At week 196 (KS1: 22JAN2024; KS2: 15MAR2024), there were 749/964 (77.7%; continuous risankizumab, 372/483 [77.0%]; placebo/risankizumab, 377/481 [78.4%]) ongoing patients in KS1. In KS2, there were 289/443 (65.2%; continuous risankizumab, 152/224 [67.9%]; placebo/risankizumab, 137/219 [62.6%]) ongoing patients. Baseline characteristics were similar between groups and were reported previously [12, 13]. The most common primary reasons for discontinuing the trial drug were withdrawal of consent in KS1 (75/964 [7.8%]) and lack of efficacy in KS2 (51/443 [11.5%]), respectively.

EfficacyJoint-Related Efficacy Through Week 196KEEPsAKE 1

Maintenance of ACR20/50/70 achievement at week 196 from week 52 was observed for 71.0%, 66.5%, and 64.0% of patients receiving continuous risankizumab and 72.7%, 74.2%, and 71.9% receiving placebo/risankizumab, respectively (Supplementary Materials Table S1).

At week 196, 57.1% of patients receiving continuous risankizumab and 56.5% of patients receiving placebo/risankizumab achieved ACR20 (Fig. 1). ACR50 and ACR70 were achieved by 39.4% of patients receiving continuous risankizumab and 38.1% of patients receiving placebo/risankizumab, and 27.8% of patients receiving continuous risankizumab and 25.5% receiving placebo/risankizumab at week 196, respectively (Fig. 1).

Fig. 1

Achievement over time of ACR20/50/70 through week 196 for KS1 (A, C, E) and KS2 (B, D, F). ACR20/50/70 ≥ 20%/ ≥ 50%/ ≥ 70% improvement in American College of Rheumatology criteria for symptoms of rheumatoid arthritis, CI confidence interval, KS1 KEEPsAKE 1, KS2 KEEPsAKE 2, NRI-C non-responder imputation incorporating multiple imputations to handle missing data due to COVID-19 (weeks 0 through 24), NRI-MI non-responder imputation incorporating multiple imputations to handle missing data due to COVID-19 and geopolitical conflict in Ukraine, Russia, and/or Israel, and all other missing data treated as non-responders, including patients who received rescue therapy or concomitant medications for PsA that could meaningfully impact efficacy assessment (weeks 28 through 196), PBO placebo, PsA psoriatic arthritis, RZB risankizumab

At week 196, resolution of enthesitis was achieved by 60.1% of patients receiving continuous risankizumab and 63.4% receiving placebo/risankizumab, while 72.3% receiving continuous risankizumab and 77.6% receiving placebo/risankizumab achieved resolution of dactylitis. All joint efficacy findings for KS1 were similar to week 100 and numerically greater when reported as observed for all endpoints (Figs. 1 and 2, Supplementary Materials Figs. S1 & 2, Supplementary Materials Tables S1 & 2).

Fig. 2

Achievement over time of resolution of enthesitis and dactylitis through week 196 for KS1 (A, C) and KS2 (B, D). Resolution of enthesitis and dactylitis reported for patients with enthesitis (LEI > 0) or dactylitis (LDI > 0) at baseline. CI confidence interval, KS1 KEEPsAKE 1, KS2 KEEPsAKE 2, LDI Leeds Dactylitis Index, LEI Leeds Enthesitis Index, NRI-C non-responder imputation incorporating multiple imputations to handle missing data due to COVID-19 (weeks 0 through 24), NRI-MI non-responder imputation incorporating multiple imputations to handle missing data due to COVID-19 and geopolitical conflict in Ukraine, Russia, and/or Israel, and all other missing data treated as non-responders, including patients who received rescue therapy or concomitant medications for PsA that could meaningfully impact efficacy assessment (weeks 28 through 196), PBO placebo, PsA psoriatic arthritis, RZB risankizumab

KEEPsAKE 2

Maintenance of ACR20/50/70 achievement from week 52 was observed for 68.7%, 63.9%, and 70.3% of patients receiving continuous risankizumab and 73.0%, 72.9%, and 63.0% receiving placebo/risankizumab, respectively, at week 196 (Supplementary Materials Table S1). These findings were similar to week 100 and numerically greater when reported as observed (Supplementary Materials Table S1 & 2).

At week 196, 54.5% of patients receiving continuous risankizumab and 50.2% of patients receiving placebo/risankizumab achieved ACR20 (Fig. 1). ACR50 and ACR70 were achieved by 35.3% of patients receiving continuous risankizumab and 37.0% receiving placebo/risankizumab, and 23.7% of patients receiving continuous risankizumab and 22.8% receiving placebo/risankizumab, respectively at week 196 (Fig. 1). The findings for ACR20/50/70 were similar to week 100 and numerically greater when reported as observed (Fig. 1 and Supplementary Material Fig. S1).

At week 196, resolution of enthesitis was achieved by 49.0% of patients receiving continuous risankizumab and 50.0% receiving placebo/risankizumab, while 75.0% receiving continuous risankizumab and 59.6% receiving placebo/risankizumab achieved resolution of dactylitis. All joint efficacy findings for KS2 were similar to week 100 and numerically greater when reported as observed for all endpoints (Figs. 1 & 2, Supplementary Materials Figs. S1&2, Supplementary Materials Tables S1 & 2).

Skin and Nail Clearance Through Week 196KEEPsAKE 1

The proportion of patients who had affected BSA ≥ 3% at baseline and achieved PASI 90 at week 196 was 65.2% among patients receiving continuous risankizumab and 62.9% among those receiving placebo/risankizumab at week 196 (Fig. 3). These findings were similar to week 100 and numerically greater when reported as observed (Fig. 3 & Supplementary Material Fig. 3).

Fig. 3

Achievement over time of PASI 90 for KS1 (A) and KS2 (B) PASI 90 reported only for patients with psoriasis-affected BSA ≥ 3 at baseline. BSA body surface area, CI confidence interval, KS1 KEEPsAKE 1, KS2 KEEPsAKE 2, NRI-C non-responder imputation incorporating multiple imputations to handle missing data due to COVID-19 (weeks 0 through 24), NRI-MI non-responder imputation incorporating multiple imputations to handle missing data due to COVID-19 and geopolitical conflict in Ukraine, Russia, and/or Israel, and all other missing data treated as non-responders, including patients who received rescue therapy or concomitant medications for PsA that could meaningfully impact efficacy assessment (weeks 28 through 196), PASI 90 ≥ 90% improvement from baseline in Psoriasis Area and Severity Index, PBO placebo, PsA psoriatic arthritis, RZB risankizumab

Maintenance of PASI 90 at week 196 from week 52 was achieved by 81.1% of patients receiving continuous risankizumab and 81.2% receiving placebo/risankizumab (Supplementary Materials Table S1). These findings were similar to week 100 and numerically greater when reported as observed (Supplementary Material Tables S1 & 2).

Mean changes from baseline in modified Nail Psoriasis Severity Index (mNAPSI) scores, Physician Global Assessment of Fingernails (PGA-F) scores, and percentage of patients achieving PGA-F of “Clear (0)” or “Minimal (1)” with ≥ 2-grade improvement were assessed only in KS1. Patients receiving continuous risankizumab experienced a – 14.99-point mean reduction in mNAPSI score; patients receiving placebo/risankizumab experienced a – 14.73-point reduction (Fig. 4). Mean reduction from baseline for mNAPSI scores was similar to week 100 (continuous risankizumab [– 14.02 points] and placebo/risankizumab [– 13.57 points]; Fig. 4).

Fig. 4

Change from baseline in mNAPSI through week 196 for KS1. mNAPSI reported only for patients with nail psoriasis at baseline. CI confidence interval, MMRM mixed-effect model for repeated measures (weeks 0 through 196), mNAPSI modified Nail Psoriasis Severity Index, KS1 KEEPsAKE 1, PBO placebo, RZB risankizumab

Mean reduction in PGA-F score from baseline was – 1.5 for both continuous risankizumab and placebo/risankizumab and was similar to week 100 (continuous risankizumab [– 1.4] and placebo/risankizumab [– 1.3]; Table 2). The proportion of patients with a PGA-F score of “Clear (0)” or “Minimal (1)” with ≥ 2-grade improvement was 78.8% in patients receiving continuous risankizumab and 73.5% in patients receiving placebo/risankizumab and was similar to week 100 (continuous risankizumab [69.9%] and placebo/risankizumab [67.9%] Table 2). Excluding week 100 for patients receiving placebo/risankizumab, findings were numerically greater for reduction in mNAPSI when reported as observed (Supplemental Materials Table S3). Findings for reduction in PGA-F were similar when reported as observed as well (Supplemental Materials Table S3).

KEEPsAKE 2

At week 196, 66.7% of patients receiving continuous risankizumab and 56.3% of patients receiving placebo/risankizumab achieved PASI 90 (Fig. 3). These findings were similar to week 100 and numerically greater when reported as observed (Fig. 3 & Supplementary Material Fig. 3).

Maintenance of PASI 90 at week 196 from week 52 was achieved by 84.8% of patients receiving continuous risankizumab and 74.6% receiving placebo/risankizumab (Supplementary Materials Table S1). These findings were similar to week 100 and numerically greater when reported as observed (Supplementary Material Tables S1&2).

MDA Achievement Through Week 196KEEPsAKE 1

The proportion of patients achieving minimal disease activity (MDA) was 39.6% receiving continuous risankizumab and 35.2% receiving placebo/risankizumab (Fig. 5). These findings were similar to week 100 and numerically greater when reported as observed (Fig. 5 & Supplementary Materials Fig. S4).

Fig. 5

Achievement over time of MDA through week 196 in KS1 (A) and KS2 (B). CI confidence interval, MDA minimal disease activity, NRI-C non-responder imputation incorporating multiple imputations to handle missing data due to COVID-19 (weeks 0 through 24), NRI-MI non-responder imputation incorporating multiple imputations to handle missing data due to COVID-19 and geopolitical conflict in Ukraine, Russia, and/or Israel, and all other missing data treated as non-responders, including patients who received rescue therapy or concomitant medications for PsA that could meaningfully impact efficacy assessment (weeks 28 through 196), KS1 KEEPsAKE 1, KS2 KEEPsAKE 2, PBO placebo, PsA psoriatic arthritis, RZB risankizumab

Maintenance of MDA at week 196 from week 52 was achieved by 72.8% of patients receiving continuous risankizumab and 72.2% of patients receiving placebo/risankizumab (Supplemental Materials Table S1). These findings were similar to week 100 and numerically greater when reported as observed (Supplementary Materials Tables S1 & 2).

KEEPsAKE 2

The proportion of patients achieving MDA at week 196 was 35.3% of patients receiving continuous risankizumab and 37.4% of patients receiving placebo/risankizumab (Fig. 5). These findings were similar from week 100 and numerically greater when reported as observed (Fig. 5 & Supplementary Materials Fig. S4).

Maintenance of MDA at week 196 from week 52 was achieved by 77.0% of patients receiving continuous risankizumab and 70.3% of patients receiving placebo/risankizumab (Supplemental Materials Table S1). These findings were similar to week 100 and numerically greater when reported as observed (Supplementary Materials Tables S1&2).

Patient-Reported Outcomes Through Week 196KEEPsAKE 1

The mean reductions in Health Assessment Questionnaire–Disability Index (HAQ-DI) score from baseline were – 0.40 for patients receiving continuous risankizumab and – 0.33 for patients receiving placebo/risankizumab at week 196 (Table 1). Clinically meaningful improvements in HAQ-DI score were achieved by 45.7% of patients receiving continuous risankizumab and 42.3% in patients receiving placebo/risankizumab.

Table 1 Efficacy results for KEEPsAKE 1 & 2 at week 196

At week 196, the mean reduction in patient-reported pain scores (VAS) from baseline was – 27.9 points for continuous risankizumab patients and – 24.6 for placebo/risankizumab patients (Table 1). Maintenance of clinically meaningful improvement in pain (≥ 10 mm reduction from baseline) at from week 52 to week 196 was achieved by 72.2% of patients receiving continuous risankizumab and 75.9% of patients receiving placebo/risankizumab (Supplementary Materials Table S1).

Mean improvements from baseline in the 36-item Short Form Physical Component Summary (SF-36 PCS) were 9.0 points for continuous risankizumab and 7.2 points for placebo/risankizumab. Mean Functional Assessment of Chronic Illness Therapy–Fatigue (FACIT-Fatigue) scores improved from baseline by 7.7 points for continuous risankizumab and 5.6 points for placebo/risankizumab (Table 1).

A low Disease Activity in Psoriatic Arthritis Low Disease Activity (DAPSA LDA) score was achieved by 66.3% of patients in continuous risankizumab and 64.3% of patients in placebo/risankizumab (Fig. 6). Maintenance of DAPSA LDA at week 196 from week 52 was achieved by 85.6% of patients in continuous risankizumab and 86.3% of patients in placebo/risankizumab (Supplementary Materials Table S1). All patient-reported outcomes for KS1 were similar to week 100 and numerically greater when reported as observed (Fig. 6, Table 1, Supplementary Materials Tables S1 to S3).

Fig. 6

Achievement over time of LDA as assessed by DAPSA through week 196 in KS1 (A) and KS2 (B). CI confidence interval, DAPSA LDA Disease Activity Index for Psoriatic Arthritis Low Disease Activity (≤ 14), NRI-C non-responder imputation incorporating multiple imputations to handle missing data due to COVID-19 (weeks 0 through 24), NRI-MI non-responder imputation incorporating multiple imputations to handle missing data due to COVID-19 and geopolitical conflict in Ukraine, Russia, and/or Israel, and all other missing data treated as non-responders, including patients who received rescue therapy or concomitant medications for PsA that could meaningfully impact efficacy assessment (weeks 28 through 196), KS1 KEEPsAKE 1, KS2 KEEPsAKE 2, PBO placebo, PsA psoriatic arthritis, RZB risankizumab

KEEPsAKE 2

The mean reductions in HAQ-DI score from baseline were – 0.26 for patients receiving continuous risankizumab and – 0.34 for patients receiving placebo/risankizumab (Table 1). Clinically meaningful improvements in HAQ-DI score were achieved by 36.7% of patients receiving continuous risankizumab and 40.6% receiving placebo/risankizumab (Table 1).

At week 196, the mean reduction in patient-reported pain scores (VAS) from baseline was – 24.5 points for continuous risankizumab and – 23.6 points for placebo/risankizumab (Table 1). Maintenance of clinically meaningful improvement in pain VAS scores at week 196 from week 52 was achieved by 75.2% of patients receiving continuous risankizumab and 67.2% receiving placebo/risankizumab (Supplementary Materials Table S1).

Mean improvements from baseline in SF-36 PCS for patients receiving continuous risankizumab were 6.5 points and 6.2 points for patients receiving placebo/risankizumab (Table 1). Mean FACIT-Fatigue scores improved from baseline by 6.0 points for patients receiving continuous risankizumab and 6.6 points for patients receiving placebo/risankizumab (Table 1).

A low DAPSA LDA score was achieved by 61.9% of patients in continuous risankizumab and 60.7% of patients in placebo/risankizumab (Fig. 6). These findings were numerically greater compared to week 100 and when reported as observed (Fig. 6 & Supplementary Materials Table S3). Maintenance of DAPSA LDA at week 196 from week 52 was achieved by 85.6% of patients in continuous risankizumab and 86.3% of patients in placebo/risankizumab (Supplementary Materials Table S1). All patient-reported outcomes, excluding DAPSA LDA, for KS2 were similar to week 100 and numerically greater when reported as observed (Fig. 6, Table 1, Supplementary Materials Tables S1 to S3).

SafetyKEEPsAKE 1 and 2

Pooled long-term safety data were available for 1365 patients (4466.7 PY of exposure) who received risankizumab 150 mg during either the double-blind or open-label phase of the trials. TEAE and serious TEAE rates were 136.7 E/100 PY and 8.3 E/100 PY, respectively (Table 2). COVID-19 infections (8.6 E/100 PY) and nasopharyngitis (4.6 E/100 PY) were the most common TEAEs (defined as any TEAE with ≥ 4.0 E/100 PY) at week 196. Serious infections occurred at a rate of 1.7 E/100 PY at week 196, numerically lower than rates of serious infections in patients with PsA based on results from US MarketScan claims data (3.98 E/100 PY) [19]. The rate of serious infections was also similar to patients with PsA using IL-17 (2.0 E/100 PY), IL-12/23 (2.0 E/100 PY), and tumor necrosis factor (TNF) inhibitors (2.7 E/100 PY) [20].

Table 2 Safety summaries for KEEPsAKE 1&2 through week 196

Opportunistic infections (0.1 E/100 PY) and herpes zoster (0.3 E/100 PY) were minimal, and there were no reported cases of active tuberculosis at week 196 (Table 2). The rates of opportunistic infections and herpes zoster were numerically lower than what has been reported for patients with PsA in multiple epidemiological studies (2.5 and ~ 1.0–2.0 E/100 PY, respectively) [21,22,23].

A total of 76 (1.7%) TEAEs leading to trial drug discontinuation occurred at week 196. Psoriatic arthropathy (n = 166 [2.7%]) was the most common TEAE leading to discontinuation of the trial drug at week 196. A total of 101 (2.3 E/100 PY) hypersensitivity events occurred at week 196, with only two (allergic reaction and immune thrombocytopenia) considered serious (Table 2). Both events were deemed to have no reasonable possibility of being related to the trial drug.

Seventeen adjudicated major adverse cardiovascular events (MACE; 0.4 E/100 PY) were reported at week 196, including two sudden cardiac deaths, 1 heart failure, 7 non-fatal myocardial infarctions, and 6 non-fatal strokes (Table 2). The rate of MACE was similar to the rate reported in epidemiological studies for patients with PsA (0.3–0.5 E/100 PY) and for patients with PsA using IL-17 (ixekizumab [0.6 E/100PY]; secukinumab [0.4 E/100 PY]) and IL-23 inhibitors (0.4 E/100 PY) [24,25,26,27,28,29].

A total of 24 malignant tumors, excluding non-melanoma skin cancers (NMSCs; 0.5 E/100 PY) and 21 NMSCs (0.5 E/100 PY), were reported at week 196 (Table 2). The rate of malignant tumors was similar to the epidemiological reference rate of patients with PsA (0.5 E/100 PY) [30], while the rate of NMSCs was similar compared to the epidemiological reference rate of patients with PsA (0.6 E/100 PY) [30] and patients with PsA utilizing another IL-23 inhibitor (ixekizumab [0.3 E/100 PY]) [25]. A total of 14 deaths were reported, all of which were deemed unrelated to the trial drug (Table 2).

With additional patient years of exposure, the rates of TEAEs remained stable. The long-term safety data at week 196 were generally consistent with the rates noted in previous publications at weeks 24 and 100.