Food · Misleading · 被健康光环包装的甜食
Dark Chocolate & Cocoa
可可里的类黄酮 (flavanols) 确实对血管有一点点好处，但它总是和糖、脂肪、热量打包在一起卖。COSMOS 大试验 (Sesso 2022)：每天 500 毫克可可类黄酮补充剂没有降低总心血管事件，只在心血管死亡这个次要指标上小幅下降。Cochrane (Ried 2017)：血压平均降约 2 毫米汞柱。机制是类黄酮促进一氧化氮、让血管舒张。但大多数巧克力经过重加工和碱化 (Miller 2008)，类黄酮所剩无几。真相：一小块 70% 以上的黑巧或无糖可可粉是可以的享受，但它不是保健品。
Story path
- 1The claim · 'chocolate is a health food'The claim · 'chocolate is a health food'
- 2What flavanols are · mostly lost in processingWhat flavanols are · mostly lost in processing
- 3Evidence grade · limited in the big trialEvidence grade · limited in the big trial
- 4Mechanism · flavanols → nitric oxide → vasodilationMechanism · flavanols → nitric oxide → vasodilation
- 5Reality check · sugar, fat, calories, portionReality check · sugar, fat, calories, portion
- 6What to do · a treat, not a supplementWhat to do · a treat, not a supplement
Chapter 1
The claim · 'chocolate is a health food'
The claim · 'chocolate is a health food'
Let's untangle something easily muddled first: cocoa beans do contain a class of molecules that are mildly good for blood vessels, called flavanols; but the bar of chocolate in your hand is flavanols packaged together with a lot of sugar, fat, and calories. Pinning the modest benefit of the former onto the whole candy bar of the latter is the core trick behind the 'healthy chocolate' claim.
In recent years, headlines love to say: dark chocolate 'protects the heart', 'lowers blood pressure', 'improves mood', 'is antioxidant', even 'protects the brain'. So chocolate transformed from a sweet into a 'functional food' you can eat with a clear conscience — especially the ones labeled high-cacao, dark, or 'raw cacao', which sell for more.
This narrative bundles several assertions:
Dark chocolate is good for the heart and lowers blood pressureIt is rich in antioxidants and fights agingIt is a 'healthy sweet' you can eat freely
This chapter separates the genuinely somewhat useful component in cocoa from the candy bar it is packed into: how strong the flavanol evidence really is, how large the effect is, and why most retail chocolate has already processed that benefit away.
In recent years, headlines love to say: dark chocolate 'protects the heart', 'lowers blood pressure', 'improves mood', 'is antioxidant', even 'protects the brain'. So chocolate transformed from a sweet into a 'functional food' you can eat with a clear conscience — especially the ones labeled high-cacao, dark, or 'raw cacao', which sell for more.
This narrative bundles several assertions:
Dark chocolate is good for the heart and lowers blood pressureIt is rich in antioxidants and fights agingIt is a 'healthy sweet' you can eat freely
This chapter separates the genuinely somewhat useful component in cocoa from the candy bar it is packed into: how strong the flavanol evidence really is, how large the effect is, and why most retail chocolate has already processed that benefit away.
Chapter 2
What flavanols are · mostly lost in processing
What flavanols are · mostly lost in processing
To discuss 'chocolate is healthy', first know which molecule the benefit comes from, and how fragile it is.
The vessel-friendly components in cocoa beans are a class of flavan-3-ols, of which the most studied is epicatechin. All the evidence for 'cocoa protects blood vessels' is about these, not about cocoa butter or sugar.
The key problem: these molecules are heavily lost across the long journey from cocoa bean to chocolate. Fermentation, roasting, and especially alkalization (Dutch processing) — which darkens the color and softens the taste — all degrade flavanols. Miller 2008 directly measured retail cocoa powders: natural cocoa powder has the highest flavanols, while alkalized cocoa powder has sharply lower flavanols and antioxidant capacity, with heavy alkalization losing more than half.
This leads to a counterintuitive reality:
Dark, smooth, non-astringent chocolate is often exactly the kind whose flavanols were processed away the mostConversely, flavanol-rich cocoa is bitter and astringent — precisely the taste the processing works to removeSo the labels dark and high-cacao do not guarantee the flavanols are still there
In other words: cocoa's small genuine merit is real, but easily processed out before you even buy it. The next scene looks at how large an effect this remaining flavanol actually produces in clinical trials.
The vessel-friendly components in cocoa beans are a class of flavan-3-ols, of which the most studied is epicatechin. All the evidence for 'cocoa protects blood vessels' is about these, not about cocoa butter or sugar.
The key problem: these molecules are heavily lost across the long journey from cocoa bean to chocolate. Fermentation, roasting, and especially alkalization (Dutch processing) — which darkens the color and softens the taste — all degrade flavanols. Miller 2008 directly measured retail cocoa powders: natural cocoa powder has the highest flavanols, while alkalized cocoa powder has sharply lower flavanols and antioxidant capacity, with heavy alkalization losing more than half.
This leads to a counterintuitive reality:
Dark, smooth, non-astringent chocolate is often exactly the kind whose flavanols were processed away the mostConversely, flavanol-rich cocoa is bitter and astringent — precisely the taste the processing works to removeSo the labels dark and high-cacao do not guarantee the flavanols are still there
In other words: cocoa's small genuine merit is real, but easily processed out before you even buy it. The next scene looks at how large an effect this remaining flavanol actually produces in clinical trials.
Chapter 3
Evidence grade · limited in the big trial
Evidence grade · limited in the big trial
Cocoa flavanol evidence is better than most 'superfoods', but reading it correctly means watching two things: the direction of the effect, and its size.
The COSMOS trial (Sesso 2022, evidence A): the largest to date, with over 20,000 older adults randomized to either a 500 mg cocoa-flavanol supplement daily or placebo, followed for about 3.6 years. The honest result: the primary endpoint of total cardiovascular events did not fall significantly; only in the secondary measure of cardiovascular death was the cocoa group about 27% lower (hazard ratio 0.73). A secondary-endpoint signal must be read cautiously and cannot be treated as a nailed-down 'protects the heart' conclusion.
Blood pressure (Ried 2017 Cochrane, evidence A): pooling dozens of randomized trials, flavanol-rich cocoa lowers blood pressure slightly, with systolic pressure dropping about 1.8 mmHg on average. Real, but small — far less than weight loss, salt restriction, exercise, or blood-pressure drugs.
Endothelial function (Flaviola, Sansone 2015, evidence B): in healthy people, a higher daily dose of cocoa flavanols improved the vessel's flow-mediated dilation (FMD). This is mechanistic support, but it used purified high-dose flavanols, not just any bar of chocolate.
Evidence summary: lowering blood pressure is a real A-grade but very small effect; lowering cardiovascular death is only a secondary-endpoint signal; and all of this used high-dose flavanols, which is not the same as eating more chocolate.
The COSMOS trial (Sesso 2022, evidence A): the largest to date, with over 20,000 older adults randomized to either a 500 mg cocoa-flavanol supplement daily or placebo, followed for about 3.6 years. The honest result: the primary endpoint of total cardiovascular events did not fall significantly; only in the secondary measure of cardiovascular death was the cocoa group about 27% lower (hazard ratio 0.73). A secondary-endpoint signal must be read cautiously and cannot be treated as a nailed-down 'protects the heart' conclusion.
Blood pressure (Ried 2017 Cochrane, evidence A): pooling dozens of randomized trials, flavanol-rich cocoa lowers blood pressure slightly, with systolic pressure dropping about 1.8 mmHg on average. Real, but small — far less than weight loss, salt restriction, exercise, or blood-pressure drugs.
Endothelial function (Flaviola, Sansone 2015, evidence B): in healthy people, a higher daily dose of cocoa flavanols improved the vessel's flow-mediated dilation (FMD). This is mechanistic support, but it used purified high-dose flavanols, not just any bar of chocolate.
Evidence summary: lowering blood pressure is a real A-grade but very small effect; lowering cardiovascular death is only a secondary-endpoint signal; and all of this used high-dose flavanols, which is not the same as eating more chocolate.
Chapter 4
Mechanism · flavanols → nitric oxide → vasodilation
Mechanism · flavanols → nitric oxide → vasodilation
The small genuine benefit of cocoa flavanols has a fairly clear mechanism, landing on your blood-vessel lining.
The inner layer of a vessel has endothelial cells that produce a signaling molecule called nitric oxide (nitric oxide: A small signal molecule from the vessel lining that relaxes the vessel-wall muscle so the vessel widens.). NO relaxes the smooth muscle in the vessel wall, so the vessel dilates and widens, blood pressure drops a little, and blood flows more smoothly. The epicatechin in cocoa raises the available NO in the endothelium, producing that measurable but small vasodilation and blood-pressure effect (this is exactly how Ried 2017 explains it).
Spelling out the mechanism also exposes several marketing overreaches:
A valid mechanism does not mean a large effect: the NO pathway is real, but the flavanols in a bar of chocolate give a millimeters-of-mercury blood-pressure drop, not a 'replace your BP drug' drop'Antioxidant equals anti-aging' is an over-extrapolation: flavanols have antioxidant activity in a test tube, but 'eating them slows aging' is a different claim lacking reliable human hard-endpoint evidenceThe trials used purified high-dose flavanols with zero sugar and fat; the chocolate you eat is low in flavanols and high in sugar and fat — the two cannot be equated
So the mechanistic conclusion is: cocoa flavanols travel a real but gentle vascular pathway — they can make vessels a bit more comfortable, but this is a nice bonus, not a rescue.
The inner layer of a vessel has endothelial cells that produce a signaling molecule called nitric oxide (nitric oxide: A small signal molecule from the vessel lining that relaxes the vessel-wall muscle so the vessel widens.). NO relaxes the smooth muscle in the vessel wall, so the vessel dilates and widens, blood pressure drops a little, and blood flows more smoothly. The epicatechin in cocoa raises the available NO in the endothelium, producing that measurable but small vasodilation and blood-pressure effect (this is exactly how Ried 2017 explains it).
Spelling out the mechanism also exposes several marketing overreaches:
A valid mechanism does not mean a large effect: the NO pathway is real, but the flavanols in a bar of chocolate give a millimeters-of-mercury blood-pressure drop, not a 'replace your BP drug' drop'Antioxidant equals anti-aging' is an over-extrapolation: flavanols have antioxidant activity in a test tube, but 'eating them slows aging' is a different claim lacking reliable human hard-endpoint evidenceThe trials used purified high-dose flavanols with zero sugar and fat; the chocolate you eat is low in flavanols and high in sugar and fat — the two cannot be equated
So the mechanistic conclusion is: cocoa flavanols travel a real but gentle vascular pathway — they can make vessels a bit more comfortable, but this is a nice bonus, not a rescue.
Chapter 5
Reality check · sugar, fat, calories, portion
Reality check · sugar, fat, calories, portion
Eating chocolate as a 'health food', the biggest problem is not the flavanols but the things packaged in with them.
A typical milk chocolate bar is roughly half sugar and cocoa butter with almost no flavanols left, and commonly 500 to 600 kcal per 100 g. Even dark chocolate is not low in sugar or calories — just more bitter with slightly less sugar. Using 'it has flavanols' to rationalize 'eat more chocolate' means swallowing a lot of sugar and calories for a tiny vascular benefit — usually a losing trade.
A few contrasts to hold in mind at once:
High cacao does not mean high flavanols: as noted, alkalization and heavy processing strip flavanols; the label percentage refers to cocoa-solids share, not active flavanolsTrial dose is not snack dose: COSMOS used a 500 mg purified-flavanol capsule daily. Eating that much via chocolate would bring a large load of sugar and fat alongThe cost of the health halo: once people believe it is 'healthy', they tend to eat more with less guilt, so net sugar and calorie intake actually rises
The honest positioning: dark chocolate is a slightly better sweet than many sweets — usually lower in sugar, with a lasting satisfaction from its bitterness — and is perfectly fine as a treat. But 'a slightly better sweet' and 'a food you should eat for health' are two entirely different categories.
A typical milk chocolate bar is roughly half sugar and cocoa butter with almost no flavanols left, and commonly 500 to 600 kcal per 100 g. Even dark chocolate is not low in sugar or calories — just more bitter with slightly less sugar. Using 'it has flavanols' to rationalize 'eat more chocolate' means swallowing a lot of sugar and calories for a tiny vascular benefit — usually a losing trade.
A few contrasts to hold in mind at once:
High cacao does not mean high flavanols: as noted, alkalization and heavy processing strip flavanols; the label percentage refers to cocoa-solids share, not active flavanolsTrial dose is not snack dose: COSMOS used a 500 mg purified-flavanol capsule daily. Eating that much via chocolate would bring a large load of sugar and fat alongThe cost of the health halo: once people believe it is 'healthy', they tend to eat more with less guilt, so net sugar and calorie intake actually rises
The honest positioning: dark chocolate is a slightly better sweet than many sweets — usually lower in sugar, with a lasting satisfaction from its bitterness — and is perfectly fine as a treat. But 'a slightly better sweet' and 'a food you should eat for health' are two entirely different categories.
Chapter 6
What to do · a treat, not a supplement
What to do · a treat, not a supplement
Pulling this chapter into take-away judgments:
Not supported: chocolate is a health food, can be eaten freely, fights aging — all overstated. Cocoa flavanol benefits are real but small, and mostly lost in processing.
Partly real but amplified: high-dose cocoa flavanols do slightly lower blood pressure and improve endothelial function (Ried 2017 / Sansone 2015); but that comes from purified flavanols, not from the candy bar.
The real positioning: a treat slightly better than most sweets, not a supplement.
Actionable low-risk steps:
Eat it if you want, as a treat: a small square of 70% to 85% dark chocolate, eaten slowly, is satisfying and relatively lower in sugarIf you are genuinely after the flavanols, unsweetened natural cocoa powder is the more honest source — mixed into water, oatmeal, or yogurt, low in sugar and caloriesPrefer non-alkalized (natural) cocoa powder; alkalization processes the flavanols awayDo not use 'it's healthy' to give yourself a bigger portion; portion size and total sugar and calories are what actually move weight and metabolismTo manage blood pressure, rely first on salt restriction, weight loss, exercise, and your doctor's plan, and put chocolate back in the 'occasional pleasure' box
Atlas connections: for how sugar and calories affect weight, see the related sugar-and-metabolism chapters; for how another health-halo food gets marketed, see the superfoods debunk.
Not supported: chocolate is a health food, can be eaten freely, fights aging — all overstated. Cocoa flavanol benefits are real but small, and mostly lost in processing.
Partly real but amplified: high-dose cocoa flavanols do slightly lower blood pressure and improve endothelial function (Ried 2017 / Sansone 2015); but that comes from purified flavanols, not from the candy bar.
The real positioning: a treat slightly better than most sweets, not a supplement.
Actionable low-risk steps:
Eat it if you want, as a treat: a small square of 70% to 85% dark chocolate, eaten slowly, is satisfying and relatively lower in sugarIf you are genuinely after the flavanols, unsweetened natural cocoa powder is the more honest source — mixed into water, oatmeal, or yogurt, low in sugar and caloriesPrefer non-alkalized (natural) cocoa powder; alkalization processes the flavanols awayDo not use 'it's healthy' to give yourself a bigger portion; portion size and total sugar and calories are what actually move weight and metabolismTo manage blood pressure, rely first on salt restriction, weight loss, exercise, and your doctor's plan, and put chocolate back in the 'occasional pleasure' box
Atlas connections: for how sugar and calories affect weight, see the related sugar-and-metabolism chapters; for how another health-halo food gets marketed, see the superfoods debunk.