Personalised prescribing in psychiatry: Has pharmacogenomics delivered on its promise?

Advances in genetic technology in the past two decades promised to transform the way drug treatments are matched to individual patients. However, the anticipated shift towards genetically informed personalised prescribing for mental disorders has not yet occurred.

To understand the potential future benefits of pharmacogenomics to psychiatry, it is useful to reflect on results to date. In 2005, the United States Federal Drug Administration approved use of the AmpliChip CYP450 Test (Roche Molecular Systems, Inc.), heralding a new era of personalised drug prescribing. The AmpliChip test was designed to identify common variants in the CYP2D6 and CYP2C19 genes, which encode two cytochrome P450 (CYP) enzymes influencing hepatic drug metabolism. The test was made available for less than US$1000, bringing cutting edge genetics technology within the reach of patients in routine treatment settings. Since then, genomic technology has progressed enormously and costs have fallen dramatically. Now the main barrier to using pharmacogenetic technology is clinical utility rather than cost.

Genetic factors influencing the pharmacokinetics of common psychiatric drugs are increasingly well understood. Most psychotropic drugs are metabolised in the liver by enzymes of the CYP family. CYP2D6 (responsible for metabolising about 25% of all drugs, including several antidepressants and antipsychotics) and CYP2C19 (about 10% of drugs, including tricyclic antidepressants, selective serotonin re-uptake inhibitors [SSRIs] and serotonin–norepinephrine reuptake inhibitors [SNRIs]) are particularly important. CYP enzymes are highly polymorphic, and genetic variation contributes greatly to drug metabolism and therefore total drug exposure. Consequently, it is logical CYP polymorphisms would also influence drug response and the liability to adverse effects, although demonstrating this has been surprisingly difficult. For example, in the Sequenced Treatment Alternatives to Treat Depression (STAR*D) study, common known polymorphisms in genes relevant to citalopram metabolism, including CYP2D6, CYP2C19, CYP3A4 and CYP3A5, did not predict retention of subjects in the trial, final citalopram dosage or depression response (Peters et al., 2008).

There are several obvious reasons why identifying genetic determinants of pharmacokinetics has not greatly impacted prescribing practices for mental disorders. First, most commonly used psychotropic drugs have a wide therapeutic index, meaning increased drug exposure does not invariably lead to more adverse effects. Furthermore, many drugs with a narrow therapeutic index, including tricyclic antidepressants, are routinely monitored with plasma drug assays; this negates much of the benefit of pharmacogenetic testing. Second, for most drugs used to treat mental disorders, the correlation between drug exposure and clinical response is low. Third, many psychotropic drugs are metabolised by multiple breakdown pathways, meaning complete loss of functioning in one enzyme system may not greatly reduce drug clearance.

Despite the disappointing results noted above, individual case reports have highlighted the potential benefits of genetic testing in selected patients with unusual adverse reactions or patterns of drug response. As an example, deficient CYP2D6 and CYP2C19 enzyme activity has been reported in patients experiencing adverse effects from antidepressants while excessive CYP enzyme activity has been reported in association with inadequate clinical response (Dinama et al., 2014).

Genetic factors influencing the transport of psychotropic drugs across the blood brain barrier have also attracted attention. In particular, ABCB1, which encodes P-glycoprotein, has been closely studied in relation to antidepressant response. Although results for ABCB1 have been mixed, a recent meta-analysis identified that the ABCB1 single nucleotide polymorphism (SNP) rs2032583 predicted antidepressant treatment outcome (Breitenstein et al., 2015).

Progress has also been made with identifying genetic predictors of adverse drug effects including antipsychotic-induced weight gain and tardive dyskinesia. Variants in genes encoding human leukocyte antigen (HLA) proteins are associated with some idiosyncratic adverse drug reactions including clozapine-induced agranulocytosis and carbamazepine-induced hypersensitivity reactions. The HLA-B*1502 allele is an important predictor of serious carbamazepine hypersensitivity reactions in Han Chinese, prompting the United States Federal Drug Administration to mandate genotyping before starting carbamazepine in this ethnic population. This represents one of the first examples of a regulator requiring genetic testing as part of routine prescribing practice for a psychiatric drug. Similarly, testing to predict severe cutaneous reactions to lamotrigine is anticipated in the future.

Genetic prediction of the pharmacodynamic activity of psychiatric drugs has proved more difficult. For example, the overlap between major depression, adjustment disorders and the long-term mood effects of childhood adversity dilutes the ability to discover useful biomarkers of antidepressant response. Many other factors limiting progress in psychiatric pharmacogenetics can be identified. These include the heterogeneity of disease presentations, our nascent understanding of functional effects for most human gene variants and the as yet unquantified role of epigenetic modifications and microbiome effects (impacts of the gut microflora) on pharmacogenetic processes.

Nonetheless, there are examples of success with identifying genetic predictors of pharmacodynamic activity. For example, naltrexone has only modest efficacy as a treatment for alcohol use disorder, but there is evidence an SNP in the gene encoding the mu-opioid receptor (OPRM1 Asn40Asp) influences naltrexone response, with several studies demonstrating carriers of the minor allele Asp40 respond better than non-carriers (Chamorro et al., 2012). This will potentially allow naltrexone treatment to be targeted towards patients with most likelihood of benefit.

Despite mixed success to date, the promise of pharmacogenetics to improve clinical decision-making in psychiatry is not lost. A proportion of adverse drug reactions is likely to be predictable with prior knowledge of a patient’s genotype, as exemplified by carbamazepine and HLA-B*1502. Current data do not in most cases support routine testing prior to drug initiation, but as genome analysis becomes more widely used for other medical reasons, it will be possible to easily incorporate genetic data into prescribing decisions. In the future, sequencing of patients’ whole exome (the part of the genome encoding functional proteins) or genome is likely to become more widespread, and possibly universal in affluent countries. This genetic information will be increasingly available in treatment settings, whether a clinician has requested it or not. This clinical imperative will drive advancements in decision support systems, augmented by published expert guidelines. Notably, the Clinical Pharmacogenetics Implementation Consortium (CPIC; www.pharmgkb.org/page/cpic) is an international collaborative providing publicly available, standardised guidelines to help clinicians incorporate genetic information into prescribing decisions. Clinicians will also need robust, secure and accessible electronic health record systems in which to store and retrieve patients’ genetic information.

Progress in pharmacogenetics has not yet resulted in the widespread routine use of genetic testing to inform psychotropic drug prescribing. Nonetheless, a new era of genetically informed prescribing is approaching. There is already randomised trial evidence that this approach can lead to better treatment outcomes (Singh, 2015). The new era promises improvements in patient care via greater drug efficacy, better tolerability and prevention of serious adverse effects. A better understanding of genetic factors underpinning drug response may also provide unexpected insights into the neurobiology of mental illness.