May 30, 2022

expert reaction to study looking at sweetened and unsweetened coffee consumption and mortality

A study published in Annals of Internal Medicine looks at sugar-sweetened, artificially sweetened, and unsweetened coffee consumption and mortality.

Dr Annette Creedon RNutr, Nutrition Manager at the British Nutrition Foundation, said:

“The researchers collected up to five 24-hour dietary recalls from participants asking whether they drink coffee and if so, whether they added sugar or artificial sweeteners or neither of these to their coffee drink.

“It is not clear whether any information was gathered about the choice of coffee drinks (other than ground, instant or decaffeinated) and whether they may have, for example, had flavoured syrups already added to their coffee which may already contain sugar or sweeteners. It is important that messages relating to the potential health benefits of coffee should be clear about the types of coffee that the presented findings relate to.

“Coffee consumption habits reported are almost 10 years old and habits may have changed since then, not only in relation to the types of coffee but also the average volume of coffee consumed, and this may be relevant in terms of health and all-cause mortality associated with coffee drinking. However, Liu and colleagues acknowledge that respondents may have underestimated the volume of coffee they consumed. Other dietary factors that weren’t reported on may have also changed over the study period and this could have had an influence on the outcomes of the study. It should be noted that the amount of added sugar reported in this study was very low and may not reflect the amount of sugar that people typically add to their coffee.  

“Overall, the study does not indicate a need to change our average coffee consumption habits, but it is important to remember that coffee with sugar added to it, including in the form of flavoured syrups, will be adding unnecessary calories to the diet and potentially damaging teeth too.”

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

“This is a pretty carefully conducted study of its general type, but I’m really not sure that it has made much advance in our knowledge of the effects of coffee drinking on health. It looks quite reassuring that whether you sweeten your coffee (with sugar or with artificial sweetener) doesn’t seem to make much difference to the pattern of association between coffee consumption and the chance of dying in the next few years, so maybe that’s one less thing to worry about. But I’m not confident that the study is capable of showing that beyond doubt.

“There’s an overall issue that the study is observational. It’s difficult to conduct a study of the effects of any food or drink consumption on health, over a substantial period of time, that isn’t observational. To do that, you’d have to allocate people to particular diets – say, different amounts of coffee each day – and then somehow ensure that they drink the required coffee amounts every day for many years. That’s hardly every practicable. So instead, in studies like this, the researchers just record the type and amount of coffee that people say they drink daily, then follow them up for a long time (seven years on average in this case), and record who dies. The problem is that people who drink different types and amounts of coffee also differ from one another in terms of many other factors. Some of these other factors might also influence the chance of early death. If so, these other factors, known as potential confounders, might be what cause differences in death rates between people who drink different amounts or types of coffee, and not the coffee at all. Of the potential confounders could mask a difference between coffee types, so that it could look as if adding sugar makes no difference to death rates when really it does.

“It’s possible to make statistical adjustments that aim to remove the effect of potential confounders and give a clearer picture of what’s going on, in terms of cause and effect, and these researchers did just that for a wide range of factors. But that process isn’t perfect. You can’t adjust for factors on which you have no data, and there can be questions about how to make the adjustments. Overall, that’s why this study can’t do more than give an indication that, possibly, it doesn’t make much difference to you chance of death in the next few years whether you sweeten your coffee or not. The researchers, rightly, make this inevitable doubt explicit.

“There are some more subtle issues arising from the statistical adjustment. For instance, the accompanying editorial points out that the participant in this study come from a country, the UK, where people do drink tea. The researchers were perfectly aware of this and made a statistical adjustment to allow for tea consumption. How this adjustment works is that comparisons of death rates, between people who drink different amounts of sweetened or unsweetened coffee, are essentially being made on the basis of what would happen if they all drank the same daily amount of tea. But we know from data in the study report that, unsurprisingly, people who drink coffee regularly drink less tea, on average, than do those who don’t drink any coffee. Without the adjustment for tea consumption, any difference in death rates between coffee drinkers and those who don’t drink coffee could be caused by the difference in tea consumption, and not directly have anything to do with coffee. The adjustment accounts for that possibility, in part at least. But if people see these new results, and change how much coffee they drink or whether they add sugar, that might also lead to a change in how much tea they drink, so that the adjusted comparison wouldn’t apply to them because their tea consumption has changed too. It’s really rather difficult to go beyond “it’s complicated”.

“Other complications arise too. On the face of it, the study looks impressive because there were over 170,000 participants – but that’s not the most crucial number. What’s really important in a study of mortality is how many people died during follow-up, and that’s considerably smaller, at 3,177, or under 1 in every 50 participants. (Not surprising given that they were middle-aged and were followed up for only about 7 years on average.)  That’s enough to give useable estimates of comparisons of death rates during follow-up, but there’s quite a lot of statistical uncertainty in those estimates. So, in a group of 1,000 people, like the people in this study who didn’t drink coffee, about 19 would die during the average 7 year follow up period, on the basis of these results. If we could somehow find another group of 1,000, who are just like the first 1,000 in terms of all the factors that were analysed (same ages, same tea consumption, and the same in a great number of other ways), except that they all drank about three daily cups of unsweetened coffee, then the study results indicate that somewhere between 12 and 16 might die during the 7-year follow-up. That’s less than the 19 who didn’t drink coffee, but it’s not that much less, the reduction isn’t very accurately estimated. And we just can’t be sure that the reduction is caused by the coffee drinking, because it might instead be caused by some combination of other factors that weren’t adjusted for. The corresponding figure for people drinking similar amounts of sugar-sweetened coffee would be between about 11 and about 19 deaths in 1,000 people – perhaps no reduction at all compared to the non-coffee-drinkers, or perhaps quite a sizeable reduction. (The error bounds are wider than for people drinking unsweetened coffee, because there were fewer people in this study who drank sugar-sweetened coffee than drank unsweetened coffee.)

“There are also questions of whether people’s coffee consumption changed over time – if it did, the study might not have picked the change up, because half the people studied provided only one records of coffee consumption during the entire study, and only about a quarter of them provided more than two records. And I do wonder about how exactly measured the various categories of people are. The accompanying editorial points out that the average amount of sugar added to a cup of coffee by the study participants wasn’t large – about a teaspoonful, on average, or around four grams, whereas the amount of sugar in caramel macchiato from Starbucks in the US is about 15 grams, nearly four times as much. The researchers point out that participants may not have assessed very accurately how much coffee they drank daily – do you know how big your home coffee mugs are, in millilitres, because I don’t? And I do wonder about milk in coffee too. Milk does contain sugars – according to my milk bottle, 100ml of milk contains almost 5 grams of sugars. I believe a latte from a coffee shop could typically contain about four times as much milk as coffee, so there can be a substantial amount of sugars in it even if no sugar or other sweetener is added by the customer. (The UK Starbucks website says there are about 12 grams of sugar in a tall caffe latte, for instance.) I’m not sure that this question of adding milk was taken into account in this study – the researchers did make adjustments for total milk consumption, but it seems to have been measured in terms of ‘drinks’ of milk rather than milk added to other drinks.

“Maybe those details aren’t all important – but they do indicate that, despite the large numbers of participants and the careful statistical analysis, we still can’t be at all sure exactly what’s going on. I’m not planning to change my coffee consumption, or start putting sugar or sweeteners in it, on the basis of these results.”

Dr Duane Mellor, Registered Dietitian and Senior Teaching Fellow, Aston Medical School, Aston University, said:

“This study looking at the British Biobank data which collected food and drink data from in this analysis just over 170,000 individuals (of a total of over half a million in the whole study) by asking what they consumed on up to 5 different 24 hour periods at the start of the study. It then looked at if coffee drinking was associated with death on average 7 years later. The findings suggest that drinking 1-3.5 cups of coffee is associated with a lower risk of dying over a period of 7 years. The authors also suggest that artificially sweetened coffee might not have the same association with a lower risk of dying. This is a little confusing and it may be that it’s harder to interpret this relationship as there were fewer participants who consumed coffee with sweeteners than those who drank coffee without any sweeteners or sugar. It also appears that those who consumed unsweetened coffee did drink slightly more of other drinks with sweeteners (such as diet soft drinks) than those who added sweeteners to their coffee, which makes it harder to fully interpret the effect of sweeteners in this study.

“The suggestion that this may be due to antioxidant effects, as suggested in the journal is not completely correct, as although these effects have been seen in laboratory studies in test tubes, the available does of these compounds in 1-3.5 cups of coffee would not be enough to have that effect in a human. Whilst it is possible that there may be a biological effect of some of the compounds in coffee, most likely through interactions with the gut microbiome or by regulating enzyme activity in the liver, it is also plausible moderate consumers of coffee may share other traits, including being more relaxed about how they live their lives, including what they eat and drink. A combination of the biological and behavioural effects are perhaps the best explanation of any reduction in associated risk of death over the 7 year period of the study analysis.

“This type of study is limited as it is unable to show any causal link between coffee and risk of death, and can only suggest any link over the period participants were followed up for. Additionally, there is possibility of confounding factors, as coffee habits were only assessed at the beginning of the study and other factors can influence choice of sweetener in coffee, for example men were more likely to have sugar in their coffee and other studies have shown artificial sweetener users may have a lower quality diet, consuming more highly processed foods and are more likely to have tried to lose weight. These along with other factors could explain the unclear results seen in coffee drinkers who used artificial sweeteners.

“This study supports the results seen in a number of other studies which suggest that moderate coffee intake may be beneficial, but it is important to remember what else you put in it e.g. syrups, cream etc as well as what you might have with it, this study did not look to see if dunking a biscuit in your coffee was good or not, let alone having a slice of cake with your preferred type of coffee.”

Prof Gunter Kuhnle, Professor of Nutrition and Food Science, University of Reading, said:

“This is a large observational study based on the UK biobank. The data for outcomes are very good, but data on diet are limited – in this instance, they have not many details about the type of coffee consumed which can have an overall impact on health. Indeed, whether coffee is Robusta or Arabia, filtered or extracted by steam makes a big difference to the potential health effect of coffee. However, the overall consensus from this study and the existing body of research is that coffee consumption is not detrimental to health and there might even be some benefits.

“However, what about sugar and sweeteners? This study tried to find out whether they matter – but the results are far from clear. People who use sweeteners tend to be more likely to be overweight or obese and/or to have diabetes, but it is very likely that this is the very reason they use sweeteners; a reverse causality. The amount of sugar added to coffee is difficult to estimate, even when using teaspoons. It is therefore very difficult to make any further inferences about whether the sugar added to coffee has an impact to health. However, it might be indicative of a lifestyle that favours other sweet foods – a marker of a ‘sweet tooth’ – and thus might be associated with an overall higher risk to health. Indeed, the authors find a J-shaped relationship between sugar-sweetened coffee intake and the risk of premature death, where 2 cups of sugar-sweetener coffee a day is associated with a drop in the risk of premature death but after that increasing sugar-sweetener coffee intake is associated with an increase in the risk of premature death – but it needs a lot more data to find a useful explanation for this.

“In summary, the study is informative and interesting, but does not justify any recommendation to change behavior.

‘Association of Sugar-Sweetened, Artificially Sweetened, and Unsweetened Coffee Consumption With All-Cause and Cause-Specific Mortality. A Large Prospective Cohort Study’ by Dan Liu et al. was published in Annals of Internal Medicine at 22:00 UK time on Monday 30th May 2022.

DOI: 10.7326/M21-2977

Declared interests

Dr Annette Creedon: “Funding to support the British Nutrition Foundation’s charitable aims and objectives comes from a range of sources including membership, donations and project grants from food producers and manufacturers, retailers and food service companies, contracts with government departments; conferences, publications and training; overseas projects; funding from grant providing bodies, trusts and other charities. Further information about the British Nutrition Foundation’s activities and funding can be found at http://www.nutrition.org.uk/aboutbnf/”

Prof Kevin McConway: “I am a Trustee of the SMC and a member of its Advisory Committee.  My quote above is in my capacity as an independent professional statistician.”

Dr Duane Mellor has previously worked with the International Sweetener Association.

Prof Gunter Kuhnle: “No conflicts of interest.”