Key points

Introduction

Axial spondyloarthritis (axSpA) is a chronic, inflammatory disease characterized by back pain, spinal ankylosis, peripheral arthritis, and extra-articular manifestations. AxSpA has generally been considered an autoinflammatory disease in the past, because it was thought to lack the autoantibodies that characterize autoimmune rheumatic diseases, such as rheumatoid arthritis and systemic lupus erythematosus. In those conditions, preclinical detection of diagnostic autoantibodies has made very early detection of incipient disease a possibility, although disease does not develop in all those individuals with autoantibodies. Identification of preclinical and early axSpA is therefore a more complex challenge.

Introduction

Axial spondyloarthritis (axSpA) is a chronic, inflammatory disease characterized by back pain, spinal ankylosis, peripheral arthritis, and extra-articular manifestations. AxSpA has generally been considered an autoinflammatory disease in the past, because it was thought to lack the autoantibodies that characterize autoimmune rheumatic diseases, such as rheumatoid arthritis and systemic lupus erythematosus. In those conditions, preclinical detection of diagnostic autoantibodies has made very early detection of incipient disease a possibility, although disease does not develop in all those individuals with autoantibodies. Identification of preclinical and early axSpA is therefore a more complex challenge.

Why should axial spondyloarthritis be diagnosed early?

The estimated prevalence of axSpA is between 0.2% and 1.2% in white European populations. The diagnosis is often delayed by up to 8 years, mainly because sacroiliitis, which is considered a hallmark of ankylosing spondylitis (AS), is not visible on plain radiographs in the early stages of the disease. In recent years, the challenge of early diagnosis of axSpA has become a high priority area of research. Although nonsteroidal anti-inflammatory drugs (NSAIDs) and physiotherapy are the cornerstones of management, tumor necrosis factor inhibitors (TNFi) are effective in patients who are resistant to standard therapy and are more effective in younger patients and those with shorter disease duration. It has also been shown that TNFi can suppress the inflammation detected on magnetic resonance imaging (MRI). Recent studies have demonstrated that TNFi agents can slow radiographic progression of the disease and the effect is most pronounced when treatment is started earlier in the course of disease. This has added to the imperative for early identification of patients with axSpA.

What is “early ankylosing spondylitis”? The concept of axial spondyloarthritis

Historically AS has been defined by the modified New York classification criteria, which require the presence of radiographic sacroiliitis. More recently the use of MRI for earlier detection of inflammation in the sacroiliac joints has led to the identification of patients with features of axSpA who do not fulfill the modified New York criteria, and this has introduced the terminology of nonradiographic axSpA (nr-axSpA). Criteria have been developed for the classification of patients with axSpA including those with nr-axSpA. It has been proposed that nr-axSpA may represent an early form of AS. This concept is supported by evidence that some patients do progress to AS over time. Data from early axSpA cohorts have demonstrated that over a 2-year period, approximately 10% of patients with undifferentiated SpA or nr-axSpA progress to radiographic AS. However, several patients did not progress to AS over the duration of these studies. This finding, along with the identification of genetic and gender differences between AS and nr-axSpA, has led to the concept of nr-axSpA as a distinct disease entity and not just early AS. A lower frequency of HLA-B27 in nr-axSpA than in AS has been observed in the German Spondyloarthritis Inception Cohort (GESPIC) and in two trials of adalimumab in nr-axSpA, one performed in Germany and one an international multicenter trial. In GESPIC the rates of HLA-B27 carriage were 72.6% in nr-axSpA and 84.3% in AS. Gender differences between the two groups have been demonstrated in three German studies with the percentage of males ranging from 31% to 43% in nr-axSpA compared with 65% to 77% in AS. In general, disease activity and burden of symptoms did not differ, although C-reactive protein levels were lower in nr-axSpA.

Early recognition of ankylosing spondylitis in primary care: the “back pain population”

The first presenting feature in AS is usually low back pain, which is typically characterized by such features as early morning stiffness, relief with exercise, lack of improvement with rest, and nocturnal pain. These constellations of clinical features are key elements of inflammatory back pain (IBP). A description of IBP in association with AS was first published in 1977 and several criteria for IBP have been proposed since then. The sensitivity and specificity of these IBP criteria for axSpA are broadly similar. For example, the Assessment of Spondyloarthritis International Society (ASAS) criteria for IBP have a sensitivity of 79.6% and specificity of 72.4%. However, chronic low back pain is very common in the general population and only approximately 5% of such patients are considered to have SpA. Inflammatory low back pain is itself not uncommon: the current prevalence of IBP in the US population as a whole is estimated at 5%. Various strategies and referral algorithms have been proposed to improve the timeliness of referral of patients with suspected SpA from primary care to rheumatologists. Because axSpA is a chronic disease and greater than or equal to 95% of patient are symptomatic by the age of 45 years, the target population for early diagnosis should be those with back pain for more than 3 months and age of onset less than 45 years. Several combinations of parameters have been proposed for recognition of patients with early SpA, which commonly include presence of HLA-B27, IBP, family history of SpA, raised acute-phase reactants, characteristic extra-articular features, suspicion of sacroiliitis on imaging, and response to NSAIDs. Although the optimal algorithm has not yet been defined, using the presence of either IBP or HLA-B27 in this population as a trigger for referral to a rheumatologist has proved effective in several studies in identifying patients with axSpA. It should, however, be noted that although approximately 8% of the white population are HLA-B27 positive, only a small proportion (<5%) develop axSpA and the use of HLA-B27 alone as a screening test is not appropriate.

Imaging for earlier recognition of spondyloarthritis

Although MRI has revolutionized the imaging of axSpA, the specificity of some MRI lesions has not been clearly defined. Sacroiliac MRI lesions considered characteristic of SpA are bone marrow edema, capsulitis, synovitis, and enthesitis. In the spine, corner-based inflammatory lesions, end-plate lesions, diffuse vertebral body lesions, and spinous process bone marrow edema lesions are characteristic. However, an evaluation of 174 patients with back pain caused by SpA, degenerative arthritis, or malignancy demonstrated that any of the spinal lesions listed may be present in degenerative arthritis or malignancy. The spinal inflammatory lesions most suggestive of axSpA, as concluded by ASAS/OMERACT, are corner-based inflammatory lesions, anterior/posterior spondylitis in three or more sites, and end-plate lesions. With regard to sacroiliitis, a definition of a positive MRI has been proposed by the same group and is widely used. When applied to a cohort of 29 patients with early IBP and 18 control subjects, the ASAS definition of sacroiliitis gave a sensitivity of 79%, specificity of 89%, and likelihood ratio of 7.1. The MORPHO study found that the ASAS definition of a positive MRI using bone marrow edema only had a sensitivity of 67% and specificity of 88%. In this study, inclusion of erosions in the definition of a positive MRI improved sensitivity. Studies that have included needle biopsy of sacroiliac joints to confirm histologic sacroiliitis have found the specificity of MRI in confirming early sacroiliitis to be high (100%) but sensitivity was low (37.7%). The concept of preclinical (asymptomatic) radiologic disease has not been investigated in detail. However, sacroiliac joint lesions are observed in control subjects with no back pain and in one study of 35 patients with AS and 35 healthy control subjects, inflammatory SpA-like lesions occurred in 26% of control subjects. Asymptomatic radiographic sacroiliitis has been observed in up to 32% of patients with inflammatory bowel disease (IBD). Further prospective studies are required to investigate long-term outcomes in young patients presenting with back pain and in asymptomatic individuals with apparent inflammatory lesions on MRI. The definition of an inflammatory lesion may need to be reconsidered in the future.

Biomarkers for earlier recognition of spondyloarthritis

An association between AS and IBD has been recognized for many years. Evidence of intestinal inflammation, which may be subclinical, is present in up to 65% of patients with SpA. IBD is associated with a variety of serologic antibodies that suggest loss of tolerance to a subset of commensal microorganisms. These include anti– Saccharomyces cerevesiae antibodies (ASCA), antineutrophil cytoplasmic antibodies, anti-I2 (associated with anti- Pseudomonas activity), anti– Escherichia coli outer membrane porin C, and anti-flagellin (anti-CBir1) antibodies. Circulating antibodies may be useful in distinguishing IBD from healthy control subjects and from other gastrointestinal disorders. For example, sensitivity of ASCA for IBD ranges from 31% to 45% and specificity from 90% to 100%. The role of circulating antibodies in the pathogenesis of IBD is not understood but it is generally accepted that they reflect an aberrant immune response rather than the recognition of specific or pathogenic bacteria. The presence of these antibodies in AS has been investigated in several studies. Patients with AS have been shown to have higher titers than control subjects of anti-I2, ASCA IgG, and total ASCA. Anti-CBir1 (anti-flagellin) antibodies are also elevated in patients with AS compared with patients with mechanical back pain and are associated with elevation of acute phase reactants in patients who have AS without coexisting IBD. It is possible that these patients have subclinical IBD and further studies are required to answer this question. In IBD, the presence of certain antibodies has clinical significance; in Crohn disease, ASCA are associated with younger age of onset and small bowel involvement, whereas antineutrophil cytoplasmic antibodies are associated with ulcerative colitis–like features. Anti-CBir1 antibodies are associated with fibrostenosing disease and complicated small bowel disease. The frequency of subclinical bowel inflammation in axSpA has suggested that a possible triggering event in the immunopathogenesis of axSpA is gut microbial dysbiosis and altered immune response to gut flora. Future work might investigate whether the presence of serum antibodies, such as anti-CBir1, in asymptomatic individuals might predict development of axSpA, or might provide prognostic information in early AS.

Other areas of interest for biomarkers in axSpA are the Wnt/β-catenin and bone morphogenic protein signaling pathways. There is evidence that abnormalities of Wnt/β-catenin signaling may contribute to spinal fusion in axSpA. Sclerostin is an antagonist of Wnt/β-catenin signaling, contributing an inhibitory signal in normal bone homeostasis. The expression of sclerostin is impaired in osteocytes of patients with AS and lower serum levels of sclerostin are seen in patients with axSpA compared with patients with osteoarthritis and healthy control subjects. Enhancement of Wnt/β-catenin signaling has been demonstrated in a mouse model of peripheral and spinal ankylosis. Bone morphogenic protein signaling is a key molecular pathway in bone formation. Noggin is an antagonist of bone morphogenic protein signaling and in a mouse model of ankylosing enthesitis, systemic overexpression of noggin prevented ankylosis.

Recent work has demonstrated that IgG autoantibodies to noggin and sclerostin are present in healthy individuals, but are present at significantly higher levels in patients with AS. Preclinical autoantibody detection has proved valuable for identification of incipient rheumatoid arthritis and systemic lupus erythematosus. It remains to be determined whether immune complexes consisting of autoantibodies to sclerostin and noggin might similarly prove valuable in the identification of preclinical axSpA.

Genetic testing for earlier recognition of spondyloarthritis

AS is a highly heritable disease. A large proportion of the genetic risk is defined by HLA-B27, the major genetic variant of the disease. Approximately 8% of white populations are positive for HLA-B27 and 80% to 90% of patients with AS carry HLA-B27. However, only a small proportion (1%–5%) of individuals positive for HLA-B27 develops AS. Testing healthy individuals for HLA-B27 is therefore not a useful method for identifying patients with preclinical AS. Could other genes contributing to the heritability of AS improve the predictive use of genetic testing? Strong associations have been identified with several genes including IL23R and ERAP1 . The association of IL23R with AS has been confirmed in several studies of white populations and provides evidence for the role of the T-helper 17 (Th17) lymphocyte pathway in AS. Similar studies have confirmed the association of AS with ERAP1 , although the mechanism by which ERAP1 influences AS susceptibility is not clear. In a mouse model of inflammatory arthritis with enthesitis, it was demonstrated that the entheses and aortic root contain a resident population of interleukin (IL)-23R+ T cells, which allow the tissue to respond to IL-23. A further study of a mouse model in SpA and IBD reported that enthesitis and ileitis were IL-23 dependent and that enthesitis was specifically dependent on IL-17A and IL-22. These findings have contributed to the understanding of the etiopathogenesis of AS. Recent findings from high-density genotyping of immune-related loci have identified more than 20 genetic loci associated with AS. One research challenge at hand is the application of this expanded scope of genetic susceptibility to AS, with a view to addressing whether a distinctive signature of genetic variants would have potential diagnostic use in the detection of preclinical AS.

Future directions: targeting high-risk populations