Patients with axial spondylarthritis (SpA) who have structural changes in the sacroiliac joints and/or the spine have been classified as having ankylosing spondylitis (AS), while those without such changes are now classified as having nonradiographic axial SpA (nr-axSpA). The differentiating features are incompletely understood.
Data from 100 consecutive patients with axial SpA not treated with tumor necrosis factor antagonists were compared clinically and with laboratory parameters, spinal radiographs, and magnetic resonance imaging (MRI) of the spine. Standardized clinical assessment tools were used to assess health status.
AS was diagnosed in 56 patients and nr-axSpA in 44 patients. Signs of inflammation were significantly higher in patients with AS than in patients with nr-axSpA, with a median C-reactive protein level of 8.0 versus 3.8 mg/liter, a median Ankylosing Spondylitis Disease Activity Score of 2.2 versus 2.8, respectively, and a median amount of spinal inflammatory lesions on MRI of 2.0 versus 0.0, respectively. Significant differences between these 2 groups were seen in sex (76.8% male AS patients versus 31.8% male nr-axSpA patients). Clinical variables did not differ between patients with AS and nr-axSpA (Bath Ankylosing Spondylitis Disease Activity Index, Bath Ankylosing Spondylitis Functional Index, Ankylosing Spondylitis Quality of Life questionnaire, Short Form 36 health survey).
Patients with nr-axSpA were characterized by the low proportion of male patients and the low burden of inflammation compared to patients with AS. While both groups did not differ regarding health status, disease activity, and physical function, they did differ in signs of inflammation; all were higher in patients with AS. Since many patients with nr-axSpA had not developed structural changes after years of symptoms, we propose that those patients should not be regarded as having preradiographic AS but rather as having nr-axSpA.
Patients with axial spondylarthritis (SpA) constitute a partly heterogeneous group of patients with a common genetic background and specific clinical characteristics such as spinal inflammation (1). Sacroiliac (SI) joint involvement is considered the hallmark of the disease and its structural consequences visible on radiographs (using the modified New York criteria) are required for the diagnosis of ankylosing spondylitis (AS), a major subgroup of SpA (2). The spectrum of SpA includes patients with psoriasis, inflammatory bowel disease, and preceding infections (reactive arthritis). However, the majority of patients with axial and peripheral SpA do not have such manifestations; therefore, they are usually classified as having undifferentiated SpA (3). The disease course of axial SpA is highly variable and often characterized by ongoing axial inflammation and radiographic progression associated with restricted mobility of the spine and decreased function (4). However, axial SpA patients that never develop structural changes such as syndesmophytes and ankylosis in the spine have also been described (5).
The tool that quantifies chronic structural changes in the SI joint has limitations in early disease because these structural changes rather reflect advanced disease stages. Active inflammation of the SI joint can be visualized using magnetic resonance imaging (MRI) technology in early and advanced disease stages. Therefore, new classification criteria for axial and peripheral SpA have recently been developed by the Assessment of SpondyloArthritis international Society (ASAS) (6, 7). The ASAS criteria for axial SpA are applied in patients with predominant axial involvement with or without peripheral manifestations, and the ASAS criteria for peripheral SpA are useful when used in patients with arthritis or enthesitis or dactylitis. The ASAS classification criteria for axial SpA require a history of chronic back pain for ≥3 months and an age at onset of <45 years as entry criteria. Next, either sacroiliitis on radiographs or MRI in addition to ≥1 typical SpA feature or the presence of HLA–B27 in addition to ≥2 typical SpA features needs to be present (6, 7). Using this set of criteria, patients can be classified as having established radiographic changes in the SI joint (classified as AS) or as not having developed radiographic changes in the SI joint (classified as nonradiographic axial SpA [nr-axSpA]). The introduction of this novel concept of axial SpA has led to an increasing interest in the examination of patients with nr-axSpA, the presumably early form of axial SpA (8). These patients are best identified by MRI, an established technology to assess inflammation in the axial skeleton (9).
These recently recognized disease stages and the subentities, their nature, their interrelationship, the future course of the disease, and their comparative burden of inflammation are incompletely understood (8). Undifferentiated SpA, although often considered as a provisional diagnosis, has been most frequently studied. For example, in a Brazilian cohort of patients classified as having undifferentiated SpA, 24.3% of >100 patients developed AS within 10 years (10). Therefore, the majority of the patients did not develop radiographic changes, and 23% went into remission. Another study compared 2 patient groups with inflammatory back pain who either had structural changes in the SI joint or did not have structural changes in the SI joint (11). Of interest, there was no evidence that the latter group of patients had a milder disease course or less severe clinical symptoms; however, using the new criteria, patients with axial SpA who have structural changes (AS) or not (nr-axSpA) have not been directly compared.
The primary objective of this study was to systematically compare patients with AS and nr-axSpA using standardized clinical tools, including patient-reported outcomes, C-reactive protein (CRP) level as a biomarker of inflammation, and imaging to assess both spinal inflammation by MRI and structural changes by conventional radiography.
Significance & Innovations
Patients with non-radiographic axial spondylarthritis differ from patients with established ankylosing spondylitis (AS), although the burden of disease is similar.
Patients with non-radiographic axial spondylarthritis are more often women and have less spinal inflammation.
Many patients with non-radiographic axial spondylarthritis do not have developed structural changes after years of symptoms.
We propose that this group should not be regarded as having preradiographic AS but rather as having non-radiographic axial spondylarthritis.
PATIENTS AND METHODS
Consecutive patients diagnosed with axial SpA who had never been treated with tumor necrosis factor (TNF) blockers were included. The disease was diagnosed according to the ASAS classification criteria for axial SpA, which include patients with established AS who fulfill the modified New York criteria (2, 6). Since the ASAS classification criteria for axial SpA were published only 2 years ago, we also included in our cohort patients that had been diagnosed with undifferentiated SpA based on the lack of radiographic changes in the SI joint. Nevertheless, at the time of inclusion, all patients fulfilled the classification criteria for axial SpA. In order to obtain a representative cross-sectional sample of patients with axial SpA, there were no limitations (for example, related to age or disease duration) for the patients to be included in the study. The only exclusion criteria were current or previous therapy with biologic agents, bone neoplasma, osteomyelitis, prior surgery of the spine, and major spinal deformity. Additional exclusion criteria were pregnancy and contraindications for MRI. All patients consecutively attended the outpatient clinic of our hospital (Rheumazentrum Ruhrgebiet) in Herne, Germany, and/or the private offices of 4 cooperating rheumatologists in the surrounding areas between January and December 2010. The local ethics committee approved the study protocol, and all patients gave informed consent for participation.
Data on demographics (sex, age, symptom duration since diagnosis, comorbidities, and the use of concomitant medication) were collected, and standardized assessment tools were applied.
All patients underwent a physical examination and had laboratory parameters measured (CRP and HLA–B27 status [if unknown]). Standardized assessment tools had to be completed by the patients (see below). In addition, the Ankylosing Spondylitis Disease Activity Score (ASDAS) was calculated (12). The published cutoffs between disease activity states are <1.3 for inactive disease, <2.1 for moderate disease activity, <3.5 for high disease activity, and ≥3.5 for very high disease activity (13).
The patients were asked to complete numerical rating scales for global pain (range 0–10) and a visual analog scale for physician's and patient's global assessment of disease activity (range 0–10 cm). The patients completed the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) to assess disease activity (range 0–10, with higher values indicating more active disease) (14) and the Bath Ankylosing Spondylitis Functional Index (BASFI) to assess physical functioning (range 0–10, with higher values indicating worse functioning) (15). Health-related quality of life (HRQOL) outcomes were assessed using the Short Form 36 health survey (SF-36) and the Ankylosing Spondylitis Quality of Life (ASQoL) questionnaire (16, 17). The SF-36 is a generic health status instrument and has 2 summary scores, the physical component summary (PCS) score and the mental component summary (MCS) score. Domain scores range from 0–100, with greater scores reflecting a better health status. Normative values for the German population have been published (18). The ASQoL, an AS-specific instrument to assess HRQOL, has a total range from 0–18, with lower scores reflecting a higher HRQOL.
Radiographs of the SI joint and the lumbar and cervical spine were obtained at baseline for all patients (if no images from the past 6 months were available). Assessments included staging of the initial disease, assessment of chronic spinal changes (using conventional radiographs), and assessment of inflammatory spinal changes (using MRI). Conventional radiography was scored by the modified Stoke Akylosing Spondylitis Spine Score (mSASSS) (19). The mSASSS has a total score ranging from 0–72, with higher scores reflecting more structural changes. The SI joints were scored according to the modified New York criteria (2). MRI examinations were performed on the spinal area with clinical symptoms using the routine imaging protocol of our department, based on the T1-weighted and STIR sequences. Since MRIs were performed for those spinal areas that were clinically affected, we could obtain images either of the upper half of the spine (the entire cervical and at least half of the upper thoracic spine) or the lower half of the spine (at least the lower half of the thoracic spine and the entire lumbar spine). Inflammation on MRI was scored based on the Berlin score grading (20). Briefly, scoring was performed on the basis of a vertebral unit (VU; defined as the region between 2 virtual lines drawn through the middle of each vertebral body) and ranged between 0 (no inflammation) and 3 (inflammation in >50% of the VU). Since MRI examinations for individual patients included only half of the spine, we based our MRI analyses on the mean number of inflamed spinal lesions per patient. All images were scored by an expert (XB) who was blinded to the clinical status of the patients.
Descriptive data are presented as absolute frequencies and percentages when referring to the qualitative variables. Continuous variables are expressed as the median with interquartile range (IQR; range 25–75%) or the mean ± SD, where appropriate. Parameters of inflammation are presented as categorical variables expressed as a proportion, and compared using the chi-square test.
We analyzed the results based on linear regression analysis. The dependent variables were MRI inflammation or CRP level, depending on the analysis. The independent variables were mSASSS baseline status, sex, and CRP level or MRI inflammation, accordingly. Correlation coefficients were calculated with Spearman's rank correlation coefficient. A P value less than 0.05 was considered statistically significant.
Demographic and clinical characteristics.
A total of 100 patients participated in the study. AS was diagnosed in 56 patients and nr-axSpA in 44 patients. The demographic data of the cohort are summarized in Table 1. The patients' demographic and clinical characteristics were largely similar between the groups, except for sex; patients with AS were more often men, and those with nr-axSpA were more often women (76.8% of the AS patients were men and 31.8% of the nr-axSpA patients were men; P < 0.001). Neither the disease duration nor the frequency of typical comorbidities differed significantly between the groups. At baseline, 58% of the patients had pure axial disease and 42% had extraspinal manifestations. A history of extraspinal symptoms was reported by 63% of the patients. There were no differences seen in both groups.
Table 1. Demographic and clinical characteristics of patients with nr-axSpA and AS*
Current treatment with nonsteroidal antiinflammatory drugs (NSAIDs) or sulfasalazine did not differ between nr-axSpA and AS patients. Almost all patients (94%) were taking NSAIDs, either nonselective or cyclooxygenase 2 selective (coxibs), with 51% of the patients taking them continuously. Five patients were taking both conventional NSAIDs and coxibs. All of the 19 patients who were taking glucocorticoids had concomitant peripheral arthritis or SpA associated with inflammatory bowel disease. Methotrexate was taken by 9% and sulfasalazine by 19% of the patients during the study period.
Disease activity, physical function, and HRQOL.
Data on disease activity, physical function, and HRQOL are shown in Tables 2 and 3 and in Figure 1. Patients with established nr-axSpA and AS did not differ much in many clinical variables (clinical parameters of disease activity and functional impairment and HRQOL), but they did differ in terms of CRP levels and ASDAS; both were higher in patients with AS.
Table 2. Clinical characteristics of patients with nr-axSpA and AS*
The median BASDAI was slightly lower in patients with nr-axSpA than in patients with AS (3.6 [IQR 2.40–4.8] versus 4.3 [IQR 2.3–5.7]), but it did not reach the significance level (P = 0.2). The relative percentage of patients with BASDAI values <4 and ≥4 was different between patients with nr-axSpA and AS; BASDAI values <4 were reported by 25 patients with nr-axSpA and BASDAI values ≥4 by 19 patients with nr-axSpA (56.8% and 43.2%, respectively), whereas BASDAI values <4 were reported by 26 patients with AS and BASDAI values ≥4 by 30 patients with AS (46.4% and 53.6%, respectively). The BASDAI did not correlate with CRP level or with the number of inflamed lesions (r = 0.09 [P = 0.37] and r = −0.84 [P = 0.40], respectively).
The median BASFI was slightly lower in nr-axSpA patients than in AS patients (1.5 [IQR 0.7–3.8] versus 2.9 [IQR 1.2–5.4]), but it did not reach the significance level (P = 0.05). The BASFI in the total cohort and also in the subgroups correlated significantly with the BASDAI level (total cohort r = 0.75, AS patients r = 0.76, and nr-axSpA patients r = 0.70), but it did not correlate with CRP level or the extent of inflammation on MRI (r = 0.2 and r = 0.06, respectively). The BASFI levels correlated significantly with age (r = 0.3 [P = 0.001]) and symptom duration (r = 0.39 [P = 0.001]) in patients with AS.
The median level of pain was not statistically different between patients with nr-axSpA and AS (for global pain, 4.0 [IQR 2.0–5.7] versus 5.0 [IQR 2.0–7.0]; P = 0.1 and for back pain, 4.0 [IQR 3.0–6.0] versus 5.0 [IQR 3.0–7.8]; P = 0.1). Physicians gave a better median rating for nr-axSpA patients than AS patients (2.5 [IQR 1.5–3.6] versus 3.2 [IQR 2.0–5.1]; P = 0.06). Patients with nr-axSpA and AS gave similar median ratings of their disease states (4.0 [IQR 1.8–6.1] versus 4.6 [IQR 2.0–6.9]; P = 0.2). Both patient groups showed similar reductions in HRQOL. The median level of ASQoL was not statistically different in patients with nr-axSpA and AS (4.0 [IQR 1.2–10.0] versus 5.5 [IQR 1.2–9.7]; P = 0.1). The median levels of the SF-36 subscales were not statistically different in patients with nr-axSpA and AS (for the PCS, 37.9 [IQR 33.2–46.7] versus 35.1 [IQR 27.1–45.4]; P = 0.7 and for the MCS, 48 [IQR 35.0–55.7] versus 50.3 [IQR 39.6–56.5]; P = 0.1).
HLA–B27 was positive in 87 (88%) of 99 patients. We did not see significant differences in HLA–B27 in sex (81.4% in women versus 92.8% in men; P = 0.16) and in the subgroups studied (86.4% in nr-axSpA patients versus 89.1% in AS patients; P = 0.8).
The median CRP level was higher in patients with AS compared to patients with nr-axSpA (8.0 mg/liter [IQR 4.0–14.0 mg/liter] versus 3.8 mg/liter [IQR 1.75–7.9 mg/liter]) (Table 3 and Figure 2). There was no strong correlation between CRP level and the number of inflamed lesions in patients with nr-axSpA and in patients with AS (r = −0.02 and r = 0.3, respectively).
Based on the structural changes that were seen on conventional pelvic or lumbar radiographs, 56 patients fulfilled the modified New York criteria for AS. Expectedly, the median mSASSS was significantly higher in AS patients compared to nr-axSpA patients (3.0 [IQR 0–18.2] versus 1.05 [IQR 0–3]; P = 0.007); the median mSASSS of all patients was 2.0 (IQR 0–6.4) (Table 3 and Figure 2). There was a correlation between mSASSS score and CRP level (r = 0.3, P = 0.006).
MRI scans of the upper half of the spine (92% of which covered the region from C2/3 to L2/L3) were available for 25 patients, while 75 patients had MRIs of the lower half of the spine (of which 97% covered the region from Th4/5 to L5/S1). The majority of the patients (60%) had spinal inflammation as assessed by MRI (21 patients with nr-axSpA [47%] versus 39 patients with AS [69%]). Patients with nr-axSpA had less spinal inflammation, but they also showed a lesser extent of inflammatory lesions as compared to AS patients (median inflamed lesion per patient 0.0 [IQR 0–1] in nr-axSpA patients versus 2.0 [IQR 0–5] in AS patients). In patients with spinal inflammation, elevated CRP levels were seen more frequently in AS patients (97.4%) versus nr-axSpA patients (61.9%; P < 0.001). Of the 31 patients with nr-axSpA and a normal CRP level, 15 had a positive MRI (48.4%), whereas of the 17 patients with AS and a normal CRP level, 11 had a positive MRI (64.7%; P = 0.24). Therefore, although the majority of patients with nr-axSpA did not have an elevated CRP level, one-half of the patients with a normal CRP level showed inflammatory lesions in the spine (Table 4).
Table 4. Relationship between elevated CRP level and positive MRI in patients with nr-axSpA and AS*
nr-axSpA (n = 44)
AS (n = 56)
Values are the number (percentage). CRP = C-reactive protein; MRI = magnetic resonance imaging; nr-axSpA = non-radiographic axial spondylarthritis; AS = ankylosing spondylitis.
MRI positive, CRP positive
MRI positive, CRP negative
MRI negative, CRP positive
MRI negative, CRP negative
In this study, we confirm the results of earlier studies (11) that showed that patients with established AS and nr-axSpA do not differ greatly in many clinical variables. However, this study also clearly shows that these groups are different in several other important aspects. While the groups did not differ regarding most patient-reported outcome parameters such as the BASDAI and HRQOL, they did differ in the extent of inflammation and radiographic damage. Indeed, there was more inflammation and radiographic damage was more extensive in patients with AS. Another major difference between patients with nr-axSpA and AS was the higher proportion of women in the nr-axSpA subgroup, while more men were affected in the AS subgroup than women (see below). Of interest, no major differences in physical functioning as measured by the BASFI between patients with established AS and nr-axSpA were found in this study. This differed from the German Spondyloarthritis Inception Cohort (GESPIC) data, in which patients with nr-axSpA had better functional values than those with AS. In contrast to our cohort, where 68% of the patients had a disease duration >5 years, these patients had a disease duration of <5 years (11). However, the reason for this difference is not entirely clear since it is known that function is mainly influenced by disease activity and structural changes in AS (21). Two aspects may be relevant in resolving this: the mSASSS does not assess all regions where structural changes may occur, and the BASFI has been evaluated in AS but not in nr-axSpA.
The fact that the clinical manifestations of the patients with nr-axSpA and AS were similar and that the HLA–B27 prevalence was not different between the groups strongly supports the diagnosis of axial SpA in the patients of this study.
Since the new ASAS classification criteria for axial SpA have been published only recently, our knowledge about the relatively new subgroup nr-axSpA is still limited. This study adds to a better understanding of this subgroup by suggesting that the majority of these patients probably cannot be characterized as having preradiographic AS, but rather an abortive form characterized by typical SpA symptoms much like established AS, but with a female predominance, better function, and less structural changes. However, it remains to be seen how many of these patients will develop definite AS over time. The data from the GESPIC cohort have suggested that the proportion of patients fulfilling the New York criteria within 2 years was ∼10% (22). The clinical comparisons of patients with nr-axSpA and AS were also consistent with the GESPIC data (11) and the Spanish inception cohort Registro Español de Espondiloartritis de la Sociedad Española de Reumatología (REGISPONDER) (23).
There are only limited data on the extent of inflammatory spinal lesions in patients with nr-axSpA. In one study of axial SpA patients who had a short symptom duration, 33% showed inflammatory lesions in the spine as detected by MRI, while all had active sacroiliitis (24). However, this result cannot be compared because it was from a therapeutic trial, while our study included all consecutive patients. Nevertheless, it is remarkable that more than one-half of the nr-axSpA patients had some spinal inflammation as detected by MRI. Of note, for pragmatic reasons, only the predominantly symptomatic area in the upper or lower half of the spine was examined in this study.
The finding of an increased proportion of women in the nr-axSpA group is consistent with the well-established fact that male AS patients have worse radiographic outcomes (25). For example, in the GESPIC, male sex and CRP levels were associated with structural radiographic changes (22). The results from the study of a Brazilian cohort (10) of patients with undifferentiated SpA who were predominantly men is possibly explained by the way the term undifferentiated SpA has been used recently: peripheral arthritis or enthesitis was included.
Despite the cross-sectional design of this study, we think it is likely that the relatively large group of female patients without structural changes after years of symptoms may not develop such changes at all. Of course, only followup examinations will show whether and how many of the nr-axSpA patients in this study will develop AS in the future. It is well known that patients with axial SpA may progress rather slowly and radiographic changes may appear only relatively late in the disease course; this has been identified as one of the reasons for the diagnostic delay of 5–10 years in AS (26). Nevertheless, a subgroup of patients in a Dutch cohort developed structural changes rather early in the disease course; 20% of the patients with a history of inflammatory back pain of <2 years in duration had chronic structural changes in the SI joints (27). Although low-grade radiographic sacroiliitis has been proposed as a prognostic factor for the development of AS in patients originally classified as having undifferentiated SpA, only a minority of patients with nr-axSpA seem to develop definitive AS (28).
Therefore, we suggest not to regard patients with nr-axSpA as having preradiographic AS but rather as an axial SpA subgroup that is less prone to develop new bone formation. However, our data also justify the inclusion of patients with nr-axSpA in the recent recommendations for anti-TNF therapy since the burden of disease was quite comparable in patients with nr-axSpA and established AS (29).
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 published. Dr. Kiltz 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. Kiltz, Baraliakos, Braun.
Analysis and interpretation of data. Kiltz, Baraliakos, Karakostas, Flörecke, Braun.
ROLE OF THE STUDY SPONSOR
Pfizer, Germany had no role in the study design or in the collection, analysis, or interpretation of the data, the writing of the manuscript, or the decision to submit the manuscript for publication. Publication of this article was not contingent upon approval by Pfizer, Germany.