14 November 2023
A few of my favourite things: Dr Claire Owen
The American College of Rheumatology Convergence is in full swing, and we’re joined by Melbourne rheumatologist Dr Claire Owen to discuss the highlights of the day.
Some highlights for Dr Owen, from Austin Health and the University of Melbourne, were the great debate about treating polymyalgia rheumatica and giant cell arteritis patients with advanced therapies immediately, how ChatGPT shaped up against a panel of rheums and thought-provoking research on long-term outcomes for giant cell arteritis patients.
The Great Debate – early treatment of PMR and GCA
First thing this morning was the Great Debate between Philip Seo and Robert Spiera on whether advanced therapies should be used up-front in polymyalgia rheumatica and giant cell arteritis patients – and I think Robert Spiera drew the short straw being on the negative.
Robert highlighted what our goals of treatment are in both PMR and GCA, and it’s quite obvious for GCA that you’re trying to prevent the patient from going blind and from developing vascular complications.
But what we’re trying to prevent is a lot hazier in PMR. There’s no known disease modification that comes with more early, aggressive treatment. You could hypothesise that you might be trying to prevent the emergence of giant cell arteritis, but that doesn’t happen for every patient. Otherwise, you’re really just talking about minimising treatment complications, and a good proportion of patients may only need steroids.
He acknowledged there was sound evidence for the use of tocilizumab, but very little evidence to support this approach early in PMR. And even in the tocilizumab space, we don’t yet know whether that early treatment is disease modifying longer term.
Then Philip Seo said he wanted to give two dot points and do a mic drop.
These were “1. biologics work, and 2. steroids are bad”, and that was the end of the argument.
He did, of course, go on to justify that. But again, the argument really wasn’t about GCA, because I think the evidence there is pretty clear that early initiation of a biologic agent is going to have better outcomes for the patients in terms of relapse and steroid exposure, et cetera.
So he focused more on the incidence and effect of steroid-related side effects in PMR. But probably the most salient point was how dangerous cumulative steroid exposure is, particularly demonstrating an increased risk of things like major adverse cardiovascular events. He showed some really interesting data that I hadn’t seen before about cerebral atrophy in patients on long-term, low-dose prednisone therapy. Even in doses of 5mg or less.
He said it made him feel very uncomfortable about this idea that has been long accepted in PMR of ‘What’s the harm in long-term, low-dose glucocorticoids?”
In the wrap up, Robert Spiera did acknowledge that, being the first author of the SAPHYR study, he was pro the idea of biologics, particularly in PMR. And he thinks that in time, we’ll have more evidence to support their use, potentially, in earlier stages.
The point about the effect of 5mg of steroids gave me food for thought, as long-term use was always something I’ve accepted for my PMR patients. I don’t push them to go on to a steroid-sparing agent, because there’s no evidence to support that approach, particularly with the use of something like methotrexate.
But I think we really need to reconsider that, because if you start to use data from other conditions such as lupus, RA, et cetera, the treatment paradigm there is one of very minimal or no steroid use long term.
AI diagnostics in rheumatology
One fascinating session set ChatGPT against a panel of rheumatologists to come to a diagnosis in a particularly difficult clinical case of myositis.
They started with the case presentation: a gentleman in his 40s who’d come with a skin rash all over his lower limbs and muscle weakness. The presenters input the data just like you would when you’re being presented a case by your registrar on the ward round.
Each time they asked ChatGPT to come up with differential diagnoses and compared these to what the clinicians were doing – and it was a fabulous insight into the art of rheumatology in terms of our clinical reasoning.
ChatGPT did quite well, in that it did arrive at the final diagnosis. But it was a lot more clunky in the way it went about it and made me recognise how fast the human brain manages to synthesise this information in a clinical scenario.
For example, ChatGPT kept trying to say one of the differentials would be cellulitis for lower limb swelling and erythema…except it was bilateral. So it could make the generalisation that redness plus swelling equals infection, but didn’t understand that this is very unlikely if those are on both sides.
Long-term vascular outcomes in GCA patients
I had the pleasure of seeing friend and colleague, Tony Sammel, present the five-year follow up of his original GAPS study – which was excellent and very well received.
While the study is on a small number of around 10 GCA patients, he’s continued to image this original group on a regular basis and tracked the development of vascular complications.
The amazing thing was that there was a significant reduction over time in the abnormalities that we’re seeing in the cranial vessels following treatment.
But in the extracranial vessels, there were patients who developed new changes up to five years after their diagnosis of giant cell arteritis – basically, aortic inflammation visualised on a PET/CT scan, years after their diagnosis. Some of these patients were no longer on any treatment, others were still on conventional disease modifying agents, because most of these patients were diagnosed in the pre-tocilizumab era.
Ultimately, what Tony showed was that the patients who were still on the DMARDs didn’t have uptake – and that was to a significant level compared with patients who were who were not on the treatment.
It’s really the first study I’m aware of that is hinting at this idea that there might be a benefit to remaining on a disease modifying agent to prevent the new development of arterial inflammation in giant cell arteritis.
So it really gives us some new insights to think about with GCA management long term.
This is important because we’re treating with tocilizumab for a period of time, and then patients are coming off their medication and we’re really not sure what we should be doing with them after that.
The other aspect to that study was to evaluate the likelihood of developing aortic aneurysm, and they did see some increase in the aortic diameter over time.
The patients that had uptake at baseline weren’t necessarily the ones who actually developed the later dilatation. So that’s a little bit concerning, although other research on a much larger cohort presented at the conference did show that association from baseline uptake.
One other take home I had from Tony’s work was that echocardiography of the aortic root appeared to be a good screening tool to detect patients that were developing changes over time – it didn’t have to be a PET CT scan.
I’ve always favoured the idea of doing an annual echocardiogram in patients who have known large vessel uptake at diagnosis. I didn’t really have a lot of evidence to support that, but certainly what Tony presented today would reinforce that that’s a sensible approach with what we know at the moment.
This interview has been edited for length and clarity.