A Matching-Adjusted Indirect Comparison of Guselkumab and Secukinumab in Patients with Psoriatic Arthritis Over 52 Weeks

Abstract

Introduction

Studies evaluating the long-term comparative efficacy between biologic therapies for psoriatic arthritis (PsA) are scarce. Two biologic therapies, guselkumab and secukinumab, were evaluated up to 52 weeks in a mixed patient population (biologic-naïve and biologic-experienced patients).

Methods

An unanchored matching-adjusted indirect comparison (MAIC) was conducted to compare guselkumab 100 mg every 8 weeks (Q8W) and every 4 weeks (Q4W) versus secukinumab 150 mg Q4W and 300 mg Q4W on American College of Rheumatology (ACR) and Psoriasis Area and Severity Index (PASI) responses from weeks 4 through 52 using pooled individual patient-level data from guselkumab trials (COSMOS, DISCOVER-1 and -2) and pooled summary-level data from secukinumab trials (FUTURE 2, 3, 4, and 5). For the primary analysis, patients from the guselkumab trials were re-weighted on six clinically relevant baseline characteristics to match those in the secukinumab trials. Additional characteristics were included as a sensitivity analysis. A scenario analysis was conducted in a biologic-naïve patient population only.

Results

For the mixed population, both guselkumab doses initially had numerically or significantly lower ACR 20 responses than both secukinumab doses prior to weeks 12–20; however, from weeks 12–24 onward, ACR 20 responses became numerically or significantly higher for guselkumab. For PASI 90 responses, both guselkumab doses showed significantly higher responses than both secukinumab doses at weeks 24 and 52. Notably, at 52 weeks, ACR 20 and PASI 90 responses for both doses of guselkumab were numerically or significantly higher than both doses of secukinumab. Results from the sensitivity and scenario analyses were similar to the primary analysis.

Conclusions

While the IL-17A inhibitor secukinumab may demonstrate more rapid and greater efficacy before weeks 12–20, both doses of guselkumab provide similar or greater efficacy on joint and skin outcomes compared to both doses of secukinumab from week 24 onward. This study provides valuable insights for treatment decisions when considering the chronic nature of PsA.

Supplementary Information

The online version contains supplementary material available at 10.1007/s40744-025-00771-9.

Key Summary Points

Why carry out this study?

This study fills a critical gap by providing comparative efficacy data between guselkumab and secukinumab, two biologic therapies for psoriatic arthritis (PsA), in the absence of direct comparisons, aiding clinical decision-making.

By focusing on outcomes at 52 weeks, this study provides valuable insights into the sustained efficacy of these therapies in a mixed biologic-naïve and biologic-experienced PsA population.

An analytical method known as matching-adjusted indirect comparison (MAIC) was used to align patients from different trials so that treatments could be compared.

What was learned from the study?

Although secukinumab 150 mg and 300 mg every 4 weeks (Q4W) may work faster and show greater joint efficacy initially (before weeks 12–20), guselkumab 100 mg every 8 weeks (Q8W) and Q4W offer similar or greater joint efficacy compared to both secukinumab doses from week 24 onward.

For skin improvement in the PsA population, both doses of guselkumab consistently outperformed both doses of secukinumab at 24 and 52 weeks.

This study provides valuable information for clinicians and patients when choosing between these biologic therapies for a chronic condition like PsA.

Introduction

Psoriatic arthritis (PsA) is a chronic immune-mediated inflammatory disease characterized by joint inflammation and skin involvement [1]. The long-term management of PsA is of critical importance to alleviate symptoms, prevent irreversible joint damage, and enhance the overall quality of life for patients [2]. To achieve these goals, having safe and effective treatment options is crucial, and their ability to maintain a sustained response over time is equally vital given the chronic nature of PsA. However, treatment selection remains challenging despite the growing number of available therapies for PsA, as comparative clinical data are lacking [3, 4].

In the absence of head-to-head trials, indirect treatment comparisons (ITCs) have emerged as a valuable method to assess comparative efficacy between treatments to aid clinical decision-making [4]. Previous ITCs in PsA, including network meta-analyses (NMA), have primarily focused on short-term outcomes [5, 6], leaving a significant gap in the understanding of long-term treatment efficacy in this chronic disease. In this context, matching-adjusted indirect comparisons (MAICs), a type of population-adjusted ITC, have gained prominence as a methodology that matches individual patient-level data from one trial to published summary-level data from a comparator trial to estimate treatment differences [7–10]. Methodological guidelines have been published describing their appropriate use [7, 11]. Unanchored MAICs are conducted when the trials compared lack a common comparator, whereas anchored MAICs use a common comparator across trials to connect the network [11]. Since most RCTs allow patients receiving placebo to switch to active treatment (i.e., placebo is no longer a common comparator between trials), unanchored MAICs can be used to compare long-term data beyond the placebo-controlled period [12]. Examining treatment comparisons beyond the relatively short placebo-controlled period will help determine whether treatment efficacy is sustained over the long term, which is critical for managing a chronic condition like PsA.

Two biologic interleukin (IL) inhibitors for PsA, guselkumab (an IL-23 p19 subunit inhibitor) and secukinumab (an IL-17A inhibitor), have both shown efficacy in their respective clinical trials when it comes to joint and skin outcomes. These assessments were based on the responses observed in terms of the American College of Rheumatology (ACR) and Psoriasis Area and Severity Index (PASI) measures [13–21]. However, no head-to-head trials have been conducted between these two treatments in PsA. Therefore, the objective of this study was to indirectly compare the relative efficacy of guselkumab versus secukinumab on joint and skin efficacy over 52 weeks in a mixed biologic-naïve and biologic-experienced PsA population.

Methods

Overview of Data Sources

A structured literature review was conducted to identify randomized controlled trials (RCT) of guselkumab and secukinumab in PsA. RCTs including treatment doses recommended by the European Medicines Agency were considered for the analyses. These included guselkumab 100 mg every 8 weeks (Q8W) and 100 mg every 4 weeks (Q4W) [22] and secukinumab 150 mg and 300 mg at weeks 0, 1, 2, 3, and 4, followed by monthly maintenance dosing of Q4W [23].

Upon review of the RCTs, it was determined that there were three distinct patient populations: biologic-naïve, biologic-experienced, and a mixed population of biologic-naïve and biologic-experienced. The secukinumab FUTURE trials included mixed populations, and patient baseline characteristics stratified by biologic-exposure were not available. As such, given that a more robust data set was available for the mixed population, the primary MAIC analyses were conducted in the mixed population. Studies included in the primary analyses for the mixed population were the guselkumab trials DISCOVER-1, DISCOVER-2, and COSMOS [18–20, 24, 25], and the secukinumab trials FUTURE 2, FUTURE 3, FUTURE 4, and FUTURE 5 [13–16, 26–31]. DISCOVER-1 included a mixed population, DISCOVER-2 included only biologic-naïve patients, and COSMOS included patients who were inadequate responders to tumor necrosis factor (TNF) inhibitors. In the primary analysis, the proportion of biologic-naïve/biologic-experienced patients were adjusted for, where possible. The biologic-naïve secukinumab trial EXCEED was excluded from the primary analysis because it only included one dose of secukinumab (i.e., 300 mg), one endpoint (i.e., ACR 20), and one assessment time point (i.e., week 52). Given the limited data available for EXCEED, a scenario analysis was conducted for the biologic-naïve population that included EXCEED, DISCOVER-2, and a subset of patients from DISCOVER-1 [18, 19, 21, 24, 25]. Since the FUTURE trials included mixed populations, the trials were not included in the scenario analysis. A flow diagram outlining the primary, sensitivity, and scenario analyses is presented in Supplementary Figure S1. Overviews of the trials are presented in Supplementary Tables S1 and S2.

Two secukinumab trials were excluded from all analyses: (1) FUTURE 1 because the doses studied are not included in the European Medicines Agency drug labels [17, 23]; and (2) MAXIMISE because it specifically included only patients who had PsA with axial manifestations [32].

MAIC Methods

An unanchored MAIC was selected as the most appropriate approach to estimate the relative efficacy of guselkumab to secukinumab because the trials assessed lacked a common comparator in the long term. Unanchored MAICs were conducted based on methods outlined by the NICE Decision Support Unit Technical Support Document 18 [11]. The primary outcomes of interest were ACR 20 and PASI 90 responses analyzed through 52 weeks of follow-up.

For the mixed population, individual patient-level data for guselkumab Q8W and Q4W were pooled from DISCOVER-1, DISCOVER-2, and COSMOS [18–20, 24, 25], and compared to pooled summary-level data for secukinumab 150 mg and 300 mg from FUTURE 2, FUTURE 3, FUTURE 4, and FUTURE 5 for the primary outcomes of interest ACR 20 and PASI 90 [13–16, 26–31]. Because COSMOS only included guselkumab Q8W, it was excluded from all analyses that assessed guselkumab Q4W [20]. Analyses for ACR 50 and ACR 70 were also conducted between guselkumab and secukinumab; the secukinumab trials with outcomes for ACR 50 include FUTURE 2, 3, 4, and 5 and FUTURE 2, 4, and 5 for ACR 70.

For the biologic-naïve population scenario analysis, individual patient-level data for guselkumab Q8W and Q4W were individually pooled from DISCOVER-2 and a subset of patients from DISCOVER-1 who were biologic-naive [18, 19, 24, 25] and compared to summary-level data for secukinumab 300 mg from EXCEED [21] for ACR 20 at week 52. As noted previously, secukinumab 150 mg data were not available in the biologic-naïve population and PASI 90 data were not available for either secukinumab dose.

Eligibility criteria were first aligned between the cohorts following the MAIC methodology [11]. Guselkumab patients who did not meet the eligibility criteria of the secukinumab trials were removed from the individual patient-level data. The remaining guselkumab patients were then re-weighted to address any imbalances in clinically relevant characteristics compared to the pooled summary-level data from the secukinumab trials. All available prognostic factors deemed clinically relevant to include for adjustment in the analyses were pre-selected and ranked based on key European guidelines and expert opinion [33–35]. In the primary analysis, these characteristics included prior use of anti-TNF inhibitors, Disease Activity Score-28, number of tender joints, number of swollen joints, enthesitis status, and psoriasis (PsO; ≥ 3% of body surface area [BSA]). For the sensitivity analysis, additional characteristics including sex, age, dactylitis, and PsA pain were included if reported in the trials. In the scenario analysis for the biologic-naïve population, characteristics included Disease Activity Score-28, number of tender joints, number of swollen joints, CRP > 10 mg/l, enthesitis status, PsO (≥ 3% of BSA), and baseline PASI score.

Propensity scores, representing the inverse odds of being in the guselkumab treatment group versus the secukinumab group, were used as weights in the matching process [8]. The propensity scores were estimated using the generalized method of moments. Estimates for the relative treatment effect of guselkumab compared to secukinumab, including odds ratios (OR), 95% confidence intervals (CIs), and p values (significance was assessed at p < 0.05), were derived from a weighted logistic regression model. Non-responder imputation methods were applied to handle missing data that resulted from patient treatment discontinuation. All analyses were performed using SAS software (version 9.4, SAS Institute Inc., Cary, NC, USA).

Ethics/Ethics Approval

This article is based on previously conducted studies and does not contain any new studies with human participants or animals performed by any of the authors.

Results

Patient Population

Baseline patient characteristics for the mixed population in the primary and sensitivity analyses are presented in Table 1. Before matching, key differences between trial populations included a slightly more severe PsA population in the guselkumab trials (e.g., higher Disease Activity Score-28, greater number of swollen and tender joints). There was a lower percentage of female patients receiving guselkumab Q4W compared to patients receiving secukinumab and a greater proportion of patients with PsO in the guselkumab trials [18–20, 24, 25] compared to the secukinumab trials [13–16, 26–31].

Table 1.

Baseline characteristics for the mixed population

	Pooled GUS 100 mg Q8W	Pooled GUS 100 mg Q4W	Pooled SEC 150 mg Q4W	Pooled SEC 300 mg Q4W
Sample (N)	564	373	572	461
Baseline characteristics informing the primary analysis
No prior anti-TNF (%)	59	90	69	68
1 or more prior anti-TNF (%)	41	10	31	32
Disease Activity Score-28 (mean)	5.0	5.0	4.7	4.6
Number of tender joints (of 78 joints; mean)a	23.3	23.9	22.0	19.9
Number of swollen joints (of 76 joints; mean)a	12.7	13.2	11.3	9.9
Enthesitis (%)	63	65	65	61
Psoriasis (≥ 3% of BSA; %)	78	81	53	47
Additional characteristics included to inform the sensitivity analysis
Sex (% female)	50	44	52	51
Age (mean)	47.2	46.5	48.5	48.6
Dactylitis (%)	40	43	33	38
Psoriatic arthritis pain (mean)	62.9	60.5	–	54.5

Note: the first seven characteristics were matched in the primary analysis (no prior anti-TNF and one or more prior anti-TNF were both included in the first match). The remaining four characteristics were included in the sensitivity analysis, if reported in the trial

BSA body surface area, GUS guselkumab, Q4W every 4 weeks, Q8W every eight weeks, SEC secukinumab, TNF tumor necrosis factor

^aFor the guselkumab trials, the mean number of tender joints was out of 68 and the mean number of swollen joints was out of 66

Baseline patient characteristics for the biologic-naïve population in the scenario analysis are presented in Supplementary Table S3. Before matching, key differences between trial populations included a slightly more severe PsA population with greater CRP levels and a higher proportion of patients with PsO in the guselkumab trials [18, 19, 24, 25] compared to EXCEED [21].

Mixed Population

Primary Analysis

Characteristics informing the primary analysis included prior use of anti-TNF inhibitors, Disease Activity Score-28, number of tender joints, number of swollen joints, enthesitis status, and PsO (≥ 3% of BSA).

ACR 20 Responses

Guselkumab Q8W vs. Secukinumab 150 mg: At weeks 4 and 8, ACR 20 responses were significantly lower for guselkumab Q8W. From week 16 onward, ACR 20 responses were numerically or significantly higher for guselkumab Q8W (week 52: 66.0% vs. 59.3%; OR = 1.33, 95% CI 1.01, 1.76; Fig. 1a).

Fig. 1.

Observed and matching-adjusted ACR 20 responses for guselkumab 100 mg Q8W versus a secukinumab 150 mg Q4W and b secukinumab 300 mg Q4W. Note: No patients were excluded from the IPD data. *The OR for GUS (adjusted) vs. SEC is statistically significant (P ≤ 0.05). Characteristics matched: prior anti-TNF inhibitor use, disease activity score-28, tender joint count (of 68 joints), swollen joint count (of 66 joints), enthesitis, and psoriasis affecting ≥ 3% of body surface. ACR American College of Rheumatology, CI confidence intervals, GUS guselkumab, IPD individual patient data, OR odds ratio, Q4W every 4 weeks, Q8W every 8 weeks, SEC secukinumab, TNF tumor necrosis factor

Guselkumab Q8W vs. Secukinumab 300 mg: ACR 20 responses were numerically or significantly lower for guselkumab Q8W until week 20, then numerically or significantly higher from week 24 onward (week 52: 64.8% vs. 64.6%; OR = 1.01, 95% CI 0.74, 1.37; Fig. 1b).

Guselkumab Q4W vs. Secukinumab 150 mg: ACR 20 responses were significantly lower for guselkumab Q4W at week 4 and 8, but numerically or significantly higher from week 12 onward (week 52: 68.3% vs. 59.3%; OR = 1.48, 95% CI 1.00, 2.19; Fig. 2a).

Fig. 2.

Observed and matching-adjusted ACR 20 responses for guselkumab 100 mg Q4W versus a secukinumab 150 mg Q4W and b secukinumab 300 mg Q4W. Note: No patients are excluded from the IPD data. *The OR for GUS (adjusted) vs. SEC is statistically significant (P ≤ 0.05). Characteristics matched: prior anti-TNF inhibitor use, disease activity score-28, tender joint count (of 68 joints), swollen joint count (of 66 joints), enthesitis, and psoriasis affecting ≥ 3% of body surface. CI confidence intervals, GUS guselkumab, IPD individual patient data, OR odds ratio, PASI Psoriasis Area and Severity Index, Q4W every 4 weeks, SEC secukinumab, TNF tumor necrosis factor

Guselkumab Q4W vs. Secukinumab 300 mg: ACR 20 responses were numerically or significantly lower for guselkumab Q4W from weeks 4 to 16, but numerically higher from week 24 onward (week 52: 68.7% vs. 64.6%; OR = 1.20, 95% CI 0.78, 1.84) (Fig. 2b).

ACR 50 Responses

Results for ACR 50 responses mirrored those for ACR 20 (Supplementary Figures S2 to S5), with a similar pattern of numerically or significantly lower ACR 50 responses for both guselkumab doses compared to both secukinumab doses at earlier timepoints (week 28 or earlier), followed by generally numerically higher responses through to week 52.

ACR 70 Responses

Initial responses (up to week 24) for ACR 70 showed a similar pattern to ACR 20 and 50 responses (Supplementary S6 to S9), with numerically or significantly lower ACR 70 responses for both guselkumab doses compared to both secukinumab doses. Both guselkumab doses had numerically higher ACR 70 responses from week 28 through to week 52 compared to secukinumab 150 mg, except at week 52, where ACR 70 response for guselkumab Q4W was slightly lower than secukinumab 150 mg. There were no data for secukinumab 300 mg between weeks 24 and 52, but at week 52, ACR 70 responses for both guselkumab doses were only slightly lower than secukinumab 300 mg.

PASI Responses

PASI 90 data were only available at weeks 16, 24, and 52 for secukinumab 150 mg Q4W, and at weeks 24 and 52 for secukinumab 300 mg Q4W.

Guselkumab Q8W vs. Secukinumab 150 mg and 300 mg: Guselkumab Q8W had numerically higher PASI 90 responses than secukinumab 150 mg at all timepoints assessed, with statistical superiority at weeks 24 and 52 (71.7% vs. 43.1%; OR = 3.35, 95% CI 2.43, 4.60; Fig. 3a). Similarly, guselkumab Q8W had significantly higher PASI 90 than secukinumab 300 mg at both assessed weeks 24 and 52 (71.0% vs. 56.3%; OR = 1.90, 95% CI 1.33, 2.71; Fig. 3b).

Fig. 3.

Observed and matching-adjusted PASI 90 responses for guselkumab 100 mg Q8W versus a secukinumab 150 Q4W and b secukinumab 300 mg Q4W. Note: No patients were excluded from the IPD data. *The OR for GUS (adjusted) vs. SEC is statistically significant (P ≤ 0.01). Characteristics matched: prior anti-TNF inhibitor use, disease activity score-28, tender joint count (of 68 joints), swollen joint count (of 66 joints), enthesitis, and psoriasis affecting ≥ 3% of body surface. CI confidence intervals, GUS guselkumab, IPD individual patient data, OR odds ratio, PASI Psoriasis Area and Severity Index, Q4W every 4 weeks, Q8W every 8 weeks, SEC secukinumab, TNF tumor necrosis factor

Guselkumab Q4W vs. Secukinumab 150 mg and 300 mg: Like guselkumab Q8W, guselkumab Q4W had numerically higher PASI 90 responses than secukinumab 150 mg at all timepoints assessed, with statistical superiority at weeks 24 and 52 (75.5% vs. 43.1%; OR = 4.07, 95% CI 2.74, 6.06; Fig. 4a). Compared to secukinumab 300 mg, guselkumab Q4W had significantly higher PASI 90 responses at weeks 24 and 52 (74.9% vs. 56.3%; OR = 2.31, 95% CI 1.49, 3.59; Fig. 4b).

Fig. 4.

Observed and matching-adjusted PASI 90 responses for guselkumab 100 mg Q4W versus a secukinumab 150 Q4W and b secukinumab 300 mg Q4W. Note: No patients are excluded from the IPD data. *The OR for GUS (adjusted) vs. SEC is statistically significant (P ≤ 0.02). Characteristics matched: prior anti-TNF inhibitor use, disease activity score-28, tender joint count (of 68 joints), swollen joint count (of 66 joints), enthesitis, and psoriasis affecting ≥ 3% of body surface. CI confidence intervals, GUS guselkumab, IPD individual patient data, OR odds ratio, PASI Psoriasis Area and Severity Index, Q4W every 4 weeks, SEC secukinumab, TNF tumor necrosis factor

Sensitivity Analysis

The sensitivity analysis for the mixed population matched for characteristics informing the primary analysis and additional characteristics, including sex, age, dactylitis, and PsA pain, if reported. The sensitivity analysis demonstrated similar results to the primary analysis for ACR 20 response (Supplementary Figures S10, S11, S12, and S13). In general, both doses of guselkumab had numerically or significantly higher ACR 20 responses compared to secukinumab 150 mg from week 16 onward and secukinumab 300 mg from weeks 24–28 onward.

For PASI 90 response (Supplementary Figures S14, S15, S16, and S17), results from the sensitivity analysis were consistent with the primary analysis with significantly higher responses for both doses of guselkumab versus both doses of secukinumab at weeks 24 and 52, with one exception. In contrast to the primary analysis, numerical, but no statistical superiority was achieved for guselkumab Q4W versus secukinumab 300 mg at week 24 (57.3% vs. 46.9%; OR = 1.51, 95% CI 0.98, 2.33) (Supplementary Figure S17).

Scenario Analysis for the Biologic-Naïve Population

Characteristics informing the scenario analysis for the biologic-naïve population included Disease Activity Score-28, number of tender joints, number of swollen joints, CRP > 10 mg/l, enthesitis status, PsO (≥ 3% of BSA), and baseline PASI score.

Compared to secukinumab 300 mg, ACR 20 responses for guselkumab Q8W (66.6% vs. 66.9%; OR = 0.98, 95% CI 0.67, 1.44) were comparable, while ACR 20 responses for guselkumab Q4W were numerically higher (71.6% vs. 66.9%; OR = 1.25, 95% CI 0.84, 1.85) (Supplementary Figure S18a and b).

Discussion

Psoriatic arthritis is a chronic condition necessitating long-term therapeutic intervention [36]. However, despite the need for long-term comparative studies to assess treatment efficacy and inform clinical decision-making, there are few head-to-head trials in PsA (e.g., EXCEED [21], SPIRIT-H2H [37], and BE BOLD [38]). In the absence of long-term head-to-head trials, the use of MAICs can provide valuable insights into comparative efficacy over time. MAICs leverage individual patient-level data to account for cross-trial differences, including study designs and prognostic factors [7, 8].

Overall, prior to week 12–20, guselkumab had lower ACR responses than secukinumab; however, beyond this period, responses were similar or greater. In the primary analysis for the mixed population, both guselkumab doses initially had numerically or significantly lower ACR 20 responses than both secukinumab doses prior to week 12–20. From week 12–24 onward however, both doses of guselkumab had numerically or significantly higher ACR 20 responses than both secukinumab doses. A similar pattern to ACR 20 responses was shown for ACR 50 responses for both doses of guselkumab and secukinumab. For ACR 70 responses, although the initial pattern of responses was similar, with numerically or significantly lower responses for guselkumab up to week 24, responses at week 52 remained slightly lower for guselkumab compared to secukinumab. For PASI 90 responses, both guselkumab doses showed significantly higher responses compared to both secukinumab doses at weeks 24 and 52. In the scenario analyses examining ACR 20 responses at week 52 in patients who were biologic-naïve, guselkumab Q8W had comparable responses to secukinumab 300 mg, while responses for guselkumab Q4W were numerically higher. In general, guselkumab efficacy was similar and robust across the sensitivity analysis that included additional clinically relevant baseline characteristics.

These findings are generally consistent with and complement the results of ITCs (i.e., NMAs) that examined joint and skin efficacy up to 24 weeks. A Bayesian NMA by Mease et al. 2023 in a mixed patient population with PsA demonstrated that both guselkumab doses were comparable to secukinumab 150 mg for ACR 20 response at 12–24 weeks; while secukinumab 300 mg showed higher responses than guselkumab Q8W but not Q4W [5]. Similarly, this MAIC showed lower ACR 20 responses for both guselkumab doses up to week 12 compared to both secukinumab doses; however, over time, both guselkumab doses were comparable or better than both secukinumab doses. For PASI 90 responses in the NMA, both guselkumab doses were superior to secukinumab 150 mg, while only guselkumab Q4W showed statistically significant superiority over secukinumab 300 mg [5]. The slight variability in these findings compared to the MAIC is likely due to differences between the analytical approaches used (i.e., NMA versus MAIC), period of assessment, and inclusion of different trials. Furthermore, another Bayesian NMA conducted by Song and Lee et al. 2021 up to 24 weeks, found that both guselkumab doses showed comparable efficacy to both secukinumab 150 mg and 300 mg for ACR 20 [27]. Additionally, both guselkumab doses had the highest probability of achieving a PASI 75 response versus both secukinumab doses and placebo, based on cumulative ranking curve analysis [27]. These findings during the induction period align well with the present study, in the absence of long-term head-to-head comparisons in PsA.

While conducting an NMA to evaluate the long-term comparative efficacy of guselkumab and secukinumab was not feasible, a MAIC offered a valuable alternative for assessing the long-term efficacy of these treatments. Such methodologically appropriate analyses can provide long-term, indirect comparisons of continuously treated active patients by leveraging individual patient-level data from a treatment’s RCT [39]. The analyses demonstrated that joint and skin efficacy were maintained over one year in patients receiving guselkumab and secukinumab, which is important for a chronic disease like PsA [2]. The consistency in the outcomes between the primary analyses and the sensitivity and scenario analyses demonstrate the robustness of the results.

There are also limitations associated with the analyses. The use of an unanchored MAIC assumes that all relevant prognostic factors and treatment effect modifiers were included in the analysis. However, adjustments cannot be conducted for unobserved differences in trial populations or study designs. To account for this, the comparability of the guselkumab and secukinumab trials was assessed to ensure alignment of patient eligibility criteria across studies. While the included trials were generally comparable across most reported summary characteristics, the adjustment via MAIC improved the alignment between trial populations by removing or reducing observed cross-trial differences [8]. In PsO, the reliability of data generated using an unanchored MAIC has been demonstrated in a case study comparing unanchored MAIC results, using the same methodology as the current analysis, with the results of a head-to-head trial [12]. This analysis demonstrated that comparative results between guselkumab and secukinumab in PsO were consistent between the unanchored MAIC and the head-to-head trial (ECLIPSE), supporting the reliability and accuracy of MAIC methods [12]. Another limitation is related to the data availability for secukinumab. Summary-level data were pooled across the secukinumab trials, and study designs and patient populations were assumed to be similar. Data for secukinumab were also limited to the mixed population for all trials except EXCEED, and data for EXCEED were limited (i.e., one dose, outcome, and timepoint). A scenario analysis that included EXCEED was conducted for the biologic-naïve population, but comparative data between guselkumab and secukinumab 150 mg, as well as PASI response data, were lacking. As well, there were no biologic-experienced data available for the secukinumab trials. Thus, it is difficult to draw reliable conclusions about the comparative efficacy in these patient populations. As such, additional studies, including those using real-world data [40], are needed to provide insight into long-term comparative efficacy for patients with and without previous exposure to biologic therapies. Future studies should also consider assessing outcomes associated with other domains of PsA, including enthesitis, dactylitis, and axial disease, as well as patient-reported outcomes.

Conclusions

While the IL-17A inhibitor secukinumab may demonstrate more rapid and greater efficacy before week 12–20, both doses of guselkumab provide similar or greater efficacy on joint and skin outcomes compared to both doses of secukinumab from week 24 onward. Notably, at 52 weeks, ACR 20 and PASI 90 responses for both doses of guselkumab were numerically or significantly higher than both doses of secukinumab. Effective and sustained treatments are critical for a chronic and life-long disease like PsA, where improvements in skin and joint efficacy also enhance patient quality of life. This analysis aims to provide valuable insights into the treatment landscape of PsA and to assist clinicians and patients in making informed decisions regarding treatment selection for this chronic and debilitating condition.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

Suzy van Sanden contributed to the conception, design, data acquisition, conducted statistical analyses, interpreting the results, and critically revised the manuscript. Agata Schubert and Fareen Hassan contributed to the conception, design, data acquisition and critically revised the manuscript; Barkha P. Patel contributed to drafting, interpreting the results, and critically revising the manuscript; Miriam Zimmermann contributed to critically revising the manuscript. All authors gave their final approval and agreed to be accountable for all aspects of the work.

Funding

Sponsorship for this study and the Rapid Service Fee were funded by Johnson & Johnson Research and Development (HEMAR Department, High Wycombe, United Kingdom). Johnson & Johnson Innovative Medicine also participated in the study design, review, and approval of the publication.

Data Availability

The datasets generated and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Declarations

Conflict of Interest

Suzy van Sanden, Agata Schubert, Miriam Zimmermann, and Fareen Hassan are employees of Johnson & Johnson Innovative Medicine. Barkha P. Patel is an employee of EVERSANA, which received funding from Johnson & Johnson Innovative Medicine for this study.

Ethical Approval

This article is based on previously conducted studies and does not contain any new studies with human participants or animals performed by any of the authors.