Atrial fib: caffeine decreases recurrences

 another study highlighting benefit of caffeinated coffee, this time finding a decrease in recurrent atrial fibrillation after ablation (see afib caffeine dec afib recurrence JAMA2026 in dropbox, or doi:10.1001/jama.2025.21056), in the DECAF study (Does Eliminating Coffee Avoid Fibrillation)

Details:

-- 200 current or previous (within past 5 years) coffee-drinking adults with persistent atrial fibrillation (AF), or atrial flutter with a history of AF, planned for electrical cardioversion from 5 hospitals in the US, Canada, and Australia between November 2021 and December 2024 in this prospective, open-label, randomized clinical trial. The date of final follow-up was June 5, 2025.

-- all patients had planned upcoming direct current electrical cardioversion

-- patients who had sustained successful cardioversion were randomized to 2 groups: one group had regular caffeinated coffee consumption, and the other group to abstain from coffee and caffeine for 6 months

    -- 100 patients who were in the coffee consumption group were encouraged to drink at least 1 cup of caffeinated coffee daily; the 100 patients in the abstinence group were encouraged to completely abstain from both caffeinated and decaffeinated coffee as well as other caffeine-containing products

-- mean age 69, 70% male, height 175 cm, weight 93 kg, BMI 30

-- 83% white/8% Asian/2% Black/10% Latino

-- Education: >undergraduate degree 44%, some college 15%, up to end of high school 35%

-- hypertension 64%, OSA 27%, heart failure 22%, cardiomyopathy 19%, diabetes 14%, stroke/TIA 10%, CAD 12%, permanent pacemaker 9%, implantable cardioverter defibrillator 6%, MI 3%, peripheral vascular disease 2%

-- AF history: paroxysmal AF 30%, coexisting atrial flutter 5%, time since first diagnosed had median AF 2.5 years, duration of current AF episode had median 60 days, CHA₂DS₂VASc score 2.5

-- AF causing physical limitation: none 55%, slight 30%, marked 15%, severe at rest 1%

-- previous AF ablation 16%

-- medication history: non-vitamin K antagonist 93%, b-blocker 70%, antiarrhythmics 52% (Class 1C 12%, Class III 40%), digoxin 6%, non-DHP calcium channel blocker 6%

-- caffeinated products: drip coffee 47%, espresso drinks 37%, tea 32%, chocolate 28%, decaf coffee 9%, baseline coffee consumption 7 cups/week

-- patients reported coffee never triggering AF in 62%, reported no coffee abstinence symptoms 62%, prior MD advice to decrease coffee 17%

-- number of alcohol drinks per week: none in 46%, 1-3 in 20%, 4-7 in 12%, 8-14 in 7%, 15-21 in 3%, >21 in 3%

-- LV ejection fraction 54%, left atrial volume index, mean 45 ml/m²

-- of note, there were significant numerical imbalances above comparing the coffee consumers vs abstainers: males 76% vs 65%, coronary artery disease 9% vs 16%, previous AF ablation 20% vs 12%, coexisting atrial flutter 3% vs 8%, and history of paroxysmal AF 25% vs 35%. however, several of these categories had very few people in them

-- primary endpoint: clinically detected recurrence of AF or atrial flutter over 6 months, in intention-to-treat analysis

    -- AF and atrial flutter recurrence was defined as being clinically detected and lasting 30 seconds or longer per physician interpretation of an ECG, wearable ECG, or implanted cardiac device electrograms, and assessed in a time-to-event analysis

-- prespecified secondary end points: recurrence of AF and atrial flutter separately, and adverse events including MI, stroke, heart failure exacerbation, emergency department visit, hospitalization, and death

Results:

-- baseline coffee intake was 7 cups (IQR, 7-18) per week in both groups

-- coffee intake during follow-up

    -- coffee consumption group: 7 (IQR, 6-11)  cups per week, which did not change during the study

    -- abstinence group: 0 (IQR, 0-2) cups per week

        -- between-group difference: 7 cups (95% CI, 7-7) per week

-- intake of other caffeinated products (tea, chocolate, energy drinks, soda) and decaffeinated coffee were numerically higher in the coffee consumption group than the 7 cups/week would represent during the trial period; this was buried in the supplemental material that coffee consumers had more tea (38% vs 25%), chocolate (32% vs 27%) and soda 13% vs 6%)

    -- however, most of the differences were not significantly different, except for the addition of sugar to coffee, which led to more sugar utilization in the coffee consumption group

-- in addition to 1-, 3-, and 6-month study follow-ups as per the study protocol, patients had a mean of 5.8 clinical and 1.9 cardiology health care encounters, including 2.1 ECGs, during the study.

    -- 53% of the patients had a continuous recording device during follow-up, such as a consumer device, wearable ECG monitor, and/or implanted cardiac device

-- recurrence of AF or atrial flutter at 6 months’ follow-up was documented in 111 patients (56%):

    -- caffeinated coffee group: 47 patients (47%)

    -- abstinence group: 64 patients (64%)

-- primary endpoint of AF or atrial flutter recurrence, in intention-to-treat analysis:

    -- coffee consumption group: 47%

    -- coffee abstinence group: 64%

        -- 39% lower recurrence, HR 0.61 [0.42-0.89], P = .01

        -- a comparable benefit of coffee consumption was observed with AF recurrences only (the largest subgroup by far, see below)

-- as per the following graph, there was an impressive very early benefit in those consuming caffeinated coffee, and this extended until the end of the study at day 180 without any evident tachyphylaxis (ie, no decreased benefit up to the 6-month time)

-- prespecified sensitivity analyses:

    -- a similar benefit for caffeinated coffee consumption was also observed including when adjusting for baseline characteristics that were either prognostic risk factors and/or the numerically imbalanced baseline differences between the groups

    -- the caffeinated coffee benefit appeared to be consistent across most analyses of subgroups except for ablation history (P = .04), "though this was not adjusted for multiplicity and should be interpreted cautiously"

-- both groups had a similar proportion of recurrent AF or atrial flutter detected by a consumer device, pacemaker or defibrillator with atrial lead, or implantable loop recorder.

    -- though more patients in the abstinence group (78%) had their AF detected by a 12-lead ECG compared with the coffee consumption group (57%)

    -- more in the coffee consumption group (21%) had their AF detected by a wearable ECG monitor compared with the abstinence group (0%)

Secondary Analyses:

--There was a similar benefit of caffeinated coffee consumption on AF recurrence only, HR 0.62 [0.43-0.91]; P = 0.01

-- there was lower hazard of atrial flutter recurrence seen with caffeinated coffee consumption, although atrial flutter recurrence occurred in only 6% of patients and the between-group difference did not reach statistical significance: HR 0.37 [0.10-1.41]; P = .14

-- there were numerically more AF or atrial flutter-related hospitalizations in the abstinence compared with the caffeinated coffee group (15 vs 10, respectively), though other adverse events were similar between the groups (however ED visits were also nonsignificantly more in the abstinence group of 16 vs 13, all other measures were essentially equal)

Commentary:

-- as we know, atrial fibrillation is the most common cardiac arrhythmia, with about 1 in 3 people developing it in their lifetimes, and 2.7 to 6.1  million having AF in the US as well as >334 million worldwide: for more details see the American Health Association scientific statement  https://www.ahajournals.org/doi/10.1161/CIR.0000000000000748

-- given the high prevalence and not insignificant treatment for AF (anticoagulation, cardioversion, ablation, as well as treatment for those who have stroke/emboli), it really makes sense to try to determine and ameliorate the modifiable AF risk factors to prevent AF cases in the first place. modifiable risk factors, elaborated in the above AHA link:

    -- obesity/high BMI: 4% increased AF risk per 1-unit increase in BMI>25, as found in Framingham study

    -- physical inactivity: seems to be independent risk factor, beyond the other risk factors associated with inactivity

        -- high-intensity interval training (HIIT) in small study  found that there was a short-term benefit (seemed to be greater improvement at 12 weeks than prior exercise done by the same patients)

        -- there is evidence that mind-body exercise (body movement, mental focus and controlled breathing), yoga, and tai chi help

        -- sleep-disordered breathing (SDB): there does seem to be a dose-response curve with increased SDB severity associated with increased AF incidence/burden, and CPAP seems to lower AF risk after AF ablation

        -- diabetes: associated with higher AF risk; glycemic control is associated with reduced AF risk

        -- hypertension: increased AF risk (though hard to disaggregate from the hypertension risk factors themselves including obesity, physical inactivity and poor diet); there is some evidence that Mediterranean diet/olive oil helps, eg https://gmodestmedblogs.blogspot.com/2014/05/olive-oil-and-atrial-fibrillation.html); evidence suggests that spironolactone seems to decrease AF risk more consistently than ACE-I/ARB in other trials, though a 2009 review did find that RAS blockade with ACE-Is and ARBs did find some benefit in AF risk reduction: https://www.ecrjournal.com/articles/atrial-fibrillation-and-renin-angiotensin-system-blockade-hypertension )

        -- tobacco use: clearly increases risk and quitting helps

        -- alcohol use: a common cause of AF, even in moderate alcohol consumers. reducing and preferably eliminating alcohol helps

        -- heart failure: a common association with AF, and this applies to either heart failure with reduced or preserved ejection fraction. treatment of the heart failure seems to decreased AF risk: https://www.ahajournals.org/doi/10.1161/HAE.0000000000000078

        -- coronary artery disease: especially a risk if the patient has heart failure, or in setting of acute MI. makes sense to lower ASCVD risk aggressively with lipid control and other risk factor reduction for prevention

    -- this study supports the use of caffeine as a negative AF risk factor (ie, it was beneficial), especially in patients with AF who had a successful electrical cardioversion:

        -- there is the long-standing concern that cardiac stimulants (and especially caffeinated products) are pro-arrhythmic for AF and should be avoided, with suggestions to drink decaffeinated coffee

        -- this study, however, found the opposite: caffeinated coffee, the most commonly consumed caffeinated beverage in the US, was beneficial

            -- as per the figure above, the benefit of coffee began right at the start of the study and persisted unabated through day 180

            -- prior studies have found mixed results on this issue, with several observational studies supporting the beneficial role of caffeinated beverages. however observational studies are prone to biases

-- as noted in the study earlier this week (see https://gmodestmedblogs.blogspot.com/2026/03/coffee-and-tea-help-cognitive-function.html), there are potentially many mechanims which support the  benefits of caffeine, including blockade of the A1 and A2a adenosine receptors. and, adenosine facilitates AF induction, an effect thought to be due to sympathoexcitatory effects, shortening of atrial refractoriness, and ectopic triggers. Caffeine may thus have adenosine-mediated antiarrhythmic properties

    -- in addition, coffee also appears to have anti-inflammatory properties (and inflammation is an AF risk factor)

    -- also AF can be vagally mediated, and perhaps the catecholamine type effects of caffeine might therefore be protective for this reason.

-- in this study, the caffeinated coffee consumers were also drinking more unhealthy drinks and having more sugar (in their coffee), which could adversely affect their weight and potentially increase AF risk; however, the caffeinated coffee consumers also did more exercise (1000 more steps/d) that should be AF beneficial

Limitations:

-- the caffeinated coffee drinkers in this study drank a small amount of coffee (7 cups/week), though they did have significantly more non-coffee caffeine (ie, this study actually compared individuals with more caffeine intake than the 7 cups/week stated in the coffee consumption group, though the actual caffeine consumed was not quantified). so, it is really unclear exactly  how the results of this study are generalizable

-- this was a pretty small sample size, and many of the demographics/medical issues were based on small numbers of people, again perhaps limiting generalizability to others. the study group was mostly white, more educated, had relatively few alcohol consumers and almost all with <7 drinks/week, and very few atherosclerotic complications (PAD, MI, known CAD) also limiting the generalizability to other groups

-- the imbalances in demographics/medical issues as noted above could well affect the outcomes found, since the large number of differences are associated with decreased accuracy of the statistical adjustments intended to decrease the differences

-- a minority of people who were candidates to be included in this study actually agreed to be involved. a selection bias??  is this the reason that the average was only 1 drink of caffeinated coffee/day? is it related to other differences in diet, medical conditions and their intensity, etc??

-- the majority had paroxysmal AF. do the results apply to others with permanent AF??

-- this was also a short study of only 6 weeks. of note, the benefit did not appear to wane at the 6-month followup, but there could well be a decline in benefit over time

-- there were more individuals in the coffee consumption group (21%) who had their AF detected by a wearable ECG monitor compared with the abstinence group (0%). conversely, 12-lead ECG detected AF in 57% of those in the coffee consumption group but in 78% in the abstinence group; ie, those in the coffee consumption group were therefore more likely to have their AF detected than in the observation group, since the latter had less aggressive AF monitoring

so,

-- an interesting study that added credence to not limiting caffeinated coffee consumption in patients who have AF that responded to cardioversion, significantly decreasing their likelihood of recurrent AF

    -- this benefit appeared early in the study and had persistent benefit for the 6 weeks of the study

    -- however, this short study was limited by several issues as noted above in the limitations, reinforcing the importance of other larger studies

    -- there still were recurrences of AF in the coffee drinkers, so both coffee drinkers and abstainers should continue to have adequate anticoagulation continued

-- so, it may well be reasonable to have coffee drinkers continue with their coffee, but the study included people drinking an average of 1 cup/day (though they did have an unquantified amount of non-coffee caffeine), suggesting that a small amount of caffeine was adequate, pending further studies

-- and, perhaps the main outcome is reinforcement that modifiable AF risk factors should be assessed and treated aggressively early, to limit the likelihood of getting AF in the first place

    -- this is especially true since the initial detection of AF clinically may well be an embolic stroke, with its potential pretty awful sequelae.....

geoff

-----------------------------------

If you would like to be on the regular email list for upcoming blogs, please contact me at gmodest@bidmc.harvard.edu

to get access to all of the blogs:  go to http://gmodestmedblogs.blogspot.com/ to see the blogs in reverse chronological order

or you can just click on the magnifying glass on top right, then type in a name in the search box and get all the blogs with that name in them