The delineation between seropositive (rheumatoid) arthritis and seronegative forms of inflammatory joint disease – in which tests for immunoglobulin G (IgG) rheumatoid factor are negative – was a crucial step in subdividing inflammatory arthritis into identifiable syndromes with distinct prognoses, pathogenic mechanisms and responses to treatments. It is now clear that a wide range of forms of inflammatory joint disease is not linked with any rheumatoid factors; therefore, the delineation has lost much of its value in categorizing rheumatic disease. Much of the arthritis once referred to as ‘seronegative rheumatoid arthritis’ is now known to be spondyloarthritis (SpA). Other forms of inflammatory arthritis, such as that associated with sarcoidosis, polymyalgia rheumatica, crystal arthritides and juvenile arthritis, may be seronegative but are part of distinct clinical syndromes and are amenable to separate identification via increasingly precise clinical criteria.
As SpA has emerged as the effective replacement of ‘seronegative arthritis’, it is clear that SpA is at least as prevalent as rheumatoid arthritis in most areas of the world and is one of the main causes of morbidity and health-related costs. In spite of this, debate and uncertainty surround proper nomenclature and classification of SpA. SpA may affect axial (spinal) joints and/or peripheral sites; therefore, division of SpA into axial SpA (axSpA) and peripheral SpA is logical. However, historical identification of disease syndromes according to associations with psoriasis [psoriatic arthritis (PsA)], inflammatory bowel disease (IBD) (enteropathic arthritis), infection at extra-articular sites [reactive arthritis (ReA)] or primary involvement of the spine [ankylosing spondylitis (AS)] has formed the basis of most clinical and therapeutic studies and, thus, retention of these familiar terms, while still recognizing the substantial overlaps between them and with ‘undifferentiated’ forms, is important.
The overlaps between the key SpA syndromes and their common associations with inflammatory eye disease, overt or covert IBD and psoriasiform lesions of the skin and mucosae were recognized by family studies that revealed a higher prevalence of each among family members of affected probands.1 Following this, identification of a strong association with the human leukocyte antigen (HLA)-B27 gene cemented the existence of the SpA ‘family’. Subsequently, further genetic markers have been associated with SpA and hypothetical common pathogenetic mechanisms based on these genes and the infective ‘triggers’ associated with ReA have been proposed.
Two further developments have brought SpA into prominence in the sphere of rheumatic disease. Magnetic resonance imaging (MRI) of the spine and sacroiliac joints, which is sophisticated and readily available, has demonstrated lesions that were previously invisible, allowing axSpA to be diagnosed and studied as never before. Similarly, ultrasound imaging has revealed peripheral lesions at joints and entheses, which has allowed much greater understanding of the disease. Simultaneous with the development of these imaging modalities, the recent introduction of biologic drugs, which are effective in the treatment of many forms of inflammatory disease, including SpA, has generated justifiable enthusiasm for identifying, treating and investigating this increasingly prominent group of disorders. This review seeks to describe the key elements of SpA, the clinical syndromes and what is known about the causation and pathogenesis, diagnosis and treatment.
Spondyloarthritis currently defies tight diagnostic criteria, being instead a family of well-defined syndromes. An initial series of family studies by Moll et al.1 established close links between a group of conditions described together initially as ‘seronegative polyarthritis’1 and subsequently as ‘spondyloathritis’. Criteria built on this description have subsequently been introduced to allow recognition of the two major ‘divisions’: axial and peripheral SpA. Recently, the Assessment of SpondyloArthritis international Society (ASAS) published widely used criteria for classification of axial and peripheral SpA;2 however, the sensitivity of these criteria remains modest for diagnostic purposes.
Predominantly axial or predominantly peripheral forms of SpA are well described; however, most data have been collected in patients with ‘polar’ syndromes. AxSpA represents a spectrum of spinal inflammation3 ranging from subtle changes detectable only by MRI to irreversible radiographic disease fulfilling the modified New York criteria4 for AS. The ASAS classification criteria to define axSpA are often used by clinicians as diagnostic criteria in the absence of better alternatives, as these criteria work reasonably well for defining the disorder.
Among predominantly peripheral SpA, the key syndromes are PsA, enteropathic arthritis and ReA, but undifferentiated and juvenile forms have also been found. The characteristic features of peripheral SpA are well known and are described further below. Peripheral SpA is differentiated from other forms of inflammatory arthritis, notably rheumatoid arthritis and connective tissue disorders, by the absence of antibodies including IgG rheumatoid factor and cyclic citrullinated peptide antibodies. Differentiation is also underlined by the distinctive genetic associations marked by HLA-B27 on the one hand and HLA-DR4/shared epitope on the other, although other genetic markers, such as interleukin 23 (IL-23) receptor polymorphisms, are shared by a wide range of inflammatory joint diseases. The classification of peripheral SpA overlaps the diagnosis of the individual syndromes that are usually characterized by oligoarthritis. The extent to which oligoarthritis, when associated with other conditions, in fact represents a single condition is unresolved and, at present, both the individual syndromes and the overarching peripheral SpA concept are widely used. A key diagnostic feature of the SpA syndromes is a set of characteristic lesions associated with enthesitis.
The key lesions of spondyloarthritis
Before describing individual clinical syndromes of SpA, it is worthwhile focusing on the key lesions that are common throughout the SpA family. These comprise skeletal lesions such as enthesitis, peripheral arthritis, dactylitis and bursitis, spondylitis (including sacroiliitis) and spinal lesions; and extra-articular manifestations such as uveitis, IBD and psoriasis. In addition to these SpA-associated lesions, several important comorbidities bear an important influence on patient management. Chief amongst these are osteoporosis, cardiovascular disease and impaired kidney function.
The key pathological lesion of axSpA is enthesitis. Entheses are complex and variable structures at the junction between the ligament, joint capsule or tendon and bone. In the spine, entheses are affected at the attachment of joint capsules around facet joints and sacroiliac joints, at the discovertebral junctions and at the attachments of the interspinous ligaments. Initially, lesions may be detectable by MRI; however, this is not always of sufficient sensitivity, as has been found in areas with a high water content. In late-stage cases, radiographs may show areas of decalcification (‘erosions’) that will subsequently induce new bone formation. Ultrasonography is increasingly used to identify peripheral entheseal lesions.
Peripheral enthesitis is a characteristic feature of axSpA. Most commonly, the heel is involved. Enthesitis of the Achilles tendon occurs at the point of tendon attachment, in marked contrast to the thickening and pain located higher up the tendon that is seen in athletes. Enthesitis is often associated with formation of an Achilles tendon bursa, seen best from behind with the patient standing. Involvement of the plantar fascia is also characteristic, with associated pain and formation of a fluffy bony spur apparent on radiographs. These changes may be impossible to distinguish from degenerative plantar fasciitis. Evaluation of the entheseal lesions can be performed using the Maastricht Ankylosing Spondylitis Enthesis score.5
Peripheral SpA is most commonly oligoarticular, principally affecting the hips, knees and metatarsophalangeal joints. In the presence of psoriasis or IBD, it is not uncommon for small joint involvement, which may be symmetrical, to occur. In 10% of PsA cases, small joint arthritis may affect the distal interphalangeal joints, contrasting with the distribution and symmetry of rheumatoid arthritis. Synovitis is histologically non-specific but MRI may demonstrate extensive entheseal lesions within the joint area.
Dactylitis and bursitis
Inflammation at a finger or toe occurs in 5–10% of patients with SpA and is highly characteristic. Dactylitis usually affects a single toe (‘sausage toe’); however, if the patient also suffers from psoriasis, the fingers may also be affected. This results from a combination of enthesitis at multiple ligamentous attachments6 and joint synovitis and may be extremely painful and disabling. Dactylitis usually resolves spontaneously but can take many months to do so.
Cases of Achilles tendon bursitis have been found to accompany Achilles tendon enthesitis and bursitis is unusual in other areas of the body.
Sacroiliitis is often the cause of buttock and/or thigh pain. Sacroiliitis is characteristically diagnosed by radiographic appearances at the sacroiliac joints, graded as 0–4, with grade 5 being applied to unilateral changes; such changes take several months or years to become diagnostic. MRI may allow detection of pre-radiographic changes and, hence, facilitate early diagnosis. Typical appearances are of juxta-articular bone marrow oedema and are best seen on short tau inversion recovery (STIR) sequences, with later fatty changes also seen on T1 images. Diagnostic appearances are defined and well illustrated in the ASAS handbook.4 Although undoubtedly valuable, MRI remains problematic because the sensitivity of the scanning is limited by the relatively dynamic bone marrow changes and distinction from degenerative change may be difficult.
Radiographic changes occur late but are highly specific, whereas changes on MRI may be detected earlier but are less specific. Inflammatory corner lesions are characteristic, appearing as bone oedema or fatty change on MRI and as sclerotic ‘shiny corners’ on radiographs. In older adults, such changes of osteitis are relatively non-specific; however, osteitis at the pedicle is strongly linked with axSpA. Osteitis may occur around the facet joints and vertebral spines and such changes may lead to new bone formation, which is eventually visible on radiographs as syndesmophytes or bony obscurations of the facet and sacroiliac joints.
The SpA family of conditions is characterized not only by the clustering of spinal and peripheral skeletal lesions but also by the co-occurrence of acute anterior uveitis (AAU) (iritis), IBD and cutaneous and mucosal psoriasis. The link with these lesions is largely explained by common genetic factors. Such lesions occur in only a minority of patients with SpA and may be asynchronous with rheumatic features but are valuable clues for the clinician as to the diagnosis.
Acute anterior uveitis occurs in approximately 30% of people with axSpA at some stage, but is usually asynchronous with clinical activity of other lesions. In unselected patients presenting with AAU, up to 50% have – or will develop – other SpA characteristics. AAU typically causes pain, redness of the eye and photophobia. Onset is acute and, unless treatment is rapid, blindness may ensue, but prompt treatment normally leads to full recovery. In most instances, topical steroid treatment is effective and, in refractory cases, treatment with oral steroids, immunosuppressive agents or tumour necrosis factor (TNF)-blocking agents can be successful, but the evidence base for the TNF-blocking agents is sketchy. Dryness of the eyes in cases of AS is also not uncommon and may cause minor irritation.
Inflammatory bowel disease
Axial or peripheral SpA occurs in 5–15% of people with ulcerative colitis or Crohn’s disease. In the majority of patients, peripheral arthritis is oligoarticular and affects large lower limb joints asymmetrically (type 1); however, for some patients, polyarticular disease more reminiscent of rheumatoid arthritis (type 2) may occur. In up to one-third of patients with IBD, non-specific arthralgia occurs. This may be aggravated by withdrawal of steroid treatment and, in patients with ulcerative colitis, may persist after total colectomy. Conversely, around 60% of people with AS have subclinical IBD detectable by ileocolonoscopy,7 although the significance of subclinical disease in this respect is not clear. Overt IBD may be exacerbated by non-steroidal anti-inflammatory drugs (NSAIDs) used to treat axSpA.
Cutaneous and/or mucosal psoriasis may accompany any of the SpA syndromes and the comorbidities of psoriasis add to those associated with SpA itself. When axSpA and psoriasis occur together, the term psoriatic spondylitis is usually used. Typically, in cases of psoriatic spondylitis, spinal involvement is asymmetrical and often with prominent, non-marginal syndesmophytes; however, sacroiliitis may be unilateral. Typical extensor plaque psoriasis is most common, but pustular psoriasis on the soles of the feet (keratoderma blennorrhagicum) and mucosal psoriasis affecting the mouth and genitalia (circinate balanitis or vulvitis) are especially characteristic of ReA
Up to 1% of patients with longstanding AS will experience aortic valve incompetence due to aortitis of the ascending aorta. Cardiac conduction abnormalities occur in approximately 5% of patients and left ventricular dysfunction has also been found.
Active inflammation, high blood pressure, smoking and NSAID consumption probably all contribute to an elevated cardiovascular morbidity in patients with longstanding AS. In several studies, AS patients have been found to have higher carotid intima–media thickness and higher atherogenic index due to lower high-density lipoprotein cholesterol than control subjects, although the actual level of cardiovascular mortality in axSpA/AS is uncertain and may vary considerably between ethnic and geographical groups.
On rare occasions, chest wall rigidity is associated with apical pulmonary fibrosis but, more commonly, patients experience breathlessness on exertion as a result of costovertebral joint fusion. Furthermore, sleep apnoea has been described as a symptom by a few patients.
Renal impairment is described especially in patients with severe longstanding AS and may be linked to long-term use of NSAIDs. Immunoglobulin A nephropathy and amyloidosis have also been described in this group of individuals.
Axial osteoporosis occurs in around 25% of patients with AS and vertebral fractures occur in up to 10%. Thus, acute spinal pain in an individual with established AS should suggest the diagnosis of spinal fracture and it is likely that treatment with TNF-α inhibitors will prevent, or reduce the severity of, osteoporosis.
The spondyloarthritis syndromes
Axial SpA usually begins in the second or third decade of life, typically presenting as persistent buttock or back pain, and is characteristically described as inflammatory back pain. The ASAS criteria for inflammatory back pain8 are described as back pain of > 3 months’ duration with at least four of the following:
age at onset < 40 years;
improvement with exercise;
no improvement with rest;
pain at night with improvement on getting up.
However, onset later in life is also well recognized. Back pain and stiffness are usually worse after inactivity and may awaken the sufferer from sleep. Symptoms improve with movement; thus, stretching and exercise are usually helpful. Sacroiliitis often causes buttock pain that radiates down the back or front of either or both thighs but not below the knee. Typically, symptoms affect one side for a period of weeks or months and then subside, only to be followed by similar symptoms on the other side. This is known as alternating buttock pain.
In up to one-third of patients, the peripheral skeleton is affected. Typically, the hips or knees are affected but extra-articular sites including entheses or the eyes may also be affected, and peripheral symptoms may precede spinal symptoms. Pain in the chest wall is also common and results from either costochondritis or referred pain from the thoracic spine. Furthermore, fatigue is often a disabling symptom.
Men and women are affected equally but males are more likely to develop typical spinal radiographic changes. Thus, in established AS, the sex ratio is approximately 3:1 while, in non-radiographic axSpa, it is equal. The level of morbidity in non-radiographic axSpA is similar to that in AS.
Poor quality of life and socioeconomic consequences are critically important. People with axSpA are less likely to be married and more likely to be divorced, and women are less likely to have children than their healthy counterparts.9 In addition, work disability is associated with heavy societal costs.
Spinal deformity is not unusual among those severely affected and flexed posture may be aggravated by hip involvement. Deformity often leads to personal isolation as well as practical difficulties.
Diagnosis of AS or axSpA is usually based on fulfilment of the modified New York10 or ASAS classification criteria,11 respectively. The modified New York criteria require the presence of radiographic sacroiliitis, whereas the ASAS criteria require the presence of characteristic SpA features with evidence of sacroiliitis demonstrated either by radiography or by MRI. Criteria for sacroiliitis detected by MRI have been published.12
A definite diagnosis of AS can be made using the modified New York criteria in the presence of one of:
low back pain and stiffness that is relieved by exercise but not by rest;
limited motion of the lumbar spine in both sagittal and frontal planes;
limited chest expansion;
radiographic sacroiliitis ≥ grade 2 bilaterally or grade 3–4 unilaterally.10
Outcome in axial spondyloarthritis
There are few reliable indicators of the outcomes for people developing AS and there are no sensitive biomarkers to allow treatment decisions to be based on prognosis. However, it is clear that a severe form of the disease early on and hip involvement bode badly for future function. It is not clear how often the disease becomes quiescent (‘burns out’), as is sometimes described, but in many instances apparent quiescence actually reflects the patient’s tolerance of long-term symptoms and low-grade persistent chronic disease.
Some indicators of poor prognosis, including hip involvement, high acute phase response, NSAID unresponsiveness, limited spinal movements, dactylitis and peripheral arthritis, have been identified13 and, more recently, smoking has also been shown to predict poor outcomes.14 A potential imaging biomarker for progression of non-radiographic axSpA to radiographic AS has been identified in the extent of MRI changes at the sacroiliac joints in the early stages of the disease.15
Most people with AS live normal lifespans; however, for those with severe forms of the disease, premature death may result from cardiovascular disease, aortic valve disease and atherosclerosis, renal impairment, complications of skeletal surgery, trauma to a rigid spine or amyloidosis.
Axial osteoporosis occurs early in many AS sufferers and leads to an increased incidence of spinal fractures later in life. Thus, acute spinal pain in established AS can suggest the diagnosis of a spinal fracture. It is likely that TNF-blockade treatment prevents or reduces the severity of osteoporosis.
Functional impairment is reflected in the levels of work disability, with one-third of patients taking premature retirement. Spinal deformity is becoming less common but is still a blight for around 5% of hospital attenders with AS. Hip and knee arthritis also necessitates surgery in a substantial minority, with heterotopic ossification commonly leading to recurrent joint restriction and surgical revision.
Although most of the available data relate to individual SpA syndromes, it is logical to group the syndromes under ‘peripheral SpA’. As a consequence, ASAS has proposed a set of classification criteria for peripheral SpA,16 although many physicians are uneasy with this ‘lumping’ approach and many data refer to individual SpA syndromes, mainly PsA. As most study data apply to the individual syndromes, this terminology is currently of limited usage in clinical practice and it is unclear whether or not data gathered from peripheral PsA cases can be applied to other SpA conditions, such as undifferentiated SpA.
Psoriatic arthritis is an inflammatory arthritis associated with psoriasis, usually with negative tests for rheumatoid factor. Classification criteria for PsA have been drawn up by the Classification Criteria for Psoriatic Arthritis (CASPAR) working group.17 In this, patients with joint, spinal or entheseal disease are classified as having PsA if there is a score of ≥ 3 features from five groups of characteristics, including clinical, serological (negative tests for rheumatoid factor) and radiological characteristics.17 PsA is not a homogeneous clinical entity and, as with other spondyloarthritides, the key features are seronegative arthritis, enthesitis and, in a minority of cases, sacroiliitis or spondylitis. It is the only form of SpA in which small joints of the hand are frequently affected. Five patterns of joint involvement are recognized, with polyarthritis, oligoarthritis, involvement of distal interphalangeal joints and arthritis mutilans constituting the peripheral categories. Involvement of the spine in up to 40% of cases accounts for the inclusion of PsA in the SpA family; spinal lesions are more likely to be asymmetrical than in typical axSpA and, characteristically, syndesmophytes tend to be thicker and non-marginal, although they may be indistinguishable from uncomplicated axSpA. However, these apparently discrete subsets most often reflect the situation at onset, the patterns merging as the condition becomes chronic.
Psoriasis occurs in 5% of most Caucasian populations, and 5–15% of sufferers develop one or another form of associated arthritis. In a small minority of patients, arthritis precedes the onset of psoriasis.
Typical psoriatic nail changes such as pitting, onycholysis and hyperkeratosis are seen in > 80% of patients with psoriatic arthritis though skin lesions may be subtle and should be sought specifically in the scalp and natal cleft. Peripheral arthritis may be mono-, oligo- or polyarticular but is most commonly oligoarticular and asymmetrical and may be associated with dactylitis of fingers or toes, often described as a ‘sausage digit’.
Distal interphalangeal joint involvement in the fingers is uncommon but highly characteristic of PsA. Enthesitis plays a role in dactylitis but may also occur at more typical sites around the patella or the heel at the Achilles tendon or plantar fascia insertion. Approximately 20% of patients with PsA develop lower back pain with sacroiliitis and may develop typical or atypical spondylitis. Conjunctivitis and anterior uveitis may occur, but less commonly than in AS.
Psoriatic arthritis may be associated with substantial morbidity over and above that associated with psoriasis per se. In addition, the management of PsA is complicated by the associated comorbidities, especially metabolic syndrome, obesity and alcohol excess.
Appropriate disease-specific outcome measures have recently been introduced by the Group for Research and Assessment of Psoriasis and Psoriatic Arthritis (GRAPPA),18 which has also published treatment guidelines.19
Reactive arthritis is aseptic arthritis that occurs in association with an extra-articular infection, the joints being sterile to conventional cultures. Epidemiological studies have linked ReA with infections of the gastrointestinal tract, notably by Salmonella typhimurium, Yersinia enterocolitica, Shigella flexneri and Campylobacter jejuni.20 Longstanding observations have linked ReA with non-gonococcal infections of the genitourinary tract and less clear data have suggested a causal link with Chlamydia trachomatis infection, even giving rise to such terms as ‘post-chlamydial arthritis’, although mechanisms of causation remain unclear.
The true incidence and prevalence of ReA are not well defined. In epidemics involving Salmonella or Yersinia, ReA develops in 2–4% of infected individuals, but in as many as 20% of HLA-B27-positive infected individuals. In such epidemic studies, HLA-B27 positivity confers risk not only for the onset of arthritis but also for axial involvement and chronicity.
Typically, arthritis begins 1–3 weeks after the gastrointestinal or genitourinary infection. As with other SpA syndromes, the pattern of joint involvement in ReA is one of asymmetrical oligoarthritis mainly affecting joints of the leg. As in AS, enthesitis may arise as Achilles tendonitis or plantar fasciitis and dactylitis may occur in one or more toes. Sacroiliitis with buttock pain may occur in the acute phase, but radiographic changes are seen largely in the patients with a chronic form of the disease. When ReA is accompanied by urethritis, conjunctivitis or mucocutaneous lesions, the term Reiter’s syndrome may be applied, but increasingly ReA is used to refer to this complex of symptoms.
Urethritis may be manifest as dysuria or discharge and psoriasiform skin and mucosal lesions include circinate balanitis and keratoderma blennorrhagicum, a painless papulosquamous eruption on the palms or soles of the feet, and painless lingual or oral ulcers may also be seen. Conjunctivitis is usually bilateral and painful and AAU is usually unilateral and is seldom synchronous with the acute episode but may be clinically indistinguishable from conjunctivitis.
The most important differential diagnosis for ReA is septic arthritis; therefore, an appropriate culture of synovial fluid should precede the diagnosis of ReA whenever possible. The course of ReA is variable and few prognostic markers are available for the clinician to predict the course in any individual case. The majority of patients have an initial episode lasting 2–3 months, but synovitis may persist for a year or longer. In patients with a chronic form of the disease, a significant minority develops some degree of functional disability.
Recent studies have led to much debate over the causation of ReA. It is now clear that while joint cultures, even using sophisticated techniques, are negative, a wide variety of bacterial components, including wall proteins, DNA21 and, in the case of chlamydiae, RNA, are detectable in joint material.22 Such findings have also been shown in joint samples from patients with longstanding arthritis and with a range of diagnoses;23 thus, the issues of causation and pathogenesis, along with the ultimate role of antimicrobial therapy in the treatment of ReA, remain tantalizingly unresolved.24
Enteropathic arthritis is an inflammatory arthritis associated with IBD, particularly ulcerative colitis and Crohn’s disease. Arthritis has also been associated, however rarely, with Whipple’s disease, coeliac disease, collagenous colitis and surgical short bowel procedures for the treatment of obesity. Two patterns of peripheral joint involvement are recognized, designated as type 1 and type 2. Type 1 arthritis affects approximately 5% of patients with IBD. Typically, the peripheral arthritis is oligoarticular and principally affects the knees. It is usually self-limiting without leading to joint deformities. Joint symptoms can occur early in the course of bowel disease and may precede the onset of bowel symptoms. Enthesitis of the Achilles tendon and plantar fascia and dactylitis may also occur. Type 2 arthropathy affects approximately 3% of patients with IBD. Arthritis is usually polyarticular, principally affecting the metacarpophalangeal joints; however, knees, ankles, elbows, shoulders, wrists, proximal interphalangeal joints and metatarsophalangeal joints may also be affected, sometimes in a migratory fashion.
Usually bowel and joint symptoms occur independently, and arthritis may wax and wane over many years. Non-specific arthralgia and myalgia without an inflammatory component, similar to that seen in fibromyalgia, is not uncommon in people with IBD and may persist in those with ulcerative colitis, even after total colectomy.
Sacroiliitis and spondylitis occur in up to 20% of patients with either form of IBD. The course of spinal involvement is completely independent of the course of the IBD and may precede by years the first manifestations of bowel disease.
Although the term ‘enteropathic’ implies causation by bowel disease, the aetiology of enteropathic arthritis is unknown. Genetic links, as throughout the SpA family, are clearly important, and current studies focusing on the gut microbiome and its interaction with mucosal and systemic immunity may provide valuable information.
The development of inflammatory back pain or peripheral large joint arthritis, often in individuals who are HLA-B27 positive, with or without other features of SpA but without fulfilling criteria for any particular subtype, is referred to as undifferentiated SpA. Most patients are young adults, although children may be affected. In a proportion of cases, the clinical picture will evolve over time to conform to a classifiable subset, particularly AS. It is not unusual for the first feature of a SpA to be an enthesitis, especially at the Achilles tendon or plantar fascia. Such lesions must be distinguished from the traumatic lesions that occur independently of any arthritic conditions, especially in athletes. In SpA, Achilles tendonitis typically affects the actual entheseal junction, often with marked bone oedema visible on MRI and sometimes with Achilles tendon bursitis. In athletes, pain and tendon swelling occur higher up the tendon closer to the muscle belly. Plantar fasciitis is not so easily differentiated but often occurs in overweight, older adults. There are no diagnostic criteria for undifferentiated SpA per se.
The extent to which the histological characteristics of peripheral joint lesions of SpA allow differentiation from other conditions, or even justify the SpA classification, is very limited. No reliable histological differentiators have been identified.
Aetiology of spondyloarthritis
The sharing of clinical–pathological lesions among SpA subtypes and the association of each SpA subtype with the HLA-B27 gene suggests common mechanisms of pathogenesis and causation. AxSpA is likely to result from the interaction of environmental and genetic factors, and evidence for involvement of genetic factors is strongest in axSpA/AS. Among white Caucasians with AS, 95% carry the major histocompatibility complex (MHC) gene HLA-B27, although this occurs in only 8% of most white populations and only approximately 2% of HLA-B27 carriers develop AS.25 The prevalence of HLA-B27 varies throughout the world, with more than 30% of individuals carrying the gene in some circumpolar populations and almost zero prevalence in equatorial populations. However, MHC genes account for considerably less than 50% of the genetic susceptibility, and recent studies have identified significant associations with the ERAP 1, ARTS-1 and IL-23 receptor genes.26 Genetic association scans have also identified several other loci elsewhere in the genome that are likely to contribute to susceptibility to AS.26 It is likely that other genes being explored may actually protect against development of AS or modulate the disease expression and account for the differential expression of AS in men and women.
Hypotheses as to the mechanisms by which HLA-B27 leads to axSpA have been reviewed by McHugh and Bowness.27 Evidence that HLA-B27 misfolds intracellularly, leading to the production of IL-23, has recently been complemented by the finding of a population of T-cells resident in the entheses that promote inflammation characteristic of SpA in response to IL-23.28 The role of gut microbiota in promoting cytokine production in both IBD and SpA is the subject of intense research. Such work may well lead to an understanding of mechanisms common throughout the spondyloarthritides. Similarly, important studies on the skin microbiome may lead to greater understanding of pathogenetic mechanisms in PsA.
The potential role of micro-organisms in triggering the onset of SpA is an issue of great current interest, with much work focusing on the gut microbiota. More compelling data about microbial causes of SpA come from studies of ReA. In this condition, studies of outbreaks of enteric infections have shown a reproducible capacity for certain infections to be followed by the acute arthritic syndrome. Micro-organisms clearly linked in this way are S. typhimurium, S. flexneri and C. jejuni. Longstanding observations of non-epidemic infections have linked ReA with Y. enterocolitica and Y. pseudotuberculosis enteric infections and non-gonococcal genitourinary infection. Within the non-gonococcal genitourinary infection group, less clear data have suggested a causal link with C. trachomatis, supported by identification of C. trachomatis cell membrane proteins as well as DNA and RNA within inflamed joint tissue. Most recently, Carter et al.29 reported an antimicrobial treatment study29 that appears to support a causal role for this micro-organism in some cases of arthritis, but further work to clarify the situation is required.
Treatment of spondyloarthritis
With the advent of more effective treatments, measurement of disease parameters and outcomes has become especially important. As a consequence, a substantial range of measures have been developed by specialty bodies focusing on axSpA and PsA. Many of these measures are patient completed, referred to as patient-reported outcome measures (PROMs), although objective measures of imaging are also available for use, especially in treatment studies. The key measures most used in SpA are set out in the ASAS handbook,6 and a more comprehensive description of the range of available instruments is provided elsewhere.30
No specific measures have been developed for ReA and enteropathic arthritis so that, in these conditions, measures devised as above are sometimes used without formal validation.
Although treatment of the spine and peripheral lesions bear some similarities, they are in fact based on fundamentally different principles and sometimes weak evidence bases; the expectation that ‘if it works in rheumatoid arthritis, it will work in SpA’ is not necessarily correct.
The treatment of axSpA should be regarded as the sum of the treatment of the individual lesions. Hence, treatment of peripheral arthritis and enthesitis follows approaches used in other peripheral arthritides and iritis, and psoriasis and IBD require specific regimens as used in individuals without AS. A recommended treatment algorithm has been published by the ASAS group in association with the European League Against Rheumatism.31
For effective treatment, predetermination of the desired outcomes is vital, and treatment should be aimed at achieving these outcomes with appropriate measures. The principle of ‘treat to target’ is well established in rheumatoid arthritis and a similar approach has recently been advocated for cases of SpA.32 However, the relative dearth of therapeutic options, especially for cases of axSpA, and poor evidence base make this desirable in principle but difficult in practice. Clinical trial data, for the most part, apply to AS (axSpA) and PsA so that much of the management of peripheral SpA syndromes is empirical. Therefore, treatment of SpA is stratified into axial and peripheral disease with consideration also being given to adjunctive treatment of comorbidities.
Treatment should be based on current symptoms and function and on the predicted progression of the disease. Success of treatment should be gauged by regular measures of outcome including the PROMs listed under ‘Axial spondyloarthritis’. Work capacity and receipt of incapacity benefit are critically important outcomes, both being indicators of severe disease and of personal and societal cost.
Treatment of spinal disease is aimed at reducing pain and stiffness, increasing well-being and function, and delaying bony ankylosis. No treatment has been proven to prevent bony progression; however, this remains an important aim.
The key to maintaining comfort, posture and spinal mobility is regular exercise. Intermittent physiotherapy may be necessary to correct or minimize deformity or spinal joint restrictions as well as to maintain motivation; however, regular sporting activities are highly desirable and remain so throughout life. Some national patient organizations such as the UK National Ankylosing Spondylitis Society help to promote regular exercise though local groups and exercise guidance apps, which are of substantial benefit in supporting appropriate activities and maintaining motivation. Clear evidence exists that regular exercise promotes the preferential expression of anti-inflammatory cytokines33 so that the benefit may extend beyond simply keeping the spine moving.
Non-steroidal anti-inflammatory drugs
The reduction of pain and stiffness and the ability to exercise are usually best achieved by regular treatment with NSAIDs. Both non-selective COX2 inhibitors such as diclofenac and naproxen and COX2-selective inhibitors such as etoricoxib and celecoxib may be helpful. Regular NSAID medication should be accompanied by use of a proton pump inhibitor such as omeprazole, especially in older subjects or in the presence of dyspepsia. The risk of enhanced cardiovascular and renal disease in regular NSAID takers should also be borne in mind in treatment planning and pre-existing risk factors should be controlled to minimize risk. Regular NSAID usage may help to slow ankylosis,34,35 but on-demand treatment may be associated with reduced drug toxicity.
Disease-modifying antirheumatoid drugs
Traditional disease-modifying antirheumatoid drugs (DMARDs), in particular sulphasalazine and methotrexate, exert little or no effect on spinal disease but may be helpful in the treatment of peripheral joints and enthesial disease.
The introduction of TNF-blockade treatments has radically changed the formerly bleak therapeutic landscape in AS. Adalimumab, certolizumab, etanercept, golimumab and infliximab have all been shown to substantially reduce symptoms and improve function and quality of life.36,37 These responses are achieved as early as 2 weeks, and a substantial minority can expect to achieve partial remission at 2–3 years. Response to anti-TNF treatment in cases of AS is greatest in those with short-duration disease, high Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) and acute phase markers, in particular C-reactive protein (CRP), but more precise biomarkers predictive of responsiveness to treatment or other outcomes have not clearly been identified.
Treatment with each of the available anti-TNF agents has been associated with significant improvement in the physical component of the Short Form questionnaire-36 items score and improvement in well-being. Fatigue and sleep disturbance are important features of active AS and all anti-TNF agents are associated with improvement in fatigue. TNF-blockade therapy is associated with only modest improvement in metrology, as indicated by improvement in the Bath As Metrology Index score, reflecting both the extent of irreversible disease at the time when such treatment is usually instituted and insensitivity of this measure. Maintenance of improvements in spinal mobility requires sustained regular exercise.
Improvement in function, as measured by the Bath Ankylosing Spondylitis Functional Index (BASFI), is seen a few weeks after initiation of TNF-blockade therapy, but this is rapidly reversed on early discontinuation of treatment. Greater functional improvement is more likely to occur in those with an early form of the disease. Anecdotal reports suggest improvement in work capacity and return to work of some patients previously disabled by their disease; however, taken together, the small studies reported thus far have been unable to demonstrate significant reductions in work capacity other than a marked reduction in days lost from work.38
Disease modification of radiographic disease progression
Using MRI, it is clear that the acute changes of spinal inflammation often respond well to TNF-blockade therapy, although some residual spinal inflammation remains in most patients. Assessing disease progression in axSpA is not perfect, and methods for scoring disease progression are problematic. The modified Stoke Ankylosing Spondylitis Spine Score30,39 is currently the most sensitive of the methods to change and is therefore the radiographic method of choice for detecting radiographic progression. Reliance on anterior changes at two segments of the spine and exclusion of the posterior elements and thoracic segment are, however, undoubted limitations. Assessment of disease progression has also been hampered by lack of long-term follow-up of randomized controls on both ethical and practical grounds. Therefore, treatment groups have been compared with historical control groups such as the Outcome Assessments in Ankylosing Spondylitis International Study cohort,40 in which patients received NSAIDs, analgesics and regular exercise therapy. When these limitations are acknowledged, only one study has suggested that TNF-blockade treatment might inhibit disease progression in patients with active AS.41
New drugs and treatment regimes
The potential value of a range of newer biologic agents has been assessed in AS/axSpA but, thus far, success has been limited by comparison with TNF-inhibitor agents.42 Rituximab was as effective as a TNF inhibitor in anti-TNF-naive patients with active AS, but appeared ineffective in patients who had failed such treatment. Anakinra and tocilizumab have been associated with anecdotal benefits only and ustekinumab is of only modest benefit in axial disease. Recent studies of secukinumab have demonstrated efficacy in axSpA; however, comparisons with TNF inhibitors cannot yet be made. Clinical trials with small-molecule anti-inflammatory agents are proceeding. Notably, apremilast has been shown, in proof of concept studies, to be effective for symptom reduction in axSpA.43
At this time, it appears likely that most patients will require indefinite treatment; however, drop-out rates range from 8% to 16% per year. Stopping treatment appears to allow relapse in almost all patients, but everyday clinical experience indicates that some patients are able to withdraw treatment for periods of months and, occasionally, indefinitely. It is not known whether or not withdrawing treatment results in worse disease progression, although most patients respond again to retreatment. However, there is a lack of data to clarify the numbers in whom drug-free remission may be expected or characteristics of patients in whom achievement of this is likely.
Treatment of PsA should follow the GRAPPA recommendations19 for patients diagnosed according to the CASPAR criteria.17 For psoriatic spondylitis, the treatment algorithms discussed above should be followed. For peripheral joint disease, as in other forms of peripheral arthritis, symptomatic treatment with analgesics and NSAIDs is routine. As monoarthritis or oligoarthritis is not unusual in cases of PsA, local intra-articular steroid injections are often effective. DMARDs, notably methotrexate and sulfasalazine, are widely used in the treatment of PsA, although clear evidence of symptomatic efficacy or prevention of joint damage is lacking. One study, in 2012, of treatment with 15 mg methotrexate weekly44 failed to demonstrate efficacy, although clinical practice continues to favour such treatment at that dose or higher. Cyclosporin A and leflunamide have demonstrable efficacy in peripheral PsA, but gold salts, chloroquine and hydroxychloroquine are not recommended and there is no clear evidence base for use of combination DMARDs.
Systemic corticosteroid therapy is also not recommended by GRAPPA because of the potential to worsen psoriasis on withdrawal and the especially undesirable long-term toxicity in patients with PsA. However, judicious use of oral steroid treatment, taken on an occasional basis, may be effective.
Biologic treatment with TNF inhibitors has been very effective at achieving symptom control and at inhibiting joint damage. In clinical trials, most data have been obtained from patients receiving TNF blockers and methotrexate in combination but it is now clear that TNF-blockade monotherapy is an effective treatment for PsA.45 The place of monotherapy and combination therapy for PsA has yet to be fully defined.
In normal clinical practice, TNF-blockade therapy is associated with concomitant improvements in the Psoriasis Area Severity Index, nail involvement (Nail Psoriasis Severity Index) and enthesitis and also with improvements in physical disability, quality of life and functional status. Corticosteroid therapy is associated with poorer clinical response rates in older women.46 Available data do not differentiate between the available TNF-inhibitor agents, either as monotherapy or in combination with methotrexate, in terms of inhibition of radiological progression; however, it is not clear whether concomitant methotrexate enhances this effect or helps to maintain efficacy.
Ustekinumab, an anti-IL-12/IL-23 monoclonal antibody, has been shown to exert beneficial effects in both cutaneous psoriasis and PsA, although the effects appear to be less marked than those of TNF inhibitors and, thus, its place in treatment of PsA is not yet clear. Apremilast, a phosphodiesterase 4 inhibitor, has also shown promise in clinical trials and such trials continue.
Undifferentiated peripheral spondyloarthritis
Few treatment studies have focused specifically on peripheral SpA as a generic entity and it is not clear to what extent data on PsA are applicable to the generalization of peripheral SpA or vice versa. Equally, data on small-joint polyarthritis may not be applicable to large-joint oligoarthritis or monoarthritis. Treatment of peripheral SpA is usually influenced or constrained by the associated key SpA conditions.
In clinical practice, large-joint monoarthritis, especially of the knee, remains a challenging problem that may not respond well to conventional therapy.47 Although NSAIDs, DMARDs and oral corticosteroids are sometimes of value, no controlled clinical trials guide therapeutic decisions. In these circumstances, when the likelihood of benefit from first- and second-line oral drugs is likely to be modest, an intra-articular corticosteroid injection is often the first-line choice once sepsis has been excluded. Intra-articular injection of corticosteroids is a relatively safe and well-tolerated procedure, widely used in everyday clinical rheumatological practice. Intra-articular osmic acid has been used with some success as an alternative to corticosteroids in large joints but intra-articular methotrexate and TNF inhibitors have not, anecdotally, been shown to be superior to corticosteroids.
Current guidelines generally do not recommend systemic biologic treatment for monoarthritis, but its potential value in this context is unknown with anecdotal examples of benefit.
When large-joint monoarthritis is recurrent in spite of local treatment, arthroscopic lavage may be of some value. Surgical synovectomy usually has short-term benefits, but outcomes are best when patients have a low pain threshold, good range of movement and minimal radiographic change preoperatively. Patients with higher scores for pain, poor function and radiographic changes are more likely to require total arthroplasty in the future.
Clinically relevant enthesitis lesions are common in cases of SpA and may occur in isolation in an undifferentiated context. For many patients, conservative local measures, including padding, shoe raises (for Achilles tendon enthesitis) or local corticosteroid injections are adequately effective. Generally, the small benefit afforded by DMARD treatment does not justify the side-effects. TNF-blockade treatment has been shown to exert beneficial effects on entheseal lesions both with and without other SpA features, but such treatment is not recommended by any national guidelines.
Treatment of peripheral SpA as a homogeneous entity with adalimumab has been shown to be effective,48 but arthritis usually relapses on cessation of treatment and, in reality, much information is still missing. In particular, data on the efficacy of TNF blockade with respect to rapidly progressive hip destruction, which is a key indicator of bad prognosis in SpA, would be of great value. No randomized controlled trial data of abatacept or tocilizumab in the context of SpA have yet been reported.