Ustekinumab in psoriatic arthritis and related phenotypes

Introduction

Psoriatic arthritis (PsA) is an inflammatory arthritis that occurs in approximately 20–30% of patients with psoriasis. Psoriasis is a chronic relapsing skin disease characterized by global epidermal thickening. Psoriasis usually precedes the inflammatory arthritis in 70% of patients, with inflammatory skin and joint disease occurring simultaneously in 15% of patients and the inflammatory arthritis occurring before the dermatosis in the remaining patients. ¹ Eventually, almost all patients with PsA will develop psoriasis. Unlike most other inflammatory rheumatic diseases, there is no sex predilection, as it occurs equally among males and females. The clinical presentation and course of PsA is quite heterogeneous. Moll and Wright described five patterns of PsA based on joint distribution: polyarticular arthritis, oligoarthritis, distal interphalangeal arthritis, spondyloarthritis and the very destructive, arthritis mutilans. ² With disease evolution, patients may exhibit multiple patterns and are not limited to one subset of arthritis. At presentation, oligoarticular disease is the most common subtype but as the disease evolves, the polyarticular variant becomes the most prevalent. Periarticular involvement, particularly enthesitis and dactylitis, frequently occur in PsA and the presence of dactylitis has been associated with increased disease severity, particularly worse radiographic outcomes. Extra-articular features include anterior uveitis and inflammatory bowel disease. There is a growing list of comorbid entities associated with psoriatic disease including obesity, type II diabetes, cardiovascular and cerebrovascular disease and depression. ³

Prior to 2006, the Moll and Wright criteria were primarily used to diagnose PsA. To satisfy these criteria, patients needed to have inflammatory arthritis with psoriasis and negative rheumatoid factor. ² The CASPAR (ClASsification Criteria for Psoriatic ARthritis) criteria are now used, which require the presence of inflammatory arthritis, spondylitis or enthesitis and three of five categories, which include psoriasis or family history of psoriasis, dactylitis, absence of rheumatoid factor and radiographic abnormalities suggestive of inflammatory arthritis. ⁴ Scalp lesions, and intergluteal/perianal psoriasis are proposed risk factors the development of PsA in psoriasis patients. ⁵

Management of PsA

In managing patients with psoriasis, it is important to address all key domains of PsA (i.e. articular disease, enthesitis, dactylitis, psoriasis, nail disease) and to be aware of the comorbidities. For the purpose of this review, we will primarily review the musculoskeletal involvement in PsA. The articular disease is quite heterogeneous and can range from minimal inflammation, pain and disability to a rapidly destructive and deforming arthritis that severely impacts patients’ quality of life. The two most referenced guidelines for the management of PsA are from EULAR (EUropean League Against Rheumatism) ⁶ and GRAPPA (Group for Research and Assessment of Psoriasis and Psoriatic Arthritis). ⁷ Although these consortia share synergies, the primary focus of EULAR is musculoskeletal (MSK) manifestations, whereas the GRAPPA guidelines address the MSK and cutaneous manifestations. In both guidelines, peripheral arthritis is managed using step-up therapy, starting with nonsteroidal anti-inflammatory drugs (NSAIDs), then synthetic disease-modifying antirheumatic drugs (DMARDs). Tumour necrosis factor (TNF)-α inhibitors are presently considered to be the first-line biologic therapy as long as there is no contraindication to anti-TNF agents. If a patient fails to respond to anti-TNF therapies or a contraindication to therapy is present, then interleukin (IL)-12/23 or IL-17A inhibition is suggested. As anti-TNF agents are established treatments for PsA, we will focus this article on the emerging role of IL-12/23 blockade in PsA. We will briefly review the role of IL-12/23 for the clinical management of psoriasis and inflammatory bowel disease, which are two extra-articular manifestations of PsA.

Th-17 signalling pathway and PsA pathogenesis

Genetic, immunologic and environmental factors activating both the innate and acquired immune response appear to have an important role in the pathogenesis of PsA. Recently, the T-helper (Th)-17 signalling pathway and the IL-23/IL-17 axis has become a prominent player in PsA and has considerably increased our understanding of disease pathogenesis.

IL-23 is a heterodimeric cytokine that binds IL-23R and IL-12Rβ1, ⁸ which may be triggered by multiple factors (e.g. biomechanical stress and intestinal dysbiosis). Activation induces IL-23-dependent differentiation and activation of Th-17 cells, IL-17 and IL-22 cytokine production and secretion, culminating in synovium and skin inflammation as well as bone remodeling. ⁹

Th-17 effector signalling (i.e. IL-17, IL-21, and IL-22) also contributes to the autoimmune and autoinflammatory response and PsA pathogenesis. IL-17, which binds to IL-17RA and IL-17RC receptors, ¹⁰ stimulates fibroblasts, endothelial cells, macrophages and epithelial cells to release proinflammatory mediators resulting in destructive tissue damage.^11,12 Of particular relevance to PsA pathology, IL-17 promotes bone erosion through the upregulation of RANKL ¹³ a key regulator of osteoclastogenesis. IL-21 is a potent immunomodulatory cytokine that has a key role in Th-17 cell differentiation, leading to increased IL-17 production and IL-23R expression,^14,15 and affords the ability to create an autocrine amplification loop by which Th-17 cells can augment the rate of differentiation. ¹⁶

Evidence to support genetic susceptibility of PsA arises from family investigations, human leukocyte antigen (HLA) studies, genome-wide association scans and candidate gene studies. Altered IL-23 signalling secondary to genetic aberration may enhance Th-17 cell expansion, with consequent IL-17-mediated upregulation of cytokines, chemokines, and tissue-degrading matrix metalloproteases, culminating in an inflammatory environment. The four genome-wide association studies have identified 13 regions associated with PsA at a genome-wide level of significance (p ⩽ 5 × 10⁻⁸) including HLA-B, HLA-C, IL-12B, IL-23R, IL23A, TNIP1, TRAF3IP2, CSF2, FBXL19, REL, RUNX3, TYK2, NOS2, PTPN22, IFNLR1, IFIH1 and NFKBIA.^17–20 Genes that encode proteins crucial for Th-17 signalling include IL-12B, IL-23R, IL23A, TRAF3IP2 and TYK2 as previously reviewed; ²¹ these findings suggest a strong genetic association of PsA with variants of the IL-23/IL-17 axis.

Ustekinumab in PsA: clinical trials

Ustekinumab, is a fully human monoclonal immunoglobulin (Ig)G1 antibody that binds specifically to the p40 subunit of IL-12 and IL-23 primarily inhibiting downstream Th-17 signalling pathways. ²² It is administered subcutaneously and approved for the treatment of adults with active PsA in the United States of America (USA) and Europe.

Several clinical trials have examined ustekinumab as a treatment for PsA. A phase II, randomized, double-blinded, placebo-controlled crossover trial was conducted to establish its efficacy and safety. ²³ The successful phase II study was followed by two larger phase III studies: PSUMMIT-1, ²⁴ which involved patients with moderate-to-severe disease that failed NSAIDs or synthetic DMARDs; and PSUMMIT-2, ²⁵ which included approximately 60% of patients that had failed anti-TNF agents. Subsequent publications followed that specifically addressed radiographic outcomes, physical function and health-related quality of life and long-term extension studies, including safety.

The aim of the initial phase II study was to assess the efficacy, tolerability and safety of ustekinumab in adult patients diagnosed with active PsA, who had failed DMARDS, NSAIDS, anti-TNF biologic therapy, or a combination of the three. Patients were divided into two groups: (1) patients that received ustekinumab 90 mg at weeks 0, 1, 2 and 3 and placebo at weeks 12 and 16; and (2) patients that received placebo at weeks 0, 1, 2 and 3 and ustekinumab at weeks 12 and 16. Then both groups were followed for 36 weeks. The primary outcome was a 20% improvement in a composite score, ACR20 (American College of Rheumatology criteria for a 20% improvement in disease), at week 12 compared with the baseline score. ACR20 was reached at week 12 in 42% of the active group and 14% of placebo (p = 0.0002). Higher endpoints for ACR scores were also reached including ACR50 (ACR criteria for a 50% improvement in disease) and ACR70 (ACR criteria for a 70% improvement in disease) significantly favouring the ustekinumab arm at week 12 (18% difference from placebo, p = 0.0038; and 11% difference from placebo, p = 0.0055; respectively). Favourable results were also noted for most secondary outcomes including HAQ-DI (Health Assessment Questionnaire-Disability Index, p = 0.0005), PASI 75 (75% improvement in the psoriasis area and severity index; p < 0.0001), DLQI (Dermatology Life Quality Index; p < 0.0001) and enthesitis (p = 0.0163). In contrast, there was no significant difference in the dactylitis score (p = 0.5429). Ustekinumab was generally well tolerated for the first 12 weeks evidenced by a lack of serious adverse events (AEs) or serious infections in the treatment group. At 36 weeks, there were six serious AEs including one patient with malignancy (recurrent basal cell carcinoma).

The success of the phase II trial paved the way for two phase III double-blind, randomized, placebo-controlled clinical trials, PSUMMIT-1²⁴ and PSUMMIT-2. ²⁵ All patients had PsA for at least 6 months and satisfied the CASPAR classification criteria and the study designs were complementary allowing certain prespecified analyses to be combined to increase statistical power.

PSUMMIT-1 participants were all naïve to anti-TNF therapy and had active disease despite 3 or more months of DMARDs or 4 weeks of NSAIDs or both. Active PsA was defined as the presence of at least five or more swollen and five or more tender joints and baseline C-reactive protein (CRP) concentration ⩾3.0 mg/l. Patients were randomized to receive either 45 mg of subcutaneous ustekinumab or 90 mg of subcutaneous ustekinumab at week 0, 4 and every 12 weeks onwards. Early escape occurred at week 16 if patients demonstrated <5% improvement and all placebo patients were switched to 45 mg of subcutaneous ustekinumab at week 24. ²⁴

A total of 205 patients were randomized to receive ustekinumab (45 mg), 204 patients received ustekinumab (90 mg) and 205 received placebo. The primary endpoint was the proportion of patients with ACR20 at week 24. There was a significant difference in the active arms as 42.2% of the 45 mg ustekinumab group (p < 0.001) and 49.5% of the 90 mg ustekinumab group (p < 0.001) improved compared with the placebo (22.8%). There was also a statistically significant improvement in the major secondary endpoints which included HAQ-DI, PASI 75, ACR50 and ACR70 at week 24. Generally, scores in the 90-mg ustekinumab group exceeded that of the 45-mg group; however, the study was not powered to detect a difference between groups. The primary and secondary response rates were essentially maintained or improved to week 52. Approximately half the patients were on concomitant methotrexate in each of the treatment groups. The percent difference appeared lower in those on concomitant methotrexate however there were no tests of significance performed on this subgroup. ²⁴

Periarticular features of the patients with baseline enthesitis and dactylitis significantly improved with treatment with ustekinumab. Fewer patients in the active groups exhibited digits with dactylitis at week 24 and an enthesitis score ⩾1. For patients with at least a 3% body surface area, a significantly greater number of patients in the ustekinumab group achieved PASI 75 as compared with placebo; 57.2% in 45 mg ustekinumab (p < 0.001) and 60.8% in the 90-mg group (p < 0.001) compared with placebo (11.0%). The safety findings from PSUMMIT-1 were similar to that noted in the larger psoriasis studies.^26–29 The three most common AEs were nasopharyngitis, upper respiratory tract infections and headache, all of which were <5%. Overall, three patients had major cardiovascular events after starting ustekinumab. No opportunistic infections (including tuberculosis), malignancies or death were reported up to 52 weeks. ²⁴

PSUMMIT-2, which had a similar design as PSUMMIT-1, included patients that were previously exposed to anti-TNF agents (⩾8 weeks of etanercept, adalimumab, golimumab, certolizumab pegol or 14 weeks of infliximab) along with those that failed NSAID or DMARD therapy as noted in PSUMMIT-1. Also at the start of PSUMMIT-2, inclusion criteria indicated that CRP should be at least 6.0 mg/l (compared with 3.0 mg/l in PSUMMIT-1) and early escape and crossover were conducted in the same manner as in PSUMMIT-1. Compared with PSUMMIT-1, PSUMMIT-2 was a smaller study with 103 patients randomized to the 45 mg ustekinumab group, 105 patients to the 90 mg ustekinumab group and 104 patients in the placebo arm. ²⁵ A total of 180 patients were anti-TNF-experienced and 132 were anti-TNF-naïve in PSUMMIT 2.

At week 24, 43.7% of patients receiving the 45-mg dose (p < 0.001) and 43.8% at the 90-mg dose (p < 0.001) were ACR20 responders as compared with 20.2% of patients in the placebo group. When stratified by anti-TNF use, the response was better for the anti-TNF-naïve group as compared with the TNF-exposed group (45 mg ustekinumab: 53.5 versus 36.7%; 90 mg ustekinumab: 55.3 versus 34.5 %; placebo: 28.6 versus 14.5%). Overall, there was no significant difference among patients with or without concomitant methotrexate therapy. There was also a statistically significant improvement in the major secondary endpoints which included ACR50, ACR70, PASI 75 and HAQ-DI score at week 24. Generally, the 90 mg ustekinumab group scores were numerically greater than the 45 mg ustekinumab group for the primary and secondary major outcomes. Similar to the primary endpoint, the response rates for most secondary endpoints were higher for anti-TNF-naïve patients than anti-TNF-experienced patients. For example, the overall PASI 75 response was substantially better with patients that were not exposed to anti-TNF. The PASI 75 for the anti-TNF-naïve group was 58.3% in the 45 mg ustekinumab group and 62.5% in the 90 mg ustekinumab group compared with PASI 75 of 45.5% and 48.8% respectively in the anti-TNF-experienced group. Also, the larger ustekinumab dose (90 mg) generally achieved slightly better response than the 45-mg dose in several secondary outcomes. For instance, the change in HAQ-DI was numerically higher for the 90 mg ustekinumab group [−0.25 (−0.50–0.00)] as compared with the 45 mg ustekinumab group [−0.13 (−0.38–0.00)]; however, the study was not powered to detect differences between the two treatment arms.

The PSUMMIT-2 trial revealed no new safety signals. The AEs were similar between the groups with the most common being nasopharyngitis and headache. Within the first 24 weeks, serious AEs occurred in 4.8% in the placebo group and 1.3% in the combined ustekinumab group. Through week 60, two ustekinumab patients reported serious infections, three patients had myocardial infections, and there were no cases of tuberculosis or deaths.

A prespecified analysis permitted integrated analysis of the PSUMMIT-1 and PSUMMIT-2 dataset to assess radiographic outcomes. ³⁰ Plain radiographs of the hands and feet at baseline and weeks 24 and 52 were analyzed. At week 24, no radiographic progression was noted in 91.7% of the combined ustekinumab group as compared with 83.8% of the controls (p = 0.005). Minimal radiographic progression of PsA was maintained between 24 and 52 weeks with minimal progression of 0.2 and 0.3 units of the total PsA modified vdH-S (van der Heijde-modified Sharp score).

Thus, both phase III studies clearly demonstrated that ustekinumab is effective in treating most domains of PsA and exhibits an acceptable safety profile. Efficacy was noted as early as 4 to 8 weeks and plateaued around 24 to 28 weeks. Ustekinumab worked with and without methotrexate and anti-TNF, although the response among patients previously exposed to anti-TNF was lower. The lower response rate among biologic-exposed patients has been a consistent feature among most studies in rheumatoid arthritis, ankylosing spondylitis and psoriasis. Clinical improvements also led to significant improvement in multiple patient reported outcomes including fatigue and mental component of SF36 (36-item short form health survey). ³¹ Collectively, IL-12/23 blockade resulted in a decrease rate of radiographic progression in PsA.

Ustekinumab in PsA: real-world studies

Only one study to date has been published on drug survival and effectiveness of ustekinumab using real-world data. ³² PsA patients in this study were retrieved from the BIOlogic aPUlian REgistry (BIOPURE), an Italian longitudinal cohort of patients with rheumatic diseases that received biologic treatments. They enrolled 160 PsA patients initiating ustekinumab from 2014 to 2016, of which 54 were biologic-naïve and 106 were biologic-experienced. Among the biologically experienced patients, nearly half the PsA patients experienced two or more anti-TNF agents. The entire cohort and the two relevant subsets, biologic-naïve and biologic-experienced, all demonstrated a statistically significant reduction in global disease activity as measured by the DAPSA (Disease Activity Psoriatic Arthritis), disability as assessed by the HAQ, and skin scores (PASI) at 6 months. Not surprisingly, biologic-naïve patients achieved the best outcomes with respect to clinical effectiveness at 12 months (DAPSA remission: 34% for biologic-naïve and 15% for biologic-experienced; p = 0.05) and drug survival at 12 months (biologic-naïve: 87%; biologic-exposed 68%; p = 0.01). Concomitant treatment with methotrexate did not increase the survivability of ustekinumab. This real-world study mirrors what has been reported in phase III clinical trials. Ustekinumab is an effective agent for the management of anti-TNF naïve and experienced PsA patients, and the best outcomes are achieved among PsA patients that are biologic-naïve.

Role of ustekinumab in psoriasis

Plaque psoriasis was the first indication for which ustekinumab received United States Food and Drug Administration (US FDA) approval, supported by two well-designed, phase III, multi-center, double-blind, placebo-controlled trials (PHOENIX-1 and PHOENIX-2), which were performed in parallel and together enrolled nearly 2000 patients with moderate-to-severe plaque psoriasis.^28,29 Both studies consisted of three phases: (1) a placebo-controlled phase (both trials); (2) a placebo-crossover and active treatment phase (both trials); and (3) a randomized withdrawal phase (PHOENIX-1) ²⁸ or dose intensification phase (PHOENIX-2). ²⁹ At study onset, patients were randomly assigned to one of three groups including placebo, ustekinumab (45 mg or 90 mg) subcutaneously at week 0, 4 and then every 12 weeks onwards. At week 12, patients receiving placebo were re-randomized to cross over to the 45 mg or 90 mg group. In PHOENIX-1, patients entered the randomized withdrawal phase at week 40, whereby those achieving a PASI 75 were re-randomized to continue with maintenance ustekinumab or discontinue active treatment. Those with a 50% loss in PASI improvement after discontinuing ustekinumab were re-treated. ²⁸ PHOENIX-2 offered increased dosing flexibility, whereby patients entering the dose intensification phase at week 28 who were partial responders were re-randomized to continue with their current dosing every 12 weeks or increased to an 8-week dosing regimen. ²⁹ Both studies met their primary endpoint of proportion of patients achieving PASI 75 at week 12, which was significantly higher in the active treatment groups in both studies when compared with placebo with a percent difference of 63.9% (PHOENIX-1) and 63.1% (PHOENIX-2) for the 45 mg ustekinumab group (p < 0.0001) and 63.3% (PHOENIX-1) and 72% (PHOENIX-2) for the 90 mg ustekinumab group (p < 0.0001).

Similar response rates were achieved in those originally randomized to the placebo group who crossed over to active treatment at week 12, in both trials. In PHOENIX-1, those randomized to withdrawal at week 40 noted a gradual decrease in PASI 75 over time, with significantly more patients in the maintenance group experiencing a sustained PASI 75 response (p < 0.0001), suggesting the need for long-term maintenance therapy. ²⁸ Encouragingly, 85.6% of those requiring re-treatment attained a PASI 75 within 12 weeks. ²⁸ In PHOENIX-2, 22.7% of the 45 mg ustekinumab group and 15.8% of the 90 mg ustekinumab group were nonresponders at week 28. ²⁹ Dosing intensification did not translate into significant gains in PASI 75 in the 45 mg ustekinumab group; however, in the 90 mg ustekinumab group, increasing the dosing frequency to every 8 weeks generated a higher PASI 75 response rate (35.4% at 52 weeks; p = 0.004) when compared with standard maintenance dosing. ²⁹ PHOENIX-2 highlighted important characteristics associated with poor response to treatment including, higher bodyweight, longer duration of disease, presence of PsA and previous resistance to biologic agents. Significant improvements in patient reported outcomes (PROs) as assessed by the DLQI were also noted in both trials for active treatment patients versus placebo. Ustekinumab was generally well tolerated, with no major safety concerns and side-effect profiles similar to those seen in the PSUMMIT trials when followed out to 76 (PHOENIX-1) and 52 weeks (PHOENIX-2) with no dose-dependent difference in rates or types of AEs. Both the efficacy and safety profile were maintained with data out to 5 years for both studies.^26,27

A further study comparing ustekinumab to etanercept revealed superior efficacy of ustekinumab at a dose of both 45 mg and 90 mg subcutaneously at weeks 0 and 4 when compared with twice weekly dosing of 50 mg of etancercept. ³³ At week 12, 67.5% of the 45 mg ustekinumab group and 73.8% of the 90 mg ustekinumab group achieved a PASI 75 response compared with 56.8% in the etanercept group (p = 0.01 and p < 0.001, respectively). In summary, there is robust clinical evidence supporting the use of ustekinumab in psoriasis as an effective therapeutic agent with no increased safety signal when patients are followed out to 5 years.

Role of ustekinumab in irritable bowel disease

To date, there have been three phase III randomized, placebo-controlled trials of ustekinumab in moderate-to-severe Crohn’s disease, totalling 1369 patients. Of these, two were 8-week induction trials (UNITI-1 and UNITI-2), and one was a 44-week maintenance trial (IM-UNITI). ³⁴ In UNITI-1, 741 patients who previously failed, or were intolerant to anti-TNF therapy, were randomized to one dose of intravenous ustekinumab at either a fixed dose of 130 mg or weight-based dosing of 6 mg/kg versus placebo. The primary endpoint of clinical remission at week 6 was achieved in 33.7% of the weight-based dose group, 34.3% of the fixed-dose group and 21.5% of placebo (p = 0.003 and 0.002, respectively). ³⁴ In UNITI-2, 628 patients who had previously failed conventional therapy, excluding anti-TNF agents were randomized as in UNITI-1. Clinical response at week 6 was achieved in 55.5% of the weight-based dose group, 51.7% of the fixed-dose group and 28.7% of placebo (all p < 0.001). ³⁴ Lastly, a total of 397 patients who achieved a clinical response at week 8 from both UNITI-1 and 2 trials entered the maintenance IM-UNITI trial and were randomized to maintenance therapy with 90 mg of subcutaneous ustekinumab, either every 8 (q8) or 12 (q12) weeks, or to placebo. The primary endpoint of sustained remission at week 44 was achieved in 53.1 % of the q8-week dosing, 48.8% of the q12-week dosing and 35.9% of placebo (p = 0.005 and p = 0.040, respectively). ³⁴ Similar to the data from previous phase III trials using ustekinumab for other indications, ustekinumab was generally well tolerated with rates of AEs similar between the active treatment group and placebo. In those who received at least one dose of ustekinumab, there were four cases of nonmelanoma skin cancer, two opportunistic infections, one case of active pulmonary tuberculosis, one nonfatal stroke and no deaths. ³⁴ A phase III randomized-controlled trial assessing the efficacy of ustekinumab in moderate-to-severe ulcerative colitis (UNIFI) is currently ongoing. ³⁵

In summary, ustekinumab is a promising new therapeutic agent in both induction and maintenance treatment of active Crohn’s disease, but its place in the treatment algorithm with respect to anti-TNF and anti-integrin agents remains to be further elucidated.

Conclusion

Psoriasis is a stubborn chronic skin disease affecting 1–3% of the population. PsA occurs in up to 40% of patients with psoriasis and usually follows the skin manifestations. Consequently, PsA can be considered as a disease within a disease, sharing many common pathogenic pathways with psoriasis. In particular, the Th-17 axis (inhibited by IL-23) is considered to play an important role in the immunopathogenesis of psoriasis and PsA. The efficacy of ustekinumab, a human monoclonal antibody that targets the p40 subunit common to IL-12 and IL-23, is mostly due to its effect on Th-17 cells, but it may also be influenced by inhibition of Th-1 cells, through its inhibition of IL-12. Ustekinumab has demonstrated sustained efficacy and good tolerability and safety in the management of psoriasis. Two key phase III ustekinumab studies in PsA, further strengthen the use of this biologic agent in psoriatic disease. The PSUMMIT-1 and PSUMMIT-2 studies, met all important clinical endpoints in PsA, including articular and periarticular outcomes. The drug was effective with and without methotrexate and more importantly was effective in the treatment of patients that were previously exposed to anti-TNF agents. In psoriatic disease, ustekinumab represents a clear alternative and satisfies an unmet need for patients that have either failed or are contraindicated to anti-TNF agents. Robust patient reported and radiographic data support the use of ustekinumab in PsA. While duration of follow up in PsA is still limited, the longevity of its safety record in psoriasis patients is reassuring. Data from the PSOLAR (Psoriasis Longitudinal Assessment and Registry) study suggests that rates of serious infections are actually lower with ustekinumab in comparison with other biologic agents, specifically anti-TNF inhibitors. ³⁶ At present, the data support the use of ustekinumab in the treatment of PsA after the failure of NSAIDs and conventional DMARDs as an alternative to, or after failure of an anti-TNF agent.