The reactive arthritides constitute a group of diverse inflammatory arthropathies in which the joint and extraarticular manifestations are caused by a preceding extraarticular infection with specific bacteria. This chapter reviews arthritis related to enteric or genitourinary bacterial infections. Rheumatic fever, caused by a pharyngeal infection with Streptococcus , is discussed in Chapter 44 .
Definition and Evolving Concept
In this chapter, use of the term reactive arthritis (ReA) is restricted to the arthritides triggered by enteric infections, with one of the so-called arthritogenic bacteria ( Yersinia, Salmonella, Shigella, or Campylobacter ) and genitourinary infections with Chlamydia trachomatis . Although ReA has predominantly occurred in individuals who are human leukocyte antigen (HLA)-B27 positive, a significant proportion of patients are HLA-B27 negative.
The finding of viable Chlamydia within the joints of patients with ReA challenges the idea that ReA is a nonseptic arthritis and raises the possibility that it can be considered to be an infectious arthritis or an infectious nonsuppurative arthritis. The current hypothetical relation between ReA, spondyloarthritis (SpA), and septic arthritis as result of Chlamydia arthritis according to Gracey and Inman is shown in Fig. 43-1 . The triad of arthritis, conjunctivitis, and urethritis (or cervicitis) was called Reiter’s syndrome in the past; currently the triad is still being recognized, but the eponym has been deleted from contemporary literature.
Classification and Diagnostic Criteria
The diagnosis of ReA is a clinical challenge because it requires the demonstration of an infection preceding the appearance of arthritis. Diagnosis is currently based on the presence of lower limb, asymmetrical oligoarthritis, and clinical or laboratory evidence of a preceding infection. In 1995, a proposal for diagnostic criteria was developed in an expert’s meeting in Berlin ( Boxes 43-1 and 43-2 ). Based on existing definitions and diagnostic criteria, Pacheco-Tena and colleagues proposed alternative classification criteria. A preceding bacterial infection is most frequently documented serologically, but its actual value is still controversial. In general, the value of diagnostic tests depends on the pretest probability of the diagnosis.
Typical Peripheral Arthritis
Predominantly lower limb, asymmetric oligoarthritis
Evidence of Preceding Infection
If there is a clear history of diarrhea or urethritis within the preceding 4 weeks, laboratory confirmation is desirable but not essential.
Where no clear clinical infection is identified, laboratory confirmation of infection is essential.
Patients with other known causes of monoarthritis or oligoarthritis (such as other defined spondyloarthropathies, septic arthritis, crystal arthritis, Lyme disease, and streptococcal reactive arthritis) should be excluded.
Culture of stool and urethra
Serology: antibodies against specific arthritogenic bacteria
Urethral swab for detection of chlamydial DNA by polymerase chain reaction (PCR) *
* A potential diagnostic test.
Synovial fluid or synovial membrane biopsy for detection of bacterial DNA by PCR *
Immunofluorescence microscopy for detection of bacteria in synovial biopsy specimen †
† Research tools, not suitable for routine diagnostic use.
Stimulation of synovial fluid lymphocytes with antigens from arthritogenic bacteria †
There appears to be a progressive decline in the number of publications dealing with ReA, perhaps reflecting a decreasing frequency of the disease. The prevalence of ReA appears related to that of HLA-B27 and probably to the rate of infections by arthritogenic bacteria in the general population. However, ReA may occur in HLA-B27–negative patients, including the majority of those reported in series of arthritis triggered by Salmonella and Yersinia .
The frequency of arthritis following infection with arthritogenic organisms varies widely. In adults, rates of ReA following Salmonella infection ranged from 3.2% and 4.4% to 43%. In a Salmonella outbreak in Germany, no cases of ReA occurred among 286 infected children, although 6 children had brief arthralgia. In contrast, 20% of 207 children reported joint, eye, or mucocutaneous symptoms after an outbreak of Salmonella typhimurium phage type 135a in Australia. ReA develops in 5% to 10% of children with Yersinia . ReA occurred in 43% of adults infected; the role of HLA-B27 was considered to be of minor relevance. A foodborne outbreak of Yersinia Pseudotuberculosis O:1 led to ReA in 22% of infected adults; 67% of those adults who were tissue typed had HLA-B27, but none of 12 infected children did. However, 42% of the children had erythema nodosum. In studies summarized by Keat, ReA was estimated to occur in 1% of patients with sexually acquired infections. Recent studies suggest that the prevalence of Chlamydia- related ReA is underestimated and that it is the most common of the arthritogenic bacteria-related arthritides. Keat reported that 2.4% of those with either Shigella or Campylobacter infections developed ReA. ReA has been also linked to Mycoplasma and Chlamydia pneumoniae, which was responsible for approximately 10% of cases of ReA in a Finnish study. HLA-B27 ReA has also been associated with Clostridium difficile.
The relative frequency of ReA among patients in pediatric rheumatology clinics in the United States, United Kingdom, and Canada ranged between 4.1% and 8.6%. This wide variation is consistent with differences in the stringency of diagnostic and classification criteria used in each study. Despite reports from other sources suggesting an increase in the recognition of ReA, that does not seem to be the case today. Most cases of ReA occur in boys between the ages of 8 and 12 years, but sex and age distribution vary according to the causative organism. In an Italian study of children with Yersinia -triggered ReA, most cases occurred between the ages of 3 and 7 years, and there was a slight predominance of females. Enteric infections are responsible for ReA at all ages, but ReA following genital infections with Chlamydia occurs more frequently during adolescence and in adults.
Although the susceptibility to the primary infection is not related to any known genetic marker, it is considered that ReA occurs most frequently in individuals with HLA-B27. Moreover, the severity of joint pain after intestinal infections by Salmonella, Shigella, and Yersinia may be associated with the presence of HLA-B27. Nonetheless, the frequency of HLA-B27 in children with ReA varies widely; in children with mild forms of Yersinia-, Campylobacter -, Chlamydia-, and Mycoplasma pneumoniae– related ReA, the frequency of HLA-B27 is similar to that of the population.
Other associations have been reported in specific populations. An association with the tumor necrosis factor (TNF) c1 allele independent of B27 was reported in a predominantly adult Finnish population with ReA. In a similar population, TAP2J, a polymorphism of transporters associated with antigen processing (TAP2), was more frequent in B27-positive patients with ReA. An association between the Toll-like receptor 2 and ReA has been described after an outbreak of S. enteritidis in Canada. Single nucleotide polymorphisms in the interferon (IFN)-γ gene (rs2430561 and rs1861493) appear to predispose the Dutch to ReA, and solute carrier family 11 member A1 gene polymorphisms are increased in the Chinese. In contrast, the chemokine receptor-5 (CCR5)-delta 32 mutation does not play a role in the susceptibility to ReA in patients with Chlamydia trachomatis infection.
Etiology and Pathogenesis
Several bacteria are involved in the etiology of ReA. In preadolescent patients, Salmonella, Shigella, Yersinia, or Campylobacter enteric infections precede the onset of arthritis in 80% of instances. Shigella flexneri, Yersinia enterocolitica, and Salmonella enteritidis have all been isolated from children with the postdysenteric ReA conjunctivitis and urethritis triad. In at least two youths and three children, Chlamydia trachomatis was identified in synovial fluid. Respiratory tract infection with Mycoplasma pneumoniae has been associated with ReA in a few children.
Role of HLA-B27
The role of HLA-B27 in the pathogenesis of ReA is still unknown. The arthritogenic peptide hypothesis postulates that the HLA-B27 molecule is able to bind a unique bacterial or self-antigenic peptide (not yet identified, but supposedly present in the joints), which is then presented to an HLA-B27–restricted cytotoxic (CD8 + ) T cell. CD8 + T cells’ cross-reaction with bacterial epitopes may then lead to inflammation and tissue damage. Several bacterial amino acid sequences homologous to HLA-B27 amino acid sequences have been described. Most recently, analysis of the HLA-B*2705 peptidome identified a number of peptides derived from various sources, mainly cartilage or bone-related proteins such as osteoprotegerin, annexin 2, chondrocyte-derived metalloproteinase, and proteoglycans, which share certain homologies with human HLA-B27.
Three of the peptides related to cartilage or bone shared 6-8 amino acids with amino acid sequences from arthritogenic bacteria.
Equally relevant is the B27 misfolding hypothesis, which refers to the accumulation of misfolded heavy chains of HLA-B27 in the endoplasmic reticulum mediated by the E3 ubiquitin ligase HRD1 (SYVN1) and the ubiquitin conjugating enzyme UBE2JL. Endoplasmic reticulum (ER) stress signaling pathways may be activated by HLA-B27 upregulation and misfolded heavy chains. Misfolding leads then to the activation of nuclear factor-κB and proinflammatory cytokines. The HLA-B27–free heavy chain and homodimer hypothesis refers to abnormal B27 molecule folding and the formation of HLA-B27 homodimers. Such homodimers may turn into receptors for humoral or cell-mediated responses mediated by class I HLA molecules or as proinflammatory targets, which lead to excessive cytokine production. The deposition of β2 microglobulin in the synovium, and perhaps other tissues, induces inflammation.
HLA-B27 modulates the production of cytokines and influences both bacterial invasion of cells and killing of bacteria. As a result, intracellular survival of arthritogenic bacteria is prolonged. Arthritogenic bacteria invade the gut mucosa and replicate within polymorphonuclear cells and macrophages. Studies in murine fibroblasts transfected with B27, not replicated in human cells, indicated that the expression of this antigen inhibited cell invasion by arthritogenic bacteria. Persistence of the organism within B27 cells was prolonged. Various bacterial components (including lipopolysaccharide, DNA, and RNA) have been identified in both synovial fluid cells and synovial membranes of patients with ReA. Regarding Chlamydia , a microorganism with both intracellular and extracellular life cycles, there is now clear evidence it may turn into a nonreplicative, nonculturable, but viable state within the cell. In this sense, it is now possible to consider that Chlamydia ReA may actually be a form of septic arthritis.
The antibody response against arthritogenic bacteria in ReA lasts longer than in infected patients who do not develop arthritis. Nevertheless, the role of antibodies in the pathogenesis of ReA is probably minimal. Additional findings suggest a role for heat-shock proteins and bacterial peptidoglycan in the pathogenesis of ReA.
The course and severity of ReA vary considerably. Symptoms of infection usually precede the onset of arthritis, enthesitis, or extraarticular disease by 1 to 4 weeks. After an active period of weeks to months, the arthritis subsides and the patient then enters a sustained remission or a phase of recurrent disease activity, which may evolve into enthesitis-related arthritis or SpA, including ankylosing spondylitis.
Characteristics of the Primary Infection
An appreciation of the characteristics of the enteric or genitourinary infections, which trigger ReA, could possibly aid in the identification of the bacteria involved in the pathogenesis of the disease.
A period of high fever, with or without watery diarrhea, and cramping abdominal pain lasting 48 to 72 hours, may be followed in 7 to 21 days by the sudden onset of nonmigratory oligoarthritis (knees and ankles) lasting from several weeks to 3 or 4 months. Diagnosis requires a history, the presence of agglutinins to Shigella flexneri serotype 2 or 2a, and an attempt to isolate the organism from the stool. Because of the long interval between the diarrhea and the joint complaints, blood cultures are positive in less than 4% of patients.
The acute onset of oligoarthritis, mostly in the knees and ankles, may follow an enteric infection with Salmonella typhimurium or Salmonella enteritidis by 1 to 3 weeks. The enteric infection may be mild, but the onset of arthritis is usually accompanied by low-grade fever. Because Salmonella infection can also result in osteomyelitis and septic arthritis, it is important to make certain that the synovial fluid is sterile. The erythrocyte sedimentation rate (ESR) is usually elevated, and the leukopenia that may accompany the acute infection is generally followed by leukocytosis. Stool cultures are usually positive, even late in the disease course, but seroconversion to Salmonella H and O antigens occurs in only 50% of patients.
ReA triggered by Yersinia may affect some children. The interval between infection and onset of arthritis in 18 children with Yersinia -triggered ReA was 7 to 30 days. The diarrhea preceding ReA was notably very mild, much more so than in the usual Yersinia enterocolitis. Contact with the organism is through infected drinking water or milk. Yersinia enterocolitica causes gastroenteritis in young children and a syndrome of abdominal pain similar to that of appendicitis in older children. In a study of children hospitalized because of Yersinia infection, 35% had arthritis lasting 3 to 22 months (average, 6.5 months). Of those with arthritis, 85% had HLA-B27. Yersinia can occasionally cause septic arthritis.
In an epidemic outbreak of Campylobacter jejuni enteritis in Finland, 2.6% of patients—all adults—developed oligoarthritis or polyarthritis 4 days to 4 weeks after infection. Synovial fluid cultures were negative, and 33% of the patients with arthritis were positive for HLA-B27.
Genitourinary tract infection with Chlamydia trachomatis is often asymptomatic but may cause dysuria, frequency, and a urethral or vaginal discharge. ReA may also be related to upper respiratory tract infections with Chlamydia pneumoniae . Artamonov and colleagues found evidence of nasopharyngeal infection in 45 of 52 children with ReA. Although the prevalence of HLA-B27 was higher than that in the control population (relative risk, 2.5), it was much lower than that in those who developed ReA after intestinal infection. Chlamydia infection is increasingly recognized in teenagers, and particularly in young adults.
Acute arthritis with marked pain and sometimes erythema over the affected joints is characteristic of ReA, but some children show only slight to moderate joint pain and swelling over several weeks. Enthesitis may occur alone or with arthritis, tenosynovitis, or bursitis ( Figs. 43-2 and 43-3 ). In other children, arthralgias appear before the onset of arthritis for a variable amount of time. The initial episode of arthritis usually affects the knees or ankles. The pattern of arthritis in the metatarsophalangeal (MTP) joints and the proximal and distal interphalangeal (IP) joints of the feet may be that of a dactylitis and may involve two or three joints in one or more digits in combination with tenosynovitis and bursitis. Arthritis of the small joints of the hands caused by Yersinia and Salmonella has also been described in ReA .
The synovial fluid effusion is usually marked, but proliferative synovitis is uncommon. In addition to involvement of peripheral joints and entheses, there may be inflammation of joints of the axial skeleton resulting in spinal and sacroiliac pain, stiffness, and reduced mobility of the lumbar and cervical spine.
In a study of 11 children with ReA followed for 0.9 to 6.7 years, Hussein observed recurrent episodes of arthritis in most patients: 4 children had severe arthritis, 5 had sacroiliitis, but none had significant disability. Cuttica and colleagues found that, at a mean follow-up of 28.6 months, 18 of 26 children with the triad of arthritis, conjunctivitis, and urethritis developed oligoarthritis, 7 developed polyarthritis, and 1 developed monoarthritis, with axial symptoms in 6. Five patients followed for around 7 years developed radiographic sacroiliitis. Symptoms remitted in most patients, but some had either a sustained or a fluctuating course. In a group of nine Greek children with Salmonella -triggered ReA, disease was active at 4 to 13 months, and there were one to four recurrences in four patients during a period of 48 to 78 months, but there were no axial symptoms.
Constitutional Signs and Symptoms
Apart from the infection itself, children with ReA may continue to have fever, weight loss, fatigue, and muscle weakness during active periods of disease. Polyarthralgia, muscle pain, and joint stiffness affecting peripheral joints and the axial skeleton sometimes accompany these symptoms. Myocarditis and pericarditis have been described during the active phase of the disease in children with Salmonella enteritidis -triggered ReA.
Mucocutaneous and Ocular Disease
Painless, shallow ulcers of the oral mucosa and palate are common and often asymptomatic. Aphthous stomatitis occurs in some patients. Urethritis and cervicitis are rare manifestations, occurring more frequently in adolescents with sexually acquired ReA caused by Chlamydia . These conditions are often mild, and girls tend to have no symptoms; they are detected only because of the presence of sterile pyuria. Diarrhea occurs in association with bacterial infection but may also be part of a generalized episode of mucositis.
Skin lesions in ReA include erythema nodosum in some children with Yersinia -triggered ReA, circinate balanitis ( Fig. 43-4 ), and keratoderma blennorrhagicum ( Fig. 43-5 ), with or without conjunctivitis or urethritis. Keratoderma may be clinically and histologically indistinguishable from psoriasis. Mucocutaneous involvement in ReA tends to parallel disease activity in the peripheral joints.
Conjunctivitis occurs in about two thirds of children at onset. In Yersinia -triggered ReA, conjunctivitis may be purulent and severe. Acute iridocyclitis in these cases is characterized by flare and cells in the anterior chamber, small keratic precipitates, cells in the vitreous, and occasionally fibrinous exudates, posterior synechiae, and macular edema in a unilateral or bilateral pattern. Acute anterior uveitis has also been described in ReA triggered by Salmonella typhimurium. Although there are few studies of the visual prognosis in children with ReA, the frequency of patients with permanent ocular sequelae appears to be low.
In the early inflammatory phase, there may be a slight decrease of hemoglobin and hematocrit, mild leukocytosis, and neutrophilia. The platelet count and serum levels of immunoglobulin M (IgM), IgG, and occasionally IgA may be elevated. The ESR and C-reactive protein (CRP) correlate with disease activity. In patients with severe disease—particularly in those with polyarthritis and polyenthesitis, fever, weight loss, fatigue, mucositis, or dermatitis—these laboratory abnormalities may be extreme. In particular, the hemoglobin concentration may fall to 8 to 10 g/dL, and the platelet count rise well above 400,000/mm 3 . The ESR and CRP values may remain elevated for a protracted period. Autoantibodies (e.g., rheumatoid factor, antinuclear antibodies) are usually absent. Synovial fluid analysis and culture helps to distinguish between ReA and septic arthritis.
With the exception of epidemics and some isolated reports, the clinical and laboratory confirmations of infection as a trigger in children with ReA are seldom made. When available, cultures obtained at the time of the infection may be helpful. Salmonella, Yersinia, Shigella, and Campylobacter may be isolated from the gut during an episode of diarrhea, or Chlamydia may be cultured from the urethra, but negative results do not exclude the diagnosis of infection-related arthritis. Because Salmonella and Chlamydia may also be present in asymptomatic carriers, these organisms can occasionally be cultured from patients who have arthritis not directly related to these organisms. Nonetheless, it is important to bear in mind that viable but nonculturable Chlamydia might be found in the joints of patients with arthritis. Electron microscopy, immunochemistry, and DNA studies might help in identifying extracellular elemental and intracellular replicative bodies and other constituents of Chlamydia.
More frequently, ReA is diagnosed in the appropriate clinical setting because of the presence of high titers of serum antibodies against arthritogenic bacterial antigens. Hemagglutination tests are useful in documenting recent infections with Salmonella or Yersinia . Both the sensitivity and the specificity of circulating IgA and IgM antibodies to Salmonella, Yersinia, and Campylobacter detected by enzyme-linked immunoassay are acceptable, but results must be compared with those in the control population. IgG antibodies are useful if levels change significantly; a rising titer of IgA antibodies may be noted. Overall, the use of these types of tests and their interpretation are still somewhat controversial.
Lymphoproliferation assays performed on cells from peripheral blood or synovial fluids also have some use as diagnostic tools. Unfortunately, these tests are not easy to perform and often demonstrate nonspecific responses to several antigens.
Yersinia or Chlamydia antigens may be detected in intestinal or genital smears or biopsies. By using electron and immunofluorescence microscopy and immunohistochemistry, it has been possible to identify intraarticular chlamydial extracellular elemental bodies and intracellular replicative bodies, Yersinia 60-kD heat-shock proteins and the urease β subunit, and Salmonella lipopolysaccharide. Likewise, bacterial DNA or RNA from several bacterial species, including Chlamydia sp., Salmonella sp., Shigella sp., and Campylobacter sp. has been identified in synovial fluid cells or the synovial membrane by PCR. The role of these tests as diagnostic tools is restricted at present.