What are cocoa flavanols? A plain-English guide

Cocoa is rich in flavanols, which are one group within a larger family of substances called flavonoids [1]. The COSMOS team says cocoa comes from cacao tree beans, and that tea, grapes and wine also have flavanols [5].

Mars Edge, the nutrition research arm of the company Mars, gave COSMOS a grant and donated the study pills [3].

A Cochrane review notes flavanols have been shown to help blood vessel walls make a substance that widens the vessels [2]. That is a theory, not proof that a supplement works.

Researchers have studied cocoa flavanols for memory, thinking and blood pressure [2][6].

The Mars Edge-backed COSMOS team says the amount of flavanols in cocoa foods can vary a lot [5]. Their trial used capsules of cocoa extract instead [5].

FDA says it does not intend to object to certain health claims about high-flavanol cocoa powder in foods [7]. But it says the evidence is very limited, and the claims do not apply to chocolate [7].

Most studies below are randomized controlled trials. In these, people are split by chance into groups, and one group gets a placebo, a dummy pill. A meta-analysis pools the results of many trials on one question. "No clear" effect means the result was not statistically significant and could be chance.

Do cocoa flavanols help memory?

Evidence grade for episodic memory: C, limited. Why this grade

COSMOS-Web was a Mars Edge-backed trial of 3,562 older adults [3]. They took cocoa extract with 500 mg of flavanols a day, or a placebo, for 3 years [3]. Two authors worked for Mars, and two had a research grant from Mars [3]. After 1 year, it found no clear overall memory benefit compared with placebo [3].

Memory did improve in people whose usual diet was lower in quality or in flavanols [3]. But that was a subgroup, not the main result [3].

Two other parts of COSMOS, also backed by Mars Edge, found no clear memory benefit compared with placebo after 2 to 3 years [4][8].

Do cocoa flavanols help with thinking speed and planning?

Evidence grade for executive function: C, limited. Why this grade

A 2026 meta-analysis pooled 16 randomized controlled trials with 2,847 people, including healthy adults, older adults and people with health conditions [6]. In it, regular dark chocolate or cocoa improved results on a timed connect-the-dots test and a word-naming test [6].

But only 8 of those 16 trials found an effect on at least one area of thinking [6]. Two large COSMOS trials in older adults, backed by Mars Edge, found no clear effect on planning and attention scores [4][8].

Do cocoa flavanols lower blood pressure?

Evidence grade for blood pressure: B, moderate. Why this grade

A Cochrane review pooled 35 trials with 1,804 mainly healthy people [2]. Compared with control products, flavanol-rich cocoa lowered blood pressure slightly, by about 2 mmHg, in trials of 2 to 18 weeks [2].

In people whose blood pressure was already normal, there was no clear difference [2]. The reviewers rated the evidence as moderate quality, because results varied between trials in ways they could not explain [2]. The effect shrank slightly without trials that had authors from the sponsor, a sign of some reporting bias [2]. Two reviewers had run trials in the review [2].

A 12-week trial run by Mars researchers gave 74 healthy adults with normal blood pressure up to 2000 mg a day or a flavanol-free control [9]. It found no clear change in blood pressure between groups [9].

To see what each grade on this page means, read how we grade evidence.

Table: Doses used in studies (not a recommendation)

These are amounts used in research, not advice. Talk to a clinician before using any supplement.

In the Cochrane review, cocoa was well tolerated [2]. Nausea and stomach complaints were reported by 1% of people taking cocoa and 0.4% of people in control groups [2].

In the main COSMOS trial, backed by Mars Edge, 21,442 older US adults took 500 mg of flavanols a day or a placebo [5]. They were tracked for a median of 3.6 years [5]. The team reported no safety concerns, though nausea was slightly more common with cocoa extract [5].

In a 12-week trial run by Mars researchers, 46 healthy adults took up to 2000 mg a day and 28 took a control [9]. Side effects were mild, with no clear difference between groups [9]. The Cochrane reviewers say longer trials are still needed to check long-term safety [2].

As of November 11, 2026 we found no controlled trials of how cocoa flavanol supplements interact with medicines (PubMed and Europe PMC search). In a 1-week pilot study with no comparison group, 20 people with stable heart disease ate 30 g of dark chocolate a day [10]. They were taking two medicines that make blood less likely to clot, aspirin and clopidogrel [10]. Cocoa boosted clopidogrel's effect on a lab test of platelets, which help blood clot, but not aspirin's [10]. No serious side effects occurred [10]. The study used cocoa beans donated by Ortinola Estates Limited [10]. Results may not apply to healthy adults.

If you take any of these medicines, talk to your doctor or pharmacist before combining them with cocoa flavanols.

The Mars Edge-backed COSMOS trial left out people who reported extreme sensitivity to caffeine [5]. Its cocoa extract held about 50 mg of theobromine and 15 mg of caffeine a day [5].

We checked LiverTox (as archived November 5, 2026) and the NIH Office of Dietary Supplements (as archived November 1, 2026). As of November 12, 2026, neither has a cocoa entry. Our sources give no safety information for pregnancy, breastfeeding or children. Chocolate trials in pregnancy were not reviewed here. Ask a clinician first if you are in one of these groups, take medicines or have a health condition.