Coffee

What it is

Brewed coffee supplies caffeine and chlorogenic acids, and unfiltered coffee also cafestol, which raises cholesterol ¹²³.

A 237 g cup contains about 95 mg caffeine ¹. Filter coffee averages 70 mg of 5-caffeoylquinic acid, or chlorogenic acid, per 100 ml ². Cafetière coffee, 0.9 litres a day, supplied 38 mg of the diterpene cafestol, paper-filtered coffee less than 1 mg ³.

In the body

Caffeine is absorbed completely, blocks adenosine receptors in the brain and is broken down in the liver, mainly by one enzyme, CYP1A2 ⁴.

CYP1A2 performs over 95% of primary caffeine metabolism, mostly N3-demethylation to paraxanthine, 70% to 80% of a dose ⁵. The half-life is about 4 hours (range 2 to 8), and 6 to 16 hours in pregnancy, when CYP1A2 activity falls ⁴.

Caffeine breakdown through paraxanthineHepatocyte endoplasmic reticulum, then cytosol

Caffeine (1,3,7-trimethylxanthine)
Cytochrome P450 1A2 (CYP1A2), N3-demethylationO2, NADPH via cytochrome P450 reductase70% to 80% of a dose; CYP1A2 performs over 95% of primary metabolism
Paraxanthine (1,7-dimethylxanthine)
Cytochrome P450 1A2 (CYP1A2), N7-demethylationO2, NADPH via cytochrome P450 reductaseParaxanthine also yields 1,7-dimethyluric acid (CYP2A6) and, via NAT2, AFMU
1-Methylxanthine
Xanthine dehydrogenase/oxidase (XDH)Mo-molybdopterin, FAD, Two [2Fe-2S] clustersMajor urinary metabolite, as are paraxanthine, 1-methylxanthine and AFMU
1-Methyluric acid

Smoking induces CYP1A2 and raises caffeine clearance up to 1.72-fold ⁴. In two studies, CYP1A2 activity was higher with the −163AA genotype (rs762551) in smokers only ⁴. Caffeine antagonises adenosine A1 and A2A receptors ⁴.

What studies show

Unfiltered coffee raises LDL cholesterol; coffee drinkers die at slightly lower rates, but gene variants that raise coffee intake are not associated with lower mortality ³⁶⁷.

In a randomised trial in 46 adults, 24 weeks of cafetière coffee raised LDL cholesterol by 0.26 mmol/L (95% CI 0.04 to 0.47) against filtered coffee ³.

In UK Biobank, 14,225 of 498,134 adults died over a median 7 years; against non-drinkers, the adjusted hazard ratio at 2 to 3 cups a day was 0.88 (0.84 to 0.93), similar for decaffeinated coffee and across genetic caffeine-metabolism scores ⁶. In Copenhagen, coffee intake had a U-shaped association with death, but alleles near AHR and CYP1A1/2 that raise intake were not associated with it (hazard ratio per allele 1.01, 0.99 to 1.03; 12,656 deaths among 112,509) ⁷.

Among 2635 pregnant women in England, fetal growth restriction occurred in 10.9% below 100 mg caffeine a day and 18.2% at 300 mg or more (adjusted odds ratio 1.4, 1.0 to 2.0); tea supplied 62% of the caffeine ⁸.

References

Textbook or database
Fukagawa NK, McKillop K, Pehrsson PR, Moshfegh A, Harnly J, Finley J. USDA's FoodData Central: what is it and why is it needed today? American Journal of Clinical Nutrition, 2022.
doi:10.1093/ajcn/nqab397 PubMed 34893796
Funding and finding
Funding
All authors are employed by the USDA Agricultural Research Service and have not received external support in the preparation of this manuscript.
Finding
FoodData Central record read 2026-10-21: 'Beverages, coffee, brewed, prepared with tap water' (SR Legacy, FDC 171890, NDB 14209; includes fast food and home brewed coffee; scientific name Coffea), per 100 g: caffeine 40 mg (analytical, 8 data points, range 30 to 44 mg); theobromine 0 mg (assumed zero); water 99.4 g; energy 1 kcal; potassium 49 mg. Portions: 1 cup (8 fl oz) 237 g, 6 fl oz 178 g.
Textbook or database
Neveu V, Perez-Jiménez J, Vos F, Crespy V, du Chaffaut L, Mennen L, et al. Phenol-Explorer: an online comprehensive database on polyphenol contents in foods Database (Oxford), 2010.
doi:10.1093/database/bap024 PubMed 20428313
Funding and finding
Funding
French Ministry of Research (grant NUTRIALIS, 2002-04) and by Unilever, Danone and Nestlé (2005-09). Open access charges paid by the Institut National de la Recherche Agronomique.
Finding
Coffee beverage [Filter] (food 552), Phenol-Explorer version 3.6, read 2026-10-21; chromatography, mean content per 100 ml (minimum to maximum; n): 5-caffeoylquinic acid 70.03 mg (47.89 to 96.00; 4); 4-caffeoylquinic acid 59.60 mg (53.00 to 66.21; 2); 3-caffeoylquinic acid 51.80 mg (40.00 to 63.60; 2); 5-feruloylquinic acid 11.69 mg (1); 4-feruloylquinic acid 8.57 mg (1); 3-feruloylquinic acid 4.17 mg (1); 3,4-, 4,5- and 3,5-dicaffeoylquinic acids 2.66, 2.05 and 1.55 mg (1 each); caffeic acid 0.03 mg. Total polyphenols by Folin assay 266.70 mg (245.54 to 281.82; 3). Compound 475, 5-caffeoylquinic acid: synonym chlorogenic acid; phenolic acid, hydroxycinnamic acid subclass; C16H18O9.
Independent research
Urgert R, Meyboom S, Kuilman M, Rexwinkel H, Vissers MN, Klerk M, et al. Comparison of effect of cafetière and filtered coffee on serum concentrations of liver aminotransferases and lipids: six month randomised controlled trial BMJ, 1996.
doi:10.1136/bmj.313.7069.1362 PubMed 8956701
Funding and finding
Participants
46
Follow-up
24 weeks of treatment after a 4-week run-in, then 12 weeks of follow-up
Funding
Netherlands Heart Foundation through the Netherlands Organisation of Scientific Research (grant No 900-562-091).
Finding
Participants drank 0.9 litres a day of strong cafetière or paper-filtered coffee (Douwe Egberts Roodmerk). Cafetière coffee provided 38 (SD 6) mg cafestol and 33 (5) mg kahweol a day; filtered coffee less than 1 mg of either. After 24 weeks LDL cholesterol was raised by 0.26 mmol/l (95% CI 0.04 to 0.47; P = 0.03), 9% over baseline, relative to filtered coffee; LDL rose 9-14% during treatment; HDL cholesterol was not affected. Triglycerides rose 26% within 2 weeks and 7% at 24 weeks. Alanine aminotransferase rose by up to 80%, and by 9 U/l (3 to 15), 45%, at 24 weeks; it exceeded the upper limit of normal at least once in 8 of 22 cafetière and 1 of 24 filtered-coffee drinkers. All increases were reversible after the intervention stopped.
Regulator
EFSA Panel on Dietetic Products, Nutrition and Allergies (NDA). Scientific Opinion on the safety of caffeine EFSA Journal, 2015.
doi:10.2903/j.efsa.2015.4102
Funding and finding
Funding
European Food Safety Authority, on request from the European Commission (Question No EFSA-Q-2013-00220), adopted on 23 April 2015.
Finding
Caffeine is rapidly (tmax 30-120 minutes) and completely absorbed and crosses the blood-brain and placental barriers. N-3 demethylation to paraxanthine by hepatic CYP1A2 is the main route (70-80%); CYP1A2 accounts for about 95% of caffeine clearance. Plasma half-life about 4 hours (range about 2 to 8). In a study of saliva caffeine clearance, smoking increased clearance 1.22-, 1.47-, 1.66- and 1.72-fold for 1-5, 6-10, 11-20 and more than 20 cigarettes a day, and each litre of daily coffee 1.45-fold; CYP1A2 induction in smokers is reported; CYP1A2 activity was higher in smokers with the -163AA genotype (CYP1A2*1F, rs762551) than with AC or CC, without differences in non-smokers. In pregnancy the half-life increased to 6-16 hours, and at the end of pregnancy is three to four times longer than in the non-pregnant state; CYP1A2 activity was progressively lower than post partum, by 65.2% at weeks 36-40. The fetus and placenta cannot metabolise caffeine. Caffeine acts as an antagonist at adenosine A1 and A2A receptors and is a non-specific phosphodiesterase inhibitor with a Ki of 48 µM, the peak concentration after a 450 mg dose. Habitual intake up to 400 mg a day raises no safety concern for non-pregnant adults; up to 200 mg a day in pregnancy raises no safety concern for the fetus, based on two prospective cohorts (CARE Study Group 2008; Sengpiel 2013) showing a dose-dependent association with fetal growth restriction and small for gestational age, observed at all levels of intake with no threshold and judged clinically relevant at about 200 mg; the Panel noted that cohorts cannot provide evidence of causation but assumed the relationship causal for the assessment.
Textbook or database
Thorn CF, Aklillu E, McDonagh EM, Klein TE, Altman RB. PharmGKB summary: caffeine pathway Pharmacogenetics and Genomics, 2012.
doi:10.1097/FPC.0b013e3283505d5e PubMed 22293536
Funding and finding
Funding
This work was supported by the NIH/NIGMS (R24 GM61374).
Finding
Caffeine is almost completely metabolised, 3% or less being excreted unchanged in urine. The main route, 70% to 80% of the dose, is N3-demethylation to paraxanthine by hepatic CYP1A2; 1-N-demethylation to theobromine and 7-N-demethylation to theophylline each account for about 7% to 8%, and C-8 hydroxylation to 1,3,7-trimethyluric acid for the remaining 15%. CYP1A2 is responsible for more than 95% of the primary metabolism of caffeine, which is therefore used as a probe drug for CYP1A2 activity. The secondary metabolism of paraxanthine to 1-methylxanthine, 1-methyluric acid, 5-acetylamino-6-formylamino-3-methyluracil and 1,7-dimethyluric acid is carried out by CYP1A2, CYP2A6, N-acetyltransferase 2 and xanthine dehydrogenase. CYP2C8, CYP2C9 and CYP3A4 also take part in primary metabolism, with lower intrinsic clearance than CYP1A2.
Independent research
Loftfield E, Cornelis MC, Caporaso N, Yu K, Sinha R, Freedman N. Association of Coffee Drinking With Mortality by Genetic Variation in Caffeine Metabolism: Findings From the UK Biobank JAMA Internal Medicine, 2018.
doi:10.1001/jamainternmed.2018.2425 PubMed 29971434
Funding and finding
Participants
498,134
Follow-up
up to 10 years (median 7 years), 3.4 million person-years, 2006 to 2016
Funding
UK Biobank resource (established by the Wellcome Trust, the Medical Research Council, the UK Department of Health and the Scottish Government, with funding from the Welsh Assembly Government, the British Heart Foundation and Diabetes UK); Intramural Research Program of the NIH, Division of Cancer Epidemiology and Genetics, National Cancer Institute; Department of Health and Human Services.
Finding
14,225 deaths occurred. Against non-coffee drinkers, multivariable-adjusted hazard ratios for all-cause mortality were 0.94 (95% CI 0.88 to 1.01) for less than 1 cup a day, 0.92 (0.87 to 0.97) for 1, 0.88 (0.84 to 0.93) for 2 to 3, 0.88 (0.83 to 0.93) for 4 to 5, 0.84 (0.77 to 0.92) for 6 to 7 and 0.86 (0.77 to 0.95) for 8 or more cups a day. Similar associations were observed for instant, ground and decaffeinated coffee and regardless of genetic caffeine metabolism scores built from variants near AHR, CYP1A2, CYP2A6 and POR; for 6 or more cups a day hazard ratios across score strata ranged from 0.70 (0.53 to 0.94) to 0.92 (0.78 to 1.10), P = 0.17 for heterogeneity. Adjustment for detailed smoking history had the largest effect on estimates; participation was about 5.5%.
Independent research
Nordestgaard AT, Nordestgaard BG. Coffee intake, cardiovascular disease and all-cause mortality: observational and Mendelian randomization analyses in 95 000-223 000 individuals International Journal of Epidemiology, 2016.
doi:10.1093/ije/dyw325 PubMed 28031317
Funding and finding
Participants
112,509
Follow-up
median 6.0 years for all-cause mortality; to November 2014
Funding
This work was supported by Herlev and Gentofte Hospital, Copenhagen University Hospital and the Danish Heart Foundation. The funding organizations had no role in the design and conduct of the study, the collection, analysis and interpretation of the data or in the writing of the paper.
Finding
Observationally, among 95,366 people with 5422 deaths, coffee intake had U-shaped associations with ischaemic heart disease, ischaemic stroke and all-cause mortality, lowest at medium intake; current smokers were 12% of non-drinkers and 36% of those drinking more than 5 cups a day. A caffeine intake allele score (variants near AHR and CYP1A1/2) was associated with 42% higher coffee intake from 0 to 4 alleles (2.2 against 3.1 cups a day; R2 0.009). Among 112,509 people with 12,656 deaths, hazard ratios per allele were 1.01 (95% CI 0.99 to 1.03) for all-cause mortality, 1.02 (0.99 to 1.06) for cardiovascular mortality and 1.02 (1.00 to 1.03) for ischaemic heart disease. With 80% power the genetic analysis could exclude a per-allele odds ratio of 0.97; about 225,000 cases and 225,000 controls would be needed to exclude the observational hazard ratio of 0.86 for 4 to 5 against 0 cups a day.
Independent research
CARE Study Group. Maternal caffeine intake during pregnancy and risk of fetal growth restriction: a large prospective observational study BMJ, 2008.
doi:10.1136/bmj.a2332 PubMed 18981029
Funding and finding
Participants
2,635
Follow-up
From 8 to 12 weeks of pregnancy to delivery; recruitment September 2003 to June 2006
Funding
Food Standards Agency, United Kingdom, Grant contract No T01032/33. The steering group included Food Standards Agency staff.
Finding
Mean caffeine intake in pregnancy was 159 mg a day; tea supplied 62%, coffee 14%, cola 12% and chocolate 8%. Fetal growth restriction (birth weight below the 10th customised centile) occurred in 122 of 1122 women below 100 mg a day, 90 of 691 at 100-199 mg, 63 of 376 at 200-299 mg and 63 of 347 at 300 mg or more. Odds ratios adjusted for salivary cotinine and alcohol were 1.2 (95% CI 0.9 to 1.6), 1.5 (1.1 to 2.1) and 1.4 (1.0 to 2.0) against below 100 mg (trend P = 0.02). More than 200 mg a day was associated with about 60-70 g lower birth weight. Risk rose with intake with no threshold. The association appeared stronger with faster caffeine clearance (interaction P = 0.06).
Industry-funded
Kempf K, Herder C, Erlund I, Kolb H, Martin S, Carstensen M, et al. Effects of coffee consumption on subclinical inflammation and other risk factors for type 2 diabetes: a clinical trial American Journal of Clinical Nutrition, 2010.
doi:10.3945/ajcn.2009.28548 PubMed 20181814
Funding and finding
Registration
ISRCTN12547806
Participants
47
Follow-up
Three consecutive 1-month stages: 0, 4 and 8 cups (150 ml) of paper-filtered coffee a day
Funding
Supported by the Institute of Scientific Information on Coffee, which is a consortium of major European Coffee Companies. The Institute of Scientific Information had no role in drafting the design, in implementation of the study, in data analysis and interpretation. Additional support from the German Federal Ministry of Health and the Ministry of Innovation, Science, Research and Technology of North Rhine-Westphalia.
In-kind support
Juhla Mokka coffee provided by the Paulig Group, Helsinki.
Finding
Single-blind (investigator), single-group, non-randomised trial without placebo. After 8 against 0 cups a day, serum total cholesterol rose 12%, HDL cholesterol 7% and apolipoprotein A-I 4%, and the LDL/HDL cholesterol and apoB/apoA-I ratios fell 8% and 9%; LDL cholesterol and triglycerides did not change significantly. Interleukin-18 fell 8%, 8-isoprostane 16% and adiponectin rose 6% (medians); markers of glucose metabolism did not change. The authors concluded that coffee appears to have beneficial effects on subclinical inflammation and HDL cholesterol.
Expert consensus
IARC Working Group on the Evaluation of Carcinogenic Risks to Humans. Drinking Coffee, Mate, and Very Hot Beverages IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, volume 116, 2018.
PubMed 31310458
Funding and finding
Funding
This programme has been supported since 1982 by Cooperative Agreement U01 CA33193 with the United States National Cancer Institute, Department of Health and Human Services. Additional support has been provided since 1986 by the European Commission Directorate-General for Employment, Social Affairs, and Inclusion, and since 1992 by the United States National Institute of Environmental Health Sciences. The contents of this volume are solely the responsibility of the Working Group.
Finding
There is inadequate evidence in humans and in experimental animals for the carcinogenicity of drinking coffee. There is evidence suggesting lack of carcinogenicity of drinking coffee in humans for cancers of the pancreas, liver, female breast, uterine endometrium and prostate; inverse associations were observed with cancers of the liver and uterine endometrium. Drinking coffee is not classifiable as to its carcinogenicity to humans (Group 3).
Industry-funded
Brito Nunes C, Huang P, Wang G, Lundberg M, D'Urso S, Wootton RE, et al. Mendelian randomization study of maternal coffee consumption and its influence on birthweight, stillbirth, miscarriage, gestational age and pre-term birth International Journal of Epidemiology, 2023.
doi:10.1093/ije/dyac121 PubMed 35679582
Funding and finding
Funding
The UK Medical Research Council and Wellcome (217065/Z/19/Z) and the University of Bristol provide core support for ALSPAC; genotyping of ALSPAC mothers was funded by the Wellcome Trust (WT088806) and offspring samples were genotyped by the Wellcome Trust Sanger Institute and LabCorp using support from 23andMe. Author support from the Norwegian Research Council, Nils Normans minnegave, NHMRC, South-Eastern Norway Regional Health Authority, British Heart Foundation, UK Medical Research Council, University of Bristol, University of Queensland and CSIRO. The funders have no role in the study design, analyses or interpretation of results.
In-kind support
23andMe provided summary statistics for gestational age and pre-term birth under a data-access agreement.
Finding
In two-sample and one-sample genetic analyses, genetically predicted coffee consumption showed no change in risk of sporadic miscarriage, stillbirth or pre-term birth, or effect on gestational age. Birthweight associations were positive and inconsistent in size (two-sample, six variants: 0.050 SD per cup a day, 95% CI 0.001 to 0.098). Assuming the genetic score explained 0.37% of variance in coffee intake, UK Biobank had 80% power to detect a 26.9 g difference in birthweight or 0.03 more stillbirths per daily cup. In ALSPAC, self-reported coffee at week 32 was associated with -9.63 g birthweight per cup (-17.57 to -1.69), attenuated to -4.52 g (-14.92 to 5.88) after adjustment for smoking and alcohol. The authors note that a null association may reflect a lack of power.
Industry-funded
Ricketts ML, Boekschoten MV, Kreeft AJ, Hooiveld GJ, Moen CJ, Müller M, et al. The cholesterol-raising factor from coffee beans, cafestol, as an agonist ligand for the farnesoid and pregnane X receptors Molecular Endocrinology, 2007.
doi:10.1210/me.2007-0133 PubMed 17456796
Funding and finding
Funding
This work was supported by the United States Department of Agriculture (Grant ARS, CRIS 6250-51000-034); National Institutes of Health (Grant RO1-DK53366); Wageningen Centre for Food Sciences (project A201C and A500); the Dutch Organization for Scientific Research (project NWO 98010001); travel grant R94-202; NWO 903-39-174; and Netherlands Heart Foundation (NHS grant 97.116). The Wageningen Centre for Food Sciences is described in Olthof et al., PLoS Medicine 2005 (PMC1140947), as an alliance of six industrial partners (AVEBE, Cosun, CSM, DSM, Unilever and NZO) and four research organisations, with 29% of its funding from the industrial partners; partial industry funding sets the class to industry. Roles of funders are not stated.
In-kind support
GW4064 was a gift from GlaxoSmithKline.
Finding
In APOE3Leiden mice, cafestol raised serum cholesterol by 40% and triglycerides by 62%. Cafestol activated the ligand-binding domains and full-length forms of FXR and PXR in cell assays and recruited the coactivator SRC-1. In wild-type but not FXR-null mice, cafestol feeding repressed hepatic CYP7A1, CYP8B1 and NTCP, without inducing hepatic SHP or other positive FXR targets; it induced intestinal IBABP and FGF15, the latter through both FXR and PXR, and PXR-dependent intestinal CYP3A11, ABCA1 and CYP27A1. Cafestol does not raise serum cholesterol in wild-type mice; the relevance of the FXR-SHP repression of CYP7A1 to humans remains to be established.

Updated 2026-10-21