Curriculum
Module 04 · 60 min

Pharmacogenomics in Practice

CYP2C19, TPMT, DPYD, HLA — when a genotype actually changes prescribing.

FoundationsClinicalResearch
Core topics

What's covered

  • T01CPIC guideline tier system; CPIC, DPWG, FDA Table of Pharmacogenomic Biomarkers
  • T02CYP2D6 + CYP2C19: SSRIs, opioids, clopidogrel, voriconazole
  • T03DPYD before 5-FU/capecitabine — EU-mandated; US uptake patchy
  • T04TPMT and NUDT15 before thiopurines
  • T05HLA-B*57:01 (abacavir), HLA-B*15:02 (carbamazepine), HLA-A*31:01
  • T06Star-allele nomenclature, structural variation in CYP2D6, copy-number callers (Stargazer, Cyrius)
Learning objectives

By the end of this module you will be able to

  • L01List the pharmacogenes with CPIC Level A guidelines and at least one drug each.
  • L02Explain why CYP2D6 genotyping is technically harder than other PGx genes.
  • L03Describe the rationale and evidence for pre-emptive vs reactive PGx testing.
Key takeaways

What you should walk away believing

  • CYP2D6 has >150 star alleles, copy-number variation, and hybrid alleles with CYP2D7 — short-read panels routinely miscall it.
  • DPYD pre-testing reduces severe fluoropyrimidine toxicity by ~50% (Henricks 2018); EMA mandates it, FDA does not.
  • Pre-emptive PGx (panel at primary-care entry) is more cost-effective than reactive testing for any patient likely to receive ≥2 PGx-actionable drugs.
  • HLA-B*57:01 testing before abacavir is one of the best-validated PGx interventions in medicine — NPV essentially 100% for hypersensitivity.
Lesson · Foundations emphasis

What this means at your level

Foundations

Pharmacogenomics uses a person's genotype to predict drug response. Some examples are routine: testing HLA-B*57:01 before abacavir, TPMT before azathioprine, DPYD before 5-fluorouracil. Others (CYP2D6 for codeine, CYP2C19 for clopidogrel) are increasingly recommended but inconsistently implemented.

Clinician deep-dive

CPIC publishes free, peer-reviewed guidelines mapping star-allele genotype to phenotype to dose recommendation. For each new prescription, check whether a CPIC Level A guideline exists. The PREPARE trial (Lancet 2023) showed pre-emptive 12-gene PGx panel reduced clinically relevant adverse drug reactions by ~30%. Pediatric oncology and psychiatry are the highest-yield specialties for upfront PGx.

Research note

The hardest open problem is CYP2D6 calling. Hybrid CYP2D6/CYP2D7 alleles, deletions, duplications, and tandems require either long-read sequencing or specialized callers (Stargazer, Cyrius, Aldy). Short-read PGx panels often report '*1/*1' as a default when they fail to resolve the locus — a false negative dressed as a normal result. Always check the lab's calling methodology.

Myth-buster

Pharmacogenomic testing is experimental.

Reality

FDA labels for >100 drugs reference PGx biomarkers. CPIC Level A guidelines exist for >25 gene–drug pairs. DPYD pre-testing is EU standard of care for fluoropyrimidines. The bottleneck is implementation, not evidence.

Evidence-graded claims

What the data say

Quick check

Test yourself

Q1A patient is starting capecitabine. EU standard of care is to test:
Q2A short-read PGx panel reports CYP2D6 *1/*1. Reasonable next thought:
Glossary

Key terms & abbreviations

CPICClinical Pharmacogenetics Implementation Consortium
Publishes free peer-reviewed gene–drug guidelines with strength ratings (A/B/C/D).
Star allele
Standardized haplotype nomenclature for pharmacogenes (CYP2D6*1, *4, *10, etc.) maintained by PharmVar.
Activity Score
Sum of allele activity values used to translate diplotype into predicted CYP2D6 phenotype (PM, IM, NM, RM, UM).
Further reading

Anchor references