Cocoa flavanol supplementation and incident atrial fibrillation in the COSMOS trial

Several large-scale, long-term cohort studies have explored the potential impact of cocoa flavanols on cardiovascular health, but the specific association between cocoa flavanols and atrial fibrillation (AF) is not fully understood. Studies of chocolate intake and AF provide indirect support for the role of cocoa flavanols but are limited by its caloric and saturated fat content and unreliable flavanol content. In the Danish Diet, Cancer and Health cohort study, participants who consumed chocolate up to 6 times a week had a 20% lower risk of incident AF compared to those who consumed it less than once a month [hazard ratio (HR) 0.80, 95% confidence interval (CI): 0.71–0.91].¹ However, the Women’s Health Study² and the Physicians Health Study³ did not find an association between chocolate consumption and incident AF. Further, a recent meta-analysis combining five studies (which included the above cohorts) found no overall association between chocolate and incident AF.⁴ Given the limited data on cocoa flavanols specifically and the heterogeneity of the available results regarding chocolate intake, large primary prevention trials are required to better understand the role of cocoa flavanols in AF.

We therefore present results from the COcoa Supplement and Multivitamin Outcomes Study (COSMOS), a recently completed randomized clinical trial of a cocoa extract supplement for cardiovascular disease (CVD) prevention.⁵ The COSMOS trial design has previously been described.⁵ In brief, COSMOS is a randomized, double-blind, placebo-controlled, two-by-two factorial trial of cocoa extract [containing 500 mg flavanols/day, including 80 mg (–)-epicatechin] and multivitamin supplementation for the prevention of CVD and cancer among 21 422 US adults. Women aged ≥65 years and men aged ≥60 years who were free of major CVD and had not been recently diagnosed with cancer were randomized after a placebo run-in to demonstrate ≥75% compliance. The primary total CVD endpoint was a composite of myocardial infarction, stroke, CVD mortality, revascularization procedures, unstable angina requiring hospitalization, peripheral artery disease, and carotid artery disease. Semi-annual questionnaires that enquired about new physician diagnoses of AF were similarly sent to all participants to ascertain the AF endpoint. After excluding participants who reported irregular heart rhythms at baseline prior to randomization, participants were analysed according to randomization group, and Cox proportional hazards models estimated HRs of incident AF in the cocoa extract group vs. placebo after controlling for age, sex, and multivitamin treatment assignment.

The study population included 18 188 COSMOS trial participants without a history of an irregular heart rhythm at baseline. Comorbidities and risk factors associated with AF remained well balanced by randomization except for heart failure which, although rare, was more prevalent in the placebo arm (see Supplementary material online, Table S1). Over a median of 5.5 years (Q1–Q3, 5.1–6.1) of follow-up, 1029 (5.7%) participants reported being diagnosed with incident AF. Of these, 694 cases were reported during a median of 3.5 years (Q1–Q3, 3.2–4.2) of randomized treatment and another 335 cases during a median of 2.0 years (Q1–Q3, 2.0, 2.0–2.1) of post-intervention follow-up. During the intervention phase, incident AF did not significantly differ for the cocoa extract (333 cases) vs. placebo (361 cases) group (HR = 0.92, 95% CI: 0.79–1.06, P = 0.24). However, over the 5.5 years of cumulative follow-up, a significantly lower HR for incident AF emerged in the cocoa extract compared to placebo group (HR: 0.85, 95% CI: 0.75–0.96, P = 0.01; 477 vs. 522 cases). During the post-intervention phase, participants that were assigned to cocoa extract supplement had a statistically significant reduction in incident AF as compared to those assigned to placebo (HR: 0.75, 95% CI: 0.60–0.93, P = 0.009; 144 vs. 191 cases; Figure 1). These results were not materially altered after multivariable adjustment for baseline imbalances in heart failure as well as other comorbidities and risk factors associated with AF risk factors (Table 1).

Figure 1.

Risk of incident atrial fibrillation.

Table 1.

Multivariate adjusted models to assess association between cocoa flavanol and incident atrial fibrillation

	HR (95% CI)	P-value
Model 1
Intervention phase	HR 0.92 (95% CI: 0.79–1.07)	0.281
Post-intervention phase	HR 0.75 (95% CI: 0.60–0.93)	0.009
Cumulative follow-up phase	HR 0.86 (95% CI: 0.76–0.97)	0.016
Model 2
Intervention phase	HR 0.94 (95% CI: 0.81–1.10)	0.452
Post-intervention phase	HR 0.76 (95% CI: 0.61–0.95)	0.015
Cumulative follow-up phase	HR 0.88 (95% CI: 0.77–0.99)	0.04

Model 1: adjusted for age, sex, and treatment assignment. Model 2: adjusted for age, sex, treatment assignment, body mass index, race, smoking, hypertension, heart failure, diabetes, CVD, and cholesterol medication.

While these results come from a large randomized clinical trial, several limitations warrant consideration. Firstly, AF was not a predefined endpoint of the primary COSMOS trial. Secondly, incident AF events did not undergo adjudication but were self-reported on annual follow-up questionnaires. This may lead to non-differential misclassification and could have biased results towards the null but is unlikely to account for the differential treatment effect observed in this randomized trial where most AF risk factors were balanced and multivariable adjustment for imbalances and risk factors did not modify the results. Thirdly, COSMOS enrolled relatively healthy individuals without a history of major clinical CVD events, introducing potential volunteer bias and concerns about generalizability to younger adults or those with CVD. It is also important to note that cocoa flavanols are difficult to translate into specific forms of chocolate as processing methods and other factors greatly impact flavanol content.⁶

The primary results of the COSMOS trial found that cocoa extract supplementation led to a 27% reduction in cardiovascular death (HR 0.73, 95% CI: 0.54–0.98), but no significant reduction in the primary endpoint of total cardiovascular events (HR 0.90, 95% CI: 0.78–1.02).⁵ In this secondary analysis, a potentially beneficial effect of cocoa extract on incident AF emerged over an extended follow-up of 5.5 years inclusive of both the intervention and post-intervention periods. This signal in COSMOS may be driven by the flavanol and (–)-epicatechin content of the cocoa extract intervention. Cocoa flavanols exhibit vasodilatory, anti-inflammatory, antioxidant, antiplatelet, and angiotensin-converting enzyme properties⁷ that may result in electrical and structural atrial remodelling, which are potentially protective against AF development. This may explain the delay in which this signal has emerged, with the reduction of incident AF only becoming apparent over time with atrial remodelling. Based upon these promising trial results, the hypothesis that cocoa flavanols may reduce the risk of incident AF warrants further study in future randomized trials.

Supplementary material

Supplementary material is available at European Journal of Preventive Cardiology.

Author contribution

M.E.M., J.E.M., A.K.A., H.D.S., and C.M.A. contributed to the conception or design of the work. J.E.M., A.K.A., and H.D.S. contributed to the acquisition, analysis, or interpretation of data for the work. M.E.M. and C.M.A. drafted the manuscript. J.E.M., A.K.A., A.C., and H.D.S. critically revised the manuscript. All gave final approval and agree to be accountable for all aspects of work ensuring integrity and accuracy.

Funding

The COcoa Supplement and Multivitamin Outcomes Study (COSMOS) is supported by an investigator-initiated grant from Mars Edge (J.E.M., H.D.S.), a segment of Mars dedicated to nutrition research and products, which included infrastructure support and the donation of study pills and packaging. Pfizer Consumer Healthcare (now Haleon) provided support through the partial provision of study pills and packaging (J.E.M., H.D.S.). COSMOS is also supported in part by grants AG050657, AG071611, EY025623, and HL157665 from the National Institutes of Health, Bethesda, MD. The Women’s Health Initiative (WHI) programme is funded by the National Heart, Lung, and Blood Institute, National Institutes of Health, U.S. Department of Health and Human Services, through contracts 75N92021D00001, 75N92021D00002, 75N92021D00003, 75N92021D00004, and 75N92021D00005. Neither company had a role in the trial design or conduct, data collection (other than blinded assays supported by Mars Edge and completed independently), data analysis, or manuscript preparation or review.

Data availability

The data set(s) will be de-identified prior to release for sharing. We will make the data and associated documentation available to users only under a data-sharing agreement. Details on the availability of the study data to other investigators will be on our study website at https://cosmostrial.org/.