Food · Beverages · 饮品
Coffee
烘焙过的种子 · 咖啡因=腺苷受体拮抗剂 · 现代证据: 适量 (3-4 杯) 关联中性偏低的全因死亡与糖尿病风险 · 拆穿咖啡有害/脱水/影响长高· 不过滤 (法压/煮) 的咖啡醇升 LDL, 滤纸能挡 · 半衰期与午后断点 · CYP1A2 个体差异
Story path
- 1What is coffee · a roasted seedWhat is coffee · a roasted seed
- 2Caffeine · an adenosine-receptor antagonistCaffeine · an adenosine-receptor antagonist
- 3Modern evidence · moderate coffee is neutral-to-favorableModern evidence · moderate coffee is neutral-to-favorable
- 4Debunked · 'coffee is bad / dehydrates / stunts growth'Debunked · 'coffee is bad / dehydrates / stunts growth'
- 5Diterpenes & LDL · the filter's hidden jobDiterpenes & LDL · the filter's hidden job
- 6Individual variation · why some can't sleep after one cupIndividual variation · why some can't sleep after one cup
- 7How to drink · how much · who should careHow to drink · how much · who should care
Chapter 1
What is coffee · a roasted seed
What is coffee · a roasted seed
That cup of black coffee actually starts as a seed. The coffee tree (mainly Coffea arabica and Coffea canephora / robusta) bears red coffee cherries, and the seed inside the pulp is the 'coffee bean.' Green beans have almost none of the aroma we know — that scent and dark color are created during roasting, a high-heat step where sugars and amino acids undergo Maillard reactions and caramelization.
So what you drink is 'an infusion extracted with hot water from ground, roasted seeds.' A few commonly confused points:
Arabica vs robusta: arabica is milder and more acidic; robusta is higher in caffeine and harsher, common in instant and blendsRoast level: affects flavor, but has little to do with caffeine content (dark roast is not necessarily higher in caffeine)Brewing method: drip, pour-over, French press, espresso, boiled — this affects not only taste but also a direct health marker (a later scene covers coffee diterpenes and LDL)
Coffee is one of the most-studied dietary components in the world and one of the most misunderstood. This island's job is to pull both the old 'coffee is harmful' belief and the new 'coffee is a superfood' hype back to the middle — explaining exactly how caffeine works, what the modern evidence really says, and which cautions are genuine.
So what you drink is 'an infusion extracted with hot water from ground, roasted seeds.' A few commonly confused points:
Arabica vs robusta: arabica is milder and more acidic; robusta is higher in caffeine and harsher, common in instant and blendsRoast level: affects flavor, but has little to do with caffeine content (dark roast is not necessarily higher in caffeine)Brewing method: drip, pour-over, French press, espresso, boiled — this affects not only taste but also a direct health marker (a later scene covers coffee diterpenes and LDL)
Coffee is one of the most-studied dietary components in the world and one of the most misunderstood. This island's job is to pull both the old 'coffee is harmful' belief and the new 'coffee is a superfood' hype back to the middle — explaining exactly how caffeine works, what the modern evidence really says, and which cautions are genuine.
Chapter 2
Caffeine · an adenosine-receptor antagonist
Caffeine · an adenosine-receptor antagonist
Coffee's core active compound is caffeine. Its wake-up mechanism is elegant — not 'revving you up,' but 'blocking the signal that makes you sleepy.'
Adenosine: the brain's 'sleepiness timer'
Every hour you are awake, your brain accumulates a molecule called adenosine. It binds adenosine receptors on neurons and acts as a 'brake' — damping neural activity and making you progressively drowsy. This is one way the body tracks 'time to sleep.'
Caffeine: it takes the seat but doesn't push
Caffeine is structurally similar to adenosine, so it can sit in the adenosine receptor — but once seated it does not activate the receptor; it just occupies the spot, blocking real adenosine from binding. That is what 'adenosine-receptor antagonist' means. The sleepiness brake is held off, and you feel alert. (Full mechanism: dive to caffeine-l-theanine; how caffeine masks sleep debt during all-nighters: dive to all-nighter.)
The key: caffeine does not erase fatigue — it only delays your perception of it. Adenosine keeps accumulating; once caffeine is metabolized, the suppressed drowsiness returns all at once.
Half-life ≈ 4-5 hours (IOM 2001, Pharmacology of Caffeine): in healthy adults the elimination half-life of caffeine is usually 4 to 5 hours. So a noon coffee still has half its caffeine working by early evening — which directly sets the 'afternoon cutoff' (last scene).
Besides caffeine, coffee also carries chlorogenic acids and other polyphenols — the leading active candidates in most of coffee's antioxidant / metabolic research, and likely relevant to the diabetes association below.
Adenosine: the brain's 'sleepiness timer'
Every hour you are awake, your brain accumulates a molecule called adenosine. It binds adenosine receptors on neurons and acts as a 'brake' — damping neural activity and making you progressively drowsy. This is one way the body tracks 'time to sleep.'
Caffeine: it takes the seat but doesn't push
Caffeine is structurally similar to adenosine, so it can sit in the adenosine receptor — but once seated it does not activate the receptor; it just occupies the spot, blocking real adenosine from binding. That is what 'adenosine-receptor antagonist' means. The sleepiness brake is held off, and you feel alert. (Full mechanism: dive to caffeine-l-theanine; how caffeine masks sleep debt during all-nighters: dive to all-nighter.)
The key: caffeine does not erase fatigue — it only delays your perception of it. Adenosine keeps accumulating; once caffeine is metabolized, the suppressed drowsiness returns all at once.
Half-life ≈ 4-5 hours (IOM 2001, Pharmacology of Caffeine): in healthy adults the elimination half-life of caffeine is usually 4 to 5 hours. So a noon coffee still has half its caffeine working by early evening — which directly sets the 'afternoon cutoff' (last scene).
Besides caffeine, coffee also carries chlorogenic acids and other polyphenols — the leading active candidates in most of coffee's antioxidant / metabolic research, and likely relevant to the diabetes association below.
Chapter 3
Modern evidence · moderate coffee is neutral-to-favorable
Modern evidence · moderate coffee is neutral-to-favorable
Over recent decades coffee's health reputation underwent a major reversal. It was once suspected of causing cancer and harming the heart; but once studies separated 'coffee' from confounders like 'also smoking,' the modern large-scale evidence reached a fairly consistent conclusion.
All-cause mortality: neutral-to-lower (Poole 2017, *BMJ*): this umbrella review pooling many meta-analyses found that, versus no coffee, about 3-4 cups a day was associated with roughly 17% lower all-cause mortality (relative risk ~0.83) and lower cardiovascular mortality. Note this is a non-linear curve — the benefit is clearest around 3-4 cups, not 'more is better.'
Type 2 diabetes: dose-responsive lower risk (Ding 2014, *Diabetes Care*): a meta-analysis of 28 prospective cohorts and over 1.1 million people found coffee intake inversely and dose-responsively associated with type 2 diabetes risk — each extra cup per day lowered risk further (the association held for both caffeinated and decaffeinated coffee, suggesting it is not caffeine alone but possibly chlorogenic acids too). Full mechanism: dive to type-2-diabetes.
Safety, qualitatively (one-hand authority) (IARC Monograph Vol. 116 / Loomis 2016, *Lancet Oncology*): the International Agency for Research on Cancer re-classified coffee from its 1991 'Group 2B (possibly carcinogenic)' to Group 3 (not classifiable as to carcinogenicity) — i.e. 'no adequate evidence that coffee itself causes cancer.' The same evaluation flagged the real risk as 'very hot (>65°C) beverages, Group 2A (probably carcinogenic) to the esophagus' — a temperature issue, not a coffee issue. So 'don't drink it scalding' is real advice; 'coffee causes cancer' is not.
One honest point: nearly all of the above is observational, proving 'association,' not 'causation.' It is enough to say 'moderate coffee is safe for most healthy adults and may even carry some benefit,' but not enough to say 'drink coffee for your health.' Non-drinkers have no need to start over this association.
All-cause mortality: neutral-to-lower (Poole 2017, *BMJ*): this umbrella review pooling many meta-analyses found that, versus no coffee, about 3-4 cups a day was associated with roughly 17% lower all-cause mortality (relative risk ~0.83) and lower cardiovascular mortality. Note this is a non-linear curve — the benefit is clearest around 3-4 cups, not 'more is better.'
Type 2 diabetes: dose-responsive lower risk (Ding 2014, *Diabetes Care*): a meta-analysis of 28 prospective cohorts and over 1.1 million people found coffee intake inversely and dose-responsively associated with type 2 diabetes risk — each extra cup per day lowered risk further (the association held for both caffeinated and decaffeinated coffee, suggesting it is not caffeine alone but possibly chlorogenic acids too). Full mechanism: dive to type-2-diabetes.
Safety, qualitatively (one-hand authority) (IARC Monograph Vol. 116 / Loomis 2016, *Lancet Oncology*): the International Agency for Research on Cancer re-classified coffee from its 1991 'Group 2B (possibly carcinogenic)' to Group 3 (not classifiable as to carcinogenicity) — i.e. 'no adequate evidence that coffee itself causes cancer.' The same evaluation flagged the real risk as 'very hot (>65°C) beverages, Group 2A (probably carcinogenic) to the esophagus' — a temperature issue, not a coffee issue. So 'don't drink it scalding' is real advice; 'coffee causes cancer' is not.
One honest point: nearly all of the above is observational, proving 'association,' not 'causation.' It is enough to say 'moderate coffee is safe for most healthy adults and may even carry some benefit,' but not enough to say 'drink coffee for your health.' Non-drinkers have no need to start over this association.
Chapter 4
Debunked · 'coffee is bad / dehydrates / stunts growth'
Debunked · 'coffee is bad / dehydrates / stunts growth'
Coffee carries several decades-old myths. Let's take them one by one.
Myth 1: 'coffee dehydrates you'
The seed of truth: caffeine has a mild diuretic effect — large doses do slightly increase urination. But Killer 2014 (*PLoS ONE*), a crossover trial, tested it directly: habitual coffee drinkers consumed coffee delivering ~4 mg/kg caffeine a day (a few cups) versus an equal volume of water, comparing 24-hour urine output and multiple hydration markers — no difference. The conclusion: for habitual drinkers, moderate coffee hydrates comparably to water and counts toward daily fluid intake. So 'drink extra water to avoid coffee dehydration' is over-worry.
Myth 2: 'coffee stunts children's growth / dissolves bone'
This has no reliable human evidence. It may come from the observation that 'caffeine slightly increases urinary calcium loss,' but that calcium loss is tiny — a splash of milk offsets it — and it does not make people short or osteoporotic. (What actually affects bones and height is overall nutrition, protein, calcium, vitamin D, and sleep — not coffee.) That said, caffeine is unnecessary for children and more likely to cause sleep and mood problems, so 'kids don't need coffee' is reasonable — but the reason is sleep and palpitations, not the 'stunted growth' pseudoscience.
Myth 3: 'coffee harms the heart / causes cancer'
This is exactly the reversal from the previous scene: modern large-scale evidence links moderate coffee to neutral-to-lower all-cause / cardiovascular mortality, and IARC has removed coffee from the 'possibly carcinogenic' list (now 'not classifiable'). The old 'cancer' impression was largely a failure to disentangle confounders like smoking.
The half-true core: these myths are not invented from nothing — each has a 'partly-true fact' (diuresis, urinary calcium, palpitations in sensitive individuals) inflated into 'harmful.' Once the mechanism is clear, they return to their real scale.
The verdict: for healthy adults, moderate coffee does not dehydrate, does not cause cancer, does not harm the heart, and does not stunt children's growth. What truly matters is individual sensitivity, pregnancy, dose, and timing (next two scenes).
Myth 1: 'coffee dehydrates you'
The seed of truth: caffeine has a mild diuretic effect — large doses do slightly increase urination. But Killer 2014 (*PLoS ONE*), a crossover trial, tested it directly: habitual coffee drinkers consumed coffee delivering ~4 mg/kg caffeine a day (a few cups) versus an equal volume of water, comparing 24-hour urine output and multiple hydration markers — no difference. The conclusion: for habitual drinkers, moderate coffee hydrates comparably to water and counts toward daily fluid intake. So 'drink extra water to avoid coffee dehydration' is over-worry.
Myth 2: 'coffee stunts children's growth / dissolves bone'
This has no reliable human evidence. It may come from the observation that 'caffeine slightly increases urinary calcium loss,' but that calcium loss is tiny — a splash of milk offsets it — and it does not make people short or osteoporotic. (What actually affects bones and height is overall nutrition, protein, calcium, vitamin D, and sleep — not coffee.) That said, caffeine is unnecessary for children and more likely to cause sleep and mood problems, so 'kids don't need coffee' is reasonable — but the reason is sleep and palpitations, not the 'stunted growth' pseudoscience.
Myth 3: 'coffee harms the heart / causes cancer'
This is exactly the reversal from the previous scene: modern large-scale evidence links moderate coffee to neutral-to-lower all-cause / cardiovascular mortality, and IARC has removed coffee from the 'possibly carcinogenic' list (now 'not classifiable'). The old 'cancer' impression was largely a failure to disentangle confounders like smoking.
The half-true core: these myths are not invented from nothing — each has a 'partly-true fact' (diuresis, urinary calcium, palpitations in sensitive individuals) inflated into 'harmful.' Once the mechanism is clear, they return to their real scale.
The verdict: for healthy adults, moderate coffee does not dehydrate, does not cause cancer, does not harm the heart, and does not stunt children's growth. What truly matters is individual sensitivity, pregnancy, dose, and timing (next two scenes).
Chapter 5
Diterpenes & LDL · the filter's hidden job
Diterpenes & LDL · the filter's hidden job
This scene holds coffee's most-overlooked yet most practical fact: how you brew it directly changes its effect on blood lipids.
Two molecules: cafestol and kahweol
The oils in coffee beans contain two diterpenes — cafestol and kahweol. Cafestol is one of the most potent cholesterol-raising compounds known in the human diet; it interferes with the liver's regulation of cholesterol metabolism and raises blood LDL cholesterol.
The paper filter is the key switch
These molecules dissolve in coffee's oils, and a paper filter traps most of them. That explains the large difference between brewing methods:
Drip / pour-over (through paper): cafestol is caught by the filter, leaving very little in the cup, with almost no LDL effectFrench press, Turkish, Scandinavian boiled coffee: no paper filter, so cafestol goes straight into the cup at high levelsEspresso: in between, and because servings are small the total exposure is usually not high
How large is the effect (Cai 2012, *European Journal of Clinical Nutrition*): this meta-analysis of 12 randomized controlled trials and 1,017 subjects found coffee (mostly cafestol-containing unfiltered coffee) raised total cholesterol by ~8 mg/dL, LDL by ~5 mg/dL, and triglycerides by ~13 mg/dL on average — while filtered coffee produced essentially no such rise.
How to use this: if your LDL runs high or you watch your lipids and drink several cups a day, prefer paper-filtered coffee (drip, pour-over) over French press / boiled. A French press or two is no big deal for most people; but 'lots of unfiltered coffee daily + already-high LDL' is a combination worth adjusting (dive to dyslipidemia).
This is another payoff of 'understanding the why': same coffee, different filter, different meaning for your blood lipids.
Two molecules: cafestol and kahweol
The oils in coffee beans contain two diterpenes — cafestol and kahweol. Cafestol is one of the most potent cholesterol-raising compounds known in the human diet; it interferes with the liver's regulation of cholesterol metabolism and raises blood LDL cholesterol.
The paper filter is the key switch
These molecules dissolve in coffee's oils, and a paper filter traps most of them. That explains the large difference between brewing methods:
Drip / pour-over (through paper): cafestol is caught by the filter, leaving very little in the cup, with almost no LDL effectFrench press, Turkish, Scandinavian boiled coffee: no paper filter, so cafestol goes straight into the cup at high levelsEspresso: in between, and because servings are small the total exposure is usually not high
How large is the effect (Cai 2012, *European Journal of Clinical Nutrition*): this meta-analysis of 12 randomized controlled trials and 1,017 subjects found coffee (mostly cafestol-containing unfiltered coffee) raised total cholesterol by ~8 mg/dL, LDL by ~5 mg/dL, and triglycerides by ~13 mg/dL on average — while filtered coffee produced essentially no such rise.
How to use this: if your LDL runs high or you watch your lipids and drink several cups a day, prefer paper-filtered coffee (drip, pour-over) over French press / boiled. A French press or two is no big deal for most people; but 'lots of unfiltered coffee daily + already-high LDL' is a combination worth adjusting (dive to dyslipidemia).
This is another payoff of 'understanding the why': same coffee, different filter, different meaning for your blood lipids.
Chapter 6
Individual variation · why some can't sleep after one cup
Individual variation · why some can't sleep after one cup
'My friend drinks coffee before bed and sleeps fine; one afternoon cup keeps me up all night' — this is not an illusion. Behind it is a genetically set difference in metabolic speed.
CYP1A2: the main enzyme that breaks down caffeine
Once caffeine enters the body, over 95% is broken down by a liver enzyme, CYP1A2 (IOM 2001). The gene encoding it has a common variant that roughly splits people into:
Fast metabolizers: high CYP1A2 activity, caffeine cleared quickly, shorter half-life, little disruption, can still sleepSlow metabolizers: low enzyme activity, caffeine lingers far longer, much longer half-life — an afternoon cup can stretch into the small hours
This is why the earlier '~4-5 hour half-life' is an average — your personal value can be markedly faster or slower. External factors shift it too: smoking markedly speeds caffeine metabolism (it induces CYP1A2), while oral contraceptives and pregnancy slow it (clearance in pregnancy can more than double — one reason to limit caffeine then).
A real burden for sensitive people: slow-metabolizing / highly sensitive individuals may get palpitations, tremor, anxiety, and trouble falling asleep even from daytime coffee. This is not 'fussiness' — it is a genuine pharmacological response.
How to use this (self-observation): there is no need to genotype. Just watch your own response — if one afternoon cup wrecks your sleep or makes you jittery, you are likely a slower metabolizer / more sensitive, so cluster coffee in the morning and cut the total. People who feel fine need not worry about this.
The population-average 'moderate coffee is beneficial' conclusion always yields, at the individual level, to your own response.
CYP1A2: the main enzyme that breaks down caffeine
Once caffeine enters the body, over 95% is broken down by a liver enzyme, CYP1A2 (IOM 2001). The gene encoding it has a common variant that roughly splits people into:
Fast metabolizers: high CYP1A2 activity, caffeine cleared quickly, shorter half-life, little disruption, can still sleepSlow metabolizers: low enzyme activity, caffeine lingers far longer, much longer half-life — an afternoon cup can stretch into the small hours
This is why the earlier '~4-5 hour half-life' is an average — your personal value can be markedly faster or slower. External factors shift it too: smoking markedly speeds caffeine metabolism (it induces CYP1A2), while oral contraceptives and pregnancy slow it (clearance in pregnancy can more than double — one reason to limit caffeine then).
A real burden for sensitive people: slow-metabolizing / highly sensitive individuals may get palpitations, tremor, anxiety, and trouble falling asleep even from daytime coffee. This is not 'fussiness' — it is a genuine pharmacological response.
How to use this (self-observation): there is no need to genotype. Just watch your own response — if one afternoon cup wrecks your sleep or makes you jittery, you are likely a slower metabolizer / more sensitive, so cluster coffee in the morning and cut the total. People who feel fine need not worry about this.
The population-average 'moderate coffee is beneficial' conclusion always yields, at the individual level, to your own response.
Chapter 7
How to drink · how much · who should care
How to drink · how much · who should care
How to drink / how much
For most healthy adults, about 3-4 cups a day sits in the 'safe and neutral-to-favorable' range (Poole 2017). That is a reference ceiling, not a 'target' — non-drinkers need not start over an associationWatching lipids / high LDL: prefer paper-filtered coffee (drip / pour-over) over French press / boiled (previous scene, diterpenes)Don't turn coffee into a 'liquid dessert' with heavy syrup and cream — then the risk comes from the added sugar, not the coffee (dive to sugary-drinks)Don't drink it scalding: IARC flags very hot (>65°C) beverages as linked to esophageal risk; let it cool a little
Caffeine and sleep (important): caffeine's half-life is ~4-5 hours, longer in slow metabolizers. Afternoon caffeine may still be 'occupying adenosine receptors' when you want to sleep, delaying onset and weakening deep sleep. A practical cutoff: no caffeine after early afternoon (about 8-10 hours before bed), earlier still for slow metabolizers / insomniacs (dive to melatonin / all-nighter).
Who should pay attention
Pregnancy / breastfeeding: caffeine clearance slows in pregnancy. ACOG advises under 200 mg caffeine per day (about 1-2 cups of coffee), counting the total of coffee + tea + chocolate + energy drinksAnxiety / panic / palpitations / known arrhythmia: caffeine can worsen symptoms; such people should cut back or avoid it, following physician adviceInsomnia / poor sleep: hold the afternoon cutoff strictly, and switch to decaf if needed (it keeps chlorogenic acids; the diabetes association in Ding 2014 holds for decaf too)Certain medications (some antibiotics, ephedrine-type stimulants, lithium, etc.): may interact with caffeine; if unsure while on medication, ask your physician / pharmacistChildren and adolescents: unnecessary, and more likely to cause sleep and mood problems
This scene provides general information only and does not replace a physician's judgment of your medications and condition.
For most healthy adults, about 3-4 cups a day sits in the 'safe and neutral-to-favorable' range (Poole 2017). That is a reference ceiling, not a 'target' — non-drinkers need not start over an associationWatching lipids / high LDL: prefer paper-filtered coffee (drip / pour-over) over French press / boiled (previous scene, diterpenes)Don't turn coffee into a 'liquid dessert' with heavy syrup and cream — then the risk comes from the added sugar, not the coffee (dive to sugary-drinks)Don't drink it scalding: IARC flags very hot (>65°C) beverages as linked to esophageal risk; let it cool a little
Caffeine and sleep (important): caffeine's half-life is ~4-5 hours, longer in slow metabolizers. Afternoon caffeine may still be 'occupying adenosine receptors' when you want to sleep, delaying onset and weakening deep sleep. A practical cutoff: no caffeine after early afternoon (about 8-10 hours before bed), earlier still for slow metabolizers / insomniacs (dive to melatonin / all-nighter).
Who should pay attention
Pregnancy / breastfeeding: caffeine clearance slows in pregnancy. ACOG advises under 200 mg caffeine per day (about 1-2 cups of coffee), counting the total of coffee + tea + chocolate + energy drinksAnxiety / panic / palpitations / known arrhythmia: caffeine can worsen symptoms; such people should cut back or avoid it, following physician adviceInsomnia / poor sleep: hold the afternoon cutoff strictly, and switch to decaf if needed (it keeps chlorogenic acids; the diabetes association in Ding 2014 holds for decaf too)Certain medications (some antibiotics, ephedrine-type stimulants, lithium, etc.): may interact with caffeine; if unsure while on medication, ask your physician / pharmacistChildren and adolescents: unnecessary, and more likely to cause sleep and mood problems
This scene provides general information only and does not replace a physician's judgment of your medications and condition.