Some people drink a double espresso at 4pm and sleep fine. Others get jittery off a single cup of tea. This isn't willpower or tolerance built up over years — for a meaningful share of it, it's genetics. A specific gene, CYP1A2, controls how fast your liver breaks caffeine down, and the difference between "fast" and "slow" metabolizers is large enough to change how caffeine should factor into your day. It is also, as you'll see, not the biggest thing changing how fast you clear it.
The gene that does the work
The overwhelming majority of caffeine clearance happens through a single liver enzyme, cytochrome P450 1A2, encoded by the CYP1A2 gene — depending on how the pathway is counted, somewhere between 90% and 95%, with N3-demethylation to paraxanthine alone accounting for over 80%.15 A common variant, often referred to by the marker rs762551 (the *1A/*1F polymorphism), affects how much of this enzyme your liver produces and how readily production can be induced. People with the "fast" variant clear caffeine efficiently; people with the "slow" variant clear it much more gradually, meaning a given dose stays active for longer and produces a stronger, longer-lasting effect from the same cup of coffee.
The honest scale of that effect is worth stating. In heavy coffee drinkers, rs762551 accounted for 14% of the person-to-person variation in CYP1A2 activity in a Serbian sample and 22% in a Swedish one.5 Real, and a minority share — anyone selling a caffeine plan off a single-gene readout is overselling one variable.
The same 95mg of caffeine can act like a light dose for one person and a strong one for another — purely based on how fast their liver clears it.
The numbers that let you do the arithmetic
Half-life is the figure that actually changes behaviour. In non-smoking adults, caffeine's elimination half-life is around 6 hours; in smokers it is roughly 3.5, because smoking increases caffeine clearance by about 40–65%.5 At a 6-hour half-life, a 2pm coffee still has a quarter of its caffeine circulating at 2am. That is the whole explanation for a large share of "I just don't sleep well," and it's why the dose-and-timing trial is worth reading before you change anything else about your evenings.
Two other modifiers dwarf most genotype effects. Combined oral contraceptives containing ethinylestradiol reduced caffeine clearance by roughly 55% after a single treatment cycle — effectively doubling how long a dose lingers. And in pregnancy CYP1A2 activity falls progressively: about 33% in the first trimester, 48% in the second and 65% in the third, stretching caffeine's half-life to as long as 10.5 hours in the final weeks.5 If either applies to you, that matters more than your rs762551 result; fuller guidance is in caffeine and supplements during pregnancy and breastfeeding.
It isn't just about feeling jittery
The clearest evidence that this genetic difference might matter clinically comes from cardiovascular research, not subjective jitteriness. A 2006 case-control study in JAMA compared 2,014 survivors of a first non-fatal heart attack with 2,014 matched population controls in Costa Rica.2 Among slow metabolizers, the odds ratio for myocardial infarction rose with coffee intake: 1.36 (95% CI 1.01–1.83) at 2–3 cups a day and 1.64 (1.14–2.34) at four or more. Among fast metabolizers the same intakes showed nothing — 0.78 (0.56–1.09) and 0.99 (0.66–1.48).
Then the part that keeps this unsettled. The authors' own caveat is that because their cases were survivors of non-fatal heart attacks, the interaction could partly reflect who survives rather than who has one. And a much larger 2019 prospective analysis of UK Biobank data — up to 347,077 people and 8,368 cardiovascular cases — did not find the CYP1A2-by-coffee interaction at all.4 One 4,000-person case-control study found it; one 347,000-person prospective study didn't. Read it as an open question. If you have cardiovascular risk factors, the more useful list is in the ten things that cause nine in ten strokes.
Clearance isn't the only gene — the receptor matters for sleep
How fast you clear caffeine and how strongly you respond to it are separate questions. Caffeine works by blocking adenosine receptors, and variation in ADORA2A, which codes for the adenosine A2A receptor, has been linked specifically to caffeine-related sleep disturbance. A genome-wide association study in 2,402 coffee-drinking adults from 1,470 Australian twin families replicated a previously reported ADORA2A association with caffeine-induced sleep disruption.6 The same paper is blunt that no variant in it reached genome-wide significance — suggestive, awaiting independent replication, which is where most consumer caffeine genetics sits.
Why "just build a tolerance" doesn't fully apply here
Regular exposure blunts some of caffeine's subjective effects over days to weeks, but it doesn't change the enzyme activity your genes set — the mechanism we go into in whether a daily formula builds caffeine tolerance. An older study in Caucasian populations also found the classical *1F polymorphism didn't fully predict enzyme activity on its own, suggesting other genetic and environmental factors are at work.3 Either way: caffeine sensitivity is real, physiological, and not something everyone can train their way out of.
What this means for Cognitive Support — and what we can't tell you
This is where we have to be straight with you about our own label. Cognitive Support contains caffeine, named on the ingredient panel. But caffeine sits inside a 255mg proprietary blend alongside six other actives, and the per-capsule caffeine amount is not individually disclosed. We checked with our manufacturer in September 2026: per-ingredient figures for that blend aren't available for us to publish, so this isn't a gap we can close on the current formula.
The consequence, stated plainly: if you are a slow metaboliser, on a combined oral contraceptive, pregnant, or simply caffeine-sensitive, you cannot calculate your total daily caffeine intake from our label. That is a genuine limitation of this product for exactly the readers this article is written for. What we can do is show our working — how it compares to coffee, tea and energy drinks and the direct answer on whether it contains caffeine — and point out that of our range, only Sleep Formula discloses every ingredient at its own dose. If undisclosed blends are a dealbreaker, that's a reasonable position, and how to read a supplement label explains how to spot them elsewhere. Our general standard, including where we fall short of it, is in why we don't sell "brain blends".
What to actually do with this
If caffeine reliably disrupts your sleep or makes you anxious at doses that don't bother other people, the useful response isn't tolerance training — it's less, earlier, without treating your reaction as a personal failing. Practically: set a cut-off by working back roughly two half-lives from bedtime (for a non-smoker on no interacting medication, around midday for an 11pm bed), count every source including tea, cola, pre-workout and any supplement with caffeine on the label, and change one variable at a time. If you want calm focus without more caffeine, the pairing with the best trial base is L-theanine with caffeine.
Where the evidence stops
rs762551 explains a minority of the variation in enzyme activity — less than smoking or a combined oral contraceptive does. The headline cardiovascular interaction failed to replicate in a cohort 85 times larger. The ADORA2A sleep finding did not reach genome-wide significance. Direct-to-consumer CYP1A2 tests are not validated for producing individual caffeine advice, and a genotype result will not tell you your half-life. The reliable signal underneath is smaller than the marketing but still real: people differ physiologically in how caffeine behaves in them, that difference is not a matter of discipline, and your own sleep is a better instrument for measuring it than a saliva kit.
References
- Grzegorzewski J, Bartsch F, Köller A, König M. "Pharmacokinetics of Caffeine: A Systematic Analysis of Reported Data for Application in Metabolic Phenotyping and Liver Function Testing." Frontiers in Pharmacology, 2021. frontiersin.org
- Cornelis MC, El-Sohemy A, Kabagambe EK, Campos H. "Coffee, CYP1A2 genotype, and risk of myocardial infarction." JAMA, 2006;295(10):1135-41. 2,014 cases and 2,014 controls. pubmed.ncbi.nlm.nih.gov
- Sachse C, et al. "Detailed modelling of caffeine metabolism and examination of the CYP1A2 gene: lack of a polymorphism in CYP1A2 in Caucasians." Pharmacogenetics, 1999. pubmed.ncbi.nlm.nih.gov
- Zhou A, Hyppönen E. "Long-term coffee consumption, caffeine metabolism genetics, and risk of cardiovascular disease: a prospective analysis of up to 347,077 individuals and 8368 cases." American Journal of Clinical Nutrition, 2019;109(3):509-516. pubmed.ncbi.nlm.nih.gov
- "Unraveling the complexities of caffeine: metabolism, genetics, evolution, and health." Hereditas, 2026. Source for the half-life, smoking, oral contraceptive and pregnancy figures. link.springer.com
- "A Genome-Wide Association Study of Caffeine-Related Sleep Disturbance: Confirmation of a Role for a Common Variant in the Adenosine Receptor." SLEEP, 2012;35(7):967-75. n = 2,402. academic.oup.com
This article is for informational purposes only and has not been evaluated by the FDA. It is not intended to diagnose, treat, cure, or prevent any disease. Speak with a healthcare provider before making changes to your caffeine intake, especially if you have a cardiovascular condition or take medication.
Related reading: Does Cognitive Support Contain Caffeine?
If you're sensitive to caffeine, know what's disclosed and what isn't: Cognitive Support's full ingredient list is at procohq.com.