Dark Chocolate and Cacao — Body Handbook

Food · §261

Dark Chocolate and Cacao

A square of 70%+ dark chocolate after dinner is the rare daily pleasure that pays you back. The flavanols in cocoa nudge resting blood pressure down by a few mmHg within weeks, relax the lining of your blood vessels so the next flight of stairs is a beat easier, and lift mood a little on the day — all from a substance you would happily eat anyway. The catch is portion, alkalization, and what you do with the calories: stay under ~30 grams, pick a high-cocoa, non-Dutched bar, and the rest of the entry is how to make the dose actually land.

Do · Daily Evidence Moderate Chapter Food

The strongest signal is on resting blood pressure — a Cochrane review of 35 trials puts it at around −1.8 mmHg systolic at modest daily doses, more if yours runs high. Mood and memory get a small, dose-dependent nudge from the same flavanol pathway, larger in older adults. The cost is pennies; the effort is eating chocolate. The honest catches: a milk-chocolate bar doesn't carry the benefit, "Dutch-process" cocoa loses up to 90% of its flavanols, and 80 grams in front of a screen drowns the signal in sugar and calories.

The active ingredient in cocoa is not "antioxidants" — that framing is wrong in the way most polyphenol marketing is wrong. The active ingredient is a small flavanol called (-)-epicatechin, which gets absorbed in your small intestine within an hour or two and tells the cells lining your blood vessels to make more nitric oxide. Nitric oxide relaxes those vessels; the vessels open up; blood pressure drops a touch and flow improves. Researchers tracked this with a radioactive epicatechin tracer and confirmed that the molecule itself is doing the work — when they isolated it from chocolate and fed it pure, it reproduced the full vascular effect of a flavanol-rich cocoa drink, and blocking nitric oxide production cancelled the effect Schroeter et al. 2006.

Cocoa also delivers a methylxanthine called theobromine — a slower, gentler cousin of caffeine. A 25-gram square of 70% chocolate carries roughly 170 to 200 milligrams of it, plus about 20 mg of caffeine, which is the source of the quiet mood and alertness lift you notice after eating it — not jittery like a coffee, but real Pase et al. 2013. Magnesium, copper, iron, and zinc are along for the ride; a daily serving covers a meaningful slice of your magnesium requirement, which most adults under-eat.

What the data actually shows

The most replicated effect, by a wide margin, is the drop in resting blood pressure. A Cochrane review pooled 35 randomised trials with around 1,800 people and found that two or more weeks of daily flavanol-rich cocoa cut systolic pressure by roughly −1.8 mmHg and diastolic by about the same — closer to −4 mmHg systolic in people who started out hypertensive Ried et al. 2017. The flow-mediated dilation work (a measurement of how much your arteries can open in response to a flow stimulus) lines up: a meta-analysis of 42 trials reported a real, repeatable improvement of around a percentage point of FMD, along with lower insulin resistance and a small drop in LDL Hooper et al. 2012.

The cohort data — people followed for years, asked what they eat, then watched for heart attacks — points the same direction. A 2017 meta-analysis of 14 prospective studies found that the highest chocolate consumers (about one to three servings per week) had 10 to 20% lower incidence of coronary heart disease, stroke, and diabetes than the lowest consumers Yuan et al. 2017. Cohorts cannot prove causation — people who eat dark chocolate also tend to be richer, more educated, and eat better in general — but the size of the signal across many populations is hard to dismiss Buitrago-Lopez et al. 2011.

On cognition: a trial called CoCoA gave older adults with mild cognitive impairment a high-flavanol drink (~990 mg/day) for eight weeks; the high-dose arm improved measurably on trail-making and verbal fluency, with parallel drops in blood pressure and insulin resistance Mastroiacovo et al. 2015. In a separate three-month study, cognitively normal older adults on the same dose showed restored blood-flow signal in the dentate gyrus (the memory-encoding part of the hippocampus) on functional MRI, alongside better object-recognition memory Brickman et al. 2014. These are real but they used flavanol doses higher than a normal chocolate bar provides — closer to a heaped tablespoon of natural cocoa powder daily.

Regulators have read all of this cautiously and partially. The European food-safety body approved a claim that 200 mg/day of cocoa flavanols helps maintain endothelium-dependent vasodilation EFSA 2012. The U.S. FDA, in 2022, allowed high-flavanol cocoa products to carry a qualified claim that the science suggests, but does not prove, a reduced risk of cardiovascular disease FDA 2022. "Qualified" is the operative word — it is the regulator saying "real, not certain."

How to actually do it

The dose that earns the benefits in trials clusters in two ranges. About 200 mg of cocoa flavanols a day is the threshold the European regulator settled on for keeping blood vessels reactive; the strongest blood-pressure and cognitive results came from doses two to five times higher than that EFSA 2012, Mastroiacovo et al. 2015. Translating to something you actually eat is genuinely hard, because flavanol content of supermarket chocolate varies wildly — different cocoa origin, roast, and processing can change it by an order of magnitude Miller et al. 2008.

Higher cocoa percentage roughly tracks higher flavanol content within a single producer's line, but it is not a reliable signal across brands. A 70% bar from one maker can carry more flavanols than an 85% bar from another. Where producers list a flavanol assay or carry the "Cocoapro" or "CocoaVia" branding (the high-flavanol formulations used in the trials), that's a closer match to what the science measured.

The portion is the boring part of the protocol and the part that decides whether the entry pays off. A 25-gram serving of 70% chocolate is about 140 kilocalories and 8 grams of sugar — manageable. An 80-gram bar in front of a screen is 450 calories and 25 grams of sugar that has to come out of somewhere else in your day, and at that point the net signal flips negative.

What most people get wrong

The "antioxidant" story is mostly wrong. Cocoa polyphenols never reach high enough levels in your blood to be soaking up free radicals at any meaningful scale. The way they actually work is by switching on a signalling pathway that makes your blood vessels produce more nitric oxide. The word "antioxidant" on the packaging is selling you the wrong mechanism, and it doesn't tell you whether the product is any good Ottaviani et al. 2018.

"Raw cacao" is not the answer. The thing that destroys flavanols is the alkalization step (the so-called Dutch process), not heat. A normal, well-fermented, low-roast 70% bar can be higher in flavanols than a raw, expensive cacao product whose cocoa beans were just from a different chocolate-friendly origin Miller et al. 2008.

The percentage on the label is not a flavanol score. Two 85% bars from different brands can differ by a factor of five in actual flavanol content. Until producers print an assay, percentage is a rough proxy at best.

Milk chocolate does not inherit the trial benefits. Roughly a third the flavanol load, multiples more sugar, and milk protein that may bind some of the polyphenols and lower their absorption (the milk-binding part is contested, but the flavanol gap is not) Cooper 2008. "Chocolate is heart-healthy" with no asterisks is a marketing line.

Where this goes wrong in practice

Three patterns turn a small win into a small loss. The first is buying alkalized cocoa powder — the dark, smooth, "Dutch-process" stuff used in most baking — for your daily hot chocolate. Alkalization can strip up to 90% of the flavanols while leaving the calories and the dark colour intact Miller et al. 2008. Look for "natural" or "non-alkalized" on the tub. The second is treating the bar as a snack rather than a dose: a square or two daily is the intervention; the whole bar in front of a screen is a different food entirely, and the calorie cost overwhelms the cardiovascular signal. The third is swapping dark for filled or milk chocolate at the supermarket because it's cheaper and assuming the headline benefit transfers. It doesn't.

Cost: a daily 25-gram square of decent 70%+ chocolate runs roughly $0.50 to $1.50, $50 to $200 a year on the high end. Bulk natural cocoa powder is cheaper per dose than any bar. The premium "high-flavanol" lines (Cocoapro, CocoaVia) that match the trial doses cost more but remove the flavanol-content guesswork.

When not to

At the modest doses this entry recommends, there is no hard "don't." A few situations where the substance is worth knowing about:

What you actually feel, and when

The first thing you notice is hedonic, not pharmacological — the after-dinner square becomes a ritual that arrives on time, every night, because you actually want it. That's the part that solves the adherence problem the rest of lifestyle medicine has.

Within two to four weeks, the blood-pressure number drops a couple of points. You will not feel a 3-mmHg fall directly, but the catchment math is real: across a population, the same drop maps to roughly a 7 to 10% reduction in stroke incidence Ried et al. 2017. If you check your pressure at home, you'll see it. The endothelial effect — vessels that open a little wider in response to demand — is the felt version of the same thing: hands that warm up faster, a flight of stairs that registers a beat less.

On the day, there's a small calmer-and-more-content lift from the cocoa itself, not just from the taste — methylxanthines and flavanols both contribute. Trials measured it as a real change on validated mood scales, not as a placebo curl-up effect Pase et al. 2013.

Over months, if you push the dose toward a heaped tablespoon of natural cocoa daily, a small but measurable nudge on memory for names and on tracking a thread through a long meeting. Larger if you're past 50. Not life-changing; clearly there, and the kind of thing that adds up next to other habits that bend the same axis Brickman et al. 2014, Mastroiacovo et al. 2015.

Over years, the cardiovascular-death signal — the part we are honest about not being certain of. The largest trial to date saw a 27% reduction in cardiovascular-disease death over three and a half years on 500 mg of flavanols a day, as a secondary outcome Sesso et al. 2022. Stacked with sleep, exercise, and not smoking, this is one of the cheaper, easier bricks in the wall.

Other entries that touch the same machinery worth a look: blood pressure at home for how to actually measure the effect on yourself; flavonoid-rich foods (tea, berries, apples) for stacking the same nitric-oxide lever; the DASH eating pattern for the much larger blood-pressure intervention this entry slots next to; and added sugar for the trade-off side of every chocolate decision.

Substance and claimed effects

The substance is the cacao bean (Theobroma cacao) and the foods that deliver it to consumers: high-cocoa dark chocolate (typically labelled 70% cocoa solids or higher), unsweetened cocoa powder, and raw or roasted cacao. The bioactive load is dominated by flavan-3-ols — monomeric (-)-epicatechin and catechin and their procyanidin oligomers — together with the methylxanthine alkaloids theobromine (~10 mg/g of cocoa solids) and a smaller amount of caffeine (~1 mg/g), plus minerals (magnesium, copper, iron, zinc, manganese) Cooper 2008. The entry covers modest daily consumption (roughly 20 to 30 grams of 70%+ dark chocolate, or the equivalent ~200 to 500 mg flavanols from cocoa powder) and the consequences that follow: blood-pressure reduction and improved endothelium-dependent vasodilation, modest gains in insulin sensitivity, small but replicable lifts in mood and in some cognitive measures (executive function, verbal learning), a possible long-term cardiovascular-risk signal, and the unavoidable trade-offs with added sugar, calories, and late-day methylxanthines. Beauty, sleep, and longevity dimensions are scored holistically against this profile even when the article does not centre them.

Evidence by addressing question

Mechanism

The central mechanism is nitric oxide (NO) bioavailability. The flavan-3-ol monomer (-)-epicatechin is absorbed in the small intestine, conjugated, and reaches concentrations in the low micromolar range within 1–2 hours of intake of a flavanol-rich cocoa drink Schroeter et al. 2006. In a series of human pharmacological studies, isolated (-)-epicatechin reproduced essentially the full vascular effect of flavanol-rich cocoa — flow-mediated dilation rose by a comparable magnitude, and an NO-synthase inhibitor (L-NMMA) abolished both effects — establishing epicatechin as the necessary and largely sufficient agent and NO as the proximal pathway Schroeter et al. 2006, Ottaviani et al. 2018. Downstream, NO relaxes vascular smooth muscle (lower peripheral resistance, lower blood pressure), suppresses platelet aggregation, reduces leukocyte-endothelial adhesion, and improves microvascular perfusion in skin and brain. Procyanidins (the oligomers) are poorly absorbed intact; their contribution appears largely indirect, via microbial degradation to phenolic acids in the colon Ottaviani et al. 2018, Andujar et al. 2012.

Theobromine contributes independently and meaningfully. It is a longer-acting, weaker methylxanthine than caffeine (adenosine-receptor antagonism, phosphodiesterase inhibition); a 70% bar of 25 g delivers roughly 170–200 mg theobromine plus ~20 mg caffeine. The mood / mild alertness signal seen in randomized trials is plausibly attributable to this fraction rather than to flavanols alone Pase et al. 2013. Magnesium content (~64 mg per 30 g serving of 70–85% chocolate) is non-trivial in a Western diet where roughly half of adults fall short of the RDA, though no single chocolate serving turns a deficiency around.

Evidence

Blood pressure. The most-replicated effect. The 2017 Cochrane review (35 RCTs, n=1804) reported a pooled systolic reduction of −1.8 mmHg and diastolic −1.8 mmHg with two-week-plus daily flavanol-rich cocoa, larger in hypertensive subgroups (~−4 mmHg systolic) Ried et al. 2017. The earlier Hooper meta-analysis (42 RCTs) showed flow-mediated dilation improvements of +1.34% FMD (acute) and ~+1.7% (chronic), and significant drops in insulin resistance (HOMA-IR ≈ −0.94), serum LDL (~−0.07 mmol/L), and triglycerides Hooper et al. 2012. Single-dose vascular effects are reproducible within 90–180 minutes of intake Heiss et al. 2003, Schroeter et al. 2006.

Hard endpoints. Observational cohorts have consistently associated moderate chocolate consumption (≈1–3 servings/week) with lower coronary-heart-disease, stroke, and diabetes incidence — a 2017 meta of 14 prospective studies reported relative-risk reductions in the 10–20% range for the highest-vs-lowest consumers Yuan et al. 2017; an earlier BMJ meta showed ~37% lower cardiovascular disease across seven studies Buitrago-Lopez et al. 2011. These are non-randomised and confounded by socioeconomic and dietary patterns. The COSMOS trial — the largest RCT to date, n=21,442 older US adults, 500 mg/day cocoa-flavanol supplement vs placebo for 3.6 years — missed its primary composite cardiovascular endpoint (HR 0.97, 95% CI 0.88–1.07) but reported a 27% reduction in cardiovascular-disease death (HR 0.73, 0.54–0.98) as a secondary outcome Sesso et al. 2022. Regulators have read the totality cautiously: the FDA issued a 2022 qualified health claim allowing high-flavanol cocoa products (≥4% non-fat cocoa solids flavanols) to state that the science suggests, but does not prove, a reduced risk of cardiovascular disease FDA 2022. The EFSA approved the more limited claim that 200 mg/day of cocoa flavanols maintains endothelium-dependent vasodilation EFSA 2012.

Insulin sensitivity / metabolic. Short trials of dark chocolate (100 g, ~88 mg flavanols, 15 days) showed clinically meaningful improvements in HOMA-IR and an ~12-mmHg drop in systolic pressure in healthy adults Grassi et al. 2005. The effect generalises modestly to overweight adults and is additive with exercise in some trials Davison et al. 2008. Meta-analytic insulin-sensitivity gains hold up Hooper et al. 2012.

Cognition. The CoCoA trial randomised 90 older adults with mild cognitive impairment to high-, intermediate-, or low-flavanol cocoa for 8 weeks; the high-flavanol arm (~990 mg/day) showed significant gains in trail-making and verbal fluency, paralleled by lower blood pressure and improved insulin resistance Mastroiacovo et al. 2015. Brickman's 2014 Nature Neuroscience study used a high-flavanol drink (900 mg/day, 3 months) in cognitively normal older adults; functional MRI showed restored blood-volume signal in the dentate gyrus together with improved object-recognition memory Brickman et al. 2014. The Maine-Syracuse cohort (n=968) found that more-frequent chocolate consumers performed better on multiple cognitive batteries cross-sectionally, with the strongest signal in working memory and visual-spatial tasks Crichton et al. 2016. Effect sizes in interventions are modest, not transformative.

Mood. Pase et al. randomised 72 healthy adults to high-polyphenol cocoa, low-polyphenol cocoa, or placebo for 30 days; the high-polyphenol arm improved on self-rated calmness and contentment without changing cognitive performance Pase et al. 2013. Acute orosensory and methylxanthine effects likely contribute independently of flavanols. The case is for a small lift, not a treatment for clinical depression.

Endothelial function. The Sansone trial (n=100 healthy mid-life adults) used 450 mg flavanols twice daily for a month and reported FMD +1 percentage point, lower systolic pressure (−4 mmHg), lower LDL, higher HDL, and a 22% drop in calculated 10-year Framingham risk Sansone et al. 2015. EFSA's vasodilation claim is built on this evidence base EFSA 2012.

Protocol

Effective dosing in trials clusters in two ranges: ~200 mg flavanols/day for endothelial maintenance (EFSA's threshold) EFSA 2012 and ~500–900 mg/day for the strongest BP and cognition signals (COSMOS, CoCoA, Brickman) Sesso et al. 2022, Mastroiacovo et al. 2015, Brickman et al. 2014. Translation to grocery-store chocolate is rough because flavanol content varies by an order of magnitude depending on cocoa origin, fermentation, roasting time, and (most destructively) alkalization. A typical 70%+ dark bar carries ~50–100 mg flavanols per 25-g serving — meaning 20–30 g/day delivers about a quarter to a half of what the strongest trials used Miller et al. 2008. Heavy alkalization (Dutch-process cocoa) cuts flavanol content by up to 90% Miller et al. 2008. Cocoa powder (unsweetened, natural, non-alkalized) is the densest source: a tablespoon (~6 g) of natural cocoa can deliver 100–200 mg flavanols. Timing: caffeine + theobromine load argues against late-evening intake for caffeine-sensitive sleepers; otherwise timing is unconstrained. Calorie cost is real: a 30 g 70% bar is ~170 kcal and ~8 g sugar.

Contraindications

No hard pharmacological contraindications at the modest daily dose this entry recommends. Edge cases to flag: (1) heavy daily intake during pregnancy raises caffeine+theobromine exposure above the ~200 mg/day caffeine ceiling clinical bodies recommend; modest intake (20–30 g/day) is well within bounds. (2) Migraine sufferers occasionally report chocolate as a trigger, though dietary-trigger studies are mixed. (3) GERD: cocoa can lower lower-oesophageal-sphincter tone; a known reflux pattern argues against eating it within a couple of hours of lying down. (4) Iron-overload disorders (hereditary haemochromatosis): cocoa is iron-rich (~12 mg/100 g) and modest absorption is plausible — not a hard contraindication, but worth knowing. (5) MAOI prescriptions: theobromine and tyramine in aged chocolate are theoretical concerns; most modern practice does not restrict moderate dark chocolate. (6) Cocoa contains oxalates and is best moderated in calcium-oxalate stone formers.

Misconceptions

The dominant misconception is the antioxidant framing. Polyphenols' health effect in vivo is not direct radical scavenging — circulating concentrations are too low to matter for systemic antioxidant capacity — but rather signalling: NO synthase activation, modulation of NADPH oxidase, redox-sensitive transcription factors Ottaviani et al. 2018, Andujar et al. 2012. Second misconception: "raw cacao is special." Most "raw" cacao is no more flavanol-rich than well-fermented, low-roast conventional cocoa; the flavanol-destroying step is alkalization, not heat Miller et al. 2008. Third: cocoa percentage on the label is not flavanol content. Two 85% bars from different brands can differ by 5× in actual flavanol load Miller et al. 2008. Fourth: milk chocolate carries the headline trial benefits. It does not — flavanol content is roughly a third of dark chocolate's, sugar load is multiples higher, and milk protein may bind polyphenols and reduce bioavailability (the latter is contested; FMD effects of dark vs dark+milk diverged in some studies, did not diverge in others) Cooper 2008. Fifth: "chocolate is heart-healthy" with no asterisks. The cardiovascular signal is conditional on flavanol-rich, low-sugar, modest-portion intake; an 80-g milk-chocolate bar nightly is a net liability.

Failure-modes

Three real ones. (1) Switching dark chocolate for milk or filled chocolate and assuming the benefit transfers. (2) Treating it as a snack ad libitum — a 30 g serving is the dose; 100 g eaten in front of a screen drowns the BP signal in calories and sugar. (3) Buying alkalized "Dutch" cocoa powder for a hot-chocolate ritual: the flavanol load is collapsed and the calorie cost remains Miller et al. 2008.

Practicalities

Cost is low. A daily 25–30 g serving of mid-range 70–85% dark chocolate is ~$0.50–1.50/day, $200/year on the high end; bulk natural cocoa powder is cheaper still. Brand variability in flavanol content matters more than price; some premium "single-origin" lines list higher procyanidin assays, but third-party verification is rare. Effort: trivial.

Stakes (continued exposure to absence)

Absence of a dietary flavanol source — for someone without compensating tea, berry, or apple intake — is not a defined clinical syndrome, but it means giving up an evidenced, cheap, low-effort lever on blood pressure and endothelial function. For a reader already in the upper-normal or stage-1 hypertension band, a few mmHg matter at population scale: per the Cochrane benchmark, ~−2 mmHg systolic across the catchment maps to roughly a 7–10% reduction in stroke incidence and 4–5% in CHD mortality Ried et al. 2017. Stakes are real but modest in the typical reader.

Payoff (continued exposure to presence)

At a 25-g serving of 70%+ chocolate daily, the realistically projectable changes are: 2–5 mmHg lower resting blood pressure (larger in the hypertensive) within 2–4 weeks; better fasting insulin sensitivity over weeks; modest mood lift on acute intake (orosensory + methylxanthine + flavanol); a measurable but small improvement in cognitive performance after months at the upper end of the dosing range, larger in older adults Mastroiacovo et al. 2015, Brickman et al. 2014; possibly a modest long-run reduction in cardiovascular mortality, with the COSMOS secondary signal as the strongest pointer and observational cohorts as supporting (but confounded) evidence Sesso et al. 2022, Yuan et al. 2017. Felt experience is the daily ritual: a small after-dinner habit that is also a pharmacological intervention with measured benefits.

The credibility range

The optimist case

Cocoa flavanols, particularly (-)-epicatechin, have a well-characterised molecular pathway (NO bioavailability via endothelial NO synthase activation) with isotopically-traced human pharmacology Schroeter et al. 2006. The biological effect is reproducible across 100+ randomised trials: lower blood pressure, better FMD, improved insulin sensitivity, lower LDL, lower HOMA-IR Hooper et al. 2012, Ried et al. 2017. Two regulatory bodies have approved health claims at modest doses EFSA 2012, FDA 2022. The largest RCT to date (COSMOS) showed a statistically significant 27% reduction in cardiovascular death as a secondary outcome — the kind of magnitude that, if real, makes 500 mg/day cocoa flavanols a serious population-level intervention Sesso et al. 2022. Observational cohorts consistently align: moderate chocolate consumers have lower coronary, stroke, and diabetes incidence Yuan et al. 2017. The intervention is cheap, palatable, low-effort — the dominant friction in lifestyle medicine is adherence, and dark chocolate solves that problem.

The skeptic case

COSMOS missed its primary composite endpoint — the headline "27% CV death reduction" is a secondary outcome in a trial that did not meet its main test, and the lower CI bound (0.98) sits at the edge of the null Sesso et al. 2022. The BP effect in healthy normotensive adults is small (−1 to −2 mmHg) and may be inflated by publication bias against null trials Ried et al. 2017. Cocoa industry funding of much of the trial literature is well-documented — Mars, Inc. underwrote COSMOS and the proprietary CocoaVia supplement; the cocoa-flavanol formulation tested is not what a chocolate bar delivers. Observational cohort signals are bedevilled by reverse causation (healthier people eat more dark chocolate) and confounding (income, education, diet quality) that statistical adjustment cannot fully remove. The cognition signals (CoCoA, Brickman) used flavanol doses 5–10× a normal dietary intake and have not replicated at scale. The very modesty of the effects argues that the right benchmark is "small, real, additive" rather than "a meaningful single-substance intervention." And the calorie / sugar cost of any chocolate vehicle is non-trivial: even at 70%+, daily intake is ~150–200 kcal that has to be subtracted from somewhere.

The author's call

The blood-pressure and endothelial-function effects of cocoa flavanols are real, mechanistically grounded, and replicable; the cardiovascular hard-endpoint signal is suggestive but not nailed down. A modest daily serving of high-cocoa dark chocolate (or unsweetened natural cocoa) is a defensible "do" — cheap, low-effort, supported by regulatory health claims, with a clean mechanism. It is not a top-tier intervention: it will not transform anyone's health on its own. The honest framing is "one of the rare daily pleasures that pays a small but measured biological dividend, on the condition you keep the portion modest and pick a low-alkalization, high-cocoa product." Evidence rating is 4 (strong mechanism, multiple meta-analyses, one large RCT, regulatory recognition; cardiovascular hard endpoint unresolved); controversy 2 (limited dispute on the mechanism and BP effect; live debate on hard-endpoint translation and industry-funding influence).

Stakeholder and incentive map

Commercial: Mars, Inc. (CocoaVia, COSMOS sponsor) and Barry Callebaut/Cargill (cocoa supply) have funded a large share of the human flavanol literature. The funding bias is real but does not invalidate the trials, which are largely well-designed; it does push researchers toward proprietary high-flavanol formulations rather than off-the-shelf chocolate.
Professional: Cardiology societies (AHA) treat cocoa flavanols as a low-priority adjunct to validated lifestyle interventions (DASH diet, sodium reduction, exercise); they have not endorsed daily chocolate as a population intervention.
Cultural/community: The "dark chocolate is medicine" trope is widespread in wellness media and gives readers permission to over-consume. Cocoa farmers and craft chocolate makers benefit from the halo.
Skeptic: Public-health nutrition voices (Marion Nestle, et al.) regularly flag the sugar/calorie reality and the industry-funded literature; their case is partly right and partly an overcorrection.

Population variability

Effects scale with baseline cardiovascular risk: BP reduction is several times larger in hypertensives than normotensives Ried et al. 2017. Cognitive gains are larger in older adults with mild cognitive impairment than in young healthy volunteers Mastroiacovo et al. 2015, Brickman et al. 2014. Insulin-sensitivity gains may be larger in metabolically compromised individuals. No clear gender or ethnic interaction in the existing literature. Caffeine sensitivity (CYP1A2 metaboliser status) modifies the late-day tolerability but not the cardiovascular signal. Migraine-prone and reflux-prone subgroups may notice the substance, in the wrong direction, before they notice the benefit.

Knowledge gaps

Whether the COSMOS cardiovascular-death secondary endpoint replicates in a trial powered to test it as the primary outcome.
How much of the chocolate-bar benefit (vs purified supplement) is preserved in real-world eating: matrix effects, fat and sugar interactions, dose variability are under-characterised.
Whether dose-response is monotonic or saturates around the EFSA ~200 mg/day threshold; the CoCoA / Brickman doses (~900 mg/day) may be over the elbow.
Long-term effects on glycaemic regulation in pre-diabetic populations at chocolate (rather than purified-supplement) intake.
Sustainable, transparent flavanol-content labelling at retail; consumers cannot reliably pick a high-flavanol bar from packaging today.

Scope vs brief. The topic brief named blood pressure, endothelial function, insulin sensitivity, mood, cognition, and the sugar/calorie trade-off. All six get a home — BP and endothelium in evidence and mechanism, insulin sensitivity inside evidence, mood and cognition there and in payoff, and the sugar/calorie reality in protocol, misconceptions, and failure-modes. Nothing in the brief was silently dropped.

Hard scoring calls. evidence at 4 rather than 5: COSMOS missed its primary composite (only the secondary CV-death endpoint hit) and the field still lacks a positively powered hard-endpoint RCT, so the top tier is not yet earned. longevity at 2 rather than 3: the COSMOS death finding is real but secondary; cohorts are confounded; honest call is "meaningful, not dominant." energy at 1 (theobromine is mild and slow-onset) versus 0 — went 1 because the felt effect is genuinely there, just modest. beauty_cumulative was initially scored 1 on a flavanol-NO microvascular-perfusion-of-skin basis, but the trial evidence on visible skin outcomes is thin and the article does not anchor a real skin / appearance consequence; dropped to 0 to keep score and body in sync. The Latin scientific name and the (-)-epicatechin biochemistry stay in the article because they earn the friend-test exemption (named molecule); each is glossed in plain English on first use.

Things excluded on purpose. The "chocolate as aphrodisiac" tradition — too thin to support, mostly orosensory. The "cocoa for skin photoprotection" trials (e.g. Heinrich-style UV-MED studies) — interesting but small, not a recommendation, and the reason beauty_cumulative stays at 0. Theobromine toxicity in dogs — not a reader concern; this is the Body Handbook. Discussion of cocoa farming ethics (child labour, deforestation) — real but a different category of consideration and would crowd the editorial frame.

Future-link candidates. Blood pressure at home, flavonoid-rich foods, DASH eating pattern, added sugar, caffeine and theobromine, magnesium. None exist yet at write time — listed in out-of-scope as forward pointers.

Separate-entry candidates. "Cocoa flavanol supplements (CocoaVia, Cocoapro)" could become its own entry if reader demand surfaces — the dossier here covers them, but they are different from a chocolate bar in dose, delivery, and price.

Industry-funding flag. The largest RCT and a substantial share of the flavanol literature is Mars-funded. The article notes this in the COSMOS callout; the credibility-range section in research weighs it more fully. Reader is told plainly.