The flavanols in cocoa lower blood pressure by about two points, more in people who start high, and within an hour or two of a dose they relax the artery lining so it widens. Those short-term effects rest on dozens of controlled trials.
The distinction that governs the whole topic is that the benefit belongs to the flavanol, not to chocolate: most bars carry little flavanol and a lot of sugar and fat. The largest long-term trial used a concentrated cocoa flavanol supplement, not candy, and it missed its main heart target while lowering cardiovascular death in a secondary measure that needs its own trial to confirm.
Findings & Outcomes
What It Is
Cocoa flavanols are a family of plant compounds, the flavan-3-ols, found in the cocoa bean and in the same class of chemicals that give green tea, apples, and berries their share of the benefit. The best studied of them is (-)-epicatechin, the molecule most of the vascular research tracks. They are not unique to chocolate.
The amount of flavanol in a food depends heavily on how it was grown and processed. Fresh cocoa is rich in flavanols; roasting, fermenting, and especially the alkalizing (Dutch-processing) that turns a bean into a smooth dark bar strip most of them out, while the sugar and fat stay. So this topic is about the flavanol, not about chocolate as a food.
Anatomy of the Practice
1Within hours
A single flavanol-rich dose raises nitric oxide in the blood-vessel lining, and the arteries relax and widen. This shows up as improved flow-mediated dilation within one to two hours of eating it, the fastest and most reproducible effect in the research.
2Over two to eighteen weeks
With daily flavanol-rich cocoa, blood pressure settles a little lower, by about two points on average and more in people who began high. This is the window the blood-pressure trials measured, and the effect is small but consistent.
3Over years
Whether the short-term vascular improvements add up to fewer heart attacks and strokes was tested directly in one large trial, and the main answer was no clear reduction, with a signal on cardiovascular death and on memory in people who ate few flavanols to begin with. The long-term payoff is smaller and less certain than the quick vascular numbers suggest.
How It Works
The mechanism is unusually well understood for a food compound. Flavanols, and epicatechin in particular, increase the amount of nitric oxide the endothelium (the single-cell lining of the blood vessels) can make and keep available. Nitric oxide is the signal that relaxes the vessel wall, so more of it means arteries that dilate more easily, which is what the flow-mediated dilation measurements show within hours of a dose, and it is the most credible route to the small blood-pressure drop seen over weeks. Trials that included more than about 50 mg of epicatechin a day produced the larger blood-pressure effects, the dose-response you would expect if the mechanism is causal.
The dose that produced these effects in the trials is a flavanol dose: around 670 mg of flavanols a day on average in the blood-pressure studies, and a standardized 500 mg (with 80 mg of epicatechin) in the large long-term trial. A typical chocolate bar, even a dark one, is an unreliable and usually poor source of that, because processing removes flavanols and the bar delivers sugar, fat, and calories alongside whatever is left.
Reading "cocoa lowers blood pressure" as "eat more chocolate" is the most common way this evidence gets misused.
The food-first case runs through whole foods, where cocoa is one vascular-active food among many.
What Changed
For years the message was simple: chocolate is good for your heart. It rested on short-term trials showing lower blood pressure and better artery function, and on population studies where people who ate more cocoa had less heart disease. People who eat chocolate also differ from those who do not in income, activity, and diet, so that population pattern is an association, not proof the cocoa did the work. The short-term vascular results were encouraging on their own terms. No one had yet run the long trial: does taking cocoa flavanols for years prevent heart attacks, strokes, and deaths?
That trial was run. COSMOS gave 21,442 older adults a daily cocoa extract standardized to 500 mg of flavanols, or a placebo, and followed them for a median 3.6 years. Its main result, a composite of total cardiovascular events, did not reach statistical significance (hazard ratio 0.90, 95% CI 0.78 to 1.02, P=0.11). The effect leaned toward benefit, and among participants who took their pills as directed the reduction was significant (hazard ratio 0.85); the primary analysis, the one the trial was built around, came up short. One of its secondary measures, cardiovascular death, was 27% lower in the cocoa group (hazard ratio 0.73, 95% CI 0.54 to 0.98).
The cognition results ran the same way. In COSMOS-Mind, cocoa extract did not change overall thinking and memory in older adults over three years. In COSMOS-Web, the supplement did not improve memory across everyone, but it did restore memory in the subgroup who ate the fewest flavanols to begin with, and rising flavanol levels in the urine tracked with improving memory. The benefit sat where a deficiency was being corrected, in people whose diets were low in these compounds, which fits cocoa flavanols better than a blanket promise of sharper memory.
That leaves three separate verdicts, each at the strength of its own evidence:
- Blood pressure and artery function, at the flavanol doses the trials used: reasonably solid over the short term.
- Cardiovascular events over years: the largest trial did not show a benefit on its main measure, with a secondary signal on cardiovascular death that still needs its own trial to confirm.
- Memory and cognition: not an across-the-board effect, and clearest in people whose diets were low in flavanols.
None of it is a reason to eat more sugar.
Ways to Do It
The goal is flavanols, not chocolate. Most of what you can buy in a candy aisle is a poor source, so the rungs below run from the free food-first version to the one standardized way to hit the dose the trials used.
Cocoa is one source among several. Green tea, apples, and berries carry flavan-3-ols from the same family, and a varied diet with these in it covers the compound class without any purchase or supplement. This is the fair answer for most people, and it is the diet pattern the population studies were describing.
Natural, non-alkalized cocoa powder keeps far more of its flavanols than Dutch-processed (alkalized) cocoa, so check the label for which one you are buying. A tablespoon stirred into oats, milk, or a smoothie gives you the compound without the sugar and fat of a bar. It tastes bitter, and that bitterness is the flavanols.
A higher cocoa percentage tends to mean more flavanols, but processing varies so much that no bar reliably delivers a trial-sized flavanol dose, and every bar delivers sugar and fat. Treat dark chocolate as a food you enjoy that carries some flavanol, not as a way to dose yourself for your blood pressure.
The only dependable way to take the amount the trials used is a supplement standardized to its flavanol and epicatechin content, such as the 500 mg of flavanols with 80 mg of epicatechin used in COSMOS. This reproduces the studied dose instead of guessing at it. It is a considered purchase; weigh it against what that trial actually found, in the findings below.
The Evidence, Graded
Each finding below is graded at the strength of its own evidence. The blood-pressure and artery-function effects rest on meta-analyses of many short trials and sit highest; the event, mortality, and memory results come from the single large supplement trial and are graded lower and more tentatively.
The Research & Studies
Everything here is based on the research we have collected and checked, sorted into groups and ordered with the strongest evidence first. Click any claim to open the studies behind it.
Heart And Vascular
Flavanol-rich cocoa lowers blood pressure about 2 mmHg, near 4 in people who start high
Eating flavanol-rich cocoa nudged blood pressure down by about 2 points on average, a systolic drop of 1.76 mmHg, and closer to 4 points in people who started with high blood pressure.
Cochrane review of 35 randomized trials (40 comparisons, 2 to 18 weeks, mean 9 weeks) in 1,804 mainly healthy adults, mean flavanol dose 670 mg/day. Graded moderate quality; heterogeneity was unexplained by prespecified subgroups, and a sensitivity analysis excluding trials with industry-employed authors reduced the effect size.
Who this may not transfer to:Meta-analysis of 1,804 mostly healthy adults of both sexes over 2 to 18 weeks. The 4 mmHg figure is the hypertensive subgroup; a reader with normal blood pressure should expect the smaller 2 mmHg end.
The study · 1
Ried 2017, Cochrane Database Syst Rev · Cochrane Database Syst Rev
Cocoa relaxes the artery lining: flow-mediated dilation rises about 3% within hours, 1.3% over weeks
Cocoa made the lining of blood vessels relax and widen more, a sign of healthier arteries, both within hours of eating it and over weeks.
Meta-analysis of 42 acute or short-term (up to 18 weeks) randomized trials in 1,297 participants. GRADE rated the evidence low to moderate, lowered for unclear allocation concealment and biomarker heterogeneity. Doses above 50 mg epicatechin/day produced larger blood-pressure effects.
Who this may not transfer to:1,297 adults of both sexes across 42 short trials. Flow-mediated dilation is a marker of arterial function measured over hours to 18 weeks, so it speaks to how the vessels behave, not to whether events fall over years.
The study · 1
Hooper 2012, Am J Clin Nutr · Am J Clin Nutr
COSMOS did not cut total cardiovascular events (HR 0.90, P=0.11) over 3.6 years
In the largest trial, a daily cocoa flavanol supplement did not clearly cut the overall rate of heart attacks, strokes and related events, though the result leaned toward benefit and was significant among people who took it as directed.
COSMOS randomized 21,442 US adults (12,666 women aged 65+ and 8,776 men aged 60+) free of major cardiovascular disease and recent cancer to cocoa extract or placebo. The primary composite included MI, stroke, coronary revascularization, cardiovascular death, carotid artery disease, peripheral artery surgery, and unstable angina.
Who this may not transfer to:21,442 US adults, women 65+ and men 60+, free of major cardiovascular disease at entry. The null describes this older, initially healthy group taking a supplement, and does not speak to younger or higher-risk people.
The study · 1
Sesso 2022, Am J Clin Nutr (COSMOS) · Am J Clin Nutr
Counts once: this finding and 1 other here come from the same source, so they are one body of evidence, not separate confirmations.
Cardiovascular death fell 27% in COSMOS (HR 0.73), a secondary endpoint
In the same trial, fewer people in the cocoa group died of heart disease, though this was one of several secondary measures rather than the main result the trial was built to test.
Prespecified secondary endpoint of COSMOS among 21,442 older US adults over a median 3.6 years. Other individual secondary endpoints (MI HR 0.87, stroke 0.91, revascularization 0.95) were neutral.
Who this may not transfer to:Same 21,442 older US adults as the primary analysis. This is one prespecified secondary endpoint among several, so it needs a trial built to test it before it carries to practice.
The study · 1
Sesso 2022, Am J Clin Nutr (COSMOS) · Am J Clin Nutr
Counts once: this finding and 1 other here come from the same source, so they are one body of evidence, not separate confirmations.
Cognition
Cocoa extract did not improve global cognition over 3 years (COSMOS-Mind)
A daily cocoa flavanol supplement did not improve overall thinking and memory in older adults over three years.
COSMOS-Mind, a substudy of 2,262 older adults (mean age 73, 60% women) using telephone-administered cognitive testing over 3 years. This is a substudy of a trial powered for cardiovascular and cancer outcomes.
Who this may not transfer to:2,262 older adults, mean age 73, 60% women, tested by telephone. The null is for overall thinking and memory in this general older sample.
The study · 1
Baker 2023, Alzheimers Dement (COSMOS-Mind) · Alzheimers Dement
Memory improved only in the lowest-flavanol subgroup; the overall endpoint was null (COSMOS-Web)
The supplement did not improve memory across everyone, but it did help the subgroup who ate the fewest flavanols to begin with, suggesting the benefit is about filling a gap rather than adding a boost.
COSMOS-Web randomized 3,562 older adults to 500 mg cocoa flavanols/day or placebo for 3 years, with a urine flavanol biomarker in 1,361 of them. Habitual flavanol intake and diet quality correlated with hippocampal-dependent memory at baseline.
Who this may not transfer to:3,562 older adults of both sexes. The memory gain applies to the subgroup who ate the fewest flavanols to begin with, not to the whole group, whose primary endpoint was null.
The study · 1
Brickman 2023, Proc Natl Acad Sci U S A · Proc Natl Acad Sci U S A
Go Deeper
- Whole foods: the food-first case, where cocoa is one flavanol source among many.
- Omega-3 and fish oil: another supplement whose short-term markers looked strong and whose long-term event trials came in mixed, and a useful comparison for reading this kind of evidence.
The Chinese Medicine View
Cocoa is a New World food and has no place in the classical Chinese materia medica, so nothing here is a traditional use. What the tradition offers is a way of reading any food by its flavor, its temperature, and what it does to a given constitution, and cocoa can be looked at through that lens as interpretation, held apart from the trial data above.
By taste, cocoa is bitter, and the bitter flavor is associated in the tradition with draining and drying and with clearing heat downward. The bitterness of raw cocoa is one thing; the chocolate most people actually eat is another, because sugar and rich fat turn it into a sweet, cloying, moistening food, and the tradition reads heavy sweet rich food as damp-forming and phlegm-forming.
For someone already damp or phlegm-laden, sluggish, heavy, prone to a coated tongue, that kind of food is understood to feed the accumulation rather than the person. Cocoa also carries caffeine and theobromine, mild stimulants that read as warming and stirring, which the tradition would not hand freely to someone already hot, restless, or short on sleep.
The traditional reading and the modern one meet, without merging, on one point: the difference between the flavanol compound and the sweet bar. A tradition that weighs a food by who is eating it would not treat a bitter, drying, flavanol-rich cocoa the same as a sweet, damp-forming bar, any more than the research treats the flavanol dose the same as the sugar. It would ask who is in front of you: the person who could use a little bitter, downward-moving quality, or the damp, sweet-loving person for whom another rich sweet food is the last thing needed.
That instinct, that the right form and amount depend on the person, sits beside the modern reading, where the benefit belongs to the flavanol and the harm to the sugar. If you work with a practitioner, this is a reasonable thing to ask about.
Cautions For This Practice
Everything to be aware of is here, in one place. This practice suits most healthy people; a few situations call for real care.
The sugar and fat in most chocolate
This is the practical one. The vascular benefits belong to the flavanols, while most chocolate delivers them alongside a large amount of sugar, fat and calories, and Dutch-processed (alkalized) cocoa has had most of its flavanols removed. Eating more chocolate to chase a two-point blood-pressure drop can easily cost more than it returns. If the goal is flavanols, unsweetened natural cocoa or a standardized supplement gets you there without the sugar.
Caffeine and theobromine
Cocoa contains caffeine and a related stimulant, theobromine. Dark chocolate and cocoa powder carry more than milk chocolate, enough to affect sleep, reflux or a sensitive stomach in some people, and enough to matter if you are already limiting caffeine. Theobromine is also toxic to dogs and cats at doses they reach easily, so keep cocoa and chocolate away from pets.
Heavy metals in some dark chocolate and cocoa
Independent testing has repeatedly found cadmium and lead in a share of dark chocolate and cocoa products, with cadmium taken up from soil by the cacao plant and lead arriving mostly after harvest. For an occasional treat this is a minor concern, but it matters more if you plan to eat dark chocolate or cocoa daily, so it helps to vary brands rather than lean on one.
Blood pressure that is already low or medicated
The blood-pressure effect is small, so for most people it is not a problem. If you already run low, or take medication for high blood pressure, a daily flavanol dose adds a small, predictable downward nudge to raise with your clinician.
A migraine trigger for some people
A minority of people with migraine find chocolate can precede an attack, though whether it triggers the headache or is a craving during the early phase of one is unsettled. If you notice a pattern for yourself, that observation is worth more than the group data.
Talk to a clinician before daily high-dose use if you are unwell or pregnant
A daily standardized supplement, or use aimed at a specific condition, is worth raising with a clinician who knows your situation, especially in pregnancy or alongside other treatment, so the dose and your own circumstances get weighed together.
Start slow, be smart, read the research, and consult a professional if you have any concerns. This is here to inform your choice, not make it for you.
Common Questions
Does eating dark chocolate actually lower blood pressure?
The flavanols in cocoa do, by a small amount. Across 35 short trials, flavanol-rich cocoa lowered blood pressure by about two points on average, and by around four points systolic in people who started high. The problem is that most chocolate bars are a weak and unreliable source of flavanols and a strong source of sugar and fat. So the effect holds for high-flavanol cocoa, while eating ordinary dark chocolate for your blood pressure is unlikely to be a good trade.
How much cocoa flavanol do the trials actually use?
More than a bar provides. The blood-pressure trials averaged around 670 mg of flavanols a day, and the large long-term trial used a supplement standardized to 500 mg of flavanols with 80 mg of epicatechin. Because roasting and especially alkalizing strip flavanols out, you cannot reliably reach that from chocolate, and even unsweetened cocoa varies. A supplement labeled for its flavanol and epicatechin content is the only dependable way to take the studied amount.
Did the big trial prove cocoa prevents heart disease?
No, not on its main measure. COSMOS followed 21,442 older adults for about 3.6 years and found no statistically significant drop in its primary measure of total cardiovascular events, though the result leaned toward benefit and was significant among people who took the pills as directed. One secondary measure, cardiovascular death, was 27% lower. A secondary finding like that can point the way, but it needs a trial built to test it before it counts as established.
Is cocoa good for memory and thinking?
The picture is mixed and points to filling a gap rather than a boost. A daily cocoa flavanol supplement did not improve overall cognition in older adults over three years, and it did not improve memory across everyone in a second large trial. It did restore memory in the subgroup who ate the fewest flavanols to begin with, and their memory tracked with rising flavanol levels. So the likeliest benefit is for people whose diets are low in these compounds, not an across-the-board enhancement.
Explore Related
Other pages this one connects to, by the evidence they share, the outcomes they touch, and the ground they cover.
All 6 sources on this page independently checked and cross-referenced.
Thomas Dehli, Founder & Editor, Sacred Lotus
Sacred Lotus has published Chinese medicine reference material since 2001. Integrative pages are held to the same standard as the herb and formula library: cite the source, grade the claim at its real strength, and say where the research has not looked. This page is educational and it is not medical advice. Last reviewed and updated August 9, 2026.
Evidence strength
How confidently the research supports a claim. Strength describes the evidence, not our endorsement.