ILD in RA - Recent Advances. Dr. Jeffrey Sparks Save

You're listening to a podcast session from RoomNow live recorded on 02/07/2026. These lectures on rheumatoid arthritis are made available to the rheumatology community by AbbVie, a premier sponsor of RNL twenty twenty six. I hope you enjoy the lectures, but first, a message from our sponsor.

For the latest topics in rheumatology, listen in on room replay. This podcast series for healthcare professionals features leading rheumatologists as they share their insights and firsthand experiences in one on one conversations with their colleagues. Tune in to Room Replay on Spotify and Apple Podcasts brought to you by AbbVie.

Okay. Our third speaker in this pod is Doctor. Jeffrey Sparks from Brigham and Women's. Jeff is a worldwide authority on so many things, but especially on lung disease and RA, and he's going to update us with the recent advances. Jeff?

All right. Thanks so much, Jack and Artie, and the whole RheumNow team for inviting me. There's 35 inches of snow and counting in Boston, and it's 10 degrees there, so I was very thankful. Stay the the week. I have to say go Pats, course, too.

All right, so I think this is a really great pod, obviously. Think a lot of the similar themes here and we're going to, you know, this is going be pretty clinical focusing really just on ILD. And we have a lot to talk about so unfortunately I didn't put any quiz questions, but I'm looking forward to the Q and A. Here are my disclosures and funding. All right, so here are the main themes and they told me to talk about some recent advances and actually this has been a pretty hot topic and over the last five or ten years a lot has transpired in this field.

So I decided to break this down into four different buckets. The first being RE ILD epidemiology and risk factors. The second, outcomes progression and mortality that was already alluded to before. There were two guidelines just released in the last year or so, and then end on treatment. So a lot to get through here.

So the first section is about epidemiology and risk factors. And probably the overarching theme here is that, as we all know, RA has gotten a lot better related to treatment options. We got a lot more things to choose from for the joints. Overall, we feel like things are in better shape there. And the notion is that RA, you know, we're doing pretty well.

Unfortunately ILD is the exception to that rule. It's probably one the one outcome within RA that's not getting better even with expanded treatment options. So we still have a lot of work to do in RA particularly related to ILD. I wanted to set up this related to ILD prevalence in rheumatic diseases and this is our group is interested across all of the rheumatic diseases. And you can see here that the pooled prevalence in this meta analysis of RA for ILD is eleven percent and that's compared to forty seven percent in systemic sclerosis, pretty similar forty one percent in myositis and you can see some of the other systemic rheumatic diseases here.

So RA, in a way, the prevalence is high, but not so high that we would screen everyone. And I'll mention that the denominator of patients here is probably more than all of the others combined. And of course, since we aren't scanning everyone, there's a lot of patients with subclinical ILD. So this is kind of a challenge and an opportunity. Do we just wait to see until patients hit our door and they develop symptoms?

By that point, patients often have lung damage that's irreversible. So I think this is kind of a really tricky area because the prevalence is high, but not so high that we'd want to screen everyone. Of course we have to mention Olmsted County, has done a lot of great population based research here. And I think this was pretty striking. When I asked other rheumatologists how common is RILD, they say, Yeah, I see it every once in a while.

I'm not sure it's that common. So this was a population based study where they basically found every patient with ILD and this is clinically diagnosed, not screened. One notion that I was told as a fellow is that ILD only occurs in long standing RA that's untreated. And I think this really counters that notion. First, if you see right at the bottom here, zero, that means that's when RA is diagnosed from an articular standpoint.

And day one, we've already got about four percent of patients who have ILD. And that of course means the ILD happened before RA and that probably could go into the mucosal origins hypothesis that was just detailed. The other thing is that ILD steadily goes up with duration of RA. So probably the first five years here there's still patients being diagnosed with ILD. So ILD for sure can happen in early RA as well.

And you can see that it ticks up steadily and all the way up to twenty years of RA duration. The cumulative incidence is fifteen point three percent, which is nearly one in six. And I'm betting if you ask your neighbors, family members if they think one in six is high, I think most people would say yes. So I think the prevalence of ILD is a bit higher than we suspect. And this is clinical ILD.

This is looking at ILD prevalence over calendar time. This is a Danish study. Basically this is the biologic era. And the gray bars are prevalence of ILD. The black is incidence of ILD.

The incidence of ILD is not getting better, maybe slightly worse. But you can see that the prevalence of ILD is steadily going up. And probably lots of mechanisms for this, easier to get imaging, maybe more recognition, but for sure the prevalence of ILD seems to increasing and this one was increased almost twofold even during the biologic era. So back to the clinic, and this is a bit of an issue with ILD is this alphabet soup of subtypes. So I wanted to mention this as well.

For RA, the most common subtype is usual interstitial pneumonia, which is the quintessential fibrotic subtype. This looks very similar to idiopathic pulmonary fibrosis or IPF. And this is the most common subtype within RA ILD, about fifty to sixty percent of patients. So you can think about you can remember this in RA related to usual most common. The next most common is NSIP or nonspecific interstitial pneumonia, around thirty to forty percent.

Confusingly, there's two sub subtypes, a more pure inflammatory version called cellular NSIP and another version that has some fibrotic features called fibrotic NSIP. But overall, NSIP is thought to be a bit more inflammatory phenotype. Of course, there's a large spectrum in between and these often take multidisciplinary panels. And honestly, those are by far the most common subtypes within RA ILD. I only mentioned organizing pneumonia, which are kind of fluffy infiltrates that can look like infectious pneumonia, but are actually inflammatory and often can be pretty responsive to glucocorticoids.

But other ones, LIP, DIP, RB ILD, DAD, and then rapidly progressing ILD are pretty uncommon overall in RA ILD. A very interesting thing about RA is that the UIP has a predilection for RA rather than other systemic autorheumatic diseases. This is from the same meta analysis. And basically they looked at the proportion of patients with ILD that have different subtypes. The dark blue is UIP, the dark green is NSIP, the light green is OP and then some other subtypes here.

And you can see that RA, most of the or about half the patients have UIP. And interestingly, systemic sclerosis, even though it's the prototypic fibrotic rheumatic disease, the most common subtype is actually NSIP. And same thing with myositis, NSIP as well as OP being pretty common. So there's something about RA that puts people down the path to UIP. And as we'll see later, UIP actually has the worst prognosis and is most progressive.

And luckily there seem to be therapies that are on the horizon for this. So now I'll detail some RAILD risk factors. And again, there's been lots of developments over the past few years. And some of these are very easily attained. For example, men have a higher risk of RA ILD.

Obviously women are more likely to have RA, but within RA if you're a man you're more likely to have ILD. We'll go over a genetic risk factor called the MUC5B promoter variant, smoking, obesity and several RA characteristics including age at diagnosis, articular disease activity, inflammatory markers, functional scores, as well as autoantibodies. I first wanted to highlight this genetic factor called the MUC5B promoter variant. And this is a single common variant. It's outside the coding region of this gene that codes for mucin 5B, which is expressed in the distal airways.

And this is pretty common. People with European ancestry, around ten to fifteen percent of people have at least one of the copies of this MUC5B promoter variant. So for sure there's going to be some people in this audience that have the MUC5B promoter variant. And this single genetic factor is strongly associated with RAILD with over nearly fivefold increased odds compared to the general population and within RA over threefold increased odds. And it's specifically associated with the UIP subtype of RA ILD and this is a figure from the New England Journal paper that showed that this co localizes with honeycomb tissue in end stage RA ILD.

Our group is also associated with ILD earlier in the RA course and interestingly it's also associated with older onset RA. And it seems to be a little bit protective against acute respiratory insults like infections and smoking and other inhalant injuries, but there seems to be a long term disadvantage as far as remodeling and recruitment of fibrotic factors within the lung. So again, this is a single genetic factor. This is a really nice Finnish study nationwide that basically on the left are men, on the right are women. On the axis is age.

And whether you have RA or the general population with or without the MUC5B promoter variant. And you can see that there's a striking association with age. There's almost no risk of ILD until you hit the age of around 50 or 55, both in men and women. And if you have RA and the MUC5B promoter variant and you're a man, the risk is fairly marked. And obviously even if you have RA without the MUC5B promoter variant, the risk is high.

But I'll say that women are still susceptible to this even though their absolute risk is a bit lower than men. Women who have RA and the MUC5B promoter variant do have a high risk of ILD and our group is also interested about whether there could be menopausal factors that are associated with this. And this is an important factor and we actually think it could be clinically available in the near future. So this is something to think about. Another thing that we think a lot about in RA is disease activity.

So our group looked to see whether disease activity was associated with incident RA ILD risk. And you can see that in a dichotomized version, basically people who have moderate or high disease activity compared to remission or low had over a two fold increased risk of incident ILD, adjusting for some known ILD risk factors. And you can see the marked dose effect as well. On a four level ordinal variable compared to remission, both the moderate and high disease activity categories were associated. The high disease activity category had over threefold increased risk.

And on a continuous level, every point increase in the DAS28 increased ILD risk by thirty five percent. And this opens up a lot of questions. So first off, we already want to get people into remission or low disease activity. This is maybe some more opportunity to try to do that. And the other question is are there specific medications that would put people at either higher risk or lower risk?

And that's a really important clinical question and there's lots of literature both for and against that that perhaps we'll get to at the Q and A. Did But not have time to detail all the DMARDs risk. However, I did want to mention methotrexate, one of my favorite medications. And when I was a fellow I was told that methotrexate might cause some fibrosis and ILD. And actually this seems to have been debunked in recent years.

I will mention that methotrexate does rarely cause pneumonitis. We helped to do a very large, complete control trial where we blindly adjudicated outcomes and within this study sample of 4,800 patients, there were seven cases on methotrexate which was point three percent and one case on placebo or less than one percent. So pneumonitis does happen. This is a bit different than ILD. These people have a timing of exposure to methotrexate usually within the first six months of starting.

They often have fever, systemic symptoms and the imaging looks a bit different as well. As far as adjudicated incident RAILD, there's now been seven observational studies. Our group published a meta analysis recently and you can see that the odds ratio here is actually below one. We think probably related to reducing inflammation as opposed to anything specific about methotrexate, but certainly the odds ratio is not above one and you can agree with that. So in my view, methotrexate does not cause incident RAILD.

So a prelude to a more detailed section later about guidelines. The ACR CHEST released it was actually last year, but they released the 2023 ACR chest guidelines. And as part of the guidelines, they recommended screening strategies. And I'll say that this really is going to open up a lot of research avenues and I don't have time to talk about all the other diseases, but related to RA, they related to consideration for screening men, people with high titer antibodies, smokers, older age at RA onset, obesity. And actually if you take all of these in isolation, you might be screening, you know, three fourths or even higher of all your patients with RA.

So we try to we're trying to come up with a little bit more tighter screening strategy. So there's been other risk scores that are a bit more nuanced. One that was recently published was the VA RA ILD risk score. This was developed in this mostly male smoker population, but a lot of the factors came out that were similar to what we've talked about. MUC5B, they also include four other genetic factors that are more modestly associated with older age, smoking, male sex, disease activity and rheumatoid factor positivity.

And you can see that the curve overall looks pretty good with an AUC of about point seven five. Another pretty similar study, this is in a more generalizable cohort, granted they're French, but there were two separate cohorts where they were prospectively screened for subclinical ILD. And they found subclinical ILD in around twenty percent within each of the cohorts. And the factors that drew out were the MUCVY B promoter variant, male sex, older age at RA onset, and higher disease activity. And they actually had even higher AUCs at point eight two and point seven one.

And I actually really love this figure from the paper because it really talks about how we would risk stratify based on these factors. So to orient you, the columns have disease activity kind of low, moderate, or high, female, male. And on the rows are age of RA onset, kind of young, middle age, older. And then the MUC5B promoter present or absent. So if you're in the top left corner, you basically have no or very few risk factors.

And if you're in the bottom right corner, have many risk factors. So if you had all four risk factors, the chances that you had ILD in that CT scan were ninety five percent. Whereas if you had none of the risk factors, the chances were almost zero percent. And certainly you'd have to figure out where to draw the line, but I think it makes sense to screen people who are over here. I think it probably makes sense to not worry about ILD for people over here.

So this is kind of risk stratification that would be very helpful in clinic. And besides the MUC5B promoter variant, these are factors that are all very accessible. So our group is also interested in this and we have a study called SAIL RA looking at ILD prevalence in early RA. This is five sites and we recently hit our recruitment goal. Basically we try to get patients in for a study visit as soon as they're diagnosed with rheumatoid arthritis.

It's also longitudinal so we're looking at progression over time. But we get research CT scan, PFTs, etcetera. I'll also mention we're really proud of our logo. It's got some Easter eggs here. The sails are actually lungs.

You can see the fissures there, the right and the left lung. And the boat is actually a joint and that is synovitis, the wave. So we just published the baseline results here and these patients had RA for a median of around six months. And eleven percent of them had subclinical ILD. These people did not have symptoms.

And again, I think that's a bit eye opening. Six months of RA, eleven percent of ILD. And we haven't didn't have the genotyping information, but quite back yet for this paper at least. But older age, male sex, although not statistically significant, but and then the thing that was quite marked was RA disease activity with seven fold increased odds if they were in moderate or high disease activity at the study visit. And we compared some risk strategies between the Ankara study that's ongoing, the ACR chest guideline and another study called the four factor risk score.

And we found overall these people these strategies had pretty high sensitivity, modest specificity, but for a screening strategy you'd probably want to maximize sensitivity. And then we also decided if you just count the risk factors and if they had two or more risk factors that the sensitivity dropped slightly to seventy nine percent but the specificity rate went up seventy three percent. And this was pretty efficient meaning that you'd only have to screen 3.6 people to get one ILD case. As an example, in our cell array model, you would only have to do scans on 54 people to find 15 of the cases which was almost nearly all of the cases. And these really simple factors all had AUCs in the point eight range.

So we think we're very close to actually coming up with a model that would be able to identify ILD with a simple risk score. I will mention that screening is inevitably a risk and benefit analysis and part of it is that once you find something you want to find have an intervention that modifies the natural history. Maybe that's more closer follow-up, maybe it's more PFTs, more scans, but you really wish there's a medication that would prevent the progression. So we've probably gotten progress here about early detection. We really need to figure an early intervention.

We need to quantify improved quality of life. We need to quantify cost. And also know that there are potential harms related to over diagnosis, incidental findings, excessive radiation and of course doing CT scans on a lot more people as a cost to the healthcare system. So I'll touch on progression and mortality. I'll probably go pretty fast through this since it kind of was already covered previously.

I think this figure was actually shown as well. But I think as already alluded to, from a relative risk standpoint, ILD is the thing that really stands out compared to the general population. And this figure was also shown earlier as well that the outcomes that are not getting better, even with the improved treatment outcomes, are infection and lung related mortality. Looking at survival, this is another Mayo Clinic study. This was an older study looking at ILD survival.

And you know, these patients were diagnosed well before the biologic era and most of them followed just to the start of that. And you can see that the median duration was only two point six years. Luckily the duration has improved and my thought would be that we're catching them earlier, but overall there's still a marked difference in mortality in RA compared to, in ILD compared to RA patients without ILD. And the hazard ratio is two point four two. So in this current era, if someone gets diagnosed with ILD, their hazard ratio of dying is two to three fold increase even with all of the ex all the treatment options that we have for RA.

And I've already alluded to this before. Part of this is related to the RA predilection of getting UIP. Within ILD, all of which have poor prognosis, UIP seems to have one of the worst prognoses. And this is within RA. The blue patients have NSIP, the red patients have UIP.

And you can see that the survival is much worse and kind of more progressive compared to NSIP. So you really worry about patients with RA who have a UIP phenotype. I'll briefly touch that we're also analyzing this smoker cohort called COPD gene that prospectively obtained over 10,000 CT scans and not a huge surprise, but there haven't been a lot of controlled studies, the odds of RA patients having subclinical ILD was eightfold increased and had predominant fibrotic phenotype. And there were also those patients who had subclinical ILD also had threefold increased mortality. So subclinical ILD definitely does progress.

Our group is also interested in using AI, who isn't, but we have computers now look at the scans and quantify the amount of ILD and the type of ILD that's within the scan. And that could be something that could be scaled and of course it's really cost and labor intensive to read these scans and to figure out who's progressing and to have kind of automated way to do this would be really a game changer. Briefly, we have some Korean colleagues that have an excellent prospective ILD cohort and these patients all have clinical ILD. We recently came up with a risk prediction model to see within ILD who are more likely to progress because not all ILD progresses. And we came up with a risk score related to UIP pattern, ILD extent, low DLCO, high titer CCP and a fibrotic marker called KL6.

And this actually worked quite well to risk stratify. You can see the high risk group had over 30 fold increased hazards of progressing compared to the low risk group of which almost no one progressed and then a moderate group. So even within ILD, there's probably some risk stratification that could happen and these probably really would indicate for early treatment. And that segues into guidelines and treatment. I was part of this process and I have to say this was one of the most complicated guidelines that I think the ACR has put out because not only was it focused on a complicated disease, it looked across many underlying rheumatic diseases and also looked at screening, monitoring.

And also the one of the issues is that a lot of the trials have happened in systemic sclerosis and they've basically used the transitive property back from grade school algebra and used those data to see to inform other diseases. And it kind of left us in a little bit of a box because some of the treatments for NSIP with systemic sclerosis may not apply to patients with rheumatoid arthritis who are more likely to have UIP. So for example, it recommends mycophenolate as first line and we know that it has poor efficacy for joints and it's pretty immune suppressing. So if you have to put it on mycophenolate for the lungs and something else for their joints, they're going to be even more immune suppressed and they have lung damage and you worry about infection risk. Azathioprine that we'll get to and then Rituximab, which we certainly use for RA joint activity, usually not first line.

They also have some recommendations for progression after first line. Many of the same medications end up on this, as well as some anti fibrotic, entinanib and profanidone. And we'll briefly go over some of the data here. Our group recently put together a narrative review detailing the data that informed RA ILD guidelines within the ACR. And it was really eye opening.

This is our most common rheumatic disease. This is probably one of the most serious manifestations. And believe it or not, there have only been four trials that have randomized patients with RA ILD. Only a total of three thirty five patients had been enrolled in a pharmacologic ILD trial for with RA. And many of them were basket trials across many different indications.

And also many RA drugs were not were recommended against or were not addressed in the guidelines. So definitely a lot of work and we wish there were more trials. I'll say that the ERS and ULA recently also released some guidelines and probably the biggest difference was within rheumatoid arthritis. They actually recommend tocilizumab, glucocorticoids, abatacept, JAK inhibitors, and those weren't even on the list for the ACR guidelines. So there's some disagreement within these big professional societies about how to treat these patients.

I'll end on RA ILD treatment and then we'll go to the Q and A. The first that I'll mention is that this is pretty exciting that there's a new medication, Narendomolast, that was just approved for idiopathic pulmonary fibrosis and progressive pulmonary fibrosis of which some patients could have RA ILD. And this was a pretty large trial with eleven over eleven hundred patients. It was placebo controlled and a large portion of them had autoimmune disease, twenty nine percent, and one hundred and eighteen patients had RAILD. And the primary outcome was FVC decline.

And you can see that Narendoma last did seem to slow the decline. And a really exciting on a secondary outcome is that there was lower mortality with a hazard ratio of point four eight to point six zero. I will say that this was a secondary outcome so it would be nice to see that verified. Nantinenib is you're probably more familiar with. This has been approved for and IPF for several years now.

And the INBUILD trial was another basket trial that did include some RAILD, thirteen percent of them. Also looked at FVC decline. This trial did not show mortality benefit. As you probably know, they have a lot of GI tolerability issues with over sixty seven percent of patients having diarrhea. Clinically, at least half the patients can't tolerate it due to GI side effects.

There was an in build subgroup analysis of RA ILD. Even in this small subgroup there were statistically significant findings of lowering FVC within RA ILD. Perfindodone is actually the only trial that's ever been done just for RA ILD the TRILO-one trial, I should say. And they're they instead of FVC, they chose a composite of death or decline of FVC by 10% or more, which was a kind of lofty expectation. And unfortunately, they under enrolled related to the COVID pandemic.

While numerically they had lower risk of this primary outcome, it was not statistically significant. However, if they had chosen FVC as a primary outcome, it did slow the decline. And interestingly, if you looked by underlying ILD pattern, profanidone really worked quite well in patients who had UIP where there was almost no decline in those who received the study drug, whereas the placebo group with UIP there was a marked decline. Interestingly, the profanidone did not seem to work in the non UIP pattern. I'll end briefly on immunosuppression, just saying that there have not been any trials in RAILD and we really need to perform those and our group is working towards that.

This is overall the level of evidence that we're looking at, kind of before and after starting a medication for RILD and you can see the FEC decline seems to stabilize but there's a lot of bias that can come into these types of analyses of pre post. I'll maybe end here that the PANTHER trial was actually done in IPF but it really informs RA ILD treatment because this was a trial looking at immune suppression in patients with IPF which again is very analogous to UIP. And the combination therapy of prednisone, azathioprine, and acetylcysteine actually had increased risk of death. So interestingly, we use immunosuppression all the times in our ILD with UIP. However, I do try to avoid prednisone and azathioprine.

I think I've already alluded about CellCept, how this is not a very satisfying medication for RAILD. And then Rituximab, there is a trial for this, however did not include RA. The RECITAL trial compared rituximab and IV cyclophosphamide and did show basically equivalent for cyclophosphamide and obviously was much better tolerated. So this probably has the best data for RA ILD, but of course you would still worry about immune suppression in these patients that have lung damage. And, you know, we're using this medication, but it hasn't necessarily made outcomes.

Tocilizumab actually has FDA approval for SSC ILD and we use it all the time RA ILD. There's been almost no studies in RA ILD and that would be low hanging fruit, I think, for another study. And I think I'll stop here. So anyway, in conclusion, hopefully you can agree that RA ILD is a really important outcome that's not getting better. We need better treatments.

There's lots of heterogeneity here. We've made some progress and there are lots of guidelines now but this is only informed by very few trials that have enrolled people with RAILD. So I'm really excited to be here and, really looking forward to the Q and A that I think is happening now. So thank you.

If you enjoyed this podcast session, there are other RNL twenty twenty six podcasts to listen to. Also, you can purchase unrestricted, on demand, full access to all of the RNL twenty twenty six content. This includes lectures, videos, podcasts, slide downloads, pre learns, and test materials. Go to roomnow.live for more info. Room Now Live, a truly dynamic and highly interactive two day educational conference.

We'll see you at our next R and L, January 3031 in 2027.