Treatment-Response Research and Genetics

Start here

This page covers research on why people respond differently to the same medicine, and where genetics fits. It is education only. Never start, stop or change a prescription because of a genetic result. Ask the prescriber. For how consumer reports fit in, see direct-to-consumer genetic tests.

What pharmacogenomics is

Pharmacogenomics is the study of how genes affect a person's response to drugs. Researchers combine pharmacology and genomics to understand who will benefit, who will not respond and who may have adverse reactions. [1] Many other factors also matter: sex, age, weight, diet and the use of other medications. [2] The explainer is at pharmacogenomics.

What the FDA says it has evaluated (United States, read 2026-10-22)

The FDA keeps a Table of Pharmacogenetic Associations. For listed pairs, it believes there is sufficient evidence that subgroups with certain variants are likely to have altered drug metabolism and, in some cases, different therapeutic effects or adverse event risks. [3] The agency adds that listing a pair does not mean it advocates testing before prescribing, unless the test is a companion diagnostic. [3] Most associations in the table have not been evaluated for the impact of testing on clinical outcomes. [3]

Examples from the table as read: for abacavir, people positive for HLA-B*57:01 have higher risk of hypersensitivity reactions, and abacavir should not be used in them. For clopidogrel, CYP2C19 intermediate or poor metabolizers have lower antiplatelet response and may have higher cardiovascular risk, and the table says to consider another platelet P2Y12 inhibitor. [3] These are decisions for prescribers using the drug label.

What trials show

Evidence label: promising but limited, varying by drug and condition.

Safety panel. In the PREPARE study, 6,944 patients in seven European countries starting a medicine covered by Dutch guidelines were enrolled; the design was a cluster-randomized crossover implementation study, open-label, funded by the EU Horizon 2020 program. A 12-gene panel guided treatment. Among patients with an actionable result, clinically relevant adverse drug reactions occurred in 21.0% with testing versus 27.7% with standard care (odds ratio 0.70, 95% CI 0.54 to 0.91). [4] Limits: it was open-label, 97.7% of participants reported European, Mediterranean or Middle Eastern ethnicity, and follow-up was 12 weeks. [4]

Antidepressants. A 2022 meta-analysis of 13 trials (4,767 patients, 10 randomized) found remission 1.41 times more likely (95% CI 1.15 to 1.74) with pharmacogenomic-guided therapy than unguided therapy; the authors called the increase modest and noted test composition varied across trials. [5] Another meta-analysis of 11 studies found higher response and remission at 8 and 12 weeks but no significant difference at 4 or 24 weeks. [6] So the result is mixed on durability.

Trials of "genetic matching" for diets, supplements or fitness are a different matter; see genetics as a nutrition input.

Consumer tests and drug response

The FDA authorized one DTC pharmacogenetic report (DEN180028) covering variants in genes such as CYP2C19 and CYP2D6, but it states the report does not say whether a person will respond to a particular medicine and should not be used to start, stop or change treatment. The FDA also said it had not authorized any DTC pharmacogenetic test that predicts response to a specific drug. [7] The FDA has warned about tests with unapproved claims to predict response to a specific drug where claims may not be supported by clinical evidence. [7]

What genes cannot do

Genes are one input. Risk and response also depend on the other medicines you take, your liver and kidney function, and dosing. Predicting response from a gene is a different task from choosing a treatment; see risk prediction vs treatment selection. Keep a medicine list and share any genetic report with your prescriber.