Coffee and Tea

TL;DR. Coffee is the most consumed psychoactive substance on Earth and, after decades of suspicion, the epidemiology now consistently associates moderate intake with lower all-cause mortality. Three to four cups a day sits at the bottom of the curve. Tea carries L-theanine alongside caffeine, which changes the subjective experience, and green tea's catechins have better evidence than most plant compounds while green tea extract has caused liver injury. The single most actionable fact in this chapter is that tea and coffee with a meal can cut non-haem iron absorption by more than half.

1. What they are

Coffee is the roasted seed ("bean") of the fruit of Coffea shrubs. Two species matter:

Tea is the leaf of a single species, Camellia sinensis. Every "true" tea is the same plant; the differences are entirely processing:

Herbal "teas" are not tea: rooibos, chamomile, peppermint, hibiscus, and the rest are infusions of other plants and contain no caffeine.

Oxidation, not fermentation. The browning of tea leaf is enzymatic oxidation by polyphenol oxidase, the same reaction that browns a cut apple. Only pu-erh involves genuine microbial fermentation.

2. Where they come from

Coffee: Ethiopia, where C. arabica grows wild in the highland forests. The origin story of Kaldi the goatherd is folklore. It spread through Yemen (where Sufi orders used it for night-long devotions and where the port of Mocha gave its name), then to the Ottoman world, then to Europe in the seventeenth century, where coffee houses became centres of commerce and political argument. Lloyd's of London began as a coffee house.

Tea: southwestern China, cultivated for over 2,000 years. Its trade shaped the Opium Wars, the Boston Tea Party, and the British colonisation of Assam and Ceylon, undertaken specifically to break China's monopoly.

Production: coffee around 10 million tonnes of green beans (Brazil, Vietnam, Colombia, Indonesia, Ethiopia); tea around 6.5 million tonnes (China, India, Kenya, Sri Lanka, Turkey). Both are grown overwhelmingly by smallholders and both are highly exposed to climate change: projections suggest a substantial reduction in land suitable for arabica by mid-century.

3. How they are grown and processed

Coffee is grown between the tropics, arabica at altitude. The fruit ("cherry") is picked, and processing decides much of the flavour:

• Washed/wet: pulp removed, fermented to strip mucilage, washed, dried. Cleaner, more acidic.
• Natural/dry: whole cherry dried in the sun. Fruitier, heavier, more variable.
• Honey/pulped natural: intermediate.

Then hulling, grading, and roasting, which is where the Maillard reactions build almost all of the 800-plus aroma compounds. Roast level trades acidity and origin character for body and bitterness. Darker roast contains slightly less caffeine by weight of bean (caffeine degrades a little and beans lose mass), which is the opposite of what most people assume.

Decaffeination removes 97 percent or more of the caffeine. Methods: Swiss Water (water and carbon filtration, no solvent), CO₂ (supercritical carbon dioxide), and solvent (ethyl acetate, marketed as "natural" or "sugarcane" process, or methylene chloride). Methylene chloride residues in finished coffee are far below regulatory limits and the method remains contentious.

Tea is grown at altitude in the tropics and subtropics, picked as "two leaves and a bud" for quality grades. Processing is withering, rolling or cutting (the industrial CTC method, crush-tear-curl, produces the fast-brewing dust in tea bags), oxidation for the time and temperature the style requires, and firing to stop it.

4. What is in them

Per 240 mL (8 fl oz) cup:

Caffeine itself is covered in Chapter 88. The rest:

Coffee contains chlorogenic acids (a major dietary polyphenol source, in fact the largest single source in many Western diets), trigonelline, diterpenes (cafestol and kahweol), and melanoidins formed during roasting.

Cafestol and kahweol are the reason brewing method matters for cholesterol. These diterpenes raise LDL, and they are removed by paper filters. So:

That is a real and under-appreciated finding: several cups a day of unfiltered coffee measurably raises cholesterol, and the same coffee through a paper filter does not. A large Norwegian cohort study found the mortality pattern tracked filtration.

Tea contains catechins (green tea is highest, dominated by EGCG), theaflavins and thearubigins (formed by oxidation in black tea, and what makes it dark), L-theanine, and fluoride (tea plants accumulate it; heavy tea drinking is a meaningful fluoride source).

L-theanine is an amino acid found almost nowhere else in the diet. It crosses the blood-brain barrier, increases alpha wave activity, and, in combination with caffeine, has reasonably consistent evidence for improving attention while reducing the jitteriness caffeine alone produces. It is the best explanation for why tea feels different from coffee at similar caffeine doses.

5. What the evidence shows

Coffee: the reversal. Coffee was suspected for decades, largely because early studies failed to adjust adequately for smoking, which correlated strongly with coffee drinking. Better-adjusted studies changed the picture.

Established or strong:

• All-cause mortality: meta-analyses of large cohorts consistently find a U-shaped relationship, with the lowest mortality at around 3 to 4 cups a day, roughly 12 to 17 percent lower than non-drinkers. Found in UK Biobank, in EPIC across ten European countries, and in US cohorts, and it holds for decaffeinated coffee too, which points at compounds other than caffeine.
• Type 2 diabetes: each additional cup a day is associated with roughly 6 percent lower risk, one of the more consistent findings in nutritional epidemiology. Also holds for decaf.
• Liver: substantial and consistent associations with lower liver fibrosis, cirrhosis, hepatocellular carcinoma, and liver enzyme levels. The liver evidence is the strongest for any organ.
• Parkinson's disease: consistent inverse association, dose-dependent, likely caffeine-mediated via adenosine A2A receptors.
• Endometrial and liver cancer: inverse associations. Coffee was removed from the IARC "possibly carcinogenic" list in 2016; the residual finding was that very hot beverages above 65 °C are probably carcinogenic to the oesophagus, regardless of what they are.

Tea:

• Green tea catechins and cardiovascular markers: modest reductions in LDL and blood pressure in meta-analyses.
• Cohort associations with lower cardiovascular disease and stroke, strongest in Chinese and Japanese populations where consumption is high.
• Black tea and blood pressure: small reductions.
• Green tea for weight loss: repeatedly tested, effects are very small and clinically insignificant.

Both remain observational for the big outcomes. Coffee drinkers differ from non-drinkers, and residual confounding cannot be excluded. What can be said is that the accumulated evidence gives no support to the idea that moderate coffee or tea consumption is harmful, and considerable support to the idea that it is at worst neutral.

6. Iron: the most actionable fact here

Polyphenols in tea and coffee bind non-haem iron in the gut and block its absorption.

• A cup of tea with a meal can reduce non-haem iron absorption by 60 percent or more. Coffee by around 40 to 60 percent.
• The effect applies to plant iron, fortified foods, and supplements. Haem iron from meat is largely unaffected.
• Timing solves it. Drinking tea or coffee an hour before or after a meal, rather than with it, largely removes the effect.

For anyone who is iron deficient, vegetarian or vegan, menstruating heavily, pregnant, or taking iron supplements, this single change is worth more than most dietary tinkering (Chapter 16).

7. Who should be careful

• Pregnancy. Guidance is generally under 200 mg of caffeine a day (about two mugs of instant coffee or one strong filter coffee), based on associations with miscarriage and low birth weight. Caffeine clearance slows substantially in later pregnancy, so the same dose produces higher and longer exposure.
• Anxiety and panic disorder. Caffeine is a reliable trigger for panic in susceptible people.
• Sleep. Caffeine's half-life is around 5 hours, so a 4pm coffee still has a quarter of its dose active at bedtime. Even when people fall asleep, deep sleep is measurably reduced. This is covered in detail in Chapter 88.
• Reflux. Both coffee and tea relax the lower oesophageal sphincter.
• Arrhythmia. Long-standing advice to avoid caffeine with palpitations has weakened; several large studies find no increase in arrhythmia with moderate coffee, and some find less. Individual sensitivity varies.
• Green tea extract supplements have caused acute liver injury, sometimes severe and requiring transplantation. They appear repeatedly in drug-induced liver injury registries. Drinking green tea does not carry this risk; the concentrated extract does. EFSA identified doses of 800 mg EGCG a day or more as a concern.
• Osteofluorosis from extremely heavy tea drinking (several litres a day for years) is a documented rarity.
• Drug interactions. Caffeine is metabolised by CYP1A2, so fluvoxamine, ciprofloxacin, and oral contraceptives slow its clearance substantially; smoking speeds it up, which is why smokers who quit often become suddenly caffeine-sensitive. Tea and coffee also reduce absorption of levothyroxine and some antibiotics (Chapter 84).

8. Practical guidance

• Three to four cups of coffee a day sits at the bottom of the mortality curve in the cohort data, and there is no reason to start drinking coffee for health if you do not want to.
• Filter your coffee if your cholesterol is a concern. Cafetière and boiled coffee raise LDL; paper-filtered does not.
• Do not drink it with a meal if you need the iron. Wait an hour.
• Cut it off 8 to 10 hours before bed if you sleep badly. Most people underestimate this.
• Do not let it get to 65 °C in your mouth: very hot drinks are the one part of the coffee/tea IARC finding that survived.
• Watch what goes in it. A large flavoured coffee shop drink can carry 400 to 600 kcal and 50 g of sugar. The coffee is not the problem.
• Green tea, brewed, is fine and useful. Green tea extract capsules are not worth the liver risk.
• Steeping: green tea at 70 to 80 °C for 1 to 2 minutes (boiling water makes it bitter and astringent); black tea at 95 to 100 °C for 3 to 5 minutes. Longer steeping extracts more caffeine and more tannin.
• Matcha delivers the whole leaf, so more catechins, more caffeine, and more of everything else in the leaf including any contaminants.

9. The bottom line

• Coffee's reputation reversed once studies adjusted properly for smoking. Three to four cups a day is associated with the lowest all-cause mortality, and the liver and type 2 diabetes findings are the most consistent.
• Brewing method changes the cholesterol effect. Unfiltered coffee raises LDL through cafestol and kahweol; a paper filter removes them.
• All true teas are one plant, differing only in oxidation. L-theanine is what makes tea feel different from coffee.
• Tea and coffee with a meal cut plant iron absorption by half or more. Move them an hour away from meals if iron matters to you.
• Green tea is good. Green tea extract capsules have caused liver failure.
• Keep caffeine under 200 mg a day in pregnancy, and stop it 8 to 10 hours before bed if you value sleep.

Sources and notes

Botany, processing, and decaffeination methods follow standard beverage technology references. Caffeine content ranges follow USDA data and published surveys showing wide variation between outlets. Diterpene (cafestol and kahweol) removal by paper filtration and the resulting LDL difference follow Urgert and Katan's work and the Norwegian cohort analysis by Tverdal et al., European Journal of Preventive Cardiology, 2020. Coffee and all-cause mortality follows Gunter et al., Annals of Internal Medicine, 2017 (EPIC), Loftfield et al. (UK Biobank), and Poole et al.'s umbrella review, BMJ, 2017. Type 2 diabetes follows Ding et al.'s meta-analysis. Liver outcomes follow Kennedy et al.'s systematic reviews. Parkinson's disease follows Ross et al. and the adenosine A2A literature. IARC's 2016 reclassification of coffee and its finding on very hot beverages is IARC Monograph 116. Green tea catechins and cardiovascular markers follow Hartley et al.'s Cochrane review. L-theanine with caffeine follows Haskell et al.'s cognitive studies. Tea and coffee inhibition of non-haem iron absorption follows Hurrell, Reddy, and Cook, British Journal of Nutrition, 1999. Green tea extract hepatotoxicity follows EFSA's 2018 assessment and LiverTox.

Open questions. Almost all the coffee outcome evidence is observational, and although it is unusually consistent, residual confounding by smoking and socioeconomic position cannot be excluded. Whether decaffeinated coffee shares the benefits is suggested by several cohorts and not established.

👉 Next: alcohol, where the science moved substantially in the last decade.