SEAM-PsA: Seems like methotrexate works in Psoriatic Arthritis?

In this issue of Arthritis and Rheumatology, Mease and colleagues report the results of an important trial, ‘SEAM-PsA’, evaluating the efficacy of methotrexate monotherapy, etanercept monotherapy, or methotrexate/etanercept in combination, in early psoriatic arthritis (PsA).(1) The primary endpoint in this study was the American College of Rheumatology 20% Response Criteria (ACR20) and minimal disease activity (MDA) was a key secondary/co-primary endpoint. MDA requires achievement of five of seven points of low disease activity including swollen and tender joint counts, psoriasis, enthesitis, patient global assessment, patient pain assessment and the health assessment questionnaire. Among patients in the etanercept monotherapy arm, 61% achieved ACR20 and 36% achieved MDA. This was very similar to the combination therapy arm (65% and 36% respectively). Both etanercept monotherapy and etanercept+methotrexate combination therapy were significantly more efficacious than the methotrexate monotherapy arm (ACR20 51% and MDA 23%).

The SEAM-PsA trial brings to the forefront several important concepts that remind us of where we are in the management of PsA in 2019. This trial represents many “firsts” in PsA: the first appropriately-powered randomized double blind trial to test a head-to-head comparison of therapies (although more head-to-head trials are in progress); the first large RCT not to include a placebo arm; the first study to include MDA as a key secondary (co-primary) outcome and one of the first trials to examine the management of early PsA and initial treatment selection. SEAM-PsA represents one of a few handful of trials to infer methotrexate monotherapy does indeed have benefit in the treatment-naïve patient with PsA, although without a placebo control this is not possible to confirm definitively.

Methotrexate in PsA: a mixed history

While the benefits of methotrexate alone and in combination with tumor necrosis factor inhibitor (TNFi) therapy have been clearly demonstrated in RA,(2, 3) the same cannot be said for PsA.(4) Studies of methotrexate monotherapy in PsA have been largely underwhelming(5–7) or at least mixed(8–11).

The Methotrexate in Psoriatic Arthritis (‘MIPA’) study, a key placebo-controlled trial of methotrexate published in 2012, found no significant advantage of methotrexate over placebo in ACR20.(5) On the other hand, the TIght COntrol in PsA (TICOPA) study in the United Kingdom, a treat to target study which used methotrexate as a backbone therapy (at higher doses than in MIPA), found benefits with oral DMARDs frequently used in combination to achieve ACR20 responses.(10, 12). In the open label RESPOND trial which compared infliximab plus methotrexate to methotrexate monotherapy, an ACR20 response was achieved in 66.7% of patients in the methotrexate arm, but the combination therapy arm was significantly more effective than monotherapy.(8)

Aside from the (3) trials mentioned above, support for methotrexate in PsA has mainly been drawn from observational data including one study demonstrating similar persistence on methotrexate among patients with RA and PsA in Norway(11) and another demonstrating that initiation of methotrexate resulted in an ACR20 response in 41% of patients with active PsA in clinical practice.(13) In fact, the relative lack of efficacy of methotrexate in RCTs led to the ACR/National Psoriasis Foundation (NPF) PsA Treatment Guidelines deviating from prior treatment recommendations to conditionally recommend the use of a TNFi biologic over methotrexate or other oral small molecules in the patient with active treatment naïve PsA.(14)

SEAMs like methotrexate works? Etanercept as active comparator

Taken even at face value, a methotrexate monotherapy arm ACR20 response rate of 50.7% should bring some enthusiasm. Etanercept may serve as a very useful “active comparator” in informing our understanding of methotrexate efficacy in this study. Given the breadth of studies to date for which we have efficacy data on etanercept in PsA, etanercept efficacy here mirrored that seen in the previous pivotal phase 3 trial, among others.(15) With that anchor, the methotrexate monotherapy in SEAM-PSA did relatively well: ACR20 50.7%, MDA 22.9%, enthesitis resolution in 43.1% and psoriasis was clear or almost clear in 66.3% at 24 weeks. Even with a potentially high placebo rate (placebo rates are presumed to be higher in trials enrolling treatment naïve patients who know they will receive a treatment), these values might suggest that we not close the door on the clinical story around methotrexate in the treatment of psoriatic disease. Again, however, it is unfortunately impossible for us to tell with certainty the effect of MTX monotherapy without a placebo arm in SEAM-PsA.

What was different in this study compared to prior methotrexate studies?

What was different in previous studies, in particular the MIPA trial, has been the under-dosing of methotrexate and under-powering of a study for the expected response.(16) In SEAM-PsA, the median oral methotrexate dose was 20mg, a dose considered appropriate by most experts, whereas MIPA target dose was 15mg weekly with as many as 11% on lower doses.

The future of methotrexate in PsA?

Even though methotrexate seems to have ‘worked’ in SEAM-PsA, only 50% of patients starting methotrexate improved by 20%, (ACR20) and only 30% improved by 50% (ACR50) at 24 weeks (compared to 61% and 44% for patients initiating etanercept). These results were produced within the ideal setting of a typical phase III trial in which adherence to therapy is generally higher than what we expect in the ‘real world.’ It is safe to say we still have a long way to go in achieving better outcomes. Furthermore, at 48 weeks, there is suggestion of more radiographic progression among those treated with methotrexate (although few patients progressed in either arm). Thus, there is still limited evidence for methotrexate halting radiographic progression in PsA; this requires further exploration.

Additionally, the results of SEAM-PsA suggest that, unlike combination therapy in RA, there was no meaningful benefit to combination therapy in PsA. That being said, SEAM-PsA does not test the potential benefit of combination therapy in prolonging persistence of etanercept or effect on etanercept blood levels and cannot be necessarily extrapolated to other TNFi mechanisms given etanercept’s unique structure as a fusion protein.(17–19)

With regard to patient satisfaction, more limited tolerability was seen with methotrexate including nausea and vomiting, and from clinical experience, we believe more fatigue would likely have been reported if it had been measured. There was otherwise no appreciable difference in infection risk or other major adverse events between etanercept and methotrexate. A further consideration is that of methotrexate safety in the long term among a patient population with a known high prevalence of obesity, fatty liver disease and metabolic co-morbidities. (20–22) In the era of multiple effective target therapies for PsA, these factors should be weighed in the patient-physician discussion of appropriate treatment options.

Still a lot to learn…

The SEAM-PsA trial will continue to contribute to our knowledge of PsA management over the coming months and years as important sub-analyses are performed. This trial of over 800 patients is an ideal dataset in which to address many questions including predictors or biomarkers of treatment response among methotrexate and TNFi users, and how to design comparative efficacy and/or effectiveness trials. For example, the ACR20 is the primary outcome used in all PsA regulatory trials but fails to capture important elements in PsA, including enthesitis, axial disease and skin psoriasis. The gap between ACR response criteria and MDA in this study underscores the difference in measured domains of disease in a psoriatic disease trial. Subsequent analyses in this trial will help us understand whether one of the alternative outcome measures may in fact be a better primary outcome measure for future comparative efficacy trials.

SEAM-PsA greatly underscores the need for more head-to-head trials in PsA to inform treatment guidelines, comparative effectiveness, and cost research. Of equal importance are more comparative effectiveness studies in the ‘real world’ to appropriately inform clinical practice.(23, 24) SEAM-PsA enrolled patients typical of trial patients but that represent less than half of our patients in clinical practice.(25) This trial was mainly composed of patients with a polyarticular phenotype (with an average of 12 swollen joints and 20 tender joints). In contrast, pragmatic trials enroll patients representative of those seen in clinical practice - approximately half of which would have an oligoarticular phenotype.(25) Such trials are needed to inform the treatment of ‘real-world’ patients, including phenotypic subsets of disease that may not be captured in traditional trials or by traditional outcome measures extrapolated from RA such as the ACR response criteria. We still have very little data to inform how to treat the patients with PsA that we see in our day to day practices.

In summary, SEAM-PsA represents an important piece of the puzzle in our larger understanding of treatment approaches to PsA. We learned that combination therapy offers little benefit over etanercept monotherapy and gained important, albeit inferred and without placebo control, insights into the efficacy of methotrexate monotherapy. These data should invigorate the community to take a deeper look into methotrexate efficacy, effectiveness and safety/tolerability in our patients with psoriatic disease; such data are certain to inform future treatment guidelines. We hope that this is a turning point in trial design and drives the initiative for more head-to-head trials, pragmatic trials and other novel clinical trial designs in PsA. It is from these direct comparisons that we actually learn how to treat our patients.