Ankylosing spondylitis and axial spondyloarthritis: recent insights and impact of new classification criteria

Abstract

Development of the Assessment in Spondyloarthritis International Society (ASAS) classification criteria for axial spondyloarthritis (SpA) was one of the major breakthroughs in the field over the past decade. Despite some concerns related to the specificity of the criteria, they stimulated research into the early stage of the disease. This resulted in major advances in the understanding of the course of the disease, revealing predictors of progression, improvement in early diagnosis and treatment in axial SpA. In this review, we summarize the recent developments resulting from the introduction of the ASAS classification criteria for axial SpA and the implications for research and clinical practice.

Introduction

A group of chronic inflammatory diseases is subsumed under the generic term ‘spondyloarthritis’ (SpA) (Figure 1). They share common clinical, genetic and pathophysiological features,^1–3 such as involvement of the axial skeleton (sacroiliac joints and spine), characteristic peripheral manifestations (dactylitis, enthesitis and asymmetric mono- or oligoarthritis predominantly affecting the lower extremities) and particular extra-articular manifestations, such as anterior uveitis, psoriasis and inflammatory bowel disease,^1,2 as well as association with the major histocompatibility complex class I human leucocyte antigen-B27 (HLA-B27).^1,4,5

Figure 1.

The spectrum of spondyloarthritis (SpA) and overlap (cross-sectional and longitudinal) between different SpA forms.

IBD, inflammatory bowel disease.

Depending on the leading manifestation, two major SpA groups are defined: axial SpA with predominant involvement of the sacroiliac joints and/or spine, and peripheral SpA with predominant peripheral manifestations such as arthritis, enthesitis and dactylitis (Figure 1). In this review, we will focus on recent developments resulting from the introduction of the Assessment in Spondyloarthritis International Society (ASAS) classification criteria for axial SpA,^6,7 and the implications for understanding, diagnosing and treating the disease.

Current classification criteria

In the early years of this century, the concept of axial SpA as one disease with two subsets has arisen: with radiographic changes in the sacroiliac joints (ankylosing spondylitis [AS] or radiographic axial SpA) and without (nonradiographic axial SpA [nr-axSpA]).^8–11 With the introduction of new and effective treatment options, such as tumour necrosis factor (TNF) blockers,^12–16 it has become crucial to identify the disease as early as possible to give patients these opportunities in the early course of the disease. As the commonly used classification criteria, the modified New York (mNY) criteria for AS,¹⁷ were not able to capture individuals with early disease (i.e. without structural damage in the sacroiliac joints on X-rays) and evidence of the ability of magnetic resonance imaging (MRI) to detect active inflammation of the spine and sacroiliac joints early in the disease course emerged,^11,18–21 the new ASAS classification criteria were developed and published in 2009.^6,7 They enabled classification of patients as having axial SpA even before relevant structural damage in the sacroiliac joints occurred. According to the new ASAS classification criteria,⁶ at an early disease stage this is possible either by using the ‘imaging’ arm with signs of active sacroiliitis in MRI with at least one other SpA feature or by using the ‘clinical’ arm, where the presence of HLA-B27 is mandatory with an additional two or more SpA features (Figure 2). Furthermore, classification is also possible at a later stage when already advanced chronic changes of the sacroiliac joints can be seen in the conventional radiograph according to the mNY ¹⁷ and at least one other SpA feature is present. These criteria showed a sensitivity of 82.9% and a specificity of 84.4%, clearly outperforming the original European Spondyloarthropathy Study Group²² and Amor criteria.²³ Focusing on the so-called ‘imaging arm’ alone, the ASAS classification criteria showed a sensitivity of 66.2% and a specificity of 97.5%. Moreover, when these criteria were applied to other cohorts and tested against the rheumatologist’s diagnosis as the gold standard, acceptable results for sensitivity and specificity were shown. These were comparable to the initial results with a sensitivity between 68% and 87% and a specificity between 62% and 95%.^24–27 At follow up of the initial cohort⁶ after 3–5 years by the treating rheumatologist, the positive predictive value that patients who initially fulfilled the criteria would still be diagnosed with axial SpA was excellent with 93.3%.²⁸ These classification criteria have been accompanied by a new definition of inflammatory back pain,²⁹ a new definition of active sacroiliitis on MRI³⁰ and new classification criteria for peripheral SpA and SpA in general.²

Figure 2.

Assessment in Spondyloarthritis International Society (ASAS) classification criteria for axial spondyloarthritis (SpA).⁶

*Sacroiliitis on imaging refers to definite radiographic sacroiliitis according to the modified New York criteria or active sacroiliitis on magnetic resonance imaging according to the ASAS consensus definition.

CRP, C-reactive protein; HLA-B27, human leucocyte antigen-B27; IBD, inflammatory bowel disease; NSAIDs, nonsteroidal anti-inflammatory drugs.

Controversial appraisal of the ASAS classification criteria of axial SpA

The development of the new axial SpA classification criteria⁶ was a significant step towards a better understanding, definition and unification of the disease concept. Undoubtedly, these classification criteria draw attention to an early stage of the disease and stimulated interventional trials that resulted in approval of effective drugs for patients who had had no approved therapeutic options in the past. However, the concept of axial SpA as one disease and the terminology used, especially when it comes to the term ‘nonradiographic axial SpA’, is still a matter of debate.^3,31–33

First, it has to be pointed out that the assessment of conventional X-rays of the sacroiliac joints is challenging³⁴ and high relevant inter- and intra-observer variations have been reported.^35–38 The inter-observer variability was even more pronounced when the grading from local readers and trained central readers was compared,^39,40 but was also evident between central readers.^41,42 The highest variability was observed in the evaluation of grades 1 and 2.³⁴ This may be clinically important, as the difference in the grading of one sacroiliac joint between grade 1 and grade 2 can change the diagnosis from nr-axSpA to AS and vice versa, with potentially prognostic and therapeutic implications. Surprisingly, the reproducibility of assessing radiographic sacroiliitis could not be substantially improved by training.³⁴ Interestingly, in a follow up of the previously described ASAS cohort, which was used to develop the ASAS classification criteria, a net progression from nr-axSpA at the baseline visit to AS of 5% was seen after a mean follow up of 4.4 years (range: 1.9–6.8).⁴³ However, cross-tabulation revealed that more than every second patient (36 out of 62), who was assessed mNY positive at baseline was classified as mNY negative at the radiological follow up, while 54 out of 295 mNY negatives at baseline ‘progressed’ to the radiological form at follow up.⁴³ The explanation of Sepriano and colleagues is that the subtle radiographic changes cannot reliably be discriminated from measurement errors.⁴³

Second, the term nr-axSpA indicates that no radiographic changes are seen on the conventional radiograph, but this does not always have to be the case. There can be minor but obvious radiographic changes in the sacroiliac joints (e.g. sclerosis and few small erosions of the right sacroiliac joint = grade 2 and suspicious changes in the left sacroiliac joint = grade 1), which are just not enough to fulfil the radiological part of the mNY criteria and therefore would still be labelled as mNY negative and thereby classified as nr-axSpA, although unambiguous changes are present. On the other hand, it can be the case – even though rare – that in the absence of radiographic sacroiliitis there are clear radiographic changes of the spine attributable to AS (e.g. syndesmophytes or complete ankylosis). In this case, the classification term would be still nr-axSpA.

The third point is related to the use of classification criteria as an aid in the diagnostic process. The sensitivity and specificity of the ASAS classification criteria for axial SpA were estimated to be 82.9% and 84.4%, respectively.⁶ This was different for separate analysis of the ‘imaging’ arm (sensitivity of 66.2% and specificity of 97.3%)⁶ and the ‘clinical’ arm (sensitivity of 56.6 % and specificity of 83.3%).⁴⁴ Thus, the multi-arm construct of the ASAS criteria increases the sensitivity of the criteria, at the price of a decrease in specificity (as compared with the imaging arm). In the absence of diagnostic criteria for axial SpA, the ASAS classification criteria may be used by physicians in the clinical routine to diagnose axial SpA in their patients. Hence, it has to be stressed, that just ‘ticking boxes on a checklist’ of the ASAS classification criteria for making a diagnosis means an inappropriate use of these criteria,⁴⁵ which, like all classification criteria, should be applied to patients with an established diagnosis only. Classification criteria, in contrast to the diagnostic approach, give a ‘yes’ or ‘no’ answer with a certain sensitivity and specificity (instead of disease probability with a diagnostic approach), do not consider negative results of diagnostic tests and do not take other explanations of symptoms/differential diagnoses into account. As a result, a young patient with unspecified back pain, who is HLA-B27 positive, could be misdiagnosed with axial SpA because of other subjective items like a ‘good response to nonsteroidal anti-inflammatory drugs (NSAIDs)’ or heel pain suggesting enthesitis, for instance. The same is, however, also true for the imaging arm of the criteria, which requires the presence of bone marrow oedema on the MRI scan of the sacroiliac joints in combination with one additional SpA parameter.

Finally, in the first years after publication of the ASAS classification criteria, the specificity of the ‘clinical’ arm was the main target of criticism.^31,32 In the last years, the focus changed to the ‘imaging’ arm and the specificity of ‘positive’ MRI,^46–50 that led to an update of the ASAS definition of active sacroiliitis on MRI,⁵¹ stressing the importance of the contextual interpretation of bone marrow oedema as ‘highly suggestive of SpA’ if used for classification. Despite the limitations related to specificity, MRI still remains the most important imaging method, especially for the early diagnosis of axial SpA.

In order to address these concerns, to test the performance of the ASAS classification criteria in a large prospective cohort of patients including those recruited in North America, and finally to modify the classification criteria (if needed) in order to increase the specificity, ASAS and the SpA Research and Treatment Network (SPARTAN) have recently announced the CLASSIC (Classification of Axial SpondyloarthritiS Inception Cohort) study that is going to start recruitment in 2018.⁵²

Similarities and differences between nr-axSpA and AS

nr-axSpA and AS share common epidemiological, genetic and clinical characteristics^{4,16,27,53–56} that support the concept of axial SpA as one disease with two stages.^1,53 The prevalence of nr-axSpA is comparable with the prevalence of AS, each with 0.3–0.6% in the population.^4,27 The ratio of nr-axSpA to AS in patients first diagnosed with axial SpA is close to 1:1.^57–60 Cohort studies suggest similar characteristics for age, prevalence of HLA-B27, occurrence of peripheral manifestations and extra-articular symptoms as well as disease activity parameters (based on patient-reported outcomes) and burden of the diseases.^9,55,56 However, AS is characterized by a significantly higher prevalence of men and higher C-reactive protein (CRP) levels in comparison with nr-axSpA.^9,55,56 This might be explained by the fact that female patients and patients with lower activity of inflammation (low CRP, less inflammation on MRI) are less likely develop radiographic changes, and therefore might sustain at the nonradiographic stage.⁵³ This implies the existence of a certain subgroup of patients with axial SpA with a rather mild, nonprogressive disease course.^1,53 Approximately 12% of the patients with nr-axSpA progress to AS over a period of 2 years, with elevated CRP levels and active sacroiliitis on MRI being the strongest predictors for this progression^41,42. Concerning the response to anti-TNF-alpha treatment, there seems to be no significant difference between the two groups when the level of active inflammation at baseline is comparable.^16,54,56,61

Treatment of axial SpA

Axial SpA is at present a ‘hot topic’ in the field of rheumatology and new therapeutic options are in the pipeline and will be available in the near future. The exact role of new therapies in the treatment algorithm in axial SpA still has to be defined.

NSAIDs and TNF blockers are well-known effective treatment options in patients with axial SpA. With secukinumab, a monoclonal antibody against interleukin-17A (IL-17A), a new compound has been available since 2015 as an important treatment option for patients with ankylosing spondylitis, while phase III studies for nr-axSpA are still ongoing.

In the last years, the major treatment target in SpA was defined as clinical remission/inactivity of musculoskeletal (arthritis, dactylitis, enthesitis, axial disease) and extra-articular manifestations.⁶² The current versions of the leading international management recommendations for axial SpA (ASAS and the European League Against Rheumatism recommendations from 2017⁶³) and treatment recommendations (from the American College of Rheumatology/Spondylitis Association of America /SPARTAN from 2016⁶⁴) are similar (Figure 3).

Figure 3.

Treatment algorithm for axial spondyloarthritis (SpA) based on the Assessment in Spondyloarthritis International Society/European League Against Rheumatism⁶⁵ and American College of Rheumatology/Spondylitis Association of America/SpA Research and Treatment Network⁶⁴ recommendations.

bDMARDs, biologic disease-modifying antirheumatic drugs; csDMARDs, conventional synthetic disease-modifying antirheumatic drugs; IL-17A, interleukin-17A; NSAIDs, nonsteroidal anti-inflammatory drugs; TNF, tumour necrosis factor.

According to these recommendations, first-line therapies for patients with symptomatic axial SpA are NSAIDs, including selective cyclooxygenase-2 (COX-2) antagonists, together with exercise/physiotherapy and education of the patients.⁶⁵ Therapy with conventional synthetic disease-modifying antirheumatic drugs (csDMARDs) such as sulphasalazine may have some beneficial effect in patients with peripheral joint involvement, but are not effective in the majority of patients with axial involvement.^66–68 If these patients have a poor response to NSAIDs, contraindications or intolerance for NSAIDs, the only effective treatment currently available is therapy with biological DMARDs (bDMARDs): TNF-alpha inhibitors⁶⁵ or with secukinumab, the recently introduced monoclonal antibody against IL-17A.⁶⁵

Currently there are five TNF inhibitors (infliximab,¹² etanercept,¹⁴ adalimumab,¹³ golimumab¹⁵ and certolizumab pegol¹⁶) and one monoclonal antibody against IL-17A (secukinumab⁶⁹) approved for the treatment of patients with AS in the EU, the USA and most other parts of the world. In the EU, four TNF inhibitors (etanercept,⁷⁰ adalimumab,⁷¹ golimumab⁷² and certolizumab pegol¹⁶) also have approval for the treatment of nr-axSpA after failure of the standard treatments, whereas the label is restricted to such patients with objective signs of active inflammation: elevated CRP or active sacroiliitis on MRI (Figure 4).

Figure 4.

(a) Approval status and year of approval for biological disease-modifying antirheumatic drugs (bDMARDs) by the European Medicines Agency for axial spondyloarthritis (SpA) in the EU and a number of other countries. (b) Approval status and year of approval for bDMARDs by the US Food and Drug Administration for SpA in the USA.

The European Medicines Agency (EMA) further demanded a therapy-withdrawal trial design to investigate whether patients with nr-axSpA can achieve a disease state of ‘drug-free remission’ after a treatment cycle like an ‘induction therapy’. These trials are still ongoing.

Contrary to the EMA decision, the US Food and Drug Administration rejected the approval for TNF inhibitors for nr-axSpA, due to questions about the specificity of the ASAS criteria and the natural history of the disease entity, hinting that a significant proportion of these patients may show spontaneous remission and therefore might not require the costly and potentially harmful anti-TNF therapy.⁷³

There is some evidence that NSAIDs, in particular celecoxib, might possess not only a symptomatic efficacy but also disease-modifying properties in AS, retarding the progression of structural damage (syndesmophytes and ankylosis) in the spine if taken continuously.⁷⁴ However, for diclofenac, a nonselective COX inhibitor, such an effect was not proven in a recently published trial.⁷⁵ The data on the effect of TNF inhibitors on radiographic spinal progression in ankylosing spondylitis, despite their high anti-inflammatory efficacy, remains controversial. While some studies could not show a retardation of radiographic spinal progression in AS over a period of 2 years^76–78 or 4⁷⁹ years, three observational studies suggested that it may take more than 4 years to detect such an effect^80–82 and that early (within the first 5 years or 10 years of the disease) initiation of anti-TNF therapy might play a key role.^81,82 Also, achievement of sustained remission with TNF inhibitors was associated with retardation of radiographic spinal progression within a 2-year time frame as shown in a recent analysis of an observational cohort study.⁸³ However, no prospective controlled trials have been conducted so far to confirm these observations.

Concept of axial SpA in the context of other inflammatory diseases

In many other systemic inflammatory-rheumatic diseases, patients classified as one generic diagnosis might present with relevant prognostic variations among subgroups. This is obviously the case for rheumatoid arthritis, with or without anti-citrullinated protein antibodies, where no one would separate this heterogeneous patient cohort and limit the existing effective treatment options, for example, for patients without radiographic erosions. It is generally true that capturing inflammatory disease (such as rheumatoid arthritis, psoriatic arthritis, Crohn’s disease, etc.) at an early stage implies a broader spectrum of possible outcomes including both milder forms with no or only slow progression of structural damage and more severe forms with a rather more rapid progression of structural damage.

A higher heterogeneity of the nr-axSpA subgroup compared with AS is also acknowledged in the wording of the current approval of TNF inhibitors for nr-axSpA in the EU: in addition to the lack of response to NSAIDs, objective signs of inflammation (elevated CRP and/or positive MRI) are required to start the treatment. So far, only one phase III trial with a bDMARD (a TNF inhibitor certolizumab pegol) has been performed in patients with axial SpA (both nonradiographic and radiographic) fulfilling the ASAS classification criteria. In this study, all patients were required to have objective signs of inflammation (positive CRP or active sacroiliitis on MRI). As a result, the response rates in both subgroups (nr-axSpA and AS) were very similar supporting again the concept of axial SpA as one entity.¹⁶ We expect that the number of clinical trials including the entire population of axial SpA will increase in future.

Conclusion

Development of the ASAS classification criteria for axial SpA has been one of the major breakthroughs in the field over the past decade. Despite some concerns related to the specificity of the criteria and some uncertainties around new nomenclature, the criteria stimulated research related to the early stage of the disease. This resulted in major advances in understanding the course of the disease, revealing predictors of progression, and improvement in the early diagnosis and treatment of axial SpA.