In recent years, a significant upsurge in the development of new laboratory tests for use in psychiatric practice has occurred. This column discusses salient questions about those tests to help mental health practitioners better understand the potential role of such tests in their clinical practices. These issues are particularly relevant for mental health professionals who prescribe psychotropic medications. The column does not provide an exhaustive list of tests nor does it review the pros and cons of any specific company’s tests.
What Has Been the History of the Development of Laboratory Tests in the Field of Psychiatry?
During my almost-40-year academic medical career, I have been interested in the incorporation of laboratory tests into psychiatry.1 This interest initially focused on therapeutic drug monitoring—which I have long believed is vastly underutilized in psychiatry—and the genetics of drug responsiveness, with an emphasis on drug metabolism. Beyond therapeutic drug monitoring, there have been many failed attempts to develop diagnostic tests, including tests to distinguish between what were postulated to be serotoninergic and noradrenergic forms of major depression in the 1970s2,3 and the dexamethasone suppression test for melancholia in the 1980s.4 Recently, a 51-analyte immunoassay test was marketed by Rules-Based Medicine Inc., as an aid in the diagnosis of schizophrenia, but the test was found to have a high false-positive rate when used in a real world population and was withdrawn from the market.5 Given this track record, caution is warranted when examining claims for new tests, and practitioners should probably take an attitude of “buyer beware” over the next few years as for profit companies endeavor to offer various tests to clinical practice based solely or principally on “analytic validity” rather than “clinical validity” and “clinical utility”—terms which are described later in this column.
What Types of Tests Are Currently Being Developed?
Most new tests are pharmacogenomic (PG)-based or immunoassay (IA)-based.
PG tests are principally based on single nucleotide polymorphisms in genes that code for pharmacokinetic mechanisms, primarily cytochrome P450 (CYP) enzymes responsible for drug metabolism and P-glycoprotein, responsible for drug transportation. The next most common type of PG tests are based on pharmacodynamic mechanisms, such as single nucleotide polymorphisms of specific receptor genes, including promoter region for serotonin (or the 5-hydroxytryptophan [5-HT] transporter [SET or 5-HTT]) or the gene for the 5-HT2A receptor.
The fact that CYP enzymes lead the list of new tests is not surprising. These enzymes and their role in the metabolism of specific drugs have been extensively studied since the late 1980s. Considerable data have been accumulated regarding variants of CYP enzymes, which result in clinically meaningful differences among individuals in their ability to metabolize drugs via these pathways. Individuals are commonly grouped into 4 phenotypic categories: ultra-rapid, extensive (or normal), intermediate, and poor metabolizers (PMs). On the basis of these phenotypes, clinical consequences can be quantified in terms of changes in drug concentration, concentration-dependent beneficial or adverse effects, and associated/recommended changes in dosing. All of this research contributes to clinical validity and utility for these tests.
Research into the role of pharmacodynamic variants, however, is still in its infancy and such variants are more difficult to measure in terms of assessing endpoints, with related limitations in knowing the clinical validity and utility of such tests.
IA assays generally measure a variety of proteins, particularly those reflecting inflammatory processes (eg, various cytokines, such as interleukin-6).6 As with pharmacodynamic measures, research into the role of inflammatory biomarkers is in the early stages. The clinical validity and utility of these tests is, therefore, less certain—witness the recent study mentioned in the first paragraph of this column.5 In some instances, even the normal or expected values in healthy controls are not yet established. Nevertheless, considerable research is being conducted in all of these areas so that new developments may lend themselves to greater clinical validity and utility.
(Note that PG biomarkers are trait measures, whereas IA biomarkers are state measures, so that complementary use of both types of tests might prove useful in diagnosis and clinical management, although such integrative use of these 2 different types of tests is generally not done today.)
What Does It Take to Market These Tests?
At a minimum, offering these tests for sale requires that the laboratory be certified by the Centers for Medicare & Medicaid Services, according to the Clinical Laboratory Improvement Amendments (CLIA) standards (http://www.fda.gov/medicaldevices/deviceregulationandguidance/ivdregulatoryassistance/ucm124105.htm). CLIA-certified laboratories are required to demonstrate the analytical validity of tests that they offer—ie, the accuracy and reliability of the test in measuring a parameter of interest—but not the clinical validity or utility of those tests. The fact that a test in fact measures what it claims to be measuring in and of itself does not mean it has clinical validity or utility (see the discussion below).
Must the US Food and Drug Administration (FDA) Approve Laboratory Tests?
No, but that situation might be changing. Currently, only tests used in a setting considered high risk—for example, a test intended to detect or diagnose a malignancy or guide its treatment—require formal approval from the FDA. In some areas (eg, oncology), the test is developed and evaluated in conjunction with the treatment and is marketed as a companion diagnostic or monitoring test to be used in conjunction with the new treatment. The approval of such a test requires submission to the FDA of clinical data supporting its clinical validity and utility, in addition to evidence of analytic validity.
Even in such cases, the degree and quality of the clinical data required may not be as high as would be required for approval of a drug, especially for tests that are developed independently after the development of a new treatment but which are still intended to be used as companion diagnostic or monitoring tests. The distinction between the data required for tests compared with drugs is understandable, given the type and quantity of data necessary for drug approval and the many years and billions of dollars it takes to accumulate such data. For most laboratory tests, providing the same level of data required to have a drug approved would be neither necessary nor feasible given the business model underlying most laboratories providing laboratory tests.
The relevance of examples from other therapeutic areas (eg, oncology) is that the same approaches will likely be used in psychiatry as the biology underlying psychiatric illnesses becomes better understood, thus making the development of truly novel treatments and companion tests possible. In other words, developments in psychiatry are likely to parallel developments in other medical areas.
What Do "Clinical Validity" and "Clinical Utility" Mean?
These are higher evidence thresholds than is needed for analytic validity, although the latter is a necessary first step on the path to achieving these higher thresholds.
Clinical validity is the ability of a test to detect one of the following:
a clinically meaningful measure, such as clinical response,
an adverse effect,
a biologically meaningful measure (eg, a drug level or a change in the electrocardiographic pattern).
Above that threshold is clinical utility, which is proof that the test can reliably be used to guide clinical management and thus meaningfully improve outcomes, such as guiding drug or dosage selection.
Is the Use of PG Testing Recommended? If So, In What Instances?
More than 30 psychotropic drugs have PG information in their labels; some of those drugs’ labels contain specific recommendations, such as obtaining PG information before selecting or starting a drug in a specific patient. An example is carbamazepine, for which the recommendation is to obtain human leukocyte antigen testing before starting the drug in patients of Han Chinese ancestry, because members of this large ethnic group are at greater risk of serious dermatologic adverse effects, including Stevens-Johnson syndrome.
In other instances, the recommendation is to do the testing before increasing beyond a specific dose. Examples of psychiatric drugs the labels of which contain such PG information are pimozide and iloperidone. (To find PG labeling in the package insert for these drugs, visit http://www.accessdata.fda.gov/scripts/cder/drugsatfda/index.cfm.) In the FDA-approved label, guidance is provided that these drugs can be started without testing if prescribed at a starting dose or dosage range. If efficacy is not achieved at this dose or dose range, the guidance recommends that testing for genetic CYP2D6 PM status be done before dosing above that starting dosage range. In other words, the recommended dosing range is higher for individuals as long as they are not genetically determined CYP2D6 PMs.
The rationale for this guidance is to reduce the risk that the patients in question (1) will achieve an excessively high plasma drug level that can cause significant prolongation of intracardiac conduction (eg, QTc prolongation) and thus (2) develop the potentially fatal arrhythmia torsades de pointes. This guidance is based on thorough QTc studies that were performed on each drug,7,8 so that the labeling exemplifies a situation in which the test has clinical validity and utility in addition to just analytical validity.
What About Data for Other Tests That Are Marketed And Promoted By Developers?
Sometimes, there are—literally—no data on available tests beyond the analytic validity of the test. Other times, the amount and quality of clinical data are quite variable, ranging from results of ≥1 small retrospective studies without controls to results of prospective, randomized, controlled studies. Even in the latter situation, the gathering and analysis of data may have been conducted solely by the developer without oversight by an independent agency, such as the FDA.
This situation raises concern that study results are not independent of the developer’s business interests and, as one might expect, leads to considerable controversy about whether the data are compelling—or not.9–12
What Is a Critical Difference Between PG Test Results and Results of Most Laboratory Tests?
PG tests are, as noted, trait rather than state characteristics. That means that the results do not change except for a phenomenon known as phenoconversion, discussed below. (Of course, advances in gene therapy might make it possible someday to change a person’s genetic makeup; for mitochondrial genes that is already possible.)
For this reason, PG test results, therefore, should not be buried in the medical record, as might happen with, say, a patient’s serial serum potassium level which is a test from a specific point in time that can change substantially from one point in time to another. Instead, PG test results need to be carried forward continuously. Results should also be given to the patient as part of his or her personal health record which can then be given to all other health care providers that the patient is currently seeing or will be seeing in the future. For the same reason, each health care provider who obtains PG tests results should consider sending the results to all of the other providers who are currently providing care for the patient.
Is Your Functional Status At a Given Moment the Same As Your Genetic Status?
No. There is a phenomenon known as phenoconversion in which a person’s functional status under specific conditions at a given point in time may be different from what would be expected on the basis of the person’s genetic status.
CYP2D6 functional status is susceptible to phenoconversion. For example, the administration of fluoxetine or paroxetine at doses of 20 or 40 mg/d converts 66% or 95%, respectively, of patients who are CYP2D6 extensive (ie, normal) metabolizers into phenocopies of people who, genetically, lack the ability to metabolize drugs via CYP2D6 (ie, genotypic CYP2D6 PMs). On the basis of a recent study of 900 participants in routine clinical care who were taking an antidepressant, 4% of the general population of the United States are genetically CYP2D6 PMs; an additional 24% are phenotypically CYP2D6 PMs because of concomitant administration of a CYP2D6 substantial inhibitor, such as bupropion, fluoxetine, paroxetine, or terbenafine.13
That is the reason a provider needs to know what drugs a patient is taking concomitantly—to consider the possibility of phenoconversion and, when necessary, to adjust the dose accordingly. On the basis of the study described above, knowing the genotype of the patient alone would be insufficient to correctly adjust the dose of a drug principally metabolized by CYP2D6 in 1 of 4 patients (ie, 24%).
What Does the Future Hold?
Developments from the human genome project and related molecular biology research have dramatically increased the ability of drug developers and subsequently clinicians to provide much more targeted and personalized treatment in other areas of medicine. This has been especially true in oncology, where laboratory testing enables clinicians to understand how specific cancers differ genetically from normal tissue and then to select treatments that target the specific genetic profile of the patient’s cancer. In the future, the development of such tests for use in psychiatric practice is likely to grow substantially, for at least 2 reasons.
There is a huge unmet need for clinically meaningful tests to aid in the provision of optimal patient care and, therefore, a tremendous business opportunity.
Knowledge of the biological basis of psychiatric disorders is growing exponentially; with that knowledge comes the ability to develop new tests.
A recent example comes from a research group that devised a test that could predict suicidality.14 Time will tell whether this test or a derivative of it enters practice. Nevertheless, it is a harbinger of likely dramatic changes in the landscape of clinical medicine, particularly psychiatry.
Given these developments, the syndromic diagnoses in DSM-5 will in the future likely be replaced by a new diagnostic schema that breaks down existing heterogenous syndromic diagnoses into pathophysiologically and etiologically meaningful entities using insights gained from genetic and biomarker data and as well as functional brain imaging. Theoretically, those insights will lead to new modalities of treatment, including somatic treatments with novel mechanisms of action, coupled to more effective psychosocial therapies—with the selection of both types of therapies guided by newly developed diagnostic tests that can also play a role in monitoring response to these more targeted treatment interventions.
During this transition from the present to the future, answers to the questions I’ve posed here about laboratory testing in psychiatry will, I hope, help the practitioner understand, evaluate, and incorporate these changes into their practices.
1. Preskorn SH, Biggs JT. Use of tricyclic antidepressant blood levels. N Engl J Med. 1978;298:166.
2. Schildkraut JJ. Biogenic amines and affective disorders. Annu Rev Med. 1974;25:333–348.
3. Maas JW. Biogenic amines and depression. Biochemical and pharmacological separation of two types of depression. Arch Gen Psychiatry. 1975;32:1357–1361.
4. Carroll BJ, Feinberg M, Greden JF, et al. A specific laboratory test for the diagnosis of melancholia. Standardization, validation, and clinical utility. Arch Gen Psychiatry. 1981;38:15–22.
5. Wehler C, Preskorn S. High false-positive rate of a putative biomarker test to aid in the diagnosis of schizophrenia. J Clin Psychiatry. 2016;77:e451–e456.
6. Savitz J, Preskorn S, Teague TK, et al. Minocycline and aspirin in the treatment of bipolar depression: a protocol for a proof-of-concept, randomised, double-blind, placebocontrolled, 2×2 clinical trial. BMJ Open. 2012;2:e000643.
7. Rogers HL, Bhattaram A, Zineh I, et al. CYP2D6 genotype information to guide pimozide treatment in adult and pediatric patients: basis for the US Food and Drug Administration’s new dosing recommendations. J Clin Psychiatry. 2012;73:1187–1190.
8. Potkin S, Preskorn S, Hochfeld M, et al. A thorough QTc study of 3 doses of iloperidone including metabolic inhibition via CYP2D6 and/or CYP3A4 inhibition and a comparison to quetiapine and ziprasidone. J Clin Psychopharmacol. 2013;33:3–10.
9. Howland RH. Pharmacogenetic testing in psychiatry: not (quite) ready for primetime. J Psychosoc Nurs Ment Health Serv. 2014;52:13–16.
10. Rosenblat JD, Lee Y, McIntyre RS. Does pharmacogenomics testing improve clinical outcomes for major depressive disorder? A systematic review of clinical trials and cost-effectiveness studies. J Clin Psychiatry. 2016. (In press).
11. Nassan M, Nicholson WT, Elliott MA, et al. Pharmacokinetic pharmacogenetic prescribing guidelines for antidepressants: a template for psychiatric precision medicine. Mayo Clin Proc. 2016;91:897–901.
12. Altar CA, Carhart JM, Allen JD, et al. Clinical validity: combinatorial pharmacogenomics predicts antidepressant responses and healthcare utilizations better than single gene phenotypes. Pharmacogenomics J. 2015;15: 443–451.
13. Preskorn S, Kane C, Lobello K, et al. Cytochrome P450 2D6 phenoconversion is common in patients being treated for depression: implications for personalized medicine. J Clin Psychiatry. 2013;74:614–621.
14. Niculescu AB, Levey DF, Phalen PL, et al. Understanding and predicting suicidality using a combined genomic and clinical risk assessment approach. Mol Psychiatry. 2015;20:1266–1285.
The content and information contained on this site is being provided as a convenience and for informational purposes only. The posting of sponsored content on this site should not be considered an endorsement or recommendation of the sponsor's products, services, policies, or procedures by the American Association of Nurse Practitioners (AANP). The information and opinions expressed on this page are those of the paid sponsors and do not necessarily reflect the view of the AANP. AANP is not responsible for the content of third-party websites linked from this page; moreover, any links on this page to third-party websites where goods or services are advertised are not endorsements or recommendations by AANP of the third-party sites, goods, or services.