Drug interaction resources: mind the gaps

Transcript

[intro music]

Welcome to the Australian Prescriber Podcast. My name is Ashlea Broomfield, your host for this episode and it's a pleasure to be speaking to Louise Grannell, who is a Senior Medicines Information Pharmacist at Alfred Health in Melbourne.

Welcome Louise.

Thank you, it's a pleasure.

Louise, what is a Senior Medicines Information Pharmacist?

Our role is to look at a variety of things. We're obviously a pharmacist that's had generally extensive clinical experience. We might get asked particular questions, a little bit more in detail, such as drug interactions. We're involved in lots of guidelines and we have our hands in lots of mini pies. There's different specialties of medicine information pharmacists at different centres and some may specialise more in pregnancy or breastfeeding or paediatrics. We're able to access a lot of different resources and really tap into those to help provide more information in the clinical context for the prescriber.

Awesome. So what sparked your interest to write an article on drug interaction resources?

The original focus of the article was actually targeting enzalutamide. Now this is a relatively new anti-androgen for metastatic castration resistant prostate cancer, so it’s generally going to be seen more in GP-land, but it's probably actually under-recognised as a really major contributor to drug interactions. We think of it like rifampicin - it’s really complex in terms of potent CYP induction amongst other things, but it also has got a very long half-life. So it's really hard to determine when exactly it will start interacting with things and again if it was needed to be stopped, when that induction potential will finish because of the long half-life.

So we started looking at the references because a lot of times patients on enzalutamide are going to be elderly patients, will also be on direct acting oral anticoagulants or DOACs and they themselves have got quite complex metabolism.

So this led to a review of all the different drug interaction resources. We just really started to tease apart where some of the gaps existed and why some of those gaps existed. And some of it is actually, it almost stems from the product information for the enzalutamide, which is a little bit conflicting in some of the component of PGP.

So that sort of sparked a review of all the different interaction resources available and tried to give some guidance to prescribers about where best to look and what the advantages and disadvantages are, and just to think big picture with some of these and not be reliant on the fact that every resource is going to have every interaction in there.

So my next question is for general practitioners, often the interactions checker that we use is the eMIMs that's linked to our prescribing software. What are some of the limitations of using that as a resource only?

Well I think the challenge is broadly with drug interaction resources and not just limited to eMIMs. There is just a lot of new drugs that actually have very complex metabolism and there's a lack of real world data for the drug combinations. So often when the drug was done in clinical trials, often drugs are excluded, all the possible combinations of drugs often weren't studied in the clinical trials.

And so when you get it being used in the real world, those combinations aren't always available. And therefore references such as eMIMS and Stockley's and other resources that we use don't always have the interactions listed for every drug. So we just need to be mindful when we're thinking about a drug interaction. If it isn't listed, we've still got to go back to that first principles of thinking. Okay, how does the drug work from a pharmacology perspective, what are the issues of pharmacokinetics? And start thinking about cytochromes and PGP and actually thinking about some of those things at the beginning and using that in combination with the resources and not necessarily being reliant on one resource. I'm looking around in multiple resources, especially for some of those newer drugs where the information isn't just readily as accessible.

And so apart from the interaction checkers linked with our software, what other resources are there available that are easy to access?

I'd highly recommend the AMH and I have no conflict of interest in recommending them, but they have relatively recently updated their online checking ability. So you can put in a couple of drugs, a couple of classes of drug, and they often then give very practical information about if an interaction exists. They also comment that although there's no documented interactions, they then will quickly provide for you that pharmacokinetic information to allow you as a clinician to go, "Okay, these ones may interact or not because of the cytochrome effects or PGP." So it allows you to start thinking about the possibilities of drug interactions.

Certainly in hospitals we also have access to other dedicated resources like Stockley's, like Lexicomp. And in the article actually I also provide lots of links to amazing free resources that are available for dedicated areas. So you've got your antiretroviral, your hep C, antifungals, your oncology. And as Dr. Geraldine Moses has also recommended there’s herbal references and one of the big ones that we use is natural medicines, which although it is a subscription, it's actually a relatively cheap subscription and that's a really great reference source as well.

What about when we're prescribing a new medication that may not have all of that real world data out there and we're not looking at an interaction between one drug but multiple drugs. We know that when anybody is on multiple drugs, it's more the additional effects that actually contribute to adverse outcomes and they can be difficult to predict. Are drug interaction resources useful for when you've got multiple drugs on board or is it usually just one drug compared to another?

Yeah, look, I think there's still quite a big gap in that particular area. I mean, fundamentally most of the key interaction checkers really do rely on one-on-one drug interactions and so yes, you've got to look and see some of the online resources such as even Stockley's and eMIMS to allow you to put in multiple drugs, but you're still looking one against the other. There is some discussion with my colleagues about a new one, YouScript, which may have a little bit more potential to do that. But I think that's still very much a work in progress and I think it's very difficult for any drug interaction resource to be able to reliably put all of that data together for the interactions and make the best suggestion of what can happen.

And also being mindful that interactions are so variable per patient. There's so many other patient factors at play. There's your pharmacogenomics, if it's a QT issue that you're worried about, it's whatever your current sort of QT ECG status may be or other comorbidities. There's so many factors. So again, even being mindful that if there was an interaction resource that actually was able to determine all of those, you still have to put everything in the context of your patient. So the short answer is no, there's not a place to go to that's going to weigh up all of these factors and give you the answer that you want. You've really got to be thinking per drug or sometimes per class of drug. Per class of drug may be good if you're thinking of what is the best azole antifungal to choose, you might want to look in that class and go, "Okay, these ones are less likely to interact."

So we're pretty safe from AI in terms of drug interaction checkers at this stage in medicine?

Look, look, there's a lot of discussion and I'm sure with the ones that exist in electronic prescribing and certainly in hospital systems there's a lot of discussions about electronic decision support, whether you'd say that's AI is a bit debatable, but I think you know certainly in the US where they've had a lot of decision support in electronic systems for a long time and I link to one of those articles in my paper that there's a big thought process or a think-tank about, well what is actually the best alerts to actually have in there. You don't want to override people and give alert fatigue and have too many interactions that aren't clinically relevant. So how you actually find tune that. That sort of work’s starting to happen as well in Australia and we’re trying to come to some consensus about what is it that prescribers really need on the ground in terms of interactions and what are the key resources to support that?

Yes, definitely. Often you can have multiple alerts per medication and you can turn on and off alerts as you would like to as well in software, and so if someone doesn't like having the alerts come up, you can actually turn them off, which is risky but also less time consuming in terms of the amount of clicks to get rid of all the alerts that come up and definitely something when I read your article that I responded to is that a lot of the alerts that come up are not necessarily clinically relevant and they're common known interactions, and so it's something that I'm taking into consideration anyway.

Something that I have noticed is things that we're aware of that the prescribing software isn't aware of, such as the triple whammy effect of ACE inhibitors, non-steroidals and diuretics, they don't often bring up an alert like that. And the other thing is the pharmacogenomics. So it's quite common, particularly in psychiatric medicine, for people to have some pharmacological genomics done, and then there's no way of putting that into our prescribing software to allow for alerts to occur. So that's something that the clinician needs to be aware of, which is very interesting to me, because you're looking at it for a specific class of medications such as psychotropics, but then it has a potential impact on all the other medications that the patient may need to be prescribed as well. And it makes prescribing a lot more complex.

Yeah, and that's definitely the case and that's why the one that I have mentioned, which is YouScript, which is a subscription-based resource, does actually tend to now provide that information based on phenotypes, so you can look and see if you were going to use these particular combinations and your patient had a pre-existing, you know, they were slower or faster status for a particular enzyme, that may give you some extra advice. But you're right, there's no perfect interaction system in any decision support that can accommodate for all those individual patient factors. But it’s certainly something of great interest, and what you were saying before that some of the alerts that come up may be quite obvious to a lot of prescribers, but then we have to think about, well what about prescribers that are newly registered medical practitioners that may not have seen all different combinations.

And so some things that are obvious to one may not be obviously up to the other. And that's also that big challenge. You want to have the best possible drug interaction database that accommodates everyone. And you might have, if you were having a debate about which ones you include and don't include that is a challenge because someone may be very familiar with those and go "oh no, that's okay.", but the next person goes "well, I'm not familiar with those." And that is the challenge that exists in building a lot of those decision support alerts.

Given that there's so many limitations of interactions checkers, there almost needs to be that final check in case it didn't get picked up. What are the kind of medications that we need to be more wary of doing that third process with?

So certainly the DOACs come up high in my list in terms of interactions. Certainly all the antifungals and all the azoles really high up there with drug interactions, transplant populations, often on lots of the immunosuppressants that are very prone for interactions. Antiretrovirals, again lots of interactions with those but also your complex oral antibiotic options. So your patients on rifampicin are prone to lots of drug interactions and even some of the other oral antibiotics, linezolid and ciprofloxacin, things that would be more commonly seen now that you just have to be thinking extra hard about - what is the potential for interactions. Certainly lots of other drugs that I should add in. There’s good old fashioned carbamazepine and some of the anti-epileptics. Look there's lots of other drugs but in terms of the recent years and the expanded use of those in community, they're ones that really sort of come to the forefront for drug interactions.

What other resources are there available to primary care providers apart from interaction checkers in terms of medicine support?

Well it's like in my article I talk about lots of free interaction resources that are available and some specialised ones that are available from hospitals. Hopefully most GPs certainly have access to a pharmacist in the community. Some pharmacists operate in GP practices that can probably sit and do that full in-depth review of all the drug interactions and consider all the options. Best practice is to always get the full drug history, and that's obviously all the prescribed medications, your over-the-counter or your complimentary medicines and just a reminder to actually consider all routes of therapy and such as inhaled and topical.

So there's been some recent cases of interactions with patients who are on fluticasone inhalers such as seretide, and patients who are then also on very potent 3A4 inhibitors such as protease inhibitors as part of an antiretroviral therapy. That combination has actually resulted in patients experiencing Cushings because the protease inhibitors increased the fluticasone. Nowadays intra-nasal fluticasone is available over-the-counter. It’s just a reminder to actually have those discussions with the patient very early on in working out what potential interactions could occur, just so that we've got that full picture of the scenario.

Once we've ceased a medication, is there anything we need to be wary of in terms of the interactions?

A lot of this is dependent on the drug's half-life. So as a general rule, if drugs have got a short half-life, they'll be cleared easily from the body, so you might not find that the impact on the interaction is prolonged. When we think specifically about drugs that actually induce cytochrome enzymes, it's really important to think that the induction effect only takes a couple of days to actually turn on, but actually can take weeks to reverse at the other end. So in the case of carbamazepine or rifampicin, you could be looking at reversal of those inductions about two to three weeks later. So be really mindful if you're starting to change doses of drugs based on induction effects, just to be really mindful that you may need some dose changes two to three weeks later after those agents are ceased.

That's unfortunately all the time that we've got for this episode. Thanks for joining us today, Louise.

Thank you. Thanks for having me.

[music]

The views of the hosts and guests on the podcasts are their own and may not represent Australian Prescriber or NPS MedicineWise. I'm Ashlea Broomfield and thanks for joining us on the Australian Prescriber Podcast.