Psychiatric Pharmacogenetic Test: Depression, Anxiety, Medication Response, and Results

A psychiatric pharmacogenetic test examines inherited variants that may affect how the body processes certain medications. For depression and anxiety treatment, the most useful evidence involves drug-metabolizing genes such as CYP2D6, CYP2C19, and, for selected drugs, CYP2B6. These results can sometimes help a prescriber adjust a dose or choose an alternative when metabolism is unusually slow or fast. The test does not diagnose depression or anxiety, reveal their cause, or reliably identify the medication that will work best. Commercial reports often combine well-supported gene–drug guidance with proprietary color categories and genes that lack prescribing evidence. Medication response also depends on symptoms, previous treatment, dose, adherence, other drugs, liver and kidney function, age, smoking, and personal preference. A result is most useful when a clinician interprets the exact gene, drug, phenotype, and guideline instead of treating a green, yellow, or red box as a prescription.

• Pharmacogenetic testing is mainly about drug exposure and adverse-effect risk, not psychiatric diagnosis.
• CYP2D6 and CYP2C19 have the strongest antidepressant prescribing evidence for selected medications.
• SLC6A4, HTR2A, COMT, and MTHFR results do not reliably identify the best antidepressant.
• Current medications can change enzyme activity and override the genetically predicted phenotype.
• Commercial combinatorial reports are not interchangeable and may use undisclosed algorithms.
• Never stop, start, or rapidly change a psychiatric medication based only on a test report.

Table of Contents

• What psychiatric pharmacogenetics can and cannot do
• Genes with the most relevant evidence
• From star alleles to metabolizer phenotypes
• How to read a commercial report
• Drug interactions and phenoconversion
• What clinical trials have and have not shown
• When testing may be most useful
• Using results safely in depression and anxiety care

What psychiatric pharmacogenetics can and cannot do

Pharmacogenetics studies how inherited DNA differences affect medication handling or effects. In psychiatry, the clearest role is to estimate whether a person metabolizes selected drugs more slowly or quickly than average. This can alter blood concentration at the same dose and may influence side effects, lack of benefit, or the time needed to reach a stable level.

The test does not determine whether someone has major depressive disorder, generalized anxiety disorder, panic disorder, obsessive-compulsive disorder, bipolar disorder, post-traumatic stress disorder, or another condition. Diagnosis requires a clinical assessment of symptoms, duration, impairment, medical causes, substance use, medication effects, and safety.

It also does not measure the biological cause of a person’s depression or anxiety. These conditions reflect many genetic and non-genetic factors. A panel of metabolism genes cannot show that serotonin is “low,” that trauma is irrelevant, or that one brain pathway explains the illness.

The test’s most defensible question is narrow: Given that a clinician is considering or already using a particular medication, does a validated gene–drug relationship suggest a different dose, slower titration, closer monitoring, or another medication? That differs from asking the test to rank every psychiatric drug from best to worst.

A pharmacogenetic association can involve pharmacokinetics or pharmacodynamics. Pharmacokinetics describes what the body does to a drug—absorption, metabolism, and elimination. Pharmacodynamics describes what the drug does at its target. Current antidepressant guidelines provide more actionable evidence for several pharmacokinetic genes than for proposed receptor or transporter response markers.

Even a strong metabolism association does not guarantee response. A person can have an average CYP2C19 phenotype and still experience severe side effects from escitalopram, or have a slow phenotype and respond well at a carefully chosen dose. Depression severity, anxiety type, coexisting bipolarity, sleep, pain, medical illness, psychotherapy, adherence, and expectations all affect outcome.

Genes with the most relevant evidence

The presence of a gene on a commercial panel does not mean that professional guidelines recommend using it. Evidence should be considered for each gene–drug pair.

Gene or marker	Potential relevance	Key limitation
CYP2D6	Metabolism of selected antidepressants and other psychiatric drugs, including paroxetine, venlafaxine, and vortioxetine	Complex copy-number and hybrid alleles; strong inhibitors can change the functional phenotype
CYP2C19	Metabolism of citalopram, escitalopram, sertraline, and some tricyclic antidepressants	Effect and recommendation differ by drug; genotype does not predict whether the drug will treat the disorder
CYP2B6	Contributes to sertraline metabolism	Evidence and recommendations are more limited than for some CYP2C19 relationships
SLC6A4 and HTR2A	Studied as antidepressant response or side-effect markers	Current CPIC evidence does not support their clinical use for serotonin-reuptake inhibitor prescribing
COMT	Studied in dopamine biology, cognition, and treatment response	No established antidepressant selection rule from common COMT variants
MTHFR	Influences folate metabolism; commonly marketed in psychiatric panels	Common variants do not identify the best antidepressant or diagnose a folate-related cause of depression

CYP2D6 is highly variable. Some people have no functional copies, while others carry duplications of functional copies. The gene lies near related pseudogenes, and hybrid structures can complicate testing. A limited assay may miss rare variants or misclassify copy number. The laboratory should report which alleles and structural changes it can detect.

CYP2C19 variants can produce reduced, absent, normal, increased, or rapid function. For citalopram and escitalopram, slow metabolism may increase drug exposure and side-effect risk, while rapid metabolism can reduce exposure. Recommendations are drug-specific; a CYP2C19 phenotype should not be applied to medications that are not meaningfully metabolized by that enzyme.

CYP2B6 contributes to sertraline metabolism. The 2023 Clinical Pharmacogenetics Implementation Consortium guideline includes CYP2B6 recommendations for sertraline, but the strength and action differ by phenotype. A panel should not present CYP2B6 as a universal predictor for all antidepressants.

SLC6A4 encodes the serotonin transporter, and HTR2A encodes a serotonin receptor. Both have been widely studied, but inconsistent findings, population differences, and unclear clinical action limit their use. The updated CPIC guideline reviews these genes and does not support using them to guide serotonin-reuptake inhibitor prescribing.

MTHFR results are often misunderstood. Common MTHFR variants can modestly affect folate-related laboratory values, especially with nutritional factors, but they do not prove that depression was caused by impaired methylation. They do not establish a need for a particular antidepressant or supplement. Folate status and supplementation should be evaluated clinically, including diet, pregnancy status, medications, anemia, and measured laboratory findings when indicated.

Other pharmacogenetic markers may matter for medications used in psychiatry. HLA-B15:02 and HLA-A31:01 can affect severe skin-reaction risk with carbamazepine in relevant populations. CYP2D6 can guide atomoxetine dosing. CYP2C9 and HLA markers are relevant to some antiseizure medications. These are separate gene–drug relationships and should not be converted into broad claims about depression or anxiety response.

From star alleles to metabolizer phenotypes

CYP genes are commonly reported with “star allele” names. One copy is inherited from each parent, and the pair is called a diplotype. Each allele is assigned a function value based on known sequence and structural variants. The laboratory then converts the diplotype into a predicted metabolizer phenotype.

Common phenotype categories include:

• **Poor metabolizer:** little or no enzyme function.
• **Intermediate metabolizer:** lower function than normal.
• **Normal metabolizer:** expected function for the reference population.
• **Rapid metabolizer:** higher function, used for genes such as CYP2C19.
• **Ultrarapid metabolizer:** markedly increased function, often related to duplicated functional CYP2D6 copies.

These labels describe enzyme activity, not personality, illness severity, or likelihood of recovery. “Poor metabolizer” does not mean poor treatment response. For an active drug, slower metabolism can increase exposure and side effects. For a prodrug that requires activation, the effect can be different. The direction must be checked for the specific medication.

A result can be indeterminate when the laboratory cannot resolve which variants are on the same chromosome, when copy number is uncertain, or when an allele’s function is unknown. The report should not force an uncertain genotype into a confident phenotype.

CYP2D6 is technically challenging. A result such as “1/2xN” indicates duplication or multiplication, but the clinical interpretation depends on which allele was duplicated. Some assays determine total copy number without establishing the duplicated allele. Hybrid genes involving CYP2D7 can also affect function. A reliable report states technical limitations and whether structural variants were assessed.

Allele frequencies differ across ancestry groups. A panel designed around variants common in one population may misclassify a person from another population if it does not test relevant alleles. Full-gene sequencing and copy-number analysis can improve coverage, but classification still depends on current functional knowledge.

Phenotype translation systems can change. Professional groups periodically update activity scores and allele function. The DNA does not change, but the predicted phenotype or recommendation may. Keep the original genotype and star-allele result, not only the color-coded medication page, so it can be reinterpreted later.

The relationship between phenotype and dose is not a simple multiplier. Guidelines may recommend considering an alternative medication, starting lower, titrating more slowly, monitoring more closely, or using therapeutic drug monitoring. The choice depends on evidence, indication, available alternatives, and the person’s prior experience.

A genotype should be confirmed in a clinical laboratory before it drives a major decision if it originated from research, consumer raw data, or a wellness report. General principles for interpreting genetic variants do not fully capture star-allele systems, but they reinforce the need to distinguish validated findings from uncertain calls.

How to read a commercial report

Many psychiatric panels place medications into green, yellow, and red bins. The colors are convenient but can obscure the underlying evidence. “Green” does not mean the drug will work, “red” does not always mean the drug is unsafe, and “yellow” does not indicate a moderate chance of response.

A medication may be placed in a caution category for several different reasons: predicted high concentration, predicted low concentration, a label-based dose limit, a proprietary pharmacodynamic marker, or an interaction among multiple genes. Those reasons have different clinical importance. The report should identify the exact gene–drug relationship and evidence source.

Combinatorial tests use algorithms that combine several genotypes and sometimes clinical variables into a recommendation. The formula may be proprietary. Two companies can place the same drug in different categories because they test different alleles, weight genes differently, use different interaction rules, or update at different times. A branded category is not a universal pharmacogenetic standard.

Read the report in layers:

1. Verify the patient and sample information.
2. Find the raw genotype, star alleles, copy number, and predicted phenotype.
3. Identify which recommendation comes from CPIC, an FDA label, another professional guideline, or a proprietary algorithm.
4. Check whether the recommendation applies to the exact medication and indication.
5. Review current medications and substances for enzyme inhibition or induction.
6. Compare the result with previous response, side effects, dose, adherence, and blood levels if available.

Some reports include medications that the person has never taken and may never need. This can be useful for future planning, but it can also distract from the present decision. The prescriber should focus on the current treatment options and avoid changing an effective, tolerated medication merely because it appears in a caution bin.

A report may list “use as directed” when no gene-based adjustment is known. That does not guarantee safety. Standard contraindications, drug interactions, pregnancy considerations, age-related warnings, bipolar screening, seizure risk, cardiac effects, and suicide-risk monitoring still apply.

Results should be incorporated into the medication list and medical record in a form that future clinicians can understand. Recording “red drug” without the genotype and reason can lead to inappropriate lifelong avoidance.

Drug interactions and phenoconversion

Genotype predicts baseline enzyme capacity, but the functional phenotype can change because of other medications. This is called phenoconversion. A genetically normal CYP2D6 metabolizer who takes a strong CYP2D6 inhibitor may function like a poor metabolizer.

Fluoxetine, paroxetine, and bupropion can inhibit CYP2D6. Other inhibitors vary in strength. If a person takes one of these medications, a laboratory’s genetically predicted CYP2D6 phenotype may not reflect current drug metabolism. The prescriber must evaluate the whole medication list, including nonpsychiatric prescriptions and over-the-counter products.

Enzyme induction can increase metabolism. Smoking induces CYP1A2 activity through smoke exposure rather than nicotine itself, affecting medications such as clozapine and olanzapine. Stopping smoking can raise drug concentrations even though the genotype remains unchanged. CYP1A2 genotype is generally less clinically decisive than smoking status and interacting drugs for these medications.

Liver disease, kidney disease, inflammation, pregnancy, age, body composition, and nutrition can also alter exposure. Genetic guidance developed in healthy adults may not apply directly to an older person with multiple illnesses or to a child.

Adherence is a common and important explanation for apparent rapid metabolism or treatment failure. A pharmacogenetic report cannot determine whether doses were taken consistently, absorbed, or combined with substances that alter concentration. A careful, nonjudgmental adherence discussion may be more informative than another gene result.

Therapeutic drug monitoring can sometimes measure actual blood concentration. This directly captures genotype, interactions, dose, adherence, and physiology at the time of sampling. It is not available or routinely useful for every antidepressant, but it may clarify unexpected effects for selected medications.

The active metabolites of a drug can complicate predictions. Faster conversion may lower the parent drug while raising an active metabolite. For some gene–drug pairs, evidence about clinical outcomes is weaker than evidence about concentration. Guidelines grade recommendations accordingly.

Polypharmacy can create competing effects. One drug may inhibit an enzyme while another induces a different pathway. Commercial reports generated from the medication list at one visit may become outdated when the regimen changes. The genotype is stable, but the clinical interpretation is dynamic.

Before acting on a result, a pharmacist or prescriber should perform a conventional interaction review. Pharmacogenetics adds to that review; it does not replace it.

What clinical trials have and have not shown

Randomized trials and meta-analyses of pharmacogenomic-guided antidepressant care have produced mixed results. Some studies report modest improvements in response or remission, especially among people whose current medication has a predicted gene–drug interaction. Other studies find small, short-lived, or statistically uncertain benefits.

The large PRIME Care trial in U.S. veterans found that access to a commercial pharmacogenomic report reduced prescribing of medications with predicted drug–gene interactions. Effects on remission were small and not persistent at the end of follow-up. This suggests that the test can change prescribing behavior more clearly than it changes long-term symptoms.

Meta-analyses sometimes find higher remission rates with guided care, but combining trials is difficult. Products use different genes and algorithms. Studies vary in blinding, clinician training, treatment history, outcome timing, and sponsorship. A positive average effect for a group does not prove that every commercial platform or every patient benefits.

Blinding is a major challenge. Clinicians who receive a report know the treatment group and may give more attention, make more medication changes, or communicate greater confidence. Patients may also know or infer that they received personalized testing. These effects can improve care without proving that the algorithm’s genetic predictions are accurate.

The American Psychiatric Association’s 2024 workgroup review concluded that available evidence did not support widespread use of current combinatorial pharmacogenomic tools to select depression treatment. This position does not reject all gene–drug guidance. It distinguishes targeted, guideline-supported pharmacokinetic relationships from claims that a proprietary panel can broadly identify the best antidepressant.

CPIC guidelines answer a different question from outcome trials. CPIC generally explains how to use an existing genotype when a specific drug is being considered; it does not say that everyone with depression should be tested. A strong CPIC recommendation for a gene–drug pair can coexist with uncertainty about universal preemptive panel testing.

FDA pharmacogenetic tables and medication labels also vary in purpose. Some describe exposure differences, some provide dose actions, and some warn that relationships are not established. The absence of a gene–drug pair from an FDA table does not prove no association exists, but a commercial claim should not be represented as FDA-endorsed when it is not.

Evidence is thinner for primary anxiety disorders than for major depressive disorder. Although the same medications and metabolic pathways are used, direct proof that panel-guided care improves anxiety outcomes is limited. Marketing should not stretch depression trial findings into certainty across every psychiatric condition.

When testing may be most useful

Testing may provide the most value when there is a concrete medication question. Examples include repeated side effects at low doses, failure at adequate doses despite good adherence, an unusual dose requirement, several prior medication trials, a planned drug with strong gene-based guidance, or complex prescribing where avoiding an interaction would be useful.

A previously generated result can be helpful before prescribing citalopram, escitalopram, sertraline, paroxetine, venlafaxine, vortioxetine, or certain tricyclic antidepressants, provided the relevant allele coverage and guideline are current. The prescriber should confirm that the report’s phenotype matches contemporary CPIC translation.

Testing can be useful when a person has had severe adverse reactions and clinicians want to avoid repeating a plausible exposure problem. It cannot prove retrospectively that genotype caused the reaction. Dose, interactions, illness, and chance remain possible explanations.

It may be less useful when the person is doing well on a stable medication. An effective, tolerated treatment is direct evidence from the individual and should not be disrupted because an algorithm predicts caution. Testing may still inform future options, but there is no need to manufacture a problem.

Testing also has limited value when the main barrier is not medication metabolism. Untreated bipolar disorder, substance use, severe psychosocial stress, poor access to psychotherapy, incorrect diagnosis, missed doses, inadequate trial duration, or an untreated medical condition will not be solved by a CYP panel.

Before ordering, ask whether the result will change a decision. A useful pretest discussion includes:

• Which medications are under consideration?
• Which genes have actionable guidance for those drugs?
• Does the laboratory test copy number and ancestry-relevant alleles?
• Will insurance cover the test, and what are the out-of-pocket costs?
• Who will interpret the result and update it later?
• How are DNA and data stored, shared, or used for research?

A focused test may be preferable to a large proprietary panel when the question involves one validated gene–drug pair. A broad panel can be reasonable for preemptive use, but unnecessary genes can add confusing claims. Bigger is not automatically better.

Testing children and adolescents requires particular caution because evidence, dosing, developmental pharmacology, and approved indications differ from adults. Results should be interpreted by clinicians experienced in pediatric psychopharmacology.

Using results safely in depression and anxiety care

The safest use begins with the current clinical situation. Review the diagnosis, symptom severity, suicide risk, history of mania or hypomania, substance use, medical conditions, previous medication trials, psychotherapy, and patient goals. Pharmacogenetics is one data point within this assessment.

Do not stop an antidepressant abruptly because it appears in a red category. Sudden discontinuation can cause withdrawal symptoms, rebound anxiety, sleep disruption, physical symptoms, or relapse. Some medications require slow tapering, and the schedule must be individualized.

Do not assume that a green medication is safe to start without standard screening. Antidepressants can worsen mania in susceptible people, interact with other drugs, affect blood pressure or heart rhythm, cause sexual or gastrointestinal effects, and require monitoring for worsening mood or suicidal thinking, especially during treatment changes.

A practical result review can use the following sequence:

1. Confirm the genotype and predicted phenotype.
2. Check whether an independent guideline addresses the exact drug.
3. Account for inhibitors, inducers, smoking, organ function, age, and pregnancy.
4. Compare the recommendation with the person’s actual prior response.
5. Discuss reasonable alternatives, dose changes, and monitoring.
6. Make one clinically understandable change at a time when possible.
7. Track symptoms and adverse effects with a consistent scale and follow-up plan.

A result can reduce uncertainty without eliminating trial and observation. Medication selection still depends on the symptom profile. A sedating medication may be helpful or harmful depending on insomnia, daytime fatigue, falls, and work. Weight, sexual effects, pain, migraine, obsessive symptoms, pregnancy plans, and cost may matter more than a modest metabolism difference.

Psychotherapy, exercise, sleep treatment, social support, and management of medical contributors remain important. A genetics report should not narrow care to medication alone.

Keep the result available for future prescribers, including the raw star alleles and laboratory date. Ask for reinterpretation if guidelines change or if a new medication is considered. A reusable genotype is valuable only when paired with current clinical knowledge.

Urgent psychiatric symptoms require immediate clinical attention, not genetic testing. New suicidal intent, inability to stay safe, severe agitation, psychosis, or mania should be treated as a safety emergency. A pharmacogenetic result cannot determine immediate risk or replace crisis assessment.

References

• CPIC Guideline for CYP2D6, CYP2C19, CYP2B6, SLC6A4, and HTR2A Genotypes and Serotonin Reuptake Inhibitor Antidepressants (2023, Clinical Pharmacology & Therapeutics)
• Table of Pharmacogenomic Biomarkers in Drug Labeling (2026, U.S. Food and Drug Administration)
• Pharmacogenomic Clinical Support Tools for the Treatment of Depression (2024, American Journal of Psychiatry)
• Effect of Pharmacogenomic Testing for Drug–Gene Interactions on Medication Selection and Remission in Major Depressive Disorder (2022, JAMA)
• Pharmacogenomic Testing to Guide Treatment of Major Depressive Disorder: A Systematic Review (2024, Current Psychiatry Reports)
• Antidepressant Pharmacogenetics in Major Depressive Disorder: A Systematic Review (2025, Pharmacogenomics and Personalized Medicine)

Disclaimer

This article is for general education and does not replace psychiatric, medical, or pharmacy care. Pharmacogenetic results must be interpreted with the exact medication, dose, interactions, organ function, diagnosis, and treatment history. Do not start, stop, or change a psychiatric medication based only on a test report; urgent safety concerns require immediate professional assessment.