CYP3A4

What it metabolises human data

CYP3A4 has an unusually large and flexible active site, which is why its substrate list reads like a formulary rather than a chemical class. It handles most of the statins that are not renally or CYP2C9 cleared, the dihydropyridine calcium channel blockers, the macrolides, the azole antifungals, the HIV protease inhibitors, the calcineurin inhibitors, most benzodiazepines other than the directly glucuronidated ones, many opioids, most kinase inhibitors, the ergot alkaloids, and both Δ9-THC and cannabidiol in part. The clinically important distinctions inside a drug class are the ones worth memorising: midazolam, triazolam and alprazolam are CYP3A4 substrates while lorazepam, oxazepam and temazepam are glucuronidated and largely escape; simvastatin, lovastatin and atorvastatin are CYP3A4 substrates while pravastatin, rosuvastatin and fluvastatin are not.

Sources: Flockhart DA 2021 · U.S. Food 2023 · Lilja JJ 1998 · Bailey DG 2013 · Ruschitzka F 2000 · Watanabe K 2007 · Jiang R 2011 · Stout SM 2014 · Mathews JM 2002

Intestinal CYP3A4 and the dual first pass human data

CYP3A4 is the dominant CYP in the enterocytes of the proximal small intestine, and for an oral dose that layer is encountered before the liver. Two separate barriers in series means two separate places an inhibitor can act, and they do not produce the same result. An inhibitor confined to the gut lumen raises oral bioavailability without changing systemic clearance, so peak concentration and AUC rise while the elimination half-life is unchanged. An inhibitor that reaches the hepatocyte in effective concentration prolongs the half-life as well. Grapefruit juice is the first kind: its furanocoumarins inactivate enterocyte CYP3A4 by a mechanism-based route, enterocyte CYP3A protein has been shown to fall after ingestion, and the effect on an intravenous dose of the same drug is small. Ketoconazole and ritonavir are the second kind and act at both sites. The practical reading is that the magnitude of a CYP3A4 interaction depends on how much of the dose the gut was destroying, which is why the largest grapefruit effects are all on drugs with low oral bioavailability to begin with.

• Gut-only inhibition: AUC and peak rise, half-life unchanged, intravenous route spared.
• Gut plus liver inhibition: AUC rises and half-life lengthens.
• Enterocytes are replaced on a timescale of days, which contributes to the recovery period after a mechanism-based gut inhibitor.
• Enterocyte CYP3A4 and apical P-glycoprotein sit in the same cell and act on overlapping substrates, so inhibiting both at once compounds the effect.

Sources: Paine MF 2006 · Thummel KE 1996 · Lown KS 1997 · Lundahl J 1995 · Bailey DG 2013 · Greenblatt DJ 2003

Drug inhibitors and inducers, with FDA potency classes human data

The FDA potency classes are defined by what an agent does to the AUC of a sensitive index substrate, which for CYP3A4 means oral midazolam. A strong inhibitor raises that AUC five-fold or more; a moderate inhibitor raises it at least two-fold but less than five-fold; a weak inhibitor raises it at least 1.25-fold but less than two-fold. The mirror classes for induction are a strong inducer decreasing AUC by 80 percent or more, moderate by 50 to under 80 percent, and weak by 20 to under 50 percent. Read those definitions as descriptions of the inhibitor measured against a high-fm substrate, not as a promise about a particular drug.

Sources: U.S. Food 2023 · U.S. Food 2020* · Flockhart DA 2021 · Orr STM 2012 · Roden DM 2004

Botanical, dietary and cannabinoid inhibitors contested in vitro

The botanical side of CYP3A4 is where this library earns its place, because it is exactly the part the drug-interaction databases leave empty. Grapefruit is the best-characterised food interaction in pharmacology and the mechanism is settled: furanocoumarins, principally bergamottin and 6,7-dihydroxybergamottin, are converted by CYP3A4 into reactive species that inactivate the enzyme, and a furanocoumarin-free grapefruit juice fails to reproduce the effect. Seville orange, pomelo and tangelo share the chemistry; sweet orange does not. Piperine from black pepper is characterised in the comparative in-vitro work as a relatively selective CYP3A4 inhibitor among common spice constituents, and it also inhibits P-glycoprotein, which is the pairing that produced the twenty-fold increase in curcumin bioavailability measured in human volunteers from a 20 mg piperine dose. Curcuminoids inhibit multiple CYPs plus UGT and SULT, which makes turmeric a phase-1 and phase-2 interaction at the same time. The kavalactones are the load-bearing and the most contested entry, and they are handled in their own section below.

Contested — caveat. The grapefruit, piperine, curcuminoid and goldenseal entries rest on human or well-replicated data. The garlic, black seed, peppermint, resveratrol and quercetin entries are in-vitro or animal signals whose magnitude at ordinary dietary intake is unestablished, and several of them enter this table from the operator corpus and forum material rather than from controlled human pharmacokinetics. They are listed rather than dropped because a reader is better served knowing that a claim exists and what its evidence grade is. The 47 percent carbamazepine figure comes from one small clinical study and should be treated as a single measurement rather than a constant.

Sources: Bailey DG 2013 · Bailey DG 1991 · Paine MF 2006 · Lown KS 1997 · Lundahl J 1995 · Greenblatt DJ 2003 · Edwards DJ 1996 · Bhardwaj RK 2002 · Volak LP 2008 · Shoba G 1998 · Pattanaik S 2009* · Bahramsoltani R 2017 · Gurley BJ 2005 · Mathews JM 2002 · Zou L 2002 · Yamaori S 2011* · Jiang R 2011 · Stout SM 2014 · Van Kush Family Research Institute (Temple Pharmacopoeia Project) 2026*

The kavalactone question, stated honestly contested in vitro

The operator corpus states that methysticin and dihydromethysticin are potent CYP inhibitors at CYP2C9, CYP2C19 and CYP3A4, and this claim is load-bearing for the kava material elsewhere in the library. The in-vitro literature supports it: kava extract and isolated kavalactones inhibit a range of human CYP isoforms in microsomal and recombinant-enzyme systems, with methysticin and dihydromethysticin among the more potent constituents, and the methylenedioxyphenyl group they carry is the structural motif associated with mechanism-based CYP inactivation across many natural products. The in-vivo human picture is not the same. In the Gurley probe study, which phenotyped CYP1A2, CYP2D6, CYP2E1 and CYP3A4/5 in volunteers before and after 28 days of a standardised kava supplement, kava produced a significant inhibition of CYP2E1 — on the order of 40 percent — and did not significantly change CYP3A4/5, CYP1A2 or CYP2D6. Both findings are real and they constrain each other rather than cancelling: potent in-vitro inhibition with no measurable in-vivo effect on a probe usually means the achieved hepatic concentration is too low, the constituent profile of the product differs from the isolated compounds tested, or the probe is not sensitive enough. What cannot be said is that kava is an established clinical CYP3A4 inhibitor in humans. Note also that the Gurley design did not include a CYP2C9 or CYP2C19 probe, so the in-vitro claims for those two isoforms are not contradicted by it — they are simply untested in vivo.

• In vitro: kavalactones inhibit multiple CYPs; methysticin and dihydromethysticin are among the more potent, and carry a methylenedioxyphenyl motif associated with mechanism-based inactivation.
• In vivo human probes: significant CYP2E1 inhibition, no significant change in CYP3A4/5, CYP1A2 or CYP2D6.
• CYP2C9 and CYP2C19 were not probed in that study, so for those two the in-vitro finding stands unconfirmed rather than contradicted.
• Product identity matters enormously: acetonic and ethanolic extracts, noble versus non-noble cultivars, and aerial plant parts versus root all differ in constituent profile, and the kava hepatotoxicity literature turns on exactly that.
• Practical consequence for anyone combining kava with a prescription: the hepatotoxicity signal and additive CNS depression are better documented than the CYP inhibition, and are the reasons to be careful even if the CYP claim is weak.

Contested — caveat. A direct in-vitro versus in-vivo conflict. The in-vitro potency is well described; the one substantial in-vivo human probe study found no significant CYP3A4/5 effect. Treat kava as an established CYP2E1 inhibitor in humans, an unresolved question at CYP3A4, and an untested question at CYP2C9 and CYP2C19 in vivo. The hepatotoxicity signal is a separate and better-documented concern.

Sources: Gurley BJ 2005 · Mathews JM 2002 · Zou L 2002 · Teschke R 2011 · Van Kush Family Research Institute (Temple Pharmacopoeia Project) 2026*

Botanical inducers: St John’s wort and the PXR route human data

St John’s wort is the most clinically consequential botanical interaction ever documented, and it is an induction interaction, which is why it kills quietly. Hyperforin is a potent ligand for the pregnane X receptor, and PXR activation raises the transcription of CYP3A4 and of ABCB1, the gene for P-glycoprotein, at the same time — two mechanisms pointing in the same direction on an overlapping substrate set. The documented consequences are not theoretical. Acute heart transplant rejection has been reported from a collapse in ciclosporin concentrations after a patient began the herb. Indinavir AUC fell by a median of about 57 percent in healthy volunteers. Breakthrough bleeding and contraceptive failure, reduced digoxin exposure, and loss of anticoagulant control are all in the pharmacovigilance record. Because the mechanism is transcriptional, onset takes one to two weeks and offset takes a similar period, so the dangerous moments are both starting the herb and stopping it — stopping restores clearance to baseline and can then overshoot the dose that was titrated upward while the induction was in place.

• Mechanism: hyperforin activates PXR, inducing CYP3A4 and ABCB1 together.
• Hyperforin content varies by product and by extraction, so the size of the induction varies by product; low-hyperforin preparations induce less.
• Onset one to two weeks, offset one to two weeks. Both the start and the stop are the interaction.
• Other botanical inducers worth naming: Schisandra and some Ginkgo preparations carry PXR signals of weaker and less consistent grade.

Sources: Moore LB 2000 · Ruschitzka F 2000 · Piscitelli SC 2000 · Henderson L 2002 · Johne A 1999

Clinical consequence, concretely human data

What a CYP3A4 interaction looks like in a person depends entirely on which substrate is involved, and the point of listing them concretely is that none of these presentations announces itself as a drug interaction. Simvastatin plus a strong inhibitor: muscle pain and weakness over days, a rising creatine kinase, dark urine, and in the worst case rhabdomyolysis with acute kidney injury. Ciclosporin or tacrolimus plus an inhibitor: rising creatinine, tremor, hypertension — nephrotoxicity that looks like rejection and is treated in the opposite direction. The same drugs plus an inducer: a normal-looking patient whose graft is being rejected. Midazolam or triazolam plus an inhibitor: prolonged sedation, a fall, a respiratory event in someone with sleep apnoea. Ergotamine plus clarithromycin: cold, painful, ischaemic extremities. A direct oral anticoagulant plus a dual CYP3A4 and P-glycoprotein inhibitor: bleeding. An oral contraceptive plus an inducer: breakthrough bleeding, then pregnancy. A kinase inhibitor plus an inducer: a cancer treated at a fraction of the intended exposure, with the failure attributed to the disease. Each of these is a documented pattern and each of them has been missed repeatedly, which is the argument for a legible mechanism table rather than a memory test.

Sources: Lilja JJ 1998 · Bailey DG 2013 · Ruschitzka F 2000 · Flockhart DA 2021 · U.S. Food 2023 · Henderson L 2002

Polymorphism: CYP3A4 is quiet, CYP3A5 is not contested human data

Unlike CYP2D6, CYP2C9 and CYP2C19, CYP3A4 has no common null allele, and the large interindividual variation in CYP3A4 activity — several-fold between healthy people — is driven mostly by expression differences, inflammation, hormonal state, diet and concurrent medication rather than by coding polymorphism. The intronic CYP3A4 star-22 variant does lower hepatic expression and has been associated with statin response and with dose requirements for some CYP3A4 substrates. The more consequential genetic variable in this subfamily is CYP3A5: the star-3 allele carries a splice defect that abolishes expression and is the majority allele in European-ancestry populations, whereas functional expressor alleles are much more common in African-ancestry populations. Because CYP3A5 shares a large part of the CYP3A4 substrate space, an expressor carries additional CYP3A capacity — which is why CYP3A5 genotype is one of the better-validated predictors of tacrolimus dose requirement, with expressors typically needing substantially more drug to reach the same trough.

Contested — caveat. Allele frequencies here are reported in the literature by broad ancestral grouping, which is a poor proxy for any individual. CYP3A4 star-22 associations are replicated for some substrates and not others, and the effect sizes are modest compared with the CYP2D6 and CYP2C19 polymorphisms.

Sources: Kuehl P 2001 · Wang D 2011 · Zanger UM 2013

Cannabinoids on this enzyme, specifically contested in vitro

Both major phytocannabinoids interact with CYP3A4 from both sides, which is the detail that gets lost when cannabis is described simply as a CYP3A4 inhibitor or simply as a substrate. Δ9-THC is oxidised by CYP2C9 and CYP3A4, with CYP2C9 dominant for the 11-hydroxylation that produces the more potent 11-hydroxy-THC and CYP3A4 contributing other oxidations; cannabinol and the other classical cannabinoids are handled by an overlapping set. Cannabidiol is oxidised principally by CYP2C19 and CYP3A4 and is also an inhibitor of CYP3A isoforms in vitro, with the free phenolic hydroxyls of its resorcinol ring implicated in that inhibition. Cannabinoid metabolites, not only the parent compounds, have themselves been characterised as CYP inhibitors, which means the interaction potential of a cannabis exposure is not fully described by the parent cannabinoid concentration. The practical consequences are that a strong CYP3A4 inhibitor raises cannabinoid exposure, a strong inducer lowers it — which matters most for a titrated pharmaceutical cannabinoid rather than for a recreational one — and that a high-dose oral CBD product is itself an interacting substance in a medication list, not a neutral supplement.

• THC: CYP2C9 dominant, CYP3A4 contributory. Route matters — oral dosing generates far more 11-hydroxy-THC on first pass than inhalation.
• CBD: CYP2C19 and CYP3A4 principal oxidative routes, then extensive glucuronidation.
• CBD is an inhibitor as well as a substrate at CYP3A, so a single product can raise the level of a co-administered CYP3A4 drug.
• Cannabinoid metabolites have their own documented CYP-inhibitory activity.
• High-dose pharmaceutical CBD is the exposure where these effects have actually been measured in humans; casual low-dose supplement use is a much weaker and largely unquantified case.

Contested — caveat. Most of the cannabinoid CYP inhibition data are in-vitro IC50 determinations in microsomes or recombinant enzymes. The clinical translation is established for high-dose pharmaceutical cannabidiol (the clobazam case) and is largely unquantified for ordinary consumer CBD and for inhaled cannabis, where systemic cannabinoid concentrations and the hepatic exposure they produce differ greatly.

Sources: Watanabe K 2007 · Jiang R 2011 · Yamaori S 2011* · Stout SM 2014 · Nasrin S 2021 · Huestis MA 2007 · Sachse-Seeboth C 2009