GeneSight & Pharmacogenomics

What the FDA and CPIC Guidelines Say About Pharmacogenomic Testing

By Dr. Stacey Forbes, DNP, APRN, PMHNP-BC · Updated October 12, 2026

Neither the FDA, CPIC nor the American Psychiatric Association’s task force recommends a pharmacogenomic test for everyone starting a psychiatric medication. The FDA lists specific gene-drug pairs with good evidence and puts some of that advice in drug labels, but says a listing is not a call to test. CPIC writes detailed guidance for using results you already have and rates only some pairs as actionable. An American Psychiatric Association task force judged the evidence insufficient for widespread use of combination panels. Together they support targeted, thoughtful use rather than routine testing.

Three different questions

Most confusion about pharmacogenomic “guidelines” comes from mixing up three separate questions:

1. Is this particular gene-drug link real? For example, do CYP2C19 poor metabolizers really get higher citalopram levels? The FDA and CPIC answer this pair by pair.
2. If I already have a result, how should it change prescribing? CPIC guidelines and FDA labels answer this.
3. Does testing a whole panel before choosing a drug lead to better outcomes? Only clinical trials can answer this, and professional bodies weigh those trials.

A strong yes to the first two does not automatically mean a yes to the third. This page covers what the regulators and guideline writers say. For the trial evidence in more detail, see does GeneSight testing work?.

The FDA’s Table of Pharmacogenetic Associations

The FDA lists gene-drug pairs where it believes there is enough evidence that people with certain variants may have altered drug metabolism or response. The table, current as of 10 September 2026, has three sections:

The FDA attaches clear caveats. Inclusion “does not necessarily mean the FDA advocates using a pharmacogenetic test before prescribing,” unless the test is a companion diagnostic. Each patient’s genetic makeup is described as only one of many factors that affect drug levels and response, and the page says patients should not adjust their medications without consulting their prescriber. The table is updated periodically; its 10 September 2026 revision added several drugs, including lamotrigine and doxepin, to Section 1.

For a handful of drugs the label itself goes further. The atomoxetine label, for example, tells prescribers to consider genetic testing to determine CYP2D6 status; pharmacogenomic testing and ADHD medication explains why.

The FDA’s warnings about test claims

In 2018 the FDA issued a safety communication warning consumers and clinicians about pharmacogenetic tests, sold directly or through providers, that claim to predict how a patient will respond to specific medications where that link has not been established. In April 2019 it issued a warning letter to Inova Genomics Laboratory over tests claiming to predict response to named drugs, including antidepressants, saying it was unaware of data showing they help patients or clinicians make better decisions.

The same 2019 announcement said most firms it contacted after the 2018 warning addressed its concerns by removing specific medication names from their labeling, promotion and patient reports. It also explained that when the FDA finds the evidence sufficient, instructions for using a genetic result appear in that drug’s labeling.

GeneSight’s own report states that the test has not been cleared or approved by the FDA; it is a laboratory-developed test interpreted from published literature, internal research and FDA labels. That is common for this type of test, and it is one reason the report insists results should supplement, not replace, clinical judgment.

CPIC: how to use a result, not whether to order one

The Clinical Pharmacogenetics Implementation Consortium (CPIC) publishes peer-reviewed gene-drug guidelines. They are written to interpret genetic results that already exist: CPIC’s 2023 antidepressant guideline describes itself as informing the interpretation of existing genotype results, and says which patients to test, which test to choose and cost-effectiveness are beyond its scope.

CPIC assigns each gene-drug pair a level from A to D. Its leaders have described levels A and B as actionable and levels C and D as not actionable. Some examples from the CPIC database as of 12 October 2026:

• Level A: CYP2C19 with citalopram, escitalopram and sertraline; CYP2D6 with paroxetine, vortioxetine and atomoxetine; CYP2D6 and CYP2C19 with amitriptyline.
• Level B: CYP2B6 with sertraline; CYP2D6 with venlafaxine and fluvoxamine.
• Level C (no recommendation): SLC6A4 and HTR2A with the SSRIs and SNRIs, where CPIC calls the evidence “mixed and insufficient to support clinical utility.”
• No CPIC guideline: ADHD stimulants. Amphetamine with CYP2D6 is rated B/C and methylphenidate with CYP2D6 is rated C.

On commercial combination panels, which merge many genes into one ranking, CPIC says trial data are accumulating with “collectively positive but modest results,” but that the proprietary, opaque nature of those algorithms puts them beyond what its process can evaluate.

Professional society positions

In 2018 the American Psychiatric Association’s Task Force for Biomarkers and Novel Treatments reviewed combinatorial pharmacogenetic tools for depression. It concluded there was “insufficient evidence to support widespread use” of these tools at that time, while noting there are situations where the technology may be informative, particularly in predicting side effects.

That review predates the two largest randomized trials. The VA’s PRIME Care trial, published in JAMA in 2022, randomized 1,944 veterans with depression at 22 medical centers. Testing shifted prescribing away from drugs with predicted gene-drug interactions, and remission over 24 weeks was modestly higher overall, but the difference at week 24 itself was not significant. The authors concluded that providing results “had small nonpersistent effects on symptom remission.”

What this means if you are deciding whether to test

If you already have a result, how to read your GeneSight report explains where the FDA-recognized interactions are marked. At Willow & Stone, testing is ordered when it is clinically useful, most often after medications have not worked or have caused side effects, and the result is reviewed with you as one data point among many. Our GeneSight testing overview describes that approach, and treatment-resistant depression and genetic testing covers the situation where people most often ask.

Common Questions

Sources & Further Reading

• FDA — Table of Pharmacogenetic Associations
• FDA — Warning letter to genomics lab over genetic test claims (April 2019)
• CPIC guideline for CYP2D6, CYP2C19, CYP2B6, SLC6A4 and HTR2A and antidepressants (Bousman et al., 2023)
• Relling et al., The Clinical Pharmacogenetics Implementation Consortium: 10 years later (2020)
• Zeier et al., APA Task Force review of pharmacogenetic decision support tools (Am J Psychiatry, 2018)
• Oslin et al., PRIME Care randomized trial (JAMA, 2022)

This article is general education, not advice for your situation. Do not start, stop, or change a medication based on a genetic test result without talking to the clinician who prescribes it.