Is dark chocolate bad for blood sugar, or good for insulin sensitivity?
Both claims circulate and they are about different things. Cocoa flavanol preparations have shown small effects on some metabolic markers in short trials; chocolate as sold is a sugar-containing food, and eating it raises blood glucose the way sugar-containing foods do.
Plausible mechanism, limited human evidence
There is a credible biological or chemical reason it might be true, and that is different from evidence that it is.
What the evidence shows
Two separate literatures get merged here and separating them resolves most of the confusion.
The first concerns COCOA FLAVANOLS. Short trials have reported small improvements in measures of insulin sensitivity and related markers, with mechanisms proposed through vascular function and through effects on glucose handling. The findings are modest, the trials are short, and results are inconsistent between studies.
The second concerns CHOCOLATE AS A FOOD. A bar contains sugar — a large proportion of it in milk chocolate, a smaller but still substantial proportion in most dark chocolate. Eating it produces a glycaemic response. The fat and the fibre in cocoa solids moderate the rate of that response relative to sugar alone, which is a real effect and is not a neutralising one.
Observational studies associating chocolate consumption with metabolic outcomes exist and are heavily confounded, in both directions: people who eat dark chocolate differ systematically from those who do not, and people who are unwell frequently change what they eat.
What it does not show
It does not show that dark chocolate is suitable for managing diabetes or insulin resistance, and nothing in this catalogue should be read as suggesting a dietary approach to either.
It does not show that a flavanol finding transfers to a bar. The trials use preparations; a bar delivers an uncertain flavanol dose alongside a certain sugar dose, and the two do not cancel.
And it does not show that reverse causation has been excluded from the observational work. It generally has not.
How good is the evidence
The flavanol trials are small, short and inconsistent, and the metabolic markers used differ between them. The observational literature is subject to confounding severe enough that its direction is not informative on its own. The glycaemic behaviour of a sugar-containing food is not in dispute and is ordinary nutrition rather than a finding about chocolate. Placed at `plausible_mechanism` because the favourable half rests on mechanism plus thin human evidence.
Who this may not apply to
- Anyone managing diabetes or insulin resistance should treat this as background reading only and follow clinical advice.
- Trials have generally recruited people without diabetes, so findings do not transfer to those managing it.
- Anyone managing a health condition, taking medication, or making decisions about their diet should speak to a clinician rather than treat this as guidance.
Related claims
“Dark chocolate is a health food — the darker the better.”
Research on cocoa flavanols and short-term measures of vascular function is real, and a narrowly scoped EU health claim exists on that basis. It concerns a specified daily intake of flavanol compounds, not chocolate. Flavanol content is not indicated by cocoa percentage, is substantially reduced by ordinary processing, and is not declared on mainstream products. Meanwhile the non-cacao mass of a dark bar is overwhelmingly sugar.
The kernel of truth: The underlying research exists and is not junk. A regulatory body assessed a claim in this area and authorised one, which very few food claims achieve. The overreach is in the distance between 'a specified flavanol intake affects a measure of vessel dilation' and 'chocolate is good for you'.
Where it comes from: Traceable in large part to industry-funded research programmes and the press coverage built on them. The Cochrane review of cocoa and blood pressure notes industry funding across a substantial share of the trials it includes.
Related
Health topics
- Sugar in chocolateRead alongside Chocolate, insulin sensitivity and metabolic markers. The most reliably calculable fact on this page, and the one that most often goes unmentioned when chocolate is discussed as a health food.
- Cocoa and blood pressureRead alongside Chocolate, insulin sensitivity and metabolic markers. Cocoa flavanol preparations have produced small average reductions in blood pressure in short trials, and the effect is real, modest and measured over weeks. Whether eating chocolate delivers it is a separate question the trials mostly do not answer.
- Cocoa products are not interchangeableRead alongside Chocolate, insulin sensitivity and metabolic markers. No, and the difference is the single largest source of error in chocolate health writing. Most favourable cocoa findings come from defatted, flavanol-standardised cocoa preparations, and are then reported as though they were about a bar of chocolate.
- Energy density and portionRead alongside Chocolate, insulin sensitivity and metabolic markers. Chocolate is energy-dense because it is mostly fat and sugar, and that is true of dark as well as milk. Dark chocolate is not a low-energy food, and the differences between types are smaller than the differences between portions.
Sources
- Guideline: Sugars intake for adults and children — World Health Organization, 2015(citation identity confirmed; passage not re-read)
- Effect of cocoa on blood pressure — Cochrane Database of Systematic Reviews, 2017(citation identity confirmed; passage not re-read)
- FoodData Central — United States Department of Agriculture, Agricultural Research Service(citation identity confirmed; passage not re-read)
Nothing on this page is health advice or dietary advice. See the editorial policy.