Llop et al. Arthritis Research & Therapy (2019) 21:126 https://doi.org/10.1186/s13075-019-1913-z RESEARCH ARTICLE Open Access Incorporation of the anteroposterior lumbar radiographs in the modified Stoke Ankylosing Spondylitis Spine Score improves detection of radiographic spinal progression in axial spondyloarthritis Maria Llop1,2, Valeria Rios Rodriguez1, Imke Redeker1,3, Joachim Sieper1, Hildrun Haibel1, Martin Rudwaleit1,4 and Denis Poddubnyy1,3* Abstract Background: To evaluate the performance of the extended modified Stoke Ankylosing Spondylitis Spine Score (mSASSS) incorporating information from anteroposterior (AP) lumbar radiographs as compared to the conventional mSASSS in detection of radiographic spinal progression in patients with axial spondyloarthritis (axSpA) Methods: A total of 210 patients with axSpA, 115 with radiographic axSpA (r-axSpA), and 95 with non-radiographic axSpA (nr-axSpA), from the GErman SPondyloarthritis Inception Cohort (GESPIC), were included in the analysis based on the availability of spinal radiographs (cervical spine lateral, lumbar spine lateral, and AP views), at baseline and year 2. Two trained readers independently scored lateral cervical and lumbar spine images according to the mSASSS system (0–3 per vertebral corner, 0–72 in total). In addition, all vertebral corners of vertebral bodies visible on lumbar AP radiographs (lower T12 to upper S1) were assessed according to the same scoring system that resulted in a total range for the extended mSASSS from 0 to 144. Reliability and sensitivity to detect radiographic spinal progression of the extended mSASSS as compared to the conventional mSASSS were evaluated. Results: The reliability of conventional and extended scores was excellent with intraclass correlation coefficients (ICCs) of 0.926 and 0.927 at baseline and 0.920 and 0.933 at year 2, respectively. The mean ± SD score for mSASSS and extended mSASSS at baseline were 4.25 ± 8.32 and 8.59 ± 17.96, respectively. The change score between baseline and year 2 was 0.73 ± 2.34 and 1.19 ± 3.73 for mSASSS and extended mSASSS, respectively. With the extended mSASSS, new syndesmophytes after 2 years were detected in 4 additional patients, new syndesmophytes or growth of existing syndesmophytes in 5 additional patients, and progression by ≥ 2 points in the total score in 14 additional patients meaning a 25%, 28%, and 46% increase in the proportion of patients with progression according to the respective definition as compared to the conventional score. Conclusions: Incorporation of lumbar AP radiographs in the assessment of structural damage in the spine resulted into detection of additional patients with radiographic spinal progression not captured by the conventional mSASSS score. Keywords: Axial spondyloarthritis, mSASSS, Radiographic spinal progression, Radiographs, X-rays * Correspondence: denis.poddubnyy@charite.de 1 Department of Gastroenterology, Infectious Diseases, and Rheumatology, Charité - Universitätsmedizin Berlin, Hindenburgdamm 30, 12203 Berlin, Germany 3 Epidemiology Unit, German Rheumatism Research Centre, Berlin, Germany Full list of author information is available at the end of the article © The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Llop et al. Arthritis Research & Therapy (2019) 21:126 Page 2 of 9 Introduction introduction of tumor necrosis factor inhibitors (TNFi) in Axial spondyloarthritis (axSpA) can be classified into the clinical practice, the vast majority of the patients radiographic SpA (r-axSpA, also known as ankylosing were off TNFi therapy at inclusion and during first spondylitis—AS) or non-radiographic SpA (nr-axSpA) 2 years of observation. depending on the presence or absence of definite radio- graphic sacroiliitis according to the grading system of Radiographic assessments the modified New York (mNY) criteria [1]. New bone Radiographs of cervical (lateral view) and lumbar spine formation with development of syndesmophytes between (lateral and AP views) were obtained at baseline and after vertebrae represents the morphological substrate of 2 years. Images were centrally collected, digitized, ano- structural damage in the spine in axSpA. Since develop- nymized, and scored independently by two trained readers ment of structural damage is usually assessed on conven- (DP and HH), who were blinded to the chronological tional radiographs, it is referred to as radiographic spinal order of the images and to all clinical data. According to progression. Structural damage in the spine and disease the mSASSS, anterior corners of the vertebral bodies activity are two major determinants of spinal mobility and from lower C2 to upper T1 (cervical spine) and from physical function in axSpA [2–4]. Currently, the modified lower T12 to upper S1 (lumbar spine) were scored as Stoke Ankylosing Spondylitis Spine Score (mSASSS) [5] is follows: 0 = normal, 1 = erosion, sclerosis, and/or squaring, considered as the gold standard for assessment of struc- 2 = non-bridging syndesmophyte, and 3 = bridging syndes- tural damage in the spine in axSpA [6]. Compared to the mophyte, giving a range for the entire score from 0 Bath Ankylosing Spondylitis Radiology Index (BASRI) [7] to 72 [5]. No adjudication was performed. The final and SASSS [8], mSASSS has a better reliability and mSASSS score included in the analysis was calculated sensitivity to change [9]. as a mean of the mSASSS scores of both readers. In The mSASSS assesses the presence of erosions, scle- addition, left and right, upper and lower vertebral rosis, squaring, non-bridging syndesmophytes, and bony corners of vertebral bodies visible on lumbar AP radio- bridges affecting the anterior vertebral corners of the graphs from lower Th12 to upper S1 were scored ac- cervical and lumbar spine on radiographs in lateral views cording the same grading system. The combination of [5]. Previous studies suggested that anteroposterior (AP) the scores from both lateral cervical and lumbar view of the lumbar spine, which are frequently performed views and lumbar AP views composed the “extended routinely, could provide additional information on the mSASSS” with a total range from 0 to 144. The final presence of structural damage in the spine [7, 9]. The extended mSASSS was calculated as a mean of the added value of the AP lumbar radiographs in the assess- scores of both readers. Syndesmophytes were considered ment of radiographic spinal progression in axSpA as to be present if both readers gave a score of ≥ 2 to a compared to lateral views only is, however, not known. vertebral corner. The aim of the current study was to evaluate the per- Missing scores of single vertebral corners were formance of the extended mSASSS score incorporating substituted with scores for the respective vertebral information of the AP lumbar radiographs compared to corners at other time point, if available, or with a score of the conventional mSASSS. zero if scores for both time points were missing. Patients and methods Data analysis Patients To assess the applicability of the extended mSASSS as a Patients with axSpA from the GErman SPondyloarthritis tool for the measurement of radiographic spinal progres- Inception Cohort (GESPIC) were included in the current sion, it was compared to the conventional mSASSS analysis based on the availability of clinical data and radio- according to the three aspects of the Outcome Measures graphs at baseline and year 2. A detailed description of the in Rheumatology (OMERACT) filter: feasibility, discri- entire cohort as well as of the radiographic subset has been mination, and truth [13]. reported previously [10–12]. Briefly, patients included in GESPIC had a definite clinical diagnosis of axSpA accord- Feasibility ing to the treating rheumatologist with symptom duration The feasibility aspect of the OMERACT filter refers to of up to 5 years for the non-radiographic form and up to the pragmatic reality on applying or using a specific 10 years for the radiographic form of axSpA. The final measure; which is decisive on its success: “Can the classification as nr-axSpA or r-axSpA was performed based measure be applied easily, given constraints of time, on the central reading of the radiographs as previously re- money, and interpretability?” [13]. The feasibility was ported [10]. Treatment of the patients was defined by local assessed in the context of time, radiation exposure, and rheumatologists and followed national and international cost excess associated with extended mSASSS compared guidelines; however, since the cohort had started prior to to the conventional one. Llop et al. Arthritis Research & Therapy (2019) 21:126 Page 3 of 9 Discrimination Statistical analysis was performed with IBM SPSS The discrimination aspect addresses the question: “Does Statistics version 24 (IBM, Armonk, NY, USA) and SAS the measure discriminate between situations of interest?” version 9.4 (SAS Institute Inc., Cary, NC, USA). This aspect of the OMERACT filter focuses on reliability and sensitivity to change [13]. Results We used the intraclass correlation coefficient (ICC) A total of 210 patients with axSpA, 115 with r-axSpA and Cohen’s kappa coefficient to measure the inter-ob- and 95 with nr-axSpA, were included in the analysis server reliability for continuous and categorical variables, based on availability of the full sets of radiographs: respectively. The analysis was performed for the overall lumbar (AP and lateral views) and cervical (lateral view) scores as well as separately for the components of the spine at baseline and after 2 years. Demographic, clinical, scores (cervical spine lateral view, lumbar spine lateral view, and radiographic data of the patients at baseline and lumbar spine AP view). Additionally, a Bland and Altman year 2 are presented in Table 1. plot was drawn and the smallest detectable change (SDC) was determined to assess the sensitivity to change of Feasibility both methods (mSASSS and extended mSASSS). The In this study, lumbar AP radiographs were available in SDC expresses the smallest change in scores that can all 210 patients that was a requirement of the inclusion be detected without measurement error, and it was in the radiographic subset. Although AP lumbar radio- calculated as follows: graph is a routine projection in the clinical practice, it SDC = 1.96 × SEM, increases the costs of the radiographic investigation of where SEM denotes the standard error of measure- the spine by approximately 50% and the time required ment of the change score obtained from a two-way for the reading according to the extended mSASSS by analysis of variance by taking the square root of the 50–100% (since additional 24 vertebral corners should error variance [14]. be assessed) as compared to conventional mSASSS. To measure the sensitivity to change, we used the Furthermore, AP lumbar radiographs are associated with following definitions of progression: (1) change of the additional radiation exposure, increasing the overall absolute score (calculated as means of the scores of both exposure associated with the assessment by appro- readers), (2) change of the score by ≥ 2 points after ximately 60–70%. 2 years, (3) development of at least one new syndesmo- phyte (score of 0 or 1 at baseline and score of 2 or 3 Discrimination after 2 years at the same vertebral corner in the opinion The status scores and progression data for the both of both readers), and (4) development of new syndesmo- scores is presented in Table 2. The mean ± standard phytes (as described above) or growth of existing syndes- deviation (SD) score at baseline for the whole group mophytes (score of 2 at baseline and 3 after 2 years at included in the analysis was 8.59 ± 17.96 and 4.25 ± 8.32 the same vertebral corner) in the opinion of both for the extended mSASSS and conventional mSASSS, readers. The sensitivity to change was evaluated in the respectively. The absolute change score between baseline entire group and in r-axSpA and nr-axSpA subgroups. and year 2 was 1.19 ± 3.73 and 0.73 ± 2.34 for the ex- We also analyzed both status and progression scores in tended mSASSS and conventional mSASSS, respectively the components of the mSASSS and the extended (Table 2). In the separate analysis of the components of mSASSS (cervical spine lateral view, lumbar spine lateral the score, the progression rates were comparable across view, lumbar spine AP view). A variance component the components (Table 3). analysis was performed. There was an excellent agreement between readers on the status scores with ICCs of 0.93 (95% CI 0.91–0.94) Truth and 0.93 (95% CI 0.90–0.94) at baseline and 0.93 (95% The truth aspect addresses the question: “Is the measure CI 0.91–0.95) and 0.92 (95% CI 0.90–0.94) at year 2 for truthful, does it measure what is intended? Is the the extended and conventional mSASSS, respectively. result unbiased and relevant?” [13]. Earlier studies The agreement for the change score was moderate with showed an association between the mSASSS and ICC of 0.45 (95% CI 0.33, 0.55) and 0.43 (95% CI 0.31, spinal mobility/functional status in axSpA [2–4]. With a 0.53) for the extended and conventional mSASSS, linear regression analysis, we investigated the relationship respectively. For the outcome progression by ≥ 2 points between both scores and spinal mobility (assessed by the after 2 years, there was a fair agreement for both Bath Ankylosing Spondylitis Metrology Index—BASMI, methods: Cohen’s kappa 0.20 (95% CI 0.06, 0.34) and an original two-step definition) and function (assessed by 0.21 (95% CI 0.06, 0.36) for extended and conventional the Bath Ankylosing Spondylitis Functional Index— mSASSS, respectively. The agreement for development BASFI). of new syndesmophytes was moderate with Cohen’s Llop et al. Arthritis Research & Therapy (2019) 21:126 Page 4 of 9 Table 1 Baseline demographic and clinical characteristics of the included patients with axial spondyloarthritis Parameter Non-radiographic axial SpA (n = 95) Radiographic axial SpA (n = 115) All patients (n = 210) Age, mean ± SD years 38.7 ± 9.9 36.1 ± 11.0 37.3 ± 10.6 Symptom duration, mean ± SD years 3.1 ± 2.2 5.0 ± 2.8 4.2 ± 2.7 Duration since diagnosis, mean ± SD years 1.0 ± 1.3 2.0 ± 2.0 1.5 ± 1.8 Male sex, n (%) 32 (33.7) 75 (65.2) 107 (51.0) HLA-B27 positive, n (%) 69 (73.4) 97 (84.3) 166 (79.4) Peripheral arthritis, n (%) 16 (16.8) 15 (13.0) 31 (14.8) † Enthesitis , n (%) 23 (24.2) 23 (20.0) 46 (21.9) Uveitis, ever, n (%) 15 (15.8) 27 (23.5) 42 (20.0) Psoriasis, ever, n (%) 11 (11.6) 17 (14.8) 28 (13.3) Inflammatory bowel disease, ever, n (%) 1 (1.1) 3 (2.6) 4 (1.9) Family history for SpA, n (%) 16 (16.8) 19 (16.5) 35 (16.7) BASDAI, mean ± SD, 0–10 4.2 ± 2.0 3.8 ± 2.2 4.0 ± 2.1 BASFI, mean ± SD, 0–10 2.8 ± 2.2 3.0 ± 2.4 2.9 ± 2.3 Treatment with NSAIDs, n (%) 65 (68.4) 78 (67.8) 143 (68.1) Treatment with csDMARDs, n (%) 26 (28.6) 35 (32.4) 61 (30.7) Treatment with systemic steroids, n (%) 6 (6.7) 6 (5.6) 12 (6.1) Treatment with a TNFα blocker, n (%) 1 (1.1) 4 (3.7) 5 (2.5) Smoking, current, n (%) 24 (25.3) 39 (33.9) 63 (30.0) † Twelve enthesitis sites of the lower limbs plus optional symptomatic sites elsewhere were assessed BASDAI Bath Ankylosing Spondylitis Disease Activity Index, BASFI Bath Ankylosing Spondylitis Functional Index, csDMARDs conventional synthetic disease- modifying antirheumatic drugs, NSAIDs nonsteroidal antiinflammatory drugs, SD standard deviation, SpA spondyloarthritis, TNFα tumor necrosis factor α Table 2 Status and change scores of mSASSS and extended mSASSS All patients (n = 210) Radiographic axial SpA (n = 115) Non-radiographic axial SpA (n = 95) mSASSS Extended mSASSS mSASSS Extended mSASSS mSASSS Extended mSASSS Status score at baseline, 4.2 ± 8.3 8.6 ± 18.0 5.9 ± 10.3 12.8 ± 22.5 2.3 ± 4.2 3.5 ± 7.4 mean ± SD Status score at year 2, 5.0 ± 9.6 9.8 ± 20.0 6.8 ± 11.7 14.5 ± 24.9 2.8 ± 5.3 4.1 ± 8.9 mean ± SD Change score, mean ± SD 0.7 ± 2.3 1.2 ± 3.7 1.0 ± 2.8 1.7 ± 4.6 0.5 ± 1.6 0.6 ± 2.2 Smallest detectable change 4.1 6.4 4.7 7.5 3.3 4.6 Change score ≥ smallest 14 (6.7) 12 (5.7) 7 (6.1) 8 (7.0) 4 (4.2) 6 (6.3) detectable change, n (%) Presence of syndesmophytes 48 (22.9) 64 (30.5) 35 (30.4) 48 (41.7) 13 (13.7) 16 (16.8) at baseline, n (%) Presence of syndesmophytes 50 (23.8) 65 (31.0) 36 (31.3) 50 (43.5) 14 (14.7) 15 (15.8) at year 2, n (%) Total number of syndesmophytes 0.9 ± 2.6 1.9 ± 5.2 1.5 ± 3.3 3.2 ± 6.6 0.3 ± 1.1 0.5 ± 1.9 at baseline, mean ± SD Total number of syndesmophytes 1.1 ± 3.1 2.2 ± 6.0 1.7 ± 3.8 3.5 ± 7.6 0.4 ± 1.4 0.6 ± 2.4 at year 2, mean ± SD Change score ≥ 2 points after 30 (14.3) 44 (21.0) 23 (20.0) 35 (30.4) 7 (7.4) 9 (9.5) 2 years, n (%) Development of new 16 (7.6) 20 (9.5) 13 (11.3) 17 (14.8) 3 (3.2) 3 (3.2) syndesmophytes after 2 years, n (%) Development of new /growth 18 (8.6) 23 (11.0) 15 (13.0) 20 (17.4) 3 (3.2) 3 (3.2) of existing syndesmophytes after 2 years, n (%) mSASSS modified Stoke Ankylosing Spondylitis Spine Score, SD standard deviation Llop et al. Arthritis Research & Therapy (2019) 21:126 Page 5 of 9 Table 3 Status and change scores of the components of the The Bland and Altman plots illustrating the inter- mSASSS and the extended mSASSS reader reliability of the extended and conventional Cervical spine Lumbar spine Lumbar scores are presented in Fig. 1. The SDC was higher for lateral view lateral view spine the extended mSASSS than for the conventional one AP view (Table 2). Regarding the presence of syndesmophytes in Change score ≥ 2 points 29 (13.8) 18 (8.6) 25 (11.9) after 2 years, n (%) the opinion of both readers at baseline, extended mSASSS detected syndesmophytes in 64 patients (30.5%) and con- Total number of 0.5 ± 1.6 0.4 ± 1.5 1.0 ± 3.0 syndesmophytes at ventional mSASSS in 48 patients (22.9%); the mean total baseline, mean ± SD number of syndesmophytes at baseline was 1.92 ± 5.24 Total number of 0.6 ± 1.7 0.6 ± 1.8 1.1 ± 3.3 and 0.94 ± 2.61 for the extended mSASSS and conven- syndesmophytes at tional mSASSS, respectively. Importantly, the extended year 2, mean ± SD mSASSS detected new syndesmophytes in 4 additional Development of new 8 (3.8) 8 (3.8) 9 (4.4) patients, new syndesmophytes or progression of exist- syndesmophytes after 2 years, n (%) ing syndesmophytes in 5 additional patients, and pro- gression by ≥ 2 units in the total score in 14 additional Development of 8 (3.8) 10 (4.8) 10 (4.8) new/growth of existing patients meaning a 25%, 28%, and 46% increase in the syndesmophytes after proportion of patients with progression according to the 2 years, n (%) respective definition as compared to the conventional mSASSS includes scores of the cervical and lumbar lateral views; extended mSASSS (Table 2). mSASSS includes additionally the lumbar spine AP view AP anteroposterior, mSASSS modified Stoke Ankylosing Spondylitis Spine When the two subgroups of patients (r-axSpA and Score, SD standard deviation nr-axSpA) were analyzed separately, the mean ± SD score at baseline as well as progression rates over 2 years were lower in the nr-axSpA group for both scoring systems (Table 2). kappa of 0.39 (95% CI 0.28, 0.57) and of 0.43 (95% CI 0.28, 0.57) for extended and conventional mSASSS, Truth respectively. The agreement was moderate also for the In the univariable analysis, both extended and conven- outcome development of new syndesmophytes or tional mSASSS were significantly associated with BASMI growth of existing syndesmophytes with Cohen’s kappa and BASFI (Table 7). In the multivariable analysis that of 0.44 (95% CI 0.29, 0.58) and 0.42 (95% CI 0.29, 0.55) was adjusted for disease activity parameters (BASDAI for the extended and conventional mSASSS, respectively. and CRP), the association of both scores with spinal An analysis of the score components revealed a com- mobility/function remained significant (Table 7). parable variability of the scores obtained in the cervical spine and in both planes of the lumbar spine (Table 4). Discussion The variance component analysis revealed that the In the present study, we evaluated the added value of patient-related variance contributed most to the total incorporation of AP lumbar radiographs in the mSASSS variance of both mSASSS and extended mSASSS score scoring system for the assessment of structural damage (Table 5); the same finding was observed when the in the spine in patients with axSpA. The extended variance of score components was analyzed separately mSASSS showed a good reliability and provided some (Table 6). improvement of the detection of radiographic spinal Table 4 Inter-observer reliability of the components of the mSASSS and the extended mSASSS Cervical spine lateral view Lumbar spine lateral view Lumbar spine AP view ICC for the status score at baseline 0.91 (0.88, 0.93) 0.91 (0.88, 0.93) 0.86 (0.83, 0.90) ICC for the status score at year 2 0.86 (0.82, 0.89) 0.92 (0.90, 0.94) 0.88 (0.84, 0.90) ICC for the change score 0.49 (0.39, 0.59) 0.43 (0.32, 0.54) 0.30 (0.19, 0.44) Kappa for the change score ≥ 2 points 0.13 (− 0.04, 0.29) 0.38 (0.18, 0.57) 0.36 (0.18, 0.54) after 2 years Kappa for the development of new 0.38 (0.20, 0.55) 0.59 (0.42, 0.77) 0.40 (0.24, 0.57) syndesmophytes after 2 years Kappa for the development of new/growth 0.34 (0.17, 0.52) 0.62 (0.46, 0.79) 0.48 (0.32, 0.64) of existing syndesmophytes after 2 years Data are presented as intraclass correlation coefficient (ICC) and Cohen’s kappa coefficient to measure the inter-observer reliability for continuous and categorical variables, respectively, with a 95% CI AP anteroposterior, mSASSS modified Stoke Ankylosing Spondylitis Spine Score, CI confidence interval, ICC intraclass correlation coefficient Llop et al. Arthritis Research & Therapy (2019) 21:126 Page 6 of 9 Table 5 Inter-observer reliability of the mSASSS and the extended mSASSS, expressed in variance components Status score at baseline Status score at year 2 Change score Residual Observer Patient Residual Observer Patient Residual Observer Patient variance variance variance variance variance variance variance variance variance component component component component component component component component component mSASSS 7.45 0.03 92.52 7.95 0.05 92.00 57.42 0.00 42.58 Extended 7.28 0.01 92.71 6.67 0.00 93.33 54.99 0.00 45.01 mSASSS mSASSS modified Stoke Ankylosing Spondylitis Spine Score progression in patients with axSpA as compared to the in more advanced disease the added value of the lum- conventional mSASSS that has to be weighed against bar AP assessment might be greater than in early additional costs, additional radiation exposure (although disease [15]. In our analysis, the subgroup of patients AP lumbar radiographs are often performed routinely), with r-axSpA had a higher level of structural damage and additional time investment associated with reading at baseline and higher progression rates after 2 years of additional 24 vertebral corners. compared to nr-axSpA patients. At the same time, in Extended mSASSS can be considered as supplemen- advanced disease with extended ankylosis, one can tary to the conventional mSASSS as it adds information expect a larger extent of structural damage that from the AP lumbar spine view. The information ob- would be captured on both lateral and AP views that tained from the AP lumbar radiograph can also be to might result into a reduction of the added value of some extent redundant, since the same structural the AP lumbar radiographs. damage (same syndesmophyte) can be visible on both The extended mSASSS—similarly to the conventional lateral and AP views. This fact explains, for instance, mSASSS—captures only a part of the structural damage a larger difference in status scores (the extended visible on radiographs. Syndesmophytes with posterior mSASSS was double as high as the conventional one) localization or in the lateral aspects of the cervical spine, than in percentages of patients with syndesmophytes as well as involvement of posterior structures (i.e., facet (30.5% with extended mSASSS vs. 22.9% with conven- joints), are not captured by any score. In the majority of tional mSASSS at baseline). cases, however, structural damage develops more or less The most relevant question is, however, if the extended simultaneously in different aspects of the spinal column mSASSS is able to detect patients with radiographic spinal that allows the assumption that the mSASSS/extended progression not captured by the conventional method. mSASSS is as a proxy for the overall structural damage Indeed, with the extended mSASSS, new syndesmophytes in the spine. In the univariable and multivariable regres- after 2 years were detected in 4 additional patients, new sion analyses, both scores showed a significant asso- syndesmophytes or growth of existing syndesmophytes in ciation with spinal mobility (BASMI) and function 5 additional patients, and progression by ≥ 2 points in the (BASFI) that confirms the construct validity of both total score in 14 additional patients that gave a 25%, 28%, scores as parameters truly reflecting structural damage and 46% increase in the proportion of patients with in the spine in axSpA. A potentially important limitation progression according to the respective definition as of any radiographs is the exclusion of the thoracic spine compared to the conventional score. These proportions from scoring due to anatomical overlap with other were even higher in patients with r-axSpA meaning that thoracic structures. However, active inflammatory [16] Table 6 Inter-observer reliability of the components of the mSASSS and the extended mSASSS, expressed in variance components Cervical spine lateral view Lumbar spine lateral view Lumbar spine AP view Residual Observer Patient Residual Observer Patient Residual Observer Patient variance variance variance variance variance variance variance variance variance component component component component component component component component component Status 8.87 0.06 91.07 9.05 0.00 90.95 13.23 0.22 86.55 score at baseline Status 13.71 0.19 86.10 7.49 0.00 92.51 12.14 0.13 87.73 score at year 2 Change 50.49 0.18 49.33 57.11 0.00 42.89 69.54 0.00 30.46 score AP anteroposterior, mSASSS modified Stoke Ankylosing Spondylitis Spine Score Llop et al. Arthritis Research & Therapy (2019) 21:126 Page 7 of 9 Fig. 1 Bland and Altman plots for mSASSS (a) and extended mSASSS (b) and structural changes [17] seem to occur here at least methods of radiographic spinal progression in axSpA in as frequent as in other spine regions. For this reason, the future [17]. another modification of the mSASSS was proposed by Baraliakos et al: the Radiographic AS Spinal Score (RASSS), which adds to the mSASSS the score of the Conclusions lower thoracic vertebrae (T10–T12), under the hypo- The incorporation of the AP radiographs of the lumbar thesis that most of progression is found in these spine in the assessment of structural damage in the segments, and the quantification of new bone formation spine provides improvement of detection of radiographic is superior to the conventional mSASSS [18]. Never- spinal progression in axSpA. theless, a recent study that compared the mSASSS with the RASSS [19] concluded the limitation of the RASSS Abbreviations in the feasibility aspect (in 64% of radiographs of the AP: Anteroposterior; AS: Ankylosing spondylitis; axSpA: Axial spondyloarthritis; BASDAI: Bath Ankylosing Spondylitis Disease Activity Index; BASFI: Bath lumbar spine, the lower thoracic vertebrae were not vis- Ankylosing Spondylitis Functional Index; BASMI: Bath Ankylosing Spondylitis ible, and therefore, the RASSS could not be calculated). Metrology Index; BASRI: Bath Ankylosing Spondylitis Radiology Index; Due to low reliability, no thoracic spine radiographs and CI: Confidence interval; CRP: C-reactive protein; CT: Computed tomography; GESPIC: GErman SPondyloarthritis Inception Cohort; ICCs: Intraclass no AP cervical radiographs were collected in GESPIC correlation coefficients; mNY: Modified New York; mSASSS: Modified for the assessment of structural damage. Stoke Ankylosing Spondylitis Spine Score; nr-axSpA: Non-radiographic Other imaging methods without technical limitations axial spondyloarthritis; OMERACT: Outcome Measures in Rheumatology; r-axSpA: Radiographic axial spondyloarthritis; SD: Standard deviation; of plain radiographs—such as low-dose computed tom- SDC: Smallest detectable change; SEM: Standard error of measurement; ography (CT)—have a potential to become the primary TNFi: Tumor necrosis factor inhibitor Table 7 Association between mSASSS or extended mSASSS score with BASFI or BASMI at baseline in the linear regression analysis Univariable analysis Multivariable model 1* Multivariable model 2* β (95% CI) β (95% CI) β (95% CI) Outcome: BASFI mSASSS 0.046 (0.008, 0.083) 0.033 (0.010, 0.056) – Extended mSASSS 0.021 (0.003, 0.038) – 0.014 (0.004, 0.025) Outcome: BASMI mSASSS 0.093 (0.068, 0.117) 0.082 (0.059, 0.106) – Extended mSASSS 0.045 (0.034, 0.056) – 0.040 (0.029, 0.050) *Adjusted for BASDAI and CRP at baseline BASDAI Bath Ankylosing Spondylitis Disease Activity Index, BASFI the Bath Ankylosing Spondylitis Functional Index, BASMI Bath Ankylosing Spondylitis Metrology Index, mSASSS modified Stoke Ankylosing Spondylitis Spine Score Llop et al. Arthritis Research & Therapy (2019) 21:126 Page 8 of 9 Acknowledgements Author details 1 We thank Prof M Leirisalo-Repo, Finland, Prof D van der Heijde, The Netherlands, Department of Gastroenterology, Infectious Diseases, and Rheumatology, and Prof M Dougados, France, for scientific advice on the design of the cohort. Charité - Universitätsmedizin Berlin, Hindenburgdamm 30, 12203 Berlin, We are grateful to Beate Buss and Petra Tietz for monitoring the cohort; Germany. 2Parc Taulí Hospital Universitari, I3PT, Universitat Autònoma de Johanna Callhoff, Anja Weiss, and Martina Niewerth for the data management Barcelona, Sabadell, Barcelona, Spain. 3Epidemiology Unit, German support; Janis Vahldiek and Georg Heine for handling the X-ray images and for Rheumatism Research Centre, Berlin, Germany. 4Klinikum Bielefeld the development of the image scoring interface; and to all patients who Rosenhöhe, Bielefeld, Germany. voluntarily participated in this cohort. Further, we would like to thank the following rheumatologists for inclusion of their patients: J Brandt, H Brandt, G-R Received: 7 January 2019 Accepted: 9 May 2019 Burmester, H Deister, E Edelmann, J Emmerich, M Enderlein, A Gauliard, E Gromnica-Ihle, F Heldmann, S Hermann, U von Hinüber, Ü Hübner, K Karberg, H Nüßlein, R Pelle-Lohfink, D Pick, G Reichmuth, E Riehers, M Rihl, R Schmidt, S Schnarr, U Schneider, I-H Song, I Spiller, U Syrbe, V Walz, S Wassenberg, H M References Wisseler, and S Zinke. 1. van der Linden S, Valkenburg HA, Cats A. Evaluation of diagnostic criteria for ankylosing spondylitis. A proposal for modification of the New York criteria. Arthritis Rheum. 1984;27(4):361–8. Funding 2. Machado P, Landewe R, Braun J, Hermann KG, Baker D, van der Heijde D. As part of the German Competence Network in Rheumatology (Kompetenznetz Both structural damage and inflammation of the spine contribute to Rheuma), GESPIC has been financially supported by the German Federal impairment of spinal mobility in patients with ankylosing spondylitis. Ann Ministry of Education and Research (Bundesministerium für Bildung und Rheum Dis. 2010;69(8):1465–70. Forschung—BMBF). As funding by BMBF was reduced according to schedule in 3. Landewe R, Dougados M, Mielants H, van der Tempel H, van der Heijde D. 2005 and stopped in 2007, complementary financial support has been obtained Physical function in ankylosing spondylitis is independently determined by also from Abbott/AbbVie, Amgen, Centocor, Schering–Plough, and Wyeth. both disease activity and radiographic damage of the spine. Ann Rheum Since 2010, GESPIC is supported by AbbVie. Dis. 2009;68(6):863–7. Additional support has been obtained also from ArthroMark (grants number 4. Poddubnyy D, Listing J, Haibel H, Knuppel S, Rudwaleit M, Sieper J. FKZ 01EC1401A) and METARTHROS (grant number FKZ 01EC1407A) projects Functional relevance of radiographic spinal progression in axial funded by BMBF. spondyloarthritis: results from the GErman SPondyloarthritis Inception The work of Maria Llop Vilaltella was supported by EULAR Scientific Training Cohort. Rheumatology (Oxford). 2018;57(4):703–11. Bursary and by the FER Institution (Fundación Española de Reumatología). 5. Creemers MC, Franssen MJ, van't Hof MA, Gribnau FW, van de Putte LB, van Riel PL. Assessment of outcome in ankylosing spondylitis: an extended Availability of data and materials radiographic scoring system. Ann Rheum Dis. 2005;64(1):127–9. For information of the availability of the data included in the GESPIC 6. van der Heijde D, Landewe R. Selection of a method for scoring database, contact Prof. Dr. Poddubnyy. radiographs for ankylosing spondylitis clinical trials, by the assessment in ankylosing spondylitis working group and OMERACT. J Rheumatol. Declarations 2005;32(10):2048–9. We confirm that the article has not been published before and is not under 7. MacKay K, Mack C, Brophy S, Calin A. The Bath ankylosing spondylitis consideration for publication elsewhere, and that it has been approved radiology index (BASRI): a new, validated approach to disease assessment. by all authors. Arthritis Rheum. 1998;41(12):2263–70. 8. Averns HL, Oxtoby J, Taylor HG, Jones PW, Dziedzic K, Dawes PT. Authors’ contributions Radiological outcome in ankylosing spondylitis: use of the Stoke Ankylosing ML, VR, and IR contributed to the data analysis and interpretation and Spondylitis Spine Score (SASSS). Br J Rheumatol. 1996;35(4):373–6. drafting of the manuscript. JS and MR contributed to the initiation of GESPIC, 9. Wanders AJ, Landewe RB, Spoorenberg A, Dougados M, van der Linden S, data collection, data interpretation, and drafting of the manuscript. HH Mielants H, van der Tempel H, van der Heijde DM. What is the most contributed to the data collection, data interpretation, and drafting of the appropriate radiologic scoring method for ankylosing spondylitis? A manuscript. DP contributed to the data collection, analysis and interpretation, and comparison of the available methods based on the Outcome Measures in drafting of the manuscript. All authors read and approved the final manuscript. Rheumatology clinical trials filter. Arthritis Rheum. 2004;50(8):2622–32. 10. Poddubnyy D, Rudwaleit M, Haibel H, Listing J, Marker-Hermann E, Zeidler Ethics approval and consent to participate H, Braun J, Sieper J. Rates and predictors of radiographic sacroiliitis All patients signed an informed consent to participate in GESPIC, and the progression over 2 years in patients with axial spondyloarthritis. Ann Rheum study was approved by the central ethics committee. Dis. 2011;70(8):1369–74. 11. Rudwaleit M, Haibel H, Baraliakos X, Listing J, Marker-Hermann E, Zeidler Consent for publication H, Braun J, Sieper J. The early disease stage in axial spondylarthritis: Not applicable. results from the german spondyloarthritis inception cohort. Arthritis Rheum. 2009;60(3):717–27. Competing interests 12. Poddubnyy D, Haibel H, Listing J, Marker-Hermann E, Zeidler H, Braun J, ML: none declared Sieper J, Rudwaleit M. Baseline radiographic damage, elevated acute- VRR: consulting fees, speaking fees, and/or honoraria from AbbVie, MSD, and phase reactant levels, and cigarette smoking status predict spinal Novartis radiographic progression in early axial spondylarthritis. Arthritis Rheum. IR: none declared 2012;64(5):1388–98. JS: consulting fees, speaking fees, and/or honoraria from AbbVie, Janssen, 13. Boers M, Brooks P, Strand CV, Tugwell P. The OMERACT filter for Outcome Lilly, MSD, Novartis, Pfizer, Roche, and UCB Measures in Rheumatology. J Rheumatol. 1998;25(2):198–9. HH: consulting fees, speaking fees, and/or honoraria from AbbVie, Janssen, 14. Bruynesteyn K, Boers M, Kostense P, van der Linden S, van der Heijde MSD, and Novartis D. Deciding on progression of joint damage in paired films of MR: consulting fees, speaking fees, and/or honoraria AbbVie, BMS, Celgene, individual patients: smallest detectable difference or change. Ann Chugai, Eli Lilly, Janssen, MSD, Novartis, Pfizer, UCB Rheum Dis. 2005;64(2):179–82. DP: consulting fees, speaking fees, and/or honoraria from AbbVie, BMS, Lilly, 15. Baraliakos X, Listing J, Rudwaleit M, Haibel H, Brandt J, Sieper J, Braun MSD, Novartis, Pfizer, Roche, and UCB J. Progression of radiographic damage in patients with ankylosing spondylitis: defining the central role of syndesmophytes. Ann Rheum Dis. 2007;66(7):910–5. Publisher’s Note 16. Baraliakos X, Landewe R, Hermann KG, Listing J, Golder W, Brandt J, Springer Nature remains neutral with regard to jurisdictional claims in published Rudwaleit M, Bollow M, Sieper J, van der Heijde D, et al. Inflammation in maps and institutional affiliations. ankylosing spondylitis: a systematic description of the extent and frequency Llop et al. Arthritis Research & Therapy (2019) 21:126 Page 9 of 9 of acute spinal changes using magnetic resonance imaging. Ann Rheum Dis. 2005;64(5):730–4. 17. de Koning A, de Bruin F, van den Berg R, Ramiro S, Baraliakos X, Braun J, van Gaalen FA, Reijnierse M, van der Heijde D. Low-dose CT detects more progression of bone formation in comparison to conventional radiography in patients with ankylosing spondylitis: results from the SIAS cohort. Ann Rheum Dis. 2018;77(2):293–9. 18. Baraliakos X, Listing J, Rudwaleit M, Sieper J, Braun J. Development of a radiographic scoring tool for ankylosing spondylitis only based on bone formation: addition of the thoracic spine improves sensitivity to change. Arthritis Rheum. 2009;61(6):764–71. 19. Ramiro S, van Tubergen A, Stolwijk C, Landewe R, van de Bosch F, Dougados M, van der Heijde D. Scoring radiographic progression in ankylosing spondylitis: should we use the modified Stoke Ankylosing Spondylitis Spine Score (mSASSS) or the Radiographic Ankylosing Spondylitis Spinal Score (RASSS)? Arthritis Res Ther. 2013;15(1):R14.