who had no cirrhosis or hepatocellular carcinoma at baseline. Coffee consumption, type — caffeinated or decaffeinated — and additives such as sugar or sweeteners were captured by questionnaire, and incident cirrhosis, hepatocellular carcinoma and liver-related mortality were identified through linked health records over a median 13 years of follow-up. In a subcohort of 28,961 people who underwent magnetic resonance imaging, hepatic fat was measured as proton density fat fraction, along with liver iron and fibroinflammation using iron-corrected T1. Proteomic profiling with Olink assays was available in 44,633 participants. Models were adjusted for demographic, behavioural, metabolic and genetic covariates.

Higher coffee intake showed a graded reduction in liver outcomes. In the group drinking 5 or more cups a day, the hazard ratio was 0.68 (95% confidence interval 0.58-0.79) for cirrhosis, 0.53 (0.34-0.83) for hepatocellular carcinoma and 0.58 (0.45-0.74) for liver-related mortality. Protective associations were similar for caffeinated and decaffeinated coffee and persisted among people adding sugar or artificial sweeteners, though additive use correlated with modestly higher iron-corrected T1. Imaging showed lower hepatic fat, iron and fibroinflammation with higher intake. Proteomics pointed the same way: higher hepatocellular synthesis and complement proteins including transthyretin, selenoprotein P and complement factor H-related protein 4/5, and lower fibrogenic and macrophage-activation markers such as microfibril-associated protein 4, colony stimulating factor 1 receptor and ectonucleotide pyrophosphatase/phosphodiesterase 2.

This confirms the direction of earlier coffee-and-liver research and adds two layers it largely lacked: imaging of the organ itself and circulating protein signatures that align with the clinical findings. The similarity between caffeinated and decaffeinated coffee argues against caffeine being the whole explanation, which complicates the simplest mechanistic account. The signal that additives correlated with modestly higher fibroinflammation on imaging is a hint rather than a demonstrated harm, and the authors' framing of unsweetened coffee reflects that. What no observational cohort can resolve is whether coffee itself protects the liver or whether coffee drinkers differ in ways that adjustment cannot fully capture — that would need randomised evidence.

This is a reason to think regular coffee drinking is at least not a liver liability, and possibly part of a favourable picture, with the association tracking upwards through higher intakes. Because decaffeinated coffee showed similar associations, people who avoid caffeine are not obviously excluded from whatever is going on. The finding on sugar and sweeteners is the softest part of the analysis: the protective pattern held, but imaging markers of fibroinflammation were modestly higher in people using additives. None of this is a prescription to change your intake, and nobody should read a hazard ratio from an observational cohort as a promise about their own liver.