Research published in the journal Clinical Nutrition demonstrates a link between more coffee consumption and greater cancer risk. However, evidence for this link was only observed for oesophageal cancer.

While there was some evidence for coffee consumption increasing multiple myeloma risk, results for other cancer types were null, suggesting no effect of increased coffee consumption on most cancers. A potential explanation is that any harmful effect of coffee consumption on cancer risk acts through physical damage to the body’s digestive system from hot liquid consumption, rather than a cancer-causing effect of coffee or caffeine.

There have been many claims of both health risks and health benefits of coffee consumption from observational studies. However, such associations are subject to bias from different sources – does coffee consumption increase or reduce the risk of disease, or is it that coffee drinkers are healthier or less healthy than non-drinkers? This is a classic question of correlation versus causation. As it is difficult to conduct long-term randomized studies of coffee consumption, it is important to investigate any potential link with cancer risk using a variety of approaches. These include so-called “natural experiments”, in which we consider factors leading to natural variability in coffee consumption levels.

A team led by researchers at the University of Cambridge and Karolinska Institute studied data on over 580,000 volunteers from the UK and Finland using a technique known as “Mendelian randomization”. Participants were compared in a statistical analysis not according to their measured level of coffee consumption, but according to whether they have inherited genetic variants predisposing them to increased or decreased coffee consumption. By making comparisons based on genetic variations rather than coffee consumption levels directly, several of the pitfalls of judging causation from observational data are avoided, as the genetically-defined groups compared are similar with respect to factors other than coffee consumption. For example, although coffee consumption is correlated with alcohol consumption, genetic predisposition to coffee consumption is not, meaning that these results cannot be explained by alcohol consumption. Hence statistical links demonstrated by Mendelian randomization are more likely to represent causal relationships than those from conventional approaches.

A genetic score associated with a 50% increase in coffee consumption was linked to a 2.8-fold increase in risk of oesophageal cancer. In UK-based individuals, this association was stronger in those who reported a drinking temperature preference for “hot” (5.5-fold increase) or “very hot” (4.1-fold increase), as opposed to “warm” (2.7-fold increase). The genetic score was similarly associated with oesophageal cancer risk in coffee drinkers as in coffee non-drinkers. However, the genetic score also increases tea consumption, so a similar mechanism may hold here as many coffee non-drinkers will be tea drinkers. The genetic association with oesophageal cancer was stronger in those who reported a preference for caffeinated hot drinks (4.8-fold increase) versus decaffeinated hot drinks (no increase), although there was substantial uncertainty in the latter estimate due to the smaller number of individuals with a preference for decaffeinated drinks.

Importantly, the genetic score was not associated with overall risk of cancer, or with the majority of specific cancer types. Oesophageal cancer is a relatively uncommon form of cancer, with around 9,300 diagnoses in UK residents per year. This represents around 2.5% of the 375,000 new cancer diagnoses in UK residents per year.

Limitations of the investigation are that the research method cannot provide direct insights into the mechanism by which cancer risk is increased, and that subgroup analyses relied on self-reported measures of coffee consumption and temperature preference. However, the impact of hot drink consumption on oesophageal cancer risk has been previously observed, particularly in countries where cultural practice is to consume hot drinks at scalding-hot temperatures. Further work is needed to understand the null finding in individuals with a preference for decaffeinated drinks. Another limitation is that genetic predisposition to coffee drinking affects lifelong consumption levels, and so we do not have suitable evidence to estimate the change in cancer risk that could occur from a short-term change in coffee drinking behaviour.

Dr Stephen Burgess, MRC Investigator at the MRC Biostatistics Unit, said:

While a link with any cancer type is unwelcome, our investigation generally provides good news for coffee drinkers, as there was no evidence supporting a causal effect of coffee consumption on any major cancer type.”

Read full paper published in Clinical Nutrition: Coffee consumption and cancer risk: a Mendelian randomisation study – ScienceDirect