Hierarchy of Impairment of Spinal Mobility Measures in Ankylosing Spondylitis: Twelve-Year Data

Arthritis Care & Research Vol. 67, No. 11, November 2015, pp 1571 1577 DOI 10.1002/acr.22614 VC 2015, American College of Rheumatology ORIGINAL ARTICLE Hierarchy of Impairment of Spinal Mobility Measures in Ankylosing Spondylitis: Twelve-Year Data SOFIA RAMIRO, 1 ROBERT B. M. LANDEW E, 2 D ESIR EE VAN DER HEIJDE, 3 CARMEN STOLWIJK, 4 MAXIME DOUGADOS, 5 FILIP VAN DEN BOSCH, 6 AND ASTRID VAN TUBERGEN 4 Objective. To investigate which spinal mobility measures (SMMs) are most frequently impaired in patients with ankylosing spondylitis (AS), whether a hierarchy of impairment can be established, and whether assessing fewer measures sufficiently captures impairment in spinal mobility. Methods. Patients from the Outcome in Ankylosing Spondylitis International Study were followed up for 12 years. SMMs were considered impaired when falling below predefined cutoffs, derived from normal individuals. The proportion of patients in whom each SMM was impaired was calculated using baseline observation. In patients with 1 impaired SMM, we investigated how often impairment in spinal mobility would be missed if only a fixed number of SMMs was assessed. Analyses were repeated using all 12-year observations. Results. A total of 216 patients were included (70% males). Lateral spinal flexion (LSF) was the most frequently impaired measure, followed by the modified Schober (mschober) test, tragus-to-wall, cervical rotation, intermalleolar distance, and chest expansion measures, respectively. This hierarchy was strikingly consistent over time, and independent of sex, symptom duration, and presence of syndesmophytes. In patients with 1 impaired SMM, LSF was impaired most frequently (86%), followed by the mschober test (58%). If only LSF was measured, 14% of patients with impairment in any SMM would be missed; if additionally the mschober test was measured, 9% would be missed. Conclusion. LSF followed by the mschober test are the most frequently impaired mobility measures in AS, reflecting an earlier involvement of the lumbar spine, followed by involvement of the thoracic and cervical spine. In clinical practice LSF and the mschober test suffice to screen impairment in spinal mobility. INTRODUCTION Impairment of spinal mobility is widely recognized as an important clinical sign and a hallmark of axial spondyloarthritis (SpA). This is reflected by its inclusion in the modified New York Criteria for the classification of ankylosing spondylitis (AS) and in the core set of domains for the evaluation of patients with SpA in clinical practice Dr. Ramiro s work was supported by the Fundaç~ao para a Ci^encia e Tecnologia (grant SFRH/BD/68684/20). 1 Sofia Ramiro, MD, MSc: Amsterdam Rheumatology Center, University of Amsterdam, Amsterdam, The Netherlands, and Hospital Garcia de Orta, Almada, Portugal; 2 Robert B. M. Landewe, MD, PhD: Amsterdam Rheumatology Center, University of Amsterdam, Amsterdam, and Atrium Medical Center, Heerlen, The Netherlands; 3 Desiree van der Heijde, MD, PhD: Leiden University Medical Center, Leiden, The Netherlands; 4 Carmen Stolwijk, MD, Astrid van Tubergen, MD, PhD: Maastricht University Medical Center and University of Maastricht, Maastricht, The Netherlands; 5 Maxime Dougados, MD: Paris Descartes University, Cochin Hospital, AP-HP, INSERM (U1153), PRES, Sorbonne Paris-Cite, France; 6 Filip van den Bosch, MD, PhD: Ghent University Hospital and Ghent University, Ghent, Belgium. and trials, as defined by the Assessment of Spondylo- Arthritis international Society (ASAS) (1). A study on the natural course of AS conducted in the pre biologic agent era described that in the long term more than half of all patients will develop moderate to severe impairment of spinal mobility (2). Spinal mobility is one of the central outcomes included in observational and treatment studies in axial SpA (3,4) and is a predictor of poor prognosis (3). The spinal mobility measures (SMMs) recommended by ASAS are chest expansion, modified Schober test (mschober; 10 cm), occiput-to-wall distance (OTW), cervical rotation, and either lateral spinal flexion (LSF) or Bath Ankylosing Spondylitis Metrology Index (BASMI) (5,6). The Dr. van den Bosch has received speaking fees, consulting fees, and/or honoraria (less than $10,000 each) from AbbVie, Celgene, Pfizer, Novartis, and UCB. Address correspondence to Sofia Ramiro, MD, MSc, Department of Clinical Immunology and Rheumatology, Amsterdam Rheumatology Center, University of Amsterdam, Meibergdreef 9, Amsterdam 1105 AZ, The Netherlands. E-mail: sofiaramiro@ gmail.com. Submitted for publication October 7, 2014; accepted in revised form May 5, 2015. 1571

1572 Ramiro et al Significance & Innovations In ankylosing spondylitis, impaired spinal mobility occurs first in the lumbar spine, followed by the thoracic and cervical spine. This order of occurrence persists independently of sex, symptom duration, or presence of baseline syndesmophytes. In clinical practice, the measurement of lateral spinal flexion and the modified Schober test suffice to screen for impairment in spinal mobility. If lateral spinal flexion and modified Schober test are impaired, additional spinal mobility measures are necessary to assess the total level of spinal involvement. BASMI, in turn, includes LSF, the mschober test, cervical rotation, tragus-to-wall distance (TTW), and intermalleolar distance (IMD) (5). LSF and BASMI have shown the best discrimination between patients with and without structural damage (7). A normal LSF finding is the best predictor of the absence of radiographic damage, while an abnormal mschober test finding best predicts its presence. Furthermore, LSF and cervical rotation more consistently discriminated between active treatment and placebo, and between responders and nonresponders in randomized controlled trials with tumor necrosis factor (TNF) inhibitors (8,9). However, to our knowledge it has never been investigated whether specific SMMs are more frequently impaired than others in patients with AS compared to normal individuals, and whether there is a certain sequential or hierarchical order in the occurrence of impairment. Such an analysis was hampered by the lack of normal values for SMMs. We have recently described reference intervals and percentile curves for SMMs obtained in normal individuals (10). In the present study, we have used these percentile curves for SMMs to perform a formal age-matched comparison with values obtained from patients with AS. The aim of this study was to investigate which SMMs are most frequently impaired in patients with AS, whether a hierarchical order in the occurrence of impairment could be established, and whether this would be dependent on characteristics such as symptom duration. Furthermore, we have investigated whether assessing fewer measures could still capture sufficient information for detecting impairment of spinal mobility. approved the study. All patients have provided written informed consent. Spinal mobility measures. Six SMMs were analyzed in this study: chest expansion, mschober test, cervical rotation, LSF, TTW, and IMD. Measurements were performed according to the ASAS recommendations, and the reference values were obtained using these same assessments (6,10). The better of 2 tries for each measurement was recorded. All measurements were recorded in centimeters, except for cervical rotation, which was recorded in degrees. All measurements were rounded to decimals, except for IMD and cervical rotation, which were rounded to units. Additionally, the BASMI (5) (range 0 10, with a higher score representing a worse mobility) was computed according to the BASMI linear formula (12). In the first 2 years, LSF has erroneously been recorded as the distance from middle fingertip to floor on maximum flexion for Dutch patients. This means that the fingertip to MATERIALS AND METHODS Study population. For this study, data from the Outcome in Ankylosing Spondylitis International Study (OASIS) were used. OASIS is a prevalence cohort that started in 1996 and included 217 consecutive patients with AS from the Netherlands, Belgium, and France (11). Clinical data, including spinal mobility, were collected regularly during 12 years. In the first 2 years, spinal mobility was measured every 6 months and afterward every 2 years. All patients were included in the present study. The ethics committees of all participating hospitals have Figure 1. Lateral spinal flexion in function of age and with the percentile curves and help lines derived from measurements in normal individuals (10). Solid lines represent the percentile curves derived from normal individuals (percentiles indicated). Broken lines (a 0 to d 0 ) are help lines parallel to the percentile curves, derived to accommodate impaired values of patients with ankylosing spondylitis. Dots represent the real measurements of lateral spinal flexion of patients with ankylosing spondylitis at baseline (A) orinall12-yearobservations(b).

Impairment of Spinal Mobility in AS 1573 floor distance in a neutral position was missing, which hampers the calculation of the LSF. To replace this missing value and enable the computation of the LSF, we derived the neutral fingertip to floor distance from height according to an equation proposed for this purpose (13). Impairment in spinal mobility. Percentile curves (2.5th to 97.5th percentiles) were obtained for each of the SMMs from our previous study in normal individuals, the MOBIL- ITY study (10). Figure 1 illustrates such percentile curves for LSF as an example. These curves represent age-specific spinal mobility in normal individuals and its evolution by increasing age. The 2.5th and 97.5th percentile curves obtained in normal individuals correspond with Z scores of 21.96 and 11.96 (10). In patients with AS impaired mobility is expected in comparison with normal individuals. In order to better describe this level of impairment, we have introduced 4 additional parallel curves (help lines) to the mobility curves, corresponding with Z scores of 22.50 (a 0 ); 23.00 (b 0 ); 23.50 (c 0 ); and 24.00 (d 0 ), respectively (Figure 1). For every observation in OASIS, an SMM was defined as impaired if its value was below a certain cutoff (either the 2.5th percentile or one of the 4 help lines). OTW was not applied in this study, because percentile curves for this measure could not be derived. The reference value for OTW is zero and the vast majority of normal individuals indeed have an OTW of zero (10). We could have applied a value of zero here, but this value would be a static figure and would not change by increasing age or with the different percentile curves used, as opposed to what happens with the other measures. This would make the evaluation of the hierarchy for the impairment of the SMM across the different cutoffs complex or even impossible. We therefore used the TTW, which is another measure for cervical kyphosis. However, for TTW, no percentile curves had been derived in the MOBILITY study, with the rationale that this measure is strongly dependent on the conformation of the head and would therefore not be so relevant to derive reference intervals. For the purpose of the present study, percentile curves for TTW were derived following the methodology described for the MOBILITY study (for more details, see Supplementary Appendix A and Supplementary Figure 1, available in the online version of this article at http://onlinelibrary.wiley. com/doi/10.1002/acr.22614/abstract) (10). Subgroup analysis. Impairment of spinal mobility was also investigated in subgroups of patients, taking into account variables that may influence spinal mobility. Subgroups were created based on sex, symptom duration (above and below the median [,18 years versus $18 years]), as well as in tertiles (,14 years, 14 25.29 years, and $25.3 years), and the presence of baseline syndesmophytes (0 versus $1 and,5 versus $5). The scores on syndesmophytes from 2 readers (separate scores from each of the readers) were used (14). Subgroup analysis. We conducted 2 types of analyses. First, for each of the SMMs, the proportion of patients in which the SMM was impaired according to each of the cutoffs was calculated, using the baseline observation. For each of the cutoffs, measures were ranked according to the propor- Table 1. Impairment of each of the spinal mobility measures (SMMs) in patients with AS, defined according to age-specific cutoffs derived from normal subjects* 2.5th percentile, line a 0, line b 0, line c 0, line d 0, Total observations, no. Patient level (baseline only) Lateral spinal flexion (cm) 149 (73) 138 (68) 119 (59) 106 (52) 55 (27) 203 Modified Schober test (cm) 111 (55) 94 (46) 78 (38) 63 (31) 36 (18) 203 Tragus-to-wall distance (cm) 85 (42) 62 (31) 43 (21) 203 Cervical rotation (degrees) 55 (27)# 44 (22)# 38 (19)# 28 (14)# 26 (13)# 203 Intermalleolar distance (cm) 39 (19)** 27 (13) 22 (11) 12 (6) 8 (4) 203 Chest expansion (cm) 54 (27) 17 (8)** 4 (2)** 0 (0)** 0 (0)** 203 Observational level (all observations) Lateral spinal flexion (cm) 1,035 (73) 919 (65) 804 (57) 661 (46) 314 (22) 1,422 Modified Schober (cm) 845 (59) 696 (49) 571 (40) 453 (32) 282 (20) 1,422 Tragus-to-wall distance (cm) 622 (44) 490 (34) 342 (24) 1,422 Cervical rotation (degrees) 439 (31)# 349 (25)# 274 (19)# 215 (15)# 167 (12)# 1,422 Intermalleolar distance (cm) 320 (23)** 207 (15) 152 (11) 88 (6) 55 (4) 1,422 Chest expansion (cm) 386 (27) 139 (10)** 32 (2)** 4 (0)** 1 (0)** 1,422 * The first column refers to impairment defined as a value below the 2.5th percentile in normal individuals. The remaining columns refer to cutoffs defined on the basis of help lines, parallel to the percentile curves, derived to accommodate impaired values of patients with ankylosing spondylitis (AS; Figure 1). Impairment in spinal mobility is presented both at the patient level (using baseline observation) and observation level (all 12-year observations). SMM most frequently impaired. Second SMM most frequently impaired. Third SMM most frequently impaired. Not possible to derive these values due to mathematical characteristics of the equation for the percentile curves for tragus-to-wall distance. # Fourth SMM most frequently impaired. ** SMM least frequently impaired. Fifth SMM most frequently impaired.

1574 Ramiro et al tion of patients with impairment in each of the SMMs. Subsequently, we conducted a similar analysis, but now taking all observations over the 12 years per patient into account. Analyses were repeated in all the subgroups. The main analysis was conducted with patients (respectively, observations) in whom assessment of SMM was complete, i.e., all SMMs were assessed. Sensitivity analyses were conducted with patients (respectively, observations) in whom any SMM was available. For the second analysis, the number of patients with at least 1 impaired SMM was calculated, using the baseline observation. Each of the SMMs was defined as impaired if the value was below help line a 0 (first cutoff below normal individuals). Among the patients with at least 1 impaired SMM, the proportion of patients with each of the SMMs impaired was calculated. Furthermore, we investigated in how many cases impairment in spinal mobility (defined as impairment in at least 1 SMM) would be missed if only a fixed number of SMMs was assessed. Subsequently, we conducted the same analysis, but now taking all observations over the 12 years per patient into account. For this second analysis, only patients with complete assessments of SMMs were included. Analyses were done using Stata SE, version 12. RESULTS A total of 216 patients with the following baseline characteristics were included: 71% males, mean 6 SD age 44 6 13 years, mean 6 SD symptom duration 21 6 12 years, 85% HLA B27 positive, 68% treated with nonsteroidal antiinflammatory drugs, and none with biologic agents. Patients were on average followed up for 7.7 6 4.1 years with a mean 6 SD of 7.2 6 2.1 observations per patient with assessment of SMMs. At baseline, the mean 6 SD LSF was 9.6 6 5.8 cm, mschober test was 2.8 6 1.4 cm, TTW was 14.3 6 4.8 cm, cervical rotation was 65 6 23 degrees, IMD was 105 6 22 cm, chest expansion was 4.7 6 2.2 cm, and the BASMI was 3.8 6 1.6. Hierarchy of impairment in spinal mobility. LSF was always the most frequently impaired measure, sequentially followed by the mschober test, TTW, cervical rotation, IMD, and chest expansion (Table 1). This order was strikingly similar for all cutoffs (with the only exception being the 2.5th percentile, for which chest expansion was slightly more frequently impaired than IMD). Using as a cutoff the help line a 0 (first cutoff below normal individuals), LSF was impaired in 68% of the patients, followed by the mschober test, which was impaired in 47% of the patients (Table 1). Impairment of BASMI, also defined according to help line a 0, occurred in 49% of the patients. The same hierarchy was found in the analyses including all 12-year observations (Table 1). Figure 1 (panel A at patient level, panel B at observation level) shows the percentile curves and help lines for LSF and the measured values for LSF (i.e., dots) below each cutoff. Table 2. Impairment of each of the spinal mobility measures (SMMs) in patients with AS, compared to cutoffs defined according to age-specific cutoffs derived from normal subjects and stratified for symptom duration (median)* 2.5th percentile, help line a 0, help line b 0, help line c 0, help line d 0, Total observations, no. Patients with shorter symptom duration (,18 years) Lateral spinal flexion (cm) 66 (69) 58 (60) 46 (48) 41 (43) 24 (25) 96 Modified Schober (cm) 47 (49) 41 (43) 30 (31) 25 (26) 15 (16) 96 Tragus-to-wall distance (cm) 29 (30) 21 (22) 14 (15) 96 Cervical rotation (degrees) 18 (19)# 14 (15)** 12 (13)** 10 (10)** 10 (10)** 96 Intermalleolar distance (cm) 15 (16) 9 (9)# 5 (5)# 3 (3)# 2 (2)# 96 Chest expansion (cm) 20 (21)** 4 (4) 0 (0) 0 (0) 0 (0) 96 Patients with longer symptom 96 duration ($18 years) Lateral spinal flexion (cm) 80 (83) 77 (80) 70 (73) 62 (65) 28 (29) 96 Modified Schober (cm) 60 (63) 50 (52) 45 (47) 35 (36) 18 (19) 96 Tragus-to-wall distance (cm) 51 (53) 38 (40) 27 (28) 96 Cervical rotation (degrees) 34 (35)** 28 (29)** 24 (25)** 17 (18)** 15 (16)** 96 Intermalleolar distance (cm) 21 (22) 17 (18)# 16 (17)# 8 (8)# 5 (5)# 96 Chest expansion (cm) 33 (34)# 13 (14) 4 (4) 0 (0) 0 (0) 96 * The first column refers to impairment defined as a value below the 2.5th percentile in normal individuals. The remaining columns refer to cutoffs defined on the basis of help lines, parallel to the percentile curves, derived to accommodate impaired values of patients with ankylosing spondylitis (AS; Figure 1). Impairment in spinal mobility is presented at the patient level (for observation level, all 12-year observations, see Supplementary Table 2, available in the online version of this article at http://onlinelibrary.wiley.com/doi/10.1002/acr.22614/abstract). SMM most frequently impaired. Second SMM most frequently impaired. Third SMM most frequently impaired. Not possible to derive these values due to mathematical characteristics of the equation for the percentile curves for tragus-to-wall distance. # Fifth SMM most frequently impaired. ** Fourth SMM least frequently impaired. SMM least frequently impaired.

Impairment of Spinal Mobility in AS 1575 Table 3. Impairment of each of the spinal mobility measures (SMMs) in patients/ observations with at least 1 impaired SMM* Patient level (baseline assessment), (n 5 161) Observation level, (n 5 1,111) Impairment in each of the SMMs Lateral spinal flexion 138 (86) 919 (83) Modified Schober test 94 (58) 696 (63) Tragus-to-wall distance 62 (39) 490 (44) Cervical rotation 44 (27) 349 (31) Intermalleolar distance 27 (17) 207 (19) Chest expansion 17 (11) 139 (13) Number of SMMs impaired 1 49 (30) 282 (25) 2 53 (33) 355 (32) 3 27 (17) 203 (18) 4 15 (9) 168 (15) 5 16 (10) 91 (8) 6 1 (1) 12 (1) Impairment in spinal mobility captured by fewer measures Lateral spinal flexion 138 (86) 919 (83) Lateral spinal flexion 1 modified Schober test 147 (91) 1,008 (91) Lateral spinal flexion 1 modified Schober test 1 154 (96) 1,067 (96) tragus-to-wall distance Lateral spinal flexion 1 modified Schober test 1 160 (99) 1,097 (99) tragus-to-wall distance 1 cervical rotation Lateral spinal flexion 1 modified Schober test 1 tragus-to-wall distance 1 cervical rotation 1 intermalleolar distance 160 (99) 1,108 (100) * Each of the SMMs was defined as impaired if the measurement fell below help line a 0 (first cutoff below normal individuals; see Figure 1). Among the patients (n 5 161) or observations (n 5 1,111) with at least 1 impaired SMM, the proportion of patients (respectively of observations) with each of the SMMs impaired was calculated. A sensitivity analysis, conducted with patients (respectively, observations) in whom any SMM was available, provided similar results (Supplementary Table 1, available in the online version of this article at http://onlinelibrary. wiley.com/doi/10.1002/acr.22614/abstract). Also, after stratification for symptom duration (median), sex, or the presence of baseline syndesmophytes this hierarchy in general persisted, both at the patient level (Table 2 and Supplementary Tables 2 and 3) and at the observation level (example shown for symptom duration in Supplementary Table 4). Stratification for symptom duration in tertiles also retrieved the same results, as well as for the presence of baseline syndesmophytes according to the second reader (data not shown). Impairment in spinal mobility captured by fewer measures. Any impairment in spinal mobility was present in 161 of the 203 patients (79%) and in 1,111 of 1,422 observations (78%) with complete assessment of SMMs. Among patients with at least 1 impaired SMM, LSF was impaired in 86%, followed by the mschober test in 58%. One SMM was impaired in 30% of the patients, 2 SMMs were impaired in 33%, and 3 SMMs were impaired in 17% (Table 3). If only LSF was measured, 14% of the patients with impairment in SMMs would be missed since in these patients LSF was normal but another SMM was impaired. If the mschober test was measured in addition to LSF, only 9% of patients would be missed. When TTW was measured in addition to LSF and the mschober test, only 4% of the patients with impairment in mobility would be missed. Analyses taking into account all 12-year observations yielded similar results (Table 3). DISCUSSION In the present study we have shown that there is an order of involvement of the spine in AS; LSF and the mschober tests are the most frequently impaired mobility measures in AS, reflecting an earlier involvement of the lumbar spine in impairment in spinal mobility, followed by involvement of the thoracic and cervical spine. This order of involvement of the spine persists across different subgroups of patients based on sex, disease duration, or the presence of baseline syndesmophytes. To our knowledge the order of impairment in SMMs has never been investigated before. Therefore, we cannot compare our findings with others. Previous studies comparing spinal mobility in normal individuals and in patients with AS have focused on a single SMM and were performed in

1576 Ramiro et al the 1970s (15,16). Several factors, most importantly different measurement techniques, hamper a proper comparison with our results (15,16). The lack of studies comparing impairment in spinal mobility between patients with axial SpA and normal individuals can partly be explained by the fact that percentile curves from normal individuals did not exist until recently. With the age-adjusted percentile curves from the MOBILITY study (10), we now have better tools to compare spinal mobility between patients with axial SpA and normal individuals and across studies. Interestingly, the SMMs most frequently impaired, LSF and the mschober test, have also been identified as the ones best discriminating between patients with and without spinal radiographic damage (7). This reinforces their importance when assessing spinal mobility. Furthermore, LSF is considered the SMM that is most sensitive to change (8). The relationship between LSF (as a single measure) and radiographic damage has been shown to be as good as the relationship between BASMI (composite measure) and radiographic damage, and substantially better than the relationship between other SMMs and radiographic damage (7). Therefore, ASAS has recommended in the core set that the LSF can be used instead of the BASMI. Our findings endorse this idea, as we have found a higher proportion of observations with impaired LSF than with impaired BASMI, when comparing patients with normal individuals. Furthermore, LSF has been selected as one of the items of the ASAS 5/6 response criteria (17). From our study, it can be deduced that the detection of impairment of spinal mobility can be done by assessing only 2 SMMs: LSF and the mschober test. With this screening method, impairment of spinal mobility is only missed in 9% of the cases. We hereby recommend that these 2 measurements are minimally performed in order to determine impairment in spinal mobility in individual patients with AS. If these 2 are impaired, it is important to assess additional SMMs to be informed about the total level of spinal involvement. If LSF and the mschober test findings are normal, it is unlikely that other SMMs will be found abnormal. This is particularly useful for daily clinical practice, where it may sometimes be unfeasible, due to time constraints, to perform a complete BASMI or all SMMs, even though these are recommended by ASAS (1). We were unable to find any study investigating the feasibility of spinal mobility assessment in clinical practice, but it may well be the case that spinal mobility measurement is frequently skipped for time reasons. Of note, if impairment in LSF and/or the mschober test is identified, then a full assessment of all SMMs is still recommended, which will ultimately inform the rheumatologist about the severity of the disease and will likely guide clinical decisions, such as referral to physiotherapy and/or rehabilitation medicine. An important limitation of the study is that we have used a prevalence cohort of patients with AS and have neither taken into account the development of impairment of spinal mobility over time, or whether this also occurs in a fixed order. We have rather analyzed impairment in spinal mobility cross-sectionally at multiple time points. We have found a hierarchy in the occurrence of impairments per SMM and it is plausible that impairment in spinal mobility over time occurs in an order described in this study. Formally, this sequence needs to be prospectively validated, but that will be extremely difficult, since the results of our study indicate that for that purpose, an inception cohort is required in which patients virtually without impairments at baseline are followed for 30 or more years in order to prove the sequence we have proposed here in a longitudinal fashion. Moreover, a comparison was established between patients with AS and normal values derived from normal individuals. We can therefore not know whether these findings are specific for AS, or whether the same type of impairment would be found in other diseases affecting the spine, e.g., mechanical discopathy. Another possible critique is the imputation of LSF values of Dutch patients during the first 2 years. Even though the imputation method performed very well (13), we cannot fully exclude some possible errors. Strengths of our study are the formal comparison between patients with AS with percentile curves derived from normal individuals. We included a large number of measurements, and the hierarchy proposed persisted across subgroups, which adds credibility to the findings. Confirmation of our findings in another cohort, and also in earlier phases of the development of mobility impairment, e.g., in patients with nonradiographic axial SpA, are warranted. If confirmed, it would be interesting to explore what factors are associated with this higher impairment in mobility of the lumbar spine, compared to the cervical or thoracic spine. Inflammation may hypothetically be one of the factors involved, as we know that treatment with TNF inhibitors, which control inflammation and disease activity, is associated with improvement in spinal mobility (8,9). In summary, this study has shown that there is an order for impairment in spinal mobility and provides clinicians with a data-driven approach to screen impairment in spinal mobility by assessing the LSF and the mschober test in patients with AS. AUTHOR CONTRIBUTIONS All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Dr. Ramiro had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study conception and design. Ramiro, Landewe, van der Heijde, Dougados, van Tubergen. Acquisition of data. Ramiro, Landewe, Stolwijk, Dougados, van den Bosch. Analysis and interpretation of data. Ramiro, Landewe, van der Heijde, Dougados, van den Bosch, van Tubergen. REFERENCES 1. Van der Heijde D, Calin A, Dougados M, Khan MA, van der Linden S, Bellamy N. Selection of instruments in the core set for DC-ART, SMARD, physical therapy, and clinical record keeping in ankylosing spondylitis: progress report of the ASAS Working Group. Assessments in Ankylosing Spondylitis. J Rheumatol 1999;26:951 4.

Impairment of Spinal Mobility in AS 1577 2. Carette S, Graham D, Little H, Rubenstein J, Rosen P. The natural disease course of ankylosing spondylitis. Arthritis Rheum 1983;26:186 90. 3. Amor B, Santos RS, Nahal R, Listrat V, Dougados M. Predictive factors for the long term outcome of spondyloarthropathies. J Rheumatol 1994;21:1883 7. 4. Laurent MR, Buchanan WW, Bellamy N. Methods of assessment used in ankylosing spondylitis clinical trials: a review. Br J Rheumatol 1991;30:326 9. 5. Jenkinson TR, Mallorie PA, Whitelock HC, Kennedy LG, Garrett SL, Calin A. Defining spinal mobility in ankylosing spondylitis (AS): the Bath AS Metrology Index. J Rheumatol 1994;21:1694 8. 6. Sieper J, Rudwaleit M, Baraliakos X, Brandt J, Braun J, Burgos-Vargas R, et al. The Assessment of SpondyloArthritis international Society (ASAS) handbook: a guide to assess spondyloarthritis. Ann Rheum Dis 2009;68 Suppl 2: 1 44. 7. Wanders A, Landewe R, Dougados M, Mielants H, van der Linden S, van der Heijde D. Association between radiographic damage of the spine and spinal mobility for individual patients with ankylosing spondylitis: can assessment of spinal mobility be a proxy for radiographic evaluation? Ann Rheum Dis 2005;64:988 94. 8. Van der Heijde D, Dijkmans B, Geusens P, Sieper J, DeWoody K, Williamson P, et al. Efficacy and safety of infliximab in patients with ankylosing spondylitis: results of a randomized, placebocontrolled trial (ASSERT). Arthritis Rheum 2005;52:582 91. 9. Van der Heijde D, Kivitz A, Schiff MH, Sieper J, Dijkmans BA, Braun J, et al. Efficacy and safety of adalimumab in patients with ankylosing spondylitis: results of a multicenter, randomized, double-blind, placebo-controlled trial. Arthritis Rheum 2006;54:2136 46. 10. Ramiro S, van Tubergen A, Stolwijk C, van der Heijde D, Royston P, Landewe R. Reference intervals of spinal mobility measures in normal individuals: the MOBILITY study. Ann Rheum Dis 2015;74:1218 24. 11. Spoorenberg A, van der Heijde D, de Klerk E, Dougados M, de Vlam K, Mielants H, et al. Relative value of erythrocyte sedimentation rate and C-reactive protein in assessment of disease activity in ankylosing spondylitis. J Rheumatol 1999;26:980 4. 12. Van der Heijde D, Landewe R, Feldtkeller E. Proposal of a linear definition of the Bath Ankylosing Spondylitis Metrology Index (BASMI) and comparison with the 2-step and 10- step definitions. Ann Rheum Dis 2008;67:489 93. 13. Ramiro S, van Tubergen A, Stolwijk C, van der Heijde D, Landewe R. Neutral lateral fingertip-to-floor distance can be derived from height [letter]. Ann Rheum Dis 2014;73:1748 9. 14. Ramiro S, Stolwijk C, van Tubergen A, van der Heijde D, Dougados M, van den Bosch F, et al. Evolution of radiographic damage in ankylosing spondylitis: a 12 year prospective follow-up of the OASIS study. Ann Rheum Dis 2015;74:52 9. 15. Moll JM, Wright V. An objective clinical study of chest expansion. Ann Rheum Dis 1972;31:1 8. 16. Moll JM, Liyanage SP, Wright V. An objective clinical method to measure lateral spinal flexion. Rheumatol Phys Med 1972;11:225 39. 17. Brandt J, Listing J, Sieper J, Rudwaleit M, van der Heijde D, Braun J. Development and preselection of criteria for short term improvement after anti-tnf a treatment in ankylosing spondylitis. Ann Rheum Dis 2004;63:1438 44.