Can a genetic test tell you which antidepressant will work?

The saliva tube is sitting on your kitchen table looking remarkably confident for something that hasn’t met you. The brochure says your DNA may help guide antidepressant treatment. After one medication made you nauseated and another did approximately nothing, a shortcut sounds less like science fiction and more like basic customer service.

Then the report arrives with green, yellow, and red boxes. Green feels like permission. Red feels like danger. Your genes have apparently become a tiny traffic department.

Pharmacogenetic testing can provide useful information. It can’t do the most tempting thing advertised around it: look into your future and name the antidepressant that will definitely work.

Your genes may affect the trip, not choose the destination

Pharmacogenetics looks at genetic variants that may influence how your body handles medication. For some antidepressants, enzymes such as CYP2D6, CYP2C19, and CYP2B6 help break the drug down. A genetic result may suggest that an enzyme works more slowly or more quickly than average.

That can matter. A slower pathway may raise exposure to a medication at a given dose and sometimes increase side-effect risk. A faster pathway may lower exposure enough to make a usual dose less useful. The Clinical Pharmacogenetics Implementation Consortium provides recommendations for using certain results to inform prescribing for particular serotonin-reuptake antidepressants.

Notice the verbs: may influence and inform. Metabolism isn’t the same as response. A gene can help explain how much drug reaches your system without knowing whether the treatment will lift your mood, quiet panic, improve sleep, or feel acceptable in your actual life.

The test isn’t reading the whole person

Your response to an antidepressant can be shaped by diagnosis, symptom pattern, dose, duration, other medicines, liver and kidney function, age, pregnancy, smoking, substances, sleep, medical conditions, expectations, adherence, and plain biological complexity. Genes are part of that room. They aren’t the entire meeting.

Other medications can also inhibit or induce the same enzymes a test measures. CPIC calls the resulting mismatch between genetically predicted metabolism and observed metabolism “phenoconversion.” In ordinary language, your DNA report may say the road is clear while another medicine has quietly parked a delivery truck across it.

Commercial panels may test different variants and use proprietary algorithms to combine them. That means two reports don’t necessarily sort every medication the same way. A polished category isn’t a universal law, and a medication in a red box isn’t automatically dangerous or forbidden.

I’ll be blunt: if a report is presented as a definitive answer to which antidepressant your brain needs, the confidence has outrun the evidence. Useful tools survive accurate limits. Sales pitches usually prefer mood lighting.

The best evidence is encouraging in a fairly uncinematic way

The PRIME Care randomized trial studied almost two thousand veterans with major depression who were starting or switching an antidepressant. Clinicians who received pharmacogenetic results prescribed fewer medications with predicted drug-gene interactions. That’s a real effect and a sensible use of the information.

Symptom outcomes were less dramatic. Across 24 weeks, remission was modestly more likely in the testing group, but the difference was small and wasn’t statistically significant at the final 24-week point. The authors described the remission effect as small and nonpersistent. A 2024 umbrella review and updated meta-analysis found possible benefits across studies, while also noting inconsistency among tests and the evidence base.

So the fair summary isn’t “genetic testing is useless” or “genetic testing ends trial and error.” It may help clinicians avoid some predicted drug-gene conflicts and may modestly improve outcomes in some settings. It doesn’t reliably place a golden arrow over the one right antidepressant.

A green box doesn’t promise comfort

Color categories are easy to overread. Green may simply mean the test didn’t identify a major gene-drug issue in its algorithm. You can still have side effects. The medicine can still be ineffective. Yellow may mean a dose adjustment, closer monitoring, or an alternative is worth considering, not that the medication is bad. Red may reflect a predicted interaction, not a command to stop immediately.

The FDA says pharmacogenetic information should be considered alongside other patient and clinical information. It also warns that inclusion of a gene-drug association in its table doesn’t automatically mean everyone should be tested before receiving that medicine. Most listed associations haven’t been proven to improve clinical outcomes through testing.

Direct-to-consumer results add another wrinkle. The FDA has not authorized a direct-to-consumer pharmacogenetic test to predict whether someone will respond to a specific therapeutic drug. Results that could affect treatment may need confirmation in a clinical laboratory. Your mail-in kit doesn’t get independent prescribing privileges.

Testing can be useful when it answers a real question

A test may be worth discussing when you’ve had unusual side effects at low doses, several disappointing medication trials, a complex medication list, or a known result that could inform a specific choice. It may also be useful when a clinician is already considering a medication with a well-supported gene-drug guideline.

It may add less when you’re doing well on a stable medication without meaningful side effects. CPIC specifically notes that people stable on effective doses may not benefit from changing those doses based on a newly discovered result. Don’t let a laboratory category talk you out of a treatment that’s actually helping.

Cost matters too. Ask what the test costs, whether insurance covers it, which genes it includes, whether the laboratory is qualified, how results will change the plan, and who will interpret them. If the answer to “What would we do differently?” is vague, the test may be collecting curiosity rather than useful information.

Your medication history is still premium data

Before paying for a new test, build a one-page medication map. List each psychiatric medicine, approximate dose, how long you took it, what improved, what didn’t, side effects, why it stopped, and any other medicines or substances present at the time. Family response can sometimes add context, although it isn’t destiny either.

That record is less futuristic than a cheek swab and often more immediately actionable. A trial that ended after five days because of nausea tells a different story from an adequate trial that never changed symptoms. “It didn’t work” is a conclusion; duration and dose are evidence.

If you already have pharmacogenetic results, bring the full report, not a screenshot of the colorful page. Ask which recommendations come from established gene-drug guidance and which come from the company’s combined algorithm. Then decide together whether the result changes dosing, monitoring, or the shortlist.

Don’t let the report make an urgent decision alone

Never abruptly stop an antidepressant because a box is red. Stopping can cause discontinuation symptoms and can allow depression or anxiety to return. Contact the prescriber who knows your history, especially if you’re having side effects, worsening symptoms, agitation, or suicidal thoughts.

If you might act on suicidal thoughts, can’t stay safe, or have a severe reaction such as major breathing difficulty, swelling, collapse, or another medical emergency, call 988 when appropriate or use emergency services. Genetic testing is a planning tool. It isn’t emergency care.

Today’s experiment doesn’t require saliva. Spend 10 minutes starting your medication map. One honest page can turn a fuzzy history into a better next decision, whether genetic testing eventually joins the conversation or not.

Sources: Clinical Pharmacogenetics Implementation Consortium, guideline for serotonin reuptake inhibitor antidepressants (2023); U.S. Food and Drug Administration, Table of Pharmacogenetic Associations and Direct-to-Consumer Tests; Oslin and colleagues, “Effect of Pharmacogenomic Testing for Drug-Gene Interactions on Medication Selection and Remission,” JAMA (2022); Tesfamicael and colleagues, umbrella review and updated meta-analysis, Frontiers in Psychiatry (2024).