Long-term coffee consumption associated with reduced risk of total and cause-specific mortality
- Correspondence to: Dr Esther Lopez-Garcia Department of Preventive Medicine, Universidad Autonoma de Madrid, Avda Arzobispo Morcillo 2, Madrid 28029, Spain;
The effect of coffee consumption on health is being re-examined based on new evidence suggesting a beneficial effect of components in coffee other than caffeine,1–3 an effect that may be seen in the long term after the acute harmful effects of caffeine have disappeared. Supporting this idea, several recent well-designed cohort studies have found an inverse association between long-term coffee consumption and the risk of all-cause mortality in different populations.4–6 On the contrary, because coffee consumption can acutely increase the risk of several health problems (insomnia, anxiety and hypertension), it is still possible that the above association reflects the fact that coffee drinkers tend to be the healthiest members in the cohorts, and thus those with a lower risk of all-cause mortality.
To examine the association between coffee and mortality, the authors performed a cohort study with a sample of participants in the National Institutes of Health (NIH)-American Association of Retired Persons Diet and Health Study.7 The study involved more than 400 000 participants followed for a median of 13.6 years, during which over 52 000 died. Coffee consumption was assessed once at baseline using a food frequency questionnaire with 10 possible categories, ranging from zero to six or more cups per day, with an additional question about the type of coffee (caffeinated vs decaffeinated). Validation of this information was done using a 24-h dietary-recall questionnaire on two non-consecutive days.8
Although coffee consumption was associated with increased mortality in age-adjusted analyses, in fully adjusted analyses, there was an inverse, dose–response relationship between coffee consumption and mortality. Among men, those who drank between two and three cups per day had a 10% decreased risk, those who drank between four and five cups per day had a 12% decreased risk and those who drank six or more cups per day had again 10% decreased risk of mortality compared to non-coffee drinkers. These figures were even higher for women (13%, 16% and 15%, respectively). These results were found after adjusting for many confounders, including age, body mass index, race, level of education, lifestyle factors (eg, current and prior smoking) and comorbidities. In addition, the authors found a similar inverse association for analyses including only never-smoking participants or participants with very good health at baseline. Coffee consumption was associated with reduced mortality across several disease categories (including heart disease, stroke, injuries and accidents and diabetes).
This study adds compelling evidence to the hypothesis that coffee consumption is associated with lower risk of premature mortality among a general population, including participants with different ethnic origin and different educational level. The large numbers of participants and events provided this study with strong statistical power to detect even small effects of coffee. A limitation was that coffee consumption was only measured at the beginning of the 13 years of follow-up. The same is true for confounders notably weight and diet.
Some important questions regarding the effect of coffee on health still need to be addressed. For example, it is plausible that different types of coffees or preparation methods may affect the results, although mortality appears to be reduced among drinkers of both caffeinated and decaffeinated coffee. The effects of coffee may also be different for people who have experienced a prior cardiovascular event, as people with these histories were excluded from the cohort. Finally, an intriguing result of this study was the inverse association between coffee consumption and all specific causes of mortality. The ubiquity of this association might suggest that coffee is only a marker for healthier people. Unmeasured psychosocial and environmental factors associated with coffee consumption might play a role in this association.