Benefits
Small blood pressure and blood vessel changes
A 2017 Cochrane review of 35 randomized trials in 1,804 people found a small drop of about 2 mmHg in systolic and diastolic blood pressure in short-term studies, which the reviewers described as a small effect based on moderate-quality evidence. A separate randomized trial in healthy adults (the Flaviola study) improved flow-mediated dilation, a surrogate marker of blood vessel function, plus a calculated risk score rather than actual heart events. The one large outcomes trial, COSMOS, did not meet its primary endpoint: in 21,442 older US adults taking 500 mg/day of cocoa flavanols for a median of 3.6 years, total cardiovascular events were not significantly reduced (hazard ratio 0.90, 95% confidence interval 0.78 to 1.02, P = 0.11). Heart attack, stroke, artery-opening procedures and death from any cause were all non-significant too. One secondary result, a 27% lower rate of cardiovascular death, was one of many secondary endpoints and is not proof that cocoa flavanols prevent heart disease.
Memory and brain blood flow (not established; the one large trial was negative)
The idea that cocoa flavanols protect memory was tested directly in COSMOS-Mind, a 3-year randomized trial in 2,262 older adults, where cocoa extract had no effect on global cognition (P = 0.28). That trial is not among this page's references, so nothing cited here supports a cognition benefit. What is cited is a small 2014 Nature Neuroscience study: over 3 months, a high-flavanol diet improved blood flow in one part of the hippocampus on brain imaging along with memory scores in adults aged 50 to 69. It was a small early-stage study, one author was affiliated with Mars Inc., and cocoa research is heavily industry funded. Treat any memory benefit as unproven: the one large randomized trial was negative.
Exercise performance (mechanism only, not tested in the cited studies)
In laboratory work, epicatechin from cocoa can activate eNOS, the enzyme that makes nitric oxide, which in theory would widen blood vessels and send more blood to working muscle. That is a mechanism, not a result. None of the studies cited on this page measured endurance, oxygen cost or recovery, so there is no evidence here that this extract improves athletic performance.
Blood sugar and metabolism (not measured in the cited studies)
Effects on insulin sensitivity and glucose handling (AMPK activation, insulin signaling in blood vessels, lower inflammatory signals) are ideas from laboratory work, not findings from the studies cited here, none of which measured blood sugar or insulin. Diabetes-risk findings from observational research describe an association with people's usual eating patterns and cannot show that taking a cocoa flavanol supplement lowers anyone's risk.
Mechanism of action
eNOS activation and nitric oxide production
In laboratory studies, epicatechin and procyanidins from cocoa activate endothelial nitric oxide synthase (eNOS) through PI3K/Akt signaling — increasing NO production in vascular endothelium. NO drives vasodilation, reduces platelet aggregation, inhibits vascular smooth muscle proliferation, and improves blood flow. This is the leading explanation for the small blood pressure and blood vessel changes seen in trials. It remains a proposed mechanism, and no study cited here shows that it translates into athletic performance benefits.
Clinical trials
Randomized, double-blind, placebo-controlled trial of cocoa extract (500 mg/day cocoa flavanols including 80 mg epicatechin) versus placebo in 21,442 older US adults, with a median follow-up of 3.6 years. Main goal: total cardiovascular events. (Sesso et al. 2022, Am J Clin Nutr, COSMOS)
21,442 US adults (women 65 and older, men 60 and older). Median follow-up 3.6 years.
The trial did not meet its main goal. Total cardiovascular events were not significantly reduced (hazard ratio 0.90, 95% confidence interval 0.78 to 1.02, P = 0.11). Heart attack (0.87), stroke (0.91), artery-opening procedures (0.95) and death from any cause (0.89, 0.77 to 1.03) were all non-significant. In the authors' own words, cocoa extract supplementation did not significantly reduce total cardiovascular events among older adults but reduced cardiovascular death by 27% (hazard ratio 0.73, 0.54 to 0.98), a result from one of several secondary endpoints. A more favorable figure sometimes quoted (hazard ratio 0.85) comes from an analysis that drops people once they stop taking the capsules, which breaks the randomization and can make the supplement look better than it is. By the standard measure this is a negative trial, despite its size.