First-Line Biologics for Psoriatic Arthritis: Have Patient Selection Criteria Evolved?
Background
The introduction of biologics into the treatment of psoriatic arthritis (PsA) has dramatically improved clinical outcomes. Across the approximately 2 decades since the first tumor necrosis factor α (TNF-α) inhibitors were approved in the early 2000s, available drug options for patients with PsA have continued to evolve. In particular, biologics targeting the interleukin 17 (IL-17) and IL-23 pathways have facilitated more selective targeting of specific body systems improved the ability to control enthesitis, skin manifestations, and joint damage.
Notably, in a recent report, physician-scientists from the rheumatology department at University Hospitals Leuven in Belgium suggested that the broadening for PsA has actually spurred greater understanding of its pathophysiology.1
Given these transformations, this Journal Club will focus on a study by Ishchenko et al that investigated ways that treatment escalation for patients with PsA have evolved since biologics first became available and whether clinical and disease factors or demographic and lifestyle variables have changed among patients eligible for first-line treatment with biologics.1
Methods
Ishchenko and colleagues analyzed KU Leuven registry data for a prospective cohort (BioSPAR) of patients with spondyloarthritis and PsA undergoing treatment at University Hospitals Leuven. Patients with only spondyloarthritis were excluded from the analysis.
The study investigators defined 3 epochs, or study periods, between September 2000 and February 2018, which corresponded to the availability of the first and second generation of TNF inhibitors and the third generation of biologics available for the treatment of PsA. Patients were placed into the epoch based on the date their PsA treatment was initiated. Patients who began PsA treatment from September 2000 to June 2006 were placed in period 1, from June 2006 to March 2016 were placed in period 2, and from March 2016 to February 2018 were placed in period 3.
Aggregated intake data included presence of clinical characteristics, such as dactylitis, enthesitis, and skin or nail psoriasis; disease activity scores, including swollen joint count (SJC), tender joint count (TJC), C-reactive protein (CRP), and erythrocyte sedimentation (ESR); and history of previous treatment with disease-modifying antirheumatic drugs (DMARDs), nonsteroidal anti-inflammatory drugs (NSAIDs), and/or glucocorticoids. The number of patients receiving and discontinuing biologics, DMARDs, NSAIDSs, and immunosuppressants as first-line treatment were recorded.
These outcome variables were analyzed using one-way analysis of variance and unpaired t-test to compare continuous and normally distributed data, Kruskal-Wallis test to compare continuous and non-normally distributed variables, and χ2 test to analyze discrete variables. Drug survival was compared across study periods 1 and 2 using the log rank test.
Results
Of the 407 patients included in the BioSPAR registry, a total of 185 patients (White, 100%) were recruited for this 18-year study. In the first epoch (n=65), all patients received either infliximab or etanercept as a first-line treatment. Similarly, of patients in the second epoch (n=84), 76 received either infliximab or etanercept and the remaining patients received golimumab and adalimumab. By contrast, approximately two-thirds of the patients in the third epoch (n=36) received apremilast as a first-line treatment, with infrequent adoption of ustekinumab and secukinumab.
The authors reported that in the first epoch, patients had a longer duration of disease prior to participating in the study. SJC scores, TJC scores, and CRP levels decreased across the 3 epochs. Skin and nail psoriasis were seen more frequently in the first epoch compared with the second or third epoch. Patient age, weight and BMI, and ESR were unchanged across the 3 epochs. Dactylitis and enthesitis were relatively infrequent with no statistically significant differences observed among epochs.
Drug survival curves were similar in epochs 1 and 2, both for biologics administered as monotherapy and given in combination with a conventional DMARD. Mean biologic survival among patients treated with monotherapy was 12.7 years in the first epoch vs 9.69 years in the second epoch. For patients who received biologics given in combination with a conventional DMARD, mean survival durations were 10.9 years and 9.2 years in the first and second epochs, respectively. The authors concluded that in epochs 2 and 3, patients were escalated to treatment with biologics earlier in the disease process and with less severe symptoms than were seen in the first study epoch.
Discussion
Allan Gibofsky, MD, JD
Kevin Yip, MD
Genna Braverman, MD
Allan Gibofsky, MD, JD, professor of medicine at Weill Cornell Medicine and an attending rheumatologist at the Hospital for Special Surgery, and his colleagues, Kevin Yip, MD, and Genna Braverman, MD, who are currently pursuing clinical fellowships in rheumatology, discuss this study and the clinical implications of published research in this area.
Allan Gibofsky, MD, JD: I think that this is nice work by Ishchenko and colleagues, looking at the evolution of patient characteristics in the area of biologic treatment for PsA over 3 distinct time periods. Furthermore, the authors analyzed cohort data that were specifically collected for research purposes. This type of data collection is becoming increasingly useful and more familiar to us, with every disease entity that we encounter as rheumatologists.
This study represents the investigators’ attempt to understand the characteristics of patients with PsA — when and how they were treated as well as the duration of certain treatments — to elucidate the implications for future care. With that overview, Kevin, please highlight the methods and results.
Kevin Yip, MD: Sure. These researchers evaluated all patients within the BioSPAR registry at KU Leuven. This was a prospective cohort of patients with either axial spondyloarthritis or psoriasis who had been treated with biologics or targeted synthetics such as apremilast and Janus kinase (JAK) inhibitors. As you mentioned, the authors effectively divided the cohort into 3 treatment epochs. I quite liked that these epochs were characterized based on the use of specific medications during these 3 distinct time periods.
The investigators collected an impressive amount of data on intake. In addition to race, age, sex, ethnicity, and so on, they report all of the potentially useful clinical data that a rheumatologist could think of in this setting — including SJC, TJC, different ways of looking at enthesitis, dactylitis, global assessments, as well as the usual inflammatory markers — and the kind of information that we would use to gauge the disease status of any patient. Then they try to understand who these patients were on a more granular scale: What kind of medications have they tried? How many patients took glucocorticoids? How many received NSAIDs? From that snapshot, the authors then assessed what patient characteristics were evident when therapy was initiated with either a biologic or a targeted synthetic.
On the whole, the statistical analysis seemed reasonable given the study findings. It is interesting that the 3 epochs do not represent equal periods of time. The first is approximately 6 years long, the second lasts 10 years, and the third is only 2 years. The number of patients per group is also quite different, and if you break it down on a per-year basis, you see an out-of-proportion number of patients recruited in the third epoch compared with the first and second epochs. That raises the question of how and why that happened. Were more people being diagnosed during the third epoch? Were more people being started on therapy with biologics? Furthermore, given that this third epoch group is part of the larger cohort entered into the BioSPAR registry, I would quite like to know what proportion of the global registry they represented. What about patients in that cohort who had axial spondyloarthritis or skin psoriasis without joint involvement and were not starting a biologic? What proportion do they represent and how has that changed over these 3 time periods?
AG: I agree. The investigators do pose almost as many questions as they attempt to answer and you have identified some important ones to consider. We will get to a discussion of how representative the 3 cohorts are given the time periods identified by the authors. Notably, while they do comment on the agents available for use during each epoch, I do not recall patient stratification in terms of previous therapy or the duration or durability of each agent, but rather differentiation based only on the global Kaplan-Meier plot. With that in mind, Genna, please tell us what this all means. In other words, address the “so what” question.
Genna Braverman, MD: This is a nice study. It really rounds out what we see in randomized controlled trials (RCTs) by providing a “real-world” snapshot of practice, prescribing patterns, and patients in rheumatology. We will talk about generalizability a bit later when we are considering the study’s strengths and weaknesses, namely to what degree the findings can be extrapolated to the management of patients in our clinic in New York City, for example. That said, the results are encouraging because biologics have revolutionized the treatment of rheumatic diseases, particularly PsA.
We have fairly new guidelines from the American College of Rheumatology and when I am in clinic, I refer to these 2018 guidelines frequently.2 Therefore, this is a timely moment to evaluate how the treatment of PsA is changing. Again, when you think about the population entering a RCT, these are patients who, generally speaking, have more severe disease and potentially represent a more homogenous group than patients in daily practice. The benefit of the cohort studied by Ishchenko and colleagues is that it is going to show us how real cohorts of patients are being treated over time. Besides providing us with a real-world picture, the current study is also much more longitudinal compared with a RCT. Although it does not necessarily show us some of the harder clinical outcomes in terms of response to treatment, the authors do address that in their therapy survival statistics, which complement the demographic picture and provide an informative, real-world practice perspective.
AG: I may have missed this, but did the investigators enroll sequential patients who presented, or just patients who met the criteria for their BioSPAR research data? I did not really see clarification of that in the enrollment criteria, and clearly it makes a difference if you are enrolling some patients but not others. The other difference that came to mind — and Genna, you certainly touched on it — is the question of whether the results are extrapolatable. All we are told is that these patients were eligible for reimbursement of their biologic under the Belgian health care system and that is a far cry from the hoops and ladders we encounter to get a patient started on a biologic, a targeted synthetic, or even a conventional synthetic in some instances. So that should be kept in mind as well. Kevin, what are your thoughts about this study’s strengths?
KY: First, I want to get clarification on the question of eligibility for reimbursement of biologics. In their discussion, the investigators wonder if one reason they found such a high rate of conventional DMARD use in these patients is that patients needed to have undergone treatment with DMARDs in order be eligible for therapy with the biologics. When you look at the relevant tables in this study, however, only about 60% to about 70% of patients had received conventional DMARDs. That indicates a missing 30% or 40%, where patients were eligible for a DMARD but had not been treated with one.
AG: Even in our system, yes, that is a prerequisite, as you well know. However, there is an exemption for patients who are allergic or intolerant, such that they would not be required to be on a conventional synthetic as a prerequisite for getting the biologic. That is something we do not know in this study, Kevin.
KY: Right. Regarding the strengths of this study, reiterating what Genna said, 18 years of data all from a single center, are evaluated for better or worse. I see this as a strength. For example, one would expect the demographics on intake to be fairly consistent, assuming that their prescribing practices or department viewpoints on use of these biologics stay fairly similar over time. I quite like the fact that there are a lot of data and the data are well defined. I like how the authors have distinguished groups based on the different kinds of DMARDs with which they are being treated. In addition, for the first cohort and somewhat for the second one, there is a long period of follow-up. Those are all good details and then the authors present a very nice-looking survival curve. It might be useful if they could break it down further to see at which points they were getting dropout.
AG: Kevin, is it a lot of patients? There were 407 patients in the cohort from which they extracted 185 in an 18-year period, so we end up with about 9 to 11 patients per year?
KY: That is a fair point. It was a good longitudinal timeline. In terms of sample size, it’s a bit weak from that perspective.
AG: Genna, what would you like to add?
GB: So, to evaluate this study from the side of weaknesses — and not to be too negative because as I mentioned, this was a really nice enterprise — there are some issues, with generalizability being one that we have touched upon in our discussion. This always comes up when you read a paper. You think about whether the findings can be applied to your own practice. Another weakness is that this single-center study is set in Europe, which is a very different health care system. For example, in order to qualify for biologic therapy under the Belgian system, patients had to have erosive arthritis. Axial involvement was not reimbursed. Then, looking at the number of patients in the cohort, as you were saying, Dr Gibofsky, I think 185 is not terrible, but breaking it down over the 18-year study period, the number is not as robust as one might hope for, just to be a little bit “greedy.” In addition, in terms of the demographic factors, this is a White, relatively homogenous population. I would have been very interested to know just a bit more about some of the other socioeconomic determinants of health.
AG: Yes, you are absolutely right. If we think about the point that the investigators are trying to make about clinical characteristics, disease duration, and disease activity being more pronounced in the first group than in the second and third, the findings make sense. When all you had was conventional synthetics and then the biologics came along, there was a virtual explosion of use in rheumatology patients who had no other alternatives and were generally not doing well, having lived with their disease for a long time. We are now seeing that patients are certainly treated earlier because of our recognition over the last 25 years since the introduction of biologics that the earlier you treat, the better the outcome. These days, we do not let patients with very active disease remain in that state. In rheumatoid arthritis, for example, a patient may undergo a 6-month trial of methotrexate — or maybe a bit longer after the dose builds up — and is switched to a different agent at the end of that period of time. Clearly, that paradigm seems to have been followed in the 2018-2019 recommendations for treatment of PsA2 and in the GRAPPA [Group for Research and Assessment of Psoriasis and Psoriatic Arthritis] recommendations as well3: Clinicians should not wait that long to switch agents, because we now have many therapeutic alternatives that are going to yield the best outcome for a given patient.
Here is another thing about generalizability: I am not even sure that, were we to evaluate our population of our patients at Hospital for Special Surgery, it would be the same as the population of patients seen at New York University or the population of patients at University of California, Los Angeles, because of different admixtures of racial migrations in different parts of our country, which is a United States. In contrast, in Belgium there is a relatively restricted background of genetic drift. Their population is more homogeneous than the population seen at our center in New York or at many other centers around the country. We know, for example, that the population of patients with lupus in our center is different from that in Johns Hopkins, which is different from similar populations elsewhere.4,5 Do you have other perspectives to share, Kevin?
KY: Yes, I want to discuss a few more specific weaknesses of this study. While I appreciate what the investigators are trying to do, it is difficult to extrapolate the findings to difference clinical settings. For example, clinicians today are more comfortable using biologics because they have been on the market for a while. The question then arises regarding whether changes in prescribing reflect rheumatologists’ increased comfort level with the anti-TNFs? We can consider, for example, that between the second and third prescribing periods or epochs, there was a massive jump into the use of apremilast and patients were barely receiving treatment with anti-TNFs. I wonder why. Did rheumatologists think that apremilast was a cheaper drug? Did the prescribing change occur because it became more available? Did rheumatologists think apremilast was better for treating milder disease? The answer could be any of the above possibilities. In the end it’s a hard to identify what drove the change in preference for certain agents over time.
I would be surprised if the authors’ local guidelines for reimbursement stayed stable for the last 20 years. I know, for example, the NICE [National Institute for Health and Clinical Excellence] guidelines6 were stating that treatment with 1 DMARD needs to fail to generate a response and then it was 2 DMARDs needs to fail. The guidance changed over the course of a decade and eventually, patients for whom treatment with 1 DMARD failed could try an anti-TNF drug. Then the guidance changed again so that it became appropriate for selected patients to receive anti-TNF agents as first-line therapy. In other words, there are many different confounding factors to consider rather than just the question of whether it is really the base population. As I alluded to before, I would be very interested to see how sizeable this group of patients was, as a proportion of the entire cohort of PsA patients that the authors treated, and to see how that changed over time. Was the proportion of cases higher, such that patients were receiving biologics earlier? I think that would be very a useful data point. Then, as Genna mentioned previously, what about the contribution of other simple comorbidities? Smoking is just one example: we know that the prevalence of smoking has decreased over the last few decades. We also know that smoking is a huge risk factor for worsening psoriasis. Perhaps, then, some of the results are accounted for by the simple possibility that smoking-cessation efforts were super-effective in the local area? We do not know. Those are just some thoughts that came to mind.
AG: Right, Kevin, we do not know anything about comorbidities in this population and we do not know why certain drugs were chosen over others. I was intrigued by Figure 1, showing the Kaplan-Meier curves of drug survival in epochs 1 and 2. They seem to be superimposable, which is generally not our experience. You know, the rule of thumb seems to be, at least in rheumatoid arthritis, that the more agents you have, the less time you keep a patient on any particular agent. Back in the old days, when all we had were 3 anti-TNFs, drug survival on each drug was very long. Now that we have about a dozen biologics and targeted synthetics to choose from, it is unusual for a patient to be treated longer than 6 months with any one of them. Thus, I was intrigued by the drug-survival result.
GB: I agree! I was floored to see that the mean survival was about a decade for both of these time periods. It made me wonder, again, about that issue of generalizability: What is it about the Belgian health care system that may underlie this finding? What are the criteria for switching agents? How challenging is it to get approval for a different biologic? You know, it just does not seem to square with data from other countries, even in Europe. I think the authors mentioned in their discussion studies of a Danish cohort and then a British cohort. There is also a retrospective analysis from France and I saw an evaluation of a French claims database7 that looked at 3-year survival for patients with psoriasis and PsA, which is about 40%. Therefore, survival findings in the study by Ishchenko et al do seem to be a bit of an outlier. They make me wonder whether there is something different about the authors’ center and/or the Belgian health care system as a whole. I think my job as a fellow would be very easy in clinic if my patients had a mean drug survival of a decade!
AG: Yes, well, it would be easier for all of us. It would make my job as an attending easier as well if I only had to interact with my fellows once every 10 years, pardon my humor!
KY: The authors do mention that. They point out that, for better or worse, all patients were from their single-center university cohort. Also, the authors are the same experienced academic clinicians who had worked at this center for 20 years. I do wonder to what extent their findings are driven by the prescribing practices of these 2 physicians and what influenced those practices? Maybe they got more experienced with these drugs because the follow-up is from beginning to end over the study period. Maybe they liked what they saw, and maybe they did not; it would be very interesting to see how those patients’ disease experiences developed. What would happen to the other half who did not withstand up to 50 years of that same treatment? What would they do next?
GB: Also, while this is not a stated goal of this study, this is sort of the closest we get to an outcome measure. It does make you wonder how the patients are actually doing on their biologic therapy. When we are thinking about those drug survival data, are they doing well on their drugs or are they not doing well but still continuing on them?
AG: Yes, this is an interesting observation, but do you think this really changes our current practice or is it simply a look back in time to see what happened 18 years ago vs what we are doing now? Did you need this paper to tell you that patients should be seen and treated earlier to attain a more optimal outcome?
KY: While I agree that this study represents an interesting historical commentary, we all know traditional DMARDs did not do the best job with this kind of disease. When the biologics came along, patients fared much better. We also already knew that it was hard to obtain good clinical result with DMARDS, and to this day still somewhat is, albeit less so than before.
AG: Genna and Kevin, can you each offer 2 takeaways?
GB: I think one takeaway is that this study confirms what we would suspect about prescribing patterns over time. As biologic therapies come on market, the earlier adopters tend to have more severe disease. Another takeaway is that there is something about this university hospital in Leuven, such that patients experience excellent drug survival.
KY: We are seeing that, with the passage of time, rheumatologists seem to have become more comfortable using these agents for patients with less severe disease. That probably speaks somewhat to how bad the traditional DMARDs are or how much more available these biologics now are.
AG: You are both right. I will make an additional observation which is that younger physicians tend to be more aggressive and willing to use newer medications earlier than some of us older “fossils,” who tend to be set in our ways. So that might be something to factor in as well. I do think that the study by Ishchenko and colleagues was very interesting. Rather than changing practice, I believe it falls into the category of “Where were we then vs where are we now?”
This article was edited for clarity and length.
Disclosure
Allan Gibofsky, MD, JD, reported affiliations withAbbVie, Inc.; Acquist Therapeutics, Inc.; Amgen Inc.; Biosplice Therapeutics, Inc.; Boehringer Ingelheim Pharmaceuticals, Inc.; Eli Lilly and Company; Pfizer Inc.; Janssen Therapeutics; and Regeneron Pharmaceuticals, Inc.
References
1. Ishchenko A, Joly J, Neerinckx B, Lories R, de Vlam K. Evolution of patient characteristics in the era of biologic treatment of psoriatic arthritis: 18-year Belgian experience from the Leuven Spondyloarthritis Biologics Cohort (BioSPAR). Rheumatol Adv Pract. 2021;5(3):rkab085. doi:10.1093/rap/rkab085
2. Singh JA, Guyatt G, Ogdie A, et al. 2018 American College of Rheumatology/National Psoriasis Foundation guideline for the treatment of psoriatic arthritis. Arthritis Rheumatol. 2019;71(1):5-32. doi:10.1002/art.40726
3. Coates LC, Corp N, van der Windt DA, O’Sullivan D, Soriano ER, Kavanaugh A. GRAPPA treatment recommendations: 2021 update. J Rheumatol. 2022;jrheum.211331. doi:10.3899/jrheum.211331
4. Alarcon GS, McGwin Jr G, Petri M, Reveille JD, Ramsey-Goldman R, Kimberly RP; PROFILE Study Group. Baseline characteristics of a multiethnic lupus cohort: PROFILE. Lupus. 2002;11(2):95-101. doi:10.1191/0961203302lu155oa
5. Redmond C, Pamuk O, Hasni SA. Lupus cohorts. Rheum Dis Clin. 2021;47(3):P457-P479. doi:10.1016/j.rdc.2021.04.009
6. Samarasekera E, Sawyer L, Parnham J, Smith CH. Assessment and management of psoriasis: summary of NICE guidance. BMJ. 2012;345:e6712. doi:10.1136/bmj.e6712
7. Vegas LP, Penso L, Claudepierre P, Sbidian E. Long-term persistence of first-line biologics for patients with psoriasis and psoriatic arthritis in the French health insurance database. JAMA Dermatol. Published online March 23, 2022. doi:10.1001/jamadermatol.2022.0364
Posted by Haymarket’s Clinical Content Hub. The editorial staff of Rheumatology Advisor had no role in this content’s preparation.
Reviewed April 2022
