Psychotropic pharmacogenetics – Distraction or destiny?

Like most challenges in life, the quest for clinically translatable biomarkers in psychiatry has been arduous and frustrating. Dexamethasone suppression testing that generated enthusiastic research interest in the 1980s and had some salience in characterizing melancholia failed to translate to clinical practice. Similarly, brain imaging has yet to achieve diagnostic or prognostic use clinically. So, will genetically guided prescribing – the alluringly named field of pharmacogenetics – be just another promising technological prospect that fails to eventuate, or will it migrate from being a controversial and somewhat ‘specialised’ technology to broader mainstream adoption?

Heuristic prescribing is a significant component of everyday clinical practice. Good patient communication, education, expectation setting, clinical acumen and prescriber experience all enhance the tailoring of medications without a need for genetics. But there are still very significant variable and unaccounted-for factors involved in prescribing outcomes. This can result in a drawn-out process of trial-and-error prescribing that risks patients becoming demoralized and may reduce both their compliance and engagement with care. The challenge, then, is whether pharmacogenetics can increase the likelihood of getting a ‘hole in one’ when prescribing.

While psychiatric pharmacogenetics offers potential to facilitate the identification of the right medication or dosage, it remains controversial because of mixed findings in the limited empirical literature. The basic premise of pharmacogenetics seems sound – matching the pharmacological profile of the medication to the associated genetic profile of the patient – specifically, gene variants related to drug targets, clinical efficacy and tolerability. Critically, the pathways from the modulation of drug targets, to cascades of molecular and cellular changes, to changes in clinical presentation are yet to be fully elucidated. These vast gaps in our understanding have thus far hampered efforts to derive predictive value on the basis of genetic profiles for treatment response. It may be that the complex physiological changes that underpin treatment responses need to be represented in genetic profiles prior to showing potential applications.

Hence, it is unsurprising that the clinical utility for psychotropic pharmacogenetics has yet to be robustly determined and replicated in randomized control trials (RCTs) (Bousman and Hopwood, 2016; Peterson et al., 2011), especially given that many well-documented non-genetic factors are at play both in terms of predictors of response and risk of reported adverse events, such as the nocebo phenomenon. Consequently, evidence is needed to demonstrate superiority over the existing model of trial-and-error prescribing (Bousman and Hopwood, 2016; Peterson et al., 2011).

In the literature, there are promising new signals emerging, such as replicated gene variant associations of robust effect size and positive comparator trials of genetically guided prescribing (Bousman and Hopwood, 2016; Peterson et al., 2011). There is particular interest in the potential of such technology from third-party payers, although this is likely to add to the overall cost of treatment. However, health insurers (including governments) are actively seeking technologies to help reduce the total cost of care and improve patient (and population) health and productivity outcomes (Peterson et al., 2011). In this regard, pharmacogenetics is a prime technological prospect, given the exponential reduction in genotyping costs. Should clinical benefits be evidenced, health economic utility is likely. Remarkably, a whole genome can now be sequenced for under US$1000, and a poly-gene variant panel can be run for under US$100. These costs continue to drop, principally because of efficiencies from the automation of genotyping and DNA mass-arrays. Furthermore, as an individual’s genotype is (essentially) static over time, only one test per lifetime is needed. Thus, pharmacogenetics is very well placed to reach health economic utility if its clinical utility is adequately demonstrated.

Psychiatric pharmacogenetics is currently (as stipulated in many international guidelines) uniformly considered standard of care for just one clinical situation (Malhi et al., 2015; US Food and Drug Administration [FDA], 2017). When considering prescribing carbamazepine to patients with Asian ancestry, human leukocyte antigen (HLA) genotyping should be offered to better assay the risk of a Stevens–Johnson Syndrome (SJS) emerging (Malhi et al., 2015; FDA, 2017). Beyond this clinical situation, routine use of pharmacogenetics is not yet recommended in clinical psychiatric practice guidelines (Malhi et al., 2015), but some other academic and governmental bodies are already advocating for its mainstream adoption (Hicks et al., 2015; FDA, 2017). Nevertheless, despite enthusiasm and potential, caution is necessary, given that previous pharmacogenetic research programmes (such as that proposed for warfarin) thus far have failed to provide sufficiently compelling evidence to spur clinical adoption.

Recently, the Clinical Pharmacogenetics Implementation Consortium (CPIC) has published guidelines on the use of phase I hepatic metabolizer status genotyping (cytochrome P450 CYP2D6 and CYP2C19) to optimally dose various antidepressants (Hicks et al., 2015). The consortium comprises a non-industry funded group of international academics, but unfortunately the CPIC guidelines are not based on prospective RCTs comparing genetically guided prescribing to unguided prescribing and therefore cannot shed light on the clinical benefits of pharmacogenetics over traditional prescribing. Despite this, if CYP2D6 and CYP2C19 metabolizer status are known, FDA guidelines recommend prescribers consider such information in making prescribing decisions (FDA, 2017). This has significant clinical and medico-legal implications. If genetic information is at hand, yet not considered, adverse medication outcomes could be blamed on failure to take such information into account. What should a reasonable clinician do in such circumstances?

Given that knowledge of pharmacogenetics is increasingly available on the Internet, clinicians must be prepared for patients inquiring about pharmacogenetic testing. Other than for carbamazepine in some Asian patients, it is reasonable to advise patients that pharmacogenetics is not yet established as a valid tool in clinical psychiatry. Independently replicated RCTs are needed to draw firmer conclusions on the merits of such testing. To date, only two published RCTs (both industry sponsored) exist – both focusing on antidepressant pharmacogenetics (Bousman and Hopwood, 2016; Peterson et al., 2011). More RCTs are underway which should help shed greater light on the clinical merits of antidepressant pharmacogenetics – presently the area of keenest focus given the huge burden of disease globally from depression.

As we await further empirical clinical utility data, one can reasonably speculate that the story will be considerably more complex than gene variants alone, with environmental factors (and associated epigenetic modifications), drug–drug interactions, hepatic and renal impairments, age-related changes at the blood–brain barrier, co-morbid complexities, morbidities, placebo and nocebo reactions and medication compliance issues all important to optimal prescribing. Pharmacogenetics will – at best – be a helpful guide to optimal prescribing, at worst a distraction and unnecessary cost (Bousman and Hopwood, 2016). However, with a growing focus on ‘patient empowerment’ and ‘shared decision-making’ to enhance engagement, compliance and outcomes, pharmacogenetics continues to attract considerable community interest.

Pharmacogenetics has the potential to reduce trial-and-error prescribing and improve patient outcomes. While the area has drawn mixed feelings by the profession, it seems premature to ‘throw the baby out with the bath water’ before more informative and methodologically appropriate clinical trials are conducted and more comprehensive gene panels are developed. In the meantime, it is important for clinicians to keep abreast of this rapidly evolving area – in order to best facilitate informed discussions with their patients. If the current evidence base strengthens and clinical acceptance reaches a ‘tipping point’, genetically guided prescribing may become part of routine clinical practice. It seems to be an area that will either reach such a tipping point of evidence and adoption over the next few years or start to fade away as many ‘tests’ have done in the past. Psychiatry urgently needs new ways to improve clinical outcomes, given the accelerating burden of disease from mental illnesses. In the near future, pharmacogenetics may prove to be a cost-effective and widely useful tool to fast-track optimal prescribing and enhance patient outcomes. However, until more evidence comes to light, it is difficult to predict what will happen to this field concerned with prediction itself.