Sir Munir Pirmohamed: ‘Now is the time to implement pharmacogenomic testing into the NHS’

The evidence that pharmacogenomic testing should form a routine part of NHS care is mounting.

A report published in March 2022 by the Royal College of Physicians and British Pharmacological Society noted that a genetic cause has now been established for variation in the response to more than 40 medicines.

The forthcoming results of the PREPARE (Preemptive Pharmacogenomic Testing for Preventing Adverse Drug Reactions) study, which has looked at the power of using pre-emptive pharmacogenomic testing before prescribing commonly prescribed medicines, are expected to show a reduction in adverse drug reactions of almost a third.

Sir Munir Pirmohamed, who is UK lead for that study and also NHS chair of pharmacogenetics at the University of Liverpool, is of the opinion that the time for waiting to implement pharmacogenomic testing in the NHS is over.

In October 2022, The Pharmaceutical Journal revealed that the first-ever NHS pilot of routine genetic testing to guide drug choice in primary care will begin in England by early 2023, and similar plans are under way in Scotland and Wales. But speaking at the Royal Pharmaceutical Society (RPS’s) 2022 annual conference on 11 November 2022, Sir Pirmohamed told delegates: “You can do pilot after pilot after pilot, but actually, now is the time to implement it.”

The Pharmaceutical Journal sat down with Sir Pirmohamed after his presentation at the RPS conference to get his take on why the UK needs to make progress on implementing pharmacogenomic testing into routine patient care and what role, if any, pharmacists should play.

Some countries, including the United States and the Netherlands, have been using pharmacogenomic testing for commonly prescribed medicines for years, so why is it taking the NHS so long to introduce it?

We do need to move towards that goal. But actually, what the NHS has to its advantage is that it’s an integrated healthcare system, which means we can introduce pharmacogenomics to the whole population. This is also a disadvantage because it means it is a massive task and, obviously, the cost of doing a full integration throughout the whole healthcare system is going to be huge, but we will save money in the longer term.

What you find with pharmacogenomics in the Netherlands and the United States is that the majority of the population don’t have access to it. If the NHS is able to do it, it will be the first integrated healthcare system to introduce it at an all-country level. That is a key thing; it underlines the leadership of the UK in genomics — this is where DNA sequencing first came through, where we discovered the structure of DNA, we’ve brought in next generation sequencing. If we can bring in pharmacogenomics for our population, we’ll be the first and that will again show our leadership.

How do you think the results from the PREPARE study will influence the development of pharmacogenomic testing in the UK?

We now have enough evidence to implement pharmacogenomics. We can never have a perfect system right away — you can’t have the perfect be the enemy of the good — so we do need to implement something and then actually make sure that things are working OK and where they aren’t, we can improve it iteratively.

You can do pilot after pilot after pilot, but actually, now is the time to implement it. The PREPARE trial results provide even more evidence of a pharmacogenomics panel, where you do multiple genes, multiple variants, multiple drugs and show — in a randomised study — that you can actually produce a clinically relevant endpoint: a reduction in adverse drug reactions.

What is your view of community pharmacists offering a private pharmacogenomic testing service, as is being planned in the UK?

The private sector will always do that and I know that there are some private medical doctors who are offering pharmacogenomic testing at the moment. I don’t do any work for private medicine at all. I work for the NHS, I’m a very big supporter of the NHS, so I would like it to be implemented without exacerbating health inequalities. If you go to private pharmacies or private clinics, then you can obviously afford to do that, but the greatest burden of disease is in our deprived populations, who are the poorest populations, and that is where we can have the greatest impact.

Eventually, do you think all newborn babies will have a pharmacogenomic panel test?

There is a pilot starting here in the UK of newborn screening and whole-genome sequencing at birth. There are also newborn screening programmes being looked at in other parts of the world, including the United States.

With whole-genome sequencing there are some ethical issues, in terms of some of the disease genes that you may identify. But, for pharmacogenetics, there isn’t that much of an ethical issue. What we do need is the ability to extract information from 3 billion data points and then work out how to make it available to the healthcare system, to the parents and eventually to the person themselves when they turn 18 years old. So, there are a lot of different things to work through — that is why we need a pilot.

The pilot is ongoing, or is going to start very soon, and I’ll be chairing the committee that will look at the pharmacogenomics content in newborn screening. It’s an important pilot that will give us a lot of information, but it’s also important to engage the general public on what they feel about it.

We already do heel prick testing where we test for about eight or nine different inborn errors of metabolism; the problem is, what do you do ethically, if you find somebody has got the Huntington’s disease gene and they’re not going to get the disease until they are 45 years old. What do you tell them? And when do you tell them?

So, there are some thorny ethical issues to work through. But the pharmacogenomics itself doesn’t present an ethical issue.

What are the barriers to integrating pharmacogenomics more widely across the NHS?

The biggest is education. We need to upskill our workforce. Some people may become specialists in pharmacogenomics and have a detailed knowledge of it, but others will have too many other things to do, so they need that ‘just in time’ information to make sure that they’re doing the right thing. The pharmacy professional is a highly trained professional and having that kind of information readily available should be very easily utilised.

I think the other major barrier is technology; the decision support systems. There are decision support systems that are being developed, but in no way is there a perfect one. However, rather than reinventing the wheel and developing something from new, we should look at what the existing systems are and develop them further.

There are five ‘Is’ that are important to make sure that we get to that perfect decision support system: interpretable, intelligent, interoperable, integrated and inclusive. ‘Intelligent’ is not there at the moment, but we will get there eventually.

How do you envisage community pharmacists will be involved in pharmacogenomic testing?

The pathway for referral may come through general practice into community pharmacy, but some community pharmacists may feel that they’re able to provide it themselves, or they may be able to liaise with the GP saying: “Look, this person’s on a particular drug, and therefore it might be worthwhile doing pharmacogenomic testing on them before we put them on that drug.”

Hopefully there’s bi-directional interaction between GP practices and the community pharmacist, but also we need to think about how we can do that crosstalk with secondary care as well.

The biggest barrier at the moment is that the majority of community pharmacists don’t actually have access to the patient care record and, if you don’t, then how do you actually do that?

Will pain management be a potential area in which pharmacogenomics could be useful?

Absolutely. There’s a lot of work going on in there. Particularly with codeine and cytochrome P450 2D6, in the mu-opioid receptor gene with morphine and some of the neuropathic pains as well — looking at amitriptyline, gabapentin, pregabalin and carbamazepine.

Pharmacogenomics will come into every area of medicine.

In the PREPARE study, there were about 40 gene pairs or so that we can say there’s enough evidence. What I want to do is to be able to put in a lot of different variants and say we can do all of this, but sometimes the evidence base isn’t good enough and what I don’t want to do is to over-promise and under-deliver, I’d rather under-promise and over-deliver.

In antimicrobial resistance, would it be useful to test which antibiotics would be most effective?

Absolutely. For example, voriconazole, which is used as an antifungal — you need to be able to get the right dose, etc., without causing toxicity. So, there are potential areas for antibiotics, antifungals, antivirals; abacavir is a typical example, not really related to the viral resistance, but related to safety.

What do you hope to see happen in the field of pharmacogenomics over the next five to ten years?

I would hope that we implement pharmacogenomics, we continuously improve the implementation by introducing new variants and identifying new drug–gene pair associations, refining what we give in terms of advice, making sure that we’re doing the right things and improving the healthcare system for outcomes.

I would also hope the pharmaceutical industry takes account of pharmacogenomics. At the moment, there’s a lot of disconnect. For example, if you look at tamoxifen, it says “avoid drugs which inhibit cytochrome P2D6” because it stops the metabolism of tamoxifen to endoxifen, it’s active metabolite. However, it just says it may affect poor metabolisers, it doesn’t tell you to genotype people before you put them on it.

So, there’s many areas of disconnect like that in some of the drug labels. Siponimod is another example, which is used in multiple sclerosis. It says genotype people for cytochrome P450 2C9, and then if you’ve got star (*) 2, *3 variants, you need to reduce the dose. *2 and *3 variants are only found in European populations. Multiple sclerosis is as common in African populations as it is in Caucasian populations, but we don’t give any advice on *5, *8, *11, which appear in the African population. All that needs to be sorted out.

However, pharmacogenomics has got a major role in identifying new drug targets. People have already shown that if you have got a pharmacogenomically-defined drug target, your chances of success of developing a new drug is two-times higher. If you consider that a drug costs US$1.5bn to develop, a two-fold increase in success rate is actually a massive saving.

But you can also use it in terms of safety as well, not only the examples I’ve given, but more generally as well. It’s about how we can utilise the genomics information to make drug therapy overall safer.