Drug metabolism transforms lipophilic compounds into more polar, water-soluble products that can be excreted by the kidneys or in bile. Phase I reactions introduce or unmask a functional group on the drug molecule, producing metabolites that may be pharmacologically active, inactive, or toxic. The cytochrome P450 enzyme family is responsible for the majority of phase I oxidative metabolism and is the central determinant of most clinically important drug-drug metabolic interactions.
Phase I reactions encompass three broad categories: oxidation, reduction, and hydrolysis. Oxidative reactions are by far the most common and are catalyzed predominantly by the cytochrome P450 (CYP) superfamily of heme-containing monooxygenases located primarily in the endoplasmic reticulum of hepatocytes, with significant additional expression in enterocytes of the small intestinal wall, and lesser expression in the lungs, kidneys, adrenal glands, and brain. The CYP reaction cycle involves binding of the substrate to the enzyme active site, reduction of the heme iron from Fe3+ to Fe2+ by electrons donated from reduced nicotinamide adenine dinucleotide phosphate (NADPH) via NADPH-cytochrome P450 (CYP) reductase, binding of molecular oxygen, and insertion of one oxygen atom into the substrate (typically forming a hydroxyl group) while the other oxygen is reduced to water. Non-CYP phase I oxidative enzymes include monoamine oxidases (MAO-A and MAO-B), flavin-containing monooxygenases (FMO), alcohol dehydrogenase, aldehyde dehydrogenase, and xanthine oxidase, each with specific substrate profiles and clinical relevance.1
The Major CYP Isoforms. Five cytochrome P450 isoforms account for the metabolism of approximately 80 to 90% of all marketed drugs. The CYP3A4 (cytochrome P450 3A4) isoform is the most abundant hepatic and intestinal CYP enzyme and metabolizes approximately 50% of clinically used drugs; its substrates include statins (simvastatin, atorvastatin), calcineurin inhibitors (cyclosporine, tacrolimus), benzodiazepines (midazolam, triazolam), most human immunodeficiency virus (HIV) protease inhibitors, many calcium channel blockers, and numerous antineoplastic agents. The CYP2D6 (cytochrome P450 2D6) isoform metabolizes approximately 25% of drugs including opioids (codeine, tramadol, oxycodone), most tricyclic antidepressants (TCAs), selective serotonin reuptake inhibitors (SSRIs; especially fluoxetine, paroxetine), most antipsychotics, beta-blockers (metoprolol, carvedilol), and tamoxifen. The CYP2C9 (cytochrome P450 2C9) isoform metabolizes warfarin (specifically the more potent S-enantiomer), most non-steroidal anti-inflammatory drugs (NSAIDs), phenytoin, glipizide, and losartan. The CYP2C19 (cytochrome P450 2C19) isoform metabolizes proton pump inhibitors (PPIs; omeprazole, lansoprazole), clopidogrel (prodrug activation), diazepam, mephenytoin, and certain antidepressants. The CYP1A2 (cytochrome P450 1A2) isoform metabolizes theophylline, caffeine, clozapine, olanzapine, duloxetine, and is the primary target for inhibition by fluvoxamine and for induction by tobacco smoke.12
Phase I Reduction and Hydrolysis. Reductive phase I reactions, catalyzed by CYP enzymes under hypoxic conditions, nitroreductases, and carbonyl reductases, convert nitro groups to amines (e.g., chloramphenicol), azo bonds to amines (e.g., prontosil to sulfanilamide), and ketones to alcohols. Hydrolytic phase I reactions cleave ester, amide, and peptide bonds, converting prodrugs to active agents: aspirin is hydrolyzed to salicylic acid by plasma and tissue esterases; procaine and other ester-type local anesthetics are hydrolyzed by plasma pseudocholinesterase (butyrylcholinesterase, BChE); heroin is sequentially hydrolyzed to 6-monoacetylmorphine and then morphine in plasma and tissues; and angiotensin-converting enzyme (ACE) inhibitor prodrugs such as enalapril are hydrolyzed by hepatic esterases to the active enalaprilat form. The clinical significance of hydrolytic metabolism is that BChE-deficient patients fail to metabolize succinylcholine normally, producing prolonged neuromuscular blockade (discussed further in the Neuromuscular Blockade chapter).1
Products of Phase I Metabolism. Phase I metabolites may be pharmacologically active, pharmacologically inactive, or toxic. Active metabolites contribute to the therapeutic effect and must be accounted for in dosing: the CYP2D6-mediated N-demethylation of codeine produces morphine, which mediates the analgesic effect; N-desmethyl-clopidogrel is an active antiplatelet metabolite formed by CYP2C19; and the hydroxylation of tamoxifen by CYP2D6 produces endoxifen, the most active anti-estrogenic metabolite. Inactive metabolites are simply the expected outcome of detoxification. Toxic metabolites are clinically important: the CYP2E1 (cytochrome P450 2E1) isoform-mediated oxidation of acetaminophen (paracetamol) produces N-acetyl-p-benzoquinone imine (NAPQI), a reactive electrophile that covalently binds hepatic proteins and causes hepatocellular necrosis when glutathione reserves are depleted; the hepatotoxicity of halothane involves CYP2E1-mediated formation of trifluoroacetyl chloride, which haptenizes hepatic proteins and triggers immune-mediated liver injury; and the nephrotoxic metabolites of methoxyflurane arise from fluoride ion release during CYP2E1-mediated defluorination.2,3
CYP3A4 (~50% of drugs): statins, calcineurin inhibitors, benzodiazepines, most human immunodeficiency virus (HIV) protease inhibitors, calcium channel blockers. CYP2D6 (~25%): codeine, tramadol, TCAs, SSRIs (fluoxetine, paroxetine), most antipsychotics, metoprolol, tamoxifen. CYP2C9: warfarin (S-enantiomer), NSAIDs, phenytoin, glipizide, losartan. CYP2C19: PPIs, clopidogrel (activation), diazepam, some antidepressants. CYP1A2: theophylline, clozapine, olanzapine, duloxetine; induced by tobacco smoke; inhibited by fluvoxamine.
Phase II reactions attach a polar endogenous molecule to a drug or its phase I metabolite, producing a conjugate that is generally pharmacologically inactive and sufficiently water-soluble for renal or biliary excretion. Phase II reactions do not universally follow phase I; many drugs with pre-existing functional groups undergo direct phase II conjugation without prior functionalization.
Glucuronidation. Glucuronidation, catalyzed by uridine diphosphate glucuronosyltransferases (UGTs) using uridine diphosphate glucuronic acid (UDPGA) as the glucuronide donor, is quantitatively the most important phase II reaction. These UGT (glucuronosyltransferase) enzymes are expressed in the liver, primarily the glucuronosyltransferase-1A (UGT1A) subfamily and the glucuronosyltransferase-2B (UGT2B) subfamily, intestine, kidney, and several other tissues. Glucuronidation attaches glucuronic acid to hydroxyl, carboxyl, amino, or sulfhydryl groups on the substrate, producing a highly polar, ionized glucuronide conjugate at physiological pH. Clinically important glucuronosyltransferase (UGT) substrates include morphine (the UGT2B7 (glucuronosyltransferase 2B7) isoform produces both the active morphine-6-glucuronide and the pro-excitatory morphine-3-glucuronide), buprenorphine, acetaminophen (the UGT1A6 (glucuronosyltransferase 1A6) and UGT1A9 (glucuronosyltransferase 1A9) isoforms produce the non-toxic glucuronide conjugate that accounts for approximately 55% of the therapeutic dose; this pathway is overwhelmed in overdose), valproic acid (UGT1A6 and UGT2B7), lamotrigine (the UGT1A4 (glucuronosyltransferase 1A4) isoform; induction by carbamazepine and phenytoin dramatically increases lamotrigine glucuronidation), and irinotecan (active metabolite 7-ethyl-10-hydroxycamptothecin (SN-38) is glucuronidated by the UGT1A1 (glucuronosyltransferase 1A1) isoform; the UGT1A1*28 polymorphism reduces glucuronidation and increases SN-38 exposure, predisposing to severe neutropenia and diarrhea).14
Sulfation. Sulfotransferases (SULTs) catalyze transfer of a sulfonate group from 3-phosphoadenosine-5-phosphosulfate (PAPS) to hydroxyl or amino groups on substrates including acetaminophen (accounting for approximately 30% of therapeutic dose), minoxidil (sulfation by sulfotransferase 1A1 (SULT1A1) produces the pharmacologically active minoxidil sulfate that opens potassium channels), dopamine and other catecholamines, thyroid hormones (triiodothyronine (T3) and thyroxine (T4) undergo sulfation as part of their inactivation), and numerous steroid hormones. Sulfation is a high-affinity but low-capacity pathway; at high substrate concentrations (as in acetaminophen overdose), sulfation saturates and a greater fraction is diverted to CYP2E1 (cytochrome P450 2E1) isoform-mediated formation of the toxic N-acetyl-p-benzoquinone imine (NAPQI) metabolite, contributing to hepatotoxicity.3
Acetylation. N-acetyltransferases (NAT1 and NAT2) transfer an acetyl group from acetyl coenzyme A (acetyl-CoA) to amino, hydroxyl, or sulfhydryl groups. NAT2 is the primary enzyme for acetylation of drugs including isoniazid (INH), hydralazine, procainamide, sulfasalazine, and some sulfonamides. NAT2 is highly polymorphic, with slow and fast acetylator phenotypes that have significant clinical consequences. Slow acetylators accumulate higher plasma concentrations of isoniazid, leading to greater risk of peripheral neuropathy (requiring pyridoxine supplementation) and drug-induced lupus erythematosus with hydralazine and procainamide, but also greater antimycobacterial activity per dose. Fast acetylators may require higher isoniazid doses for adequate activity and are at greater risk of isoniazid-related hepatotoxicity because rapid acetylation produces the hepatotoxic hydrazine metabolite more rapidly. The acetylator phenotype does not affect drugs metabolized by NAT1, which is not polymorphic to a clinically significant degree.4
Glutathione Conjugation. Glutathione S-transferases (GSTs) catalyze the conjugation of reduced glutathione (GSH) to electrophilic substrates, forming mercapturic acid derivatives that are excreted in urine. Glutathione conjugation is primarily a detoxification mechanism that protects cellular macromolecules from electrophilic attack. The most clinically important application is acetaminophen hepatotoxicity: N-acetyl-p-benzoquinone imine (NAPQI) produced by CYP2E1 and the CYP3A4 (cytochrome P450 3A4) isoform is normally detoxified by glutathione conjugation, but when glutathione stores are depleted (as in overdose, fasting, or chronic alcohol use that induces CYP2E1), NAPQI accumulates and binds covalently to hepatic proteins, causing Zone 3 centrilobular necrosis. N-acetylcysteine (NAC) treatment works by replenishing glutathione precursors (cysteine), restoring glutathione-mediated detoxification capacity. Other clinically important glutathione substrates include reactive metabolites of platinum chemotherapeutic agents and certain alkylating agents, where glutathione conjugation contributes to tumor drug resistance.1
At therapeutic doses: ~55% glucuronidation (UGT; non-toxic), ~30% sulfation (SULT; non-toxic), ~10% CYP2E1/CYP3A4 oxidation to NAPQI (detoxified by glutathione). In overdose: glucuronidation and sulfation saturate; CYP2E1 fraction increases; glutathione depleted; NAPQI accumulates; Zone 3 hepatic necrosis. Risk amplified by: chronic alcohol (CYP2E1 induction + glutathione depletion), fasting (glutathione depletion), and CYP2E1 inducers. NAC treatment: replenishes cysteine for glutathione synthesis; effective within 8–10 hours of ingestion.
Hepatic clearance is the most important determinant of drug elimination for the majority of small-molecule drugs. Understanding what governs the rate of hepatic drug removal, and how that relationship changes with disease, blood flow, and enzyme induction or inhibition, is essential for anticipating when standard dosing will produce unexpected drug exposures.
The hepatic clearance of a drug is described by the well-stirred (venous equilibrium) model: hepatic clearance (CLH) = QH × EH, where QH is hepatic blood flow (approximately 1.5 L/min at rest) and EH is the hepatic extraction ratio, the fraction of drug removed from blood during a single pass through the liver. EH in turn is determined by the intrinsic clearance (CLint), which reflects the maximal metabolic capacity of the liver in the absence of blood flow or protein binding constraints, and by plasma protein binding: EH = (fu × CLint) / (QH + fu × CLint), where fu is the unbound fraction of drug in blood. This relationship reveals that two distinct types of drugs exist with respect to hepatic clearance, and the clinical implications of enzyme induction, inhibition, and hepatic blood flow changes differ substantially between them.5
High-Extraction (Flow-Limited) Drugs. For high-extraction drugs (EH greater than 0.7), fu × CLint greatly exceeds QH, and the extraction ratio approaches 1.0. In this limit, hepatic clearance is determined almost entirely by hepatic blood flow: CLH ≈ QH. The clinical consequences are important. First, enzyme induction or inhibition has relatively little effect on the clearance or bioavailability of high-extraction drugs administered intravenously, because clearance is already flow-limited and not enzyme-limited; even if CLint doubles through enzyme induction, EH cannot exceed 1.0 and CLH cannot exceed QH. Second, for oral dosing of high-extraction drugs, first-pass extraction is the dominant determinant of bioavailability, and changes that alter hepatic blood flow or portal shunting (cirrhosis, heart failure, co-administration of vasodilators) substantially change oral bioavailability. Third, changes in plasma protein binding have minimal effect on hepatic clearance of high-extraction drugs because the unbound fraction is already efficiently extracted and total clearance is flow-limited. Examples include lidocaine, propranolol, morphine, verapamil, labetalol, and nitroglycerin.56
Low-Extraction (Capacity-Limited) Drugs. For low-extraction drugs (EH less than 0.3), fu × CLint is much smaller than QH, and hepatic clearance simplifies to: CLH ≈ fu × CLint. Hepatic clearance is now determined by the intrinsic metabolic capacity and the free drug fraction, not by blood flow. The clinical consequences also differ from high-extraction drugs. Enzyme induction increases CLint, directly increasing CLH and reducing drug exposure. Enzyme inhibition decreases CLint, directly reducing CLH and increasing drug exposure. These interactions are pharmacokinetically significant and can require dose adjustments. Changes in hepatic blood flow have minimal effect on clearance because QH greatly exceeds the metabolic capacity and is not limiting. However, changes in protein binding do affect clearance: for low-extraction drugs, an increase in fu (as in hypoalbuminemia) increases CLH, tending to restore free drug concentrations toward baseline even as total concentrations fall. Clinically important low-extraction drugs include warfarin, phenytoin, theophylline, diazepam, and valproic acid.56
Hepatic Disease and Drug Clearance. Liver disease reduces drug metabolism through three mechanisms: reduction in the number and activity of hepatic cytochrome P450 (CYP) and phase II enzymes as functional hepatocyte mass is lost, reduction in hepatic blood flow in cirrhosis due to portal hypertension and formation of intrahepatic shunts, and reduction in plasma albumin increasing the free fraction of bound drugs. For high-extraction drugs, cirrhosis reduces both hepatic blood flow and first-pass extraction, substantially increasing systemic bioavailability after oral dosing and reducing intravenous clearance. For low-extraction drugs, the reduction in CYP enzyme activity reduces CLint and thereby reduces hepatic clearance, prolonging half-life and increasing drug accumulation at standard doses. The Child-Pugh score is the most widely used clinical tool for estimating hepatic functional reserve and guiding dose adjustments, though it correlates imperfectly with specific enzyme activity. As a general principle, drugs with narrow therapeutic indices that undergo extensive hepatic metabolism require dose reduction and more frequent monitoring in patients with Child-Pugh B or C cirrhosis.6
High-extraction drugs (lidocaine, propranolol, morphine): clearance governed by hepatic blood flow. IV enzyme interactions minor; oral bioavailability sensitive to blood flow and portal shunting. Cirrhosis: increase oral bioavailability substantially. Low-extraction drugs (warfarin, phenytoin, theophylline): clearance governed by enzyme activity and free fraction. Enzyme inducers/inhibitors produce significant changes in clearance and require dose adjustment. Protein binding changes affect free drug; monitor free levels in hypoalbuminemia.
Enzyme induction and inhibition represent the two mechanisms by which co-administered drugs alter the metabolic clearance of cytochrome P450 (CYP) enzyme substrates. Their clinical consequences range from therapeutic failure through under-exposure to serious toxicity through over-exposure, and they are among the most common preventable causes of adverse drug events in polypharmacy settings.
Enzyme Induction. Enzyme induction is an increase in CYP enzyme protein content resulting from increased gene transcription. Most clinically important CYP inducers activate nuclear receptors: the pregnane X receptor (PXR) controls the CYP3A4 (cytochrome P450 3A4) and CYP2C9 (cytochrome P450 2C9) isoforms for induction and is activated by rifampicin, carbamazepine, phenytoin, phenobarbital, St. John's Wort, and glucocorticoids; the constitutive androstane receptor (CAR) controls the CYP2B6 (cytochrome P450 2B6) isoform and CYP3A4 induction; the aryl hydrocarbon receptor (AhR) controls the CYP1A2 (cytochrome P450 1A2) isoform induction and is activated by polycyclic aromatic hydrocarbons in tobacco smoke, charbroiled meat, and omeprazole. Because induction requires new protein synthesis, its onset is delayed: CYP enzyme levels begin increasing within 24 to 48 hours of inducer exposure but maximum induction requires one to two weeks of continuous exposure. Similarly, when an inducer is discontinued, enzyme levels return to baseline over one to two weeks as the newly synthesized enzyme is turned over. This time course has direct clinical implications: the drug interaction risk peaks after one to two weeks of co-administration and persists for one to two weeks after the inducer is stopped, a period during which toxicity from the previously induced substrate may emerge if dose adjustments were made during the induction period.7
Clinically Important Inducers and Their Consequences. Rifampicin is the most potent CYP3A4 inducer in clinical use and can reduce the plasma concentrations of CYP3A4 substrates by 80 to 90%. This interaction is clinically catastrophic for calcineurin inhibitors (transplant rejection due to sub-therapeutic cyclosporine or tacrolimus levels), oral contraceptives (contraceptive failure), direct-acting antivirals for hepatitis C, and many antiretrovirals. Carbamazepine induces its own metabolism (autoinduction) as well as that of many co-administered drugs, requiring dose escalation over the first weeks of therapy. Phenytoin and phenobarbital are potent inducers; their use in combination with warfarin requires careful international normalized ratio (INR) monitoring during initiation and discontinuation. St. John's Wort (hyperforin component) induces CYP3A4, CYP2C9, and P-glycoprotein; co-administration with cyclosporine has caused transplant rejection, and its use with warfarin, oral contraceptives, and antiretrovirals requires avoidance or very careful monitoring. Tobacco smoke induces CYP1A2; smokers may require higher doses of clozapine, olanzapine, and theophylline, and dose reduction upon smoking cessation must be anticipated to avoid toxicity.27
Enzyme Inhibition: Competitive and Reversible. Competitive inhibition occurs when an inhibitor drug occupies the CYP active site and competes with the substrate for binding. The degree of inhibition depends on the inhibitor concentration relative to its inhibitory constant (Ki), and inhibition is reversible and immediate in onset: it begins with the first dose of the inhibitor and resolves rapidly when the inhibitor is eliminated. The clinical magnitude of a competitive inhibition interaction depends on the inhibitor Ki, the inhibitor plasma concentration (which in turn depends on the inhibitor's own pharmacokinetics), and the contribution of the inhibited enzyme to the substrate's total clearance.7
Ketoconazole and itraconazole are among the most potent competitive CYP3A4 inhibitors; co-administration with simvastatin or lovastatin (but not pravastatin, which is not a CYP3A4 substrate) is contraindicated due to the risk of severe myopathy and rhabdomyolysis from a greater than ten-fold increase in statin exposure. Fluconazole is a potent CYP2C9 inhibitor; co-administration with warfarin requires international normalized ratio (INR) monitoring and typically a 25 to 50% warfarin dose reduction.78
Mechanism-Based (Irreversible) Inhibition. Mechanism-based inhibitors, also termed suicide inhibitors or time-dependent inhibitors (TDIs), are metabolized by the CYP enzyme to a reactive intermediate that covalently modifies the enzyme active site, permanently inactivating it. Recovery requires synthesis of new enzyme protein, which takes days to weeks depending on the enzyme. This has several important clinical implications. First, the inhibitory effect persists long after the inhibitor is cleared: erythromycin, clarithromycin, and diltiazem are mechanism-based CYP3A4 inhibitors with half-lives of several hours, but their inhibitory effect on CYP3A4 persists for several days after discontinuation. Second, the inhibition develops progressively over the first several days of dosing as a fraction of the total enzyme pool is inactivated with each dose. Third, mechanism-based inhibitors are particularly dangerous in the peri-operative period: stopping clarithromycin before surgery does not immediately restore CYP3A4 activity, and drug interactions with anesthetic agents metabolized by CYP3A4 may persist. Clinically important mechanism-based CYP inhibitors include erythromycin and clarithromycin (CYP3A4), verapamil and diltiazem (CYP3A4), and gestodene (CYP3A4).7
Fluoxetine and its active metabolite norfluoxetine inhibit the CYP2D6 (cytochrome P450 2D6) isoform through a combined competitive and mechanism-based mechanism. Because norfluoxetine has a half-life of approximately one to two weeks, CYP2D6 inhibition persists for weeks after fluoxetine discontinuation, requiring a five-week washout period before initiating monoamine oxidase inhibitors (MAOIs) or switching to drugs whose dose depends on CYP2D6 activity.7
Rifampicin + calcineurin inhibitors: avoid; if unavoidable, expect 3–5-fold dose increase and TDM daily. Rifampicin + oral contraceptives: switch to non-hormonal contraception for duration plus 4 weeks after stopping. Clarithromycin/erythromycin + simvastatin/lovastatin: contraindicated; switch to pravastatin or rosuvastatin. Fluconazole + warfarin: reduce warfarin ~25–50%; check INR in 3–5 days. Fluoxetine discontinuation: 5-week washout before MAOIs or dose-sensitive CYP2D6 substrates. St. John's Wort + cyclosporine: avoid; cases of acute transplant rejection documented.
Genetic variation in drug-metabolizing enzymes is one of the most important sources of interpatient variability in drug response. For enzymes with well-characterized functional polymorphisms, genotyping can predict which patients will experience unexpected toxicity or therapeutic failure at standard doses, enabling prospective dose individualization before adverse events occur.
Pharmacogenomic variation in cytochrome P450 (CYP) enzymes is classified by metabolizer phenotype, which describes the functional consequence of an individual's combined genotype (diplotype). Poor metabolizers (PMs) carry two non-functional alleles and have no active enzyme activity; they accumulate parent drug and may have minimal or absent formation of active metabolites. Intermediate metabolizers (IMs) carry one functional and one non-functional allele (or two reduced-function alleles) and have reduced but present enzyme activity. Extensive metabolizers (EMs), the most common phenotype in most populations, carry two functional alleles and represent the normal population standard. Ultrarapid metabolizers (UMs) carry gene duplications or multiplications, producing excessive enzyme activity and very rapid metabolism of substrates; they may fail to achieve therapeutic drug concentrations at standard doses.9
CYP2D6 Polymorphism. The CYP2D6 (cytochrome P450 2D6) isoform is among the most extensively studied pharmacogenomic targets because it metabolizes approximately 25% of drugs and is highly polymorphic, with more than 100 allelic variants identified. Approximately 7 to 10% of European populations are poor metabolizers due to homozygous non-functional alleles (*3, *4, *5, *6); poor metabolizer frequency is lower in African (approximately 2%) and Asian (approximately 1%) populations. Ultrarapid metabolizers carry CYP2D6 gene duplications and constitute approximately 1 to 2% of Northern European populations and up to 5 to 10% of North African and Middle Eastern populations. The clinical consequences of CYP2D6 poor metabolizer status depend on whether the drug requires CYP2D6 for activation or inactivation. For prodrugs activated by CYP2D6 (codeine, tramadol, oxycodone activation from oxymorphone formation): poor metabolizers have reduced or absent analgesic response because the active metabolite is not formed; ultrarapid metabolizers may have excessive opioid effects. For drugs inactivated by CYP2D6 (most TCAs, metoprolol, risperidone): poor metabolizers accumulate higher parent drug concentrations, increasing the risk of dose-related adverse effects at standard doses. Tamoxifen's conversion to the active metabolite endoxifen requires CYP2D6; poor metabolizers and patients on CYP2D6 inhibitors (paroxetine, fluoxetine) have substantially reduced endoxifen levels and may have inferior breast cancer outcomes.912
CYP2C19 Polymorphism. The CYP2C19 (cytochrome P450 2C19) isoform is highly polymorphic, with the CYP2C19*2 and CYP2C19*17 alleles being the most clinically important. The *2 allele creates a splice defect resulting in a non-functional enzyme; homozygous *2/*2 individuals are poor metabolizers, constituting approximately 2 to 5% of European and African populations but 15 to 20% of East Asian populations (a frequency difference with significant implications for drug dosing guidelines across ethnic groups). The *17 allele contains a promoter variant that increases transcription, producing the ultrarapid metabolizer phenotype. For proton pump inhibitors (PPIs), poor metabolizers have substantially higher plasma exposures than extensive metabolizers at the same dose and achieve superior acid suppression; ultrarapid metabolizers may require higher PPI (proton pump inhibitor) doses or twice-daily dosing for adequate acid control, which is clinically relevant in Helicobacter pylori eradication regimens. For clopidogrel, CYP2C19 is the key activation enzyme; poor metabolizers generate substantially less of the active thiol metabolite, have reduced platelet inhibition, and have higher rates of major adverse cardiovascular events after percutaneous coronary intervention (PCI). The US Food and Drug Administration (FDA) black-box warning on clopidogrel recommends considering alternative antiplatelet agents (prasugrel, ticagrelor) in CYP2C19 poor metabolizers undergoing PCI.11
CYP2C9 Polymorphism. CYP2C9 (cytochrome P450 2C9) polymorphisms are clinically significant primarily for warfarin and phenytoin, two narrow-therapeutic-index drugs where dose individualization is essential. The CYP2C9*2 allele (Arg144Cys) produces an enzyme with approximately 30 to 40% of wild-type activity; the CYP2C9*3 allele (Ile359Leu) produces an enzyme with approximately 5% of wild-type activity. Patients who are CYP2C9*3/*3 homozygotes or *2/*3 compound heterozygotes have markedly reduced warfarin (S-enantiomer) clearance and require substantially lower warfarin doses to maintain therapeutic anticoagulation, with the risk of life-threatening bleeding at standard doses. The FDA-approved warfarin labeling includes CYP2C9 genotype-based dosing tables, and clinical algorithms integrating CYP2C9 and VKORC1 (vitamin K epoxide reductase complex subunit 1) genotypes with clinical variables predict warfarin dose requirements more accurately than clinical variables alone.211
Non-CYP Pharmacogenomics. Pharmacogenomic dose individualization extends beyond CYP enzymes to include thiopurine S-methyltransferase (TPMT) and dihydropyrimidine dehydrogenase (DPYD), two metabolic enzymes with clinically important polymorphisms. TPMT catalyzes the inactivation of thiopurine drugs (azathioprine, 6-mercaptopurine, thioguanine) to inactive methylated metabolites. Patients with TPMT poor metabolizer phenotype (approximately 0.3% of the population are homozygous poor metabolizers; 10% are heterozygotes with intermediate activity) accumulate thioguanine nucleotides, the active metabolites responsible for both the therapeutic immunosuppression and the dose-limiting myelosuppression. Standard doses of azathioprine in TPMT poor metabolizers cause life-threatening myelosuppression; TPMT genotyping or phenotyping prior to initiating thiopurine therapy is recommended and widely practiced. DPYD is the rate-limiting enzyme for catabolism of the fluoropyrimidine drugs 5-fluorouracil (5-FU) and capecitabine; patients with DPYD poor metabolizer variants (approximately 3 to 8% of the population carry one deficient allele) are at risk for severe and potentially fatal fluoropyrimidine toxicity including mucositis, diarrhea, and myelosuppression at standard doses. European regulatory agencies now mandate DPYD genotyping before fluoropyrimidine initiation.9
CYP2D6 before codeine in children (ultrarapid metabolizers at risk for fatal respiratory depression; codeine now contraindicated under age 12 in many jurisdictions). CYP2D6 before tamoxifen in breast cancer (avoid paroxetine/fluoxetine co-prescription; consider genotyping in patients requiring antidepressant therapy). CYP2C19 before clopidogrel in high-risk PCI (consider prasugrel or ticagrelor if poor metabolizer). CYP2C9 + VKORC1 before warfarin initiation (genotype-guided dosing reduces time to stable INR). TPMT or NUDT15 before azathioprine or 6-mercaptopurine (prevent fatal myelosuppression). DPYD before fluoropyrimidines (now mandated in EU).
A prodrug is a pharmacologically inactive compound that requires metabolic conversion to yield the active drug. For prodrugs activated by cytochrome P450 (CYP) enzymes, the pharmacogenomic and drug interaction principles discussed in the preceding sections take on immediate clinical significance: patients with reduced enzyme activity may fail therapy, while those with elevated enzyme activity are at risk for toxicity from excessive active metabolite generation.
Codeine. Codeine is a weak opioid with limited intrinsic activity at mu-opioid receptors. Its analgesic effect depends primarily on O-demethylation to morphine by the CYP2D6 (cytochrome P450 2D6) isoform, a reaction accounting for approximately 10% of the administered codeine dose. Despite this small fraction, morphine has approximately 200-fold greater mu-opioid receptor affinity than codeine, making the CYP2D6-dependent conversion pharmacodynamically essential for clinically meaningful analgesia. CYP2D6 poor metabolizers produce negligible morphine from codeine and receive little or no analgesic benefit; they may be inadvertently undertreated for pain while receiving opioid-related side effects from the parent codeine and its other metabolites. CYP2D6 ultrarapid metabolizers convert codeine to morphine at an accelerated rate, producing morphine plasma concentrations several-fold higher than in extensive metabolizers at the same codeine dose. Multiple fatal cases of respiratory depression in children given post-tonsillectomy codeine by ultrarapid metabolizer mothers who excreted high morphine concentrations in breast milk led to the US FDA and European Medicines Agency (EMA) restricting codeine use in children, in nursing mothers, and in patients with known ultrarapid metabolizer status. Tramadol undergoes analogous CYP2D6-mediated O-demethylation to O-desmethyltramadol, which has substantially higher mu-opioid receptor affinity than the parent drug, and carries similar pharmacogenomic risks.910
Clopidogrel. Clopidogrel is an inactive thienopyridine prodrug that requires two sequential oxidative steps for conversion to its pharmacologically active thiol metabolite. The first step (conversion to 2-oxo-clopidogrel) is catalyzed primarily by the CYP1A2 (cytochrome P450 1A2) and CYP2C19 (cytochrome P450 2C19) isoforms; the second step (formation of the active thiol) is catalyzed predominantly by the CYP2C19, CYP3A4 (cytochrome P450 3A4), and CYP2B6 (cytochrome P450 2B6) isoforms. The active thiol metabolite irreversibly alkylates the P2Y12 purinergic receptor (P2Y12) for adenosine diphosphate (ADP) on platelets, inhibiting ADP-mediated platelet aggregation. CYP2C19 poor metabolizers generate approximately 30 to 40% less active metabolite than extensive metabolizers, resulting in measurably reduced platelet inhibition as assessed by platelet function assays. Multiple large clinical studies, including the Trial to Assess Improvement in Therapeutic Outcomes by Optimizing Platelet Inhibition with Prasugrel (TRITON-TIMI 38) and the Platelet Inhibition and Patient Outcomes trial (PLATO), demonstrated that CYP2C19 poor metabolizer status is associated with significantly higher rates of major adverse cardiovascular events (MACE) including stent thrombosis after percutaneous coronary intervention (PCI). Proton pump inhibitors, particularly omeprazole and esomeprazole (potent CYP2C19 inhibitors), reduce clopidogrel activation and were the subject of FDA advisories recommending avoidance of concomitant use; pantoprazole and rabeprazole have less CYP2C19 inhibitory activity and are preferred when proton pump inhibitor (PPI) co-prescription is required in patients on clopidogrel.11
Tamoxifen. Tamoxifen is a selective estrogen receptor modulator (SERM) used for the treatment and prevention of estrogen receptor-positive (ER+) breast cancer. Its pharmacology is complex, involving multiple metabolic pathways producing metabolites with varying degrees of anti-estrogenic activity. The two most active anti-estrogenic metabolites are 4-hydroxytamoxifen (formed by CYP2D6 and CYP3A4) and endoxifen (N-desmethyl-4-hydroxytamoxifen, formed from the major metabolite N-desmethyltamoxifen by CYP2D6). Endoxifen achieves substantially higher plasma concentrations than 4-hydroxytamoxifen at steady state and is now recognized as the primary pharmacologically active species responsible for tamoxifen's clinical benefit. CYP2D6 poor metabolizers have endoxifen plasma concentrations approximately 75% lower than extensive metabolizers, and several retrospective and prospective studies have associated CYP2D6 poor metabolizer status with inferior breast cancer outcomes including increased recurrence rates, though the evidence is not uniformly consistent across all studies. The co-prescription of potent CYP2D6 inhibitors (paroxetine, fluoxetine) with tamoxifen reduces endoxifen to levels comparable to those in poor metabolizers and should be avoided; SNRIs such as venlafaxine and the antidepressant citalopram have minimal CYP2D6 inhibitory activity and are preferred when antidepressant therapy is required during tamoxifen treatment.12
Codeine → morphine via CYP2D6: poor metabolizers have inadequate analgesia; ultrarapid metabolizers risk morphine toxicity. Codeine contraindicated in children under 12, nursing mothers, and known ultrarapid metabolizers. Clopidogrel → active thiol via CYP2C19: poor metabolizers have reduced platelet inhibition and higher MACE risk post-PCI; consider prasugrel or ticagrelor. Avoid omeprazole/esomeprazole with clopidogrel; use pantoprazole or rabeprazole. Tamoxifen → endoxifen via CYP2D6: avoid paroxetine and fluoxetine during tamoxifen therapy; use venlafaxine or citalopram for vasomotor symptoms. CYP2D6 genotyping clinically actionable for codeine, tramadol, and tamoxifen in appropriate settings.
Brunton LL, Hilal-Dandan R, Knollmann BC, eds. Goodman & Gilman's: The Pharmacological Basis of Therapeutics. 13th ed. New York: McGraw-Hill; 2018. ISBN 9781259584732.
Zanger UM, Schwab M. Cytochrome P450 enzymes in drug metabolism: regulation of gene expression, enzyme activities, and impact of genetic variation. Pharmacol Ther. 2013;138(1):103-141.
doi:10.1016/j.pharmthera.2012.12.007Hodgman MJ, Garrard AR. A review of acetaminophen poisoning. Crit Care Clin. 2012;28(4):499-516.
doi:10.1016/j.ccc.2012.07.006Relling MV, Evans WE. Pharmacogenomics in the clinic. Nature. 2015;526(7573):343-350.
doi:10.1038/nature15817Rowland M, Tozer TN. Clinical Pharmacokinetics and Pharmacodynamics: Concepts and Applications. 4th ed. Philadelphia: Lippincott Williams & Wilkins; 2011. ISBN 9780781750462.
Wilkinson GR. Drug metabolism and variability among patients in drug response. N Engl J Med. 2005;352(21):2211-2221.
doi:10.1056/NEJMra032424Cascorbi I. Drug interactions: principles, examples and clinical consequences. Dtsch Arztebl Int. 2012;109(33-34):546-556.
doi:10.3238/arztebl.2012.0546Bailey DG, Dresser G, Arnold JM. Grapefruit-medication interactions: forbidden fruit or avoidable consequences? CMAJ. 2013;185(4):309-316.
doi:10.1503/cmaj.120951Crews KR, Monte AA, Huddart R, et al. Clinical Pharmacogenetics Implementation Consortium guideline for CYP2D6, OPRM1, and COMT genotype and select opioid therapy. Clin Pharmacol Ther. 2021;110(4):888-896.
doi:10.1002/cpt.2149Relling MV, Schwab M, Whirl-Carrillo M, et al. Clinical Pharmacogenetics Implementation Consortium guideline for thiopurine dosing based on TPMT and NUDT15 genotypes. Clin Pharmacol Ther. 2019;105(5):1095-1105.
doi:10.1002/cpt.1304Mega JL, Simon T, Collet JP, et al. Reduced-function CYP2C19 genotype and risk of adverse clinical outcomes among patients treated with clopidogrel predominantly for PCI: a meta-analysis. JAMA. 2010;304(16):1821-1830.
doi:10.1001/jama.2010.1543Goetz MP, Rae JM, Suman VJ, et al. Pharmacogenetics of tamoxifen biotransformation is associated with clinical outcomes of efficacy and hot flashes. J Clin Oncol. 2005;23(36):9312-9318.
doi:10.1200/JCO.2005.03.3266