Coffee's Sweet Spot? What Large UK Studies Really Say About Three Cups, Five Cups, and Your Genes

Coffee headlines tend to arrive with the confidence of a barista calling out an order: three cups are good, five cups are bad, and your genes decide the rest. The real evidence is more interesting and less tidy. Two large UK Biobank studies found that moderate coffee intake was associated with lower risks at several stages of chronic disease, while one of them also found higher risks among people drinking more than five cups a day. These were observational findings, not proof that coffee itself caused either outcome [1].

The crucial detail is that the popular 48% figure and the warning about more than five cups came from different studies. They examined different disease groupings and different transitions through illness [1], [2]. Blending them into one smooth dose-response curve creates a stronger story than either paper actually supports. Here is what the researchers found, what remains uncertain, and how to think about your own mug without turning a population study into a prescription.

What the newer study found

The newer analysis, led by Yanghong Xue, followed 331,158 UK Biobank participants who were free at baseline of the four disease groups being studied: type 2 diabetes, cardiovascular disease, cancer, and chronic respiratory disease. During follow-up, 60,616 participants developed a first noncommunicable disease, and 9,938 later progressed to multimorbidity, meaning at least two of those disease groups. Researchers used a multistate model to examine movement from disease-free health to a first disease, then to multimorbidity and death [1].

Compared with non-drinkers, people reporting fewer than three cups a day had about a 15% lower rate of progressing from a first disease to multimorbidity. At the other end, more than five cups a day was associated with a 26% higher rate of that transition. Among people who had reached multimorbidity, more than five cups was also associated with a 19% higher rate of death. These are relative differences in modeled rates, not statements that 15 or 26 people out of every 100 will avoid or develop disease [1].

The infographic's phrase 'people who already had one chronic disease' needs a small correction. Participants entered the study without the specified diseases. The higher-risk result applied to the transition after some participants developed a first disease during follow-up. That distinction matters because the paper was studying a disease trajectory, not simply comparing coffee drinkers who were sick at enrollment [1].

Where the 48% figure came from

A separate study led by Xujia Lu analyzed coffee and tea intake in 188,091 UK Biobank participants who had no cardiometabolic disease at baseline. Here, cardiometabolic multimorbidity had a narrower definition: having at least two of type 2 diabetes, coronary heart disease, and stroke. Over a median follow-up of about 11.7 years, moderate intake showed the lowest observed risk. People reporting around three coffee drinks a day had a hazard ratio of 0.519 compared with non-drinkers, which is the source of the often-quoted 48% lower association [2].

But this study did not show that the benefit reversed beyond five cups. Its highest coffee category, more than 4.5 drinks a day, still had a hazard ratio of 0.660 compared with non-drinkers. Three drinks looked best within the categories tested, but higher intake remained associated with lower new-onset cardiometabolic multimorbidity in that analysis [2]. The 'reversal' belongs to the broader, newer Xue study and to progression after a first chronic disease, not to the 188,091-person result.

Why the studies are not necessarily contradictory

The two papers asked related but different questions. Lu and colleagues focused on diabetes, coronary heart disease, and stroke. Xue and colleagues also included cancer and chronic respiratory disease, then modeled later transitions and mortality. Different outcomes can produce different curves. Coffee might correlate one way with the development of a first cardiometabolic condition and another way with progression in people who have already developed disease. Differences in intake categories, statistical adjustments, and participant subsets matter too [1], [2].

There is broader observational support for a middle-range sweet spot. A major umbrella review found that coffee was more often associated with benefit than harm, with the largest estimated reductions for several outcomes around three to four cups a day [3]. Still, the reviewers emphasized the need for robust randomized trials. Coffee drinkers differ from non-drinkers in smoking, sleep, occupation, diet, income, health awareness, and many other ways. Statistical adjustment helps, but it cannot guarantee that coffee caused the difference.