Effectiveness of Golimumab as Second Anti-TNFα Drug in Patients with Rheumatoid Arthritis, Psoriatic Arthritis and Axial Spondyloarthritis in Italy: GO-BEYOND, a Prospective Real-World Observational Study

In this prospective observational study, data were collected from 34 rheumatology clinics in Italy in patients with rheumatoid arthritis (RA), psoriatic arthritis (PsA) and axial spondyloarthritis (axSpA) who started golimumab (GLM) as a second anti-TNFα drug. The primary objective was to evaluate the effectiveness of GLM after 6 months. Changes in quality of life using the EQ-5D-5L were also assessed. A total of 194 patients aged 53.2 ± 12 years started GLM as a second anti-TNF drug: 39 (20.1%) with RA, 91 (46.9%) with PsA and 64 (32.9%) with axSpA. After 6 months of GLM treatment, 68% of RA patients achieved low disease activity (LDA; DAS28-CRP ≤ 3.2), 31.9% of PsA patients achieved minimal disease activity and 32.5% of axSpA patients achieved LDA (ASDAS-CRP < 2.1). Good/moderate EULAR response was achieved in 61.9% and 73.8% of patients with RA and PsA, respectively, and 16% of axSpA patients achieved a 50% improvement in BASDAI. Across all indications, improvements in disease activity measures and EQ-5D-5L domains were observed over 6 months. The main reasons for GLM interruption were lack/loss of efficacy (7.2%) or adverse events (2%). This study confirms the effectiveness of GLM as a second-line anti-TNF for the treatment of RA, PsA and axSpA in a real-world setting in Italy.


Introduction
Following the availability of antibodies targeting tumor necrosis factor-alpha (TNFα), the standard of therapeutic care offered to patients with rheumatoid arthritis (RA), psoriatic arthritis (PsA) and spondyloarthritis (SpA) has substantially improved in terms of controlling signs/symptoms of inflammation, quality of life and functional outcome [1][2][3][4].
Whilst anti-TNF drugs are the most frequently used first-line bDMARDs, the response can decline over time in some patients, necessitating the switch to another biologic [5][6][7][8][9][10]. Data from real-life settings show that as many as 40-50% of patients stop their first anti-TNF treatment due to loss of therapeutic effectiveness; this is typically seen after a good initial response or due to adverse events (AEs) [8,9].
Recent guidelines for the management of RA, PsA and axial SpA (axSpA) have acknowledged the need for switching due to first-line failure and have been updated accordingly. In patients with RA, the European Alliance of Associations for Rheumatology (EULAR) guidelines recommend that if one anti-TNF therapy fails, patients may receive another anti-TNF agent or another biologic with a different mode of action [10]. Similarly, in patients with axSpA, EULAR guidelines recommend switching to another anti-TNF drug or an anti-IL-17 agent if anti-TNF treatment fails [5]. In patients with PsA who fail to respond to a biological disease-modifying antirheumatic drug (bDMARD), it is recommended to switch to another anti-TNF therapy or another bDMARD [6].
Among the other four licensed anti-TNF inhibitors approved for the treatment of RA, PsA and axSpA (adalimumab, infliximab, etanercept and certolizumab pegol), golimumab (GLM) has also been shown to be effective in RA [11], PsA [12], ankylosing spondylitis (AS) [13], and non-radiographic axSpA [14,15] from randomized clinical trials (RCTs). GLM also has the longest dosing intervals (once monthly) among subcutaneous anti-TNF drugs. Besides improving signs and symptoms in these patients, a high rate of retention on GLM is achieved in biological-naïve patients, with about 70% maintaining treatment through 5 years [16].
The efficacy of GLM after failure with other anti-TNF drugs was evaluated in the prospective GO-AFTER study [17]. GLM was shown to be effective and safe for patients with RA who had failed one or more anti-TNF drugs, with results confirmed through 5 years in the long-term extension study [18].
However, in patients with PsA or axSpA, data on the effectiveness of GLM following the first-line anti-TNF failure are scarce [19,20]. To address this gap, recent real-life retrospective studies performed in Spain [21], Turkey [22], France [23,24], Germany [25] and Italy [26] have examined the long-term effectiveness and retention rate of GLM. The present prospective study, GO-BEYOND, was designed to evaluate the effectiveness of GLM as a second anti-TNF drug in patients with RA, axSpA or PsA who discontinued a first anti-TNF drug.

Patients and Study Design
GO-BEYOND was a prospective observational study including patients with a diagnosis, by the treating rheumatologists, of RA, PsA or axSpA. Patients were enrolled between July 2017 and December 2019, with visits at baseline, 3 and 6 months. The main analysis for this manuscript evaluated the effectiveness after 6 months of GLM therapy. All patients fulfilled 2010 classification criteria for RA [27], or CASPAR criteria for PsA [28] or ASAS criteria for axial (radiographic and non-radiographic) SpA [29]. Patients were treated across 34 Italian Rheumatology Centers with GLM after first-line anti-TNFα inhibitor failure. Inclusion criteria were: male or female patients ≥18 years diagnosed with active RA, PsA or axSpA, eligible for treatment with GLM according to the SmPC [30]; patients previously treated with an anti-TNFα inhibitor with or without methotrexate (MTX) for at least 6 months and experiencing anti-TNFα treatment failure due to loss of efficacy; patients who stopped an effective initial anti-TNFα treatment for at least 3 months due to discomfort or tolerability issues after taking treatment; agreement from female subjects to use adequate contraception to prevent pregnancy. The following contraindications to GLM according to the Summary of product characteristics (SmPC) [30] prevented patients from being included in the study: moderate or severe heart failure according to New York Heart Association Class III/IV; tuberculosis or other severe infections such as sepsis, abscesses and other opportunistic infections; hypersensitivity or clinically serious adverse drug reaction to the active substance or to any of the excipients. Other exclusion criteria included: current or previous cancer in the past 5 years and current participation in an interventional trial or another non-interventional study, excluding registries. The study (protocol number: MK-8259-6415) was approved by the Regional Ethics Committee for Clinical Studies from the Tuscany Region (Sezione, Area Vasta Sud Est) on 20 March 2017. Local ethics committee approval from all participating centers and written informed consent for the anonymous use of personal data were also obtained from every patient, in compliance with Legislative Decree 196/2003. This study complies with the ethical standards laid down in the 1975 Declaration of Helsinki.
The GO-BEYOND database recorded demographic features at the baseline visit (age, sex, and time since RA, PsA or axSpA diagnosis) and at each visit (visits 0, 3, and 6 months) the following measures: Disease Activity Score in 28 joints (DAS28) [31] for patients with RA and PsA, Ankylosing Spondylitis Disease Activity Score based on CRP (ASDAS-CRP) [32,33], Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) [34], the Assessment of SpondyloArthritis international society Health Index (ASAS HI) [35] for patients with axSpA and health-related quality of life (HRQoL) assessed using the EuroQoL 5-Dimension 5-Level (EQ-5D-5L) [36] for all three patient groups. For peripheral joint assessment, 68 joints were assessed for tenderness, and 66 joints were assessed for swelling (for patients with RA and PsA). Rheumatoid factor (RF), anti-citrullinated protein antibody (ACPA) (for patients with RA) and human leukocyte antigen B27 (HLAB27) were also measured at baseline (for patients with axSpA).

Outcome Measures
The primary outcome of the present study was to assess the effectiveness of GLM treatment after 6 months. In patients with RA, this was defined as the proportion (%) of patients achieving low disease activity (LDA; DAS28-CRP ≤ 3.2) according to EULAR guidelines [37]. For patients with PsA, the proportion of patients achieving MDA was evaluated. The MDA criterion is a score of 7 outcome measures in PsA, and patients were classified with MDA when 5 out of 7 outcome measures were fulfilled as previously described [38]. The effectiveness of GLM in axSpA was defined as the proportion of patients achieving at least LDA according to ASDAS-CRP (<2.1) [32,33]. In patients with axSpA, we measured BASDAI 50, defined as a 50% improvement or more of the baseline BAS-DAI after treatment with GLM. The EULAR response criteria were used to evaluate the individual change in DAS28-CRP from baseline and the level of DAS28-CRP reached at 6 months in patients with RA and PsA classified as good responders, moderate responders or non-responders [39]. HRQoL was assessed using the EQ-5D-5L [36]. Discontinuation of treatment and causes thereof were also recorded over the 6-month treatment period.

Sample Size Calculation
Based on the prevalence of RA, PsA and axSpA and the propensity of switching between anti-TNFα therapy in clinical practice (higher for RA) [19,40], it was estimated to enroll approximately 100 RA patients, 85 PsA patients and 65 axSpA patients. However, a greater number of axSpA patients and a lower number of RA patients were observed in clinical practice after a review of the feasibility of enrolment in each study center. Considering a drop-out rate of 10% expected to be lost to follow-up, we estimated the final number of patients to be 200: 40 with RA, 85 with PsA and 75 with axSpA should have been enrolled in about 30 investigational sites treating rheumatologic diseases in Italy. Assuming a response rate of 30% for each disease and the aforementioned sample sizes, the width of the 95% confidence interval (CI) is 0.3 for RA, 0.19 for PsA and 0.21 for axSpA. Sample size calculations were performed using PASS software (NCSS, LLC, Kaysville, UT, USA).

Statistical Analysis
Continuous variables are presented as mean and standard deviation, median, 25th and 75th percentiles. Categorical variables were summarized using absolute and relative frequencies (percentages). Percentages were calculated together with the corresponding 95% confidence interval computed using Clopper-Pearson method [41]. Data were initially evaluated for normal distribution with the Shapiro-Wilk test, and then the differences in disease activity scores at 3 and 6 months compared to baseline values were assessed using the paired-samples t-test or the paired-samples Sign Test, as appropriate. Differences at baseline and 6 months in the proportion of patients with problems across the 5 domains of the EQ-5D-5L (mobility, self-care, usual activities, pain/discomfort and anxiety/depression) were tested using the McNemar Test. Missing data for the calculation of disease-specific measures (e.g., BASDAI score, ASAS HI score) were handled according to the recommendation given in the literature for the specific score. A p-value of <0.05 was considered statistically significant, and all analyses were performed using SPSS statistical software, version 23.0 (SPSS, Chicago, IL, USA). There were no changes in the planned analyses of the study due to the COVID-19 pandemic.
81.9%). At baseline, the median DAS28-CRP score was 3.92 (interquartile range, IQR: 3.5-4.79) and decreased over time to 2.72 (IQR: 2.07-3.46; p < 0.001) at 6 months ( Figure 1A). Mean SJC, TJC and PGA significantly decreased at 6 months compared to baseline values. CRP levels were also observed to decrease after 6 months, but this decrease did not reach statistical significance. ESR levels remained unchanged after 6 months compared to baseline values (Table 4).

Effect of Concomitant Methotrexate Treatment in PsA and AxSpA Patients
Sub-analysis was also performed to evaluate the efficacy of GLM on the primary efficacy outcomes in PsA and axSpA patients with and without MTX at 6 months. In PsA patients, we observed that MDA was achieved in 22.6% (7 out of 31) patients with MTX compared to 39% (16 out of 41) patients without MTX; this difference did not achieve statistical significance (p = 0.14). In axSpA patients, we observed that LDA (according to ASDAS-CRP) was achieved in 22.2% (2 out of 9) patients with MTX compared to 35.5% (11 out of 31) patients without MTX; this difference did not achieve statistical significance (p = 0.69). This sub-analysis suggests that concomitant MTX did not influence the efficacy of GLM in patients with PsA and axSpA in our cohort.

Reasons for Discontinuation
During the 6-month treatment period, a total of 29 patients (14.9%) interrupted the study; of whom 19 (9.8%) discontinued due to a definitive interruption of GLM treatment, 8 were lost to follow-up, 1 interrupted due to lack of compliance and 1 for an unspecified reason. The mean time when interruption occurred was 1.96 ± 1.7 months after commencing GLM treatment. Of the 19 GLM definitive treatment interruptions, the majority were due to an issue with efficacy, with 11 (5.7%) reporting as lack of therapeutic effect and 3 (1.5%) reporting as loss of efficacy. Four treatment interruptions were due to adverse events, with two (1%) of these AEs considered related to GLM (both events occurring in patients with PsA) and two (1%) considered unrelated to GLM (both events occurring in patients with axSpA). The remaining treatment interruption was due to an unspecified other reason (Table 6). Overall, patients with RA had the highest proportion of study interruptions; however, none of the discontinuations due to AEs occurred in these patients. AxSpA-axial spondyloarthritis; EQ-5D-5L index-EuroQoL 5-Dimension 5-Level (index); EQ-VAS-EuroQol Visual Analogue Scale; PsA-psoriatic arthritis; RA-rheumatoid arthritis. Statistically significant differences compared to baseline values are represented by p-values in bold text. Data presented as frequency and %. AEs related to GLM * 2 (1) 0 (0) 2 (2.2) 0 (0) AEs not related to GLM ** 2 (1) 0 (0) 0 (0) 2 (3.1) Other 1 (0.52) 0 (0) 1 (1.1) 0 (0) AE-adverse events; axSpA-axial spondyloarthritis; GLM-golimumab; PsA-psoriatic arthritis; RA-rheumatoid arthritis. * AEs related to GLM included eczema for one patient and epistaxis and erythema for the second patient. ** AEs not related to GLM were lung cancer and nosocomial infection.

Discussion
The main findings from the present real-life prospective study in Italy confirm the effectiveness of GLM as a second-line anti-TNFα for the treatment of RA, PsA and axSpA. In this analysis, with up to 6 months of treatment, GLM was found to be effective in achieving primary outcome measures to a similar extent in patients with RA, PsA and axSpA.
This real-life prospective study specifically focused on the effectiveness of GLM in a subgroup of patients who are not normally included in clinical trials, and our findings reveal that in patients with spondyloarthropathies (axSpA and PsA) as well as RA, who need to discontinue their first anti-TNF drug, the decision to switch to GLM can be considered a suitable therapeutic choice. These findings are in agreement with results reported in other recent real-life studies [21][22][23][24][25][26].
While several observational studies have focused on the long-term drug survival of GLM given as a second-line treatment in inflammatory arthritis, few studies have specifically evaluated the effectiveness of GLM in this setting. In a retrospective longitudinal study performed in Spain by Alegre-Sancho and colleagues, the long-term effectiveness and persistence of GLM as a second biological drug in patients with PsA (N = 131) and axSpA (N = 79) were evaluated [21]. In PsA patients, DAS28 decreased from 4.0 ± 1.3 at baseline to 2.5 ± 1.2 at 3 months. In axSpA patients, BASDAI score at baseline was 5.5 ± 2.1 and decreased to 3.9 ± 2.0 at 3 months. In our cohort, which shares similar clinical characteristics with the Spanish study in terms of disease duration, comorbidities and concomitant medication, we observed a similar reduction in DAS28-CRP by 3 months in PsA patients (3.8 ± 0.99 at baseline to 2.8 ± 1.1; p < 0.0001) as well as a decrease in BASDAI in axSpA patients (6.02 ± 2.09 at baseline to 4.9 ± 2.2; p < 0.001), although baseline disease activity was higher in our axSpA population. Patients with RA were not included in the study by Alegre-Sancho et al.; however, a similar study (in terms of design and patient characteristics) undertaken in Turkey (GO-BEYOND Turkey) did include RA patients [22]. Akar et al. evaluated persistence and changes in disease activity measures over 2 years in 60 RA and 269 axSpA patients after GLM treatment. Although the majority of patients in this study were biologic naïve, their analysis did stratify a subset of patients who were biologic experienced. In this subgroup of the RA population (N = 7), DAS28-CRP decreased from 4.8 to 3.0 after 6 months, similar to what we have observed in our larger cohort in Italy (N = 39; from 4.1 at baseline to 2.8 at 6 months; p < 0.001). In patients with axSpA previously treated with an anti-TNF drug (N = 28), a two-fold reduction was seen for ASDAS and BASDAI after 6 months, a greater improvement than what we have observed. However, patients with axSpA were considerably younger (40 vs. 51 years) and presented with lower CRP levels, ASDAS and BASDAI scores at baseline.
In the recent real-life retrospective analysis of the Italian GISEA registry, the effectiveness and persistence of GLM in biologic inadequate responder (IR) patients with RA, PsA and axSpA were investigated [26]. Disease activity in the subgroup of RA patients who were IR to 1 biologic (N = 94) improved after 6 months, as seen by a marked reduction in DAS28 from 4.9 ± 1.2 at baseline to 3.7 ± 1.1 at 6 months. These patients had a higher burden of disease activity at baseline compared to patients included in our study (DAS28 of 4.9 vs. 3.7), but the absolute decrease (1.2 units) was identical to what we observed. A good EULAR response was achieved in 52% of RA patients who had previously failed 1 biologic compared to 61.9% in our cohort. In the GISEA registry study, among PsA patients with axial involvement previously failing 1 biologic, 41% had BASDAI < 4 at 6 months. In patients with axSpA, the percentage who achieved BASDAI 50 at 6 months was 40%, while approximately 60% of all patients with SpA achieved LDA, and 23% were in remission according to ASDAS. Compared to GISEA, our findings are favorable in terms of improvement in disease activity and the clinical response observed in patients with RA and PsA; however, clinical outcomes in patients with axSpA as measured by BASDAI 50 and LDA were less favorable. In our study, patients with axSpA tended to be older (mean 51 vs. 46 years) and presented with a higher burden of comorbidities (76.6% vs. 41%), which may, in part, explain these observed differences.
The post hoc analysis of the prospective GO-NICE study undertaken in Germany also evaluated the effectiveness of GLM by line of treatment in patients with RA, PsA and AS [25]. In patients with RA (N = 104) given GLM as a second-line biologic, DAS-28 decreased from 4.9 ± 1.3 to 3.4 ± 1.5 and 19.5% were in remission at 6 months. The Psoriatic Arthritis Response Criteria [42] in patients with PsA showed a 51.2% improvement at 6 months and BASDAI score decreased from 4.9 ± 2 to 3.3 ± 2.2 in AS patients. Similar results were also seen in the prospective post hoc analysis of the prospective GO-PRACTICE study performed in France in second-line biologic patients with axSpA, where BASDAI was observed to decrease to a similar extent (from 5.7 to 3.9) at 6 months [24]. Results from GO-NICE and GO-PRACTICE corroborate with our findings up to 6 months in terms of DAS-28 improvement in patients with RA (approximately 30% reduction vs. baseline). A similar improvement was also seen in the axSpA group for BASDAI despite the fact that our cohort included slightly older patients (51 ± 13.2 years vs. 45.3 ± 12.3 years and 46.8 ± 12.3 years for GO-NICE and GO-PRACTICE, respectively).
Collectively, results from the present study, in conjunction with these real-life studies, provide evidence on the effectiveness of GLM as a second-line biologic in patients with RA, PsA and axSpA for up to 6 months.

Study Limitations and Strengths
The main limitation of this study is intrinsic to its observational design, thus potentially affecting the results of the analysis.
Strengths of this study include the prospective design, with the accuracy of data collection according to systematic protocol-specified visits at baseline, 3 and 6 months and the enrolment of a very selected study population including only patients with RA, PsA or axSpA who were switched to GLM after previous failure with another anti-TNFα.

Conclusions
Considering the high rate of failure following first-line anti-TNF treatments, rheumatologists are frequently faced with the need to consider a switch in anti-TNF as second-line therapy. In this respect, the availability of real-life data such as those in the present study provide useful information to guide therapeutic decisions. In this large real-life prospective cohort, the use of GLM after first-line anti-TNF failure was found to be an effective and valuable option in patients with RA, PsA and axSpA.
Author Contributions: S.D., A.M.G. and C.B. were involved in the design of the study and data analysis. All other authors were involved in the enrolment of patients, data collection and contributed to drafting the paper/revising it critically for intellectual content. All authors have read and agreed to the published version of the manuscript. Data Availability Statement: Data are available upon reasonable request. All data relevant to the study are included in the article. The datasets generated during or analyzed during the current study are not publicly available. All data used in this study were anonymized to respect the privacy of patients in line with applicable laws and regulations.