A common CYP1A2 variant affects how quickly the body clears caffeine. A widely repeated claim that 'slow' metabolizers face higher heart-attack risk from coffee rests on limited evidence that has not held up well in larger studies.
Caffeine is broken down in the liver almost entirely by a single enzyme, CYP1A2, which accounts for roughly 95% of caffeine's metabolism. How quickly a person clears caffeine from their bloodstream depends on how active their personal copy of this enzyme is — and that activity varies a great deal from person to person, for both genetic and non-genetic reasons.
One of the most-studied genetic contributors to this variation is a variant near the CYP1A2 gene, identified by the marker rs762551. People are sometimes described as "fast" or "slow" caffeine metabolizers based on their genotype at this position: those with two copies of the common allele tend to have more inducible, higher-activity CYP1A2 and clear caffeine relatively quickly, while carriers of the variant allele (called CYP1A2*1F in the older nomenclature) tend to clear it more slowly. In practice this means that after the same cup of coffee, a "slow" metabolizer will have measurably higher caffeine levels in their blood for longer than a "fast" metabolizer will.
A well-known 2006 study of adults in Costa Rica found that among people carrying the slower-metabolizing variant, drinking 4 or more cups of coffee a day was associated with a higher risk of a first non-fatal heart attack, while no such association was seen in the fast-metabolizer group. This finding got a lot of attention and is still frequently repeated as settled fact in articles about "coffee genes."
What is less often repeated is that this single study, while carefully done, has not replicated consistently. A later analysis using data from more than 340,000 participants in the UK Biobank — a far larger sample than the original study — found that neither CYP1A2 genotype nor a broader genetic score for caffeine metabolism changed the relationship between habitual coffee drinking and cardiovascular disease risk. In other words, the biggest and most recent look at this question did not find the interaction that made the original finding famous.
This doesn't mean the original study was wrong, or that genotype never matters. It means the evidence is genuinely mixed, and a definitive verdict either way would be premature. Any claim that "slow metabolizers get heart attacks from coffee" is stating a hypothesis as if it were an established fact.
Genotype is only one influence on CYP1A2 activity, and often not the dominant one. Smoking is a strong inducer of CYP1A2 and can roughly double enzyme activity in smokers compared with non-smokers, regardless of genotype. Pregnancy, some medications (such as certain antidepressants and the antibiotic ciprofloxacin, which inhibit CYP1A2), and even diet can shift caffeine clearance more than this particular variant does on its own. A genetic report on caffeine metabolism describes a tendency, not a fixed trait, and it says nothing about whether someone smokes, is pregnant, or takes an interacting medication.
Gene and variant: CYP1A2, marker rs762551, historically referenced as defining the CYP1A2*1F allele (an A→C substitution intragenic to CYP1A2, located in intron 1, originally characterized by Sachse et al. 1999 using a caffeine metabolic-ratio phenotyping approach). Carriers of the *1F variant show decreased CYP1A2 inducibility — measured as the ratio of plasma or urinary caffeine to its metabolites after a caffeine dose — consistent with a "slow" metabolizer phenotype, while individuals homozygous for the ancestral allele (*1A/*1A) show greater inducibility and are termed "rapid" metabolizers (nomenclature as used in Cornelis et al. 2006, JAMA).
Key primary evidence (Cornelis et al. 2006, JAMA): In a case-control study of 2,014 first non-fatal myocardial infarction (MI) cases and 2,014 matched population controls in Costa Rica, 54–55% of participants were carriers of the *1F allele. Among *1F carriers, the multivariable-adjusted odds ratio (OR) for MI associated with ≥4 cups of coffee/day versus <1 cup/day was 1.64 (95% CI, 1.14–2.34). Among *1A/*1A homozygotes, the corresponding OR was 0.99 (95% CI, 0.66–1.48); the gene×coffee interaction was statistically significant (P=.04). The effect was more pronounced in participants younger than 50 (OR 4.07, 95% CI 1.89–8.74, for ≥4 cups/day in *1F carriers).
Non-replication in a much larger cohort: Zhou and Hyppönen (2019, American Journal of Clinical Nutrition) analyzed up to 347,077 UK Biobank participants (8,368 incident cardiovascular disease cases) and found a nonlinear (U-shaped) association between habitual coffee intake and cardiovascular disease risk overall, but this association was not modified by CYP1A2 genotype or by a composite caffeine-metabolism genetic score. This is a substantially larger and more statistically powered dataset than the original case-control study, and its null interaction result is an important, frequently omitted counterpoint. The evidence base for a CYP1A2-genotype-dependent coffee–MI relationship should be characterized as limited and inconsistently replicated, not as an established mechanism.
Role of induction (smoking): CYP1A2 is substantially inducible by polycyclic aromatic hydrocarbons in cigarette smoke; smoking increases CYP1A2 activity independent of genotype (Kalow & Tang, 1991). Cornelis et al. also note that the degree of induction by smoking is blunted in *1F carriers relative to *1A/*1A individuals, meaning genotype and smoking status interact rather than act independently. Any interpretation of caffeine-metabolism genotype should account for smoking status, use of CYP1A2 inhibitors (e.g., fluvoxamine, ciprofloxacin) or inducers, sex, and pregnancy, all of which can shift phenotype more than this single variant.
Guideline status: There is no CPIC (Clinical Pharmacogenetics Implementation Consortium) guideline for CYP1A2 genotype and caffeine, and no professional body issues caffeine-intake recommendations based on CYP1A2 genotype. This topic should be framed as an area of active, unsettled research rather than an actionable clinical finding.
What a 23andMe/AncestryDNA export or raw VCF can and can't tell you about Caffeine Metabolism (CYP1A2) comes down to these specific, well-studied positions — not a diagnosis.
It means your particular version of the CYP1A2 gene tends to produce an enzyme that is less inducible and clears caffeine from your bloodstream more slowly than average, so caffeine's effects may last longer after the same cup of coffee compared with a 'fast' metabolizer.
No. A single 2006 study found this association in one population, but a much larger 2019 study of over 340,000 people found that CYP1A2 genotype did not change the relationship between coffee and cardiovascular disease risk. The evidence is mixed and does not support treating this as a settled risk factor.
Smoking is a strong, well-documented inducer of CYP1A2 and can substantially increase enzyme activity regardless of genotype. For many people, smoking status, pregnancy, or interacting medications shift caffeine clearance more than this one genetic variant does.
No professional guideline recommends adjusting caffeine intake based on CYP1A2 genotype. This information describes a general tendency, not a personalized safety threshold.
No. Unlike genes such as CYP2D6, CYP1A2 and caffeine do not have a published CPIC guideline, reflecting the limited and inconsistent state of the evidence.
Free to reuse. This page's text is original writing from freely-available research, licensed CC BY 4.0 — reuse it, including commercially, with attribution to MyGeneLog. It's general research-derived information, not medical advice or a diagnosis — see Terms of Use.