Robert H. Howland, MD
Associate Professor of Psychiatry, University of Pittsburgh School of Medicine, Western Psychiatric Institute and Clinic, Pittsburgh, PA
Dr. Howland has disclosed that he has no relevant financial or other interests in any commerical companies pertaining to this educational activity.
“Personalized medicine” is a buzzword in healthcare and stems from the idea that treatments can be designed specifically for a patient, based on his or her own biological characteristics.
In psychiatry, personalization is largely based on “pharmacogenetics,” the selection of medications based on genetic factors associated with drug response and tolerability. Could your patient’s genetic code predict which medications you prescribe?
It’s important to point out that some genes affect pharmacokinetics while others involve pharmacodynamic processes. Pharmacokinetics refers to how quickly and efficiently a drug reaches its target and how quickly it leaves the body: drug absorption, distribution, metabolism, and excretion. Clinically, the most important contributor is the cytochrome P-450 (CYP450) system, which accounts for the metabolism of approximately 60% of prescribed drugs.
Multiple CYP450 enzymes exist and are classified according to a standardized nomenclature. The major enzymes of interest in clinical psychopharmacology are 1A2, 2B6, 2C9, 2C19, 2D6, and 3A4. For instance, fluoxetine is a substrate of 2D6; increased activity of this enzyme means lower blood levels of fluoxetine, while decreased activity corresponds to higher blood levels.
Pharmacodynamics, on the other hand, refers to the mechanism of action of a drug at its particular target(s). Whenever you prescribe a psychotropic drug, you are (most likely) thinking about the drug’s targets: receptors, transporters, or enzymes. Each of these directly or indirectly regulates the synthesis, transmission, or degradation of neurotransmitters such as serotonin and dopamine. Similar to the enzymes mentioned above, pharmacodynamic targets exist as proteins produced by different genes. Slight variations in the coding for a particular gene are referred to as polymorphisms, and these can alter the amount, structure, binding, or function of these proteins. In turn, these differences in the protein targets can influence the therapeutic or adverse effects of the drugs you prescribe.
For a well-known example of pharmacodynamic variation, consider the serotonin reuptake transporter (SERT). SERT regulates the reuptake of serotonin into neurons, and is the main site of action of selective serotonin reuptake inhibitor (SSRI) antidepressant drugs. Multiple genetic polymorphisms in SERT have been identified. Some research suggests that patients carrying certain SERT polymorphisms (such as the S or “short” allele) may respond less well to SSRI drugs and may experience more adverse effects of SSRIs, but the correlation is not absolute.
Polymorphisms of other genes involved in the pharmacodynamics of drug response, such as serotonin (5-HT) receptors, dopamine receptors, and other transporters, have been studied. But no single genetic difference, as of now, is significant enough to predict an outcome when you prescribe a drug.
In recent years, numerous products have come on the market to analyze genetic polymorphisms. The first commercially available product was the AmpliChip CYP450 Test, developed by Roche Diagnostics and approved by the FDA in 2004. Using a small blood sample from the patient, it analyzes genetic polymorphisms associated with two metabolizing enzymes (2D6 and 2C19). Based on the patient’s 2D6 and 2C19 polymorphisms, his or her 2D6 metabolic activity is characterized as poor, intermediate, extensive, or ultra-rapid, and 2C19 activity as poor or extensive. This information can theoretically be used to make clinical decisions about drugs that are 2D6 or 2C19 substrates.
Many newer pharmacogenetic tests, based on similar technology, are currently available on the market. The most popular ones include Genecept and GeneSight. These tests analyze the majority of the known 450 enzyme polymorphisms, as well as various combinations of pharmacodynamic genes.
Laboratory Tests are Under-regulated by FDA
Despite their ready availability, does pharmacogenetic testing make sense for your patient? Does it really matter whether the patient in front of you is a “fast” or a “slow” metabolizer of a drug you are prescribing?
The answer to these key questions comes down to data, which I’ll review later in this article. But first, it’s important to know a bit about how these tests are regulated (or, more accurately, under-regulated). We are all familiar with the standards used by the FDA to approve medications: companies must submit double-blind, placebo-controlled trials and the FDA carefully scrutinizes the data before finally rendering a decision about approval.
Not so with laboratory tests. In fact, there are no specific federal requirements for laboratories to establish or verify the clinical validity of their tests, and laboratories generally do not have the capability to develop evidence of clinical utility. The bottom line is that the availability of a test should not be assumed to be proof that it has been proven to enhance clinical outcomes. Partly because of this problem, the FDA is currently developing draft guidelines on the regulation of laboratory tests, which would include pharmacogenetic test products.
What Does The Data Show?
Seven years ago, the federal Agency for Healthcare Research and Quality (AHRQ) reviewed existing studies to determine if testing for 450 polymorphisms in patients taking SSRIs leads to improvement in outcomes or if testing results are useful in medical, personal, or public health decision-making (Thakur M et al, Genet Med 2007;9(12):826–835). The review revealed few high-quality, clinical studies. Several studies included non-randomized design, small numbers of subjects, and a failure to account for other genetic factors that may influence SSRI response or tolerability. There were no prospective studies of P450 genotyping and its relationship to clinical outcomes. There was no correlation between P450 polymorphisms and SSRI drug levels, efficacy, or tolerability. There were no data regarding whether testing leads to improved depression outcomes; whether testing influences medical, personal, or public health decision-making; or whether any harms are associated with testing itself or with subsequent management decisions. A more recent study found no clear benefit of testing for pharmacodynamic targets (de Leon J, Pharmacol Res 2009;59(2):81–89).
All this negative data has not dissuaded testing companies from marketing their products to us, sometimes aggressively so. The “GeneSight Psychotropic” test, offered by Assurex Health, detects genetic polymorphisms associated with six metabolic enzymes (1A2, 2B6, 2C9, 2C19, 2D6, and 3A4) and two pharmacodynamic genes (5HT2A and SERT). They claim that the results are potentially relevant to the use of 22 antidepressant drugs and 16 antipsychotic drugs.
The testing process is quite simple: blood samples or mouth swabs are sent to a central laboratory for analysis, and the results (available in 36 hours) categorize each of these 38 drugs into one of three groups: 1) little or no gene-drug interaction; 2) moderate gene-drug interaction; and 3) severe gene-drug interaction. For a particular patient, the use of drugs within each group is characterized as “use as directed” (referred to as “green bin” drugs), “use with caution” (“yellow bin”), and “use with caution and with more frequent monitoring” (“red bin”). The “green bin” drugs require no special dosing considerations for the patient. For drugs within the yellow and red “bins,” additional comments about their potential use are provided in the laboratory report. These comments might explain expected changes in drug blood levels (such as too high or too low) or expected clinical effects (such as reduced efficacy or increased side effects).
Does this information lead to better clinical outcomes? Two open-label studies have reported that GeneSight Psychotropic was effective for managing patients with depression (Hall-Flavin DK et al, Transl Psychiatry 2012;2:e172; Hall-Flavin DK et al, Pharmacogenet Genomics 2013;23(10):535–548). In each study, a pharmacogenetic testing report was used to guide the selection and dosing of medication for one patient cohort, but not for the other cohort. The guided group in each study had greater depression symptom improvements. However, there were methodological problems. Patients were not randomly assigned to the groups. Also, prescribers and patients in each group were not fully blinded—potentially leading to a placebo effect that could artificially improve the outcomes for those who got the testing. Moreover, although these studies were funded by Mayo Clinic research grants, most of the authors have significant financial relationships with Assurex Health, which could have further biased the outcomes.
The company subsequently funded a prospective double-blind randomized trial, comparing the use of GeneSight Psychotropic to treatment without these test results. There was a slightly greater improvement in depression scores with guided treatment, but the difference between groups was not statistically significant (Winner JG et al, Discov Med 2013;16(89):219–227). The overall likelihood of medication switches, augmentations, or dose-adjustments did not differ between groups. However, a subanalysis showed that GeneSight subjects taking a “red bin” medication at baseline were significantly more likely to have this medication changed and, afterward, had significantly improved depression scores than unguided subjects taking a “red bin” medication. Overall, not very impressive results.
Assurex Health has commercialized two other pharmacogenetics products: GeneSight ADHD (released in May 2012) and GeneSight Analgesic (released in April 2014). Using the three-bin categorization scheme described previously, GeneSight ADHD classifies eight stimulant and non-stimulant drugs used for treating ADHD and GeneSight Analgesic classifies 22 opioid and non-opioid drugs. I am unaware of any published literature on clinical outcomes associated with the use of these tests.
Larger multi-center studies of genetic testing are currently underway. Cost-effectiveness will need to be assessed, as these tests are not cheap (eg, GeneSight Psychotropic is approximately $3,800) although they are sometimes covered by insurance. Forthcoming FDA guidelines will likely encourage, if not require, the assessment of clinical validity and utility of these tests before future tests go to market.
TCPR’s Verdict: Pharmacogenetic testing is intriguing, expensive, and unlikely to be clinically useful. Until we see better evidence, buyer beware!