Benefits
Blood pressure in short-term trials
A Cochrane review of 35 randomized trials (1,804 mainly healthy adults, 2 to 18 weeks, chocolate or cocoa products) found about 1.8 mmHg lower systolic and diastolic pressure, rated moderate-quality evidence. A wider analysis of 145 trials of flavan-3-ol foods found about 2.8/2.0 mmHg. In COSMOS, a median 3.4 years of cocoa extract did not lower the overall rate of new high blood pressure.
Blood vessel function (flow-mediated dilation)
Flow-mediated dilation measures how well an artery widens. A meta-analysis of 18 trial arms found cocoa flavanols raised it by 1.17% on average, with the largest effect near 710 mg a day, though most studies had risk-of-bias concerns. A 1-month drink trial in 100 healthy adults (450 mg twice daily) found a similar gain. The EU accepts 200 mg a day for helping maintain blood vessel elasticity.
Long-term heart outcomes in older adults (COSMOS)
COSMOS gave 21,442 older US adults a cocoa extract capsule (500 mg flavanols a day) or placebo for a median 3.6 years. Its main outcome, total cardiovascular events, was not significantly reduced (hazard ratio 0.90, P = 0.11). As a secondary outcome, cardiovascular deaths were 27% lower; heart attack, stroke and death from any cause did not differ significantly. Mars Edge partly funded the trial and supplied the capsules.
Memory and thinking in older adults
Three COSMOS studies at 500 mg flavanols a day found no overall gain in cognition: 2,262 people tested by phone for 3 years and 573 tested in person for 2 years. In 3,562 people the main 1-year memory result was not significant; memory improved only in those with lower habitual diet quality or flavanol intake, a subgroup finding. A smaller 8-week drink trial in 90 older adults reported better scores on some tests.
Cholesterol and blood sugar markers
Results are inconsistent. A 1-month drink trial in 100 healthy adults lowered LDL cholesterol by 0.17 mmol/L and raised HDL by 0.10 mmol/L, and a 1-month capsule trial in 45 young men lowered total cholesterol, but a 4-week capsule trial in 75 young adults found no lipid change. For blood sugar, COSMOS found no reduction in new type 2 diabetes among 18,381 people over 3.5 years.
Mechanism of action
Epicatechin and nitric oxide in the artery lining
Cocoa flavanols are thought to increase nitric oxide made by the artery lining, which relaxes blood vessels. In a 1-month capsule trial in 45 young men, an extract with 130 mg epicatechin plus procyanidins improved flow-mediated dilation and blood pressure, while one with mostly procyanidins (20 mg epicatechin) did not; the gain tracked blood levels of epicatechin metabolites.
Procyanidins and blood lipids
In the same capsule trial, total cholesterol fell with both the epicatechin-rich and the procyanidin-rich extract, suggesting the longer procyanidin chains, not epicatechin alone, drove that change. This is one small trial in young men.
Theobromine and caffeine come along with the extract
Cocoa extracts also contain methylxanthines. The COSMOS capsules supplied about 50 mg theobromine and 15 mg caffeine a day, and the investigators note they cannot separate these from the flavanols. A 4-week trial testing epicatechin, methylxanthines or both in young adults found no effect on arterial stiffness from any of them.
Clinical trials
Randomized, double-blind, placebo-controlled 2-by-2 factorial trial of a cocoa extract capsule (500 mg flavanols a day, including 80 mg epicatechin) and a multivitamin; median follow-up 3.6 years. Funded in part by Mars Edge, which supplied the capsules (Sesso et al. 2022, Am J Clin Nutr)
21,442 US adults free of major cardiovascular disease (12,666 women aged 65 or older and 8,776 men aged 60 or older).
Total cardiovascular events, the primary outcome, were not significantly reduced: 410 on cocoa extract vs 456 on placebo (hazard ratio 0.90, 95% CI 0.78 to 1.02, P = 0.11). Among secondary outcomes, cardiovascular death was lower (hazard ratio 0.73, 0.54 to 0.98), while heart attack, stroke and all-cause death did not differ significantly. A per-protocol analysis limited to adherent participants gave a hazard ratio of 0.85, but this breaks randomization. No safety concerns were found.
Cochrane systematic review and meta-analysis of randomized trials lasting at least 2 weeks comparing flavanol-rich chocolate or cocoa products with flavanol-free or low-flavanol controls (Ried et al. 2017, Cochrane Database Syst Rev)
1,804 mainly healthy adults in 35 trials (40 comparisons), given 30 to 1,218 mg flavanols a day (mean 670 mg) for 2 to 18 weeks.
Systolic and diastolic pressure were each about 1.76 mmHg lower than control, rated moderate-quality evidence. Systolic pressure fell about 4 mmHg in people with high blood pressure but not significantly in those with normal pressure (subgroup difference P = 0.08). Removing trials with industry-employed authors shrank the effect slightly. Gut complaints or nausea affected about 1% on cocoa vs 0.4% on control.
Systematic review and dose-response meta-analysis of randomized trials measuring flow-mediated dilation after cocoa flavanol intake; authors from the Nestle Research Centre (Sun et al. 2019, Food Funct)
15 published studies with 18 intervention arms; active groups received 80 to 1,248 mg (mean 704 mg) more flavanols than controls.
Flow-mediated dilation improved by 1.17% (95% CI 0.76 to 1.57). The dose-response curve was an inverted U, with the largest effect near 710 mg total flavanols (about 95 mg epicatechin) a day. Much of the variation between studies was unexplained, and most studies had risk-of-bias concerns.
Randomized, controlled, double-masked, parallel-group trial of a cocoa flavanol drink (450 mg) or a nutrient-matched flavanol-free drink twice daily for 1 month; one co-author was affiliated with Mars (Sansone et al. 2015, Br J Nutr)
100 healthy men and women aged 35 to 60 with no cardiovascular disease.
Flow-mediated dilation, the primary outcome, rose by 1.2% over control (95% CI 1.0 to 1.4). Blood pressure fell by 4.4/3.9 mmHg, LDL cholesterol by 0.17 mmol/L and total cholesterol by 0.20 mmol/L, while HDL rose by 0.10 mmol/L. This tested a drink, not a capsule, and lasted only a month.
Randomized, controlled, double-masked, parallel-group trial of two cocoa extract capsules and a flavanol-free control capsule taken once daily for 1 month; two co-authors were affiliated with Mars (Rodriguez-Mateos et al. 2018, Am J Clin Nutr)
45 healthy men aged 18 to 35.
The extract with 130 mg epicatechin and 560 mg procyanidins increased flow-mediated dilation (the primary outcome) and lowered pulse wave velocity and blood pressure; the extract with 20 mg epicatechin and 540 mg procyanidins did not. Total cholesterol fell with both extracts. The authors link the vascular effect mainly to epicatechin.
Randomized, placebo-controlled 5-arm trial of daily capsules for 4 weeks: flavanol-rich cocoa powder (at least 200 mg flavanols), epicatechin, methylxanthines, epicatechin plus methylxanthines, or placebo (Weigant et al. 2025, Mol Nutr Food Res)
75 healthy young nonsmoking adults (15 per group).
No group changed aortic pulse wave velocity (P = 0.410), blood pressure, endothelin-1 or blood lipids compared with placebo. The cocoa dose matched the amount in the EU blood vessel elasticity claim, and the groups were small.
Randomized, placebo-controlled 2-by-2 factorial 3-year trial of cocoa extract (500 mg flavanols a day) and a multivitamin, with cognition tested by telephone each year (Baker et al. 2023, Alzheimers Dement)
2,262 older US adults (mean age 73, 60% women).
Cocoa extract had no effect on the primary global cognition score (mean z-score 0.03, 95% CI -0.02 to 0.08, P = 0.28). The multivitamin arm, a secondary comparison, showed a small benefit; cocoa did not.
Randomized, placebo-controlled study within COSMOS of cocoa extract (500 mg flavanols a day) for up to 3 years, with a urine flavanol biomarker in a subset; two authors were Mars employees and two received a Mars research grant (Brickman et al. 2023, Proc Natl Acad Sci U S A)
3,562 older adults; biomarker measured in 1,361.
The prespecified primary outcome, a memory gain in all participants after 1 year, was not statistically significant. Memory improved in participants in the lower thirds of habitual diet quality or flavanol intake, and rises in the flavanol biomarker tracked memory gains. These subgroup results need confirmation.