Llop et al. Arthritis Research & Therapy (2022) 24:30 https://doi.org/10.1186/s13075-021-02695-5 RESEARCH ARTICLE Open Access Sustained low disease activity measured by ASDAS slow radiographic spinal progression in axial spondyloarthritis patients treated with TNF-inhibitors: data from REGISPONSERBIO Maria Llop1, Mireia Moreno1, Victoria Navarro‑Compán2, Xavier Juanola3, Eugenio de Miguel2, Raquel Almodóvar4, Eduardo Cuende Quintana5, Jesús Sanz Sanz6, Emma Beltrán7, M. Dolores Ruiz Montesinos8, Joan Calvet1, Antoni Berenguer‑Llergo9, Jordi Gratacós1* and the Regisponserbio group Abstract Background: To evaluate the influence of the disease activity on radiographic progression in axial spondyloarthritis (axSpA) patients treated with TNF inhibitors (TNFi). Methods: The study included 101 axSpA patients from the Spanish Register of Biological Therapy in Spondyloarthri‑ tides (REGISPONSERBIO), which had clinical data and radiographic assessment available. Patients were classified into 2 groups based on the duration of TNFi treatment at baseline: (i) long-term treatment (≥4 years) and (ii) no long-term treatment (< 4 years). Radiographs were scored by two readers according to the modified Stoke Ankylosing Spondy‑ litis Spine Score (mSASSS) with known chronology. Disease activity differences between patients’ groups at each time point were assessed using a linear mixed-effect model. Results: Radiographic progression was defined as an increase in ≥2 mSASSS units. At inclusion, approximately half of the patients (45.5%) were receiving long-term treatment with TNFi (≥4 years). In this group of subjects, a significant difference in averaged Ankylosing Spondylitis disease Activity Score (ASDAS) across follow-up was found between progressors and non-progressors (2.33 vs 1.76, p=0.027, respectively). In patients not under long-term TNFi treatment (54.5%) though, no significant ASDAS differences were observed between progressors and non-progressors until the third year of follow-up. Furthermore, no significant differences were found in progression status, when disease activity was measured by Bath Ankylosing spondylitis Disease Activity Index (BASDAI) and C reactive protein (CRP). Conclusions: Patients on long-term TNFi treatment with a mean sustained low disease activity measures by ASDAS presented lower radiographic progression than those with active disease. Keywords: Spondyloarthritis, Radiology, Biological therapies, Inflammation, Outcome measures Introduction Axial spondyloarthritis (axSpA) is a chronic inflamma- tory disease with predominantly axial symptoms, such as *Correspondence: jgratacosmas@gmail.com inflammation of the sacroiliac joints and spine, leading 1 Rheumatology, Parc Taulí Hospital Universitari, I3PT Research Institute to structural damage [1]. AxSpA can be classified as into (UAB), Universitat Autónoma de Barcelona (UAB), Sabadell 08208, Spain Full list of author information is available at the end of the article radiographic SpA (r-axSpA, also known as ankylosing © The Author(s) 2022. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/. The Creative Commons Public Domain Dedication waiver (http://​creat​iveco​ mmons.​org/​publi​cdoma​in/​zero/1.​0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Llop et al. Arthritis Research & Therapy (2022) 24:30 Page 2 of 9 spondylitis (AS)) or non-radiographic SpA (nr-axSpA) an inactive disease status (AS Disease Activity Score depending on the presence or absence of definite radio- (ASDAS) ≤1.3) when receiving TNFi. Currently, the graphic sacroiliitis respectively, according to the modified ASAS-EULAR management recommendations for New York criteria grading system [2]. axSpA are based on a treat-to-target approach [10]. How- Inflammation, when persists, leads to structural dam- ever, in clinical practice, it might be quite challenging to age, which can be detected on conventional radiograph achieve inactive disease in axSpA patients and low dis- as sclerosis, erosions, and new bone formation (syn- ease activity seem to be a more realistic target [11]. desmophytes). A major concern in the SpA field is the This study aimed to evaluate the influence of low dis- development of new bone formation in the spine due to ease activity measured by ASDAS (< 2.1) on radiographic its contribution to disease severity [3]. While the clini- progression in axSpA patients treated with TNFi, using cal efficacy of TNF inhibitors (TNFi) in axSpA has been the data from REGISPONSERBIO. widely shown to decrease symptoms and signs of the disease in randomized clinical trials, it is still unclear Methods whether TNFi inhibits radiographic progression. Accord- Patients and study design ing to the early data published from the pivotal studies, The study population comprised all patients included in TNFi did not seem to inhibit radiographic damage in REGISPONSERBIO (Spanish Registry of axSpA under axSpA [4–6]. Nevertheless, some subsequent studies sug- TNFi therapy) who presented at least 2 sets of cervical gested that long-term treatment with TNFi (more than and lumbar spinal radiographs, one at inclusion and one 4 years) could slow down radiographic progression in after a minimum interval of 2 years. A detailed descrip- axSpA [7, 8]. However, more data is needed to draw fur- tion of the entire cohort has been reported previously ther conclusions. [12]. Briefly, REGISPONSERBIO is a prospective mul- More recently, the Swiss Clinical Quality Manage- ticenter cohort design study that includes both patients ment cohort [9] reported their results in patients with who were already receiving TNFi treatment and those r-axSpA treated with TNFi who underwent radiographic who started on TNFi therapy at the time of recruitment assessments every 2 years during a 10-year period. The (Fig. 1). Patients in the registry were followed up for a study showed that radiographic spinal progression was 3-year period in which they underwent clinical and labo- nearly entirely inhibited in the following 2-year radio- ratory controls every 6 months. Criteria for starting TNFi graphic interval in patients with r-axSpA who achieved therapy was made according to the recommendations of Fig. 1 Flow chart describing the inclusion of REGISPONSERBIO patients in our study Llop et al. Arthritis Research & Therapy (2022) 24:30 Page 3 of 9 the Spanish Society of Rheumatology [13]. The register high disease activity (ASDAS >3.5). Active disease was and subsidiary efficacy and safety analyses complied with defined as ASDAS ≥2.1 or, alternatively, BASDAI ≥4. the Declaration of Helsinki and were approved by the Ethical Review Boards of all participating hospitals, and Statistical analysis patients signed an informed consent form to be included. Concordance in radiographic assessment between both For this study, patients were classified into 2 groups readers was explored by calculating the intraclass cor- based on the duration of TNFi treatment at recruit- relation coefficient (ICC, two-way mixed, single score ment: (i) long-term duration (patients receiving TNFi for method) and Cohen’s kappa coefficient for continuous at least 4 years) and (ii) no long-term duration (patients and categorical variables, respectively. Additionally, a receiving TNFi for less than 4 years and patients who Bland-Altman plot was drawn, and the smallest detect- started TNFi at study entry). This cut-off was selected able change (SDC) was determined to assess the sensitiv- according to previously published evidence for the poten- ity to change [19]. tial influence of TNFi on radiographic progression in For description purposes, continuous variables were axSpA from various studies [7–9]. described by their median and minimum and maximum values, while absolute and relative frequencies were used Assessment of radiographic progression for categorical parameters. Differences between included Spinal radiographs were collected centrally in the rheu- and excluded patients from the RESPONGISERBIO matology department of Hospital Parc Taulí. The images cohort were evaluated regarding variables potentially were anonymized and scored in random order by 2 inde- linked to radiographic progression, using the Mann- pendent and trained readers (ML and MM), who were Whitney and Fisher tests for continuous and categori- blinded to all clinical data but aware of the chronologi- cal variables, respectively. To assess differences between cal order. The images were scored according to the modi- patient groups, ASDAS, BASDAI, and CRP were fitted fied Stoke Ankylosing Spondylitis Spinal Score (mSASSS) to linear mixed-effects models that included a random [14]. According to the mSASSS, anterior corners of the intercept in order to model the patient’s effect, and a vertebral bodies from lower C2 to upper T1 (cervical fixed effect coding the time point for which the measure- spine) and from lower T12 to upper S1 (lumbar spine) ment had been taken. Patient group was included in the were scored as follows: 0 = normal; 1 = erosion, sclero- models as an explanatory variable, which described the sis, and/or squaring; 2 = non-bridging syndesmophyte; individuals’ characteristics according to their treatment and 3 = bridging syndesmophyte, giving a range for status and their radiographic progression in four subsets: the entire score from 0 to 72. No adjudication was per- (i) patients not under long-term TNFi treatment without formed. The final mSASSS score included in the analysis radiographic progression, (ii) patients not under long- was calculated as a mean of the mSASSS scores of both term TNFi treatment with radiographic progression, readers. Syndesmophytes were considered present if both (iii) patients under long-term TNFi treatment without readers assigned a score of ≥2 to a vertebral corner (VC). radiographic progression, and (iv) patients under long- A change of the mSASSS score by ≥2 points was taken as term TNFi treatment with radiographic progression. the definition of progression [9, 15]. This model parametrization is mathematically equiva- Only scores from radiographs with ≤ 3 missing VC per lent to include the treatment status and the radiographic cervical or lumbar segment were used. Missing scores of progression as covariates, together with the interaction single VCs were substituted with scores for the respective of these two parameters. As potential confounding fac- vertebral corners at other time point, if available, or with tors, we considered the length of the radiographic inter- a score of zero if scores for both time points were missing val and the following patients’ characteristics collected [16, 17]. at the time of recruitment: gender, age, smoking habit, radiographic sacroiliitis, mSASSS average, presence of Disease activity syndesmophytes, HLAB27, and time from onset of symp- Disease activity measurements included the Bath Anky- toms. These variables were included in the models as losing Spondylitis Disease Activity Index (BASDAI), fixed effects to obtain adjusted estimates of disease activ- ASDAS-CRP, and C reactive protein (CRP) serum levels. ity differences and group means. Binarized versions of These were collected at start of TNFi, at start of the study these disease activity measurements were analyzed simi- (recruitment) and subsequently every 6 months over a larly, using generalized linear mixed-effects models with follow-up period of 3 years. Disease activity states were a binomial distribution (logistic regression). The associa- assigned using the ASDAS cut-off levels [18]: inactive tion between mSASSS and ASDAS increases relative to disease (ASDAS <1.3), low disease activity (1.3 ≤ASDAS baseline values was also assessed using a standard linear <2.1), high disease activity (2.1 ≤ASDAS ≤3.5), and very model (no random effects included). When necessary, a Llop et al. Arthritis Research & Therapy (2022) 24:30 Page 4 of 9 Tukey’s transformation ladder was applied to continuous total mSASSS score during follow-up was observed in 30 variables in order to fulfill the assumptions of the linear patients (29.7%). model (lambda parameters: 0.5 for time from symptom- Among patients receiving long-term treatment, sig- atology onset; 0.25 for CRP; 0 for MSASSS increase at nificant mean ASDAS differences were found at baseline 3 years follow-up compared to baseline). For analyses (i.e., at time of inclusion) between progressors and non- comparing the patients groups of interest, mean differ- progressors (mean difference = 0.63; p-value = 0.046) ences, adjusted means (for continuous outcomes), odds (Fig. 2). The magnitude of this difference, while not sta- ratio (OR, for binary outcomes), and their corresponding tistically significant, was preserved after adjusting for standard errors were used to measure the magnitude of clinical potential confounders (mean difference of 0.61, effects. As long-term TNFi-treated patients kept essen- p-value = 0.058). Importantly, these patients maintained tially the same disease activity levels during the follow- essentially the same baseline ASDAS levels during the up, differences across follow-up in these patients were 3-year follow-up with an adjusted mean difference rang- assessed by averaging time point within progression ing from −0.14 to 0.16 (Fig. 2). Based on this observation, groups. Adjusted means were computed using our own we performed further analyses in which ASDAS values dataset as reference population. Statistical significance were averaged across all time points (Table 2). In this was assessed using Wald tests. All analyses were carried comparison, averaged ASDAS was significantly higher in out using R [20]. progressors compared to non-progressors mean differ- ence = 0.56, p-value = 0.027), even after controlling for potential confounders (mean difference = 0.57, p-value = Results 0.031, Table 2). Importantly, averaged ASDAS remained A total of 256 patients were included in REGISPONSER- above the 2.1 threshold (2.33) that indicates high disease BIO, of whom 101 with axSpA were included in the anal- activity in patients who progressed. In contrast, this esti- ysis based on the availability of full sets of radiographs mation was within the low activity range (1.76) for non- (lumbar and cervical lateral view) and clinical data. Of progressors (Supplementary Figure S2). these patients, 46 (45.5%) were classified in the long-term In the no long-term TNFi treatment group (<4 years), TNFi group (≥4 years) and 55 (54.5%) in the not long- similar ASDAS differences at inclusion were seen in pro- term group (<4 years). Radiographic progression was gressors and non-progressors (mean difference = 0.52; observed in 32.6% of the long-term TNFi patients and a p-value = 0.093) (Fig. 2). As expected, ASDAS scores in 27.3% for not long-term treatment (15 patients in each this group were higher than those observed for patients group). Demographic, clinical, and radiographic data of on long-term treatment, irrespective of their progres- the patients at the time of the first radiograph are shown sion status (mean difference = −0.52; p-value = 0.018). in Table 1 and Supplementary Tables S1 and S2 and Nonetheless, these patients experienced a substantial were similar to those showed by the REGISPONSERBIO decrease in their ASDAS scores after TNFi treatment patients not included in the study (Supplementary Tables was prescribed at recruitment, observed as early as after S3 and S4), mostly differing in the duration of TNFi treat- 6 months of follow-up (mean decrease of −0.53, p-value ments. This time point coincided with the onset of TNFi <0.001 in non-progressors; mean decrease of −1.10, in 18 (18%) patients. At inclusion, 52.1% patients had p-value <0.001 in progressors). At the end of follow-up, low disease activity status (ASDAS <2.1), and only 24% this decrease in disease activity induced by TNFi therapy of these had inactive disease (Table 1). Association with was conserved, or even accentuated in patients without radiographic progression was explored specifically in radiographic progression (additional mean decrease of each patient group defined by their treatment status at 0.41 during follow-up compared to the first time point, recruitment (TNFi ≥4 years or TNFi <4 years). p-value = 0.015). On the other hand, patients with radi- Agreement between both readers was “excellent” for ographic progression did not maintain this substantial mSASSS status score (ICC 0.99 and 0.98, at inclusion and decrease in ASDAS during follow-up (mean increase follow-up, respectively) and “moderate” for the change of 0.44 compared to first time point, p-value = 0.083). of the mSASSS score by ≥2 point (Cohen’s kappa 0.53). These differences in ASDAS profiles throughout follow- The Bland and Altman plots illustrating inter-reader reli- up were evident after controlling for clinical variables ability are presented in Supplementary Figure S1. More potentially linked to disease progress (additional mean inter-reader reliability data is available in Supplementary decrease of 0.30, p-value = 0.074 in the non-progressors Table S5. The mean (SD) time between radiographs was group; mean increase of 0.41, p-value 0.100 in the pro- 3.5 (1.0) years, the change score between inclusion and gressors group, compared to the first time point). follow-up was 1.98 (0.83), and the SDC of progression Nevertheless, it was not until the third year of fol- was 2.26 mSASSS units. Progression by ≥ 2 units in the low-up that significant ASDAS differences were found Llop et al. Arthritis Research & Therapy (2022) 24:30 Page 5 of 9 Table 1 Baseline characteristics and radiographic progression of the 101 patients included in the study by baseline TFNi treatment status. N TNFi treatment ≥ 4years TNFi treatment < 4years All n=46 (45.5%) n=55 (54.5%) n=101 Age, years 101 46.0 (25.0, 75.0) 47.0 (21.0, 75.0) 46.0 (21.0, 75.0) Female, sex % 101 8 (17.4%) 11 (20.0%) 19 (18.8%) Radiographic progression (mSASSS ≥ 2) 101 15 (32.6%) 15 (27.3%) 30 (29.7%) HLA-B27 positive, % 99 43 (95.6%) 43 (79.6%) 86 (86.9%) AS, % 101 41 (89.1%) 45 (81.8%) 86 (85.1%) BMI 95 26.7 (19.8, 40.8) 25.4 (19.4, 32.0) 26.0 (19.4, 40.8) Current smokers, % 101 16 (34.8%) 15 (27.3%) 31 (30.7%) Symptom duration, years 97 19.0 (2.0, 52.0) 11.0 (0.0, 50.0) 15.0 (0.0, 52.0) CRP (mg/L) 94 2.9 (0.10, 26.0) 4.3 (0.0, 88.7) 3.4 (0.0, 88.70) CRP < 5 94 27 (62.8%) 26 (51.0%) 53 (56.4%) BASDAI (0-10) 97 3.0 (0.2, 8.6) 4.2 (0.40, 8.8) 3.2 (0.20, 8.8) BASDAI < 4 97 30 (66.7%) 24 (46.2%) 54 (55.7%) ASDAS-CRP 96 1.71 (0.2, 5.1) 2.3 (0.3, 5.1) 2.1 (0.2, 5.1) Low disease (ASDAS < 2.1), % 96 28 (63.6%) 22 (42.3%) 50 (52.1%) Inactive disease (ASDAS <1.3), % 96 10 (22.7%) 13 (25.0%) 23 (24.0%) BASFI (0–10) 98 3.6 (0.0, 9.1) 4.2 (0.0, 9.3) 3.9 (0.0, 9.3) BASMI (0–10) 82 2.9 (0.5, 7.3) 3.0 (0.6, 6.4) 2.9 (0.5, 7.3) mSASSS (0–72) 101 10.0 (0.0, 72.0) 3.0 (0.0, 66.0) 5.0 (0.0, 72.0) Syndesmophytes present, % 101 26 (56.5%) 25 (45.5%) 51 (50.5%) On NSAID treatment, % 99 23 (51.1%) 34 (63.0%) 57 (57.6%) On TNFi treatment, % 100 45 (100.0%) 37 (67.3%) 82 (82.0%) Number of previous TNFi 0 101 24 (52.2%) 51 (92.7%) 75 (74.3%) 1 16 (34.8%) 4 (7.3%) 20 (19.8%) 2 6 (13.0%) 0 (0.0%) 6 (5.9%) Months of TNFi treatment in treated patients 82 84.0 (48.0, 132.0) 19.0 (1.0, 43.0) 51.0 (1.0, 132.0) Uveitis 101 10 (21.7%) 12 (21.8%) 22 (21.8%) Psoriasis 99 5 (11.1%) 2 (3.7%) 7 (7.1%) IBD 99 3 (6.7%) 4 (7.4%) 7 (7.1%) Continuous variables are described by their median and their minimum and maximum values (between brackets), while absolute and percentages are showed for categorical variables. HLA-B27, human leucocyte antigen B27; AS, ankylosing spondylitis; BMI, body mass index; CRP, C reactive protein; BASDAI, Bath Ankylosing Spondylitis Disease Activity Index; ASDAS-CRP, Ankylosing Spondylitis Disease Activity Score; BASFI, Bath Ankylosing Spondylitis Functional Index; BASMI, Bath Ankylosing Spondylitis Metrology Index; mSASSS, modified Stoke Ankylosing Spondylitis Spine Score; NSAID, non-steroidal anti-inflammatory drug; TNFi, tumor necrosis factor inhibitor; IBD, inflammatory bowel disease between progression status in patients not treated = 17.5, p-value = 0.057). Furthermore, an increase of in the long-term group, with progressors showing a 1 point in the ASDAS score at this time point (3 years) higher level of ASDAS compared to non-progressors at was associated with a 16% increase in the mSASSS, the end of follow-up (mean difference of 0.79, p-value both compared to values obtained at recruitment = 0.014). At this time point, only 15% (4 out of 27) (p-value = 0.021). with no radiographic progression showed a high level Although similar patterns were found when analyz- of disease activity (ASDAS ≥ 2.1), while this percent- ing disease activity with BASDAI score and CRP serum age rose to 50% (6 out of 12) in patients with radio- levels, we did not observe significant mean differences graphic progression (OR = 18.39, p-value = 0.037). between progressors and non-progressors in long-term Despite the small sample size available for this com- TNFi-treated patients (Table 2). Additionally, in con- parison, the above differences persisted after control- trast to ASDAS findings, in the third year of follow-up, ling statistically for potential confounders, although no significant differences were found between progres- they were not statistically significant at the selected sors and non-progressors in patients not under long- threshold (mean difference: 0.57, p-value = 0.093; OR term TNFi treatment (Supplementary Figure S3). Llop et al. Arthritis Research & Therapy (2022) 24:30 Page 6 of 9 Fig. 2 Association between disease activity measured by ASDAS-CRP every 6 months with radiographic progression. Radiographic progression was defined as an increase in the mSASSS score at ≥2 points in patients under TNFi treatment during a follow-up period of 3 years. Bold dots and intervals represent means and ± their corresponding standard errors, as provided by a mixed-effects linear model (no other covariates included) Table 2 Disease activity differences between patients under ASDAS is clearly associated with radiographic spinal long-term TNFi treatment with and without radiographic progression in patients on TNFi treatment. progression The duration of biologic treatment is an important Univariable analysis factor in assessing its effect on spinal progression in Progressors mean (SE) Non-progressors mean Difference axSpA. As previous studies showed a potential asso- (SE) ciation between the duration of TNFi treatment and ASDAS 2.33 (0.21) 1.76 (0.14) 0.56 risk of progression [7, 9, 21], with no TNFi effect being p = 0.027 observed within a follow-up period of 2 years [9, 22]. BASDAI 3.86 (0.42) 2.97 (0.29) 0.89 Moreover, the potential long-term effect of TNFi on p = 0.085 new bone formation is thought to be due to an effec- CRP 1.40 (0.09) 1.21 (0.06) 0.19 tive suppression of inflammation. In this regard, new p = 0.087 evidence suggests that inflammatory changes are fol- Multivariable model lowed by the replacement of subchondral bone marrow Progressors Non-progressors Difference (SE) (SE) by repairing tissue which then stimulates osteoblasts, ASDAS 2.38 (0.23) 1.81 (0.14) 0.57 leading to new bone formation [1]. Based on these p = 0.031 observations, the association between disease activ- BASDAI 3.83 (0.46) 3.03 (0.28) 0.80 ity and radiographic spinal progression was evaluated p = 0.144 separately for patients receiving long-term treatment CRP 1.41 (0.11) 1.22 (0.07) 0.19 (TNFi ≥4 years) and patients treated for a shorter p = 0.130 period. Our results showed that patients on long-term Results derived from a mixed-effect linear model with no other covariates (univariable analysis) or including potential confounders factors collected at treatment with a mean sustained average of low dis- time of recruitment (multivariable analysis): length of the radiographic interval, ease activity measured by ASDAS over time show sig- gender, age, smoking habit, radiographic sacroiliitis, mSASSS average, presence nificantly less progression rates than those with high of syndemosphytes, HLAB27, and time from symptomatology onset. Statistical significance was assessed using Wald tests derived from the model. SE, standard ASDAS averages. Conversely, in patients treated for error; p, p-value; CRP, C-reactive protein; BASDAI, Bath Ankylosing Spondylitis less than 4 years who experimented radiographic pro- Disease Activity Index; ASDAS-CRP, Ankylosing Spondylitis Disease Activity gression presented a progressive increase of ASDAS mean score compared with those without progression. Discussion Such differences, however, did not achieve statistical In this study, we evaluated the association between significance until the third year of follow-up. Hence, disease activity and radiographic spinal progression our data supports previous studies that suggest a slow- in axSpA patients receiving TNFi who were included down effect on radiographic progression associated in a Spanish national registry (REGISPONSERBIO). to the maintenance of disease control in patients with Our results showed that disease activity measured by axSpA treated with TNFi [3, 22]. Llop et al. Arthritis Research & Therapy (2022) 24:30 Page 7 of 9 To our knowledge, this is the first study suggesting that the future, as it has been shown to detect more progres- low disease activity is an appropriate target in the inhibi- sion in the form of new and growing syndesmophytes tion of radiographic spinal progression. As stated in the in patients with r-axSpA [29]. Currently, conventional recent ASAS-EULAR management recommendations for radiography requires a radiographic interval of at least axSpA, a treat-to-target approach is recommended [10, 2 years to detect a relevant change, so long-term treat- 23]. Nevertheless, there is no absolute consensus about ment is necessary before radiographic progression can which specific target to use. Molnar et al. [9] suggest that be detected. Another possible issue is the extrapola- inactive disease (ASDAS <1.3) might be an adequate tar- tion of our results to the study population, as only 101 get for the inhibition of spinal radiographic progression. patients out from 257 that are part of the REGISPON- However, it might be quite challenging to achieve inac- SERBIO cohort were included in the study based on tive disease in standard clinical practice, as only 25–35% the availability of full sets of radiographs (lumbar and of patients achieve inactive disease status, according to cervical lateral view) and clinical data. Nevertheless, we previously published studies [18, 24]. These observations found only minor differences regarding variables that are in line with those of our own cohort, in which only could influence disease activity and radiographic pro- 23% of patients had inactive disease at the time of recruit- gression, which discards a major impact on the results ment, while 52.1% of patients achieved low disease activ- due to a selection bias. Finally, although readers were ity. Nevertheless, our findings suggest that low disease blinded to all clinical data, radiographic scoring was activity (ASDAS <2.1) might be an appropriate target for performed with readers being aware of the chronologic preventing radiographic progression and a more reason- order. However, this approach appears to be more sen- able treatment objective to achieve in clinical practice. sitive for detecting change than reading with paired In our data, disease activity measured by BASDAI and time order [30]. CRP showed similar trends to ASDAS, although the asso- One of the strengths of our study is the prospective ciation with progression status was weaker and did not nature of the cohort used in our study, which provided show statistical significance. These findings are in line disease activity data every 6 months over a follow-up with previous publications, where despite an acceptable period of 3 years. This amount of disease activity data concordance between patient groups defined by BASDAI offered an accurate picture of disease status in patients ≥4 and ASDAS ≥2.1 [25, 26], ASDAS is the preferred treated with TNFi. Additionally, this design allowed the tool for assessing disease activity [10]. As reported by use of statistical tools to analyze the data as an experi- previous authors, a longitudinal relationship between ment of repeated measures (linear-mixed effects model). ASDAS and subsequent syndesmophyte formation was Finally, disease activity was assessed with ASDAS-CRP, found, while such a relationship between BASDAI and BASDAI score, and CRP serum levels, allowing compari- syndesmophytes was far weaker [3, 27]. Furthermore, son in terms of radiographic implications. ASDAS shows a better correlation than BASDAI when measurements of disease activity assessed by patients and physicians are compared [28]. Finally, ASDAS cutoffs Conclusions for disease activity and response criteria are well estab- Patients receiving TNFi with a mean sustained low dis- lished and are based on a thorough validation process, ease activity over time measured by ASDAS presented in contrast to the BASDAI scale where thresholds have less radiographic progression rates than those with high been defined in a more arbitrarily manner [24]. disease activity. The benefit of maintaining ASDAS at low Our study had several limitations. First, not all levels becomes apparent after long-term treatment with REGISPONSERBIO patients could be included in the TNFi (≥4 years). Moreover, our data emphasize the use current study due to the unavailability of suitable radi- of ASDAS as the preferred tool to assess disease activity, ographs. This point has negatively affected the final and suggest that low disease activity (<2.1) is an accept- number of patients available for inclusion and only 30 able threshold as a therapeutic objective in a treat-to-tar- patients with radiographic progression were available get strategy to prevent radiographic progression. for analyses (15 progressors vs 40 non-progressors in the long-term treatment group; 15 progressors vs 31 Abbreviations non-progressors in the not long-term treated patients). AxSpA: Axial spondyloarthritis; Nr-axSpA: Non-radiographic SpA; R-axSpA: Secondly, conventional radiography might be an insen- Radiographic SpA; AS: Ankylosing Spondylitis; TNFi: TNF inhibitors; mSASS: sitive tool for detecting radiographic spinal progression Modified Stoke Ankylosing Spondylitis Spine Score; ASDAS: Ankylosing Spondylitis Disease Activity Score; BASDAI: Bath Ankylosing Spondylitis compared to other imaging techniques. Low-dose CT Disease Activity Index; BASFI: Bath Ankylosing Spondylitis Functional Index; VC: has the potential to become the primary method used Vertebral corner; SDC: Smallest detectable change; CRP: C-reactive protein; SD: to detect radiographic spinal progression in axSpA in Standard deviation; OR: Odds ratio; ICC: Intraclass correlation coefficient. Llop et al. Arthritis Research & Therapy (2022) 24:30 Page 8 of 9 Supplementary Information 2 Rheumatology, Hospital Universitario La Paz, Madrid, Spain. 3 Rheumatol‑ ogy, Hospital Universitari Bellvitge, L’Hospitalet de Llobregat, Barcelona, The online version contains supplementary material available at https://​doi.​ Spain. 4 Rheumatology Unit, Hospital Alcorcon Foundation, Madrid, Spain. org/​10.​1186/​s13075-​021-​02695-5. 5 Rheumatology, Hospital Universitario Príncipe de Asturias, Alcalá de Henares, Spain. 6 Rheumatology, Hospital Universitario Puerta de Hierro, Madrid, Spain. 7 Additional file 1: Supplementary Table S1. Baseline characteristics Rheumatology, Hospital del Mar, Barcelona, Spain. 8 Rheumatology, Hospital of the 101 patients included in the study by baseline TNFi treatment Universitario Virgen Macarena, Sevilla, Spain. 9 I3PT Research Institute (UAB), status and radiographic progression. Supplementary Table S2. Baseline Sabadell, Spain. characteristics of the 101 patients included in the study by radiographic progression. Supplementary Table S3. Baseline characteristics of the 156 Received: 13 July 2021 Accepted: 8 December 2021 REGISPONSERBIO patients not included in the study. Supplementary Table S4. Comparison of baseline characteristics at first visit between patients included (101) and not included (156) in the study. Supple- mentary Table S5. Inter-observer reliability. Supplementary Figure S1 Blat and Altman Pot, inter-reader reliability. Supplementary Figure S2. References Boxplot of the ASDAS values averaged across all time points in patients 1. Poddubnyy D, Sieper J. Mechanism of new bone formation in axial spon‑ under long-term TNFi treatment. Supplementary Figure S3. Associa‑ dyloarthritis. Curr Rheumatol Rep. 2017;19(9):55. tion between disease activity measured by BASDAI (a) and CRP(b) every 2. van der Linden S, Valkenburg HA, Cats A. Evaluation of diagnostic criteria 6 months with radiographic progression defined as an increase in the for ankylosing spondylitis. A proposal for modification of the New York mSASSS score at ≥2 points in patients unter TNFi treatment during a criteria. Arthritis Rheum. 1984;27(4):361–8. follow-up period of three years. Bold dots and intervals represent means 3. Ramiro S, van der Heijde D, van Tubergen A, Stolwijk C, Dougados M, van and +/- their corresponding standard errors provided a the mixed-effects den Bosch F, et al. Higher disease activity leads to more structural dam‑ linear model (no other covariates included). age in the spine in ankylosing spondylitis: 12-year longitudinal data from the OASIS cohort. Ann Rheum Dis. 2014;73(8):1455–61. 4. van der Heijde D, Landewé R, Einstein S, Ory P, Vosse D, Ni L, et al. Radio‑ Acknowledgements graphic progression of ankylosing spondylitis after up to two years of We would like to thank the following rheumatologists for inclusion of their treatment with etanercept. Arthritis Rheum. 2008;58(5):1324–31. patients: Pedro Zarco Montejo, Beatriz Joven, Miriam Almirall, Mª Cruz Fer‑ 5. van der Heijde D, Landewé R, Baraliakos X, Houben H, van Tubergen A, nandez Espartero, Enrique Batlle Gualda, Cristina Campos, Eduardo Collantes Williamson P, et al. Radiographic findings following two years of inflixi‑ Estevez, Pilar Font, Teresa Clavaguera Poch, Luis F. Linares Ferrando, Carlos Rod‑ mab therapy in patients with ankylosing spondylitis. Arthritis Rheum. ríguez Lozano, and Beatriz Yoldi for the REGISPONSERBIO Study Group. Further, 2008;58(10):3063–70. we would like to thank V. Rios-Rodriguez for the scientific advice of the study. 6. van der Heijde D, Salonen D, Weissman BN, Landewé R, Maksymowych WP, Kupper H, et al. Assessment of radiographic progression in the spines Authors’ contributions of patients with ankylosing spondylitis treated with adalimumab for up ML, AB, and JC contributed to the data analysis and interpretation and draft‑ to 2 years. Arthritis Res Ther. 2009;11(4):R127. ing of the manuscript. ML and MM scored the images. MM, VN, XJ, and EM 7. Haroon N, Inman RD, Learch TJ, Weisman MH, Lee M, Rahbar MH, contributed to the initiation of REGISPONSERBIO, data collection, data inter‑ et al. The impact of tumor necrosis factor α inhibitors on radio‑ pretation, and drafting of the manuscript. RA, EC, JS, EB, and MR contributed graphic progression in ankylosing spondylitis. Arthritis Rheum. to the data collection and drafting of the manuscript. JG contributed to the 2013;65(10):2645–54. data collection, analysis, and interpretation and drafting of the manuscript. All 8. Baraliakos X, Gensler LS, D’Angelo S, Iannone F, Favalli EG, de Peyrecave authors read and approved the final manuscript. N, et al. Biologic therapy and spinal radiographic progression in patients with axial spondyloarthritis: a structured literature review. Ther Adv Funding Musculoskelet Dis. 2020;12:1759720X20906040. This work was supported by Pfizer and complementary financial support 9. Molnar C, Scherer A, Baraliakos X, de Hooge M, Micheroli R, Exer P, et al. for the medical writing has been obtained also from the Catalan Society of TNF blockers inhibit spinal radiographic progression in ankylosing spon‑ Rheumatology. dylitis by reducing disease activity: results from the Swiss clinical quality management cohort. Ann Rheum Dis. 2018;77(1):63–9. Availability of data and materials 10. van der Heijde D, Ramiro S, Landewe R, Baraliakos X, Van den Bosch The data that support the findings of this study are available from the cor‑ F, Sepriano A, et al. 2016 update of the ASAS-EULAR manage‑ responding author upon reasonable request. ment recommendations for axial spondyloarthritis. Ann Rheum Dis. 2017;76(6):978–91. 11. Ørnbjerg LM, Brahe CH, Askling J, Ciurea A, Mann H, Onen F, et al. Declarations Treatment response and drug retention rates in 24 195 biologic-naïve patients with axial spondyloarthritis initiating TNFi treatment: routine Ethics approval and consent to participate care data from 12 registries in the EuroSpA collaboration. Ann Rheum Dis. All patients signed an informed consent to participate in REGISPONSERBIO, 2019;78(11):1536–44. and the study was approved by the central ethics committee. 12. Moreno M, Gratacós J, Navarro-Compán V, de Miguel E, Font P, Clavagu‑ era T, et al. Should over-treatment of axial spondyloarthritis with biologics Consent for publication remain a concern after the issue of the new ASAS criteria? Data from Not applicable. REGISPONSERBIO (Spanish register of biological therapy in Spondyloarth‑ ritides). Clin Exp Rheumatol. 2018;36(6):1038–42. Competing interests 13. Gratacós J, Del Campo D, Fontecha P, Fernández-Carballido C, Juanola All authors have completed the ICMJE uniform disclosure form at www.icmje. Roura X, Linares Ferrando LF, et al. Recommendations by the Spanish org/coi_disclosure.pdf and declare no support from any organization for the sub‑ Society of Rheumatology on the use of biological therapies in axial spon‑ mitted work, no financial relationships with any organizations that might have an dyloarthritis. Reumatol Clin. 2018;14(6):320–33. interest in the submitted work in the previous 3 years, and no other relationships 14. Creemers MC, Franssen MJ, van’t Hof MA, Gribnau FW, van de Putte or activities that could appear to have influenced the submitted work. LB, van Riel PL. Assessment of outcome in ankylosing spondylitis: an extended radiographic scoring system. Ann Rheum Dis. 2005;64(1):127–9. Author details 15. Ramiro S, van Tubergen A, Stolwijk C, Landewé R, van de Bosch F, 1 Rheumatology, Parc Taulí Hospital Universitari, I3PT Research Institute Dougados M, et al. Scoring radiographic progression in ankylosing spon‑ (UAB), Universitat Autónoma de Barcelona (UAB), Sabadell 08208, Spain. dylitis: should we use the modified stoke Ankylosing spondylitis spine Llop et al. Arthritis Research & Therapy (2022) 24:30 Page 9 of 9 score (mSASSS) or the radiographic Ankylosing spondylitis spinal score (RASSS)? Arthritis Res Ther. 2013;15(1):R14. 16. Llop M, Rios Rodriguez V, Redeker I, Sieper J, Haibel H, Rudwaleit M, et al. Incorporation of the anteroposterior lumbar radiographs in the modified stoke Ankylosing spondylitis spine score improves detection of radio‑ graphic spinal progression in axial spondyloarthritis. Arthritis Res Ther. 2019;21(1):126. 17. Poddubnyy D, Rudwaleit M, Haibel H, Listing J, Märker-Hermann E, Zeidler H, et al. Effect of non-steroidal anti-inflammatory drugs on radiographic spinal progression in patients with axial spondyloarthritis: results from the German Spondyloarthritis inception cohort. Ann Rheum Dis. 2012;71(10):1616–22. 18. Machado PM, Landewé R, Heijde DV, ASAS, AoSiS. Ankylosing spondylitis disease activity score (ASDAS): 2018 update of the nomenclature for disease activity states. Ann Rheum Dis. 2018;77(10):1539–40. 19. Bruynesteyn K, Boers M, Kostense P, van der Linden S, van der Heijde D. Deciding on progression of joint damage in paired films of individual patients: smallest detectable difference or change. Ann Rheum Dis. 2005;64(2):179–82. 20. R Core Team. R: a language and environment for statistical computing. Austria: R Foundation for Statistical Computing V; 2019. URL https://​ www.R-​proje​ct.​org/ 21. Jeong H, Eun YH, Kim IY, Park EJ, Kim H, Lee J, et al. Effect of tumor necro‑ sis factor α inhibitors on spinal radiographic progression in patients with ankylosing spondylitis. Int J Rheum Dis. 2018;21(5):1098–105. 22. Boers N, Michielsens CAJ, van der Heijde D, den Broeder AA, Welsing PMJ. The effect of tumour necrosis factor inhibitors on radiographic progres‑ sion in axial spondyloarthritis: a systematic literature review. Rheumatol‑ ogy (Oxford). 2019;58(11):1907–22. 23. Smolen JS, Schöls M, Braun J, Dougados M, FitzGerald O, Gladman DD, et al. Treating axial spondyloarthritis and peripheral spondyloarthritis, especially psoriatic arthritis, to target: 2017 update of recommendations by an international task force. Ann Rheum Dis. 2018;77(1):3–17. 24. Machado P, Landewé R, Lie E, Kvien TK, Braun J, Baker D, et al. Ankylos‑ ing spondylitis disease activity score (ASDAS): defining cut-off values for disease activity states and improvement scores. Ann Rheum Dis. 2011;70(1):47–53. 25. Fagerli KM, Lie E, van der Heijde D, Heiberg MS, Kaufmann C, Rødevand E, et al. Selecting patients with ankylosing spondylitis for TNF inhibitor therapy: comparison of ASDAS and BASDAI eligibility criteria. Rheumatol‑ ogy (Oxford). 2012;51(8):1479–83. 26. Vastesaeger N, Cruyssen BV, Mulero J, Gratacós Masmitjá J, Zarco P, Almodovar R, et al. ASDAS high disease activity versus BASDAI elevation in patients with ankylosing spondylitis as selection criterion for anti-TNF therapy. Reumatol Clin. 2014;10(4):204–9. 27. Sepriano A, Ramiro S, Wichuk S, Chiowchanwisawakit P, Paschke J, van der Heijde D, Landewé R, Maksymowych WP. Tumor necrosis factor inhibitors reduce spinal radiographic progression inpatients with radiographic axial spondyloarthritis: a longitudinal analysis from the Alberta prospective cohort. Arthritis Rheumatol. 2021;73(7):1211–19. 28. Lukas C, Landewé R, Sieper J, Dougados M, Davis J, Braun J, et al. Devel‑ opment of an ASAS-endorsed disease activity score (ASDAS) in patients with ankylosing spondylitis. Ann Rheum Dis. 2009;68(1):18–24. 29. de Koning A, de Bruin F, van den Berg R, Ramiro S, Baraliakos X, Braun J, et al. Low-dose CT detects more progression of bone formation in com‑ parison to conventional radiography in patients with ankylosing spondy‑ litis: results from the SIAS cohort. Ann Rheum Dis. 2018;77(2):293–9. Ready to submit your research ? Choose BMC and benefit from: 30. Wanders A, Landewé R, Spoorenberg A, de Vlam K, Mielants H, Dougados M, et al. Scoring of radiographic progression in randomised clinical trials • fast, convenient online submission in ankylosing spondylitis: a preference for paired reading order. Ann Rheum Dis. 2004;63(12):1601–4. • thorough peer review by experienced researchers in your field • rapid publication on acceptance Publisher’s Note • support for research data, including large and complex data types Springer Nature remains neutral with regard to jurisdictional claims in pub‑ • gold Open Access which fosters wider collaboration and increased citations lished maps and institutional affiliations. • maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions