Coffee Consumption and Incident Heart Failure in Men

Abstract & Commentary

By Andrew J. Boyle, MD, PhD

Source: Ahmed et al. Coffee consumption and risk of heart failure in men: An analysis from the cohort of Swedish men. Am Heart J. 2009;158:667-672

The effects of coffee on the heart remain incompletely described. There are conflicting reports of coffee's effects on the incidence of coronary artery disease, heart failure (HF), atrial fibrillation, and glucose homeostasis. There are plausible biological reasons that coffee may increase the incidence of HF (such as predisposing to hypertension) or decrease it (such as improving calcium sensitivity of cardiomyocytes). To address the issue of coffee intake and incident HF, Ahmed et al utilized this cohort of Swedish men to determine outcomes in men who selfreported varying levels of coffee consumption.

The cohort comprised 48,850 men aged 45-79 years residing in Sweden who were followed prospectively for nine years. Dietary, lifestyle, demographic, and behavioral factors were self-reported at the beginning of the study. Clinical events of HF hospitalization and HF deaths were extracted from the Swedish inpatient dataset and the Swedish death registry, respectively. After excluding subjects with pre-existing cancer, HF, diabetes, myocardial infarction (MI), improbable self-reported energy intake, and incorrect national registration number, 37,315 men were included. The data were statistically corrected for age, activity, body-mass index, sodium intake, fat intake, smoking, dyslipidemia, education level, marital status, aspirin use, alcohol intake, tea intake, and family history of premature MI. The primary endpoint was HF hospitalization or death.

During the study period, 690 men were hospitalized with HF and 94 died of HF, which corresponds to 24.5 cases per 10,000 person-years. This is comparable to the reported overall HF hospitalization rate in Sweden of 23.7 per 10,000 patient-years. Subjects were stratified into five groups according to coffee intake: 1 cup per day, 2 cups per day, 3 cups per day, 4 cups per day, and 5 cups per day. Ahmed et al found that coffee consumption was not a significant predictor of incident heart failure. This lack of association was not changed when analyzed by number of cups per day, overweight status, current smoking, or alcohol intake. The initial analysis was not controlled for hypertension, but when Ahmed et al also controlled for this variable, there was still no relationship between coffee consumption and HF. Analysis of those with prior MI or diabetes, who were excluded at baseline, revealed there was no association between coffee consumption and HF in that population either. Ahmed et al then performed an additional analysis by excluding the incidences of heart failure in the first two years, in case baseline symptoms had affected caffeine intake. There was still no relationship between coffee consumption and incident HF. Finally, Ahmed et al analyzed all-cause mortality, and showed no difference based on coffee consumption. They conclude their results do not support the hypothesis that high coffee consumption is associated with increased rates of HF hospitalization or mortality.


I am asked relatively frequently about the effects of coffee consumption on the heart. The data have been conflicting, with some studies showing a detrimental effect, some showing a positive effect, and others showing no effect at all. This large cohort study included over 37,000 men and, thus, has strong statistical power. The multivariable model Ahmed et al used adjusted for many baseline variables. However, some limitations to this study must be acknowledged. The clinical variables are self-reported, and often binary, whereas blood pressure is really a continuous variable, and under-reporting may occur if subjects are not aware of the problem. It would be more robust to have used the subjects' measured blood pressure. Furthermore, dietary consumption (including coffee) is self-reported, but then so is our patients' coffee consumption. This study was only in men, so the results should not be generalized to women.

The types of coffee and the strengths of coffee vary by location. It is not clear from this study whether the predominant type of coffee is drip coffee, instant, or boiled coffee. Nor is the study controlled for caffeine content. Therefore, the number of cups per day quoted in this study may not translate for all types or strengths of coffee. Furthermore, the use of decaffeinated coffee was not explored in this study, nor was the use of caffeinated non-coffee drinks, such as cola or energy drinks. Until definite evidence regarding the effects of caffeine on heart health emerge, moderation continues to be appropriate, but there does not appear to be a need to tell patients to abstain from coffee to prevent incident HF.