Definition and Classification

Enthesitis-related arthritis (ERA), a term introduced in the International League of Associations for Rheumatology (ILAR) classification of juvenile idiopathic arthritis (JIA), is defined according to the inclusion and exclusion criteria shown in Box 19-1 . It is predominantly a disease affecting joints and entheses of the lower extremities and can eventually affect the spine or sacroiliac (SI) joints. It is characterized by the absence of rheumatoid factor (RF) and by a strong association with the human leukocyte antigen–B27 (HLA-B27). In many instances, the disease evolves to closely resemble ankylosing spondylitis (AS), although the spine and SI joint involvement at disease onset are seldom present in childhood or adolescence. The progression to juvenile ankylosing spondylitis (JAS) is unpredictable but possible and early identification of individuals with axial disease is important so that appropriate therapeutic interventions can be initiated. ERA’s relationship to other subtypes of JIA, such as psoriatic arthritis, is not entirely clear (see Chapters 15 and 20 ).

JAS is a chronic inflammatory arthritis of the axial and peripheral skeletons, frequently accompanied by enthesitis, characterized by RF seronegativity, paucity of ANA, and having a firm genetic basis. Unlike ERA, radiological evidence of bilateral inflammation of the SI joints is required for a definitive diagnosis. AS in adults is defined by sets of criteria that are based on clinical, laboratory, and radiographic abnormalities. Only a few children or adolescents meet these criteria, principally because of the low frequency of spinal or sacroiliac signs or symptoms in the young patient. In addition, criteria for the diagnosis of AS in adults (see Chapter 15 ) are not applicable to the younger age group for a number of reasons: Normal values for some of the required physical measurements have not been published for children, or, if reported (back range), they have not yet been validated. In addition, limitations of spine and chest motion may reflect disease duration and are therefore of little aid in facilitating early diagnosis. The fact that peripheral joint disease precedes clinical axial involvement by years in many children precludes an early diagnosis by criteria in which abnormalities of spinal mobility or radiological changes are essential diagnostic features. Reasons for these differences between adults and children, whether they have an immunological, genetic, biochemical, or structural basis, are not understood.

The ILAR classification criteria allow most but not all JAS patients to be included into the ERA category because of the requirement to exclude patients with psoriasis, or with a first-degree relative with psoriasis; those with two positive tests for RF at least 3 months apart; or those with signs of systemic JIA in addition to lack of radiological documentation of axial involvement to serve as an inclusion criterion. The term juvenile spondyloarthritis (JSpA) may be used to represent the full spectrum of arthritis that appears in late childhood and adolescents, and which has a strong association with HLA-B27 and potential axial involvement. ERA and JAS are often referred to as the undifferentiated and differentiated forms of JSpA, respectively.

Epidemiology

The data examining the epidemiology of ERA are emerging. For this reason, the discussion that follows relies heavily on the study of patients who fulfilled traditional criteria for AS. Where there is sufficient evidence, ERA-specific information is included.

Etiology and Pathogenesis

The cause of JAS or ERA is unknown. The clinical, genetic, and epidemiological similarities of these disorders and diseases such as reactive arthritis, in which enteric or genitourinary tract infections play a triggering role, suggest the possibility of an infectious etiology, although none has been proven. Although no organisms have been isolated from the joints of patients with AS, evidence of a local inflammatory response to antigen is supported by some antibody and cellular immune studies, although confirmation of these findings is lacking.

The relationship between gastrointestinal infection and HLA-B27 is complex. The HLA-B27 transgenic rat develops a disease that is remarkably like AS in the presence of intestinal bacteria. In humans with IBD, there is a strong association between HLA-B27 and sacro­iliitis. Reports of an association between HLA-B27 and gastrointestinal (GI) isolation of Klebsiella species in adults with AS remain largely unconfirmed. Mielants and associates described the presence of inflammatory gut changes in adolescents with spondyloarthropathies, supporting a pathogenic relationship between spondylitis and inflammation of the GI tract. Clinical or occult GI inflammation may be related in ERA and AS to the association with HLA antigens as well as to cellular immunity to cartilage proteoglycans.

The mechanisms whereby HLA-B27 is involved in disease pathogenesis have been the subject of much debate. As a class I major histocompatibility complex (MHC) molecule, the role of HLA-B27 is to present endogenous peptides to the T-cell receptor on CD8 ⁺ lymphocytes. It has been proposed that the HLA-B27 molecule or peptides it presents share amino acid sequences with a microbial antigen (molecular mimicry) and thereby become a target for CD8 ⁺ T cells or cross-reacting antibody, resulting in an inflammatory response. However, to date no “arthritogenic peptide” has been identified. Misfolding of the HLA-B27 heavy chain in the endoplasmic reticulum may invoke an inflammatory response. One possible mechanism arises due to a unique Cys67 residue in the HLA-B27 heavy chains that leads to self-association, and homodimerization and misfolding, which results in retention in the endoplasmic reticulum. This, in turn, causes induction of the proinflammatory unfolded protein response. Another mechanism suggests that HLA-B27 homodimers are not retained in the endoplasmic reticulum but are in fact expressed on the cell surface and lead to activation of disease relevant immunoreceptors and downstream pathways (i.e., KIR3DL1, KR3DL2, and immunoglobulin-like transcript 4 receptors on synovial and peripheral monocytes, B and T cells). In support of this mechanism, Myles et al. have demonstrated that membrane-bound toll-like receptors are overexpressed in synovial and peripheral blood monocytes in ERA, which leads to increased levels of interleukin (IL)-6 and IL-8 levels compared with healthy controls. The last hypothesized mechanism has no reference to HLA-B27 homodimerization but relates to the inability of HLA-B27–positive persons to clear intracellular pathogens. In fact, carriers of HLA-B27 are defective in the killing of intracellular bacteria such as Yersinia, Salmonella, Shigella, and Chlamydia , the same pathogens implicated in triggering reactive arthritis. Furthermore, bacterial antigens or DNA have been identified in the synoviocytes of patients with reactive arthritis supporting the role of infection in the pathogenesis of SpA. Finally, genetic microarray studies are providing new insights into the pathogenesis in ERA. Preliminary studies in the peripheral blood and synovial fluid mononuclear cells of ERA patients compared to healthy controls demonstrate a distinct genetic profile with higher expression of genes associated with antigen presentation, scavenger function, chemotaxis, and proteases, along with a lower expression of genes associated with natural killer (NK) cell function, cell adhesion, and inhibition of apoptosis.

The pathogenesis of enthesitis, the characteristic abnormality of ERA and JAS, has been extensively studied by Benjamin and Mc­Gonagle. An inflammatory infiltrate that includes CD8 ⁺ and CD14 ⁺ cells develops at sites of enthesitis in subchondral bone with bone absorption and new bone formation. Tumor necrosis factor-α (TNF-α) messenger RNA is increased in affected bone. CD2R, a T-cell activation marker, is expressed at high levels in patients with JAS and JRA. A type 1 helper T-cell (Th1) response has been suggested, with lymphocytic and mononuclear cellular infiltrates. Cells of the synovial membrane express TNF-α, TNF-β, and TNF receptors similar to those of children with other types of JIA.

Genetic Background

The major genetic factor in ERA is the association with HLA-B27, as it is with AS. Family studies have indicated that AS is inherited as an autosomal dominant trait with penetrance of about 20%. Although the risk of development of ERA or AS in an HLA-B27–positive person is not precisely known (approximately 1% to 3%), epidemiological studies suggest that AS occurs 10 to 20 times more frequently in relatives of patients with AS (20%) and 50 to 80 times more frequently in their siblings. Thus, HLA-B27–positive persons with a family history of AS have a tenfold greater risk of AS than that of HLA-B27–positive persons with no family history of AS. The general risk that an HLA-B27 heterozygous parent with AS will have a male child with the disease is approximately 5% to 10% (20% if the child is also HLA-B27 positive; close to zero if the child is HLA-B27 negative). The risk of having a female child with JAS is lower. Risk may or may not be increased in first-born children. Familial disease may not be concordant for phenotypes in each member. As less than 5% of HLA-B27–positive individuals develop SpA, there are likely other genetic or environmental influences. Other genes make minor contributions to the genetic risk of AS. Reveille has summarized the genetic factors implicated in SpA ( Table 19-2 ).

Family studies, disease concordance in twins, and modeling of genetic risk has implicated genetic factors in AS. The strongest association with AS is with HLA-B27. There are at least 105 different HLA-B27 subtypes. The most common subtypes associated with AS are HLA-B27*05 (Caucasians), HLA-B27*04 (Chinese), and HLA-B27*02 (Mediterraneans). Two subtypes, HLA-B*2706 and HLA-B*2709, seem to have no disease association.

Of 247 children with JAS, HLA-B27 was present in 91%, confirming its strong disease association. The possibility that homozygosity for HLA-B27 was responsible for the juvenile onset of AS was not supported by data in one small study, and the idea that disease severity is genetically determined is unresolved. In 56 Latvian children with JIA (44% ERA patients) who were HLA-B27 positive, 8 HLA-B27 subtypes were identified with HLA-B*2705 being the most commonly associated with ERA. Additionally, the authors suggested that the HLA-B27 subtypes may be useful in predicting the treatment response. Specific single nucleotide polymorphisms (SNPs) of ERAP1 (AS associated gene) and IL23R (AS and PsA associated gene) were studied in JIA patients (N = 1054 including 65 ERA, 76 PsA and 24 undifferentiated patients) and healthy controls (N = 5200). The ERAP1 SNP was most strongly associated with the ERA subtype ( P = 0.005) while the IL23R SNP was best associated with PsA ( P = 0.04), although there was a trend toward association in the ERA category. Neither the SNPs of ERAP1 nor IL23R were associated with the other JIA subtypes. There are few known class II associations with JAS. A higher frequency of HLA-DRB1*08 (44.9%) was reported in a Mexican population with JAS than in a control population (25.4%). Maksymowych and colleagues reported that the LMP2A allele frequency in patients with adult- and juvenile-onset AS with uveitis was twice that in those without this complication (odds ratio [OR], 2.51). Ploski and colleagues reported an increase of B*4001, DRB1*08, and DPB1*0301, and the LMP2 b/b phenotype in patients with JAS compared with HLA-B27–positive controls or adults with AS. It is very likely that JAS and ERA are polygenic disorders.

Clinical Manifestations

The onset of ERA may be insidious and characterized by intermittent musculoskeletal pain and stiffness or objective inflammation of peripheral joints, particularly those of the lower extremities, together with enthesitis at one or more sites around the knee or foot. Occasionally, the disease may have an abrupt onset. Systemic signs are often minimal, but fatigue, sleep disturbances, and low-grade fever may be present. Symptoms related to the back are usually absent at onset but become increasingly evident during the disease course in adolescents. Differing modes of presentation and course may characterize specific population groups.

Enthesitis

Entheses—the sites of attachment of ligament, tendon, fascia, or capsule to bone—are characteristic sites of inflammation, especially in the lower limbs, that occurs in 60% to 80% of ERA and JAS patients. Consequently, many patients with ERA complain of knee, foot, or heel pain. Weiss et al. noted that in 32 newly diagnosed patients with ERA, 66% had at least one tender entheseal site while 44% have more than two sites involved. The presence of an increasing number of tender entheses at presentation was associated with persistence at 6 months (OR 2.18). Although the presence of exquisite, well-localized tenderness at characteristic entheses strongly suggests ERA, it must be noted that enthesitis occurs occasionally in other disorders, including other types of JIA, other rheumatic diseases, and occasionally in children without disease. Osgood–Schlatter disease should also be excluded before making a diagnosis of enthesitis affecting the tibial tuberosities. The presence of enthesitis is, however, the most helpful clinical feature in differentiating ERA from other types of JIA. Enthesitis is a characteristic early manifestation of JAS and occurs with greater frequency in JAS than in adult-onset AS. It frequently produces severe pain and resultant disability, which may be the child’s most important complaints.

A careful history and a thorough but gentle palpation of entheses may document evidence of past or present inflammation. The entheseal exam can include any of the anatomical sites shown in Figs. 19-1 and 19-2 . Enthesitis is diagnosed clinically from the presence of marked localized tenderness or swelling at the entheseal insertion into the bone. A diagnosis of ERA is strongly supported by enthesitis in the lower extremities especially at insertional sites around the patella (10, 2, and 6 o’clock positions especially), heel, and plantar fascial attachments to the heads of the metatarsals ( Fig. 19-2 ). Less commonly, pain can also be demonstrated at insertional sites around the pelvis (including the origin of adductor longus near the pubic symphysis), over the spinous processes, and at the entheses of the upper extremities. Observation of stance and gait (including walking on the toes and heels), may reveal altered weight bearing as the child avoids pressure on inflamed entheses.

Anatomical sites for assessment of enthesitis in ERA and JA.

A, Arrows indicate the most common sites of tenderness associated with enthesitis at the insertions of the quadriceps muscles into the patella and the attachments of the patellar ligament to the patella and tibial tuberosity. B, Arrow indicates the site of tenderness at the insertion of the Achilles tendon into the calcaneus. C, Arrows indicate the most common sites of tenderness associated with enthesitis at the insertion of the plantar fascia into the calcaneus, base of the fifth metatarsal, and heads of the first through fifth metatarsals. Swelling in this area is best visualized by having the child lie prone on the examining table with the feet over the edge.

( B and C from R.E. Petty, P. Malleson, Spondyloarthropathies of childhood, Pediatr. Clin. North Am. 33 (1986) 1079–1096.)

Peripheral Joint Arthritis

The peripheral arthritis in ERA and JAS is often asymmetrical and mainly involves the lower extremities. In most instances, the number of joints involved is limited (four or fewer), although children may have a polyarticular involvement in up to 25% at disease onset and up to 45% during the disease course. Joints of the lower extremity are affected more frequently than the upper extremity. Unlike other types of JIA, hip involvement at disease onset and during the disease course is common. Isolated hip disease may be the presenting feature. Although involvement of one or both knees is characteristic of both oligoarticular JIA and ERA, the child’s age at onset, especially if a boy, is a useful distinguishing feature. Small joints of the foot and toes are commonly involved. The least commonly affected joints are the small joints of the hands. Consequently, symmetrical disease of the small joints of the hands or polyarticular disease, particularly in a girl, is more likely to be another type of JIA. Shoulders are not uncommonly affected, and even the temporomandibular joint may be involved.

Tarsitis (inflammation of the intertarsal joints, overlying tendons, entheses, and soft tissues) is an unique manifestation occurring in up to one third of children with ERA at disease onset ( Fig. 19-3 ). It is accompanied by pain, tenderness, and restriction of movement in the midfoot that, in the presence of disease of the first metatarsophalangeal joint, results in a characteristic forefoot adduction deformity. Burgos-Vargas and associates concluded that a diagnosis of JAS could be confirmed or strongly suspected shortly after onset of disease in children who displayed enthesopathy, midtarsal foot involvement, sparing of the hands, and progressive onset of lumbosacral disease.

Clinical photograph and MRI demonstrating tarsitis in a 12-year-old boy (HLA-B27 positive) with ERA.

Pain at the costosternal and sternoclavicular joints and the sternomanubrial joint, often in conjunction with tenderness over the proximal clavicle, may be associated with significant impairment of chest expansion. In the work by Schaller and associates, five of seven patients with JAS had decreased chest expansion. Aside from the number and distribution of the affected joints, there is nothing clinically to distinguish the peripheral joint disease of ERA from that of other types of JIA. Burgos-Vargas and colleagues have identified a subgroup of children with typical adult-type onset of disease. Whether this presentation represents a distinct entity or merely an extreme end of the clinical spectrum is not certain.

Axial Skeleton

Involvement of the joints of the axial skeleton (spine or SI joint), although seldom present at disease onset can occur within 5 to 10 years from disease onset. In a population of Mexican children, spinal involvement developed within 3 to 5 years, with some patients reporting axial symptoms within the first year of disease onset. Distinct from AS in adults, children seldom have symptoms of axial involvement at disease onset; only 24% of children with JAS are reported to have pain, stiffness, or limitation of motion of the lumbosacral spine or SI joints at presentation. Many children may have difficulty reporting axial symptoms, but it is important to inquire about any pain in the lower back or buttock regions that is worsened by periods of inactivity (i.e., sitting for prolonged periods). Predictors of sacroiliitis or AS include HLA-B27, DRB1*04, male sex, age at onset of symptoms, family history of AS, arthritis, polyarthritis, enthesitis, tarsitis, hip involvement, axial involvement, and psoriasis.

In children with SI arthritis, pain may be elicited by direct pressure over one or both SI joints, compression of the pelvis, or distraction of the SI joints by the Patrick test (FABER test), or Gaenslen’s test. Examination of the back may demonstrate abnormalities in contour, such as loss of the normal lumbar lordosis, exaggeration of the thoracic kyphosis, or increased occiput-to-wall distance. The contour of the back on full forward flexion may demonstrate loss of the normal smooth curve in the lower part of the thoracolumbar spine ( Fig. 19-4 ), or there may be restriction of hyperextension, signifying early axial disease. The rigid spine of long-standing AS is rare in children. Cervical spine involvement is also a late development. Although observations of abnormalities of the contour of the back are often more informative than actual numerical measurements, sequential measurement of thoracolumbar mobility is useful in documenting progression of the disease. Muscle wasting and weakness of the muscles around the hip joints may be quite prominent.

A 15-year-old boy shown in the position of maximal forward flexion. Note the flattened back ( arrow ). Radiographs demonstrated bilateral sacroiliac arthritis but no abnormality of the lumbosacral spine.

The modified Schober test provides one index of abnormality ( Fig. 19-5 ). With the child standing with the feet together, a line joining the dimples of Venus is used as a landmark for the lumbosacral junction. A mark is made 5 cm below (point A) and 10 cm above (point B) the lumbosacral junction. With the patient in maximum forward flexion with the knees straight, the increase in distance between points A and B is used as an indicator of lumbosacral spine mobility. Normal values plus or minus 1 standard deviation are indicated in Fig. 19-6 . In general, a modified Schober measurement of less than 6 cm (e.g., an increase from 15 cm to less than 21 cm) should be regarded as abnormal. However, care should be exercised in interpreting this measurement, because there are large normal variations at each age, and the data have not been adequately validated in children with musculoskeletal disease. Measurement of the distance from the fingertips to the floor on maximum forward flexion is often used to quantitate spinal motion but is poorly reproducible and does not correlate with the Schober measurement. Furthermore, finger-to-floor distance reflects hip as well as back flexion.

Schober test. A, Measurement 10 cm above and 5 cm below the lumbosacral junction (the dimples of Venus) in the upright position. B, Measurement of the distance between the upper and the lower marks when the child is bending forward.

Normal values for the modified Schober test: mean ( dashed line ) ± 1 standard deviation ( solid line ).

(Adapted from H.M. Moran, M.A. Hall, A. Barr, et al., Spinal mobility in the adolescent, Br. J. Rheumatol. Rehab. 18 (1979) 181–185.)

Thoracic disease may be reflected in limitation of chest expansion. Normal thoracic excursion varies a great deal, and normal age- and sex-adjusted ranges have not been established. However, in a specific child, sequential measurement of thoracic motion may be useful in documenting progressive loss of range. In the adolescent, any thoracic excursion of less than 5 cm (maximum expiration to maximum inspiration, measured at the fourth intercostal space) should be regarded as probably abnormal. Even in the absence of symptoms, chest expansion in children with JAS may be restricted to 1 or 2 cm. Pain and tenderness at the costosternal and costovertebral joints may be elicited by firm palpation. Sternomanubrial tenderness sometimes occurs, but sternoclavicular pain is more common.

Pathology

The pathology of JAS or ERA has not been studied, but it is probable that abnormalities are similar to those of AS. The synovitis is in general much milder, and the degree of cartilage erosion in peripheral joints much less, in AS compared with adult rheumatoid arthritis (RA). The synovitis itself is otherwise virtually indistinguishable from that of RA, although there may be relatively more polymorphonuclear leukocytes present.

The characteristic pathological changes in the apophyseal and SI joints are enchondral and capsular ossification. The earliest lesion in the SI joints is subchondral inflammation, rather than synovitis, with formation of granulation tissue with few inflammatory cells. The surfaces of the SI joints are minimally affected, and pannus is not present. Enchondral ossification on the iliac side of the joint accounts for the radiographic appearance of erosions. As Ball commented : “As a rule, it seems that in any synovial joint in AS, the outcome represents a balance of erosive synovitis and capsular and/or ligamentous ossification. In joints of low mobility the ossific process tends to be the dominant feature.”

Enthesitis is characterized by nonspecific inflammation. Granulation tissue, infiltrated with lymphocytes and plasma cells and causing a localized osteitis, undermines the bony and cartilaginous attachment of the ligament or tendon. Healing of this lesion gives rise to a bony spur, such as a calcaneal spur at the insertion of the plantar fascia into the calcaneus or a syndesmophyte at the attachment of the outer fibers of the annulus fibrosus to the anterolateral aspects of the rim of the vertebral body.

Differential Diagnosis

At onset, ERA most closely resembles oligoarticular JIA (OJIA). However, whereas OJIA is characteristically a disease of young girls, ERA typically occurs in older boys and adolescents. The presence of enthesitis is the distinguishing clinical feature in children with ERA. ERA may mimic other inflammatory arthropathies, mechanical causes of back or lower extremity pain, or, very occasionally, infection or malignancy. A history of cramping abdominal pain, diarrhea, weight loss, and fever suggests an accompanying IBD in a child who otherwise has typical ERA. A few children with ERA also have psoriasis and would fulfill criteria for psoriatic arthritis. In most instances, however, children with ERA lack SI and back symptoms at onset, but, unlike children with OJIA, may have hip joint involvement.

In older children with established ERA, signs and symptoms of spine and SI arthritis clearly differentiate the child from one with OJIA. The disease then resembles JAS or AS. Arthritis of the cervical spine is infrequent and, when present, mimics that in children with polyarticular JIA. Thoracolumbar pain may reflect Scheuermann disease. Lumbar and lumbosacral pain has a myriad of causes, including spondylolysis, spondylolisthesis, osteoid osteoma, osteomyelitis, diskitis, and (rarely) lumbar disk herniation. Trauma may cause chronic pain in the sacrum and coccyx. SI tenderness and pain occurs in many patients with JAS, but septic SI disease, osteomyelitis, Ewing sarcoma of the ilium, and familial Mediterranean fever also produce pain in and around these joints.

Pain that mimics enthesitis may result from a number of causes, including excessive running or jogging. Usually, the pain of traumatic enthesopathy is less severe and more diffuse than that caused by inflammation. Osteochondrosis of the tibial tuberosity (Osgood–Schlatter disease), of the inferior pole of the patella (Sinding–Larsen–Johansson syndrome), or of the apophysis of the calcaneus (Sever disease) may mimic inflammatory enthesitis at those sites. The coexistence of enthesitis at multiple sites usually eliminates these disorders from consideration. The absence of HLA-B27 positivity also assists in differentiating these disorders from the inflammatory enthesitis of JAS. Pressure over bony prominences, including entheses, may produce pain in children with leukemia or bone tumors. In most instances, however, the pain resulting from such infiltrative diseases is less discrete and more severe than that of inflammatory enthesitis and frequently awakens the child from sleep.

Laboratory Examination

There are few distinguishing laboratory features. Anemia is usually mild and characteristic of the anemia of chronic disease. White blood cell counts are usually normal or moderately elevated with normal differential counts. Indices of inflammation may not be a reliable measure of disease activity as they can be normal or minimally elevated even with clinically active disease. Very high values for the ESR (above 100 mm) occasionally occur but should also suggest the possibility of occult IBD.

Elevated immunoglobulin levels reflect inflammation, and selective IgA deficiency has been reported. High levels of IgA and C4 and of circulating immune complexes in adults with AS suggest an immunoreactive state. In a cohort of Chinese JAS patients, significantly higher levels of serum IgA were found compared with adult-onset AS patients and reflected the worse functional outcome and global assessment of disease activity in the JAS patients. Characteristically, RFs are absent. ANAs do not occur in children with ERA more commonly than in a healthy population. Antiphospholipid antibodies have been demonstrated in 29% of adults with AS, but children with ERA have not been studied. Although there are no reports of systematic studies of other autoantibodies in ERA or JAS, experience suggests that they are not common. Fecal calprotectin levels are high in children with ERA compared to those with other types of JIA.

HLA-B27 is present in 90% of children with JAS, 60% to 80% of ERA patients, and in approximately 8% of the overall white population; it does not constitute a diagnostic test but rather an indicator of risk. The diagnosis of ERA rests on clinical characteristics, and the use of HLA typing for diagnosis may lead to misdiagnosis, although typing is important as a criterion to classify patients for study.

There are no specific studies of synovial fluid, but the changes are probably similar to those in adults with AS, in which the differential white blood cell count includes more neutrophils and fewer lymphocytes than in RA. It has been reported that the predominant large mononuclear cell in synovial fluid in RA is lymphocyte derived, whereas that in AS is of macrophage origin. Macrophages containing degenerated neutrophils are more common in the synovial fluid of patients with AS and related diseases such as reactive arthritis than in those with RA. Descriptions of the synovial fluid are otherwise similar to those for RA, except that the complement level is usually normal or increased.