CYP2D6 and CYP2C19 pharmacogenomic testing helps in psychiatry when a patient responds unpredictably to a standard dose — no effect, dramatic side effects, or toxicity — because these two enzymes metabolise most antidepressants, antipsychotics and opioid analgesics, and common genetic variants sort patients into poor, intermediate, normal, rapid or ultrarapid metabolizer phenotypes. It is a targeted test for puzzling cases, not a routine screen before every prescription.

This article explains what the genotype actually tells you, the concrete psychiatric drug examples examiners favour, when a result should genuinely change your prescribing, and the caveats — phenoconversion and clinical titration — that keep pharmacogenomics a tool rather than a replacement for clinical judgement. It maps onto the CPIC and DPWG guidelines and onto MRCPsych, RANZCP and USMLE-style pharmacology questions.

What CYP2D6 and CYP2C19 pharmacogenomic testing measures

Both genes are highly polymorphic. Alleles are scored as no function, decreased function or normal function (for example CYP2D6 *4 and *5 carry no function; *10 carries decreased function), and CYP2D6 additionally has copy-number variation — duplicated functional alleles drive the ultrarapid phenotype that is common in parts of East Africa, the Middle East and North Africa.

The two alleles you carry are translated into an activity score and then a metabolizer phenotype:

  • Poor metabolizer (PM) — little or no enzyme activity. Parent drug accumulates: higher exposure, more adverse effects at standard doses, and for prodrugs (codeine, tamoxifen, clopidogrel) little active metabolite is formed, so the drug simply does not work.
  • Intermediate metabolizer (IM) — reduced activity; sits between poor and normal and often behaves like a poor metabolizer when a second hit (an inhibitor) is added.
  • Normal metabolizer (NM) — typical population response; standard doses behave as expected.
  • Rapid and ultrarapid metabolizer (RM/UM) — increased activity (CYP2D6 gene duplication, or highly active alleles). Prodrugs are converted too fast, risking toxicity, while rapidly cleared active drugs are eliminated before they can work.

Genotype is not the whole story. A normal genotype on fluoxetine or paroxetine behaves much like a poor metabolizer because the drug itself inhibits the enzyme — the phenomenon called phenoconversion. This is the single most common exam nuance around pharmacogenomic testing: the label on the result reflects DNA, but the patient's actual phenotype depends on their medication list too.

CYP2D6 in practice: codeine, tramadol and the tricyclics

CYP2D6 is the classic teaching gene because it handles both directions of risk:

  • Codeine and tramadol are prodrugs. CYP2D6 converts codeine to morphine and tramadol to its active metabolite. An ultrarapid metabolizer generates morphine far faster than intended — there are well-documented cases of life-threatening respiratory depression, including fatal outcomes in children and breastfeeding infants exposed via an ultrarapid-mother's milk. A poor metabolizer, by contrast, gets little analgesia at any reasonable dose.
  • Tricyclic antidepressants (nortriptyline, imipramine, desipramine, clomipramine's demethylation) are CYP2D6 substrates: poor metabolizers accumulate parent drug and can reach toxic plasma levels — anticholinergic effects, QT prolongation — on doses that a normal metabolizer tolerates easily.
  • Tamoxifen (oncology, but the standard CPIC example) requires CYP2D6 to form endoxifen; poor metabolizers have reduced exposure, which is why CYP2D6 testing appears in oncology guidance — a reminder that the same gene governs efficacy as well as toxicity.
  • Some antipsychotics and ondansetron are also CYP2D6 substrates, which is why a patient who needs unusually high doses of an antipsychotic may be a rapid metabolizer rather than "treatment resistant".

Gene-drug pairs: what a CYP2D6 or CYP2C19 result means

CYP2C19 covers the SSRI story: citalopram, escitalopram, sertraline (minor pathway) and the tricyclic clomipramine are substrates, and voriconazole and clopidogrel sit outside psychiatry as the standard non-psych examples. The table is the version worth memorising.

GeneDrugPhenotypeConsequence and typical action
CYP2D6Codeine, tramadolUltrarapidExcess morphine formation — toxicity and respiratory depression risk; avoid codeine/tramadol, use a non-prodrug analgesic.
CYP2D6Codeine, tramadolPoorLittle or no analgesia; switch to a non-prodrug (e.g. morphine, paracetamol, NSAID) rather than escalating the dose.
CYP2D6Nortriptyline, imipraminePoorHigh plasma levels, anticholinergic toxicity and QT risk; consider a lower dose or an alternative antidepressant.
CYP2C19Citalopram, escitalopramPoorHigher exposure and dose-dependent QT prolongation; citalopram dosing is capped in poor metabolizers (commonly cited as 20 mg/day max in labelling) and ECG review is prudent at higher doses.
CYP2C19Citalopram, escitalopramUltrarapidSubtherapeutic levels and apparent non-response at standard doses; consider an alternative agent or a dose increase guided by response.
CYP2C19ClomipraminePoorElevated levels; monitor ECG and consider dose reduction.
CYP2C19Voriconazole, clopidogrel (non-psychiatric)PoorVoriconazole: toxicity and dose reduction. Clopidogrel: reduced activation and loss of antiplatelet effect — the classic prodrug example.

Two structural points about the genetics: CYP2D6 poor metabolizers are roughly 5–10% of European-ancestry populations while CYP2C19 poor metabolizers reach 2–5% in Europeans but 15–20% in East Asian populations — so ancestry informs prior probability, though genotype testing gives the answer directly. The DPYD gene that must be checked before fluorouracil-based chemotherapy is the landmark example of pre-treatment pharmacogenomics, but it sits outside psychiatric prescribing and is a comparison point rather than a tool you will use.

When a pharmacogenomic result actually changes your prescribing

Guidelines (CPIC and the Dutch DPWG) publish gene–drug pairs with recommendation levels, but the result only earns its place in the notes when it answers one of three clinical questions:

  • Explaining a strange reaction at a normal dose. Toxicity on a standard dose of a CYP2D6 or CYP2C19 substrate, or clear non-response despite adequate time and adherence, is the classic trigger for testing.
  • Choosing between two reasonable options. When the genotype shows poor metabolizer status for citalopram, the practical alternatives are a dose cap with ECG review or switching to a drug outside that pathway — the result converts a coin-flip into a documented rationale.
  • Stopping the dose-creeping spiral. A patient escalated to high antipsychotic doses with no benefit may be a rapid metabolizer rather than treatment resistant; the test reframes the whole treatment history.

What testing does not do is replace titration. Genotype predicts capacity, not the effect of the last three prescriptions, smoking status, oral contraceptive use, hepatic function or the two inhibitors the patient started last month. CPIC is explicit that results are an input to — never a substitute for — clinical judgement, and a normal result does not license higher-than-needed doses.

Exam trap

The highest-yield single fact: CYP2D6 ultrarapid metabolizer + codeine means excessive morphine formation — the documented cause of fatal respiratory depression in children and in breastfed infants of an ultrarapid mother's milk. The mirror trap is phenoconversion — a genetically normal metabolizer on fluoxetine or paroxetine behaves as a poor metabolizer because the drug inhibits its own metabolism. Exam questions love this reversal: the phenotype on the report is DNA, but the phenotype at the bedside is DNA plus the medication list.

Frequently asked questions

When should pharmacogenomic testing be ordered in psychiatry?

When a patient has unexpected adverse effects or clear non-response at an adequate dose of a CYP2D6 or CYP2C19 substrate, or when choosing between agents would genuinely change with the result. It is a targeted test for puzzling cases, not a routine screen before every prescription.

What does a CYP2D6 poor metabolizer result mean for codeine?

Little or no conversion to morphine, so codeine provides weak or absent analgesia. The correct response is to switch to a non-prodrug analgesic rather than escalating the dose — and the same logic applies to tramadol.

Can a normal pharmacogenomic result still behave like poor metabolism?

Yes — through phenoconversion. CYP inhibitors such as fluoxetine, paroxetine or bupropion suppress enzyme activity so that a genetically normal metabolizer effectively functions as an intermediate or poor metabolizer. Always read the genotype together with the current drug list.

Does CYP2C19 status affect citalopram dosing?

Yes. Poor metabolizers accumulate citalopram, increasing exposure and dose-dependent QT prolongation risk, so labelling caps the dose in this group (commonly cited as 20 mg/day) and ECG review is prudent at higher doses. Ultrarapid metabolizers may show subtherapeutic levels and apparent non-response.

Key takeaways

  • CYP2D6 and CYP2C19 sort patients into poor, intermediate, normal, rapid and ultrarapid metabolizer phenotypes — each with a predictable direction of error.
  • Prodrugs fail in poor metabolizers (no effect) and poison ultrarapids (codeine toxicity); active drugs do the reverse.
  • Phenoconversion means the drug list can override the genotype — the most common exam nuance in this area.
  • Test when the clinical picture is genuinely puzzling; treat the patient, not the report.
  • CPIC levels A/B recommendations exist for the key pairs — cite them, but keep titration and monitoring in charge.

Pharmacogenomics sits naturally alongside the interaction work in Preppsy's MRCPsych pharmacology bank, where gene–drug pairs are drilled as exam-style MCQs.

Cover image: “CYP2D6 and CYP2C19 Testing (cover)” by Preppsy Editorial Team (Original artwork) via Preppsy.