Impact of deucravacitinib on arthritis in patients with systemic lupus erythematosus: results from the phase II PAISLEY SLE study

Abstract

Objective

Assess the impact of deucravacitinib, an oral, selective tyrosine kinase 2 inhibitor, on arthritis in patients with systemic lupus erythematosus (SLE) in a post-hoc analysis from the PAISLEY SLE study (NCT03252587).

Methods

PAISLEY SLE was a 48-week, global, double-blind, placebo-controlled, phase II study evaluating deucravacitinib (3 mg two times per day), 6 mg two times per day and 12 mg one time per day in SLE. Patients with ≥6 swollen, tender or active (swollen plus tender) joints and those with any joint involvement at baseline were evaluated from weeks 4 to 48. Arthritis efficacy was assessed by 100% reduction (JC-100) or ≥50% reduction (JC-50) from baseline in joint count using non-responder imputation. ORs were calculated using logistic regression. Proportions of patients demonstrating musculoskeletal involvement, as assessed by British Isles Lupus Assessment Group 2004 (BILAG-2004) and Systemic Lupus Erythematosus Disease Activity Index 2000 (SLEDAI-2K) domain scores, were also evaluated.

Results

More patients achieved JC-100 with deucravacitinib versus placebo whether they had ≥1 or ≥6 tender or active joints at baseline. At week 48, JC-100 active joint responses in patients with ≥6 joints involved were deucravacitinib 3 mg two times per day (n=38/63), 60.3%; 6 mg two times per day (n=26/65), 40.0%; 12 mg one time per day (n=28/62), 45.2%; and placebo (n=20/64), 31.3%. At week 48, ORs for JC-100 for swollen, tender or active joints with deucravacitinib 3 mg two times per day versus placebo were 3.4, 2.0 and 3.3, respectively. Comparable responses and ORs were seen with achievement of JC-50 for swollen, tender or active joints at all deucravacitinib doses. More patients receiving deucravacitinib versus placebo had musculoskeletal score improvements by both BILAG-2004 and SLEDAI-2K at week 48.

Conclusion

Deucravacitinib was associated with improvements in arthritis in patients in the phase II PAISLEY SLE study. Results will be further evaluated in ongoing phase III studies.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Arthritis is one of the most common manifestations in systemic lupus erythematosus (SLE) and is associated with impaired daily functioning and reduced quality of life. Alleviation of arthritis symptoms remains an important therapeutic goal in the management of SLE. Deucravacitinib, a selective tyrosine kinase 2 (TYK2) inhibitor, has demonstrated efficacy and was well tolerated in patients with active SLE in the phase II PAISLEY SLE study.

WHAT THIS STUDY ADDS

• In this post-hoc analysis of the phase II PAISLEY SLE study, deucravacitinib was associated with improvements in arthritis compared with placebo, with a higher proportion of patients achieving complete (JC-100) or ≥50% (JC-50) reductions from baseline in swollen, tender or active (swollen plus tender) joint counts through week 48. Comparable improvements were observed across all evaluated deucravacitinib doses.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• These findings support selective TYK2 inhibition with deucravacitinib as a therapeutic approach with the potential to address both SLE disease activity and inflammatory arthritis and provide a rationale for continued evaluation in ongoing phase III clinical trials.

Introduction

Patients with systemic lupus erythematosus (SLE) develop diverse clinical features.¹ Arthritis is one of the most common manifestations,^1–3 with a presentation and severity that varies across patients and over time.⁴ Assessment of arthritis can be challenging, as inflammation in overlapping musculoskeletal structures, such as tendons and muscles, can occur separately or together in patients with SLE. Arthritis in SLE has a significant impact on daily functioning and quality of life.⁵

Deucravacitinib is a first-in-class, oral, selective tyrosine kinase 2 (TYK2) inhibitor approved in multiple countries for the treatment of adults with moderate-to-severe plaque psoriasis and active psoriatic arthritis.^6–11 Inhibition of TYK2 by deucravacitinib prevents signalling into cells by several cytokines that are implicated in the pathophysiology of SLE, including type I and III interferons, interleukin (IL)−23 and IL-12.^12–14 Deucravacitinib selectively binds the TYK2 regulatory domain, locking the enzyme to prevent activation by cytokines. The mechanism of action of deucravacitinib distinguishes it from Janus kinase (JAK) 1, 2, 3 inhibitors that bind to the ATP–binding site on the active domain, which is conserved across JAK family members.^{12 14}

PAISLEY SLE was a 48-week, global, randomised, double-blind, placebo-controlled, phase II study (NCT03252587; https://clinicaltrials.gov/study/NCT03252587) designed to evaluate the efficacy, safety and tolerability of deucravacitinib in adults with active SLE.¹¹ The primary endpoint of SLE Responder Index-4 response at week 32 and all key secondary endpoints were met for the deucravacitinib 3 mg two times per day. Secondary and exploratory endpoints that included reductions in swollen, tender or active joint counts were also met.¹¹ Deucravacitinib was well tolerated and had a similar safety profile to that reported in patients with psoriasis and psoriatic arthritis, with no new safety signals detected in patients with SLE.^{11 15–18}

The objective of this post-hoc analysis was to expand the examination of swollen, tender and active joint findings in patients with arthritis in the PAISLEY SLE study and evaluate changes in arthritis defined as clinically significant by British Isles Lupus Assessment Group 2004 (BILAG-2004) and Systemic Lupus Erythematosus Disease Activity Index 2000 (SLEDAI-2K).

Patients and methods

PAISLEY SLE study design

Details of the PAISLEY SLE study design, key eligibility criteria, study endpoints and statistical analyses have been previously reported.¹¹ Patients who met Systemic Lupus International Collaborating Clinics classification criteria for SLE and were positive for anti-nuclear antibodies, anti–double-stranded DNA or anti-Smith antibodies and had a total SLEDAI-2K score ≥6 and a clinical SLEDAI-2K score ≥4 with joint involvement and/or rash (N=363) were randomised 1:1:1:1 to receive placebo (n=90) or deucravacitinib 3 mg two times per day (n=91), 6 mg two times per day (n=93), or 12 mg one time per day (n=89).¹¹ Patients were required to taper glucocorticoid use to ≤7.5 mg/day between weeks 8 and 20, as clinically appropriate, with stable dosing between weeks 20 and 32. An additional optional taper period occurred between weeks 32 and 40, with stable dosing required between weeks 40 and 48.¹¹ Clinical assessments occurred at 4-week intervals after baseline.

40-joint count

A 40-joint count was used to assess arthritis in joints relevant to SLE (online supplemental figure 1). These include bilateral wrists, elbows, shoulders, ankles, knees, interphalangeal joints of the thumb, individual proximal interphalangeal joints of the hand, first through fifth metacarpophalangeal joints of the hand and individual metatarsophalangeal joints of the feet. Joint counts were defined in these 40 joints by the presence or absence of swelling (ie, swollen joint count; SJC), presence or absence of tenderness (ie, tender joint count; TJC) and presence or absence of both swelling and tenderness (ie, active joint count; AJC).

Each joint was assessed independently for three separate characteristics: presence or absence of tenderness, presence or absence of swelling and presence or absence of both tenderness and swelling (active joints). Therefore, individual joints may be counted in multiple categories. For example, a joint with both tenderness and swelling would be counted in the tender joints category, the swollen joints category and the active (tender+swollen) joints category.

Joint count outcome measures

In this post-hoc analysis, patients with SLE who entered the phase II PAISLEY SLE study with six or more swollen, tender or active joints and those with any joint involvement (one or more swollen, tender or active joint) were evaluated from weeks 4 to 48 for 100% reduction (complete resolution with no joint activity; joint count=0) or ≥50% reduction in swollen, tender or active joint counts from baseline (table 1). Sustained SJC-50, TJC-50 and AJC-50 responses, defined as achievement of a ≥50% reduction in SJC, TJC or AJC from baseline for five consecutive visits between weeks 32 and 48, were also calculated in patients with six or more swollen, tender or active joints at baseline.

Table 1. Posthoc analysis patient subpopulations: joint counts at baseline.

n (%)	Patients with≥1 joint involved at baseline			Patients with≥6 joints involved at baseline
	Active	Swollen	Tender	Active	Swollen	Tender
Deucravacitinib 3 mg two times per day (n=91)	90 (98.9)	90 (98.9)	90 (98.9)	63 (69.2)	68 (74.7)	80 (87.9)
Deucravacitinib 6 mg two times per day (n=93)	93 (100)	93 (100)	93 (100)	65 (69.9)	69 (74.2)	84 (90.3)
Deucravacitinib 12 mg one time per day (n=89)	88 (98.9)	88 (98.9)	88 (98.9)	62 (69.7)	66 (74.2)	75 (84.3)
Placebo (n=90)	87 (96.7)	87 (96.7)	88 (97.8)	64 (71.1)	66 (73.3)	76 (84.4)

Patients could be counted in multiple columns. Overall PAISLEY SLE study population demographics and baseline disease activity are published elsewhere.

SLE, systemic lupus erythematosus.

BILAG-2004 index

The BILAG-2004 index scores disease activity in individual organ systems using grades ranging from A (severe disease) to B (moderate), C (mild), D (no current disease activity from previously affected) or E (never affected). BILAG-2004 scores were recorded at each analysis visit. The BILAG-2004 musculoskeletal domain score includes mild and severe myositis and mild, moderate and severe arthritis categories.¹⁹ Reclassification of the BILAG-2004 index data, in which the musculoskeletal domain was modified to remove myositis categories in order to evaluate the grade based solely on arthritis manifestations, was also performed.

SLEDAI-2K

The SLEDAI-2K score is based on weighted scores for each of 24 clinical descriptors in the past 30 days. Myositis and arthritis domains can be combined into a SLEDAI musculoskeletal domain.²⁰

Statistical analysis

The 40-joint count data (binary responses) were analysed with non-responder imputation, which considers (1) any missing data, (2) patients who used prespecified prohibited medication during the study and (3) patients who had early discontinuation or were lost to follow-up as a non-response, as previously described.¹¹ Patients who did not complete the mandatory glucocorticoid dose taper to ≤7.5 mg/day by week 20 of the PAISLEY SLE study or who required glucocorticoid rescue therapy after week 8 outside protocol-defined allowances were also considered non-responders. The two-sided 95% CI for responses was calculated based on the asymptotic method (including continuity correction). ORs (95% CI) were calculated using logistic regression, with treatment group and redefined stratification factor as fixed factors. BILAG-2004 grades and SLEDAI-2K score improvements (categorical data) were analysed in the randomised population. Data were classified as not reported if patients dropped out of the study or had incomplete data. No formal hypothesis testing was undertaken for these exploratory post-hoc analyses.

Results

PAISLEY SLE study population demographics and baseline disease characteristics were previously reported.¹¹ Table 1 summarises the patient subpopulations of those with one or more or six or more swollen, tender or active joints involved at baseline. Overall, 6.9% of patients had a medical history of arthritis at baseline (deucravacitinib 3 mg two times per day, 6.6%; 6 mg two times per day, 5.4%; 12 mg one time per day, 10.1%; placebo, 5.6%); 2.8% had polyarthritis (5.5%, 3.2%, 0% and 2.2%, respectively) and 0.3% had osteoarthritis (one patient in the deucravacitinib 6 mg group).

Patients with one or more swollen, tender or active joint involved at baseline

To evaluate the impact of deucravacitinib on all patients with arthritis at baseline, post-hoc analyses were completed to assess 100% reduction in AJC (AJC-100), SJC (SJC-100) or TJC (TJC-100). A greater proportion of patients treated with deucravacitinib had AJC-100 responses versus those receiving placebo over time, including at week 48 (deucravacitinib 3 mg two times per day, 52 of 90 (57.8%, 95% CI 47.0 to 68.5); 6 mg two times per day, 42 of 93 (45.2%, 95% CI 34.5 to 55.8); 12 mg one time per day, 38 of 88 (43.2%, 95% CI 32.3 to 54.1); placebo, 28 of 87 (32.2%, 95% CI 21.8 to 42.6); figure 1A). Similar trends were observed for SJC-100 (deucravacitinib 3 mg two times per day, 51 of 90 (56.7%, 95% CI 45.9 to 67.5); 6 mg two times per day, 39 of 93 (41.9%, 95% CI 31.4 to 52.5); 12 mg one time per day, 37 of 88 (42.0%, 95% CI 31.2 to 52.9); placebo, 28 of 87 (32.2%, 95% CI 21.8 to 42.6); figure 1B) and TJC-100 (deucravacitinib 3 mg two times per day, 41 of 90 (45.6%, 95% CI 34.7 to 56.4); 6 mg two times per day, 32 of 93 (34.4%, 95% CI 24.2 to 44.6); 12 mg one time per day, 34 of 88 (38.6%, 95% CI 27.9 to 49.4); placebo, 22 of 88 (25.0%, 95% CI 15.4 to 34.6); figure 1C). Separation between the deucravacitinib arms and the placebo arm was observed at the first postbaseline analysis visit (week 4) with numerical differences maintained after the mandatory glucocorticoid taper ended at the week 20 visit through week 48. Those treated with deucravacitinib had greater odds of achieving AJC-100, SJC-100 or TJC-100 at weeks 32 and 48 versus those receiving placebo (figure 1D).

Figure 1. Achievement of 100% reduction in joint counts over time and ORs in patients with one or more active, swollen or tender joint involved at baseline. Assessed by non-responder imputation. All randomised patients assessed; missing data, prohibited medication use and early discontinuation were analysed as a non-response. A response was defined as a 100% decrease from baseline in active (swollen plus tender), swollen or tender joint count. No p values were calculated for this post-hoc analysis. For (A–C), vertical lines represent 95% CI. The two-sided 95% CI for responses was based on asymptotic (including continuity correction) method. For (D), horizontal lines represent 95% CI. Striped boxes indicate week 32 data, and solid boxes indicate week 48 data. ORs were calculated using logistic regression, with treatment group and redefined stratification factor as fixed factors. Details on stratification factors have been previously published.¹¹ AJC-100, a 100% reduction from baseline in active (swollen plus tender) joint count; BID, two times per day; BL, baseline; GC, glucocorticoid; JC-100, a 100% reduction in joint count; QD, one time per day; SJC-100, a 100% reduction from baseline in swollen joint count; TJC-100, a 100% reduction from baseline in tender joint count.

Similar findings were observed at week 48 for achievement of AJC-50 (deucravacitinib 3 mg two times per day, 58 of 90 (64.4%, 95% CI 54.0 to 74.9); 6 mg two times per day, 51 of 93 (54.8%, 95% CI 44.2 to 65.5); 12 mg one time per day, 46 of 88 (52.3%, 95% CI 41.3 to 63.3); placebo, 39 of 87 (44.8%, 95% CI 33.8 to 55.9)), SJC-50 (deucravacitinib 3 mg two times per day, 58 of 90 (64.4%, 95% CI 54.0 to 74.9); 6 mg two times per day, 51 of 93 (54.8%, 95% CI 44.2 to 65.5); 12 mg one time per day, 45 of 88 (51.1%, 95% CI 40.1 to 62.1); placebo, 38 of 87 (43.7%, 95% CI 32.7 to 54.7)) and TJC-50 (deucravacitinib 3 mg two times per day, 57 of 90 (63.3%, 95% CI 52.8 to 73.8); 6 mg two times per day, 45 of 93 (48.4%, 95% CI 37.7 to 59.1); 12 mg one time per day, 44 of 88 (50.0%, 95% CI 39.0 to 61.0); placebo, 34 of 88 (38.6%, 95% CI 27.9 to 49.4)).

Patients with six or more swollen, tender or active joints involved at baseline

Among patients with six or more active joints involved at baseline in the PAISLEY SLE population who were initially assessed for joint count endpoints, a greater proportion receiving deucravacitinib achieved AJC-100 responses over time than those receiving placebo, including at week 48 (deucravacitinib 3 mg two times per day, 38 of 63 (60.3%); 6 mg two times per day, 26 of 65 (40.0%); 12 mg one time per day, 28 of 62 (45.2%); placebo, 20 of 64 (31.3%); figure 2A). Similarly, a greater proportion of patients receiving deucravacitinib versus placebo achieved SJC-100 responses (deucravacitinib 3 mg two times per day, 41 of 68 (60.3%); 6 mg two times per day, 25 of 69 (36.2%); 12 mg one time per day, 29 of 66 (43.9%); placebo, 20 of 66 (30.3%); figure 2B) and TJC-100 responses (deucravacitinib 3 mg two times per day, 34 of 80 (42.5%); 6 mg two times per day, 30 of 84 (35.7%); 12 mg one time per day, 29 of 75 (38.7%); placebo, 20 of 76 (26.3%); figure 2C) over time, including at week 48.

Figure 2. Achievement of 100% and ≥50% reduction in joint counts over time in patients with six or more active, swollen or tender joints involved at baseline: (A) AJC-100; (B) SJC-100; (C) TJC-100; (D) AJC-50; (E) SJC-50; and (F) TJC-50. Vertical lines represent 95% CI. Assessed by non-responder imputation. All randomised patients with six or more active, swollen or tender joints involved at baseline were assessed; missing data, prohibited medication use and early discontinuation were analysed as a non-response. The two-sided 95% CI for responses was based on asymptotic (including continuity correction) method. No p values were calculated for this post-hoc analysis. AJC-50, a ≥50% reduction in active (swollen plus tender) joint count from baseline; AJC-100, a 100% reduction in active (swollen plus tender) joint count from baseline; BID, two times per day; BL, baseline; JC-50, ≥50% reduction from baseline in joint count; JC-100, 100% reduction from baseline in joint count; QD, one time per day; SJC-50, a ≥50% reduction from baseline in swollen joint count; SJC-100, a 100% reduction from baseline in swollen joint count; TJC-50, a ≥50% reduction from baseline in tender joint count; TJC-100, a 100% reduction from baseline in tender joint count.

A greater proportion of patients treated with deucravacitinib who had six or more joints involved at baseline had AJC-50 responses over time than those receiving placebo, including at week 48 (deucravacitinib 3 mg two times per day, 43 of 63 (68.3%); 6 mg two times per day, 34 of 65 (52.3%); 12 mg one time per day, 35 of 62 (56.5%); placebo, 29 of 64 (45.3%); figure 2D). A greater proportion of patients treated with deucravacitinib who had six or more swollen joints involved at baseline had SJC-50 responses over time than those receiving placebo (week 48: deucravacitinib 3 mg two times per day, 47 of 68 (69.1%); 6 mg two times per day, 35 of 69 (50.7%); 12 mg one time per day, 36 of 66 (54.5%); placebo, 29 of 66 (43.9%); figure 2E), and more patients treated with deucravacitinib who had six or more tender joints involved at baseline had TJC-50 responses over time than those receiving placebo (week 48: deucravacitinib 3 mg two times per day, 50 of 80 (62.5%); 6 mg two times per day, 42 of 84 (50.0%); 12 mg one time per day, 38 of 75 (50.7%); placebo, 31 of 76 (40.8%); figure 2F).

Separation of achievement of ≥50% reduction and 100% reduction in swollen, tender or active joint counts between the deucravacitinib arms and the placebo arm emerged at the first postbaseline analysis visit (week 4) with numerical differences maintained after the mandatory glucocorticoid taper ended at the week 20 visit through week 48 (figure 2).

Patients treated with deucravacitinib had increased odds of achieving AJC-100, SJC-100 or TJC-100 at weeks 32 and 48 versus those receiving placebo (figure 3). Similar results were observed for achievements of AJC-50, SJC-50 and TJC-50 (figure 3).

Figure 3. ORs for achievement of 100% and ≥50% reduction in joint counts with deucravacitinib versus placebo at weeks 32 and 48 in patients with six or more joints involved at baseline. Horizontal lines represent 95% CI. Striped boxes indicate week 32 data, and solid boxes indicate week 48 data. ORs were calculated using logistic regression, with treatment group and redefined stratification factor as fixed factors. Details on stratification factors have been previously published.¹¹ No p values were calculated for this post-hoc analysis. AJC-50, a ≥50% reduction in active (swollen plus tender) joint count from baseline; AJC-100, a 100% reduction in active (swollen plus tender) joint count from baseline; BID, two times per day; JC-50, a ≥50% reduction in joint count; JC-100, a 100% reduction in joint count; QD, one time per day; SJC-50, a ≥50% reduction in swollen joint count from baseline; SJC-100, a 100% reduction in swollen joint count from baseline; TJC-50, a ≥50% reduction in tender joint count from baseline; TJC-100, a 100% reduction in tender joint count from baseline.

In a subset of patients with six or more swollen, tender or active joints involved at baseline who had available data, sustained AJC-50, SJC-50 or TJC-50 responses, inclusive of weeks 32 to 48, were assessed. Among patients treated with deucravacitinib, 3 mg two times per day, 34 of 40 (85.0%) had a sustained AJC-50 response between weeks 32 and 48 compared with 20 of 29 (69.0%) receiving placebo. Similar trends in the deucravacitinib 3 mg two times per day group were observed for sustained SJC-50 (3 mg two times per day, 38 of 44 (86.4%); placebo, 20 of 29 (69.0%)) and TJC-50 (3 mg two times per day, 39 of 46 (84.8%); placebo, 23 of 31 (74.2%)) responses between weeks 32 and 48. Achievement of sustained AJC-50, SJC-50 or TJC-50 with deucravacitinib 12 mg one time per day (AJC, 30 of 35 (85.7%); SJC, 31 of 36 (86.1%); TJC, 33 of 39 (84.6%)) was similar to those observed with deucravacitinib 3 mg two times per day and higher than that observed with placebo. With the deucravacitinib 6 mg two times per day dose, achievements of sustained AJC-50 (30 of 42 (71.4%)), SJC-50 (30 of 44 (68.2%)) and TJC-50 (34 of 49 (69.4%)) were similar to those observed with placebo.

BILAG-2004 grade shifts

The proportion of patients with severe or moderate disease (ie, A and B grades) in the musculoskeletal domain decreased over time (figure 4A). At baseline, 44.0% of patients in the 3 mg two times per day group had severe disease (A grade), 52.7% had moderate disease (B grade), 2.2% had mild disease (C grade) and 1.1% had no disease activity (D/E grades); by week 48, 1.1% of patients had severe disease, 2.2% had moderate disease, 24.2% had mild disease and 51.6% had no disease activity. Similar results were seen with deucravacitinib 6 mg two times per day at baseline versus week 48 (A, 37.6% vs 1.1%; B, 58.1% vs 10.8%; C, 4.3% vs 29.0%; D/E, 0% vs 43.0%) and 12 mg one time per day at baseline versus week 48 (A, 41.6% vs 0%; B, 56.2% vs 7.9%; C, 2.2% vs 20.2%; D/E, 0% vs 43.8%). At baseline, among patients in the placebo group, 40.0% had severe disease, 54.4% had moderate disease, 3.3% had mild disease and 2.2% had no disease activity. The placebo group had higher proportions of patients with severe and moderate grades combined at week 48 (A, 4.4%; B, 10.0%; C, 26.7%; D/E, 31.1%) compared with the deucravacitinib groups (figure 4A).

Figure 4. Musculoskeletal changes over time. (A) BILAG-2004 musculoskeletal domain grade shifts. All randomised patients were assessed at baseline, week 32 and week 48 (deucravacitinib 3 mg two times per day, n=91; 6 mg two times per day, n=93; 12 mg one time per day, n=89; placebo, n=90). The BILAG-2004 index scored disease activity by grade, defining A as severe disease, B as moderate disease activity, C as mild disease activity, D as inactive but previously affected and E as inactive with no previous involvement. (B) SLEDAI-2K musculoskeletal domain improvements. Overall, 86 patients received placebo; 90, 93 and 88 patients received deucravacitinib 3 mg two times per day, 6 mg two times per day and 12 mg one time per day, respectively. The musculoskeletal domain of the SLEDAI-2K included arthritis and myositis. Improvement in the SLEDAI-2K musculoskeletal domain score was defined as a postbaseline visit score lower than the SLEDAI-2K musculoskeletal domain score at baseline. The percentage was based on the total number of patients with a SLEDAI-2K musculoskeletal domain score of >0 at baseline. BILAG-2004, British Isles Lupus Assessment Group 2004; SLEDAI-2K, Systemic Lupus Erythematosus Disease Activity Index 2000.

Seven patients had myositis (placebo, n=1; deucravacitinib 3 mg two times per day, n=5; deucravacitinib 6 mg two times per day, n=1); all patients with myositis also had arthritis. When myositis was excluded from the standard BILAG-2004 musculoskeletal domain grade determination, one patient in the 3 mg two times per day group shifted from mild to no disease activity at week 32 and another patient in this group shifted from mild to no disease activity at week 48 (online supplemental figure 2).

SLEDAI-2K score improvements

At baseline, 98.9% of the deucravacitinib 3 mg two times per day group, 100% of the deucravacitinib 6 mg two times per day group, 98.9% of the deucravacitinib 12 mg one time per day group and 95.6% of the placebo group had activity in the musculoskeletal domain of the SLEDAI-2K. After week 4, more patients treated with deucravacitinib had improvement in the SLEDAI-2K musculoskeletal domain score than those receiving placebo (figure 4B). At week 48, 71.1% of the deucravacitinib 3 mg two times per day arm, 65.6% of the 6 mg two times per day arm and 61.4% of the 12 mg one time per day arm had SLEDAI-2K musculoskeletal score improvements compared with 47.7% of the placebo group.

Discussion

Arthritis is a common and clinically relevant manifestation of SLE, often presenting early in the disease course and contributing to patient burden.^{5 21} Cytokines that signal through TYK2, including type I and III interferons, IL-12 and IL-23, play an important role in SLE immunopathogenesis and have been linked to heightened disease activity.^{12 13} Deucravacitinib, an oral, selective TYK2 inhibitor evaluated in the phase II PAISLEY SLE trial, demonstrated improvements in multiple organ systems, including the musculoskeletal system.¹¹ In this post-hoc analysis, effects of deucravacitinib on arthritis were evaluated using complementary clinical measures, including swollen, tender and active joint counts and musculoskeletal domain scores from the BILAG-2004 and SLEDAI-2K indices.

The 40-joint count used in PAISLEY SLE included those in the 28-joint count (commonly used to assess arthritis in rheumatoid arthritis and lupus) plus bilateral metatarsophalangeal and ankle joints (online supplemental figure 1), as inflammatory foot joint manifestations have been reported in patients with SLE.^{22 23} In patients who entered PAISLEY SLE with six or more swollen, tender or active joints and the larger population with any joint involvement (one or more swollen, tender or active joint), a greater proportion of those treated with any of the three doses of deucravacitinib met 100% and ≥50% improvement thresholds than those receiving placebo. The magnitude of difference from placebo was comparable for all joint count assessments, and deucravacitinib was associated with greater odds of achieving reductions in AJC, SJC and TJC at weeks 32 and 48, particularly with deucravacitinib 3 mg two times per day, which was determined to be the optimal dose for multiple SLE disease domains.¹¹ Differences in sustained responses in consecutive visits between weeks 32 and 48 were also observed with deucravacitinib in patients with six or more swollen, tender or active joints involved at baseline, and AJC-50 responses were comparable for all subgroups evaluated. While the presence of concomitant osteoarthritis may confound the TJCs independently of inflammatory disease activity, the overall incidence of baseline osteoarthritis in this study was low and is unlikely to have impacted the interpretation of treatment outcomes.

Swollen joints represent evidence of synovitis and active inflammatory activity, while TJCs indicate the level of pain.²⁴ Clinically detectable inflammation may progress to structural damage of joints.²⁴ Therefore, in the context of SLE, swollen but non-tender joints are regarded as evidence of active inflammatory synovitis, particularly in the absence of established deforming arthropathy. The reduction of both swollen and TJCs with deucravacitinib is consistent with its mediating broad cytokine inhibition of the TYK2 pathway that mediates pain and inflammation in addition to SLE pathogenesis.^{12 14}

At week 48, more patients treated with deucravacitinib versus placebo had BILAG-2004 grades that indicated no active disease in the musculoskeletal domain, which suggests that severe and moderate disease activity was reduced with deucravacitinib. This finding was consistent whether the original BILAG-2004 musculoskeletal domain, including myositis, was evaluated (figure 4A) or whether the arthritis activity was assessed separately (online supplemental figure 2). By week 48, a greater proportion of patients treated with deucravacitinib versus placebo had improvements in the SLEDAI-2K musculoskeletal domain.

Therapies currently approved for the treatment of lupus, including anifrolumab and belimumab, and therapies recently under investigation, including litifilimab and upadacitinib, have also been shown to improve arthritis in patients with SLE.^25–31 The current findings from PAISLEY SLE also suggest that targeting TYK2 has the potential to improve arthritis manifestations in patients with lupus, supporting the continued evaluation of deucravacitinib in this disease. To date, no clinical trial in SLE has reported complete resolution of arthritis in all patients without adverse events. This highlights the heterogeneous nature of lupus and suggests that different immune targets may be needed for different patients or for the same patient over time.

A limitation of this post-hoc analysis is the small sample size for each subgroup. Inclusion of the feet and ankles in the 40-joint count has the potential to confound mechanical pain with inflammatory pain in these joints and restricts comparisons with trials in which only 28 joints were evaluated. Myositis was included in the SLEDAI-2K and BILAG-2004 index musculoskeletal domain scores; however, a reclassification of the BILAG-2004 index was completed to help address this limitation. Additionally, baseline radiographic assessments for erosive disease were not performed in the trial. Therefore, the potential influence of pre-existing structural joint damage on the arthritis treatment response of deucravacitinib could not be evaluated.

In this post-hoc analysis, deucravacitinib was associated with improvements in arthritis in patients with SLE through week 48, including those who entered with any arthritis or those who entered with significant arthritis, defined by six or more joints involved. Data from several different cut-offs for joint count improvement or sustained improvement were consistent with data using BILAG or SLEDAI thresholds for improvement. Further exploration of effects of deucravacitinib on the spectrum of inflammatory arthritis will be possible in the long-term extension of the PAISLEY SLE study (NCT03920267) and in the two ongoing global, phase III studies of deucravacitinib in SLE (POETYK SLE-1 (NCT05617677) and POETYK SLE-2 (NCT05620407)) as well as additional ongoing phase III studies in Sjögren’s disease (POETYK SjS-1 (NCT05946941)) and psoriatic arthritis (POETYK PsA-1 (NCT04908202) and POETYK PsA-2 (NCT04908189)).

Data availability statement

Data are available upon reasonable request.