Main characteristics of included studies
| Study/year | Design | Region | Population | Follow-up (years) | Age (years)/male % | Caffeine exposure | Comparator/reference category | Outcome assessment | Outcome adjustments |
|---|---|---|---|---|---|---|---|---|---|
| Mattioli 200531 | Case–control study | Italy | 116 patients hospitalised for an acute episode of lone AF 116 healthy outpatient age- and sex-matched control subjects. | N/A | 54/74% | Self-administered questionnaire | Non-coffee drinkers: 0 cups per day | ECG evaluation | None (crude OR) |
| Wilhelmsen 200132 Multifactor Primary Prevention Study | Cohort | Sweden | 7495 male individuals who participated in Multifactor Primary Prevention Study | 25.2 | 47–55/100% | Not defined | Non-coffee drinkers: 0 cups per day | ECG, hospital records and AF ICD-9 codes | Age |
| Frost and Vestergaard 200633 Danish diet, cancer, and health study | Prospective cohort | Denmark | 57 053 individuals aged between 50 and 64 years | 5.7 | 56/47% | Food frequency questionnaire was applied and quantity of caffeine was derived with FOODCALC software | First quintile: mean caffeine intake 248 mg/day (SD 91) | ICD-8 and ICD-10 codes | Age, sex, body height, BMI, smoking, consumption of alcohol, systolic blood pressure, treatment for hypertension, total serum cholesterol, and level of education |
| Mukamal 200934 Stockholm Heart Epidemiology Program (SHEEP) | Prospective cohort | Sweden | 1369 survivors of myocardial infarction and included in SHEEP | 6.9–9.9 in patients free from events | 59.8/70.2% | Cups of coffee per day assessed through questionnaire | 0 to <1 cup of coffee per day | ICD-10 codes | Age, sex, diabetes, smoking, obesity, physical inactivity, alcohol consumption, tea consumption, education, and intake of boiled coffee |
| Conen 201035 Women's Health Study (WHS) | Prospective cohort | USA | 33 638 women older than 45 and free from CV disease | ∼10 | 53/0% | Food frequency questionnaire was applied in order to determine specific amount of coffee, tea, cola and chocolate | First quintile: median caffeine intake 22 mg/day (IQR 9–44) | AF reporting by individuals prompted medical record review to retrieve ECG evidence of AF or a clear medical report indicating a personal history of AF | Age, systolic blood pressure, BMI, hypertension, diabetes, hypercholesterolaemia, smoking, exercise, alcohol consumption, parental history of myocardial infarction, treatment group, fish intake, and race/ethnicity |
| Shen 201036 Framingham Heart Study | Prospective cohort | USA | 4526 adults from Framingham Heart Study original and offspring cohort | 4 | 62/44% | 126-item semiquantitative food frequency questionnaire | First quintile: mean caffeine intake 23 mg/day (minimum 0; maximum 82) | AF was ascertained and recorded from interim medical evaluations. Afterwards it was validated by cardiologists who reviewed and classified all available ECG and clinical records | Age, sex, BMI, systolic blood pressure, hypertension treatment, ECG, PR interval, significant heart murmur, and heart failure |
| Klatsky 201137 | Prospective cohort | USA | 11 679 individuals who voluntarily underwent a health examination before subscribing to a healthcare plan and supplied information about coffee | 17.6 | 43.4% >60 y 47.2% | Questionnaire that included coffee and tea consumption patterns | Non-coffee drinkers | ICD-9 codes | Age, sex, ethnicity, BMI, education, cigarette smoking, alcohol intake and a cardiorespiratory composite covariate |
AF, atrial fibrillation; BMI, body mass index; CV, cardiovascular; ICD, International Classification of Diseases; N/A, not available.