Food and diet

Tea, coffee and brain cancer

“Caffeine helps cut brain cancer risk,” said the Daily Express, reporting that a daily cup of tea or coffee can stop tumours growing by restricting blood flow to the brain.

The large study behind this story followed 410,000 men and women across 10 European countries for 8.5 years and looked at the development of two forms of brain tumour. Brain tumours are rare, and during follow-up there were only 588 new cases in total. The researchers found no meaningful associations when they looked at each country separately, although combining all the national results showed a trend between greater caffeine consumption and lower cancer risk.

The results echo those of a previous study and are likely to lead to further research into how caffeine may affect the working processes in the brain. However, the research has numerous important limitations, including the fact that variable methods were used for measuring caffeine intake across countries. Overall, while this research is of scientific interest, it has limited implications for current medical treatment.

Where did the story come from?

The study was carried out by researchers from Imperial College London, and numerous other academic institutions in Europe and the US. Sources of funding were not reported. The study was published in the peer-reviewed American Journal of Clinical Nutrition.

This study has several important methodological limitations, which the papers have generally not reported on.

What kind of research was this?

This was a cohort study designed to examine the association between coffee and tea intake and the risk of developing glioma and meningioma, which are types of brain tumour. Gliomas are tumours of glial cells, which protect nerve cells, while a meningioma is a tumour in the meninges, which are the protective cells surrounding the brain and spinal cord. A similar US study recently noted an inverse association between caffeine intake and glioma, i.e. glioma became less common as caffeine intake increased.

This was a cohort study, which is a design used to assess the effects of an exposure on an outcome and to provide evidence to answer the question of whether one thing causes another. Cohort studies are not perfect, and the limitations of this particular study included difficulty with accurately quantifying the amount of coffee and tea someone drinks, and also the fact that brain tumours are rare, so a very large number of people must be followed over a long period of time to record new cancers.

What did the research involve?

This research involved participants from the European Prospective Investigation into Cancer and Nutrition (EPIC) cohort, an ongoing study conducted in 10 European countries and including 521,448 men and women. The participants were mostly aged between 25 and 70, and recruited between 1991 and 2000.

At the start of the study, blood samples and body measurements were taken and the participants completed a health and lifestyle questionnaire. The questionnaire recorded information on diet during the previous 12 months, and specifically coffee and tea intake. Total consumption of coffee and tea was estimated in mL per day, with methods of assessment differing between countries. The methods used were not specifically reported in the research paper.

Over an average of 8.5 years of follow-up the researchers used population registries, health insurance records and cancer mortality registries (depending on country) to determine development of brain tumours. In their statistical analyses researchers excluded all cancers present at study start and participants who were missing data on diet or missing laboratory information on the microscopic structure (histology) of the cancers that developed. Analyses between coffee and tea consumption and brain cancer risk were adjusted for age, BMI, smoking and education.

What were the basic results?

After exclusions the final cohort included 410,309 men and women. During follow-up there were 343 new cases of glioma and 245 cases of meningioma. 

Daily coffee and tea intake was highly variable across European countries, with the highest coffee consumption reported in Denmark (798mL/day) and the lowest in Italy (98mL/day). For tea, the highest consumption was reported in the United Kingdom (532mL/day) and the lowest in Spain (6.2mL/day). Higher consumption of coffee and tea was generally associated with slightly older age, higher education, current smoking and lower BMI.

The researchers split the participants into five different groups (quintiles) based on their consumption levels of tea, coffee and both drinks together. From the lowest (first quintile) to highest intake (fifth quintile), no amount of coffee, tea or combined coffee and tea was associated with either type of brain cancer.

When the researchers looked at each country separately there were no significant associations between the cancers and drinking more than 100mL of coffee and tea per day compared with drinking less than 100mL per day, although there was a non-significant association between drinking more than 100mL and decreased risk. However, when the researchers combined the results for all of the countries they found that drinking more than 100mL per day was associated with 34% reduced risk of developing glioma compared with drinking less than this (hazard ratio 0.66, 95% CI 0.44 to 0.97).

How did the researchers interpret the results?

The researchers conclude that in this large cohort they observed an inverse association between total coffee and tea consumption and the risk of glioma. They say that this was consistent with the findings of a recent study.


This research has strengths, particularly its size and duration: it followed 410,309 men and women for 8.5 years, allowing a reasonable time for brain tumours to develop. However, while it found a trend towards an association between higher coffee and tea consumption and lower risk of glioma, careful consideration should be given to the possible shortcomings of this study:

  • Brain tumours are rare cancers. As reported in the journal article, the age-adjusted incidence rates for brain and nervous system tumours in Europe range from between 4 and 6 cases per 100,000 person-years for women, and between 6 and 8 per 100,000 person-years for men. In this study, from a population of 410,309 followed for 8.5 years, there were only 343 new cases of glioma and 245 cases of meningioma.
  • Dietary questionnaires varied between countries, for example, some asked participants to specify intake of caffeinated and decaffeinated drinks, while others assessed coffee only and not tea. The questions asked in each country are not reported, but asking people to quantify the volume of tea or coffee that they drink each day is likely to involve estimation. Also, the same volume of coffee is likely to cover a variable amount of caffeine depending on whether this is, for example, freshly brewed coffee, instant coffee, espresso, latte or cappuccino.
  • Methods of ascertaining new cancer cases also differed by country and the accuracy of these registers may vary.
  • Analyses were adjusted for age, BMI, smoking and education. However, as the authors rightly point out, risk factors for brain tumours are largely unknown, so it is possible that other unknown risk factors were confounding the observed association.
  • Overall, there was no association between any level of coffee and tea consumption and risk of glioma or meningioma in individual countries, although there was a non-significant trend within each for consumption of more than 100mL to be associated with lower risk than consumption of less than 100mL. It was only when the results for all countries were combined that a significant decreased risk was found for higher consumption.

These findings are of scientific interest and, as they echo the results of a previous study, they are likely to lead to further research investigating the effect that caffeine may have on physiological processes in the brain. However, they currently have limited implications for health. The adverse effects of excess caffeine on general wellbeing are well established.

NHS Attribution