Despite the hope that personalized medicine can improve care in a notoriously difficult to treat patient population, pharmacogenomic testing in the field of psychiatry is still characterized by clinicians as an emerging technology. Commercially available tests developed by startup personalized medicine companies are emerging, as are in-house tests developed by large hospitals and laboratories. But clinical experts in the field say the tests, which are a big improvement over trial-and-error medication selection, have not yet fulfilled the dreams of truly personalized psychiatric care.
“The tests available to individual practitioners really do seem to have value, but where we are at the present time is that tests are really aiding drug selection based on pharmacokinetics. Current testing cannot yet predict responsiveness, but they really are a substantial step forward,” says Daniel Hall-Flavin, M.D., an assistant professor of psychiatry at the Mayo Clinic in Rochester, Minn., who has participated in translational research efforts with pharmacogenomic firm AssureRx (Mason, Ohio).
Trials of these early versions of pharmacogenomic tests demonstrate that these assays can enhance the safety of prescribing psychotropic medications by cutting potential adverse drug reactions and can decrease the time and cost in initiating successful treatment, particularly for patients with depression. However, the genetic knowledge gained from pharmacogenomic testing is not foolproof, as genetics is an important factor but not the sole determinant of a treatment’s effectiveness.
Genetics Is Just One Consideration
Drug response phenotypes are influenced by the complex intersection of genetics and the environment, which can include such factors as diet, smoking, and other medications.
“There are so many considerations, I liken it to the New York City subway,” Hall-Flavin tells
DTET. “There are just so many stops, so many factors influencing how individuals respond to drugs. Pharmacokinetics may be an express stop, but it is just one stop. So it is important to consider, but it is not the whole picture or the whole patient.”
But advocates for pharmacogenomic testing say that it does improve care. For example, Hall-Flavin explains that when starting treatment selection with a patient with depression, there are more than 25 antidepressants to choose from. On average there is a 30 percent to 40 percent chance the patient will respond to the first medication and a two-thirds to three-quarter chance that they will respond to some medication combination. Genetic data examining the metabolic enzyme CYP2D6 show that in the general population of Caucasians, 7 percent to 8 percent will be poor metabolizers while another 7 percent to 8 percent will fall on the opposite end of the spectrum and will be rapid metabolizers. Hall-Flavin says it is in these patients on the ends of the spectrum that the pharmacokinetic information on an individual’s drug metabolism qualities makes the biggest impact. Personally, he says he generally utilizes pharmacogenomic testing in complex patients—either those who have not responded to previous treatment selections and truly have difficult to treat depression and those who are medically difficult and are taking multiple other medications, where drug-drug interactions are a concern.
Driving Adoption
“We often lose sight in medicine that these are new products in new markets. We would see similar patterns of adoption if you compare [pharmacogenomic tests] to new technology in other industries,” says Nancy Grden, the general manager at the pharmacogenomics firm Genomind (Chalfont, Pa.). “Fundamentally there are early adopters, there is a group who will never adopt it, and there is a large middle group that needs to see data and know that their colleagues are using it. This is a normal part of adoption for all new products.”
The profile of Genomind’s early adopters includes psychopharmocologists, many of whom are largely already aware of the research but didn’t have access to a commercial test. More encouraging to the company is test adoption by psychiatric nurse practitioners, who Grden says are on the front lines every day and are happy to have a new tool in their arsenal. Other common factors among early adopters of Genomind’s Genecept assay include outpatient and hospital systems with more difficult to treat patients, including those with longer lengths of stay, as well as those who see more self-pay patients.
But test developers are currently trying to penetrate that big group of psychiatrists who really need more evidence on the benefits of pharmacogenomic testing before widespread clinical adoption will take off.
Experts
DTET spoke to point to several actions that laboratories and test developers need to undertake to both have the necessary evidence these providers expect to see and to present the tests’ information in an actionable format. For one, test developers need to continue to develop a larger body of evidence, employing high-quality study methodologies. These physicians are looking for randomized, controlled trials demonstrating that pharmacogenomic testing provides clinical benefits to the patients and can make their own practice more efficient with test utilization.
Part of the exhibition of data literally includes a presentation component—the user interface for test results. Can psychiatrists easily comprehend what to do with the results? Are the results tied to actionable clinical decision support? Hall-Flavin emphasizes the importance of presenting results in an understandable way for busy physicians who may only have 10 minutes to 15 minutes with each patient. Additionally, the test results need to be communicated to patients with no genetic background, and laboratories can assist with providing that format.
To realize increased test adoption, laboratories and test developers also need to push for some larger, systemwide changes. For instance, Hall-Flavin says the broader research community needs to better “slice and dice” psychiatric phenotypes, which tend to be very heterogeneous collections of symptoms. These efforts will be enhanced by improved electronic medical record functionality including data mining using natural language processes that can better characterize responsiveness in very specific phenotypes.
Broader efforts are also needed to improve clinician education when it comes to genomics. Physicians continue to self-report perceived deficits in their knowledge of when to order genetic tests and how to interpret genetic tests, which will both prove to be barriers to wider use of molecular technologies.
Researchers from Columbia University surveyed psychiatrists to gauge familiarity and comfort with ordering and interpreting genetic tests. According to the study results, published in the April issue of the
Journal of Genetic Counseling, 14 percent of 372 responding psychiatrists had ordered genetic testing in the past six months. These orders included pharmacogenomic tests but also tests to diagnose learning disabilities and dementia. Only one-third of respondents felt confident about how to order and where to send genetic tests, and less than one-quarter of responding psychiatrists had a genetics professional to whom to refer patients.
Laboratories can play a critical role in providing continuing medical education programs and other professional development for psychiatrists. Additionally, advocating for development of clinical practice guidelines that incorporate genomic testing and the inclusion of pharmacogenomic information in electronic medical records will also aid in improving physicians’ comfort with handling these test results.
Other stakeholders, including payers, will also influence test adoption. Evidence that the cost of pharmacogenetic testing is justified by clinical outcomes is lacking, and greater efforts are needed to generate comparative effectiveness data. Not only are there the direct treatment costs, but the financial burden of treatment failures can be substantial in terms of disability claims, decreased productivity, and missed work, resulting from less than optimal therapy.
| Pharmacogenomic Biomarkers in Psychiatric Drug Labeling
Psychiatric Drugs With Warnings and Precautions/Dosage Administration |
| Drug |
Marker
|
Referenced Subgroup
|
| Aripiprazole |
CYP2D6 |
CYP2D6 poor metabolizers (dosage only) |
| Aripiprazole |
CYP2D6 |
CYP2D6 poor metabolizers (dosage only) |
| Fluoxetine |
CYP2D6 |
CYP2D6 poor metabolizers (warning and precaution only) |
| Iloperidone |
CYP2D6 |
CYP2D6 poor metabolizers (warning and precaution only) |
| Pimozide |
CYP2D6 |
CYP2D6 poor metabolizers |
| Thioridazine |
CYP2D6 |
CYP2D6 poor metabolizers (warning and precaution only) |
| Source: U.S. Food and Drug Administration |
These data are beginning to emerge. A study, which was funded by AssureRx and published in
Translational Psychiatryin March 2013, retrospectively demonstrated that inappropriate medication selection led to increased health care utilization and cost. For psychiatric patients whose medication status for a year was identified as having been suboptimal, there were estimated higher costs of $5,188 per individual.
These costs were derived from findings of roughly 68 percent more health care visits, greater than threefold more medical absence days, and greater than fourfold more disability claims patients whose medication regimen included a medication identified by the GeneSightRx panel as most problematic for that patient, based on genomic analysis, compared to all other patients.
| Pharmacogenomic Biomarkers in Psychiatric Drug Labeling
Psychiatric Drugs With Drug Interaction Labels |
| Drug |
Marker
|
Referenced Subgroup
|
| Atomoxetine |
CYP2D6 |
CYP2D6 poor metabolizers |
| Citalopram |
CYP2C19 |
CYP2C19 poor metabolizers |
| Clomipramine |
CYP2D6 |
CYP2D6 poor metabolizers |
| Clozapine |
CYP2D6 |
CYP2D6 poor metabolizers |
| Desipramine |
CYP2D6 |
CYP2D6 poor metabolizers |
| Diazepam |
CYP2C19 |
CYP2C19 poor metabolizers |
| Fluvoxamine |
CYP2D6 |
CYP2D6 poor metabolizers |
| Iloperidone |
CYP2D6 |
CYP2D6 poor metabolizers |
| Imipramine |
CYP2D6 |
CYP2D6 poor metabolizers |
| Modafinil |
CYP2D6 |
CYP2D6 poor metabolizers |
| Nefazodone |
CYP2D6 |
CYP2D6 poor metabolizers |
| Nortriptyline |
CYP2D6 |
CYP2D6 poor metabolizers |
| Paroxetine |
CYP2D6 |
CYP2D6 poor metabolizers |
| Perphenazine |
CYP2D6 |
CYP2D6 poor metabolizers |
| Risperidone |
CYP2D6 |
CYP2D6 poor metabolizers |
| Trimipramine |
CYP2D6 |
CYP2D6 poor metabolizers |
| Venlafaxine |
CYP2D6 |
CYP2D6 poor metabolizers |
| Source: U.S. Food and Drug Administration |
Additionally, Genomind presented new data at the International Society for CNS Clinical Trials and Methodology Autumn Conference (Sept. 30-Oct. 2, 2013; Philadelphia) that patients who used the Genecept Assay had significantly increased medication adherence rates as compared to controls. This adherence translated to a relative cost savings of 9.5 percent over the four-month trial period (roughly equal to an annual cost savings of $1,827 per patient).
Takeaway: Hope for personalized psychiatric care is still outpacing the current generation of pharmacogenomics tests. These tests, though, are permeating clinics, particularly for use in complex patients. Evidence is emerging that these tests can improve the safety profile of medications prescribed for psychiatric patients, but widespread adoption will be hampered if strides are not made in producing solid evidence of the outcomes and cost-effectiveness benefits of these tests.
Side Box:
Commercially Available Pharmacogenomic Assay for Psychiatry
AssureRx Health (Mason, Ohio): GeneSightRx assay
The company’s pharmacogenomic technology is based on intellectual property licensed from the Mayo Clinic and Cincinnati Children’s Hospital Medical Center. The company’s GeneSight line of laboratory-developed tests (GeneSight Psychotropic, GeneSight ADHD, GeneSight Analgesic) analyze variants affecting the metabolism and response to psychiatric medications and are intended to serve as a clinical support tool in making treatment decisions.
The company’s latest publication, a double-blind randomized control trial published in November 2013 in
Discovery Medicine, shows that patients whose treatment decisions were pharmacogenomic-informed were more than twice as likely to respond when treatment was guided by GeneSight compared to treatment as usual. This effect was most dramatically seen in the 30 percent of patients identified by the test to have severe gene-drug interactions.
The GeneSight clinical report utilizes a binning method. The patient’s genotype is determined for each of the six genes in the panel. A composite phenotype for each drug is created based on the phenotypes predicted from each of the six genotypes. The 18 most commonly prescribed antidepressants and eight antipsychotics on the panel are positioned in a green, yellow, or red “bin.”
Genomind (Chalfont, Pa.): Genecept assay
Genomind’s Genecept assay is a saliva-based panel that includes a patented algorithm for analysis of 10 genes, including pharmokinetic genes related to drug metabolism (CYP2D6, CYP2C19, CYP3A4), as well as pharmacogynamic genes tied to serotonin transporters and receptors, gated calcium channels, ankyrin G, dopamine receptors, catechol methyl transferase, and methylenetetrahydrofolate reductase.
Grden tells
DTET that the list price (self-pay with a credit card) is $650, which includes a clinician consult with a psychopharmacologist. The company also has a patient assistance plan and commercial insurance reimbursement.
In early April the company closed a $5 million round of Series A financing, led by Claritas Capital, to fund growth in commercialization of the assay.
