What is needed to incorporate clinical pharmacogenetic tests into the practice of psychopharmacotherapy?

Abstract

This editorial considers two questions in psychopharmacotherapy: 1) What is needed to market pharmacogenetic tests in the US, since the US appears to lead other countries? and 2) What is needed for US-marketed pharmacogenetic tests to be incorporated by prescribers into long-term practice? US marketing of pharmacogenetic tests requires 1) understanding the pharmacological complexity of drug response, 2) modifying the oversight of non-FDA regulatory agencies, 3) clarifying the FDA’s role and 4) promoting innovative marketing. The incorporation of pharmacogenetic tests into long-term practice requires 1) not jeopardizing pharmacogenetic testing by short-sighted marketing of non-validated tests, 2) educating prescribers about benefits, 3) educating patients about limitations and 4) considering the differences between isolated testing and generalized testing incorporating big data.

KEYWORDS:

• Biological markers
• cytochrome P-450 enzyme system/genetics
• genotype
• histocompatibility testing
• pharmacogenetics
• psychotropic drugs/administration and dosages
• United States Food and Drug Administration

An oncologist may answer the question, “What is needed to incorporate clinical pharmacogenetics into clinical practice?” with “some time.” In contrast, pharmacogenetic tests face major obstacles in psychopharmacotherapy. In 2008,[1] four pharmacogenetic psychiatric tests had peer-reviewed articles. Two are rarely used and two are no longer available. Paradoxically, pharmacogenetic tests based on pharmacodynamic genes with limited data on clinical utility are being used by US psychiatrists [2,3] after aggressive marketing. An optimistic oncologist may see the glass half full,[4] whereas this editorial reflects the glass half empty.

The question is separated into two: “What is needed to market pharmacogenetic tests in the US, since the US appears to lead other countries?” and then “What is needed for US-marketed pharmacogenetic tests to be incorporated by prescribers into their long-term practice?” US marketing of pharmacogenetic testing requires understanding the pharmacological complexity of drug response, modification of the oversight by non-FDA regulatory agencies, clarification of the FDA’s role and promotion of innovative ways of marketing. Long-term clinical use of pharmacogenetic testing requires not jeopardizing pharmacogenetic testing by short-sighted marketing of non-validated tests, educating prescribers about its benefits, educating patients about its limitations and considering differences between isolated testing and generalized standard pharmacogenetic testing incorporating big data.

Understanding the complexity of drug response

Pharmacogenetic testing can focus on disease and/or pharmacological mechanisms.[5] Oncology is rapidly progressing in the knowledge of disease mechanisms.[4] However, most psychiatric disorders are not diseases but syndromes, making disease biomarker validation impossible.[6] Even for Alzheimer’s disease, brain mechanisms appear to be pretty hard to “crack”.[5] Consequently, complexity in psychiatry is “reduced” to addressing pharmacological mechanisms. “Reduced” is a euphemism, as pharmacological mechanisms include pharmacokinetic and pharmacodynamic mechanisms, and efficacy versus safety mechanisms, including two extreme types of safety mechanisms (idiosyncratic vs. dose-dependent). Moreover, genetic, environmental and personal factors influence all these mechanisms; one can personalize drug selection and/or dosing; and in drugs with narrow therapeutic windows, dosing is mainly influenced by pharmacokinetic mechanisms, while in drugs with wide therapeutic windows dosing is mainly controlled by prescribers.[7,8] Genetics is only one of many variables adding to the complexity of drug response. Senn, a statistician with expertise in randomized controlled trials (RCTs), explained that, despite the hype, very limited data support the hypothesis that genetics controls drug response.[9] Every drug may be different.[5] For some drugs, genetics may be extremely important in most patients, while in others it may be irrelevant when compared with environmental factors. Genetic relevance may also vary within each drug; genetic testing may be crucial for understanding some rare patients’ drug response (statisticians call them “outliers”) but irrelevant for average patients.[10] In psychiatry, two types of pharmacogenetic tests are ready for clinical practice [8,11]: HLA testing to rule out carbamazepine for some Asian patients, and CYP2D6 and CYP2C19 testing for some antidepressants and antipsychotics, mostly for dosing. As therapeutic drug monitoring (TDM) identifies phenoconversion due to inhibitors/inducers, it is crucial to conjointly implement TDM and pharmacogenetic testing in order to advance personalized dosing.[8,11]

Modifying oversight by non-FDA regulatory agencies

Clinical Laboratory Improvement Amendments (CLIA) regulate quality standards for US clinical laboratories.[3,12] Accreditation by the College of American Pathologists, the Joint Commission on Accreditation of Healthcare Organizations or state health departments assures full compliance with CLIA regulations and allows any laboratory to legally offer pharmacogenetic testing. Advisory governmental committees propose three aspects of pharmacogenetic test evaluation: analytic validity, clinical validity and clinical utility.[12] CLIA regulations focus only on the most basic aspects of analytic validity.

Clarifying the FDA’s role

In the 1990s, the FDA started pushing for increased pharmacogenetic testing [13] despite their role as a regulator only of drug marketing, not of diagnostic test marketing. In a Me-Too drug market, the reluctant pharmaceutical companies eliminated CYP2D6 drugs from their pipelines.[14] The FDA provided guidelines for pharmacogenetic testing in 2005 [14] and a draft for laboratory testing in 2014.[15]

In the 1990s, Affymetrix (Santa Clara, CA, USA) developed the GeneChip by putting DNA chains in computer microchips, producing the first DNA microarray and the first system of parallel genetic testing that simultaneously could test thousands of genetic variations. Pharmacogenomics (testing for many genes) and all other “Omics” were born. After Affymetrix marketed the first GeneChip for HIV genotyping, it developed one for CYP2D6 and CYP2C19 testing, which appeared to be cost-effective in a 1996 pilot study when first-generation psychiatric drugs were used.[16] Roche Diagnostics (Indianapolis, IN, USA) took over the marketing.[14] Some small companies developed other parallel systems of genetic testing and US labs started offering CYP2D6 and CYP2C19 testing.[1] Roche Diagnostics decided to ask for FDA approval, which was given in 2005. However, the test was too expensive and approved too late; the second-generation drugs not so dependent on CYP2D6 and CYP2C19 had come to dominate the psychiatric market. Moreover, psychiatrists did not understand how to use it.

Two examples of the FDA’s contribution to pharmacogenetic testing are atomoxetine (negative) and carbamazepine (positive). The pharmaceutical company completed RCTs for attention deficit hyperactivity disorder using atomoxetine, a non-Me-Too CYP2D6 drug. Pharmacokinetic studies indicated that it was a wide-therapeutic-window drug, but the FDA pushed for safety studies on CYP2D6 poor metabolizers, which were negative. The FDA approved atomoxetine with no need for CYP2D6 genotyping, but the approved dose may be too low for up to one-third of US patients.[17] After a small Taiwanese carbamazepine study demonstrated the presence of HLA B*15:02 had very high positive and negative prediction in Chinese Han,[18] the FDA called, in 2007, for HLA-B*15:02 screening in Asian patients before starting carbamazepine. Later, a large mirror-image Taiwanese study [19] demonstrated that the FDA was right.[8]

Promoting innovative ways of marketing pharmacogenetic tests may be helped by new legislation

In the US, CLIA regulations focusing on minor aspects of laboratory validity allow the marketing of tests with no proven clinical validity/utility, while the FDA has made an effort to promote pharmacogenetic testing with mixed results. As rapid technological advances are sharply reducing the costs of genetic analysis, one would think that the paralyzed US legislators should find a way of developing an intermediate means of approval using the basic concepts of analytic validity, clinical validity and clinical utility [12,15] without the FDA’s excessive approval costs. The large prospective RCTs needed to establish the classic proof of concept and the cost–benefit ratio of pharmacogenomics in psychiatry will not occur due to very high costs and lack of a funding mechanism.[8] However, proof-of-concept studies with comparisons to historical data (like the Taiwanese carbamazepine study) can be used to approve pharmacogenomic tests in specific contexts.[8]

Not jeopardizing pharmacogenetic testing by short-sighted marketing of non-validated tests

Psychiatrists in the US and a few European countries are being offered genotyping for multiple CYPs and pharmacodynamic genes, with no, or very limited, data on clinical validity/utility.[1,3,7] These tests recommend that clinicians use or avoid some psychiatric drugs. On the other hand, published guidelines from reputable organizations plus the authors’ research and clinical experience suggest that only CYP2D6 and CYP2C19 genotyping can help, and only with a very small number of drugs in psychiatry.[11] Testing for other CYPs and receptor/transporter genes should not be used in clinical practice, and may jeopardize the starting field of pharmacogenetic testing in psychiatry.

Educating prescribers about the benefits

Educating prescribers in how to use pharmacogenetic tests for drug selection/dosing in the context of other factors such as environmental and personal factors and pharmacological mechanisms is a daunting task, due to limited pharmacological training in medical schools.

Educating patients about the limitations

Although patients with outlier drug responses and their families are more open to new pharmacogenetic testing than clinicians, they may be unhappy when pharmacogenetic testing is unable to explain unusual drug responses. Sometimes this is due to lack of scientific knowledge, but frequently it is due to unusual psychological responses to medications.

Considering the differences between isolated testing and generalized standard pharmacogenetic testing incorporating big data

It is possible to train psychiatrists to order CYP testing in patients with unusual lack of efficacy, adverse drug reactions or unusual TDM. But it is much more complicated to teach them how to interpret the unregulated CYP tests offered in the US. Oncology appears to be getting to a point that computer programs designed by scientists may soon personalize drug selection according to the specific molecular signature of the patient’s cancer.[4] If follow-up big data is added, this may lead to true personalized prescription in oncology.

The complexity of psychiatric disorders has defeated pharmaceutical companies, leading to their abandonment of psychopharmacology.[20] Ignoring disease mechanisms and only focusing on pharmacological mechanisms requires considering multiple variables in drug selection and dosing: genetic, environmental (sometimes acting through epigenetic mechanisms) and personal factors that influence pharmacodynamic and pharmacokinetic mechanisms and TDM.

Financial & competing interests disclosure

No commercial organizations had any role in the completion or publication of this study. This article was completed without any external funding. J de Leon personally develops his presentations for lecturing, has never lectured using any pharmaceutical or pharmacogenetic company presentation, and has never been a consultant for pharmacogenetic or pharmaceutical companies. In the past, J de Leon has received researcher-initiated grants from Eli Lilly (one ended in 2003 and the other, as co-investigator, ended in 2007), from Roche Molecular Systems, Inc. (ended in 2007), and, in a collaboration with Genomas, Inc., from the NIH Small Business Innovation Research program (ended in 2010). He was on the advisory boards of Bristol-Myers Squibb (2003/04) and AstraZeneca (2003). Roche Molecular Systems supported one of his educational presentations, which was published in a peer-reviewed journal (2005). His lectures have been supported once by Sandoz (1997), twice by Lundbeck (1999 and 1999), twice by Pfizer (2001 and 2001), three times by Eli Lilly (2003, 2006, and 2006), twice by Janssen (2000 and 2006), once by Bristol-Myers Squibb (2006), and seven times by Roche Molecular Systems, Inc. (once in 2005 and six times in 2006). E Spina has participated in speakers/advisory boards and lectures supported by AstraZeneca, Bristol-Myers, Eli Lilly & Co., Janssen Pharmaceuticals, Lundbeck and Pfizer. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed. Editing assistance was provided by Lorraine Maw at the Mental Health Research Center at Eastern State Hospital, Lexington, KY, USA.