FormalPara Key Summary Points

Why carry out this study?

As psoriatic arthritis (PsA) is a chronic disease requiring long-term treatment, it is important to assess the safety and efficacy of novel therapies for clinicians and patients to decide upon the best treatment options.

Many patients with PsA cycle through various treatments due to loss of response over time. Therefore, therapies that demonstrate consistent efficacy between biologic disease-modifying antirheumatic drug (bDMARD)-naïve patients and those who experience loss of response or intolerance to tumor necrosis factor inhibitors (TNFi-IR) are of clinical importance.

The safety and efficacy of bimekizumab, which inhibits both IL-17A and IL-17F, was assessed in this study up to 2 years from the BE OPTIMAL (bDMARD-naïve) and BE COMPLETE (TNFi-IR) studies and the BE VITAL open-label extension.

What was learned from the study?

The safety of bimekizumab to 2 years was consistent with previous studies assessing bimekizumab in PsA and psoriasis; no new safety signals were observed with longer-term follow up. Bimekizumab treatment resulted in sustained efficacy response over 2 years in both bDMARD-naïve and TNFi-IR patients.

Bimekizumab is an effective, long-term treatment option for patients with PsA, irrespective of their prior biologic use. The data reported here may assist clinicians and patients with shared decision making with respect to treatment.

Digital Features

“This article is published with digital features, including a slide deck, to facilitate understanding of the article. To view digital features for this article, go to https://doi.org/10.6084/m9.figshare.26379796.”

Introduction

Psoriatic arthritis (PsA) is a chronic, inflammatory disease that affects multiple domains such as joints, skin, and nails [1,2,3]. The impact of PsA on these domains results in impaired patient function and quality of life [4]. As PsA is a chronic condition, there is a need for treatments that deliver long-term efficacy with an acceptable safety profile.

Long-term studies in PsA show the cumulative benefit to patients of maintaining high disease control over the long-term [5,6,7,8]; however, some patients may respond inadequately or lose response, therefore switching or cycling between multiple therapies becomes necessary [9]. Interleukin (IL)-17 family members and IL-23 are pivotal cytokines in the chronic inflammation seen in PsA, and treatments inhibiting IL-‍17A and IL-23 are used to treat patients with this condition [10]. IL-‍17A and IL-12/23 inhibitors have also demonstrated longer persistence than tumor necrosis factor inhibitors (TNFis) or targeted synthetic disease-modifying antirheumatic drugs (DMARDs), and similar persistence to one another [11]. However, approximately 20–50% of patients taking a biologic DMARD (bDMARD) change or discontinue treatment within 1 year [12, 13]. Furthermore, whilst 40% of patients who switched from a TNFi to another bDMARD remained on their initial treatment after 20 months, the 2-year survival rate for biologics in any treatment line was 31.9–‍‍41.7% and decreased to 17.3–‍23.3% after 5 ‍years [14].

IL-17A and IL-17F have recently been identified as drivers of joint and skin inflammation [2, 15, 16]. IL-17A and IL-17F share approximately 50% of their structural homology and have overlap in their pro-‍inflammatory function. Given IL-17A and IL-‍17F can promote pro-inflammatory signaling by acting as either homodimers (IL-17A/IL-17A or IL-17F/IL-17F) or heterodimers (IL-17A/IL-17F) [17], inhibiting both isoforms may provide greater disease control. Indeed, in vitro dual neutralization of both these interleukins is more potent than neutralization of IL-‍17A or IL-‍17F alone [16, 18]. Bimekizumab is a humanized, monoclonal IgG1 antibody that selectively inhibits IL-‍17F in addition to IL-17A [17, 19]. In the phase 3 BE OPTIMAL and BE COMPLETE (plus BE VITAL open-label extension [OLE]) studies, the efficacy and safety of bimekizumab were demonstrated to 1 year in patients who were bDMARD-naïve, or who experienced inadequate response or intolerance to TNFi (TNFi-IR) [20, 21].

Long-term tolerability, sustained efficacy across impacted disease domains, and maximizing patients’ health-related quality of life are important treatment goals of patients with PsA [3, 22, 23]. Therefore, therapies that adequately address these areas and reduce treatment cycling would provide a significant benefit to patients. Herein, the 2-‍year clinical safety and efficacy of bimekizumab from the BE VITAL OLE study are reported, including the 1-year efficacy of bimekizumab in patients who switched from TNFi (adalimumab) in BE OPTIMAL.

Methods

Study Designs

Complete methodologies were reported to week 16/24 in the primary publications of BE OPTIMAL and BE COMPLETE [24, 25], and further details can be found in the subsequent 1-year publications [20, 21]. All studies assessed subcutaneous bimekizumab 160 mg every 4 weeks (Q4W) in patients with active PsA who met the CASPAR criteria [26].

In brief, BE OPTIMAL (NCT03895203; ClinicalTrials.gov) was a 52-week, phase 3, multicenter, randomized, double‑blind, placebo‑controlled, active reference study of bDMARD-naïve patients with active PsA. The study comprised a 16-week placebo‑controlled, double‑blind period followed by a 36-week active treatment‑blind period. A reference arm (adalimumab 40 mg every 2 weeks [Q2W]) was included to provide a standard-of-care reference for bimekizumab treatment. Patients were randomized 3:2:1 to bimekizumab 160 mg Q4W, placebo or reference (adalimumab 40 mg Q2W). From week 16, placebo‑randomized patients received bimekizumab 160 mg Q4W.

BE COMPLETE (NCT03896581; ClinicalTrials.gov) was a 16-week, phase 3, multicenter, randomized, double-‍blind, placebo-controlled study of patients with active PsA and prior TNFi-IR. Patients were randomized 2:1 to bimekizumab 160 mg Q4W or placebo.

BE VITAL (NCT04009499; ClinicalTrials.gov) is a multicenter, OLE study of patients with active PsA who completed week 52 of BE OPTIMAL or week 16 of BE COMPLETE and met eligibility criteria. The study commenced on 13th August 2019 and is ongoing. All patients that entered the BE VITAL OLE received bimekizumab 160 mg Q4W, regardless of feeder study or prior randomization. Patients randomized to the reference arm in BE OPTIMAL switched to receive bimekizumab 160 mg Q4W at week 52 regardless of treatment response. There was no washout period for patients randomized to adalimumab who switched to bimekizumab. The OLE included up to 140 weeks of treatment (Treatment Period) for all subjects in all countries. Following the completion of, or early withdrawal from the Treatment Period, patients returned for a safety follow-‍up visit 20 weeks after their last dose of bimekizumab. Study designs can be found in Fig. 1.

Fig. 1
figure 1

BE OPTIMAL and BE COMPLETE study designs. The ADA 40 mg Q2W treatment arm served as an active reference. The BE OPTIMAL study was not powered for statistical comparisons of ADA to BKZ or PBO. Completion rates include patients that completed to week 52/104 in BE OPTIMAL and week 52/100 in BE COMPLETE not on randomized treatment (BE OPTIMAL week 52: 9 [1.1%], week 104: 8 [0.9%]; BE COMPLETE week 52: 4 [1.0%], week 100: 2 [0.5%]). 2 patients in BE COMPLETE were classified as ongoing at week 52 as they did not have a visit for week 52 but no formal discontinuation reason was reported. ACR50‍ ≥ 50% improvement in American College of Rheumatology response criteria, ADA ‍adalimumab, bDMARD biologic disease-‍modifying antirheumatic drug, BKZ bimekizumab, OLE open-label extension, PBO placebo, TNFi-IR prior inadequate response or intolerance to tumor necrosis factor inhibitors

Full size image

Study visits occurred Q4W from the Entry Visit (week 0 of the BE VITAL OLE) to week 12, then every 12 weeks thereafter up to week 140. In some countries, this was extended past week 140. Bimekizumab could be self-administered after week 12, as patients were required to attend study visits less frequently. Patients could still attend Q4W to receive bimekizumab if self-‍administration was not possible; study outcomes were not collected at these visits.

Patients

Inclusion and exclusion criteria for BE OPTIMAL and BE COMPLETE have been reported previously [24, 25]. Patients in both studies were recruited during the same period of time at overlapping study sites and countries. In BE OPTIMAL, patients with current or prior exposure to any biologics for the treatment of PsA or psoriasis were excluded. In BE COMPLETE, patients were required to have had a prior inadequate response or intolerance to one or two TNFis for either PsA or psoriasis; patients with current or previous exposure to any other biologics were excluded. Patients with a prior history of uveitis and/or inflammatory bowel disease were not excluded from either study.

To be eligible for enrollment in the BE VITAL OLE, patients must have completed either week 52 of BE OPTIMAL or week 16 of BE COMPLETE, been expected to benefit from the OLE per the investigator’s discretion, not met any withdrawal criteria in either feeder study, and provided separate informed consent for the OLE.

Studies were conducted in accordance with the Declaration of Helsinki and the International Conference on Harmonization guidance for Good Clinical Practice. Ethical approval was obtained from the relevant institutional review boards at participating sites, and all patients provided written informed consent in accordance with local requirements. Local institutional review board information can be found in Supplementary Appendix S1.

Outcomes

The primary objective of the BE VITAL OLE was to assess the long-term safety and tolerability of bimekizumab in adult patients with PsA over a period of up to 140 weeks. The secondary objective was to assess the long-term efficacy of bimekizumab.

Safety and efficacy outcomes are reported as the number of weeks from baseline of BE OPTIMAL and BE COMPLETE. Safety outcomes are presented by treatment group and by study period (week 0–52 or week 52–104), and efficacy outcomes are presented by randomization group (placebo-randomized, bimekizumab-randomized, and reference [adalimumab]-randomized [BE OPTIMAL only]). Results are reported by study; results for patients originally enrolled in BE OPTIMAL or BE COMPLETE are hereafter referred to as ‘BE OPTIMAL’ or ‘BE COMPLETE’, respectively.

Safety outcomes, reported from week 0 to 52 and week 52 to week 104, include treatment-emergent adverse events (TEAEs), serious TEAEs, and study discontinuations due to TEAEs. Other safety results reported include drug-related TEAEs, severe TEAEs, deaths, and other safety topics of interest (major adverse cardiovascular events [MACE], neutropenia, serious infections, fungal infections, hypersensitivity, injection site reactions, suicidal ideation and behavior, malignancies, inflammatory bowel disease [IBD] and liver function test changes/enzyme elevations). A serious adverse event must meet one or more of the following criteria: death, life-threatening event, significant or persistent disability/incapacity, congenital anomaly/birth defect (including that occurring in a fetus), initial inpatient hospitalization or prolongation of hospitalization, or an important medical event that (based on appropriate medical judgement) may jeopardize the patient and may require surgical intervention to prevent one of the other outcomes listed. A severe TEAE relates to the intensity of a TEAE, which are reported as either mild, moderate or severe based on clinical classification. A TEAE can be severe, but not considered a serious TEAE if it does not meet any of the abovementioned criteria. MACE, suicidal ideation and behavior, hepatic events and IBD events were adjudicated by an external committee.

Clinical efficacy outcomes, assessed through week 104/100 (BE OPTIMAL/BE COMPLETE), included improvements from baseline of ≥ 20%, ≥ 50%, and ≥ 70% in the American College of Rheumatology response criteria (ACR20/50/70) [27], improvements from baseline of ≥ 75%, ≥ 90%, and 100% in the Psoriasis Area and Severity Index (PASI75/90/100, in patients with psoriasis affecting ≥ 3% body surface area [BSA]) [28] and ACR50 + PASI100 (in patients with psoriasis affecting ≥ 3% BSA).

Achievement of minimal and very low disease activity (MDA, VLDA; achievement of ≥ 5/7 or 7/7, respectively, of the following criteria: tender joint count [TJC] ≤ 1, swollen joint count [SJC] ≤ 1, either PASI ≤ 1 or BSA ≤ 3%, Patient’s Assessment of Arthritis Pain visual analogue scale [Pain VAS] ≤ 15, Patient’s Global Assessment of PsA [PGA‑PsA-VAS] ≤ 20, Health Assessment Questionnaire‑Disability Index [HAQ-‍DI] ≤ 0.5 [29], and tender entheseal points ≤ 1 [measured with the Leeds Enthesitis Index, LEI]) [30], as well as resolution of TJC, SJC, enthesitis (LEI = 0), dactylitis (Leeds Dactylitis Index [LDI] = 0) [31] and nail psoriasis (modified Nail Psoriasis Severity Index [mNAPSI] = 0) were also assessed to week 104/100. Other assessed efficacy outcomes were PASI ≤ 1 responders, Disease Activity Index for Psoriatic Arthritis (DAPSA) high disease activity (HDA), moderate disease activity (MoDA), low disease activity (LDA) and remission (REM) [32], and Psoriatic Arthritis Disease Activity Score (PASDAS) REM and LDA [33].

Patient-reported outcomes (PROs) assessed through week 104/88 include the HAQ-‍DI change from baseline (CfB) and minimal clinically important difference (MCID; decrease from baseline of ≥ 0.35 in patients with baseline score of ≥ 0.35), the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) total score CfB, the Pain VAS CfB and ≥ 30/50% improvement thresholds, the 12-‍item Psoriatic Arthritis Impact of Disease (PsAID-‍12) questionnaire total score CfB and MCID (decrease from baseline ≥ 3 in patients with PsAID-12 ≥ 3 at baseline), the Functional Assessment of Chronic Illness Therapy-Fatigue (FACIT-Fatigue) CfB and MCID (CfB ≥ 4 in patients with FACIT-Fatigue subscale ≤ 48 at baseline) and the Short-Form 36-‍item Health Survey Physical Component Summary (SF-36 PCS) CfB.

Statistical Analysis

Statistical powering and sample size determination for BE OPTIMAL and BE COMPLETE were reported in the primary publications [24, 25]. There was no formal sample size for the BE VITAL OLE, as it was determined by the number of patients eligible for entry to the study.

Descriptive statistics are used to provide an overview of the safety and efficacy results. Baseline values for efficacy variables were determined from baseline values of the respective feeder studies (week 0), as per the EULAR guidance for reporting clinical trial extension data [34].

The Safety Set consisted of all randomized patients who received at least one dose of bimekizumab. Safety variables were analyzed for all patients in the Safety Set. Adverse events were coded according to the Medical Dictionary for Regulatory Activities (MedDRA version 19.0). Safety outcomes are reported as exposure-‍‍‍adjusted incidence rates (EAIRs) per 100 patient-years (PY) of exposure, with associated 95% confidence intervals (95% CI).

The Randomized Set consisted of all enrolled patients randomized in either feeder study. Efficacy variables were analyzed for all patients in the Randomized Set and reported here by treatment group at initial randomization.

Non-responder imputation (NRI) was used to impute missing data for binary outcomes and multiple imputation was used to impute missing data for continuous outcomes. For categorical data, worst-category imputation was used. Data were imputed using baseline patient numbers from randomization of the feeder studies. Any patients that did not enter the BE VITAL OLE were imputed as non-responders, as per the EULAR guidance for reporting clinical trial extension data [34]. Observed case (OC) data are also reported for binary outcomes. All analyses were done with SAS, version 9.3 or higher.

Results

Patient Disposition and Baseline Characteristics

Patient disposition from baseline is provided in Supplementary Fig. S1. Of patients who were randomized in BE OPTIMAL at week 0, 90.4% (n/N = 254/281) of placebo-randomized patients, 87.9% (n/N = 379/431) of bimekizumab-randomized patients and 86.4% (n/N = 121/140) of reference (adalimumab)-randomized patients entered the BE VITAL OLE. Of patients who were randomized in BE OPTIMAL, 85.1% (n = 239) placebo-randomized, 83.3% (n = 359) bimekizumab-randomized and 80.0% (n = 112) reference (adalimumab)-randomized patients completed week 104.

Of patients who were randomized in BE COMPLETE, 91.0% (121/133) of placebo-‍randomized patients and 95.9% (n/N = 256/267) of bimekizumab-randomized patients entered the BE VITAL OLE. Of patients that were randomized in BE COMPLETE, 80.5% (n = 107) of placebo-randomized patients and 80.5% (n = 215) bimekizumab-randomized patients completed week 100.

Patient demographics and baseline characteristics have been reported previously and were representative of patients with long-standing moderate-to-severe PsA [20, 21, 24, 25]. Treatment arms were generally comparable within each trial (Supplementary Table S1).

Safety

Generally, incidence rates of TEAEs and other safety events remained consistent from week 0–52 to week 52–104. Safety data from baseline to week 52 can be found in Table 1, Supplementary Tables S2–S6 and the 1-year reports for BE OPTIMAL and BE COMPLETE [20, 21]. After week 52 and up to week 104, patients treated with bimekizumab in BE OPTIMAL (n = 773), including patients who had switched from the reference treatment (adalimumab) to bimekizumab at week 52, had a TEAE incidence rate (EAIR/100 PY) of 152.2 (95% CI 139.6, 165.5). From week 52 to 104, patients in BE COMPLETE (n = 361) had a TEAE incidence rate of 87.3 (75.3, 100.6; Table 1). The incidence rates of serious TEAEs were 8.2 (6.2, 10.6) and 4.9 (2.8, 7.9) in BE OPTIMAL and BE COMPLETE, respectively. The incidence rates of study discontinuations due to TEAEs were similar between trials; 3.0 (1.9, 4.6) in BE OPTIMAL and 1.8 (0.7, 3.9) in BE COMPLETE. From week 52 to 104, the incidence rate of deaths for patients treated with bimekizumab was 0.1 (0.0, 0.8) in BE OPTIMAL and 0 in BE COMPLETE. In BE OPTIMAL, 1 death occurred between week 52 and week 104 in a 51-year-old female patient who switched from the reference treatment at week 52. The death was due to an acute myocardial infarction and was deemed unrelated to study treatment by the investigator. The patient had a history of cardiovascular risk factors.

Table 1 Safety for week 0–52 and week 52–104
Full size table

The three most common TEAEs for week 52–104 in BE OPTIMAL were SARS-CoV-2 (COVID-19) infection (18.6 [15.4, 22.1]), upper respiratory tract infection (6.3 [4.6, 8.4]) and oral candidiasis (5.5 [3.9, 7.5]). In BE COMPLETE, the three most common TEAEs were SARS-CoV-2 (COVID-19) infection (7.4 [4.7, 10.9]), upper respiratory tract infection (4.6 [2.6, 7.6]) and nasopharyngitis (4.3 [2.3, ‍7.2]); Table 1.

The incidence rates of fungal infections reported by the investigators from week 52 to week 104 were 10.5 (8.3, 13.2) in BE OPTIMAL and 4.6 (2.6, 7.5) in BE COMPLETE (Table 1). The rates of Candida infections were 6.5 (4.8, ‍8.7) and 3.3 (1.7, 5.9) in BE OPTIMAL and BE COMPLETE, respectively (Supplementary Table S2). Most cases were oral candidiasis (BE OPTIMAL: 5.5 [3.9, 7.5]; BE COMPLETE:‍ 3.3 [1.7, 5.9]), the next most frequent was vulvovaginal candidiasis (BE OPTIMAL: 0.8 [0.3, 1.8]; BE COMPLETE:‍ 0). There was one serious case of oropharyngeal candidiasis in BE OPTIMAL, and no serious Candida infections in BE COMPLETE. There were no systemic Candida infections in either study. Further breakdown of fungal infections can be found in Supplementary Table S2.

In BE OPTIMAL from week 0–104, 89.3% (735) of patients did not report a Candida infection, 6.4% (53) patients reported one Candida infection and 4.3% (35) of patients reported two or more Candida infections. In BE COMPLETE, 92.0% (357) of patients did not report a Candida infection, 5.9% (23) reported one Candida infection and 2.1% (8) reported two or more.

The incidence rates of fungal infections leading to patient discontinuation between week 52 and week 104 were low in both studies (all oral candidiasis; BE OPTIMAL:‍ 0.5 [0.1, 1.4]; BE COMPLETE: 0.6 [0.1, 2.2]). There were no documented cases of patient discontinuation due to recurrent fungal infections.

The incidence rate of malignancies (excluding nonmelanoma skin cancer) between week 52 and week 104 was similar between studies (BE OPTIMAL: 0.4 [0.1, ‍1.2]; BE COMPLETE: 0.6 [0.1, 2.2]); a further breakdown of malignancies can be found in Supplementary Table S3).

Incidence rates from week 52 to week 104 were similar between studies for events adjudicated as MACE (BE OPTIMAL: 0.4 [0.1, ‍1.2]; BE COMPLETE: 0; Supplementary Table S4), neutropenia (BE OPTIMAL: 1.2 [0.6, ‍2.3]; BE COMPLETE: 1.8 [0.7, 3.9]) and injection site reactions (BE OPTIMAL: 1.8 [1.0, 3.1]; BE COMPLETE: 0.3 [0.0, 1.7]). Incidence rates of thromboembolism were low in BKZ-treated patients (Supplementary Table S5). The incidence rate for hypersensitivity was 6.9 (5.1, 9.1) in BE OPTIMAL and 3.6 (1.9, 6.4) in BE COMPLETE; the majority of these were dermatitis and eczema (BE OPTIMAL: 3.6 [2.4, 5.3]; BE COMPLETE: 1.2 [0.3, 3.1]). There were low incidence rates of events adjudicated as suicidal ideation and behavior (BE OPTIMAL: 0.1 [0.0, ‍0.8]; BE COMPLETE: 0; Supplementary Table S6) and definite or probable IBD (BE OPTIMAL:‍ 0.4 [0.1, 1.2]; BE COMPLETE: 0). There were no reported cases of uveitis in either study.

Efficacy

ACR50 response rates with bimekizumab treatment were sustained from week 52 to week 104/100 in both the BE OPTIMAL and BE COMPLETE studies (Fig. 2). In BE OPTIMAL, 51.5% of bimekizumab-randomized patients, 50.5% of placebo-randomized patients, and 50.7% of reference (adalimumab)-randomized patients achieved ACR50 at week 104, respectively. In BE COMPLETE at week 100,  50.6% of bimekizumab-randomized patients and 48.1% of placebo-‍randomized patients achieved ACR50, respectively (Fig. 2). ACR20 and ACR70 responses were also sustained during the OLE in both patient populations (Fig. 2; OC data also shown).

Fig. 2
figure 2

ACR20/50/70 responses to week 104/100 (NRI, OC). Randomized Set. In BE OPTIMAL, patients randomized to ADA at baseline switched to BKZ 160 mg Q4W at week 52. aReference arm; study not powered for statistical comparisons between ADA and BKZ or PBO. ACR20/50/70 ‍ ≥ 20/50/70% improvement from baseline in American College of Rheumatology response criteria, ADA ‍adalimumab, bDMARD biologic disease-modifying antirheumatic drug, BKZ‍ bimekizumab, NRI ‍non-responder imputation, OC observed case, PBO placebo, Q4W ‍every 4 weeks, TNFi-IR prior inadequate response or intolerance to tumor necrosis factor inhibitors

Full size image

The proportion of patients achieving complete skin clearance (PASI100) was sustained from week 52 to week 104/100 in bimekizumab-randomized and placebo-randomized patients with ≥ 3% BSA affected by psoriasis at baseline (Fig. 3). An increase in the proportion of patients achieving PASI100 occurred in reference (adalimumab)-randomized patients following the switch to bimekizumab. In BE OPTIMAL, 59.4% of bimekizumab-randomized patients and 60.0% of placebo-randomized patients achieved PASI100 at week 104. The proportion of reference (adalimumab)-randomized patients that achieved PASI100 increased from 48.5% at week 52 to 60.3% at week 104. In BE COMPLETE, 63.1% of bimekizumab-randomized patients and 65.9% of placebo-randomized patients achieved PASI100 at week 100, respectively (Fig. 3). PASI75 and PASI90 responses were also sustained to week 104/100 in both trials (Fig. 3; OC data also shown).

Fig. 3
figure 3

PASI75/90/100 responses to week 104/100 (NRI, OC). Randomized Set, in patients with ≥ 3% body surface area affected by psoriasis at baseline. In BE OPTIMAL, patients randomized to ADA at baseline switched to BKZ 160 mg Q4W at week 52. aReference arm; study not powered for statistical comparisons between ADA and BKZ or PBO. ADA adalimumab, bDMARD biologic disease-modifying antirheumatic drug, BKZ bimekizumab, NRI‍ non-responder imputation, OC ‍observed case, PASI75/90/100‍ ≥ 75/90/100% improvement from baseline in Psoriasis Area and Severity Index, PBO placebo, Q4W every 4 weeks, TNFi-‍IR prior inadequate response or intolerance to tumor necrosis factor inhibitors

Full size image

The proportion of patients achieving MDA was similarly sustained in all patients from week 52 to week 104/100 (Fig. 4). In BE OPTIMAL, 52.4% of bimekizumab-randomized patients, 49.8% of placebo-randomized patients, and 50.7% of reference (adalimumab)-randomized patients were judged to be in MDA at week 104. In BE COMPLETE, 44.9% of bimekizumab-randomized patients and 46.6% of placebo-randomized patients achieved MDA at week 100 (Fig. 4). VLDA and ACR50 + PASI100 responses were also sustained to week 104/100 in both trials (Fig. 4; OC data also shown).

Fig. 4
figure 4

MDA, VLDA and ACR50 + PASI100 responses to week 104/100 (NRI, OC). Randomized Set. In BE OPTIMAL, patients randomized to ADA at baseline switched to BKZ 160 mg Q4W at week 52. aIn patients with  ≥ 3% body surface area affected by psoriasis at baseline. BE OPTIMAL: 140 PBO/BKZ, 217 BKZ, 68 ADA/BKZ; BE COMPLETE: 88 PBO/BKZ, 176 BKZ; bReference arm; study not powered for statistical comparisons between ADA and BKZ or PBO. ACR50 + PASI100 Achievement of both ≥ 50% improvement in American College of Rheumatology response criteria and 100% improvement in Psoriasis Area and Severity Index; ADA adalimumab, bDMARD‍ biologic disease-‍modifying antirheumatic drug, BKZ bimekizumab, MDA ‍minimal disease activity, NRI ‍non-‍responder imputation, OC observed case, PBO placebo, Q4W every 4 weeks, TNFi-IR prior inadequate response or intolerance to tumor necrosis factor inhibitors, VLDA ‍very low disease activity

Full size image

Also sustained from week 52 to week 104/100 was the proportion of patients achieving resolution of SJC (SJC = 0), an objective measure of inflammation. In BE OPTIMAL, 60.6% of bimekizumab-randomized patients, 58.0% of placebo-randomized patients, and 56.4% of reference (adalimumab)-randomized patients achieved SJC resolution at week 104. In BE COMPLETE, 53.6% of bimekizumab-‍randomized patients and 60.9% of placebo-randomized patients achieved SJC resolution at week 100 (Fig. 5I). Resolution of TJC (TJC = 0), as well as change from baseline in SJC and TJC values were also sustained to week 104/100 in both trials (Fig. 5II; OC data also shown for resolution data).

Fig. 5
figure 5

SJC (I) and TJC (II) resolution and change from baseline to week 104/100. Randomized Set. In BE OPTIMAL, patients randomized to ADA at baseline switched to BKZ 160 mg Q4W at week 52. Error bars represent standard errors. aReference arm; study not powered for statistical comparisons between ADA and BKZ or PBO. ADA adalimumab, bDMARD‍ biologic disease-modifying antirheumatic drug, BKZ‍ bimekizumab, CfB change from baseline, MI ‍multiple imputation, NRI non-responder imputation, OC observed case, PBO placebo, Q4W every 4 weeks, SD standard deviation, SJC swollen joint count, TJC tender joint count, TNFi-‍‍‍‍IR‍ prior inadequate response or intolerance to tumor necrosis factor inhibitors

Full size image

Results were similar for enthesitis (Fig. 6) at week 104, where of patients that had enthesitis at baseline (LEI > 0), 53.8% of bimekizumab-randomized, 67.1% of placebo-randomized, and 55.6% of reference (adalimumab)-randomized patients in BE OPTIMAL achieved resolution (LEI = 0), along with 53.8% of bimekizumab-randomized and 50.0% of placebo-randomized patients in BE COMPLETE. Dactylitis resolution (LDI = 0) and nail psoriasis resolution (mNAPSI = 0) are also reported in Fig. 6.

Fig. 6
figure 6

Enthesitis, dactylitis and nail psoriasis resolution to week 104/100 (NRI, OC). Randomized Set. In BE OPTIMAL, patients randomized to ADA at baseline switched to BKZ 160 mg Q4W at week 52. aIn patients with enthesitis at baseline (LEI > 0; BE OPTIMAL: 70 PBO/BKZ, 143 BKZ, 36 ADA/BKZ; BE COMPLETE: 36 PBO/BKZ, 106 BKZ); bIn patients with dactylitis at baseline (LDI > 0; BE OPTIMAL:‍ 33 PBO/BKZ, 56 BKZ, 11 ADA/BKZ; BE COMPLETE: 14 PBO/BKZ, 34 BKZ); cIn patients with nail psoriasis at baseline (mNAPSI > 0; BE OPTIMAL: 156 PBO/BKZ, 244 BKZ, 75 ADA/BKZ; BE COMPLETE: 83 PBO/BKZ, 159 BKZ); dReference arm; study not powered for statistical comparisons between ADA and BKZ or PBO. ADA adalimumab, bDMARD‍ biologic disease-modifying antirheumatic drug, BKZ‍ bimekizumab, LDI Leeds Dactylitis Index, LEI Leeds Enthesitis Index, mNAPSI modified Nail Psoriasis Severity Index, NRI non-‍responder imputation, OC observed case, PBO‍ placebo, Q4W every 4 weeks, TNFi-‍IR‍ prior inadequate response or intolerance to tumor necrosis factor inhibitors

Full size image

Patient-reported outcomes are also reported at 2 years (Table 2). DAPSA disease states (Supplementary Figure S2) and nail psoriasis CfB (Supplementary Figure S3) data can be found in the supplement.

Table 2 Additional efficacy outcomes (week 104/100) and patient reported outcomes (week 104/88)
Full size table

Discussion

Results from the BE VITAL OLE demonstrated that bimekizumab maintained a tolerable safety profile, whilst sustaining efficacy seen at 1 year up to 2 years in patients with active PsA who were bDMARD-naïve or TNFi-IR. The safety profile was consistent with that observed to 1 year [20, 21]. Efficacy up to 2 years, for patients treated with bimekizumab from baseline or week 16, was sustained from week 52 across a range of PsA domains, including joints, skin and nails and reflected in composite outcomes such as MDA. For patients in BE OPTIMAL who switched from the reference treatment (adalimumab) to bimekizumab at week 52, efficacy was generally sustained through 2 years, with further improvements in skin and nails, following the switch and with no washout between treatments.

The long-term safety profile, observed up to 2 years, was generally consistent between bDMARD-naïve and TNFi-IR patients. Bimekizumab was well tolerated in patients with PsA; no new safety signals were observed with long-term treatment [20, 21]. No additional safety signals were observed following direct switch from reference (adalimumab) treatment to bimekizumab without a washout. Incidence rates of TEAEs did not increase with exposure, and incidence rates for adverse events of interest were low, such as suicidal ideation and behavior, uveitis, MACE, IBD, and hypersensitivity.

The majority of fungal infections were mild or moderate, and none were systemic; the incidence of fungal infections did not increase with long-term therapy. Most fungal infections resolved with standard antifungal therapy. Given that fungal infections were categorized at the discretion of the investigator, some Candida infections may have been reported as a fungal infection “not elsewhere classified”. All fungal infections are reported in Supplementary Table S2. There were low numbers of recurrent Candida infections, low incidence rates of patient discontinuation due to fungal infections, and no instances of patients discontinuing after fungal infection recurrence.

Treatment with bimekizumab demonstrated consistent sustained clinical efficacy to 2 years in bDMARD-naïve and TNFi-IR patients with active PsA. The majority of improvements across domains observed at 1 year were sustained at similar levels for up to 2 years. TNFi-naïve and TNFi-experienced patients alike have been shown to switch to a new biologic over their treatment period [35], however, patients switching from a TNFi to another bDMARD often experience reduced persistence of the second agent [14]. Time to discontinuation has been shown to be shorter in TNFi-‍‍experienced patients, who are more likely to discontinue or switch [35]. The long-‍term efficacy demonstrated in these trials suggests that bimekizumab can provide lasting benefit in the clinical management of patients with PsA, who require lifelong treatment. bDMARD-naïve patients switching from reference (adalimumab) treatment to bimekizumab at week 52 showed sustained efficacy in joint symptoms after the switch; further improvement in skin and nail responses were observed following the switch to bimekizumab and were sustained to week 104.

A strength of this analysis was the high proportion of patients remaining in both studies for 2 years. Another strength was that response data were imputed using NRI from baseline of BE OPTIMAL and BE COMPLETE, and were not modified at any later timepoint. Additionally, there were similar long-term responses between the two study populations. In BE OPTIMAL, the reference treatment to bimekizumab switch demonstrated that a direct switch without washout is possible without an increased risk of adverse events.

A limitation of the BE VITAL OLE study is that it was not designed to compare treatment groups from BE OPTIMAL and BE COMPLETE. Additionally, in BE OPTIMAL, there was not an early visit that assessed efficacy outcomes after the switch from reference treatment to bimekizumab. Therefore, the impact of receiving two biologics without a washout period between them could not be evaluated at an early timepoint. Also, in BE OPTIMAL, caution is needed when evaluating the reference (adalimumab)/bimekizumab treatment arm, as the switch took place regardless of treatment response. This may not be applicable to what is seen in the clinic. In BE COMPLETE, as it was open-label following week 16 (versus BE OPTIMAL, which was open-label following week 52), patients received a greater duration of open-label treatment.

Further long-term efficacy and safety data are being collected and will be reported at 3 years when available. Additionally, further evaluation of subgroups for responses and parameters key for long-term responses will be assessed. The authors also acknowledge the need for real-world data for bimekizumab in patients with a wider range of inclusion criteria.

Conclusions

Bimekizumab was well tolerated, with a safety profile consistent with that previously observed in studies of bimekizumab in PsA during phase 2b and 3 studies. Bimekizumab treatment resulted in a sustained high level of efficacy response over 2 years in both bDMARD-naïve and TNFi-IR patients. Responses were consistent across these populations. Efficacy outcomes in the reference (adalimumab) arm following switch to bimekizumab were sustained to 2 years, with a further increase in efficacy observed in skin and nail outcomes.