Pharmacogenomics: three challenges to the NHS

Last week we explored the promise of pharmacogenomics in the health service, but there are good reasons why it will take time to implement

In our piece published last week, we reported on a recent event focusing on the use of pharmacogenomics. The event saw experts including England’s chief scientific officer, Professor Dame Sue Hill, outline their visions for the integration of pharmacogenomic testing into routine NHS practice as we move towards the NHS Genomic Medicine Service.

Despite its potential, however, there are several hurdles to be overcome before pharmacogenomics can become a reality.

1. A new ‘cross-pathways’ approach

Medicine has evolved into specialties along several lines, for example: organs or body systems (cardiology, endocrinology); patient age (paediatrics, geriatrics); disease type (oncology); and specific skills (anaesthesiology, radiology). Some of the most effective pharmacogenomic tests, however, span multiple specialties.

“The drug-gene pairings we’ve been looking at are broader than the specialities we’ve been looking at, even within the 100,000 Genomes Project – unlike the very targeted medicines in cancer and rare diseases,” Professor Hill said.

To give one example, the CYP2DX gene codes for an enzyme that interacts with about a quarter of all common medications. Around 8% of the population lack the enzyme entirely, meaning that they metabolise these medicines slower than average; while 2% of people have more than two copies, making them ultra-fast metabolisers.

This single gene interacts with antidepressants, antibiotics, antipsychotics, antihistamines, and medications for blood pressure, diabetes, cancer, pain management and heart conditions. It can even have obstetric and paediatric implications, as the metaboliser status of the mother can affect her child in utero or drug concentrations in breast milk.

The reality of pharmacogenomics means that a new cross-pathways approach will be required if we are to avoid duplication and unnecessary spending and to ensure patients benefit from advances in understanding. Professor Hill anticipated that by 2025 the NHS will have a “new taxonomy of medicine based on underlying drivers of disease”.

2. Joined-up patient records

In many instances, testing for single pharmacogenomic variants before prescribing medications could increase NHS costs; however, research shows that testing for a whole panel of gene variants costs a similar amount and can actually save money on average. But, crucially, this will only be the case if results from such tests can be accessed and used wherever they may be relevant.

The human leukocyte antigen (HLA) system provides a good example of the potentially wide applicability of pharmacogenomic testing. The system contains over 200 genes, variants in many of which are associated with adverse reactions (some very serious) to medications, including: carbamazepine, used in epilepsy treatment; abacavir, an antiviral used in the suppression of HIV; allopurinol for gout; and clozapine for schizophrenia.

Testing for a single HLA variant before prescribing is cost-prohibitive in many cases, but panel testing could be cost-effective because results about other variants may assist in decision-making throughout the lifetime of the patient. To get value from the test, the results need to be available to every health professional the patient sees – primary care, specialists, emergency care and so on – possibly decades after the test was done. All of those individuals will also need to be skilled in using such information in their area.

3. Prescribing outside the guidelines

Ideal doses for some patients may fall outside the guidelines issued by the National Institute of Health and Care Excellence (NICE).

A good example of this would be warfarin – a commonly used medication and the cause of the third-highest number of adverse drug reactions within the NHS according to Professor Munir Pirmohamed, who spoke at last month’s event. Warfarin sensitivity is affected by a variety of genomic variants, many of which are common. Professor Pirmohamed shared data from the PREDICT programme, led by Professor Dan Roden at Vanderbilt University in Tennessee, which looked at just two genes known to affect warfarin sensitivity. PREDICT showed that more than two-thirds of the first 10,000 patients in the study had at least one actionable variant – that is, a variant that could be taken into account when prescribing.

According to Professor Pirmohamed, when other factors such as a patient’s weight and age have been taken into account, there can be as much as a 40-fold variability in the stable dose for different patients. Frameworks are being devised for clinicians to use all the relevant factors to calculate the best dose, but, as NICE guidelines don’t currently consider pharmacogenomic factors, their conclusions could fall outside recommend guidelines, leaving them in a predicament. It is therefore vital that genomic factors are integrated into NICE guidelines as well as into clinical settings.

The Genomics Education Programme held a workshop in association with the UK Pharmacogenetics and Stratified Medicine Network in London on 22nd May 2019.

You can watch the discussions that took place here.