Is drinking coffee good for you? It’s a question that has divided the scientific community for centuries. Last month, a massive study from the UK Biobank came out in favor of drinking coffee.
Over the years there have been many publications warning the public against drinking coffee.
For example, in 1674, a group of women in London drafted a petition in favor of closing coffee houses, claiming the liquid was making their men impotent. In 1916, Good Housekeeping Magazine declared that coffee stunts growth and causes neurosis. More recently, research indicated that coffee leads to increased cholesterol and myocardial infarction. And touched with a more humorous tone, a meta-analysis conducted in 2010 found a significant correlation between coffee consumption and lung disease, despite the fact that they were not able to eliminate the confounding effects of smoking.
There have also been ample publications in favor of drinking coffee. 
A recent meta-analysis found that coffee-drinking is associated with a lower risk of heart disease. Also, when more became known about the antioxidant properties in coffee, various outlets—including the US Department of Health—started promoting coffee as a health food. Researchers, however, have raised concerns about how this affects people who have poor caffeine metabolism.
Last month, a massive new study was published in JAMA Internal Medicine that was based on data from the UK Biobank, and it found that coffee-drinking is good for you regardless of your metabolic makeup.
But is it really that conclusive? Let’s have a look at some of the details to see what this study really tells us about causality—though I think I may need another cup of coffee to digest it…
Study set up
The study’s purpose was to determine the association between coffee-drinking behavior, caffeine metabolism, and mortality. To this end, researchers selected lifestyle and genetic data from around 500,000 participants from the UK Biobank. For each participant, a ‘caffeine metabolism score’ was determined, based on genetic differences in caffeine metabolism. The participants were surveyed on their coffee-drinking behavior and then followed throughout a 10-year period to audit mortality. Mortality was recorded as ‘all-cause mortality,’ which processed mortality regardless of cause, and as ‘cause-specific mortality,’ whereby the cause fell within a category such as cancer or cardiovascular disease.
Analysis methods used
In the study, researchers used a so-called ‘multivariable-adjusted cox proportional hazard model,’ for which they used non-coffee-drinkers as a control group. The cox model estimates the effect of a specific treatment, such as drinking coffee, on overall survival during the study period. Since multiple factors are expected to influence mortality, ‘hazard-ratios’ were calculated for all factors thought to impact mortality. Some of these factors can be confounding, meaning it has an effect on both the dependent variable (mortality) and the independent variable (coffee consumption). A good example of this is when people who drink lots of coffee also smoke often, and therefore are more likely to die from smoking. The graph below shows how complex the interactions between different factors can be. In addition, as with any observational study, there is the risk of unobserved effects that are not included in the model. For example, there could be additional foods that coffee-drinkers consume which could have an effect on health. What do you think could be another variable that has been left out of this study?
What do the results seem to say?
Drinking coffee in high quantities is good for your health
The study found an inverse dose-dependent relationship between all-cause mortality and coffee consumption. This suggests that the more coffee you drink, the better for your health. But, all-cause mortality does not only look at death from diseases. Could the results perhaps mean that people who drink coffee are more awake and therefore less likely to be involved in a traffic collision?
People who drink coffee tend to live longer
Yes, drinking coffee was associated with lower mortality. Interestingly though, there was a stronger inverse association between mortality rates and coffee consumption shortly after the survey was filled out. Ok, what does this mean? This means that there could also be some reverse causality. Perhaps people who drink less coffee die sooner… or perhaps people who are dying tend to drink less coffee. Which do you think is more likely?
People with better metabolism tend to live longer
The study specifically looked at caffeine metabolism. From simply looking at the data, you could conclude that people with a better caffeine metabolism live longer. Of course, this does not necessarily make clinical sense. Though it does beg the question, what would be the likely reason for this association?
All types of coffee are healthy
Associations with coffee-consumption were observed for all types of coffee, including instant coffee and decaf coffee. The effects of decaf coffee are especially interesting since they plea for the positive health effects of non-caffeine ingredients. Based on these findings, could it be that decaf is healthier than caffeinated coffee?
Conclusion
This study adds to the growing evidence that coffee-drinking can be a part of a healthy diet. We should, however, keep in mind that this is an observational study. As in any observational study, the largest risk is that of unobserved effects. If an important effect is missed, a large sample size wouldn’t compensate for this. We should always be careful when concluding causal relationships from observational studies. Having said that, this study does make me feel a little bit safer when grabbing my fourth cup of the day.
