CYP2C19 genetic testing for Mavacamten and ischaemic stroke treatment: What does the result mean for cardiovascular prescribers in the UK and Europe?

The premise

For the first time a genetic test which is known to predict effectiveness and side effects for multiple, commonly prescribed, non-specialist medications, across speciality and subspeciality domains, is being used in routine healthcare in the United Kingdom (UK) and Europe. These results, as they begin to appear in patient records, will have an impact on all cardiovascular prescribers.¹ However, most practitioners are unaware and there are no national or international strategies for managing these data.

Mavacamten

Mavacamten, a first in class myosin inhibitor, has recently been approved in the UK and European Union (EU) for the treatment of symptomatic obstructive hypertrophic cardiomyopathy.^2,3 This is the first therapy to target the underlying cause of the condition and has been eagerly anticipated by cardiologists and patients.

CYP2C19

Cytochrome P450 2C19 (CYP2C19) is a hepatic enzyme responsible for the metabolism of mavacamten as well as many other commonly used medications.⁴ CYP2C19 is encoded by the CYP2C19 gene. Common loss of function variants in the CYP2C19 gene predict a decreased metabolism phenotype. Based on clinical trial evidence, the UK regulator, the Medicines and Healthcare products Regulatory Agency and the European Union regulator, The European Medicines Agency have advised that all patients should undergo CYP2C19 genotyping before commencing mavacamten therapy.³ This is because patients with a CYP2C19 poor metabolizer phenotype (which can be predicted from the presence of two CYP2C19 loss of function variants—one from each parent) have increased mavacamten exposure leading to increased risk of systolic dysfunction, a type A adverse drug reaction (otherwise known as too much of a good thing). Therefore, regulators have recommended both a lower a starting dose and a lower maximum dose of mavacamten for genetically predicted poor metabolizers.³ This testing is now available, and the first prescriptions have been issued in the UK and in the EU.

It should be noted, however, that CYP2C19 genotype has relevance beyond the prescription of mavacamten. Common variation in this gene can influence the safety or effectiveness of several routinely prescribed medicines, including antidepressants (poor metabolizers may be at higher risk of side effects, while ultra-rapid metabolizers may have a decreased chance of responding), antiplatelet drugs (poor metabolizers will not be able to activate clopidogrel into its active metabolite and therefore have increased risk of secondary vascular events), and proton pump inhibitors (ultra-rapid metabolizers may suffer from reduced efficacy at standard doses, while poor metabolizers may have an increased risk of side effects).^4,5

There are also more complex drug–drug–gene interactions. In this scenario, a concurrently administered medication, also metabolized by CYP2C19, can interact with the genotype to alter metabolism (a phenomenon known as phenoconversion) further impacting the efficacy and side effect profile of medication. For example, in genetically predicted poor CYP2C19 metabolizer, the co-prescription of a strong CYP2C19 inhibitor (e.g. 40 mg of the proton pump inhibitor omeprazole) doesn't impact the dosing guidance, which is a low starting and maximum dose. However, for individuals who are predicted to metabolize mavacamten normally based on genotype, the co-administration of a strong CYP2C19 inhibitor has the functional effect of causing a poor metabolizer state. Therefore, the guidance in this case is to reduce the dose of mavacamten to align with the dosing recommended for poor metabolizers.

Recently, the National Institute for Health and Care Excellence in the UK has released draft guidance recommending testing for the same gene, CYP2C19, to guide antiplatelet therapy following ischaemic stroke.⁶ This is because clopidogrel is a prodrug, which must be activated by CYP2C19 in the body before it can confer its antiplatelet activity.⁷ Patients with CYP2C19 loss of function variants are known to have high on treatment platelet reactivity and higher risk of secondary ischaemic events if prescribed clopidogrel.⁸ Clopidogrel is currently the standard of care for secondary prevention of ischaemic stroke as well as being widely used in the treatment of peripheral vascular disease and ischaemic heart disease.

The problem

If a patient has CYP2C19 testing prior to commencing mavacamten, then has a stroke a year later, in theory, this result could be immediately available within their records to guide antiplatelet prescribing. Whether this will actually happen is uncertain as, at present, there is no strategy in place for the re-use of this pharmacogenetic data and there are significant informatic and implementation challenges in achieving this. Although there are technical barriers to making pharmacogenetic data available within electronic health records, these are surmountable. Several implementation programmes around the world have successfully developed pharmacogenetic services where data are made accessible to guide future prescribing activity. However, most of these have been undertaken at single institutes. Ensuring that these data are available across an entire healthcare system, which may be made up of many thousands of interrelated institutions, is a greater challenge.

The lack of a formalized strategy for the reuse of data could create geographical and socioeconomic inequities, whilst placing an unfair burden on the cardiovascular prescribers and general practitioners who will ultimately receive the results. Most importantly, it may place patients at risk of avoidable adverse drug reactions or failed prevention.

I'm not an ICC specialist—Why do I care about this?

In contrast to disease causing genetic variants, variants in genes that impact on drug response, are extremely common. Indeed, data from the UK Biobank population shows that virtually all of us will have at least one actionable variant.⁹ As such, the storage and reuse of pharmacogenetic data is fundamentally distinct from other type of genomic data, with unique technical and logistical complexities. Furthermore, pharmacogenetic data is only relevant once contextualized within a prescribing moment, making it different from someone carrying a pathogenetic disease predisposition gene. Because of this, the ethical implications of pharmacogenetic testing and the subsequent re-use of this data are not consistent with those experienced when discussing rare disease or cancer predisposition testing. It will be critical for health systems to consider the privacy and ethical issues raised by the widespread storage and reuse of pharmacogenetic data, but these should not be conflated with the ethical paradigms faced in traditional genomic medicine services.¹⁰

English prescribing data shows that nearly nine in every ten people will have been prescribed at least one of these medications which interacts with such genes by the age of 70 years.¹¹ Over 95% of these gene-drug pairs are associated with only three genes—CYP2C19, CYP2D6, and SLCO1B1¹¹. Only two of those, CYP2C19 and CYP2D6, have implications for multiple different classes of drugs across the prescribing spectrum. Thus, the emergence for the first time of CYP2C19 results in NHS and EU care records is extremely noteworthy and has significant implications that have, so far, received little attention.

Although CYP2C19 loss of function variants are frequently observed across all populations, they are more common in some ancestry groups than others. Approximately, one in every three people of European ancestry will have impaired CYP2C19 metabolism, but this rises to two in every three people of Asian ancestry and up to nine in every ten people of Oceanic ancestry.⁵ Indeed, the high frequency of CYP2C19 loss of function variants in this latter population resulted in the case of State of Hawaii vs. Bristol–Myers Squibb, with a judge initially awarding $834 million in compensation for failing to warn them that Plavix (clopidogrel) was less effective for poor responders. Therefore, the health equality implications of ‘ignoring’ this information for secondary prescribing will be significant.

Although mavacamten is only prescribed in specialist cardiology centres, it is highly likely that the CYP2C19 genotypes generated as part of the prescribing process will be viewed by other healthcare professionals in the future. This will be accelerated by testing for the same gene in patients with ischaemic stroke.

Indeed, issues of liability may arise if the data are not accessible to future prescribers. As such, an awareness that this type of result may become increasingly prevalent within a patient's record is of significant relevance to healthcare professionals.

That sounds scary, what can I do about it?

CYP2C19 genotype and predicted phenotype should be clearly labelled and stored in a standardized way within the clinical record, ideally using a consistent data-standard such as SNOMED-CT.

Patients should be supplied with their pharmacogenomic information and information about what it means and implications for prescribing, with dedicated patient facing resources.

Pharmacogenomic information should be communicated between prescribers and reported to the patient's primary care provider. Prescribers can reference international consortia guidelines [such as Clinical Pharmacogenetics Implementation Consortium (CPIC) or The Dutch Pharmacogenetics Working Group (DPWG)], and local resources (e.g. GeNotes in the UK) to support a medication review with this pharmacogenomic information.

To reduce the burden on individual professionals, healthcare systems and policy makers should prioritize developing centralized informatic solutions to ensure these pharmacogenetic results can be made accessible across a healthcare ecosystem, irrespective of electronic healthcare record platform or geographical location. Such systems should include clinical decision support to standardize the implementation of prescribing based on pharmacogenomic results.

Key messages

• CYP2C19 genotyping is now available to support the prescription of mavacamten in the UK and Europe, but the safety or effectiveness of many other commonly prescribed medicines are influenced by the same gene.

• There are not nationally or internationally agreed strategies at present considering the reuse of this CYP2C19 data for other indications, such as antidepressants and proton pump inhibitors.

• There are existing resources to support pharmacogenomic guided prescribing which have been developed by international consortia [CPIC, DPWG] and nationally based guidance, e.g. GeNotes in the UK.¹²

Acknowledgements

This work forms part of the portfolio and was funded by the National Institute for Health Research Barts Biomedical Research Centre (NIHR BRC). E.F.M. is funded by Barts Charity. J.M. is supported by the National Institute for Health and Care Research Doctoral Fellowship Award (DRF) (NIHR 301748). W.N. and J.M. are supported by the Manchester NIHR Biomedical Research Centre BRC (NIHR 203956). N.W. is supported by Innovate UK (10058536).

Conflict of interest The authors have no conflicts of interest to declare.

Data availability

No new data were generated or analysed in support of this research.

References

Author notes