Use of TREMFYA in Patients with Active Psoriatic Arthritis Responding Inadequately to TNF Inhibitors (COSMOS)

SUMMARY

• The company cannot recommend any practices, procedures, or usage that deviate from the approved labeling.
• The efficacy and safety of TREMFYA (COSMOS) was evaluated in a randomized, phase 3b, double-blind, placebo (PBO)-controlled, multicenter trial in adult patients with active psoriatic arthritis (PsA) with an inadequate response to 1-2 tumor necrosis factor inhibitors (TNFis). The proportion of patients who achieved a response of ≥20% improvement from baseline in the American College of Rheumatology Criteria (ACR20) at week 24 was 44.4% in the TREMFYA 100 mg group and 19.8% in the PBO group (P<0.001) and they maintained response in the TREMFYA group through week 48.¹
• Through week 24, serious adverse events (SAEs) and serious infections occurred in 3.7% and 0.5% of patients, respectively, in the TREMFYA group (n=189). In the PBO group (n=96), SAEs and serious infections occurred in 3.1% and 0% of patients, respectively. A similar safety profile was observed through weeks 24 and 56.¹
• Data from post-hoc analyses of COSMOS study regarding baseline subgroups and the effect of TREMFYA on minimal clinically important improvement (MCII) in
  patient-reported outcomes (PROs) are summarized below.^2-5

CLINICAL DATA

COSMOS

Coates et al (2021)¹ evaluated the efficacy and safety of TREMFYA in adult patients with active PsA who demonstrated inadequate response to 1-2 TNFis in a randomized, double-blind, PBO-controlled, multicenter, phase 3b study.

Study Design/Methods

• Selected inclusion criteria: patients ≥18 years of age; active PsA; ≥3 swollen joints and ≥3 tender joints for ≥6 months; patients who met ClASsification criteria for Psoriatic ARthritis (CASPAR) at screening; active (≥1 psoriatic plaque of ≥2 cm) or documented history of plaque PsO or current nail PsO; inadequate response or intolerance to 1-2 TNFis.
• Stable baseline use of medications, including methotrexate, sulfasalazine, hydroxychloroquine, leflunomide, oral corticosteroids, and nonsteroidal anti-inflammatory drugs (NSAIDs) or other analgesics were permitted.
• Patients were randomized in a 2:1 ratio to receive either TREMFYA 100 mg or PBO subcutaneously (SC) at weeks 0, 4, and every 8 weeks (q8w) thereafter through week 44.
• The study comprised PBO-controlled (weeks 0-24) and active-treatment (weeks 24-48) periods, with final study intervention at week 44.
• At week 16, patients from either group who experienced <5% improvement from baseline in both swollen and tender joint counts were eligible for early escape (EE) to initiate/increase one of the permitted concomitant medications per investigator discretion. Patients in the TREMFYA group continued the randomized treatment and received PBO at week 16 to maintain blinding, and patients in the PBO group who qualified for EE received TREMFYA at weeks 16, 20, and q8w thereafter through week 44.
• At week 24, all remaining randomized PBO patients crossed over to receive TREMFYA 100 mg at weeks 24, 28, and q8w thereafter through week 44.
• The primary endpoint was an ACR20 response at week 24.
  • Patients with missing data or who met treatment failure (TF) criteria through week 24 (defined as discontinuation of study agent and/or study participation for any reason, initiation or increase in the dose of allowed conventional synthetic disease-modifying antirheumatic drugs (csDMARDs) or oral corticosteroids for PsA, and initiation of protocol-prohibited medications/therapies for PsA or met EE criteria) were considered nonresponders.
  • In the EE-correction analysis, 12 patients in the TREMFYA group did not meet any other TF criteria (ie, introduction/change in the dose of concomitant therapy) through week 24, and their response was included with other patients in the TREMFYA group; 8 patients in the PBO group received TREMFYA as EE therapy at weeks 16 and 20, met TF criteria, and were considered nonresponders.
• Efficacy response was assessed through week 48.
• Multiplicity-controlled secondary endpoints evaluated at week 24 included the following: mean change from baseline in Health Assessment Questionnaire-Disability Index (HAQDI); ≥50% improvement from baseline in the American College of Rheumatology Criteria (ACR50); mean change from baseline in the 36-item Short-Form Health Survey (SF-36 PCS) score; and 100% improvement from baseline in the Psoriasis Area and Severity Index (PASI 100) score in patients with ≥3% body surface area (BSA) psoriasis involvement and an Investigator’s Global Assessment (IGA) score ≥2 at baseline.
• Through week 56, safety was evaluated as events per 100 PY.
• Through week 44, 88% of patients treated with TREMFYA 100 mg completed the study.

Results

Patient Characteristics

• Overall, 285 patients were randomized to receive TREMFYA 100 mg (n=189) or PBO (n=96).
• Baseline demographics are summarized in Table: Summary of Patient Baseline Demographics.

Efficacy

• At week 4, the ACR20 response was significantly higher in the TREMFYA group vs the PBO group (19.0% vs 4.2%; P<0.001; not multiplicity adjusted), with TF criteria applied.
• At week 24, the primary endpoint for ACR20 response rates were 44.4% (84/189) vs 19.8% (19/96) for the TREMFYA group vs the PBO group (P<0.001).
• ACR20 response rates for TREMFYA vs PBO were consistent across subgroups defined by baseline patient and disease characteristics (gender, body weight, swollen joint count [SJC], tender joint count [TJC], C-reactive protein [CRP], enthesitis, and dactylitis) and by prior and concomitant medications (nonbiologic disease modifying antirheumatic drugs [DMARDs]), oral corticosteroids, NSAIDs, prior TNFis, reason for prior TNFi discontinuation). See Table: ACR20 Response at Week 24 by Baseline Characteristics.

• The results for the primary and secondary endpoints are summarized in Table: Summary of Primary and Select Secondary Endpoints at Week 24 and Week 48.

Safety

• The most common adverse events in patients treated with TREMFYA through week 24 included nasopharyngitis (5%) and upper respiratory tract infection (4%), with similar incidence in the PBO group (5% and 3%, respectively).
• Through week 24, SAEs and serious infections occurred in 3.7% and 0.5% of patients in the TREMFYA group (n=189), respectively. In the PBO group (n=96), SAEs and serious infections occurred at 3.1% and 0%, respectively.
• Overall, the safety profile was similar through weeks 24 and 56, with no opportunistic infections, active tuberculosis, anaphylactic/serum sickness-like reaction, confirmed inflammatory bowel disease, or death was reported.
• Through week 56, there was no increase in adverse event rates in the TREMFYA group.
• The safety results are presented in Table: Summary of Safety Results through Week 56.

Post Hoc Analyses

Gossec et al (2026)² conducted a post hoc analysis of COSMOS to assess the effect of TREMFYA on MCII in PROs through week 48 among adult patients with active PsA and inadequate response to 1-2 TNFis.

Study Design/Methods

• The PROs assessed were Functional Assessment of Chronic Illness Therapy-Fatigue (FACIT-F) (≥4-point improvement) at 8 weeks, HAQ-DI (≥0.35-point improvement) at 4 weeks, Dermatology Life Quality Index (DLQI 0/1) (0-30) (at ≥5-point improvement) at 8 weeks, DLQI 0/1 at 8 weeks, and SF-36 PCS score (≥5-point improvement) at 8 weeks.
• TREMFYA was compared with PBO using logistic regression, adjusting for baseline PRO levels, number of prior TNFis used, and concomitant csDMARDs. Missing data and patients meeting TF criteria were assessed through nonresponder imputation.

Results

Patient Characteristics

• The baseline characteristics are summarized in Table: Summary of Patient Baseline Demographics.

Efficacy

• There was a significant difference in the least-squares mean (LSM) of FACIT-F in the TREMFYA vs PBO groups from baseline to week 12 (LSM change from baseline: -11.2 vs
  -4.9, P=0.0233) and week 16 (LSM change from baseline: 4.3 vs 1.7, P=0.0059).
• Patients receiving TREMFYA showed significant improvements compared to patients receiving PBO in HAQ-DI at week 12 (LSM change from baseline:-0.2 vs 0.0, P=0.0005), SF-36 PCS at week 16 (LSM change from baseline: 3.3 vs 0.4, P=0.0002), and DLQI at week 8 (LSM change from baseline: -6.3 vs -1.5, P<0.0001).
• Results of the FACIT-F, HAQ-DI, DLQI 0/1, and SF-36 PCS responses are summarized in Table: Summary of MCII in FACIT-F, HAQ-DI, DLQI 0/1, and SF-36 PCS Score Achievements through Week 48.
• Hazard ratios (HRs) were reported for the time to achieving MCII across PRO assessments and for reaching DLQI 0/1 through week 24 in Table: Summary of HR for Time to Achievement of MCII Responses through Week 24.

Gossec et al (2025)³ evaluated the effect of TREMFYA on composite indices through 48 weeks in adult patients with active PsA and inadequate response to TNFis from the COSMOS trial.

Study Design/Methods

• The post hoc analysis assessed the following outcomes:
  • Disease Activity Index for Psoriatic Arthritis (DAPSA; evaluated at weeks 4, 8, 12, 16, 20, 24, and 48)
  • Disease Activity Score in 28 Joints (DAS28; evaluated at weeks 4, 8, 12, 16, 20, 24, and 48)
  • Psoriatic Arthritis Disease Activity Score (PASDAS) evaluated at weeks 8, 16, 24, and 48)

Results

• The baseline DAPSA, DAS28, and PASDAS scores among the TREMFYA and PBO groups are provided in Table: Selected Disease Characteristics.
• The other baseline demographic characteristics are reported in Table: Summary of Patient Baseline Demographics.

Efficacy

• Efficacy results are summarized in the following tables:
  • The achievements of DAPSA low disease activity (LDA), DAS28 LDA, and PASDAS LDA scores among the TREMFYA and PBO groups at different time points are summarized in Table. Summary of Achievements in DAPSA LDA, DAS28 LDA, and PASDAS LDA Over Time.
  • The achievements in DAPSA and PASDAS VLDA among patients not fulfilling the outcome criteria at baseline are discussed in Table Summary of Achievements in DAPSA and PASDAS VLDA Remission Over Time.

McInnes et al (2024)⁵ prospectively evaluated the effect of TREMFYA on various clinical outcomes through 48 weeks in subgroups of adult patients with active PsA from the COSMOS trial.

Study Design/Methods

• The post hoc analysis assessed the following outcomes from the COSMOS trial at weeks 24 and 48:
  • Effects on joints: ACR20/50/70 responses, enthesitis and dactylitis resolutions (Leeds Enthesitis Index=0 and Dactylitis Severity Score=0)
  • Skin: improvement in PASI 90/100 response and IGA score of 0 in patients with ≥3% BSA and IGA ≥2 at baseline, IGA score of 0/1 response and IGA ≥2 reduction from baseline
  • Patient-report outcomes: ≥4 points improvement from baseline in FACIT-F and ≥0.35 points improvement from baseline in HAQ-DI
  • Composite outcome measures: PASDAS LDA, DAPSA, and MDA
• These outcomes were evaluated by baseline patient age, sex, body mass index, SJC, TJC, PsA duration, % BSA, CRP level, pain Visual Analogue Scale, number of prior TNFi and discontinuation reasons, and csDMARD status.

Results

Efficacy

• ACR20 and ACR50 response rates at weeks 24 and 48 by baseline patient demographics, disease characteristics, prior/ongoing therapies, and reasons for prior TNFi discontinuation are reported in Figure: ACR20 and ACR50 Response Rates at Weeks 24 and 48 by Baseline Subgroups

Literature Search

A literature search of MEDLINE^®, EMBASE^®, BIOSIS Previews^®, and DERWENT^® (and/or other resources, including internal/external databases) was conducted on 23 February 2026.