Drug monitoring and genetic testing: what the evidence supports

Transcript

SoheeA client arrives excited. She swabbed her cheek, mailed it off, and got back a report that she says tells her exactly which antidepressants will work for her. She wants to bring it to her psychiatrist and switch medications.

RyanAnd you want to support her hope. But you also want to know: what does the evidence actually say about tests like that?

SoheeWelcome to EPPP Prep, the podcast. I'm Sohee.

RyanI'm Ryan. This is our last Domain 1 episode. Two topics from the same knowledge statement: monitoring drugs through blood tests, and genetic testing, from newborn screening to pharmacogenetics.

SoheeStart with therapeutic drug monitoring. MedlinePlus, from the National Library of Medicine, defines it as testing that measures the amount of certain medicines in the blood, to check whether the amount is safe and effective.

RyanMost medicines don't need it. But some must stay at a certain level in the blood to work well. The goal is to land in what's called the therapeutic range: medically helpful, but not dangerous.

SoheeWhy can't you just pick a dose and stop worrying? Because the body processes and removes medicine over time, and that can change with aging, pregnancy, illness and infection. Other medicines can change levels too.

RyanAnd MedlinePlus lists lithium and valproic acid as monitored medicines for mental health conditions. It also notes monitoring can be used to make sure a person is taking the medicine as prescribed.

SoheeLithium is the textbook case, and its label from the U.S. Food and Drug Administration, the FDA, shows why. Its boxed warning says lithium toxicity is closely related to serum lithium concentrations and can occur at doses close to therapeutic concentrations.

RyanSo the label says facilities for prompt and accurate serum lithium measurement should be available before treatment begins.

SoheeTiming is part of the method. The label tells prescribers to get a serum lithium level after three days, drawn 12 hours after the last oral dose, and regularly until the patient is stabilized. And to monitor again after any change in dose or in other medications, like diuretics or nonsteroidal anti-inflammatory drugs.

RyanAnd monitoring goes beyond the drug level. Before starting lithium, the label calls for checking kidney function, vital signs, electrolytes and thyroid function. Thyroid monitoring continues at three months and then every six to twelve months, because lithium can cause hypothyroidism.

SoheeThe National Institute of Mental Health, NIMH, sums it up: people on lithium usually take part in regular monitoring of lithium levels and kidney and thyroid function. For anticonvulsant mood stabilizers, monitoring checks medication levels, side effects and interactions.

RyanThen there's safety monitoring that isn't a drug level. NIMH says people on atypical antipsychotics may be asked to monitor weight, glucose and lipids. And clozapine requires regular blood tests for that dangerous side effect we discussed, severe neutropenia.

SoheeHere's a recent change worth knowing. In February 2025, the FDA said it no longer expects prescribers, pharmacies and patients to take part in the clozapine REMS program, the Risk Evaluation and Mitigation Strategy, or to report blood test results before dispensing. It also told the manufacturers the REMS must be eliminated.

RyanBut the FDA still recommends that prescribers monitor the absolute neutrophil count according to the prescribing information. The risk didn't go away. The reporting requirement did.

SoheePause and think. A question asks why a lithium level is drawn on a schedule rather than whenever convenient, and why it's rechecked after a new blood pressure medication. What's the principle?

RyanLithium's toxic range sits close to its therapeutic range, so small changes matter. Levels are compared at a standard time after the dose, and anything that changes clearance, like a diuretic, can push levels up. The tempting wrong answer is that monitoring is only needed at the start.

SoheeNow genetics. Pharmacogenomics, per MedlinePlus, is the study of how genes affect a person's response to drugs. The hope is to predict who will benefit, who won't respond, and who will have side effects.

RyanAnd the FDA keeps a Table of Pharmacogenetic Associations, where it believes the scientific evidence is sufficient that certain genetic variants alter how some patients metabolize a drug, and in some cases its effects or risks.

SoheeA few examples from psychiatry. For citalopram, people who are poor metabolizers through an enzyme called CYP2C19 get higher drug concentrations and a higher risk of a heart rhythm effect, QT prolongation.

RyanFor aripiprazole, CYP2D6 poor metabolizers get higher concentrations and a higher risk of adverse reactions. And for carbamazepine, people positive for a variant called HLA-B 15:02 have a higher risk of severe skin reactions.

SoheeBut read the FDA's caveats closely, because they're the heart of the evidence question in this knowledge statement. Most associations in the table haven't been evaluated for whether genetic testing improves clinical outcomes. And a patient's genetic makeup is only one of many factors that affect drug levels and response.

RyanThe carbamazepine entry even says genotyping is not a substitute for clinical vigilance.

SoheeNow our client's report. In 2019, the FDA issued a warning letter to a lab marketing tests that claimed to predict responses to specific drugs, including antidepressants. The FDA said the tests hadn't been reviewed for safety and effectiveness, and it was unaware of any data showing the tests help patients or providers make appropriate treatment decisions for those drugs.

RyanThe agency's device center director said the FDA continues to warn patients and health care professionals that they should not rely on these tests for treatment decisions, where such claims haven't been established.

SoheeAnd the risk is concrete. The FDA said patients may change a medication dose based on such a test, which may lead to incorrect treatment or worsening illness.

RyanAnd the FDA's page on direct-to-consumer tests, last updated in 2019, says that as of then the FDA had not authorized any direct-to-consumer pharmacogenetic test that predicts whether a person will respond to, or have adverse reactions from, any specific drug.

SoheeSo with our client: honor the hope, share that the evidence for these claims is limited, and encourage her not to change anything on her own but to review it with her prescriber. That matches NIMH's guidance about not stopping medication without a provider.

RyanNow genetic screening and testing more broadly. MedlinePlus lists the main uses. Newborn screening, to find treatable genetic disorders early. Diagnostic testing, to confirm or rule out a condition. Carrier testing, to find people who carry one copy of a variant that causes disease in two copies.

SoheePrenatal testing, for changes in a fetus's genes or chromosomes. Preimplantation testing of embryos created through in vitro fertilization. And predictive and presymptomatic testing, for people with a family history but no symptoms.

RyanNewborn screening has one of the clearest success stories. The National Institute of Child Health and Human Development says all 50 states require newborns to be screened for phenylketonuria, PKU, using a few drops of blood from the baby's heel.

SoheeBefore screening was possible, most infants with PKU developed severe intellectual and developmental disabilities. Since screening, and research on a restricted diet to treat it, PKU has been almost completely eliminated as a cause of those disabilities.

RyanAnd notice the logic of screening versus diagnosis. If a newborn screen is positive, the child needs additional tests to confirm it.

SoheePredictive testing is more complicated, emotionally and scientifically. Take Huntington's disease. Because of the CAG repeat counts we covered, a result of 40 or more repeats almost always means the person will develop the disorder. But 36 to 39 repeats may or may not.

RyanAnd for late-onset Alzheimer disease, MedlinePlus is careful. People with the APOE e4 allele inherit an increased risk, not the disease itself. Not everyone with Alzheimer's has e4, and not everyone with e4 develops the disease.

SoheePause and think. A client learns she carries one APOE e4 allele and says, so I'm going to get Alzheimer's. How do you respond, accurately and kindly?

RyanE4 increases risk but doesn't determine the outcome. Not everyone with it develops the disease, and not everyone with the disease has it. The tempting wrong answer treats a risk variant like a dominant disease gene, which is a different situation, like Huntington's.

SoheeMedlinePlus also lists the limits and risks of genetic testing. Many of the risks are emotional, social or financial. People may feel angry, depressed, anxious or guilty, and results can create tension in families, because they reveal information about relatives too.

RyanAnd tests often can't tell whether a person will show symptoms, how severe they'll be, or how the disorder will progress. Plus, many genetic disorders lack treatment once diagnosed.

SoheeThen the law. The Genetic Information Nondiscrimination Act, GINA, has two parts. Title I prohibits genetic discrimination in health insurance. Title II prohibits it in employment.

RyanBut GINA has gaps the exam likes. It doesn't apply to employers with fewer than 15 employees, and it doesn't protect against discrimination in life, disability or long-term care insurance.

SoheeRecap. Therapeutic drug monitoring keeps hard-to-dose drugs in the therapeutic range. Lithium needs timed blood levels, plus kidney and thyroid checks. Clozapine needs neutrophil monitoring, even though the FDA no longer expects REMS participation or reporting.

RyanPharmacogenetic associations are real for some drugs and listed by the FDA, but most haven't been evaluated for whether testing improves outcomes, and the FDA warns against relying on tests that claim to predict antidepressant response.

SoheeGenetic testing has many uses. Newborn PKU screening is a clear success. Predictive results can be uncertain, like intermediate Huntington's repeats or APOE e4, which raises risk but isn't destiny. And GINA protects health insurance and employment, not life, disability or long-term care insurance.

RyanAnd that's Domain 1. Synapses, brain regions, disorders, drugs, trials, genes and tools. If you've listened all the way through, that's a real accomplishment. This is often the domain people dread most.

SoheeGo try some Domain 1 practice questions while it's fresh, and be proud of how far you've come. Thanks for listening to EPPP Prep.

Loading the read-along timings…