select search filters
roundups & rapid reactions
before the headlines
Fiona fox's blog

expert reaction to two studies looking at coffee consumption and risk of death

Two studies publishing in the Annals of Internal Medicine assess the association between higher coffee consumption and reduced risk of death.


A Before the Headlines analysis accompanied this roundup.


Comment on both studies:

Dr Tim Chico, Reader in Cardiovascular Medicine / consultant cardiologist, University of Sheffield, said:

“These two studies followed many thousands of people for 16 years and compared death rates in coffee drinkers with non-coffee drinkers.  In keeping with previous studies, death rates were slightly lower in coffee drinkers than in non-coffee drinkers.

“The studies were well conducted and have the advantage of having observed thousands of people from many different countries and ethnic backgrounds, and the association between coffee drinking and mortality seems to be the same across all these populations.

“The authors of both studies are commendably cautious about whether or not coffee drinking is the cause of the reduced mortality in coffee drinkers, or whether there are other unknown differences between coffee drinkers and non-coffee drinkers that are the real cause for the differences observed.  Neither study seems to take into account the income of the people involved (although they do adjust for educational level) and since coffee isn’t cheap it is possible that non-coffee drinkers are less well off, which would be a potential explanation for some of the differences seen.

“What can we conclude from these studies?  The authors of both papers sensibly go no further than concluding that their results show that coffee drinking is not harmful.  I’m occasionally asked by patients whether they should drink coffee and these studies will help in advising them that coffee drinking is safe.

“The only way to be certain whether or not coffee might make people live longer is to force many thousands of people to drink it regularly, while preventing many thousands of otherwise similar people to never drink coffee.  A study like this is never going to take place, so we may never know the answer to this question.

“I don’t think this study should lead anyone to drink more coffee in search of a health benefit that might not actually exist.  It is useful to compare this to the proven benefits of physical activity.  A 20-min walk to a local coffee shop will definitely provide many health benefits, even if you don’t actually go in and buy anything.”


Prof. Kevin McConway, Emeritus Professor of Applied Statistics, The Open University, said:

“Both of these papers on coffee consumption and mortality are well-conducted and large, and show similar results across several different populations and ethnic groups.  As a coffee drinker myself, they do reassure me that my habit probably isn’t bad for me.  However, if I didn’t already drink coffee, I’m not sure that they would persuade me to take it up for the good of my health.  That’s because the size of the potential protective effect of coffee, in these studies, is not very large; because we can’t be sure what is causing what; and because, even if coffee drinking is somehow directly improving people’s health on average, neither study throws much light on exactly how it might do that.

“Like any observational study, it’s impossible to be sure what is causing what in these studies.  Indeed both studies demonstrate exactly what the problem is with conclusions based only on observing what people do.  In fact the people who drank the most coffee had the highest mortality rate in both studies, not the lowest, until observations other than coffee drinking and mortality were also taken into account.  This pretty clearly seems to be because people who drink a lot of coffee are more likely to be smokers, and smoking is well known to be bad for your health.  When the researchers allowed statistically for people’s smoking habits, the pattern reversed, and those groups that drank most coffee then had lower mortality than people who did not drink coffee.  On its own that’s pretty clear – but there are likely to be many other factors, apart from smoking, that are also separately associated with coffee drinking and with the risk of death, that may similarly be confusing the picture of cause and effect.  Both groups of researchers used statistical methods to allow for a wide range of other possible factors like this, but one can never be sure they allowed adequately for everything, so that we can’t be sure what’s causing what.

“Studies of this kind also cannot really say much about what it is about coffee drinking that might have a protective effect.  It appears that, whatever it is, it’s not the caffeine content, because the association between coffee drinking and lower mortality was much the same whether or not the coffee was decaffeinated.  It’s also worth noticing that the coffee consumption was measured in cups, and that the size of a coffee cup is not very clearly defined, and differs between different cultures.  In the European study, this shows up in the amount of coffee (in in millilitres) that defined the group who drank the most coffee.  In Italy, you had to drink only 138ml a day to be in that group, while the limit for that group was 750ml in the Netherlands (and 488ml in the UK).  Doubtless that has something to do with the type and strength of coffee that is drunk in different countries – Italian coffee cups are typically very small and very strong.  But maybe other effects are at play.  Perhaps there is some social aspect of drinking several cups of coffee a day, that has a protective effect on health, regardless of what is actually in the coffee?  These studies can’t tell us.

“In my view, anyone reporting on these studies should be sure to read the very good editorial that accompanies them.  It isn’t long and, as well as putting the results in a wider context, it points out authoritatively several issues in interpreting their results.”


Comment on just the European study:

Prof. Naveed Sattar, Professor of Metabolic Medicine, University of Glasgow, said:

“This is a very nice paper, written by an excellent group and very well done, but its conclusions will not lead me to start drinking coffee or to recommend people drink more coffee as a way to lessen their risks for heart disease.  Why not?  I remain unconvinced that the link between coffee and heart disease represents a true cause and effect relationship and that coffee is truly protective, regardless of how large a study suggests this.

“I say this because when some, perhaps many, people are unwell they cut their intake of coffee so that in any given population, many who do not drink coffee or drink less of it may be more unwell.  Whilst the authors have done their best (and better than most) to limit this potential bias, such a bias is very hard to fully overcome.

“I would only be convinced of coffee being truly protective on the basis of a well-conducted and long lasting randomised cardiovascular outcome trial which whilst difficult, is not impossible to conduct and the present study makes a good case for such a study.  In the meantime, I will continue to recommend and target proven risk factors with proven ways to lessen heart disease – stop smoking, better diets (low saturated fat etc.), treating cholesterol and blood pressure, and increasing activity levels.”


Prof Sir David Spiegelhalter, Winton Professor for the Public Understanding of Risk, University of Cambridge, said:

“This is a very good study, huge in size and carefully done, but on its own it can never confirm a causal relationship because it is an observational study.  If these estimated reductions in all-cause mortality really are causal, then an extra cup of coffee every day would on average extend the life of a man by around 3 months, and a woman by around a month.  Pro-rata, that’s as if that cup of coffee puts, on average, around 9 minutes on a man’s life, and around 3 minutes on a woman’s.  So perhaps we should relax and enjoy it.”


* European study: ‘Coffee drinking and mortality in 10 European countries: A multinational cohort study’ by Marc J. Gunter et al. published in the Annals of Internal Medicine on Monday 10 July 2017.

American study: ‘Association of coffee consumption with total and cause-specific mortality among nonwhite populations’ by Song-Yi Park et al. published in the Annals of Internal Medicine on Monday 10 July 2017.

Editorial: ‘Moderate coffee intake can be part of a healthy diet’ by Eliseo Guallar et al. published in the Annals of Internal Medicine on Monday 10 July 2017.


Declared interests

Dr Tim Chico: “I am a committee member and Treasurer of the British Atherosclerosis Society, a charity established in 1999 to promote UK atherosclerosis research.”

Prof. Kevin McConway: “Apart from being a regular coffee drinker, I have no relevant interests to declare.”

Prof. Naveed Sattar: “I am a life-long tea drinker.”

Prof Sir David Spiegelhalter: “No conflicts of interest.”

in this section

filter RoundUps by year

search by tag