WHAT IS ALREADY KNOWN ON THIS TOPICWHAT THIS STUDY ADDS

Spinal radiographic progression is associated with baseline axial involvement defined as radiographic sacroiliitis.

There is a four times higher probability of syndesmophytes development when a clinical/radiographic description of axial involvement is combined with high C reactive protein.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICYIntroduction

Psoriatic arthritis (PsA) is a heterogeneous chronic inflammatory disease commonly involving the peripheral joints and is associated with skin and nail psoriasis, enthesitis, dactylitis and, like with psoriasis, a reduced health-related quality of life and physical disability.1 The disease can also affect the axial joints in the spine and the sacroiliac joints (SIJ), which then occasionally is referred to as axial PsA (axPsA). Depending on the description axial involvement is reported in 25%–70% of the patients with PsA.2 The wide variability is a direct cause of the ongoing debate how to define axial involvement in PsA. With the lack of explicit definitions for axial involvement or criteria specifically for axPsA the classification of axPsA is subject to many interpretations. Back pain that worsens in rest or improves with exercise and morning stiffness (lasting longer than 30 min) are symptoms considered typical for axPsA.3 These features play a major role in defining inflammatory back pain (IBP). Other studies focus on radiographic damage when quantifying axial involvement in PsA. Some investigators focus on the presence of syndesmophytes or other indications for spinal damage such as sclerosis or erosions,4 while others describe axPsA as unilateral or bilateral sacroiliitis using either the modified New York criteria (mNY)3 5 or Assessment of SpondyloArthritis International Society (ASAS) definition for a positive MRI.6 7

There is a paucity of information on axial radiographic progression in patients with PsA. A pilot study reported spinal radiographic progression at a rather slow rate of 0.6 modified Stoke Ankylosing Spondylitis Spinal Score (mSASSS) units per year.8 However, this study was conducted in only 22 patients and results from larger studies are needed.

In the current study, we aim to correlate several definitions of axial involvement to the development of syndesmophytes over 2 years in patients diagnosed with PsA.

MethodsStudy design and data collection

This study included patients enrolled in the Belgian Epidemiological Psoriatic Arthritis Study (BEPAS).9 Details on the purpose and design of the study have been published before. In short, BEPAS is a national, multicentre, non-interventional, epidemiological trial. Between December 2012 and July 2014, patients age ≥18 years with a clinical PsA diagnosis also fulfilling the Classification Criteria for Psoriatic Arthritis were enrolled from 17 academic, non-academic and private practice centres across Belgium. The selection procedure was unbiased and systematic as weekly the first, third and fifth eligible PsA patient presenting at the different sites were asked to participate in the study.

Patients/public were not involved in the creation of the design nor methodology of this particular study.

Axial involvement description

There were eight definitions of ‘axial involvement’ that were tested in this study; eight are clinically oriented and three have a radiographic basis (figure 1). All clinical definitions were based on the assessment of the physician. The clinical definition of ‘physician global assessment’ represented the global assessment of the patient’s axial involvement by the treating physician. A less stringent definition is the ‘physician detailed assessment’, which contained the treating physician’s evaluation of the following features on past or present occurrence: (1) spinal pain, (2) night pain, (3) morning stiffness, (4) anterior chest wall pain, (5) alternating buttock pain, (6) improvement of pain with exercise and (7) back pain onset <45 years. If any of the features was present, it was considered ‘axial involvement’, unless feature 7 (back pain onset <45 years) was solely present, then this was not considered ‘axial involvement’. ‘Chronic back pain’ (CBP) was defined by the treating physician and IBP was defined according to the ASAS definition.10 Clinical definitions were expanded by adding a C reactive protein (CRP) of ≥10 mg/L (elevated) to ‘CBP’ and ‘IBP’.

Figure 1

Clinical and radiographic definitions of axial involvement in patients with PsA from the Belgian Epidemiological Psoriatic Arthritis Study cohort. ASAS, Assessment of SpondyloArthritis International Society; CRP, C reactive protein; IBP, inflammatory back pain.

With only radiographs available in this study, the radiographic axial involvement was purely focused on fulfilment of mNY criteria. Also here the combination of radiographic and elevated CRP was included as separate definitions.

Radiographic imaging assessment

At scheduled baseline and year 2 visits pelvic and spinal radiographs were performed. Two calibrated central readers (AI and MdH) performed two training sets and one calibration round with recurring evaluation meetings in between, resulting in substantial interreader agreement (kappa>0.70). This process was conducted separately for SIJ and spine. After calibration the imaging assessment used for the current study started. Each reader evaluated spinal radiographs separately from the pelvic radiographs. Readers were blinded for time sequence, origin of the cohort, clinical data and information from other obtained images (radiographs of hands and feet).

Pelvic radiographs were scored according to the mNY criteria. With this method SIJ are assessed on anteroposterior view plain radiographs. Radiographic changes of each SIJ are graded on a 0–4 scale. The mNY criteria are fulfilled in case of a bilateral grade 2 or unilateral grade 3 or 4 score (mNY+).11 Spinal radiographs were assessed using the mSASSS. With this method, the anterior vertebral corners (VCs) of the cervical (lower border of C2 to upper border of T1) and lumbar (lower border of T12 to upper border of S1) spine (total of 24 VCs) are scored in the lateral view as 0 if VC was normal, 1 when erosion, sclerosis and/or squaring was present in absence of syndesmophyte, 2 if a non-bridging syndesmophyte was present and 3 with the presence of bridging syndesmophytes. Total score per patient ranges from 0 to 72.12

Statistical analysis

Descriptive statistics were reported either as means with SD or as percentages. The probability of developing syndesmophytes after 2 years was described conditional on the presence or absence of axial involvement at baseline. Per definition of ‘axial involvement’ logistic regression analyses were used investigating the association between axial involvement and the development of syndesmophytes 2 year later. Consensus mNY reader scores were used in analyses for radiographic axial involvement. If individual reader scores were used this was indicated accordingly. Statistical analyses were preformed by using STATA V.17.

Results

In total 461 patients (57% males; 52±12 years age) were enrolled in the BEPAS by 17 centres between November 2012 and July 2014. Of those there were 150 patients with baseline and 2-year follow-up radiographs included in this study. Table 1 shows average age was 54.3±12.3 years, 62 patients (41.3%) were female and the mean disease duration was 8.8±10.2 years. Two-third of the patients did not have HLA-B27 typing, making its prevalence not representative. When physicians made a global assessment 47.3% of the patients had axial involvement. This number increased to 76.5% when axial involvement was defined less stringently (physicians detailed assessment). CBP was seen in 67.6% while IBP was only seen in a minority of the patients (16.7%).

Table 1

Characteristics of psoriatic arthritis patients from the BEPAS cohort

Only five patients (3.3%) showed radiographic sacroiliitis according to reader consensus score. This increased to nine patients (6.0%) when considering individual reader findings. All mNY+ patients had peripheral manifestations like dactylitis, heel enthesitis or swollen joint count. There were 19 patients (12.7%) with syndesmophytes at baseline. On average these patients had 1.9±0.9 syndesmophytes. At 2 years, 20 patients (13.3%) had syndesmophytes with an average of 2.1±0.9 syndesmophytes. In total 11 patients developed syndesmophytes over 2 years follow-up; 6 patients with no syndesmophytes at baseline (range 1–4 new syndesmophytes after 2 years) and 5 patients with ≥1 syndesmophyte at baseline whom all developed syndesmophytes over 2 years follow-up at an average rate of 1.1 mSASSS.

The vast majority of patients (~90%) show no syndesmophyte development over 2 years. Nevertheless the probability of radiographic spinal progression seems to be increased when ‘axial involvement’ is defined as ‘IBP and elevated CRP’ or with radiographic axial involvement (table 2).

Table 2

Probability of radiographic spinal progression after 2 years with and without axial involvement defined as (1) clinical axial involvement or (2) radiographic axial involvement in patients with PsA from the BEPAS cohort (n=150)

This is concurred by the regression analyses as all clinical definitions of axial involvement showed low association (ORs 0.83–1.41, table 3) with the exception of axial involvement defined as ‘CBP and elevated CRP≥10’ (OR 2.55, 95% CI 0.49 to 13.27) and ‘IBP and elevated CRP≥10’ (OR 4.08, 95% CI 0.87 to 19.20); none of which were statistically significant. Radiographic axial involvement showed a non-statistically significant OR of 3.38 (95% CI 0.34 to 33.11). When combining with elevated CRP, the OR could not be calculated as this involved a single patient (table 3).

Table 3

Association of axial involvement according to various definitions and developing syndesmophytes in 2 years in patients with PsA from the BEPAS cohort (n=150)

Discussion

In this study, the associations between different definitions of axial involvement and the development of radiographic syndesmophytes in patients with PsA are tested. In this unselected PsA population with a large percentage of physician-reported axial complaints, syndesmophyte progression is overall low. We found that despite the minimal amount of radiographic spinal damage the probability of radiographic spinal progression over 2 years increases 3–14 times in PsA patient with radiographic axial involvement defined as radiographic sacroiliitis according to the mNY criteria.

The discussion on axial involvement in PsA is ongoing and will be left without conclusive statements until a universally accepted definition for ‘axial involvement’ is found. There are numerous descriptions for axial involvement and due to this the prevalence of patients with axPsA varies tremendously in the literature. By using an extensive variety of definitions we aimed to contribute to the insights of what axial involvement in PsA is and we aimed to find possible associations with radiographic spinal progression. In this study, we interpreted axial involvement with a clinical as well as radiographic approach. The physician drove the clinical definitions of axial involvement. It is suggested that with regard to pain assessment as an outcome measure the physicians opinion is preferred over patient assessment as most patients (>85%) rate their global pain assessment higher than their physician.13 Nevertheless, it remains a prejudiced evaluation, which is subject to large fluctuations. By including an elevated CRP, the definition contains an objective component, of which we hypothesised that ‘axial involvement’ takes on a more distinctive character. Our results concur that indeed. The probability of spinal radiographic progression intensified for both clinical and radiographic axial involvement when combined with high CRP. Only with CBP as an indication for axial involvement this intensification was not so pronounced. In this study, the physician reported CBP in over two-thirds of the patients. Argument could be made that CBP in general is not sufficiently specific to serve as a proxy for axial involvement. The probability of developing syndesmophytes over 2 years increased seven times when axial involvement was considered as ‘IBP with elevated CRP’. Yet, the OR of 13.8 was not statistically significant, which presumably could be a power problem.

In this population with moderate long disease duration, approximately 10% of the patients develop new syndesmophytes over 2 years, most of them developed just one new syndesmophyte. In a study on spinal progression in axial spondyloarthritis (axSpA) patients an average rate of 1.8 mSASSS was reported in the first 2 years of follow-up. The absolute rate is slightly higher than the 1.1 mSASSS units this study reports. Nevertheless this could be expected as the OASIS cohort included radiographic axSpA patients with longer symptom duration. Therefore, comparing these data, we see similar radiographic spinal progression in PsA as in r-axSpA patients.14

Due to the low radiographic progression, the associations found in this study were not statistically significant. Still the descriptive data, especially presented in the conditional probability tables, show that the development of a syndesmophyte was three times more likely when the patient also presented with radiographic sacroiliac involvement (fulfilling mNY criteria) at baseline. When on top of that also a high CRP was established in these patients the probability of new syndesmophytes was 14 times higher. Note should be made that there was only one patient that fulfilled this condition (mNY+ and elevated CRP), which shows the scarceness of axial radiographic damage in this cohort. These results suggest that radiographic sacroiliitis is a robust description of axial involvement, especially when combined with an objective inflammatory factor such as CRP. Based on our results, we suggest that in the absence of radiographic imaging, IBP in combination with CRP probably serve as the best proxy for axial involvement in PsA.

At time of developing, the BEPAS protocol (before 2012) MRI was not a standardised procedure in daily clinical practice for patients with PsA. The lack of MRI data in this study leaves a gap in the range of definitions of axial involvement. With MRI a preradiographic stage of axial damage can be detected by assessing inflammatory lesions and structural lesions such as fat deposition. This could be a key component in future’s globally accepted nomenclature of axial involvement. A future study was recently announced to explicitly address this current breach in axPsA research.7 In addition to the inability to detect transient lesions, the radiographic imaging assessment used in this study also holds the possibility of missing spinal damage in the form of erosions or ankylosis located in the facet joints. In this study, mSASSS was used to evaluate spinal damage, as per suggestion of the ASAS group on core outcome set instruments selection.15 The mSASSS method does not take into account facet joints. Not embracing the facet joints is a major limitation of any validated scoring method assessing radiographic spinal damage.16 In this particular study, it may result in underestimating radiographic spinal damage. On the other hand, the scoring of facet joints is a very complicated endeavour, as a validated and widely used scoring system is currently missing. Also the absence of HLA-27 typing in most patients is a limitation to this study. Unfortunately at time of developing, the BEPAS protocol (before 2012) HLA-B27 was not a standardised feature in daily practice nor covered by healthcare. In hindsight, the study would have definitely benefited from adding HLA-B27 typing but at the time the decision was reasonable.

The systematic approach of having two independent readers blindly assessing all pelvic radiographs separately from the spinal radiographs is a major strength of this study. In addition, readers showed excellent agreement, which is assuring for the robustness of the data even though the statistical analyses were preformed on vertebral unit (VU) level and adjusted for reader assessment. Another strength is the selection of patients for the BEPAS. Selection bias was kept to a minimum due to the inclusion procedure: every week the first, third and fifth eligible PsA patient presenting at the different sites was asked to participate in the study. Using this method, an unbiased systematic selection of patients was obtained.

Summarising, this study shows a rarity of syndesmophyte formation over 2 years in patients with PsA, in particular without baseline radiographic sacroiliitis. Though rare, radiographic syndesmophyte development is more likely in PsA patients with radiographic sacroiliac involvement at baseline, particularly in combination with elevated CRP, compared with a purely clinical description of axial involvement.

Data availability statement

Data are available on reasonable request. The data underlying this article will be shared on reasonable request to the corresponding author.

Ethics statementsPatient consent for publication

Not applicable.

Ethics approval

All participating patients in the BEPAS gave their written informed consent and the study was conducted according to Good Clinical Practice/International Council for Harmonisation guidelines and the Declaration of Helsinki. This study involves human participants and was approved by an Ethics Committee(s) or Institutional Board(s) (University Hospitals Leuven, approved on 27 July 2012 (B322201215141)).