Outcome measures

The primary outcome in this study is the change in PASDAS score at 24 weeks. The PASDAS is a weighted composite index encompassing joints counts, an assessment of enthesitis and dactylitis, acute phase reactant, QoL, and patients’ and physician’s global assessment by visual analogue scale (VAS). The PASDAS has been shown to perform well in both oligoarticular and polyarticular forms of PsA [36] and cut-offs for disease activity and response have now been developed and validated using interventional trial data [34].

Secondary outcomes at weeks 12, 24, 36 and 52 include the Leeds enthesitis index (LEI) [38], the Leeds dactylitis index basic (LDI-B) [39, 40], the Psoriasis Area and Severity Index (PASI) score [41], the Body Surface Area (BSA) affected by psoriasis, the modified Nail Psoriasis Severity Index (mNAPSI) [42], the proportion of subjects achieving MDA, the proportion of subjects achieving the American College of Rheumatology Response Criteria [43], the proportion of subjects achieving the PASI75 response (defined as having an improvement of at least 75% in the PASI score compared to baseline levels). The proportion of subjects achieving PsARC response is a secondary outcome to be collected at weeks 8, 12, 24, 36 and 52. The Composite Psoriatic Disease Activity Index (CPDAI) [44] is a secondary outcome to be collected at weeks 24 and 52. The proportion of patients requiring additional steroid therapy, as well as the cumulative steroid dose up to week 12 of this trial, will be also recorded as a secondary outcome.

Systematic WB-MRI scanning of the axial and peripheral skeleton (Table ​(Table3)3) is performed in all suitable subjects (Table ​(Table4),4), using commercially available Siemens MAGNETOM® Verio 3 T scanner with the following sequences: (i) T1-weighted spin echo (SE) before and after an intravenous gadolinium contrast injection; (ii) Short Tau Inversion Recovery (STIR). The estimated scanning time is 60 min. Images are acquired in two phases, allowing patients to mobilize from the scanner for a short time. MRI features of inflammation (ie: synovitis, BMO or osteitis lesions) and damage (erosions, bone formation, fat infiltration, sclerosis and/or ankylosis) will be scored at peripheral joints, entheses and axial skeleton including sacroiliac joints and spine using a novel scoring system.

Body areas undergoing magnetic resonance imaging (MRI) and ultrasound (US) evaluation

Eligibility criteria to gadolinium contrast-enhanced magnetic resonance imaging at the Leeds Musculoskeletal Biomedical Research Unit

All criteria must be satisfied to fulfil eligibility

US scanning of selected joints and entheses of lower and upper limbs (see Table ​Table3),3), is performed using a multi-planar technique with symmetrical scanning by sonographer blinded to the participant’s clinical characteristics. Articular and entheseal sites will be assessed for the presence of grey scale (GS) abnormalities and power-Doppler (PD) signal. US pathological findings will be identified according to the Outcome Measures in Rheumatoid Arthritis Clinical Trials (OMERACT) definitions [45–47].

The patients’ overall assessment of PsA activity will be recorded at baseline and then at weeks 12, 24, 36 and 52, using the 100 mm horizontal VAS using the specific wording proposed by Cauli et al. [48] for PsA. To calculate the PsARC (at weeks 8, 12, 24, 36 and 52), the same assessment will be also recorded on a Likaert scale ranging from 1 to 5. All participants will fill in the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) [49] questionnaire at baseline and then at weeks 12, 24, 36 and 52. The outcomes relating to QoL and health status are: the Health Assessment Questionnaire Disability Index (HAQ-DI) [50] score; the Dermatology Life Quality Index (DLQI) [51] score; the Ankylosing Spondylitis Quality Of Life (ASQOL) [52] score; the Short Form (SF-36) [53] score. All these questionnaires will be collected at baseline and then at 12, 24, 36 and 52 weeks.

Our null hypothesis (H0) is that the difference between the two treatment arms in the PASDAS score at week 24 is equal to zero. Hence, the alternative hypothesis (H1) is that the difference between the two treatment arms in the PASDAS score at week 24 is not equal to zero. In general, summary statistics (n = number of available measurements; arithmetic mean; standard deviation; median; minimum and maximum) for quantitative variables and absolute and relative frequency tables for qualitative data will be presented. Analyses will be adjusted for the randomisation stratification factor(s) and baseline values of the outcome; 2-tailed tests will be performed and will be considered statistically significant if p < 0.05.

The primary endpoint will be assessed on an intention-to-treat (ITT) basis. Analysis of covariance by multiple linear regression will be used to compare PASDAS between the two treatment groups at week 24. Binary secondary endpoints will be analysed using multiple binary logistic regression. Continuous interval outcomes will be analysed using multiple linear regression. Severely skewed or ordinal outcomes will be analysed using quantile regression. Planned subgroup analyses will investigate differences in treatment response according to oligo/polyarthritis status, immunological status, disease duration at baseline. A per protocol analysis will also be performed.

The frequency of all Serious Adverse Events (SAEs) during the study period in patients who received at least 1 dose of study treatment will be presented for each treatment group separately. The data will be displayed as number of subjects experiencing the SAEs, percentage of subjects, and number of SAEs. Data will also be corrected for exposure by 100 patient-years.

For patients who withdraw early, data from the withdrawal visit will be imputed for subsequent visits for continuous outcomes, and non-response will be imputed for binary outcomes. For all other instances of missing data, multiple imputation will be used. A number of sensitivity analyses testing robustness of conclusions under different missing data mechanisms will be conducted.

Using in-house unpublished data, we estimated the minimum clinically important difference to be 0.7 units on the PASDAS; this is similar to a published value for smallest detectable difference of 0.8 units [34]. We will aim to detect a difference of at least 1 unit between the treatment arms in this study. The standard deviation of PASDAS in the Tight Control of Psoriatic Arthritis (TICOPA) [54] MTX rapid escalation arm at 24 weeks (restricted to patients who remained on methotrexate throughout) was 1.57. Assuming δ = 1, σ = 1.57, at alpha = 0.05 and 1-Beta = 0.8 this would require 78 patients; accounting for 10% drop-out we will aim to recruit a total of 88 (44 per group).