Ep154: Delabelling penicillin allergy FAST

Allergies to first-line antibiotics complicate care, and use of broad-spectrum alternatives is associated with prolonged hospital stays and increased rates of admissions to ICU. It also makes emergence of antimicrobial-resistant pathogens more likely. Ten to 16% of adult patients are marked in the record as being allergic to penicillins but the majority of these labels are applied inaccurately or have become outdated over time. 

While skin-prick testing is a safe and reliable way to delabel patients who can tolerate penicillin, it is resource-intensive and requires trained allergists on-site. In this podcast we describe a decision tool called PEN-FAST that helps to quickly identify patients suitable for direct oral challenge in almost any setting. Guests include two of the leading academics behind the validation and scale-up of this algorithm. 

Guests
Prof Jason Trubiano FRACP FAAAAI‍ (Austin Health, Director of Infectious Diseases; Peter Doherty Institute for Infection and Immunity)
Prof Suran Fernando FRACP FRCPA PhD OAM (recent!) (Royal North Shore Hospital; University of Sydney)
Dr Fionnuala Fagan FRACP FRCPA (Dunedin Hospital; Otago University) 

Chapters
10:42   PEN-FAST tool
22:23   Multi-site testing of PEN-FAST
28:27   Non allergist-led testing
32:42   System-wide outcomes

Production
Production by Mic Cavazzini DPhil. Music licenced from Epidemic Sound includes ‘The Bayou’ and ‘Feels like I’m going Crazy’ by Tigerblood Jewel. Feedback on this episode kindly provided by Dr Marion Leighton, Dr Robin Sia, Dr Aidan Tan, Dr Hugh Murray, Arnika Martus and Paul Cooper PhD. Image courtesy of Thirteen of Clubs at Flickr.

Visit the Pomegranate Health web page for a transcript and supporting references. Add educational activity to MyCPD. Subscribe through any podcasting app or our email alerts list. 

Transcript

Copy Transcript Copied

MIC CAVAZZINI: Welcome to Pomegranate Health, a podcast about the culture of medicine. I’m Mic Cavazzini for the Royal Australasian College of Physicians.

Allergies to first-line antibiotics complicate care because the alternatives are not as effective or well targeted. Use of broad-spectrum antimicrobials is associated with persistent reinfection, prolonged hospital stays, increased rates of admissions to ICU and mortality. 10 to 16% of adult patients are marked in the record as being allergic to penicillins, but as we’ll hear, the majority of these labels are applied inaccurately or have become outdated over time.

Just a quick refresher from med-school. An allergy to penicillins can cause anaphylaxis within an hour or two of injestion through a reaction mediated by IgE antibodies. Such aggressive Type I hypersensitivity occurs in only around 7% of labelled patients, however, and under 0.05% of the population. Angioedema also quite rare. Most presentations will include hypotension, wheezing, urticaria or hives.

Dermal testing is a well-recognised way to assess the presence of an ongoing Type I hypersensitivity. The skin is pricked with a droplet of allergen and if specific IgE antibodies are present these activate mast cells to release histamine. It is histamine that generates the itchy lump on the skin, and the size of the swelling correlates roughly to the amount of specific antibody in the system.

Type IV hypersensitivity to penicillin can not be assessed by skin prick testing. It involves a range of severe reactions that take days to manifest as they result from T-cell mediated cytokine release. We’ll talk about these later. Even where it is called for, dermal testing is a drain on time and staff resources. Direct oral challenge is a more straightforward and efficient, as long as probability of a severe reaction is low. In this podcast we describe a decision tool called PEN-FAST that helps to safely identify patients suitable for direct oral challenge.

Today’s story came to me from a haematologist on my review group, who was shocked that she and many colleagues had never been introduced to PEN-FAST. We were lucky to track down the tool’s lead developer, Professor Jason Trubiano in Melbourne. And another leader in this field of work is Professor Suran Fernando, from whose office I recorded this interview.

SURAN FERNANDO:        I'm Professor Suran Fernando. I'm an immunologist and the head of clinical immunology and also the director immunopathology and here at Royal North Shore and I'm also affiliated with the University of Sydney.

MIC CAVAZZINI:                Moving proximo-distally to Melbourne.

JASON TRUBIANO:          I'm Professor Jason Trubiano. I'm an infectious disease physician. I'm the Director of Infectious Diseases and Immunology at Austin Health and the head of infectious diseases at the University of Melbourne.

MIC CAVAZZINI:                Thank you, and all the way over the Tasman….

FIONNUALA FAGAN:       Hi, I'm Fionnuala Fagan. I'm a clinical haematologist, based in Dunedin in New Zealand and I have affiliations with the Otago University.

MIC CAVAZZINI:                Okay, so, we’ve long known that allergies to first-line antibiotics complicate care. When clinicians are forced to use broad-spectrum antibiotics instead, this is associated with prolonged hospital stays and increased admissions to ICU. Suran, can you briefly explain why that is?

SURAN FERNANDO:        It occurs first because when you use broad spectrum antibiotics, you're also wiping out a lot of the intestinal flora and other flora that allows for the emergence of drug-resistant organisms. And studies have shown that the use of those broad-spectrum antibiotics are associated with the emergence of organisms like MRSA, mancomycin-resistant enterococcus. So that's why, you know, we try to use narrow spectrum antibiotics.

MIC CAVAZZINI:                In an NHS guideline the example is presented of a penicillin-allergic patient who has acquired in hospital a severe pneumonia. They suggest that drugs outside the beta lactam class are indicated, a combination of vancomycin, metronidazole and gentamicin. What are the knock-on risks of this as compared to the first line therapy?

SURAN FERNANDO:        Patients with a bona fide Type 1 allergic response do not necessarily have to avoid beta lactams altogether. In Australia the first-line antibiotic therapy in patients who present to hospital with severe pneumonia is usually ceftriaxone and azithromycin.

I mean, there's sufficient evidence to now consider prescribing cephalosporins in patients with a Type 1 penicillin allergy. And the cross reactivity is primarily determined by the R1 side chain you know in the structure, as opposed to the beta-lactam ring. So cross reactivity rates are about 1 to 2%. So, for instance, amoxicillin and ampicillin are structurally similar with respect to the side chain with cephalexin. And these cephalosporins, should not be prescribed, but other cephalosporins such as cefazolin and ceftriaxone may be given.

Once the patient has improved, then oral therapy with amoxicillin is usually indicated. And that's where there's an ideal opportunity to address the penicillin allergy label. As if it is low risk, then the label can be removed and the amoxicillin therapy implemented. Otherwise, patients are administered a broad-spectrum cephalosporin or a fluoroquinoline, which increases the risk of drug resistant organisms such as MRSA. And also, if we can delabel, we can also clear the infection faster and reduce the risk of readmission

JASON TRUBIANO:          Yeah and, you know, some of the drugs you mentioned, like ciprofloxin, vancomycin—if you apply those drugs to any bacteria for a prolonged period of time, they will develop resistance to those, exactly like Suran's saying. So, penicillins for pneumonia—some people still use benzyl penicillin, streptococcal infections, syphilis, they still require the most narrow spectrum penicillin that we have. So, by removing somebody's penicillin allergy, we're able them to use those first-line, front-line, narrow spectrum antibiotics, and then prevent antimicrobial resistance. AMR or superbugs, whatever you like to call it.

MIC CAVAZZINI:  So, depending on the health system, 10 to 16% of adults are labelled with penicillin allergy, but in most cases we actually don’t know how serious or how accurate that label is and the real prevalence is likely much lower. Jason, before we get into the actual numbers, tell us why labels might get applied in a slapdash way or might no longer be accurate.

JASON TRUBIANO:          I get in trouble for this, because my mum told me off when she listened to a podcast previously where I said mothers are not always right. Because the vast majority of allergies to penicillins are acquired in childhood and nobody has a clue why they're applied. Most of our studies demonstrate that 40% of people applied with a penicillin allergy will be from childhood. And that will either be, “I don't know what it was, I had a rash during antibiotic therapy” that probably was a result of a viral exanthem at the same time. They had a non-immune-mediated reaction: nausea, vomiting, diarrhea, headache, dizziness. We've had people that have been applied because their mother, brother, or sister was also allergic, and then it was applied to them. So, there are a range of reasons why an allergy label may be applied, but most majority of them stem from childhood or a remote history. I don't know, Suran, would you agree with that?

MIC CAVAZZINI: And tell us briefly—we talked about IgE mediated responses, but there's this whole other area class type B reactions that are mediated by T cells. Suran, can you describe some of these delayed reactions and explain whether they are a similar veto for use of penicillin.

SURAN FERNANDO: Yeah, that's a lecture in itself. That's a one-hour discussion. But delayed T-cell mediated reactions can be broadly categorized for all intents and purposes into two groups. The first is the benign type of reaction that typically occurs after six hours or usually 24 hours without any features of anaphylaxis or severe cutaneous adverse reactions. And this type of mild reaction occurs in up to 12% of cases, during a course of a beta-lactam antibiotic.

So, this group of patients can have another penicillin or be delabelled to the penicillin in question as the T-cell response to the penicillin does tend to wane over time. But the second group, on the other hand, are potentially life-threatening—Severe Cutaneous Adverse Reactions or SCAR, and these are fortunately rare. The most severe of this is Stevens-Johnson syndrome, toxic epidermal necrolysis, and that affects about 1 to 10 individuals per million people per year. And that causes blistering sort of skin reactions and detachment and severe ocular and oral and genital involvement with a myriad of ensuing complications. And often these patients end up in the burns unit. From a couple of single-site studies in Australia, penicillins and cephalosporins are the implicated agent in about 25% of cases. The figure is currently a little less in the AuSCAR registry, which is a prospective multi-centre registry of SCARs in Australia that Jason had.

The second one is DRESS, which stands for Drug Reaction Eosinophilic Systemic reactions. And that causes a skin rash and inflammation in internal organs such as the kidney and liver, and that affects about 2 to 10 million people per year. Penicillins and cephalosporons are implicated in about a quarter to third of those reactions. And AGEP or Acute Generalized Exanthemus Pustulosis is the mildest of these reactions. And that causes the eruption of these new numerous tiny pustules and affects about 3 to 5 million individuals per million people per year. And ninety percent of those cases are due to drugs and penicillin is the most common cause.

So, SCARS are rare, but they do produce a very strong memory T cell response. And the cross reactivity between penicillin and cephalosporins for SCARS are not fully predictable. And the clinical guidelines would recommend avoiding these classes of antibiotics due to the risk of an even more severe reaction upon re-exposure. And

MIC CAVAZZINI: And I guess it's less likely that those would be mislabelled in the first place.

JASON TRUBIANO:          People remember them. It’s hard to it's hard to forget ending up at a burns unit with a drug reaction.

FIONNUALA FAGAN:       I will I will say I gave someone allopurinol, bendamustine and cotrimoxazole all at the same time and they ended up in ICU, and I was like, I don’t know which one of those gave the rash?

JASON TRUBIANO:          This is a this is every week clinic referral we get.

SURAN FERNANDO:        I know. Same here. And quite challenging to teach out.

JASON TRUBIANO:          Or lenalidomide is the other one.

SURAN FERNANDO:        Yeah.

PEN-FAST tool

FIONNUALA FAGAN:       Yeah, so clearly it's really important to allow penicillin to be given to as an option for patients. And as someone who gives a lot of antibiotics I'm very heavily invested in a microbial stewardship and being able to use those antibiotics in the future. So, the really important thing is for us to label those patients properly and to unlabel people who are no longer allergic.

Suran, you wrote a paper in 2019 with your colleagues at Royal North Short that suggested that 80% of patients with a type A reaction could be delabelled after the interview. Now I remember in med school that the classical way of establishing an allergy was  the skin prick and epidermal testing for reactivity. But because this can produce false negatives, then it has to be followed by a direct or oral challenge. But this is incredibly time-heavy, it requires specialist supervision and specialist training. So, what does your research show about the resource-intensiveness of this protocol?

SURAN FERNANDO:        Look, in terms of time, to evaluate somebody according to that existing protocol, we would have spent up to three and a half hours in the clinic doing the assessment, doing the skin testing and the incremental challenge. And then of course the follow-up with a three-day course, you know, with check-ins the day after and at the end of the three day course and even at five days. So, there was a huge cost involved there.

Whereas we could just simplify it, you know, by using our own algorithm identifying somebody as having low risk and go straight to a direct oral challenge for one hour. And we showed that we could reduce length of stay here at Royal North Shore by six days. And we also projected savings of about $1.75 million if we just delabelled one person per weekday, over the course of a year. And that was like 2020 figures.

In that study, we were particularly interested in patients admitted to hospital with an infection and a penicillin allergy label due to three different causes. 1) history of a type A reaction such as gastrointestinal symptoms, 2) a mild low-risk immune-mediated reaction, such as a rash, 3) or if the history of the reaction was unknown.

And we rigorously matched these patients to controls that were age, sex-matched, infection-matched as well, the type of infection, as well as the timing. So, our protocol involved patients initially having skin testing, followed by an incremental three-dose challenge, followed by a three-day course of a penicillin. And we found that 96% of all patients were delabelled to penicillin with only two patients having a benign rash. And of the 63 patients that underwent skin testing, only one had a positive skin test. And that patient had no prior exposure to amoxicillin despite having a positive skin test to amoxicillin. So, we challenged them anyway, and they tolerated a three-day course.

So, we suggested a simplified protocol which included direct delabelling for these type A reactions, you know, such as a history of gastrointestinal reactions or a headache or thrush or if they have a family history. And a direct oral challenge with a single dose for low-risk penicillin allergy labels. And we published some of that data in subsequent papers and, of course, now we have excellent validated clinical assessment tools to assist in risk stratification and proceeding to a direct oral challenge if the risk is low.

MIC CAVAZZINI:                Yeah that paper was one amongst a field that was sort of validating this idea of direct oral challenge. And such findings have fed into a risk algorithm for oral challenge called PEN-FAST. Jason, can you describe what scoring criteria are included in PEN-FAST and what was found from the first few validation studies (in general patients and pregnant patients] about where useful threshold for safety should sit. 

JASON TRUBIANO:          I think the key thing we're all trying to discuss is how do we identify a patient in front of us that has a penicillin allergy that is low risk that we can just offer a direct oral challenge, which is a single dose of an antibiotic without the traditional skin testing.

And actually, PEN-FAST was derived and born out of the traditional method. We were skin testing everybody that came to see us in clinic, and we then collected a cohort of 622 patients that we were able to determine what the positive rate was and therefore what the predictors were of having a true allergy. And we came up with the variables that made nicely into a catchy PEN-FAST tool.

So the PEN in it is describing anyone that reports a penicillin allergy. This could be benzyl penicillin, penicillin VK, amoxicillin. Whatever penicillin it would be, the rule applies to. The other initials are the F, which is for Five. It’s been five years or less since they’ve had the reaction. If it’s that time frame, they get two points.

The next variable is the A. The A stands for stands for anaphylaxis or angioedema. They get two points if they had any one of those two phenotypes. The other way you can get two points is for the S. For Severe Adverse Cutaneous Reaction. So you’ll get two points for any SCAR or a rash that may resembles that, blistering or desquamating or something that involves the mucosal membrane.

And the last one is T for Treatment. And we give one point if they were given treatment or treatment was unknown. So you can get a score anywhere from zero to five. Five being high risk, likelihood of being a positive reaction to zero, bugger all risk of having a true penicillin allergy.

And then we found that if you had a score of less than three, there was a ninety-six percent chance that you had no likelihood of having a penicillin allergy. And we got together with some friends in Royal North Shore, in WA and in the USA to validate those three points. And it came out exactly the same. Very high negative predicted value, high sensitivity. So great as a rule out test.

It's now been validated in 15 separate studies in 21 countries to really show that anybody can take those key clinical variables that has learnt how to take a proper history and get this score. And then you would offer that person a direct oral challenge rather than the traditional skin testing approach. And that's why I think PEN-FAST has been readily taken up by a lot of clinicians.

FIONNUALA FAGAN: So, Jason, in 2023 in JAMA Internal Medicine, there was a randomized controlled trial called PALACE that you led. That study had 380 patients in it, so it wasn’t much bigger than the other validation studies. Those patients were in Australia, at the Austin, the Peter Mac and The Royal Melbourne Hospital. It’s always interesting to hear theory translated into practice, so quickly tell us what the randomised controlled trial proved on top of thise validation studies that had come before it?

JASON TRUBIANO:          Yeah. Up until PALACE, we'd had one underpowered randomized control trial done in the US, and we had PEN-FAST and other observational data to say yes, direct oral challenge is safe. And if you pick a low-risk phenotype, it's safe. And Suran and I would have been doing that in our practice. But really, to move the needle, most of the time you need the highest level of evidence, and at this point is a randomized control trial.

So we said, okay, well, why don’t we just randomize people. You can randomize them to get the traditional skin testing, then followed by the oral challenge, which is the gold standard, versus just getting the oral challenge itself first. And we used PEN-FAST to risk stratify people, into a low risk, less than three, and then you got randomized. But the study wasn't about validating PEN-FAST. It was about validating low risk is safe for direct oral challenge. So, I think you can take from that study that yes, a PEN-FAST less than three, it's safe to directly oral challenge that patient. We now have randomized control trial evidence.

But if you look at the phenotypes, the people, they are people with childhood rashes. They are people with rash more than 10 years previous. All the same things that would have been in a Suran paper five years earlier, or our original oral observational studies. So, it's about validating the safety of direct oral challenge in low-risk patients and the ones we commonly see in an Australian setting.

MIC CAVAZZINI:  So, in that study, more patients were actually delabelled in the intervention arm than in the skin-testing control, because, it was argued, “positive skin tests may represent false positives, and this may lead to unnecessary avoidance of penicillin.” This was unwittingly corroborated by two examples of protocol violation, where patients in the study underwent oral challenge despite a positive skin test, and both turned out just fine without complications. Suran, do you want to speculate on how many people might be labelled allergic to penicillin due to a false positive skin test?

SURAN FERNANDO:        Yeah, they mention this in the PALACE paper, and Jason, please correct me if I've got this wrong. But when the pretest probability of a test is low, such as in low risk penicillin allergy, positive skin tests may re represent false positives and this may lead to unnecessary avoidance of penicillin. So, the PALACE study estimated that about a hundred thousand individuals in the USA could be labelled as penicillin allergic based on a positive predictive value of 50 to 75% because of the high false positivity rate. 

In fact, it's interesting, there was a another paper, Systemic Review and Diagnostic Accuracy Meta-analysis of Skin Testing in Penicillin Allergy, it was published around 2021, and it estimated a low positive predictive value on the pooled sensitivity and specificity data. And if the true prevalence of penicillin allergy is about 2% in a population, then the positive predictive value is probably around fifteen percent. So that figure of a hundred thousand could even be higher. So, I think we can ex extrapolate from that what the sort of figures would be in Australia. I don't know whether Jason wants to comment some more.

JASON TRUBIANO:          No, I love that. And I think it's true, you know, you do any diagnostic test when there's a low prepress probability across medicine and we get this problem, right? I think we've got a great example that's happened in China because their experience is actually doing skin testing for penicillin before a patient receives it therapeutically without a history of penicillin allergy. Now we've got some colleague and a friend, Philip Lee in Hong Kong who's been trying to stamp this out. But there's a huge number of patients rolling around that have a false penicillin allergy from a skin test. And we get them to clinic and if they migrate to Australia we have some of these that will come through my clinic and Sutan’s and they're always negative. So it just shows you what a a test done in the wrong patient can do.

SURAN FERNANDO:        And it's interesting, some of these patients that do migrate actually know that you should have a skin test before being prescribed penicillin and so they ask for the skin test.

Multi-site testing of PEN-FAST

MIC CAVAZZINI:                Jason, you and your co-authors of the PALACE study noted in the limitations of the study that 94% of participants had a PEN-FAST score of zero or one, so there would have only been dozen or so with scores of 2 or 3 who went on to oral challenge.  And you suggested that maybe numbers were too low to really observe the sort of rare reactions in the those middle scores. Also you noted that the patient sample was predominantly white, adult outpatient population, so who knows how broadly the results could be generalized. To tie of some of those loose ends, you both joined a couple of hundred contributors to the International Network of Antibiotic Allergy Nations study, published just a few months ago in Clinical Infectious Diseases. This involved 40 hospitals across 8 countries, not just our usual anglo peers but also in South Africa and Malaysia. Different settings, different ethnic groups, different ages. Suran, what more did we learn from massively scaled up project?

SURAN FERNANDO:        The iNaan study, which was really spearheaded by Elise Mitri—and I think we really should acknowledge her—as part of her doctoral studies, did an incredible job with this study. A lot of credit does go to her. But it involved multidisciplinary clinicians using a digital penicillin allergy toolkit, you know, the Naan app. To perform point-of-care penicillin allergy risk assessments using an adapted version of an antibiotic allergy assessment tool and PEN-FAST to determine the risk of performing a direct oral challenge. And that was the intervention.

And participating sites also received bi-monthly audit and feedback, which was the implementation strategy. And the study provided the largest body of evidence for the safety of adult inpatient penicillin direct oral challenges across a wide range of settings; high-income and middle-income countries; public, private, regional, remote; and acute specialized settings. There were 5,121 assessments and 1,573 direct oral challenges performed, and 95.5% of direct oral challenges were negative.

None of the positive challenges experienced anaphylaxis and there were no deaths. Six patients did have significant adverse events, with four determined as being immune-mediated to penicillin. One patient developed AGEP and that was treated with oral steroids. Another developed a diffuse rash also requiring oral steroids, but the rest required no specific treatment. So a very, very safe sort of strategy.

There was also a trial target evaluation within the study and those who underwent a direct oral challenge and removal of their penicillin allergy label had higher rates of penicillin prescribing within 90 days. Lower rates of restricted antibiotic prescribing, which are associated with all sorts of complications, including multidrug resistant organisms. And we also found a lower rate of one of these organisms, a multidrug resistant gram-negative bacteria.

FIONNUALA FAGAN: So, despite the evidence from the PALACE trial, there is often a lot of anxiety about whether to challenge an adult who has a history of urticarial reactions. And that uncertainty is often reflected in the guidelines and they do differ as to what type of distant events should be red flagged or not. I understand, Jason that you conducted a sub-analysis of the cohort, looking at those who had reported things like angioedema, immediate onset urticaria, rashes and swelling. And there’s a little bit of a spoiler in the title of that paper, it’s “Low Risk Is Low Risk”. Jason, what do you want us to take home from that paper?

JASON TRUBIANO:  Yeah, I think we owe a lot to Elise from running this. There's effectively four papers that I think have changed practice from the iNaan study. The paired CID publications about, you know, the effectiveness of implementation, and then these two posthocs about non-allergist and this low is low. And I think iNaan effectively becomes like a post-marketing phase four study where you've just got this huge cohort of people that are being challenged.

And the only way to find out about those phenotypes is huge numbers. And so we know from PALACE, and we know from observational data that childhood rash, delayed benign exanthem, rash more than 10 years, in every way you slice and dice them, they're safe. But what is the risk or reaction risk at some of those things that are controversial? Like urticaria. Like somebody says, “I blew up”. Like somebody that says, “I have no clue what my allergy is”. And if you ask different people around the globe if they're willing to challenge those, you can have a standoff argument about who's willing to challenge which. Now you've got two people that are on the same side here, Suran and I, so probably the wrong kind of people.

But what that did do in that lovely paper is actually in the supplement, and it gives the reaction rate of having a positive challenge per phenotype. And the highest is angioedema at 7.2%. So, you know what? You can still challenge that. There were no serious adverse events, but that's the allergists that are really doing those higher risk phenotypes. But you go to the bottom of that graph and there's unknown phenotype at just over 1%. You've got there benign rash at 2%, and then you've got urticaria sitting between 5 to 7%. So, for me, that's still low. That's a little 5% of people are going to be positive. 95% are going to be negative. None of those had a serious adverse reaction.

So, I think this gave us granular ideas about what phenotypes are safe, and now we've got unknown. We've got childhood, we've got benign rash, five or ten years, whichever way, and urticaria, particularly somebody that says ‘delayed urticaria’ because we know that's not a true Type 1 or immediate hypersensitivity. We've opened the window of who we can challenge dramatically by that data. So, I'm too passionate about that data because I think it…

Non allergist-led testing

MIC CAVAZZINI:  So, on the back of such findings, a guideline from the European Society of Clinical Microbiology and Infectious Diseases, just published in April, invites non-allergists to perform drug challenge tests in low-risk patients in settings where there is some backup, and reactions can be managed. But still there might be some nerves around this without dedicated training. So, to calm the farm, the iNaan study group compared the rate of adverse outcomes from this protocol when it was led by allergists as compared to other health practitioners. And from 2300 patients prospectively recruited, there were 66 immune-related adverse events observed. Suran, was there any difference in rates of events from challenges led by allergists to those performed without allergist input?

SURAN FERNANDO:        Yeah, that was part of an iNaan sub study led by one of our colleagues and friends, Neil Powell, in the United Kingdom. And we found that there was no difference in the rate of immune-mediated adverse events in sites with no allergist support and sites with allergist-led or allergist supported sites. And that sort of iNaan data really does provide evidence that penicillin direct oral challenges can be implemented in the hospital setting by non-allergists, especially given the fact that there are limited allergy immunology services. And that's not just in Australia, but that's just across the globe.

I mean, colleagues like Philip Lee in Hong Kong, who we mentioned before, that have shown very nicely that nurse-led and pharmacist-led penicillin allergy assessment delabelling programs are very, very effective. So, that was actually a very good finding from the from the iNaan study. because the global burden is just so high that it just can't be left to allergists or immunologists or people that have present drug allergy.

JASON TRUBIANO:  And it's and it's not about cutting one out, is it, Suran? It's about diversifying the workforce. And I think what that paper showed, which was nice, was broken up into three groups, right? Because there's three models. There's people that just don't have allergies available to them.

There's some that, like—I'm lucky enough that I'm working simultaneously with both in the same department. And there's others that just have allergists run. And it didn't matter across those three groups, which did it. But what it did show was who was doing what. And so it was really cool to see hey, the allergists, they're biting off a little bit more, they're doing those slightly higher risks. But then we can see the non allergist, they're really focusing on those childhood exanthem and delayed rash. And I think that's really cool because there's a place for everybody to do a sort of specialist work group. But collectively we can try and remove you know a burden that's 10 to 20 %.

FIONNUALA FAGAN:       So I think this gives us a really good model for in hospital treatment or outpatient treatment in a hospital setting. But what type of patients would you choose to do the challenges in, what resources would you want to know that you had on sight, and kind of training would you need to be able to do this safely.

JASON TRUBIANO:          Yeah. Look, this was being done for a long time outside of the hospital. In fact, iNaan was about inpatient challenges, right? But a lot of the observational data in PALACE were done in outpatient rooms off-site, not at a main hospital. So, you don't need an ICU on-site, a crit care team, a crash cart. And you have an iNaan database, and it's now in those subgroups that are more than 2,000 people, that nobody had anaphylaxis ICU from those challenges, particularly if you look at that lowest risk.

And so, a lot of us around the globe think that you know low is low. But if you're going to really stick to those low ones like childhood delayed rash unknown reactions, you don't need much available to you. Most would recommend yes, you should be prepared to treat anaphylaxis, like you give a vaccine where you would have an adrenaline or epi pen on site. But we've never used it in our entire program here for a low risk challenge. I mean Suran, what do you think about what you need?

SURAN FERNANDO:        Totally agree. We used to when performing direct oral challenges in the low-risk setting, always have rescue medications on standby, you know, with adrenaline and antihistamines, but we never used the adrenaline at all. So, in fact, it's not even part now of our guidelines and protocols when we do direct oral challenges for low for low risk you know allergies.

System-wide outcomes

MIC CAVAZZINI:  Great. Some of the sort of broader systems ramifications of the iNaan studies was, within six months of activation, tons of clinicians were adopting the Naan app. plenty of assessments performed by pharmacists and increased scope for others. I think Suran already alluded to this, there was a thirteen fold increase in penicillin usage, 1.3 fold reduction in restricted antibiotic prescribing.

And Suran, another observation was that the intervention was associated with a shift away from intravenous antibiotics to oral ones. Why should that be an immediate consequence and, for the non-clinician in the room, why are IV drugs a no-no when it comes to antimicrobial stewardship?

JASON TRUBIANO:          That's tough question.

SURAN FERNANDO:        It’s a good question, yeah. I mean there was a markedly increased prescription of oral antibiotics in patients who underwent direct oral challenges and were delabelled to penicillin. The antibiotics on the WHO access list of antimicrobials have a narrow spectrum of activity and a lower potential for the selection of antimicrobial resistance, in addition to having fewer side effects and lower costs. And many of these on the list on that particular WHO access list are oral antibiotics.

Intravenous antibiotics, such as the third generation cepahlosporins and the carbapenem like Meropenem are broad-spectrum antibiotics and used in patients with penicillin allergy labels and promote antimicrobial resistance and the emergence of you know multidrug resistant organisms.

But there are exceptions to this. You know, there are a number of IV narrow spectrum antibiotics such as ampicillin and there are also broad spectrum oral agents such as the fluoroquinolones. I don't whether Jason has anything to add, you know, to that. 

JASON TRUBIANO:          No, completely agree. And I think just the pragmatic approach is that we love you using oral Augmentin and we love using oral amoxicillin in Australia. And so it is an immediate thing to enable oral switch. and I think clinicians are comfortable with it, they're happy to switch down to it. So, it was just a natural thing that helped that IV to oral switch movement. And IV to oral switch is like a movement in Australia. It's a big AMS thing and the Kiwis I'm sure are the same. Like everybody come on, you gotta want to switch.

FIONNUALA FAGAN: I’m a haematologist, I like my IV.

JASON TRUBIANO:          I mean if they made Tamosin in a tablet, a haematologist would be happy to switch.

FIONNUALA FAGAN: We’re on Cefepime, my love, Cefepime!

JASON TRUBIANO:          Cefepime in a tablet? Now, we can't sell your Ceflex, but maybe we can sell your Cefuroxime. But look, I think that you know we loved trying as an AMS intervention to switch and this helped.

MIC CAVAZZINI:  Different fit for different systems, perhaps.

JASON TRUBIANO:  Yeah.

MIC CAVAZZINI:  Go on, Fionnuala.

FIONNUALA FAGAN:       I think it probably highlights how when we're comfortable with the oral antibiotics how quickly we are to go back down to them. Like we know how to use Cefalexin, we know how to use amoxicillin, we know how to use Augmentin, whereas some of the other drugs…

JASON TRUBIANO: Yeah.

FIONNUALA FAGAN:       Jason, so far we’ve been talking about assessing patients with an allergy label at the point of need- when you need to give them penicillin. But there was a 2025 paper in the Journal Clinical Infectious Diseases where you suggested that any hospital encounter could be used as an opportunity to delabel someone,  and then in some cohorts you might take a more deliberate approach still. I must admit this is something our department is thinking about whether or not when we can  get them into a delabelling clinic before we start them on chaemotherapy.

JASON TRUBIANO:          We could have a whole discussion on models of care, right? So I think one of the things is we now use this idea of opportunistic delabelling or targeted delabelling. Now the targeted is where: bug and drug and then it comes to a team that we need to use. That's easy and we need programs to enable that.

The opportunistic one is well they're on the ward, why don't we grab them? Now that's resource intensive. So, you kind of need to know well who are the people that are going to have the biggest bang for buck. And I love a paper we did a while back in 2021, Kyra Chua led that, where she looked at the predictors of people that were going to use a penicillin after being delabelled. And they were a surgical patient, they were somebody that had chronic respiratory disease, and they were immunocompromised.

Cancer, haematology, transplant, renal, hepatobiliary, they're the people. So, we have to find ways in which we prioritize those and target those. So, get them while they're in the hospital. You know, how's your AMS converter model of care where you can do a once-a-week round or targeted. Pre, when they're being worked up for their transplant—you know, new diagnosis of AML, they get automatically referred to us now. Being referred for renal transplantation, they get automatically referred. It's not just penicillin, it's sulfonamides in particular for PJP prophylaxis.

So we've got to get them early in care to prevent that downward cascade. That takes time and money, but it's not hard to just think about it and refer them on. So, if you refer that patient to Suran, he's going to see that on his list and go, “Crap, this patient's going to need antibiotics quickly. I'm going to put them on my clinic list”, right? So I think that's the idea too.

SURAN FERNANDO:        Yeah and I might add that there are so many specialized services that would benefit from having pathways for a rapid access to penicillin allergy to delabelling. So, Jason and I, see HIV patients for instance in a sexual health clinic. And so that's a very important group to also target, you know, patients that are risk of syphilis, for instance.

And also just even in obstetrics care, where you know, there's a great opportunity to delabel pregnant women who have a penicillin allergy label because they are at increased risk of beta haemolytic streptococcus later in pregnancy, which can cause all sorts of complications. So there are ample opportunities for opportunistic delabelling.

FIONNUALA FAGAN: Yeah. So, in earlier validation studies, the negative predictive value was lower when tested in children and teens and the iNaan study only took in patients that were above the age of 18. Suran, would the same kind of extensive testing need to be repeated in those younger age groups to confirm the safety of algorithm-guided direct oral challenge?  

SURAN FERNANDO:        Yeah, there was a very nice study, a Canadian prospective multi-centre cohort of just over two thousand children, led by Anna Maria Copescu and Jason, of course, and showed that PEN-FAST was not useful for restratification, especially in children younger than the age of twelve. The negative predictive value in children under the age of two is about 93. And 95 in children aged two to twelve as compared to adults with low risk scores of less than three, which have been shown in many studies to be sort of in it greater than 98% percent.

So, this is likely to be due to the increased prevalence of viral induced reactions in young children as compared with a true drug allergy. Children also experience different types of allergic reactions as compared to adults, and that's not always captured adequately by PEN-FAST. Now there is a penicillin allergy assessment tool for children as published in the ESCAPE-Allergy study in 2022 that demonstrated a high sensitivity for immediate reactions and a high specificity across all phenotypes. And that tool is now being utilized in the iNaan-KID study, which will audit existing clinical programs and participating hospitals to ensure you know to measure safety and antimicrobial prescribing outcomes in inpatient paediatric antibiotic allergy assessments and delabelling via direct oral challenge.

JASON TRUBIANO:          Yeah. I think the key thing is that kids are not little adults and we need we need tools specifically for them. But there is a ton of evidence that direct oral challenge is safe in paediatrics. In fact, when you look at the literature, most comes from kids in the outpatient setting. And do yourself a favor and read the paper from Vyles et al, in JAMA Pediatrics. It is the best paper demonstrating the safety of direct oral challenge in kids. And what Suran has beautifully highlighted is that we still need to see the demonstration that this is helpful in the inpatient setting. Does it improve antimicrobial stewardship? Is it opportunistically while they're in hospital rather than getting them to clinic? All those things we've talked about and learned from adults. But we know it's bloody safe. It's very safe in kids.

MIC CAVAZZINI: And Suran, this is probably a whole other podcast, but I'll give you a chance to nerd out on patients who are labelled for you know proper bona fide severe reactions in the past. then I understand that there are emerging immunologic diagnostics that could be used to identify those who are no longer no longer so allergic.

SURAN FERNANDO:        Are you talking about those that have severe cutaneous adverse reactions?

MIC CAVAZZINI:                Well those that score highly on PEN-FAST with whatever reaction. Is it worth going into that? Are we close to an another on the spot assay or that a whole other…

SURAN FERNANDO:        I can talk about the severe cutaneous adverse reactions, you know, for instance—they can be investigated further in specialized centres like Royal North Shore or and the Austin. Maybe I'll give you an example. I mean, for instance, if a patient with DRESS had two or more implicated agents, we can perform skin testing to those agents, as well as some in vitro studies, such as lymphocyte transformation tests or ELISPOT to see what the rate of reactivity or proliferation in the lymphocytes and also the production of various cytokines as a result of that activation.

So, in doing so, we may be able to identify the culprit agent and consider delabelling to those drugs that tested negative. If the benefit of this approach outweighs the risk. These strategies are primarily in the research base. I have more confidence in proceeding to delabelling if we have identified the implicated drug by skin testing and an in vitro method and excluded the other agents using the same two methods. Because, of course, the risk is high if we get it wrong.

MIC CAVAZZINI:  And just to keep Jason busy, I see that that you’ve published papers in the last year describing validation of tools called CEPH-FAST, and SULF-FAST. I couldn’t bring myself to dive into that pool of literature, but should we expect a similar story to the one we’ve heard today to build-up for those cephalosporins and sulfonamides.

JASON TRUBIANO:          Yeah. Well PEN-FAST was my oldest child, my daughter. I love her to bits. And these are kind of like two twins that came and that's it. We're not having any more as a family. So, I think, they're all based upon the same principle. They're based upon the clinical decision rule, the same variables, and what it demonstrates is that low is low across antibiotic types; penicillin, sulfonamides, and cephalosporins. Those classic phenotypic features, if you score less than 3, the negative predictive value is more than 94% across all three of those tools. And what we're trying to find is just pragmatic, practical ways where we can risk assess people across antibiotic allergy types and give them the damn drug so they can get better and have the appropriate antibiotic therapy. And I think that's just two more tools that help that.

I think CEPH-FAST needs bit more validation. Cephalosporins are tricky, you know, there's intravenous, there's oral, there's a million different types. Sulfonamides is a lot easier because it's one drug predominantly, we're dealing with an antibiotic, and we just got to be super careful we're not dealing with a SCAR. That's the key thing we've got to deal with in sulfurs. But otherwise, jee, it's very handy to give somebody that Bactrim—urinary tract infection, skin and soft tissue, MRSA, PJP, melioidosis, take something. It's a great drug.

MIC CAVAZZINI: Many thanks to Profs Jason Trubiano and Suran Fernando for sharing their time and expertise for this episode of Pomegranate Health. There’s a transcript of this conversation at our website, studded with links to every research finding alluded to. Just go to racp.edu.au/podcast and click on episode 154, or if you’re listening from a smartphone app, follow the link at the bottom of the blurb.

Don’t forget to log your CPD hours while you’re there. For some great lectures on allergy, immunology and infectious diseases go check out the College Learning Series at eLearning.racp.edu.au. The CLS is designed around the curriculum for basic physician training.

I’m really grateful to Dr Fionnuala Fagan for bringing this story to my attention and to all the members of the podcast editorial group who had a listen. They are physicians Marion Leighton, Robin Sia, Aidan Tan and Hugh Murray and also Arnika Martus and Paul Cooper PhD. Working with these curious and generous people is the best part of my job.

The music soundtrack for this podcast came from Tigerblood Jewel and was licensed through Epidemic Sound. This podcast was produced on the lands of the Gadigal clan of the Yura nation. I pay respect to their elders past and present. I’m Mic Cavazzini, thanks for listening.