Efficacy and safety of risankizumab for active psoriatic arthritis: 24-week results from the randomised, double-blind, phase 3 KEEPsAKE 2 trial.

OBJECTIVES: Risankizumab is an interleukin-23 inhibitor under study for the treatment of patients with psoriatic arthritis (PsA). The phase 3 KEEPsAKE 2 trial investigated the efficacy and safety of risankizumab versus placebo in patients with active PsA who had previous inadequate response or intolerance to ≤2 biological therapies (Bio-IR) and/or ≥1 conventional synthetic disease-modifying antirheumatic drug (csDMARD-IR). Results through week 24 are reported here.
METHODS: Adults with PsA who were Bio-IR and/or csDMARD-IR were randomised to receive subcutaneously administered risankizumab 150 mg or placebo at weeks 0, 4 and 16 during a 24-week, double-blind treatment period. The primary endpoint was the proportion of patients who achieved ≥20% improvement in American College of Rheumatology score (ACR20) at week 24. Secondary endpoints assessed key domains of PsA and patient-reported outcomes.
RESULTS: A total of 444 patients (median age 53 years, range 23-84 years) were randomised to risankizumab (n=224) or placebo (n=220); 206 patients (46.5%) were Bio-IR. At week 24, a significantly greater proportion of patients receiving risankizumab achieved the primary endpoint of ACR20 (51.3% vs 26.5%, p<0.001) and all secondary endpoints (p<0.05) compared with placebo. Serious adverse events were reported for 4.0% and 5.5% of risankizumab-treated and placebo-treated patients, respectively; serious infections were reported for 0.9% and 2.3%, respectively.
CONCLUSION: Treatment with risankizumab resulted in significant improvements versus placebo in key disease outcomes and was well tolerated in patients with PsA who were Bio-IR and/or csDMARD-IR. TRIAL REGISTRATION NUMBER: NCT03671148. © Author(s) (or their employer(s)) 2022. Re-use permitted under CC BY-NC. No commercial re-use. See rights and permissions. Published by BMJ.

Many patients with psoriatic arthritis (PsA) do not achieve an adequate response or are intolerant to conventional synthetic disease-modifying antirheumatic drugs (csDMARDs) or biological agents, highlighting a need for additional effective treatments.

This study demonstrates the efficacy of the interleukin-23 inhibitor, risankizumab, across multiple domains of PsA, including patient-reported outcomes assessing disease burden in patients who had previous inadequate responses to csDMARDs or biological agents.

Risankizumab was well tolerated based on low rates of serious adverse events (AEs), severe AEs, serious and opportunistic infections, and discontinuation of treatment due to AEs by <1% of patients receiving risankizumab.

Results from the phase 3 KEEPsAKE 2 trial demonstrate that risankizumab is effective and well tolerated to treat active PsA.

Risankizumab may provide an additional treatment option for patients with PsA who have had an inadequate response or are intolerant to currently approved therapies.

Introduction

Psoriatic arthritis (PsA) is a progressive, chronic, inflammatory condition that affects approximately 30% of patients with psoriasis.1 2 Symptoms of PsA involve the synovium, tendons, entheses and bone in axial or peripheral joints, and progression is characterised by joint degeneration, leading to disability and increased risk of mortality.3–5 Comorbid conditions such as cardiovascular disease, metabolic syndrome, obesity, diabetes and mood disorders are common among patients with PsA, contributing to functional impairment and decreased quality of life.3 6 PsA is also associated with considerable individual, societal and economic burdens, including reduced employment and increased healthcare costs compared with the general population.7

The aim of PsA treatment is to reduce symptoms, structural damage and inflammation, while restoring overall function, with a goal of remission (REM) and/or reduced disease activity and increased long-term, health-related quality of life.8 9 Initial recommended treatment for PsA is non-steroidal anti-inflammatory drugs (NSAIDs), which may be combined with local corticosteroid injections. Second-line treatment includes use of conventional synthetic disease-modifying antirheumatic drugs (csDMARDs) such as methotrexate, followed by therapy using antitumour necrosis factor medications, and/or other biological agents.8 9

Although biological agents are effective in treating PsA,10 approximately 25%–40% of patients do not achieve at least 20% improvement in American College of Rheumatology score (ACR20), and clinical REM and minimal disease activity (MDA) are often short-lived.11–19 Lack of efficacy frequently leads to treatment switching or discontinuation, which may negatively affect patients’ clinical outcomes and increase treatment costs,20–24 revealing a need for well-tolerated treatments with sustained efficacy.

Risankizumab is a humanised IgG1 monoclonal antibody that specifically inhibits interleukin (IL)-23 by binding to its p19 subunit.25 26 IL-23 is a key component driving the release of IL-17 from Th17 cells, and overexpression of IL-23 has been reported in affected skin in psoriasis and in the synovial tissue of patients with PsA.26–28 KEEPsAKE 2 is an ongoing clinical trial that is evaluating the efficacy and safety of risankizumab to treat PsA in patients with a history of inadequate response or intolerance to csDMARD and/or biological therapies. The results of the initial 24-week double-blind period of the KEEPsAKE 2 trial are reported here.

Methods

Study design and treatment

This was a phase 3, global, multicentre study assessing the efficacy and safety of risankizumab 150 mg vs placebo to treat PsA in patients with inadequate response or intolerance to biological agents (Bio-IR) and/or inadequate response or intolerance to conventional synthetic disease-modifying antirheumatic drugs (csDMARD-IRs). During a screening period of approximately 35 days, patients were stratified by current csDMARD use (0 vs ≥1), number of prior biological therapies (0 vs ≥1) and extent of psoriasis (≥3% vs <3% body surface area affected by psoriasis), then randomised using an interactive response technology system in a 1:1 ratio to receive double-blind treatment with risankizumab 150 mg or matched placebo for 24 weeks, administered subcutaneously at weeks 0, 4 and 16. Patients then received open-label risankizumab every 12 weeks through week 208. The current report presents results for the 24-week double-blind period only, which was from 7 March 2019 to 22 June 2020. Study modifications for the COVID-19 pandemic included out-of-window study visits, phone calls and/or at-home visits for patients unable to attend on-site visits due to travel restrictions, quarantine or COVID-19 infection. The study drug was not administered to patients with suspected or confirmed COVID-19 infection; study drug administration and study visits could be resumed after patients recovered from infection.

Patient and public involvement

Patients and/or the public were not involved in the design, conduct, reporting or dissemination plans of this research.

Patient eligibility

Eligible patients were adults (aged 18 years or older) with a clinical diagnosis of active PsA defined as ≥5 tender joints and ≥5 swollen joints, meeting the Classification Criteria for Psoriatic Arthritis, with symptoms of ≥6 months before screening, and active plaque psoriasis with ≥1 psoriatic plaque of ≥2 cm in diameter or nail changes consistent with psoriasis at screening. Patients were also required to be Bio-IR and/or csDMARD-IR, as described further.

Prior or concomitant medications

Stable treatment with ≤2 concomitant csDMARDs at study entry was permitted if treatment was started ≥12 weeks before baseline at protocol-approved doses. In addition, patients could remain taking stable doses of concomitant NSAIDs, oral corticosteroids (equivalent to prednisone ≤10 mg/day) and other analgesics if they were started ≥1 week before baseline. Patients with a demonstrated lack of efficacy after ≥12 weeks or those who experienced intolerance or had a contraindication to methotrexate, sulfasalazine, leflunomide, apremilast, bucillamine, iguratimod or ciclosporin A were defined as csDMARD-IR.

Patients previously treated with biologic agents, except for IL-23, IL-12/23 or IL-17 antagonists, were also eligible for enrolment. The discontinuation of biological agents was required for prespecified durations before the first study treatment (≥4 weeks for etanercept; ≥8 weeks for adalimumab, infliximab, certolizumab, golimumab and abatacept; ≥1 year (or ≥6 months with normalisation of B cells) for rituximab; or ≥5 times the mean terminal elimination half-life for any other permitted biological agent). Patients with a demonstrated lack of efficacy after ≥12 weeks of treatment, or intolerance to one or two eligible biological agents, were defined as Bio-IR.

Assessments

Efficacy

The primary endpoint was the proportion of patients who achieved ACR20 at week 24. Ranked secondary endpoints assessed at week 24, except where noted, were change from baseline in Health Assessment Questionnaire–Disability Index (HAQ-DI), proportion of patients who achieved ≥90% reduction in Psoriasis Area and Severity Index (PASI 90), proportion of patients who achieved ACR20 at week 16, proportion of patients who achieved MDA, change from baseline in 36-Item Short Form Health Survey Physical Component Summary (SF-36 PCS) score and change from baseline in Functional Assessment of Chronic Illness Therapy–Fatigue Questionnaire (FACIT-Fatigue) score.

Additional non-ranked secondary endpoints included the proportion of patients who achieved ACR50, ACR70, resolution of enthesitis (Leeds Enthesitis Index=0) and resolution of dactylitis (Leeds Dactylitis Index=0) at week 24. Post hoc analyses included the proportions of patients who achieved very low disease activity (VLDA), Disease Activity in Psoriatic Arthritis (DAPSA) REM (defined as DAPSA score ≤4), low disease activity (LDA) +REM (defined as DAPSA score ≤14), ≥50% and ≥85% reductions in DAPSA, HAQ-DI score ≤0.5, ≥10% and ≥30%, and ≥50% reductions in pain (as measured on a visual analogue scale (VAS)), and minimally clinically important difference (MCID) for PtGA (defined as a reduction of 10 mm or more from baseline as measured on a VAS).

Safety

Safety assessments were based on monitoring of treatment-emergent adverse events (TEAEs), which were defined as adverse events (AEs) with onset after the first dose of study drug and were summarised based on the Medical Dictionary for Regulatory Activities V.23.1. Findings from physical examinations, vital sign measurements and clinical laboratory tests (haematology and chemistry) were also assessed. Unblinded safety data were reviewed periodically by an external independent data monitoring committee through the week 24 interim analysis.

Statistics

Sample size determination

It was estimated that 210 patients per treatment group would have a 90% power to detect a mean difference of 0.24 for the changes from baseline in HAQ-DI between risankizumab and placebo, assuming a common SD of 0.72. This sample size would also ensure that analyses would have at least a 90% power to detect a 20% treatment difference in ACR20 at week 24, with an assumed placebo response rate of 35%, using a two-sided test at a significance level of 0.05 and accounting for a 10% dropout rate.

Efficacy and safety analyses

Efficacy and safety analyses were conducted based on the full analysis set, defined as all randomised patients who received one or more doses of the study drug. Patient demographic and medical characteristics were summarised using categorical variables or continuous variables as appropriate.

For the efficacy analyses, the Cochran-Mantel-Haenszel test adjusted for the stratification factors was used for categorical variables, and a mixed-effect model repeat-measurement method was used for continuous variables, each with a two-sided α of 0.05. Due to the smaller number of patients with enthesitis and dactylitis at baseline, it was prespecified that data for the analyses of resolution of enthesitis and dactylitis were to be pooled from the companion study, KEEPsAKE 1 (NCT03675308), and KEEPsAKE 2 to increase sample size. Pooled data for these endpoints were analysed under the multiplicity control of KEEPsAKE 1 and are reported separately.29 A multiple testing procedure was used to control the type I error rate by comparing risankizumab versus placebo in a fixed hypothesis testing procedure that began with the primary endpoint, proceeded through the ranked secondary endpoints in sequence, and continued until an endpoint did not achieve statistical significance. For categorical efficacy endpoints, missing data were handled by non-responder imputation incorporating multiple imputation to handle missing data due to COVID-19 (NRI-C). Missing data unrelated to COVID-19 were handled by non-responder imputation, and missing data due to COVID-19 (infection or logistical restrictions) were handled by multiple imputation. In addition, patients were considered non-responders after the initiation of rescue therapy or concomitant medications for PsA that could have meaningfully impacted efficacy assessments. For continuous efficacy endpoints, observations after the initiation of rescue therapy or concomitant medications for PsA that could have meaningfully impacted efficacy assessments were considered as missing and were excluded from the model. Safety results are summarised as the number and proportion of patients for whom TEAEs were reported within each treatment group.

Results

Patients

A total of 444 patients at 99 sites in 23 countries were randomised to receive risankizumab (n=224) or placebo (n=220); of these patients, 215 (96.0%) and 199 (90.5%), respectively, completed the week 24 study visit (figure 1). One patient was randomised but never received the study drug and was excluded from the efficacy analyses; therefore, 443 patients were included in the full analysis set. Reasons for study discontinuation are summarised in figure 1. No patients discontinued from the study because of COVID-19 infection during the double-blind period; however, one patient discontinued because of COVID-19-related logistical restrictions. Less than 2.5% of patients had missing efficacy data for any parameter in either treatment group because of COVID-19 (online supplemental table 1).

Figure 1

Patient disposition. *One patient was randomised but never received study drug and was therefore excluded from the efficacy analyses, resulting in 219 patients included in the PBO group in the full analysis set. PBO, placebo; RZB, risankizumab.

Demographics and baseline disease characteristics were generally balanced between treatment groups (table 1). The median age (range) was 53 (23–84) years and 55.1% were female. A total of 46.5% patients were Bio-IR. Demographics and baseline disease characteristics for the Bio-IR and csDMARD-IR subgroups are presented in online supplemental table 2. Baseline enthesitis and dactylitis were present for slightly greater proportions of patients in the placebo group compared with the risankizumab group.

Table 1

Demographics and baseline characteristics

Characteristics	RZB 150 mgN=224	PBON=219
Female, n (%)	124 (55.4)	120 (54.8)
Age (years), median (range)	53 (23–84)	52 (24 to 83)
Race, n (%)
White	218 (97.3)	210 (95.9)
Black or African–American	2 (0.9)	3 (1.4)
Asian	2 (0.9)	3 (1.4)
Other	2 (0.9)	3 (1.4)
Not Hispanic/Latino, n (%)	182 (81.3)	176 (80.4)
BMI (kg/m2), mean (SD)	31.5 (8.0)	31.2 (6.8)
PsA duration (years), mean (SD)	8.2 (8.2)	8.2 (8.3)
Swollen joint count,* mean (SD)	13.0 (8.7)	13.6 (9.0)
Tender joint count,† mean (SD)	22.8 (14.9)	22.3 (13.8)
Patient’s assessment of pain,‡ mean (SD)	55.0 (23.5)	57.0 (23.1)
PtGA of disease activity,‡ mean (SD)	56.2 (21.8)	56.2 (23.0)
PGA of disease activity,‡ mean (SD)	63.0 (17.0)	60.7 (16.4)
HAQ-DI, mean (SD)	1.10 (0.62)	1.13 (0.63)
hsCRP (mg/L),§ mean (SD)	7.5 (10.9)	8.2 (17.1)
Presence of psoriasis affecting ≥3% BSA, n (%)	123 (54.9)	119 (54.3)
BSA (%),¶ mean (SD)	12.5 (15.4)	11.7 (14.9)
PASI,¶ mean (SD)	7.7 (6.7)	8.4 (9.9)
MDA, n (%)	5 (2.2)	5 (2.3)
Presence of enthesitis,** n (%)	147 (65.6)	158 (72.1)
LEI,†† mean (SD)	3.0 (1.5)	3.0 (1.6)
Presence of dactylitis,‡‡ n (%)	40 (17.9)	57 (26.3)
LDI,§§ mean (SD)	78.9 (98.4)	109.8 (155.3)
SF-36 PCS score, mean (SD)	35.6 (8.8)	35.2 (9.1)
FACIT-Fatigue score, mean (SD)	28.2 (11.5)	27.7 (12.7)
Prior csDMARDs, n (%)
0	12 (5.4)	11 (5.0)
1	88 (39.3)	81 (37.0)
2	60 (26.8)	60 (27.4)
≥3	64 (28.6)	67 (30.6)
Any prior biologic, n (%)	105 (46.9)	101 (46.1)
Prior failed biologics, n (%)
0	137 (61.2)	132 (60.3)
1	72 (32.1)	64 (29.2)
≥2	15 (6.7)	23 (10.5)
Prior TNF antagonist, n (%)	103 (46.0)	100 (45.7)
Concomitant medication at baseline, n (%)
MTX¶¶	110 (49.1)	99 (45.2)
csDMARD other than MTX***	31 (13.8)	30 (13.7)
MTX and another csDMARD	8 (3.6)	10 (4.6)
Oral corticosteroids	28 (12.5)	22 (10.0)
NSAIDs	141 (62.9)	145 (66.2)

*Based on 66 joints.

†Based on 68 joints.

‡Scored as millimetres on a 100 mm horizontal visual analogue scale.

§Reference range: 0–10 mg/dL.

¶Among patients with ≥3% BSA affected by psoriasis (RZB, n=23; PBO, n=119).

**LEI >0.

††Among patients with LEI >0 (RZB, n=147; PBO, n=158).

‡‡LDI >0.

§§Among patients with LDI>0 (RZB, n=40; PBO, n=57).

¶¶As monotherapy or in combination with another csDMARD.

***Sulfasalazine, leflunomide or apremilast, without MTX.

BMI, body mass index; BSA, body surface area; csDMARD, conventional synthetic disease-modifying antirheumatic drug; FACIT-Fatigue, Functional Assessment of Chronic Illness Therapy–Fatigue; HAQ-DI, Health Assessment Questionnaire–Disability Index; hsCRP, high-sensitivity C reactive protein; LDI, Leeds Dactylitis Index; LEI, Leeds Enthesitis Index; MDA, minimal disease activity; MTX, methotrexate; NSAID, non-steroidal anti-inflammatory drug; PASI, Psoriasis Area and Severity Index; PBO, placebo; SF-36 PCS, 36-Item Short Form Health Survey Physical Component Summary; PGA, physician’s global assessment; PsA, psoriatic arthritis; PtGA, patient’s global assessment; RZB, risankizumab; TNF, tumour necrosis factor.

Efficacy assessments

The primary endpoint and all ranked secondary endpoints were met (table 2). For the primary endpoint, 51.3% of patients treated with risankizumab and 26.5% treated with placebo achieved ACR20 at week 24 (p<0.001). Changes from baseline in each ACR component at week 24 are summarised in online supplemental table 3. Higher ACR20 response rates were observed for patients treated with risankizumab versus placebo, regardless of whether patients received concomitant csDMARDs (50.4% vs 33.9%) or risankizumab as monotherapy (53.0% vs 16.0%), and among the csDMARD-IR (56.3% vs 36.6%) and Bio-IR (45.7% vs 14.9%) patient populations (online supplemental table 4). For the full patient population, similar results favouring risankizumab were also observed for ACR50 (26.3% vs 9.3%, nominal p<0.001) and ACR70 (12.0% vs 5.9%, nominal p<0.05) (table 2).

Table 2

Primary and secondary efficacy endpoints

	RZB 150 mgN=224	PBON=219	Difference(95% CI)	P value
Primary endpoint
ACR20 at week 24, n (%)	115 (51.3)	58 (26.5)	24.5 (15.9, 33.0)	<0.001*
Ranked secondary endpoints
Change in HAQ-DI at week 24, mean (95% CI)	−0.22 (−0.28 to –0.15)	−0.05 (−0.12 to 0.02)	−0.16 (−0.26 to 0.07)	<0.001*
PASI 90 at week 24,† n (%)	68 (55.0)	12 (10.2)	44.3 (33.9 to 54.6)	<0.001*
ACR20 at week 16, n (%)	108 (48.3)	55 (25.3)	22.6 (13.9 to 31.2)	<0.001*
MDA at week 24, n (%)	57 (25.6)	25 (11.4)	14.0 (7.0 to 21.0)	<0.001*
Change in SF-36 PCS score at week 24, mean (95% CI)	5.9 (4.9 to 6.9)	2.0 (0.9 to 3.1)	3.9 (2.4 to 5.3)	<0.001*
Change in FACIT-Fatigue score at week 24, mean (95% CI)	4.9 (3.7 to 6.0)	2.6 (1.4 to 3.9)	2.2 (0.6 to 3.9)	<0.01*
Non-ranked secondary endpoints
ACR50 at week 24, n (%)	59 (26.3)	20 (9.3)	16.6 (9.7 to 23.6)	<0.001
ACR70 at week 24, n (%)	27 (12.0)	13 (5.9)	6.0 (0.8 to 11.3)	<0.05
Resolution of enthesitis at week 24,‡ n (%)	63 (42.9)	48 (30.4)	13.8 (3.5 to 24.2)	<0.01
Resolution of dactylitis at week 24,§ n (%)	29 (72.5)	24 (42.1)	38.8 (22.9 to 54.8)	<0.001

All changes are LS mean changes from baseline.

Results for binary endpoints are based on non-responder imputation incorporating multiple imputation if there are missing data due to COVID-19 or non-responder imputation if there are no missing data due to COVID-19. Results for continuous endpoints are based on mixed models for repeated measures.

ACR20/50/70, ≥20/50/70% improvement in American College of Rheumatology score; BSA, body surface area; FACIT-Fatigue, Functional Assessment of Chronic Illness Therapy–Fatigue Questionnaire; HAQ-DI, Health Assessment Questionnaire–Disability Index; LDI, Leeds Dactylitis Index; LEI, Leeds Enthesitis Index; LS, least square; MDA, minimal disease activity; PASI 90, ≥90% reduction in Psoriasis Area and Severity Index; PBO, placebo; RZB, risankizumab; SF-36 PCS, 36-Item Short Form Health Survey Physical Component Summary.

*Statistically significant under overall type I error control.

†Among patients with ≥3% BSA affected by psoriasis at baseline (RZB, n=123; PBO, n=119).

‡Defined as LEI=0 among patients with LEI >0 at baseline (RZB, n=147; PBO, n=158).

§Defined as LDI=0 among patients with LDI>0 at baseline (RZB, n=40; PBO, n=57).

At week 4 (after a single dose), a greater proportion of patients treated with risankizumab than placebo achieved ACR20 (nominal p=0.016), and the improvement was sustained for all subsequent time points (figure 2A), including a significant difference for the secondary endpoint of ACR20 at week 16 (48.3% vs 25.3%, p<0.001; table 2). Similar patterns occurred with ACR50 (figure 2B) and ACR70 (figure 2C) results.

Figure 2

ACR and PASI response rates over time. (A) ACR20, (B) ACR50, (C) ACR70 and (D) PASI 90 response rates for RZB 150 mg and PBO over the 24-week, double-blind treatment period. PASI 90 results are among patients with ≥3% body surface area affected by psoriasis at baseline. *P≤0.05, **P≤0.01, ***P≤0.001. $Statistically significant under overall type I error control. ACR20/50/70, ≥20/50/70% improvement in American College of Rheumatology criteria score; PASI 90, ≥90% reduction in Psoriasis Area and Severity Index; PBO, placebo; RZB, risankizumab.

At week 24, a greater proportion of patients treated with risankizumab versus placebo experienced resolution of enthesitis (42.9% vs 30.4%, nominal p<0.01) and dactylitis (72.5% vs 42.1%, nominal p<0.001; table 2). These results are consistent with the pooled results from KEEPsAKE 1 and 2 previously reported.29 Additionally, a significantly greater proportion of patients treated with risankizumab versus placebo achieved PASI 90 at week 24 (55.0% vs 10.2%, p<0.001); a difference was observed as early as week 4 (figure 2D).

In the analysis of patient-reported outcomes, the change from baseline in HAQ-DI score was significantly greater in the risankizumab group compared with the placebo group (−0.22 vs −0.05, p<0.001; table 2). In a prespecified analysis of patients with HAQ-DI ≥0.35 at baseline, a greater proportion of patients treated with risankizumab achieved a clinically meaningful improvement in HAQ-DI (≥0.35 from baseline30) at week 24 versus placebo (39.9% vs 23.6%, nominal p<0.001). Significantly greater changes from baseline were also observed for patients treated with risankizumab versus placebo for both SF-36 PCS score (5.9 vs 2.0, p<0.001) and FACIT-Fatigue score (4.9 vs 2.6, p<0.01). The proportion of patients achieving MDA was significantly greater for risankizumab versus placebo (25.6% vs 11.4%, p<0.001; table 2). Additional outcomes on VLDA, DAPSA REM and LDA+REM; percentage reductions in DAPSA and pain; HAQ-DI score ≤0.5; and MCID for PtGA are reported in online supplemental table 5.

Safety

TEAEs were reported for 124 (55.4%) and 120 (54.8%) patients in the risankizumab and placebo groups, respectively (table 3). Most events reported in the risankizumab group were mild or moderate. The most frequently reported TEAE was upper respiratory tract infection (risankizumab, n=17 (7.6%); placebo, n=12 (5.5%); table 4); no other event was reported for ≥5% of patients in either treatment group. Frequencies of serious and severe TEAEs were similar between treatment groups, and, except for severe psoriatic arthropathy (risankizumab, n=1 (0.4%); placebo, n=2 (0.9%)), no severe TEAE was reported for more than one patient in either group. TEAEs leading to discontinuation of treatment were more frequent in the placebo group (n=5, 2.3%) than in the risankizumab group (n=2, 0.9%). No deaths occurred during the 24-week double-blind period.

Table 3

Safety summary

Patients, n (%)	RZB 150 mgN=224	PBON=219
TEAE	124 (55.4)	120 (54.8)
COVID-19-related TEAE	1 (0.4)	0
Serious AE	9 (4.0)	12 (5.5)
Severe TEAE	6 (2.7)	7 (3.2)
TEAE leading to discontinuation of study drug	2 (0.9)	5 (2.3)
Death	0	0
Serious infections*	2 (0.9)	5 (2.3)
Active tuberculosis	0	0
Herpes zoster	0	1 (0.5)
Any other opportunistic infections	0	0
Malignancy†	1 (0.4)	1 (0.5)
Anaphylactic reactions	0	0
Injection site reactions	3 (1.3)‡	1 (0.5)
MACE	1 (0.4)	0

*Serious infections reported in the RZB group were abscess and cellulitis (one patient) and gastroenteritis (one patient); in the placebo group, serious infections were erysipelas, gastroenteritis, postoperative abscess, upper respiratory tract infection and urinary tract infection (each reported for one patient).

†Both were non-melanoma skin cancer.

‡All were non-serious and did not result in discontinuation of the study drug.

AE, adverse event; MACE, major adverse cardiovascular event; PBO, placebo; RZB, risankizumab; TEAE, treatment-emergent adverse event.

Table 4

Frequently reported TEAEs

Patients, n (%)	RZB 150 mgN=224	PBON=219
TEAEs reported in ≥2% of patients in either group
Upper respiratory tract infection	17 (7.6)	12 (5.5)
Hypertension	10 (4.5)	6 (2.7)
Nasopharyngitis	9 (4.0)	8 (3.7)
Arthralgia	7 (3.1)	7 (3.2)
Nausea	6 (2.7)	4 (1.8)
Psoriatic arthropathy	6 (2.7)	9 (4.1)
Bronchitis	5 (2.2)	4 (1.8)
Diarrhoea	5 (2.2)	5 (2.3)
Headache	5 (2.2)	8 (3.7)

PBO, placebo; RZB, risankizumab; TEAE, treatment-emergent adverse event.

Frequencies of AEs of safety interest were low and comparable between treatment groups (table 3). However, injection site reactions were more frequently reported in the risankizumab group (n=3, 1.3%) than the placebo group (n=1, 0.5%). None of the injection site reactions occurring in the risankizumab group were serious or resulted in patient discontinuation, and no anaphylactic reactions were reported. One (0.4%) patient in the risankizumab group with a history of hypertension experienced a non-fatal stroke adjudicated as a major adverse cardiac event. Serious infections were reported for two (0.9%) patients in the risankizumab group and for five (2.3%) patients in the placebo group. There were no reports of active tuberculosis or other opportunistic infection in either treatment group, and only one case of herpes zoster was reported for a patient receiving placebo. There was one reported AE of uveitis in a patient treated with risankizumab and no reported AEs of inflammatory bowel disease.

Mean changes in haematology and clinical chemistry values were small, not clinically meaningful, and comparable between the risankizumab and placebo groups. There were no grade 3 transaminase elevations (as judged by Common Terminology Criteria for Adverse Events V.4.03) reported in either treatment group. Shifts in transaminase levels from baseline are reported in online supplemental table 6.

Discussion

In this phase 3 study, treatment with the IL-23 p19 inhibitor, risankizumab, led to significant improvements in key efficacy measures for patients with active PsA who were csDMARD-IR or Bio-IR. A significantly greater proportion of patients treated with risankizumab versus placebo achieved ACR20 at 24 weeks and all secondary endpoints, including assessments of disease activity in joints and skin, and patient-reported outcomes. Overall, the safety profile was consistent with that described in the UltIMMa-1, UltIMMa-2,31 IMMvent,32 IMMhance,33 and IMMerge34 studies of risankizumab in patients with moderate-to-severe plaque psoriasis, and no new safety signals were observed. Of note, serious infections were reported for <1% of patients treated with risankizumab through week 24, and there were no cases of opportunistic infection, including herpes zoster, systemic candidiasis, or active tuberculosis. There was one major adverse cardiac event reported among patients receiving risankizumab (a non-fatal stroke in a patient with a history of hypertension that did not result in discontinuation of risankizumab). Additionally, only two patients (0.9%) in the risankizumab group discontinued the study due to AEs as the primary reason, which is comparable to the rate in the placebo group (three patients; 1.4%). Together, these results show that risankizumab 150 mg is effective and well-tolerated in patients with PsA.

Multiple agents that target IL-23 or its downstream pathway component, IL-17, are approved to treat PsA.10 Data from the KEEPsAKE 129 and KEEPsAKE 2 trials provide further evidence that specifically targeting the p19 subunit of IL-23 is an effective therapeutic strategy to treat PsA. Notably, similar efficacy was observed with or without background csDMARDs or methotrexate for patients treated with risankizumab (ACR20 response rates, 48% for csDMARDs other than methotrexate and 51% for any methotrexate vs 53% for no csDMARD; online supplemental table 4). However, patients receiving placebo had higher ACR20 response rates when treated with concomitant csDMARDs or methotrexate than did those without any csDMARD or methotrexate use (27% for csDMARDs other than methotrexate and 36% for any methotrexate vs 16% for no csDMARD). The KEEPsAKE 2 study results also support the effectiveness of this mechanism of action for patients with a history of inadequate response to other biological therapies (ie, patients who are generally considered to be more treatment refractory as evidenced by higher rates of treatment discontinuation and switching).21 22 35 As a larger proportion of patients who had failed prior biologics achieved ACR20 at week 24 when treated with risankizumab (45.7%) versus placebo (14.9%), risankizumab may provide an additional effective treatment option for these patients. Among patients who were not Bio-IR, 56.3% versus 36.6% in the risankizumab versus placebo groups achieved ACR20 at week 24. Overall, ACR20 response rates among patients receiving risankizumab were similar among patients who were Bio-IR and those who were csDMARD-IR (45.7% and 56.3%, respectively).

Greater improvement across key domains of PsA, including psoriasis, enthesitis (among the 69% of the study population who had LEI>0 at baseline) or dactylitis (among the 22% of the study population who had LDI>0 at baseline) was observed in patients treated with risankizumab versus placebo. Importantly, risankizumab treatment significantly improved physical function, fatigue and health-related quality of life as assessed by HAQ-DI, FACIT-Fatigue and SF-36 scores, respectively. Together, results from this study demonstrate the efficacy of risankizumab across key domains of PsA for not only musculoskeletal manifestations but also patient-reported outcomes.

This study enrolled a relatively large, representative population of patients with PsA and assessed a broad range of meaningful endpoints. However, there were some limitations. First, the study was performed during the COVID-19 pandemic, which introduced health-related and logistical challenges. These were addressed by implementing specific mechanisms to handle missing data resulting from the pandemic. Overall, less than 2.5% of patients had missing efficacy data because of COVID-19, and this did not affect the overall study results. In addition, no safety concerns attributed to COVID-19 were observed. Though the study is currently limited by the relatively brief assessment period of 24 weeks, the open-label portion of this study, which remains ongoing at the time of this report, will provide safety and efficacy data for risankizumab in this patient population over a 4-year period.

In conclusion, results from the 24-week, double-blind portion of this phase 3 clinical trial in patients with active PsA reveal that risankizumab was well tolerated and effective in treating patients who have experienced previous intolerance and/or inadequate response to csDMARDs or prior biological therapies. Overall, treatment with risankizumab demonstrated efficacy in key clinical PsA domains, providing additional evidence that targeting the p19 unit of IL-23 is a rational therapeutic approach to treat PsA.