Pharmacogenomics
Also known as: pharmacogenetics, PGx, pharmacogenomic testing
Pharmacogenomics is the study of how a person's genes affect the way they respond to medications, from how fast they break a drug down to whether they're at risk of a serious reaction.
A genetic test can help doctors choose the right drug and dose the first time. While pharmacogenomic testing is backed by solid evidence for certain drugs, it's often oversold as a cure-all for trial-and-error prescribing, especially for antidepressants.
Pharmacogenomics, explained at three levels
At home
Think of your liver as a recycling plant for medicines, with workers whose speed is set partly by your genes. Some people's workers are slow, so a drug piles up and causes side effects. Others are very fast, so the drug is cleared before it works, or converted into something stronger. Pharmacogenomics reads those genes to predict how you will process certain medicines.
At work
Pharmacogenomics uses genetic test results to guide drug choice and dosing. It focuses on genes that affect how drugs are metabolized, transported, or recognized by the immune system. For health organizations, it can prevent serious adverse reactions and failed treatments for specific drug-gene pairs, reducing hospital admissions and costs, provided results reach prescribers at the moment they prescribe.
At home
Many medicines are processed by liver enzymes from the CYP family, such as CYP2D6 and CYP2C19. You inherit two copies of each gene, and variants change how well the enzyme works. Labs combine your two copies into a type: poor, intermediate, normal, or ultrarapid metabolizer. For codeine, which CYP2D6 converts into morphine, a poor metabolizer gets little pain relief, while an ultrarapid metabolizer can get a dangerous dose.
At work
Pharmacogenomic genes fall into groups. Metabolism genes, such as CYP2C19 for clopidogrel and DPYD for fluoropyrimidine chemotherapy, change drug levels. Immune genes, such as HLA-B*57:01 before abacavir, predict severe hypersensitivity. Others, like TPMT and NUDT15 for thiopurines, affect toxic metabolites. Labs report star alleles, which CPIC guidelines translate into phenotypes and prescribing actions. Preemptive panels test many genes once and store the results for future prescriptions.
At home
A gene result predicts one part of how a drug behaves, and other factors can override it. Another medication can block the same enzyme, turning a normal metabolizer into a poor one, called phenoconversion. Age, kidney and liver function, and diet also shift drug levels. For many antidepressants, response depends on brain biology no gene test captures. In my view, consumer tests promising your perfect antidepressant overstate the evidence, which is strongest for a smaller set of drugs.
At work
Implementation hinges on data plumbing. Results must arrive as discrete, coded genotypes and phenotypes in the EHR, not PDF reports, so clinical decision support can fire an alert when a matching drug is ordered. Star-allele nomenclature and CPIC or DPWG guidelines standardize interpretation. Variant knowledge is weaker for under-studied ancestries, so tests may miss relevant alleles. Plan for lab validation, coverage policy, prescriber and pharmacist education, and genetic privacy protections.
Real-world examples
At home
Nadia had a bad reaction to codeine after dental surgery. A pharmacogenomic test later showed she was an ultrarapid CYP2D6 metabolizer, so her doctor flagged her chart to avoid codeine and tramadol.
At work
Jerome is a hospital pharmacist in Alberta. His team began testing patients for DPYD variants before certain chemotherapy drugs, and adjusted doses for those at risk of severe, sometimes fatal, toxicity.
Common confusions
- A gene test will tell me which medicine works best for any condition. — Strong evidence exists for specific drug-gene pairs, not for most medications.
- Pharmacogenomics is only for rare diseases. — It affects common drugs such as some painkillers, blood thinners, and antidepressants.
- Results need repeating for each new prescription. — Your genes don't change, so one test can inform prescriptions for life.
For analysts
Pharmacogenomic results only help if prescribers see them at the right moment. I'd make EHR integration and alerts a requirement of any testing program, rather than storing results as PDFs nobody reads.
History & origin
Doctors noticed in the 1950s that some patients reacted unusually to certain drugs, and the term 'pharmacogenetics' was coined in 1959. The Human Genome Project made it possible to study these differences at scale. A tragic Toronto case published in 2006 brought the issue home: a breastfed newborn died after the mother, an 'ultrarapid metabolizer' of codeine due to her CYP2D6 genes, passed high levels of morphine through her milk. The Clinical Pharmacogenetics Implementation Consortium (CPIC), formed in 2009, began publishing guidelines on using genetic results for specific drugs. Today, regulators list pharmacogenomic information on hundreds of drug labels, and some Canadian pharmacies offer tests directly. In my opinion, pharmacogenomics works best quietly built into prescribing systems, not sold as a standalone consumer test.