In total, 203 patients were included in this analysis. Baseline demographics and disease characteristics were generally similar among patients with and without background csDMARD use (Table 1). At baseline, 132 patients (65.0%) were on background csDMARDs, with comparable proportions in the placebo (44/66; 66.7%), deucravacitinib 6 mg QD (45/70; 64.3%), and deucravacitinib 12 mg QD (43/67; 64.2%) arms. (Table 1) Among patients on background csDMARDs at baseline, MTX use was also similar across placebo (39/44; 88.6%), deucravacitinib 6 mg QD (35/45; 77.8%), and 12 mg QD (37/43; 86.0%) arms. Full details on which csDMARDs were used can be found in Table 1. The number of patients on background oral glucocorticoids was small (n = 25 total; n = 21 with baseline csDMARD use and n = 4 with no baseline csDMARD use) and did not allow for meaningful statistical analysis of the efficacy of deucravacitinib in patients with or without background oral glucocorticoids. The proportion of patients with prior TNFi use was also small (n = 32) and was generally similar for patients with or without background csDMARD use. The low numbers of patients with prior TNFi use (placebo, n = 11; deucravacitinib 6 mg QD, n = 12; 12 mg QD, n = 9) did not allow for meaningful statistical analysis of the efficacy of deucravacitinib in this subgroup of patients compared with TNFi-naïve patients.

On most clinical outcomes, patients treated with the two doses of deucravacitinib showed similar improvements at week 16 regardless of background csDMARD use, with improvements at both doses being greater than those observed in the placebo arm.

Similar improvements at week 16 were observed in the two subpopulations with both doses of deucravacitinib vs placebo on the majority of components of ACR response (Fig. 1).

ACR components (a‒g) and PASI total score (h) at week 16 by csDMARD use. Modified baseline observation carried forward was used to impute missing data. Adjusted means and 95% CI from an analysis of covariance model with factors for body weight and TNFi use and the baseline value as covariate. ACR American College of Rheumatology, CI confidence interval, csDMARD conventional synthetic disease-modifying antirheumatic drug, DEUC deucravacitinib, HAQ-DI HAQ-Disability Index, hsCRP high-sensitivity C-reactive protein, PASI Psoriasis Area Severity Index, PBO placebo, QD once daily, TNFi tumour necrosis factor inhibitor, w with, w/o without

Swollen joint counts: The adjusted mean change from baseline in swollen joint counts was similar in patients with and without background csDMARD use in both dose groups (Fig. 1a).

Tender joint counts: In the deucravacitinib 6 mg treatment arm, patients with background csDMARD use had a numerically larger adjusted mean change from baseline in tender joint count (− 11.3 [95% CI − 13.9, − 8.8]) compared with patients without background csDMARD use (− 6.4 [95% CI − 10.7, − 2.0]), with overlapping CIs. However, patients treated with deucravacitinib 12 mg showed similar improvements in tender joint count regardless of background csDMARD use (Fig. 1b).

Physician and patient global assessments: Deucravacitinib 6 mg and 12 mg showed similar improvements in physician and patient global assessments of disease activity regardless of background csDMARD use (Fig. 1c, d).

Patient global assessment of pain: There were similar improvements in patient global assessment of pain for patients treated with both doses of deucravacitinib regardless of background csDMARD use (Fig. 1e).

HAQ-DI: Patients treated with both doses of deucravacitinib showed similar improvements in HAQ-DI regardless of background csDMARD use (Fig. 1f).

hsCRP: Patients with or without background csDMARD use showed similar decreases in hsCRP with deucravacitinib at both doses. Of note, patients in the placebo arm without background csDMARD use showed increases in hsCRP levels, while those with background csDMARD use showed decreases in hsCRP levels (Fig. 1g).

PASI scores: Patients with or without background csDMARD use showed similar improvements in PASI scores with deucravacitinib at both doses. Of note, patients in the placebo arm without background csDMARD use showed worsening in PASI total scores, while those with background csDMARD use showed improvements (Fig. 1h).

ACR 20 responses: Patients treated with deucravacitinib with and without background csDMARD use showed comparable achievement of ACR 20 at week 16 (Fig. 2a).

Composite measures: ACR 20 response (a) and PASDAS (b) at week 16 by csDMARD use. aAnalyses were performed using NRI for patients with missing data. bModified baseline observation carried forward was used to impute missing data. Adjusted means and 95% CI from an analysis of covariance model with factors for body weight and TNFi use and the baseline value as covariate. ACR American College of Rheumatology, CI confidence interval, csDMARD conventional synthetic disease-modifying antirheumatic drug, DEUC deucravacitinib, NRI nonresponder imputation, PASDAS Psoriatic Arthritis Disease Activity Score, PBO placebo, QD once daily, TNFi tumour necrosis factor inhibitor, w with, w/o without

PASDAS scores: Patients with or without background csDMARD use showed similar improvements from baseline in PASDAS scores with deucravacitinib at both doses (Fig. 2b).

The most frequent AEs with deucravacitinib treatment were upper respiratory tract infection, nasopharyngitis, headache, sinusitis, and rash. No clinically relevant differences in AEs were observed in patients treated with deucravacitinib with or without background csDMARD use (Table 2).

There were no clinically relevant changes from baseline over time in laboratory parameters with deucravacitinib treatment at either dose with or without background csDMARD use, including liver transaminases (ALT and AST), neutrophil counts and lymphocyte counts (see Supplementary Fig. S1 in the electronic supplementary material for details).