Risk of Atrial Fibrillation Associated with Coffee Intake: Findings from the Danish Diet, Cancer, and Health Study

Abstract

Background:

There have been discrepant findings on whether coffee consumption is associated with the rate of developing atrial fibrillation (AF).

Methods and Results:

We used data on 57,053 participants (27,178 men and 29,875 women) aged 50–64 years in the Danish Diet, Cancer and Health study. All participants provided information on coffee intake via food-frequency questionnaires at baseline. Incident AF was identified using nationwide registries. During a median follow-up of 13.5 years, 3,415 AF events occurred. Compared with no intake, coffee consumption was inversely associated with AF incidence, with multivariable-adjusted hazard ratios of 0.93 (95% confidence interval [CI] 0.74 −1.15) for more than none to <1 cup/day, 0.88 (95% CI 0.71–1.10) for 1 cup/day, 0.86 (95% CI 0.71–1.04) for 2–3 cups/day, 0.84 (95% CI 0.69–1.02) for 4–5 cups/day, 0.79 (95% CI 0.64–0.98) for 6–7 cups/day and 0.79 (95% CI 0.63–1.00) for >7 cups/day (p-linear trend=0.02).

Conclusions:

In this large population-based cohort study, higher levels of coffee consumption were associated with a lower rate of incident AF.

Introduction

Atrial Fibrillation (AF) is the most common arrhythmia in clinical practice, affecting 2.2 million people in the United States and 4.5 million in the European Union. In the United States, AF prevalence is expected to rise to between 5 and 12 million by 2050¹; in Europe the number of patients with AF is expected to reach 14–17 million and the number of new cases of AF per year is expected to reach 120,000–215,000 in 2030.² The growing burden of AF has important public health implications because it is associated with a higher risk of stroke, heart failure, and mortality³ and it costs annually about €13.5 billion (approximately U.S. $15.7 billion) in the European Union.⁴

While coffee may result in short-term increases in blood pressure, serum cholesterol, insulin resistance and homocysteine,⁵ long-term studies have shown that regular coffee consumption is not associated with higher cardiovascular risk, and may even be protective.⁶ Although some prior prospective studies have evaluated the relationship between caffeine consumption and risk of AF,^7–9 many of the studies examined the impact of total caffeine intake rather than focusing on coffee consumption, which may have different biologic effects than caffeine from other sources. In this study, we aimed to examine whether there is an association between coffee consumption and the rate of AF after adjusting for relevant confounders among the participants in the Danish Diet, Cancer, and Health Study. We also updated a previous analysis from this cohort on caffeine and AF¹⁰ to include more cases and a longer follow-up time.

Methods

Study Cohort

In 1993 to 1997, the prospective Danish Diet, Cancer, and Health Study invited 160,725 individuals to participate. Inclusion criteria were 50–64 years of age, residence in the greater Copenhagen or Aarhus area, and no previous cancer diagnosis in the Danish National Patient Register; 57,053 participants accepted the invitation. At enrollment, anthropometric measurements were taken and biological materials were collected at one of the study centers; information on usual diet and on lifestyle was obtained using self-administered questionnaires including a semi-quantitative food-frequency questionnaire (FFQ). Using the unique personal identification number (CPR number), we linked the cohort to the Danish National Patient Register to identify hospital admissions using primary discharge diagnoses for AF or flutter (International Classification of Diseases, 10th Revision: I48) through December 2009. We excluded participants with missing information on coffee consumption or a previous record of AF in the Danish registry. A total of 55,504 participants met the inclusion criteria and had complete information on coffee and caffeine consumption, and they comprised the sample for these analyses. The Diet, Cancer and Health study was approved by the regional Ethical Committees on Human Studies in Copenhagen and Aarhus [jr.nr. (KF) 11–037/01] and [jr.nr. (KF) 01–045/93] and the Danish Data Protection Agency. All participants gave verbal and written informed consent.

Coffee Intake

Diet was assessed at baseline through a 192-item food frequency questionnaire.^{11, 12} Average intake of each food item during the last 12 months was reported in 12 possible categories from ‘never’ to ‘eight times or more per day.’ For each participant, average daily intake was calculated using the Food Calc software (version 1.3; J Lauritsen, University of Copenhagen). Standard recipes and sex-specific portion sizes were applied to calculate intake in grams per day by using data from the 1995 Danish National Dietary Survey, 24-hour diet recall interviews from 3,818 of the study participants,¹³ and several cookbooks. The questionnaire did not differentiate between caffeinated or decaffeinated coffee, but most people in Denmark drink caffeinated filtered coffee. Of the 55,504 people in our study cohort, 42,664 (76.9%) completed an additional questionnaire 5 years later. For the first 5 years of follow-up, coffee intake was based on the baseline questionnaire responses and for participants who remained alive and at risk of AF thereafter, coffee exposure was calculated as the average of their responses from the two questionnaires.

We examined the association between incident AF and intake of coffee. In a secondary analysis, we examined the association for total intake of caffeine by summing the caffeine content for a specific amount of each caffeinated food during the previous year (1 cup for coffee or tea, one 12-ounce bottle or can for carbonated beverages, and 1 ounce for chocolate) multiplied by the frequency of its use. We used the U.S. Department of Agriculture food composition sources to estimate that the caffeine content was 137 mg per cup of caffeinated coffee, 47 mg per cup of tea, 46 mg per bottle or can of cola beverage, and 7 mg per serving of chocolate candy.

Atrial Fibrillation

All citizens of Denmark have a unique personal identification number that is used in all national registries, and updated information is available on emigration, hospitalizations and death. Hospital admissions and discharge diagnoses from all hospitals in Denmark are recorded by date in the Danish National Patient Register. The registry includes discharge diagnoses from in-hospital patients since 1977 and additional discharge diagnoses from emergency rooms and outpatient visits since 1995.

The outcome in this study was defined as incident atrial fibrillation and/or atrial flutter (AFL) during the study period. Diagnoses were recorded using the Eighth International Classification of Diseases (ICD-8) until the end of 1993 (AF (427.93) and AFL (427.94) in the Danish version which is equivalent to AF or AFL (427.4) in the international version). From January 1994, the ICD-10 classification was used with the diagnosis of AF and/or AFL (I.48). The validity of the combined diagnosis of AF and/or AFL is high, with a positive predictive value of 92.6% in this cohort.¹⁴ If a patient had both an emergency room visit and a hospital admission on the same date, only the in-hospital diagnosis was considered in order to avoid possible misclassification. In line with previous observational studies, the combined diagnosis of ‘AF and/or AFL’ was referred to as AF.

Other Covariates

We obtained information on demographics and lifestyle factors using a self-administered questionnaire. Body mass index, blood pressure and total cholesterol were measured by a lab technician at the time of recruitment.¹⁵ We used self-reports, ICD codes and ATC (Anatomical Therapeutic Chemical) drug codes to obtain information on prevalent and incident hypertension, diabetes mellitus and cardiovascular disease (yes/no).¹⁶

Statistical Analysis

Individuals were considered at risk from the date of the study questionnaire (1993–1997) until the date of first hospital admission for AF, death, emigration, or end of follow-up (December 2009), whichever came first. We modeled coffee consumption using categories of 0, <1, 1, 2–3, 4–5, 6–7 and ≥7 cups per day. We constructed multivariable Cox proportional hazards models with age as the time scale to estimate the incidence rate ratio (IRR) and 95% confidence intervals (CI) for the association between coffee consumption and the rate of AF. For consistency with the previous paper on caffeine and AF in this cohort,¹⁰ we selected the same covariates a priori: From the baseline data, we included information on sex, body mass index (BMI; in kg/m²), systolic blood pressure (mm Hg), total serum cholesterol (continuous), alcohol consumption (grams/day), smoking status (never, former, current) and years of education beyond elementary school (0,<3, 3–4, >4 years). We used regularly updated information on hypertension (yes/no), diabetes mellitus (yes/no) and cardiovascular disease (yes/no) using time-varying covariates to increase the sensitivity of key covariates. We conducted tests of the linear component of trend for increasing categories of coffee and caffeine intake by assigning the median values for each category and testing the statistical significance of the term in the multivariable model. In order to non-parametrically examine whether there was a nonlinear association, we conducted sensitivity analyses to model caffeine exposure using restricted cubic splines with five knots at fixed percentiles of the distribution of caffeine (5%, 27.5%, 50%, 72.5%, 95%) with the reference value set at the median caffeine intake for participants in the 5th percentile of intake (118 mg/d) and included the covariates listed above.

To examine whether the IRRs varied by sex, history of hypertension, diabetes or cardiovascular disease, we conducted likelihood ratio test to compare models with versus without cross-product terms for categories of exposure and the potential modifier. We tested the proportional hazard assumptions using Schoenfeld residuals and interactions with the logarithm of time, and we found no significant violations. Statistical analyses were performed using Stata (version 12.1, Stata Corp, College Station, Texas, USA) with 2-tailed tests set at α=0.05 for statistical significance.

Results

Participants with higher levels of coffee consumption were more likely to be current smokers and they had higher levels of total serum cholesterol (Table 1). During a median follow-up of 13.5 years, 3,415 people were hospitalized with incident AF.

Table 1.

Baseline characteristics of the subjects in the Danish Diet, Cancer, and Health Study according to cups of coffee per day, mean± standard deviation or n (%)

				Cups of Coffee Per Day
		All Participants		None		<1		1		2–3		4–5		6–7		>7
N		55,504		2,068		4,533		2,826		14,962		15,564		9,503		6,048
Median caffeine consumption		621.6		212.8		223.4		255.1		390.7		638.1		902.8		1103.0
IQR of caffeine consumption		367.4 – 893.5		118.2 – 309.0		128.9 – 320.0		164.9 – 348.8		351.2 – 463.1		621.4 – 734.3		894.1 – 938.1		1098.5 – 1124.1
Age (y)		56.7±	4.4	56.5±	4.4	56.6±	4.3	57.2±	4.4	57.1±	4.4	56.7±	4.3	56.5±	4.4	55.8±	4.2
Sex (% men)		26,401	(47.6)	699	(33.8)	1,831	(40.4)	1,097	(38.8)	6,381	(42.6)	7,626	(49.0)	5,080	(53.5)	3,687	(61.0)
Body height (cm)		170.1±	8.9	168.4±	8.6	169.6±	8.8	169.2±	8.7	169.6±	8.8	170.3±	9.0	170.6±	8.9	171.5±	8.6
Body mass index (kg/m2)		26.0±	4.0	26.0±	4.6	26.0±	4.3	26.0±	4.2	26.0±	4.0	26.0±	3.9	26.2±	4.0	26.1±	4.0
Smoking (%)
	Never	19,481	(35.1)	1,071	(51.8)	2,090	(46.1)	1,313	(46.5)	6,323	(42.3)	5,333	(34.3)	2,462	(25.9)	889	(14.7)
	Former	15,931	(28.7)	568	(27.5)	1,461	(32.2)	899	(31.8)	4,724	(31.6)	4,665	(30.0)	2,464	(25.9)	1,150	(19.0)
	Current	20,092	(36.2)	429	(20.7)	982	(21.7)	614	(21.7)	3,915	(26.2)	5,566	(35.8)	4,577	(48.2)	4,009	(66.3)
Median alcohol consumption		12.9		7.5		11.7		14.0		13.7		13.8		13.1		11.7
IQR of alcohol consumption		5.9 – 31.1		1.7 – 18.3		5.1 – 30.5		6.1 – 32.6		6.4 – 31.8		6.5 – 31.6		6.0 – 30.7		3.6 – 26.0
Systolic blood pressure (mm Hg)		139.5±	20.5	137.9±	21.3	139.3±	21.1	140.8±	21.6	140.2±	20.5	139.7±	20.4	139.2±	20.1	137.9±	19.6
Hypertension treatment (%)		13,703	(24.7)	498	(24.1)	1,140	(25.1)	720	(25.5)	3,709	(24.8)	3,788	(24.3)	2,329	(24.5)	1,519	(25.1)
Total serum cholesterol > 6 mmol/L (%)		27,503	(49.6)	860	(41.6)	2,053	(45.3)	1,338	(47.3)	7,432	(49.7)	7,831	(50.3)	4,923	(51.8)	3,066	(50.7)
Diabetes mellitus (%)		284	(0.5)	10	(0.5)	21	(0.5)	20	(0.7)	83	(0.6)	78	(0.5)	48	(0.5)	24	(0.4)
Education (%)
	≤7 years	18,310	(33.0)	610	(29.5)	1,071	(23.6)	700	(24.8)	4,274	(28.6)	5,099	(32.8)	3,757	(39.5)	2,799	(46.3)
	8 to 10 years	25,580	(46.1)	951	(46.0)	2,141	(47.2)	1,338	(47.3)	7,253	(48.5)	7,270	(46.7)	4,199	(44.2)	2,428	(40.1)
	≥11 years	11,614	(20.9)	507	(24.5)	1,321	(29.1)	788	(27.9)	3,435	(23.0)	3,195	(20.5)	1,547	(16.3)	821	(13.6)

The rate of AF was lower with higher levels of regular coffee intake (Table 2). Compared with no regular coffee intake, the multivariable adjusted rate ratio of AF was 0.93 (95%CI 0.74 −1.15) for people who drank more than none to fewer than 1 cup per day, 0.88 (95%CI 0.71–1.10) for 1 cup per day, 0.86 (95%CI 0.71–1.04) for 2–3 cups per day, 0.84 (95%CI 0.69–1.02) for 4–5 cups per day, 0.79 (95%CI 0.64–0.98) for 6–7 cups per day and 0.79 (95%CI 0.63–1.00) for >7 cups per day. Although not all of these estimates were statistically significant, there was a statistically significant linear component of trend (p=0.02), suggesting that higher coffee intake is associated with a moderately lower rate of AF (Table 3). There was no evidence that the association between coffee consumption and AF varied by sex (p-interaction=0.29), history of hypertension (p-interaction=0.91), diabetes (p-interaction=0.56) or cardiovascular disease (p-interaction=0.27).

Table 2.

Incidence rates and incidence rate ratios and 95% confidence intervals according to cups of coffee per day in the Danish Diet, Cancer, and Health Study

	Cups of Coffee Per Day
	None	< 1	1	2–3	4–5	6–7	> 7	P-Value*
N	2068	4533	2826	14962	15564	9503	6048
No. of subjects with atrial fibrillation	124	279	173	899	933	557	344
No. of person-years of follow-up	26404	57978	36161	193576	200646	122047	77439
Incidence rate per 10,000 person-years	46.96	48.12	48.12	46.44	46.50	45.64	44.42
†Age Adjusted	1 (ref.)	0.99 [0.79 – 1.23]	0.95 [0.77 – 1.18]	0.93 [0.77 – 1.13]	0.97 [0.80 – 1.18]	0.96 [0.78 – 1.18]	1.03 [0.82 – 1.29]	0.40
‡Multivariable Adjusted	1 (ref.)	0.93 [0.74 – 1.15]	0.88 [0.71 – 1.10]	0.86 [0.71 – 1.04]	0.84 [0.69 – 1.02]	0.79 [0.64 – 0.98]	0.79 [0.63 – 1.00]	0.02

^*p-value for linear component of trend

^†Age was the time scale in the Cox models

^‡Age was the time scale in the Cox models and adjusted for sex, body mass index, systolic blood pressure (mm Hg), total serum cholesterol (continuous), alcohol consumption (grams/day), smoking status (never, former, current), years of education beyond elementary school (0, <3, 3–4, >4 years), diagnosis of hypertension (yes/no), diabetes mellitus (yes/no) and cardiovascular disease (yes/no)

Table 3.

Incidence rates and incidence rate ratios and 95% confidence intervals by percentiles of caffeine consumption in the Danish Diet, Cancer, and Health Study

	Percentiles of Caffeine Intake Per Day
	0–5	5–20	20–40	40–60	60–80	80–95	95–100	P-Value*
	0–146.8 mg	146.9–350.1 mg	350.1–554.1 mg	554.2–651.7 mg	651.7–903.4 mg	903.4–1108.6 mg	1108.7–1542.7 mg
N	2776	8302	11129	11074	11108	8347	2766
Median Caffeine (mg/d)	118	317	393	622	829	1012	1137
No. of subjects with atrial fibrillation	177	548	633	681	663	458	149
No. of person-years of follow-up	34628	106377	144463	142420	143346	107424	35563
Incidence rate per 10,000 person-years	51.12	51.51	43.82	47.82	46.25	42.63	41.90
†Age Adjusted	1 (ref.)	0.97 [0.82 – 1.15]	0.85 [0.72 – 1.00]	0.94 [0.79 – 1.10]	0.93 [0.79 – 1.10]	0.91 [0.76 – 1.08]	0.95 [0.76 – 1.18]	0.87
‡Multivariable Adjusted	1 (ref.)	1.04 [0.88 – 1.24]	0.95 [0.80 – 1.12]	0.96 [0.82 – 1.14]	0.97 [0.82 – 1.15]	0.89 [0.75 – 1.06]	0.87 [0.70 – 1.09]	0.046

^*p-value for linear component of trend

^†Age was the time scale in the Cox models

^‡Age was the time scale in the Cox models and adjusted for sex, body mass index, systolic blood pressure (mm Hg), total serum cholesterol (continuous), alcohol consumption (grams/day), smoking status (never, former, current), years of education beyond elementary school (0, <3, 3–4, >4 years), diagnosis of hypertension (yes/no), diabetes mellitus (yes/no) and cardiovascular disease (yes/no)

In analyses evaluating the association between total caffeine intake and the rate of AF, there was a lower rate of AF at higher levels of total caffeine intake (Table 3). Compared to people in the lowest 5^th percentile of caffeine intake, the multivariable adjusted rate ratio of AF was 1.04 (95%CI 0.88–1.24), and the rate ratios suggested modest protective benefits for people in the 5^th-20^th percentile, (IRR=0.95, 95%CI 0.80–1.12), 20–40^th percentile (IRR=0.95, 95%CI 0.80 – 1.12), 40–60^th percentile (IRR=0.96, 95%CI 0.82–1.14), 60–80^th percentile (IRR=0.97, 95%CI 0.82–1.15), 80–95^th percentile (IRR=0.89, 95%CI 0.75–1.06) and 95–100^th percentile (IRR=0.87, 95%CI 0.70–1.09). Although none of these estimates were statistically significant, there was a statistically significant linear component of trend (p =0.05), suggesting that higher caffeine intake is associated with a moderately lower rate of AF. In multivariable analyses allowing for potential nonlinear associations, there was a statistically significant association between total caffeine intake and the rate of AF. Compared to abstainers, there appeared to be progressively lower rates of AF until 500 mg of caffeine per day, and the inverse association appeared to plateau or weaken for higher levels of consumption (Figure).

Figure.

Restricted cubic spline of the incidence rate ratio and 95% confidence interval for the association between atrial fibrillation and caffeine per day across the full range of caffeine intake. Reference value is median caffeine intake for participants in the 5^th percentile of intake (118 mg/d).

Figure Legend Multivariable-adjusted IRRs for incident AF according to caffeine intake. IRRs were adjusted for sex, body mass index (BMI; in kg/m²), systolic blood pressure (mm Hg), total serum cholesterol (continuous), alcohol consumption (grams/day), smoking status (never, former, current) and years of education beyond elementary school (0, <3, 3–4, >4 years), diagnosis of hypertension (yes/no), diabetes mellitus (yes/no) and cardiovascular disease (yes/no). The dashed line represents the 95% confidence interval for the estimated hazard ratio.

Discussion

In this population-based prospective cohort study, both coffee consumption and total caffeine intake were associated with a lower rate of incident AF. The association was consistent across subgroups defined by potential risk factors for AF.

Coffee is one of the most widely consumed beverages in the world, and caffeine has been studied more than any other of its ingredients. However, coffee is a complex mixture of hundreds of compounds and contains other bioactive ingredients that may yield beneficial and detrimental consequences independent of the effects of caffeine.^17–19 Therefore, research on the health impacts of total caffeine intake may not highlight the physiologic effects specific to coffee, whether or not it is caffeinated. For instance, some studies have shown that the deleterious impact of caffeine on epinephrine concentrations²⁰, post-load glucose concentrations²¹ and blood pressure²² are stronger than the impact of the same amounts of coffee, suggesting that other compounds in the coffee may counterbalance the adverse effects of caffeine.

The cardiovascular health benefits of coffee and caffeine may also differ. One cross-sectional study²³ showed that higher caffeinated coffee consumption and higher total caffeine intake were associated with lower plasma concentrations of E-selectin and C-reactive protein among women with type 2 diabetes, but decaffeinated coffee consumption was only associated with lower plasma concentrations among healthy women. The lower incidence of type 2 diabetes mellitus^{24, 25} and the lower C-peptide concentrations²⁶ associated with coffee intake is similar for consumption of caffeinated and decaffeinated coffee, suggesting that some of the salubrious effects are attributable to coffee components other than caffeine. Finally, in a small randomized controlled trial comparing the glucose and insulin response to a standard oral glucose tolerance test,²¹ caffeine acutely increased the glucose and insulin response, caffeinated coffee did not affect glucose levels but tended to increase plasma insulin and decaffeinated coffee reduced the glucose response.

The compounds underlying the link between coffee and cardiovascular benefits have not been definitively identified but it is possible that the high concentration of antioxidants in coffee inhibits inflammation and thereby reduces cardiovascular risk. Coffee phenols have been shown to reduce blood pressure and improve endothelial function²⁷ and the metabolized phenols, caffeic acids, may enhance high-density lipoprotein-mediated cholesterol efflux from macrophages.²⁸

In a previous analysis on caffeine and AF from this cohort with shorter follow-up, fewer AF cases and lower statistical power,¹⁰ the authors reported similar lower rates of AF associated with higher levels of caffeine intake, but the findings did not reach statistical significance. Likewise, several other studies reported lower rates of AF associated with total caffeine²⁹ and coffee^{30, 31} that did not attain statistical significance.

In our study of people in Denmark, a majority of the participants consumed 4 or more cups of coffee per day, which is higher than the intake reported in prior studies, but the results support prior findings of a lower risk of AF with habitual coffee intake. One large study³² reported a U-shaped association between AF and both total caffeine and caffeinated coffee intake but no association for decaffeinated coffee. However, the authors only concluded that total caffeine was not associated with elevated risk, omitting the potential health benefits of coffee intake. Klatsky and colleagues³⁰ also reported a lower AF risk among people who consumed coffee, and concluded that caffeine is unlikely to increase arrhythmia risk. In a sensitivity analysis, they showed that the association between total coffee intake and arrhythmia was stronger for people who reported consuming only caffeinated coffee (IRR=0.68, 95%CI 0.42–1.10) than for those reporting only decaffeinated coffee (IRR=0.87, 95%CI 0.54–1.39) or both caffeinated and decaffeinated coffee (IRR=0.80, 95%CI 0.49–1.29). This may be due to the fact that decaffeinated coffee has been artificially treated so as to remove caffeine and other possibly cardioprotective compounds might also be removed in the process.

There are some limitations to our study that warrant discussion. Although we had extensive data on diet, lifestyle and comorbid conditions, we cannot rule out the possibility of residual or unmeasured confounding. However, we observed a statistically significant association after adjusting for age, smoking status and other potential confounders and for coffee intake the association became stronger after control for confounding. Our food frequency questionnaire asked about total coffee intake, so we cannot distinguish between the impacts of caffeinated versus decaffeinated coffee. However, in Denmark, there is a generally high level of caffeinated coffee consumption and limited consumption of decaffeinated coffee. Therefore, most of the self-reported intake was likely caffeinated coffee. As a result, we could not examine the impact of very low amounts of caffeinated coffee intake or the impact of decaffeinated coffee. In addition, we did not have information on brewing method or genetic polymorphisms that may modify this association. Coffee intake and covariate data were only available at baseline, and for most participants, in the fifth year of follow-up, and these may have changed over the 13 year average follow-up period, resulting in some degree of non-differential misclassification of exposure, which would reduce the power to detect an association. In addition, this study was limited to cases with recorded hospitalizations or deaths for AF as a primary cause. We cannot detect the incidence of silent AF, though this should not affect the validity of the current study, since the identification of hospitalizations or deaths from symptomatic AF is unlikely to be impacted by levels of habitual coffee intake. It is possible that some participants had obstructive sleep apnea, resulting in higher coffee intake and a higher risk of AF. However, this would have led to underestimate of the potential benefits of coffee on AF risk. As with any study using self-reported exposure information, there is a concern of poor recall. However, our food-frequency questionnaire was validated in a study comparing against two 7-day weighted diet records.¹² Furthermore, if the misclassification of coffee was unrelated to AF incidence, our results would likely be an underestimate of the protective effect of coffee.

On the other hand, our study has many strengths, including a large sample size, a prospective population-based design, detailed data on diet and factors potentially related to exposure and outcome, and almost complete follow-up of the study cohort over many years.

It is challenging to study the long-term health effects of caffeine in the absence of coffee since coffee is the most abundant source of caffeine. However, the results of our study and others suggest that coffee may have health benefits due to components other than caffeine. In conclusion, participants with higher levels of coffee intake had a lower rate of incident AF. Further research is necessary to identify the components of coffee that promote cardiovascular health.