CYP2C9

What it metabolises human data

CYP2C9 has a preference for weakly acidic and anionic substrates, which is why its list is dominated by the non-steroidal anti-inflammatories, the sulfonylureas, the sartans, and the two narrow-index drugs that define its clinical importance: warfarin and phenytoin. The warfarin case has a wrinkle worth stating precisely. Warfarin is administered as a racemate and S-warfarin is three to five times the more potent enantiomer for vitamin-K-epoxide-reductase inhibition; S-warfarin is cleared by CYP2C9 while R-warfarin goes through CYP1A2 and CYP3A4. So a CYP2C9 inhibitor moves the INR far more than its effect on total warfarin concentration suggests, and a CYP1A2 or CYP3A4 effect on warfarin is correspondingly muted. For this library the cannabinoid entries are the other reason the page matters: CYP2C9 performs the 11-hydroxylation of Δ9-THC that produces 11-hydroxy-THC, a metabolite at least as psychoactive as the parent, and cannabidiol is a CYP2C9 substrate and inhibitor as well.

Sources: Holbrook AM 2005 · Flockhart DA 2021 · U.S. Food 2023 · Sachse-Seeboth C 2009 · Bland TM 2005 · Watanabe K 2007 · Stout SM 2014

Drug inhibitors and inducers, with FDA potency classes human data

The FDA classes are defined the same way on every isoform: a strong inhibitor raises the AUC of a sensitive index substrate five-fold or more, a moderate inhibitor at least two-fold and less than five-fold, a weak inhibitor at least 1.25-fold and less than two-fold. CYP2C9 has few true strong inhibitors, which is a fact with a clinical sting in it — moderate inhibition of a narrow-index substrate like warfarin or phenytoin is more than enough to cause a bleed or a toxic level, so the absence of strong entries on this list is not reassurance.

Sources: U.S. Food 2023 · U.S. Food 2020* · Flockhart DA 2021 · Holbrook AM 2005

Botanical and cannabinoid inhibitors and inducers contested in vitro

The botanical picture at CYP2C9 is thinner than at CYP3A4 and the strongest entries are the two cannabinoids and the curcuminoids. Cannabidiol inhibits CYP2C9 and is also cleared by it. Curcuminoids inhibit CYP2C9 in the comparative in-vitro work alongside their action on CYP3A4, UGT and SULT. The operator corpus lists piperine as a CYP2C9 inhibitor as well as a CYP3A4 and CYP1A2 inhibitor; the controlled in-vitro comparison characterises piperine as relatively SELECTIVE for CYP3A4, so the CYP2C9 claim should be carried at lower confidence than the CYP3A4 one. Kavalactones inhibit CYP2C9 in vitro, and unlike the CYP3A4 case that claim has not been tested against a human in-vivo probe, so it is unconfirmed rather than contradicted. The clinically loudest botanical on this enzyme is not an inhibitor at all: St John’s wort induces CYP2C9 as part of its PXR programme and has been associated with loss of anticoagulant control.

Contested — caveat. Only the curcuminoid, cannabidiol and St John’s wort entries have reasonable support, and even those rest largely on in-vitro determinations plus pharmacovigilance rather than on controlled human pharmacokinetic studies with a CYP2C9 probe. The piperine CYP2C9 claim comes from the operator corpus and is at odds with the one controlled selectivity comparison. The cranberry literature is inconsistent.

Sources: Volak LP 2008 · Bahramsoltani R 2017 · Bhardwaj RK 2002 · Mathews JM 2002 · Zou L 2002 · Gurley BJ 2005 · Yamaori S 2011* · Stout SM 2014 · Nasrin S 2021 · Moore LB 2000 · Henderson L 2002 · Holbrook AM 2005 · Van Kush Family Research Institute (Temple Pharmacopoeia Project) 2026* · Bailey DG 2013

Clinical consequence, concretely human data

Warfarin plus a moderate CYP2C9 inhibitor is the archetype and the presentation is bruising, gum or nose bleeding, haematuria, black stool, or an intracranial haemorrhage with no trauma, arriving days after a new antibiotic or antifungal was started. Because the INR is monitored, this interaction is often caught — but the classic misses are the ones nobody counts as a drug: an oral miconazole gel for oral thrush, a topical or vaginal antifungal, a course of co-trimoxazole. Phenytoin plus an inhibitor gives nystagmus, ataxia, slurred speech and confusion, and because its kinetics are saturable a small clearance change at the top of the range produces a disproportionate rise. A sulfonylurea plus an inhibitor gives hypoglycaemia, which in an older patient is often recorded as a fall or delirium rather than as a drug event. Losartan plus an inhibitor gives the inversion: blood pressure rises because less active metabolite is formed. And in this library the case that matters most is the cannabinoid one — a CYP2C9 poor metaboliser taking oral THC reaches substantially higher exposure from the same dose, which is one mechanistic explanation for why identical edibles produce wildly different experiences in different people.

Sources: Holbrook AM 2005 · Flockhart DA 2021 · U.S. Food 2023 · Sachse-Seeboth C 2009

Polymorphism contested human data

CYP2C9 star-2 (an arginine-to-cysteine substitution at residue 144) and star-3 (isoleucine to leucine at 359) both reduce catalytic function, star-3 more severely, and both are common enough in European-ancestry populations to matter at the population level; additional reduced-function alleles including star-5, star-6, star-8 and star-11 occur at appreciable frequency in African-ancestry populations and were under-represented in the early dosing algorithms, which is a documented equity problem in warfarin pharmacogenomics rather than a technicality. Carriers need lower warfarin doses and have a higher risk of bleeding during initiation; genotype-informed dosing algorithms combining CYP2C9 with VKORC1 explain a substantial part of dose variance, although randomised trials of genotype-guided initiation have produced mixed results on hard outcomes. For this library the striking finding is the cannabinoid one: in a human pharmacokinetic study stratified by CYP2C9 genotype, star-3 homozygotes showed roughly threefold higher Δ9-THC exposure than star-1 homozygotes after an oral dose, with correspondingly less 11-hydroxy-THC formed. That is a genetic explanation for a variability everyone who has ever shared an edible has observed.

• Star-2 and star-3 are the classical reduced-function alleles; several further reduced-function alleles are commoner in African-ancestry populations and were missed by early algorithms.
• Lower warfarin dose requirement, higher early bleeding risk.
• Genotype plus VKORC1 explains much of warfarin dose variance; outcome trials of genotype-guided dosing are mixed.
• Δ9-THC exposure roughly threefold higher in star-3 homozygotes after oral dosing, with less 11-hydroxy-THC formed.

Contested — caveat. The CYP2C9 and THC finding is a small human study and has not been replicated widely; treat the threefold figure as one measurement rather than a constant. Randomised trials of genotype-guided warfarin initiation have not consistently improved clinical outcomes even though the pharmacokinetic association is solid, so genotype explains variance without automatically improving care.

Sources: Aithal GP 1999 · International Warfarin Pharmacogenetics Consortium (Klein TE 2009* · Zanger UM 2013 · Sachse-Seeboth C 2009 · Bland TM 2005