🟠 Moderate Evidence
The conventional advice to stop drinking coffee at 2pm applies to roughly half the population—but not the other half. A single genetic variant in the CYP1A2 gene determines whether you metabolize caffeine in 4 hours or stretch it to 7 hours or longer, according to research on caffeine pharmacogenomics published in the British Journal of Clinical Pharmacology (1999).
Key takeaways
- 95% of ingested caffeine is metabolized by a single liver enzyme, CYP1A2, whose activity varies dramatically by genotype
- Fast metabolizers (A/A genotype, ~46% of populations) clear caffeine with a 4-hour half-life; slow metabolizers (C/C genotype, ~10%) extend it to 7+ hours
- A slow metabolizer drinking coffee at 2pm retains roughly 50% of peak caffeine at 10pm bedtime, whereas a fast metabolizer has cleared 70%
- The practical cutoff time for slow metabolizers is closer to 11am or noon, not 2pm
Study at a Glance
| Source | British Journal of Clinical Pharmacology |
| Study type | Observational pharmacogenetic characterization |
| Sample size | N = 185 healthy non-smokers |
| Population | Adults without smoking history |
| Country | Not specified in source summary |
Caffeine metabolism varies sharply by genotype
Population frequency and half-life by CYP1A2 variant (rs762551), based on 185 healthy participants
Source: Sachse et al., British Journal of Clinical Pharmacology, 1999 | Georgian Medical Journal News
One gene controls 95% of your caffeine clearance
Caffeine is metabolized almost entirely by a liver enzyme called CYP1A2. According to research on caffeine metabolism, approximately 95 percent of the caffeine you ingest passes through this single enzyme (Sachse et al., British Journal of Clinical Pharmacology, 1999). The activity of CYP1A2 varies dramatically between individuals, and one of the largest sources of that variation is a single-letter DNA difference in the CYP1A2 gene called rs762551.
The rs762551 polymorphism creates three common genotypes with distinct metabolic profiles. According to Sachse and colleagues’ 1999 study of 185 healthy non-smokers, the population distribution is: A/A genotype (fast metabolizers, ~46%), A/C genotype (intermediate, ~44%), and C/C genotype (slow metabolizers, ~10%). This distribution appears consistent across diverse populations.
Peak levels are identical; the decay is what differs
When a fast metabolizer and slow metabolizer each consume 200 mg of caffeine—roughly a strong cup of brewed coffee—both reach similar peak plasma caffeine concentrations within about 60 minutes. After that convergence, the trajectories diverge sharply based on genotype.
In fast metabolizers (A/A), the average half-life of caffeine is approximately 4 hours. In slow metabolizers (C/C), the half-life extends to 7 hours or longer. This means that 8 hours after consuming a single dose, fast metabolizers have cleared roughly 70 percent of their peak caffeine level, while slow metabolizers still retain approximately 50 percent in circulation. The same dose consumed at the same time produces nearly identical initial peaks but radically different elimination patterns.
By 8 hours after a 200 mg caffeine dose, fast metabolizers have cleared 70% of peak levels, while slow metabolizers retain 50%—a four-fold difference in sleep-relevant exposure.
— Sachse et al., British Journal of Clinical Pharmacology (1999)
The 2pm rule applies only to half the population
This pharmacokinetic difference translates directly into practical sleep implications. A fast metabolizer (A/A genotype) who drinks coffee at 2pm will have metabolized most of the caffeine by a typical 10pm bedtime. For this group, the conventional 2pm cutoff is roughly accurate. A slow metabolizer (C/C genotype) consuming the same coffee at the same time remains substantially caffeinated at 10pm, with roughly half their peak still circulating.
For many slow metabolizers, sleep-protective timing requires shifting the practical “last coffee” cutoff much earlier—typically to 11am or noon. This genetic difference has direct implications for sleep quality, as population studies published in JAMA have documented associations between caffeine timing and sleep disruption (Cornelis et al., 2006).
The implications extend beyond anecdotal sleep trouble: slow caffeine metabolizers show measurably higher cardiovascular risk when consuming large amounts of caffeine, and chronotype (circadian preference) may interact with metabolic genotype to determine optimal consumption windows. For clinical practice and patient counseling, a one-size-fits-all 2pm deadline misses the reality that roughly half of any population should shift that boundary earlier based on their genetic caffeine-clearing capacity.
What this means
Frequently asked questions
Can I find out my CYP1A2 genotype?
Yes. The rs762551 polymorphism is included in many direct-to-consumer genetic tests (such as 23andMe) and can be formally ordered by a clinician through pharmacogenomic testing services. Once you know your genotype, you can adjust your caffeine cutoff time accordingly: A/A genotypes can safely consume caffeine until 2–3pm; C/C genotypes should shift to 11am–noon.
Does this affect decaffeinated or low-caffeine beverages?
The genetic variation in CYP1A2 applies to all caffeine metabolism, including decaffeinated coffee (which still contains 2–5 mg per cup) and tea. However, the total caffeine dose is so much lower that the timing effects are less pronounced. The half-life difference matters most with high doses (150+ mg).
Can other factors override my genotype?
Yes. Pregnancy, hormonal contraceptives, liver disease, and some medications slow CYP1A2 activity regardless of genotype. Smoking induces CYP1A2 and speeds caffeine clearance. Age and circadian rhythm also influence caffeine’s sleep impact. Genotype is a primary determinant but not the only one.
As personalized medicine expands into behavioral health, pharmacogenetic testing for common substances like caffeine offers a low-cost way to align lifestyle advice with individual biology. The 2pm rule remains statistically sound for a population median—but for millions of slow metabolizers, an earlier cutoff time backed by genetic evidence may be the difference between good sleep and chronic sleep disruption. Testing is increasingly available, and the knowledge itself costs nothing but a conversation with a clinician.
Source: Sachse et al., British Journal of Clinical Pharmacology (1999) and Cornelis et al., JAMA (2006)
Was this article helpful?
Disclaimer. This article is health journalism intended for general information and education. It is not medical advice and is not a substitute for professional diagnosis or treatment. Always consult a qualified healthcare provider about your individual circumstances. Full disclaimer →
Related Coverage
Editorial standards. This article was produced under the GMJ News editorial process, with oversight by the GMJ Editorial Board. Our editorial process. Spotted an error? Contact the editorial team.
