CYP2C19 and clopidogrel — what a pharmacogenomic result actually means
Most of consumer genetics is interesting rather than useful. Pharmacogenomics is the exception, and CYP2C19 is the clearest case in it: a well-studied gene, a widely prescribed medicine, and published clinical guidance connecting the two.
It is also a topic where the right action is almost never something you take on your own.
What does the CYP2C19 gene do?
It codes for a liver enzyme that processes a number of common medicines, including the antiplatelet drug clopidogrel, several proton pump inhibitors, and some antidepressants. Inherited variants change how much working enzyme a person produces, so the same dose can behave differently in different people.
Pharmacogenomics uses star allele notation rather than plain genotypes. *1 is the reference version, *2 and *3 reduce or remove function, and *17 increases it. Your two copies combine into a diplotype such as *1/*2.
That diplotype is then translated into a metaboliser status: poor, intermediate, normal, rapid or ultrarapid. It is the status, not the raw letters, that published guidance is written against.
Why does CYP2C19 matter for clopidogrel in particular?
Because clopidogrel is a prodrug — it arrives inactive and CYP2C19 is what converts it into the form that works. Someone producing little functioning enzyme converts less of the dose, so the antiplatelet effect is reduced. The drug’s labelling carries a warning about diminished effectiveness in poor metabolisers.
This is unusually direct for consumer genetics. Most findings describe a tendency; this one describes a mechanism with a known consequence and a documented alternative.
Published guidance from the pharmacogenomics consortium CPIC addresses exactly this situation, recommending that an alternative antiplatelet be considered for poor and intermediate metabolisers in certain cardiac settings. That guidance is written for prescribers, which is the important detail.
What do the CYP2C19 star alleles mean in practice?
*1/*1 is a normal metaboliser. One reduced-function copy, such as *1/*2, is intermediate. Two, such as *2/*2, is poor. *17 increases activity, so *1/*17 is rapid and *17/*17 ultrarapid. Frequencies differ substantially between populations, with loss-of-function alleles more common in East Asian ancestry.
The ultrarapid end matters too, though it gets less attention. Faster clearance can mean less benefit from medicines that need to persist, such as some proton pump inhibitors.
None of these statuses is a problem in itself. They describe how a body handles a class of molecules, and they only become relevant when one of those molecules is prescribed.
What should you do if your file shows a CYP2C19 variant?
Take it to a prescriber, and change nothing on your own. This matters more here than almost anywhere else in consumer genetics, because clopidogrel is usually prescribed after a cardiac event and stopping or altering it independently carries real risk. The result is information for the conversation, not a reason to act before it.
What it is genuinely good for is being known in advance. A result already in hand when a medicine is first considered is far more useful than one looked up afterwards, and this is one of the few genetic findings a prescriber can act on directly.
A line in a raw DNA file is not a diagnosis, and pharmacogenomic guidance is written for clinicians deciding between options with a patient in front of them.
Can a consumer DNA file read CYP2C19 reliably?
Partially, and the partial matters. Arrays commonly include the markers for *2 and *17, less often *3, and miss the many rarer alleles entirely. So a file can usually establish the common statuses, and cannot rule out a rare variant it never looked for.
That is different from CYP2D6, another important pharmacogene, which arrays cannot read reliably at all because the gene varies by whole-gene duplications and deletions rather than by single positions. We mark it as not readable rather than reporting a result we cannot support.
The honest position for CYP2C19 is that a consumer file gives a usable indication for the common alleles, stated with its coverage, and that clinical decisions rest on proper testing ordered by whoever is prescribing.
What we look at for medicines → · How we turn published research into a report →