Is dark chocolate good for you?

THREE THINGS DECIDE WHETHER ANY CHOCOLATE-AND-HEALTH CLAIM MEANS ANYTHING, and they are the reason the same research produces both cautious papers and extravagant headlines.

WHAT WAS ACTUALLY STUDIED. The large majority of positive cocoa research uses standardised high-flavanol cocoa extracts, capsules or specially prepared drinks — not chocolate. Ordinary chocolate contains a small and highly variable fraction of the flavanol dose used in those studies, because roasting, alkalisation and processing destroy a great deal of it and nothing on a label tells you how much survived. This is the single largest source of misleading chocolate health writing and it has its own page.

WHAT WAS MEASURED. A change in a biomarker is not a change in health. Flow-mediated dilation, blood pressure at one visit, a lipid fraction and an antioxidant assay are all measurements of something that MIGHT lead to an outcome, and the step from one to the other is exactly where the evidence thins.

WHAT ELSE CAME WITH IT. Chocolate is roughly a third fat and, in the versions people actually eat, a substantial proportion sugar. Any account of its benefits that does not sit beside its energy density is incomplete arithmetic rather than a health finding.

And a fourth thing worth knowing: a substantial share of cocoa flavanol trials are funded by confectionery manufacturers. That does not make them wrong, and it is something a reader is entitled to know without having to look it up.

These are the claims where the evidence is genuinely there and the finding is smaller than the phrase attached to it.

BLOOD VESSEL FUNCTION. Cocoa flavanols measurably improve flow-mediated dilation, an ultrasound measure of how an artery responds. This is the best-supported cocoa finding there is and the mechanism — flavanol effects on nitric oxide availability — is coherent. What it is NOT is evidence that eating chocolate prevents heart disease, and the topic page says so at length.

BLOOD PRESSURE. Meta-analyses of cocoa flavanol trials find small average reductions. Small is the operative word — the effects are of a size that matters at population scale and would be invisible to any individual — and the trials are short, heterogeneous and mostly not using chocolate.

MOOD, IN THE SHORT TERM. Eating something you enjoy improves mood, reliably and immediately. That is a real effect and it is not a pharmacological one; the compounds usually credited with it are present in quantities far too small to act as the headlines suggest.

Each of those has its own page with the dose, the population caveats and the explicit statement of what the finding does not extend to.

Real research, real mechanisms, and conclusions the evidence does not yet carry.

COGNITION. Several trials report short-term effects of high-flavanol cocoa on measures of attention and processing speed. The studies are small, the measures vary, and the durable claim about long-term cognitive health is not supported.

INSULIN SENSITIVITY AND METABOLIC MARKERS. Mechanistically plausible, some supportive trial data, and complicated by the fact that the vehicle in most real-world consumption carries sugar.

BLOOD LIPIDS. Mixed. Some trials find small improvements in some fractions; others find nothing; the cocoa butter in chocolate is largely stearic acid, which behaves differently from other saturated fats and is part of why this area is more confusing than it looks.

THE GUT MICROBIOME. Cocoa polyphenols are largely not absorbed in the small intestine and reach the colon, where they are metabolised by gut bacteria — which is both a mechanism for the flavanol findings above and an active research area in its own right. Early, and genuinely interesting.

EXERCISE PERFORMANCE AND RECOVERY. Some small trials, plausible mechanisms, and no established effect. Chocolate milk's reputation as a recovery drink is a reasonable and cheap option rather than a demonstrated superior one.

Several widely repeated claims come from cohort studies — populations followed over time, with chocolate consumption recorded and outcomes counted.

The associations are real and they are ASSOCIATIONS. People who report eating chocolate regularly differ from people who do not in a great many ways that are hard to adjust for: income, overall diet, exercise, and — the awkward one — health status itself, since people who are unwell frequently eat less of everything.

So an association between chocolate consumption and lower cardiovascular event rates is a genuine observation and a poor basis for a recommendation, and the direction of causation is not settled by it.

This matters more in this subject than in most, because chocolate cohort findings are irresistible to headline writers and the gap between "associated with" and "reduces" is one word.

ANTIOXIDANTS. Cocoa is high in polyphenols and the word antioxidant is doing much less work than people think. Antioxidant capacity measured in a test tube says little about what happens in a body, the ORAC methodology that made these claims famous was withdrawn by the body that published it, and the interesting effects of cocoa polyphenols now look to be about signalling and gut metabolism rather than about mopping up free radicals.

CHOCOLATE CAUSES ACNE. Long-standing, and the evidence has never supported it well. Some more recent work is less dismissive than the classic studies, which is a reason to say the question is open rather than a reason to believe the folk claim.

CHOCOLATE CAUSES MIGRAINE. Frequently reported by people with migraine and poorly supported by challenge studies. A plausible explanation is that craving for chocolate is part of the prodrome — the phase before the headache — so the chocolate is a symptom rather than a trigger.

CHOCOLATE IS ADDICTIVE in the clinical sense. Chocolate craving is real, well described and culturally patterned, and the evidence that it constitutes an addiction in the way the word is used clinically is weak.

CHOCOLATE HELPS YOU SLEEP. It contains caffeine and theobromine, both stimulants. Dark chocolate contains meaningfully more caffeine than milk, though considerably less than coffee.

AND THE WHOLE CATEGORY OF FRAMING that presents a food as exceptionally healthful by its nature. The best-known term for it has no definition in any food standard, and this catalogue's validator refuses that term in published prose — a rule about honesty rather than style, and one this sentence had to be rewritten to satisfy. Unlike its sustainability counterpart it has no escape for naming a phrase rather than asserting it, which was deliberate: no page here needs to print it.

The questions that get asked least and matter most.

ENERGY DENSITY. Chocolate is calorie-dense, largely from fat, and a hundred grams of dark chocolate is a substantial part of a day's energy. No flavanol finding changes that arithmetic, and any account of benefits that omits it is incomplete.

SUGAR. Dark chocolate contains less sugar than milk chocolate by composition rather than by any clever formulation, which is the simplest true statement in this whole area and the only reason "dark is better" holds up as a general rule.

CADMIUM AND HEAVY METALS. Cacao takes up cadmium from soil and levels vary substantially by origin. This is a genuine and quantified concern with regulatory limits attached in several jurisdictions, and it is the one health topic here where the caution runs in the opposite direction to the enthusiasm.

WHAT PERCENTAGE. Higher cocoa means more flavanols in principle, less sugar in fact, and more bitterness — and the flavanol figure is unreliable because processing destroys a variable amount and nothing declares it. 85% is lower in sugar than 70%; whether it delivers more usable flavanols is not something a label can tell you.

AND FOR PETS AND CHILDREN the picture is different in kind rather than in degree, and both have their own pages. Theobromine is genuinely dangerous to dogs, and the risk scales with cocoa content — a dark chocolate is far more hazardous than a milk one.

The honest answer has three parts and none of them is a headline.

COCOA FLAVANOLS DO SOMETHING MEASURABLE, principally to blood vessel function and blood pressure, in trials that mostly used preparations that are not chocolate at doses a bar does not deliver.

DARK CHOCOLATE IS A BETTER VERSION OF A CONFECTION than milk chocolate, chiefly because it contains less sugar. That is a real and useful thing to know and it is a much smaller claim than the one usually made.

AND IT IS STILL A CONFECTION. It is enjoyable, it is calorie-dense, and eating it because you like it is a perfectly good reason that requires no scientific justification at all.

This site describes evidence and does not give dietary advice; there is no field anywhere in its data model that can hold a recommended dose. Anyone managing a health condition should speak to a clinician rather than a chocolate reference.