This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Reprints and Permissions Copyright Information Books from ASM Press MicrobeWorld Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Colmegna, I. Articles by Espinoza, L. R. Search for Related Content PubMed PubMed Citation Articles by Colmegna, I. Articles by Espinoza, L. R.

 Previous Article  |  Next Article 

Clinical Microbiology Reviews, April 2004, p. 348-369, Vol. 17, No. 2
0893-8512/04/$08.00+0     DOI: 10.1128/CMR.17.2.348-369.2004
Copyright © 2004, American Society for Microbiology. All Rights Reserved.

HLA-B27-Associated Reactive Arthritis: Pathogenetic and Clinical Considerations

Section of Rheumatology, Department of Medicine, LSU Health Science Center, New Orleans, Louisiana 70112

SUMMARY

INTRODUCTION

HISTORICAL CONSIDERATIONS

PATHOGENESIS

Evidence of Chlamydia in the Joints

Evidence of Enteric Pathogens in the Joints

Bacterium-Host Interaction

Chlamydial ReA.

Enteric ReA.

Bacterial Target Antigens

Arthritogenic peptides.

Bacterial LPS.

Bacterial DNA.

Host

HLA-B27.

The gut.

The joints.

Immune Response to Bacteria

Miscellaneous

CLASSIFICATION CRITERIA

EPIDEMIOLOGY

ETIOLOGIC AGENTS

CLINICAL ASPECTS

Articular Findings

Extra-Articular Findings

LABORATORY FINDINGS

RADIOGRAPHIC FINDINGS

DIAGNOSIS

PCR Amplification

HLA-B27

DIFFERENTIAL DIAGNOSIS

COURSE

PROGNOSTIC FACTORS

THERAPY

Nonpharmacologic Approach

Pharmacologic Approach

Acute presentation.

Antibiotic use.

(i) Antibiotic trials in urogenital arthritis.

(ii) Antibiotic trials in enteroarthritis.

Refractory ReA: second-line therapy.

REFERENCES

Top Next References

INTRODUCTION

Top Previous Next References

 
Reactive arthritis (ReA) is defined as a sterile synovitis developing after a distant infection, usually in the genitourinary or gastrointestinal tract. The detection of microbial components (microbial DNA and RNA) in the joints of patients with ReA has led to the reconsideration of this definition (59). Currently, ReA is better defined as an immune-mediated synovitis resulting from slow bacterial infections and showing intra-articular persistence of viable nonculturable bacteria and/or immunogenetic bacterial antigens synthesized by metabolically active bacteria residing in the joint and/or elsewhere in the body (104, 204). A classification into HLA-B27-associated and nonassociated forms has also been proposed (28, 95, 206). Post-streptococcal, Lyme, and viral arthritis are HLA-B27 nonassociated and should be described as distinct entities under the general heading of "infection-related arthritides."

This article focuses on HLA-B27-associated ReA. This form of ReA belongs to the group of spondyloarthropathies (SpA), is triggered by bacteria (which enter the body through the mucosal surfaces) from the genera Campylobacter, Chlamydia, Salmonella, Shigella, and Yersinia, and is clinically associated with oligoarthritis of the lower limbs and sometimes with urethritis and conjunctivitis (203, 210).

HISTORICAL CONSIDERATIONS

Top Previous Next References

 
Hippocrates in the fourth century B.C. was probably the first to link the presence of arthritis and infection in the genitourinary tract (7, 95, 122) when he noted that "A youth does not suffer from gout until after sexual intercourse," gout meaning acute arthritis at that time. Interestingly, historical records concerning Christopher Columbus indicate that he suffered from arthritis associated with eye inflammation. If this, however, was "the first case of ReA in the old world," it remains a point of further discussion (1, 6, 72, 77).

Sir Benjamin Brodie, in 1818, was the first to describe the classic triad of urethritis, arthritis, and conjunctivitis in a series of five patients (31, 225). Gonococcal identification by Neisser in 1879 facilitated the differential diagnosis of patients with nongonococcal arthritis from the arthritis of disseminated gonococcal infection (1, 95).

On the other hand, the association of arthritis and dysentery or diarrhea was first recognized in the late 1600s (22, 224). Paronen et al. at the end of the second world war reported their experience with a large outbreak (more than 150,000 individuals) of dysentery secondary to Shigella (151). Of these, 344 (<1%) individuals developed arthritis, and four cases were associated with conjunctivitis. In 1916, several reports of this association were published almost simultaneously in Germany and France. Fiessinger and Leroy described four cases of urethritis, arthritis, conjunctivitis, and diarrhea, which they named the "oculo-urethro-synovial" syndrome (51). This clinical association is still known in France as the Fiessinger and Leroy syndrome. In Germany, Hans Conrad Reiter reported a similar clinical syndrome in a young German soldier and suggested a spirochetal infection as the triggering agent (156); however, he did not provide confirmation studies. To his credit, however, ReA secondary to the Lyme disease spirochete was recently reported from the same geographic area in which Reiter's patient resided (218).

The eponym Reiter's syndrome became widely adopted in the literature following its first description in the American medical literature by Walter Bauer and Efrain Engelman (18). In recent years, the Nazi past of Hans Reiter has come to light (216). He was responsible for involuntary sterilization, euthanasia, and medical experiments that killed thousands of concentration camp prisoners (60). Based on these considerations and Reiter's ethical background, several authors and groups including The Spondylitic Association of America, a patient advocacy group that represents people with this clinical syndrome, have strongly recommended that this syndrome should be called "reactive arthritis" (218).

PATHOGENESIS

Top Previous Next References

 
ReA is definitely related to an infectious disease, and multiple factors contribute to its development. Three central aspects of the pathogenesis of ReA are the presence of bacteria or bacterial products in the joint (Table 1), the bacterium-host interaction, and the local immune response directed against these bacteria. The fact that such different bacteria can induce arthritis, yet some bacterial subtypes (such as Yersinia O:8 or Shigella sonnei) in a single species do not, provides evidence that antigenicity alone does not determine the induction of arthritis. Rather, it may be a property of the ability of the bacterium to reach the joint or gain access to particular cell types and evade the host defenses (180, 230).

View larger version (24K): [in this window] [in a new window]   FIG.1. Events in the natural history of the arthritogenic infection. Data are from reference 175.

Chlamydial DNA is more likely to be found in the synovial tissue than in the synovial fluid (26, 143). The difficulty in growing Chlamydia from joint fluid or tissue may be explained by the fact that the organism persists as intracellular reticulate bodies rather than as infectious elementary bodies. Live but noncultivable intracellular organisms may be driving the inflammation in chlamydial ReA.

The discovery of mRNA and rRNA (nucleic acids which have a half-life of minutes in tissues) of C. trachomatis using reverse transcriptase PCR of synovial biopsy specimens provided evidence of the occurrence of transcription and hence active multiplication of the bacteria (59). C. trachomatis was also found in cells of the peripheral blood but not in the serum by PCR of samples from patients with early Chlamydia-induced ReA (104), making it likely that monocytes (143) may be involved as the vehicle transporting chlamydiae from the genital epithelium to the synovium. Chlamydial DNA in the joints of ReA patients does not correlate with the immune response, whether measured by antibodies or by lymphocyte proliferation (181). Actually, an inverse relationship between the presence of chlamydial DNA in the joint and the synovial Chlamydia-specific lymphocyte proliferation was found. This suggests that an impaired T-cell response might contribute to the persistence of bacteria. In synovial tissue, the transcript for Omp1, the major outer membrane protein of chlamydiae, which is the major target of the humoral immune response, was not detected, but transcripts for other chlamydial proteins, e.g., heat shock protein 60, were found (59).

Antigens of the enteric pathogens Yersinia enterocolitica, Salmonella enterica serovar Enteritidis, and S. enterica serovar Typhimurium have been identified in joints by using immunohistochemistry with samples of synovial tissue or synovial fluid (64, 70). Lipopolysaccharide LPS of Yersinia, Salmonella, Shigella (62, 63, 64), Chlamydia trachomatis (88), and the YadA protein (70) and 61-kDa heat shock protein (66) of Yersinia are antigens found in the joints from ReA patients. Yersinia lipopolysaccharide (LPS) and hsp in synovial fluid and peripheral blood cells were found for up to 4 years in ReA patients (66), which explains the fact that patients who develop ReA have long-lasting and strong immunoglobulin A responses. Salmonella LPS has been found in synovial tissue 2 years after the onset of infection (62), and other indirect proof of the persistence of Salmonella in the host is the presence for several months of antibodies at high concentrations in serum samples from patients with ReA, whereas antibody concentrations decline in salmonellosis patients without arthritis (129).

Chlamydia resides intracellularly in the synovial tissue (110) (rather than in the fluid [26]), displaying an aberrant morphology (irregular shape rather than the standard-size, spherical reticulate body) (143). An in vitro model showed that in the aberrant form that characterizes persistent C. trachomatis infection, copies of the chromosome of the bacterium accumulate in the apparent absence of subsequent cell division (19) and the expression of omp1 (the gene encoding the major outer membrane protein [MOMP]) is attenuated during persistence. These issues help to provide some "invisibility" from immune surveillance. The upregulation of the expression of the chlamydial hsp60 gene (which specifies a highly immunogenic protein) (59) is important for the inflammatory response that characterizes Chlamydia-associated arthritis.

Chlamydia inhibits the apoptosis of host cells by inhibiting the release of mitochondrial cytochrome c and also by directly engaging the death domains of the tumor necrosis factor (TNF), receptor family (48, 230). Chlamydia-infected monocytes induce apoptosis of autologous T lymphocytes through the release of TNF- (88). Similarly, Y. enterocolitica also induces macrophage apoptosis by suppressing the activation of NF-B, which inhibits TNF- release. Thus, both Chlamydia and Yersinia interfere with immune cell activation of T cells and macrophages, and this may critically affect the host immune response and the resulting outcome of infection with these microorganisms (48, 162, 230).

On the other hand, Chlamydia directly stimulates infected host cells to upregulate proinflammatory or immunostimulatory soluble mediators (TNF-, interleukin-6 [IL-6], IL-1, inducible nitric oxide synthase, and alpha interferon [IFN-]) (211). If elementary bodies are taken up by macrophages/monocytes, these become activated. Alternatively, organisms may be processed by dendritic cells or other antigen-presenting cells, leading to pathogenic responses by both T and B lymphocytes. Chlamydia polyclonally stimulates both T and B cells in humans and experimental animals to secrete cytokines that further promote inflammatory responses in tissues (121).

Enteric ReA. In the enteric forms of ReA, the persistence of bacterial antigens may explain the appearance of an inflammatory reaction in the synovium. It is likely that these bacteria survive at an extra-articular site, in particular in the mucosal membranes of the digestive system and/or the lymphatic ganglions, and are carried to the joint by monocytes, probably in recurrent fashion (66, 101, 102). Mucosal leukocytes collected from patients with an inflammatory bowel disease (which could be extrapolated to ReA) can bind well to synovial vessels (166), and after reaching the joint, the bacterial antigens can persist in the synovium.

Yersinia and Salmonella can persistently infect the mucosa of the intestine and the digestive ganglions but not the synovium (66). These microbes are also present in monocytes, which may serve as a reservoir and "transporter" from extra-articular sites into the joint. A recent study investigated the mechanisms involved in the invasion, degradation, and persistence of Yersinia and Salmonella in synovial fluid (138). Infection of synovial fluid with these bacteria started with the adhesion of the bacteria to synovial cell membrane and was followed by uptake of the morphologically intact bacteria into the cells. A few hours after infection, the bacteria showed very low metabolic activity and started a process that finally led to the total disappearance of the bacterial cytosol (nucleic acid-free bacterial rods known as "ghosts").

Yersinia surface structures are critical for the adherence and internalization of these bacteria. Yersinia adhesin (YadA) binds specifically to joint collagens, which may contribute to their arthritogenic potential (30). The proteins necessary for Yersinia host invasion are not constitutively expressed. They are encoded by genes in a virulence-responsible plasmid and are designated Yops (for "Yersinia outer proteins") (80). These genes are activated only when the bacteria are in proximity to host cells. YopB and YopD decrease the secretion, by cultured human intestinal and cervical cell lines in vitro, of IL-8, a potent chemoattractant of polymorphonuclear leukocytes (173). The Y. enterocolitica virulence factors invasin, protein tyrosine phosphatase, cytotoxin, and adhesin interfere effectively with the microbicidal action of neutrophils. This could enable the extracellular survival of Yersinia in host tissues (162), and such persistence may permit the initiation of an inflammatory process. In addition, Y. enterocolitica suppresses the cellular activation of NF-B, which inhibits TNF- release and triggers apoptosis in macrophages (162).

A natural epitope derived from a Yersinia protein is presented by major histocompatibility complex MHC class I even though the microbes are found mostly extracellularly and, when seen in an intracellular compartment, remain entirely within the vacuole. Such presentation may subject the infected cell to recognition by cytotoxic T lymphocytes (168). Y. enterocolitica, as well as S. enterica serovar Typhimurium, has the ability to induce the apoptosis of macrophages, in vitro, Yersinia was shown to perturb dendritic cell function (167, 189), mechanisms that might be important for bacterial survival (80).

Yersinia has two envelope proteins that are cross-reactive with the thyrotropin receptor and have a mitogenic activity on B cells (231). Patients with elevated levels of anti-Yersinia antibodies are prone to thyroid gland disturbances (116). A structural similarity between the human thyrotropin receptor and Y. enterocolitica has been described and considered a possible mechanism for autoimmunity (205).

Shigella infects only digestive epithelial cells, and transportation by monocytes is highly unlikely because this bacterium rapidly kills monocytes through apoptosis. Shigella DNA has not been convincingly demonstrated in the joints of ReA patients, making it more likely that only pieces of bacteria are transported to the joints. It has been reported that only Shigella strains containing the pHS-2 plasmid are arthritogenic (180, 191). pHS-2 influences invasiveness and intracellular and intercellular movement and hence is important for the mechanism of infection.

The potential role of the commensal microflora in the development of ReA merits discussion (38). This issue is of great relevance, considering that the majority of disease-causing bacteria from the intestine may have been derived from commensals that have acquired genes from foreign sources, turning them into pathogens (68). A variety of commensal microorganisms including Pseudomonas spp., Salmonella spp., Bacillus cereus, and Lactobacillus spp. have been identified by PCR in the synovium of patients with ReA (171), and the conventional intestinal microflora plays an important role in the development of arthritis in the HLA-B27 transgenic-rat model (195). In addition, the exact role of and/or relationship between the microflora and the compromised function of the intestinal barrier needs to be established.

For years, an intensive search for the decisive arthritogenic antigen has been going on. Bacterial hsp60 seems to be a major target of the T-cell response in ReA, and cross-reactivity against autologous hsp60 has been discussed as a cause of autoimmunity (180). Chlamydial hsp60, the Hc1 protein (a histone), and an OMP (Omp2) were found as target antigens recognized by CD4⁺ T cells (40). The T cells specific for hsp60 of Chlamydia did not cross-react with hsp60 from enterobacteria.

In Yersinia-induced ReA, patients have antibodies directed against a multitude of Yersinia antigens. Yersinia LPS, or YadA, seems to be essential for the induction of arthritis (180). A cationic 19-kDa urease B subunit, hsp60, and the ribosomal protein L23 of Y. enterocolitica O:3 were also identified as immunodominant antigens (10, 153). The urease subunit is highly conserved, and this observation lends support to the possibility that autoimmunity in reactive arthritis may be mediated by antigen mimicry between evolutionarily conserved epitopes (137). Also in Yersinia-triggered ReA, proteins secreted by the bacterium, including a protein with tyrosine phosphatase activity (YopH), are potent immunogens stimulating CD4⁺ T cells within the inflamed joint (109). The pathogen-reactive T-cell clones preferentially utilize a limited set of T-cell variable gene segments, suggesting the role of superantigens (193) in the pathogenesis of ReA.

Bacterial LPS. Bacterial LPS plays an important role in the pathogenesis of ReA (128) because it contributes greatly to the virulence of the bacteria and acts to modulate the immune system. Monocytes that phagocytose Y. enterocolitica O:3 are a constant source of a membrane-active form of LPS in their microenvironment (226). In patients with Salmonella-triggered ReA, the persisting humoral immune response and intra-articular antibodies are directed primarily against LPS (130).

Intra-articular LPS are powerful macrophagic stimulators and can trigger the synthesis of proinflammatory cytokines (TNF-, IL-1, and IL-6) through CD14-Toll-like receptor activation (208, 232). LPS also enhance lymphocyte entry into joints by their effects on the vascular endothelium, induce the synthesis of monocyte chemotactic protein 1 in human articular chondrocytes (213), enhance the secretion of the polymorphonuclear leukocyte chemotactic and activating cytokine IL-8 from chondrocytes (127), and decrease C5aR expression on monocytes (94). All these changes contribute to the recruitment of leukocytes into the synovium and lead to the persistence of mononuclear phagocytes within the inflamed synovium. LPS, together with macrophages, contributes to the degradation of cartilage matrix by collagenases and other neutral proteases secreted by chondrocytes (87). LPS also suppresses proteoglycan synthesis and induces the production of nitric oxide by chondrocytes (164). In response to LPS, monocytes produce tissue factor which is associated with thrombotic complications (226).

Bacterial DNA. Bacterial DNA, which differs from that of eukaryotes by the presence of more nonmethylated CpG motifs, can intensely stimulate monocytes/macrophages and thus could be a target antigen. Experimental intra-articular injection of bacterial DNA is sufficient to trigger arthritis in mice (42).

A direct role of HLA-B27 makes the strongest genetic contribution to the development of SpA, although how much and how it contributes to ReA are important questions. The association of B27 and ReA is illustrated by the fact that the prevalence of disease in B27-positive individuals is five times greater than in the general population. In B27-positive relatives of patients with ReA, the prevalence is another 10 times greater (49). The generation of the HLA-B27 transgene in both rats and mice leads to the development of arthritis, which is another way of validating the role of the HLA-B27 gene in arthritis (195). In transgenic animals, susceptibility to disease is clearly related to gene copy number and level of expression of B27, and the specificity of the peptides bound to B27 appears to play an important role in the pathogenesis of arthritis.

The role of HLA-B27 may be related to its major function as a class I antigen (presenting antigenic arthritogenic peptides to cytotoxic lymphocytes) (170, 174). The arthritogenic peptide hypothesis, stating that a microbial or self-antigen is presented to CD8⁺ T cells, has received some attention. Molecular mimicry between HLA-B27 and bacterial molecules has also been a consideration because similar amino acid sequences in HLA-B27 and Yersinia (194) or Shigella proteins have been found (109); thus, cross-reactivity could lead to tolerance and persistence of organisms.

Recently, it was found that the heavy chains of HLA-B27 possess an unpaired cystein at position 67 in the B pocket of the peptide binding groove, which may lead to the formation of unusual forms of HLA-B27, i.e., homodimers and heterodimers, which are subsequently recognized by CD4 T cells through their T-cell receptors and NK receptors (4, 5, 20, 131). Such dimers can bind peptides and apparently stimulate CD4⁺ T cells. HC-B27 is more abundant in patients with HLA-B27-associated disease than in healthy B27 individuals. These data could explain why a class I molecule is associated with a disease driven by T-helper cells. It also explains why mice transgenic for HLA-B27 develop arthritis only when lacking ß₂-microglobulin (99). ß₂-Microglobulin inhibits the formation of HC-B27 by binding to monomeric B27 heavy chains (23).

A different hypothesis about the role of HLA B27 arose from the observation that HeLa cells transfected with HLA-B27 respond to in vitro bacterial invasion with signals not observed in control cells (79, 227). This could explain why HLA-B27 cells accommodate bacterial survival differently from control cells. The explanation of the different signals could be that HLA-B27 differs from other HLA alleles in that when maturing inside the endoplasmic reticulum, the protein is folded significantly more slowly (136). Colbert et al. showed that in ReA during the antigen processing and assembly pathway in the endoplasmic reticulum, HLA-B27 has a tendency to misfold even without any ß₂-microglobulin or peptide deficiency. This misfolding implicates the B pocket of the molecule and can lead to a stress response, which could increase the production of proinflammatory cytokines by activation of the NF-B (35, 36)

Additional mechanisms have been proposed to explain the role of HLA-B27. In some genetically predisposed subjects, HLA-B27 appears to lack the ability to eliminate infected macrophages normally (92, 111), improving the intracellular survival of pathogens (183). Indeed, it has been shown that the adhesion molecules of certain bacteria (Yersinia and Salmonella) use HLA-B27 as a ligand to attach to cells of the synovial environment. The expression of most of the HLA-B27 epitopes is decreased on monocytes that have ingested Yersinia or Salmonella in vitro (93, 227). This change of expression is mostly posttranslational, although the synthesis of HLA-B27 is also decreased. Interestingly, S. enterica serovar Typhimurium and Y. enterocolitica also induce alternative splicing of the pre-mRNA of HLA-B27 with removal of the fifth exon. Because this fifth exon encodes the transmembrane domain, cell-free soluble HLA-B27 is generated (75).

In vitro the elimination (but not the uptake) of S. enterica serovar Enteritidis from HLA-B27-transfected U937 cells was remarkably impaired compared with that from HLA-A2 transfectants serving as controls (105). This suggests that HLA-B27 itself may influence the intracellular persistence of arthritis-triggering Salmonella strains within monocytes. In vivo endogenous HLA-B27-positive primary human fibroblasts did not differ from HLA-B27-negative cells in the ability to support invasion and in the persistence of S. enterica serovar Enteritidis (78).

The gut. The increased frequency of subclinical inflammatory lesions in the gut in patients with SpA (139) and the known association between inflammatory bowel diseases and SpA support a pathogenic link between gut inflammation and SpA that is independent of HLA-B27 (186). Animal models of SpA emphasized the role of the intestinal flora in pathogenesis. In HLA-B27 transgenic animals, both colitis and arthritis develop only in animals that have been transferred from germ-free or specific-pathogen-free conditions to normal animal facilities (196).

A suggested hypothesis is that during the acute infection in the gastrointestinal mucosa, other bacterial antigens also penetrate the destroyed epithelial layer and meet the gut-associated lymphoid tissue, leading to stimulation of the leukocytes. Lymph nodes and the gut submucosa have been suggested as potential reservoirs of arthritis-triggering bacteria, since it seems improbable that bacterial proteins could persist for long in the absence of any continuing source of infection (209). Yersinia structures have been shown to persist in the submucosa of the gut (73) and in lymph nodes of patients with prolonged or chronic Yersinia arthritis (209). Thereafter, T cells primed to local bacteria in the gut might see the same or related bacterial antigens or cross-reacting self antigens in peripheral or sacroiliac joints (45), leading to the development of the rheumatic manifestations (226).

The joints. The joints behave in many ways like part of the reticuloendothelial system and are prime sites for lodging of circulating infectious agents (169). Microbes or their components may enter the joints via blood vessels either as whole organisms that may circulate in the blood or within the cells or as a part of immune complexes (62, 172). Peripheral blood mononuclear cells that contain intracellular bacterial fragments are especially prone to bind to synovial high endothelial venules and transmigrate through the endothelial cell monolayer (101). Monocytes that have phagocytosed ReA-triggering agents can induce the expression of P-selectin on cultured endothelial cells; this molecule is important for homing to the synovium (166).

The discovery of C. trachomatis DNA by PCR in synovial samples from healthy volunteers and patients with osteoarthritis (OA) (171) and the simultaneous presence of RNA of numerous bacterial species in synovial samples from patients with rheumatoid arthritis (RA) or OA demonstrate that the synovium is not a sterile structure but more probably is an interfacial zone which can be colonized by bacteria originating from the environment and the endogenous flora (175). Such intra-articular microbes, depending on the characteristics of the host, may be eliminated or may provoke a synovial infection.

There are definite changes at the synovial level in ReA. Unlike rheumatoid arthritis, in which synovitis is the initial or primary lesion, the synovitis of SpA is a secondary event after enthesitis, at least in some joints (135). Arthroscopy shows that the synovial vascular pattern is typically tortuous and bushy (in contrast to RA, where the vessels are straight and branching) (155). A comparative morphometric analysis of cell apoptosis in various forms of arthritis showed no difference between ReA and other arthritides (32).

In ReA, the antibacterial Th1 cytokine response (production of IFN-, IL-2, and IL-12), which is necessary for the elimination of ReA-associated bacteria, is impaired. On the other hand, a Th2 response (IL-4 and IL-10) predominates and contributes to bacterial persistence in the joints (30, 228). The pathogenesis of this Th1-Th2 imbalance is unclear, but it is likely that genetic factors of the host such as the polymorphism of cytokine genes (90) are causally involved (175). There is some evidence that TNF- genotypes which seem to be associated with low TNF- production are present at a higher percentage in ReA (180). Also, mice that are knockouts of the p55 receptor are more susceptible to infection by Yersinia and develop more severe arthritis (234).

Impaired peripheral-blood T-cell responses to ReA-triggering agents have been demonstrated. In patients with Yersinia-triggered ReA, there is a lower frequency of Yersinia-reactive T cells in peripheral blood than in patients with uncomplicated yersiniosis (214). When synovial T cells from ReA patients were stimulated in vitro with the triggering bacteria, the IFN-/IL-10 and the TNF-/IL-10 ratios were clearly lower (228) compared with the response to Borrelia burgdorferi in Lyme arthritis patients. Because IFN- and TNF- are crucial for the elimination of bacteria, the insufficient production of these cytokines in ReA patients contributes to bacterial persistence. Furthermore, ReA patients with lower TNF- secretion have a longer disease duration than patients with higher secretion (30).

These findings may indicate a defective first-line defense in the patients developing arthritis that allows the bacterial antigens to reach the synovium, where they then initiate the specific T-cell responses (226). In addition, cytotoxic T lymphocytes infected with Yersinia seem to have a reduce lytic capacity against the infected target cells (2).

CLASSIFICATION CRITERIA

Top Previous Next References

 
The diagnostic, classification, and inclusion criteria of patients with ReA are heterogeneous (82). Wilkens et al. in 1981 proposed the preliminary criteria for Reiter's syndrome (219). Considering that ReA is a member of the SpA group, two sets of criteria, the Amor criteria and the European Spondyloarthropathy Study Group criteria, could be used for ReA diagnosis (43). Although studied as classification criteria, they are useful and valid as diagnostic criteria (7). In 1996, a group of experts during the Third International Workshop on Reactive Arthritis proposed for research purposes a set of nonvalidated criteria (178). These criteria are as follows: typical peripheral arthritis (predominantly lower limb, asymmetric oligoarthritis) plus evidence of preceding infection (when there has been clear clinical diarrhea or urethritis in the preceding 4 weeks, laboratory confirmation is desirable but not essential; when there has been no clear clinical infection, laboratory confirmation of infection is essential); laboratory tests for preceding infection include positive stool cultures, detection of C. trachomatis in the first portion of the morning urine or in a urogenital swab, anti-Yersinia and anti-Salmonella antibodies to LPS or other specific antigens of immunoglobulin G (IgG) plus IgA or IgG plus IgM subtypes, IgG, IgM, and IgA antibodies to C. trachomatis (sensitivity and specificity regarded as low), and detection of chlamydial DNA in the joint by PCR. The experts at the Workshop also state that patients with other known causes of mono- or oligoarthritis, such as other defined SpAs, septic arthritis, Lyme disease, and streptococcal ReA, are exclusion criteria for ReA. The presence of HLA-B27, extra-articular features of Reiter's syndrome (conjunctivitis, iritis, skin lesions, noninfectious arthritis, and cardiac and neurologic features), or typical SpA features (inflammatory back pain, alternating buttock pain, enthesitis, and iritis) are not required for diagnosis.

Another proposal for the classification of patients entering clinical and experimental studies has recently been published by a Mexican group (149). The sensitivity, specificity, and positive predictive value of any single or combination of criteria require evaluation in a prospective study.

EPIDEMIOLOGY

Top Previous Next References

 
The prevalence of ReA has been calculated as 0.1% (27, 107), with incidence rates of 4.6 to 13 per 100,000 for postvenereal ReA and 5 to 14 per 100,000 for postenteric ReA (85, 107). In a recent Swedish study, the total annual incidence of ReA was 28 in 100,000 (188). An underestimation of ReA cases is possible mainly because of two reasons. First, a clear differentiation among SpA subgroups, especially in their early stages, may not always be possible owing to overlapping clinical features (107, 160, 179). To exemplify this, in a community-wide epidemiologic study, Boyer et al. found that 36% of the ReA patients had not been correctly classified (24). Second, and very important for clinicians, is that as many as 36% of chlamydial and 26% of enteric ReA cases have asymptomatic triggering infections (107, 222).

While enteric ReA occurs most commonly following dysenteric outbreaks, sexually acquired ReA is the endemic form of ReA (81, 160). The causative infection is reported to be identified in 40 to 56% of patients with ReA. Genitourinary tract infection with C. trachomatis is the more commonly recognized cause of ReA in developed countries (in the United States it has been identified as the preceding infection in 42 to 69% of patients with urogenic ReA). Infections with enterobacteria are more common triggers in developing countries (98). Yersinia or Salmonella are the causative pathogen in 52% of patients with enteric ReA, with Salmonella being more common than Yersinia (33 and 18%, respectively). Although Yersinia is a relevant pathogen in continental Europe, it seems to be rather rare in the United Kingdom and United States (50).

Changes in the epidemiology of ReA have been reported. Recent data indicate a 30% decrease in prevalence of Yersinia arthritis in Finland, with no changes in the number of reported cases of Salmonella-triggered ReA. This decrease is probably related to a decline in the rate of sexually transmitted disease secondary to the awareness of human immunodeficiency virus (HIV) infection. In contrast, in South Africa, where ReA used to be extremely rare and where HLA-B27 has a prevalence of less than 1%, the prevalence of ReA is now increasing in the wake of the HIV epidemic (98, 142, 223).

ETIOLOGIC AGENTS

Top Previous Next References

 
With the increasing awareness of ReA, the list of causative agents is still growing, and it has become clear that different infectious agents can trigger the same clinical picture (53). Table 2 lists the ReA-triggering agents and their association with HLA-B27. Although efforts have been made to characterize the microorganisms linked to ReA, no definite common feature has so far emerged. Yersinia, Salmonella, Shigella, Chlamydia, and Campylobacter (pathogens causing HLA-B27-associated disease) all have a capacity to cause the primary infection in mucosal tissues, have LPS as a part of their outer membrane, exert intracellular parasitism, have a similar type of peripheral joint involvement, and contain virulence factors which help them to evade the host defense mechanisms and survive and disseminate in the body (61, 200, 226).

C. jejuni is the leading cause of bacterial gastroenteritis in many Western countries. The incidence of ReA following C. enterocolitis is 2 to 3% (range, 0.6 to 24%) (46). The calculated incidence of ReA after Shigella infection is 1 in 10³. In a localized outbreak with the same agent, the rate of arthritis was 1.2% (124, 185). The reported incidence of Salmonella-induced ReA is between 1.2 and 7.3%, and in outbreaks the frequency was as high as 19% (83, 125, 212). Concerning Salmonella serotypes, 60% of Salmonella-induced ReA cases are triggered by S. enterica serovar Typhimurium and 25% are triggered by S. enterica serovar Enteritidis (128, 147). The most common triggering agents in childhood bacterial reactive arthritis are Y. enterocolitica and Y. pseudotuberculosis; the incidence ranges from 5 to 35% (194).

CLINICAL ASPECTS

Top Previous Next References

 
ReA affects people in the second to fourth decades of life and occurs 1 to 4 weeks following genitourinary (male-to-female ratio, 9:1) or enteric (male-to-female ratio, 1:1) infections (12, 95). ReA not only affects the joints but also is a systemic disease, and serious visceral involvement (cardiac, renal, or neural abnormalities) may develop. In the acute phase of ReA, no simple relationship exists between extra-articular features and the nature of the inciting infection, and the clinical characteristics of joint disease associated with C. trachomatis infection and those associated with postenteric arthritis are virtually identical.

ReA combines four syndromes: (i) an enthesopathic syndrome (135), (ii) a peripheral arthritis syndrome (an asymmetric acute or subacute oligoarthritis of the lower limbs), (iii) a pelvic and axial syndrome (spinal involvement with sacroiliitis), and (iv) an extramusculoskeletal syndrome. Table 4 summarizes the clinical manifestations of ReA, and Table 5 summarizes the sensitivity and specificity of some of them.

Diffuse swelling of an entire finger or toe ("sausage digit") occurs in 16% of patients and has a sensitivity of 27% but a high specificity of 99% (103). Enthesitis (inflammation of the ligaments and tendons at the sites of their insertion into the bone) in the European Spondylarthropathy Study Group study was found in approximately 30% of ReA patients and in approximately 20% of control patients with no SpA. Secondary to enthesitis, patients may have heel pain (sensitivity of 52% and specificity of 92%), Achilles tendonitis, or pain at the insertion of the patella tendon into the tibial tubercle.

Inflammatory low back pain (sensitivity of 71% and specificity of 77%) secondary to sacroiliitis or spondylitis or due to inflammation of ligamentous or tendinous insertions in the ischial tuberosity is another distinctive feature (183). Affection of the cervical spine is unusual but may occur and may even result in atlantoaxial subluxation (54).

Urethritis may be a principal feature of ReA even in some postdysenteric cases. As a precipitating event, it precedes the symptoms of ReA by 1 to 3 weeks. Between 70 and 80% of women and a substantial proportion of men infected with C. trachomatis have no genitourinary symptoms (160). Nonspecific urethritis is mild, painless, and associated with nonpurulent clear mucoid urethral discharge. Prostatitis is common and has been found up to 80% of patients. Hemorrhagic cystitis may develop (97). In female patients, nonspecific or mucopurulent cervicitis with or without easily induced cervical bleeding and/or abdominal pain may occur (12, 144).

Gastrointestinal symptoms are absent or mild in ReA triggered by Yersinia, but more severe and of longer duration in patients with Salmonella- and Campylobacter-induced ReA (50, 124). The interval from Salmonella infection to arthritis is <3 weeks in most patients. In cases of postdysenteric ReA, the longer the duration of diarrhea, the greater the likelihood of developing ReA (83, 125, 183). This phenomenon may reflect an abnormality at the level of the gut mucosal defense, allowing for enhanced colonization or delayed clearance of the pathogen. When ileocolonoscopy is carried out in patients with established ReA, either macroscopically or microscopically detectable lesions resembling ulcerative colitis or Crohn's disease are seen in almost 70% of cases. Gut inflammation is mainly subclinical, but there is a close correlation between disease activity in the joints and activity in the gut (140, 200). Even patients with sexually acquired ReA have colonoscopic lesions in about 25% of cases, suggesting that in some cases these lesions represent yet another extra-articular manifestation of SpA (100).

Mucocutaneous lesions are very specific (9). Keratoderma blennorrhagica or pustulosis palmoplantaris (5 to 30%) begins on the palms and soles as pustules; they gradually become covered with thick, horny crusts, and neighboring lesions may become confluent. Clinically and microscopically, these cutaneous lesions are difficult to differentiate from pustular psoriasis. Circinate balanitis (20 to 40%) is a well-defined, painless erythematous lesion with small, shallow ulcers of the glans penis and urethral meatus. Oral ulcers (5 to 10%), located in the hard and soft palate, gingiva, tongue, and cheeks, are characteristically painless, erythematous, and superficial. Hyperkeratotic nail (6 to 12%) and skin lesions, not easily distinguishable from pustular psoriasis, could also be present. Erythema nodosum is a well-recognized feature of Yersinia infection that typically occurs in women, especially those who are HLA-B27 negative, and the lack of gastrointestinal symptoms in these cases is also documented. Yersinia infections can account for up to 15% of patients in erythema nodosum series, and these cases may even mimic sarcoidosis (57).

Cardiac manifestations include aortic disease and conduction delays. The aortic ring and ascending aorta are the usual sites of involvement, and aortic incompetence (dilatation-regurgitation) is usually a late finding and is often asymptomatic. ReA aortic disease may be difficult to distinguish from that of syphilis, since pathologically there is virtually no difference (74). Electrocardiographic abnormalities are recorded in 5 to 14% of patients. The most commonly reported conduction abnormality in patients with long-standing disease (more than 10 years) is first-degree atrioventricular (AV) block, which could progress to second-degree AV block (Wenckebach type) and complete heart block (41).

Proteinuria, microhematuria, or aseptic pyuria is seen in about 50% of patients with sexually acquired ReA and is usually asymptomatic. Glomerulonephritis and IgA nephropathy are rare. Severe systemic necrotizing vasculitis (21), thrombophlebitis, purpura, livedo reticularis, and amyloidosis (44) have been reported as rare complications in chronic disease (222). Transient neurologic dysfunction, such as peripheral or cranial nerve palsy, Parsonage-Turner syndrome, and hemiplegia have also been described.

LABORATORY FINDINGS

Top Previous Next References

 
Laboratory abnormalities are nonspecific and include mild normocytic anemia, moderate neutrophilic leukocytosis, and elevated erythrocyte sedimentation rate, C-reactive protein level, and serum complement levels. Urinalysis should be carried out at presentation and should be repeated during follow-up to detect possible aseptic pyuria due to urethritis. Antinuclear antibody and rheumatoid factor are negative. Antineutrophil cytoplasmic antibodies against lactoferrin and myeloperoxidase were found in one study in 12 of 22 patients with ReA, but this is not supported by others (126, 192).

Joint fluid should always be aspirated when possible. Synovial fluid may be mildly to severely inflammatory and may contain large macrophages with vacuoles that contain nuclear debris and whole leukocytes which are called Reiter's cells and are not specific for ReA. Synovial tissue and synovial fluid cultures are negative. Crystals are not present. Synovial biopsy is not of diagnostic value since the inflammation is nonspecific (inflammatory changes, including vascular congestion and perivascular polymorphonuclear cell infiltration) but may be valuable in the exclusion of alternative diagnoses.

RADIOGRAPHIC FINDINGS

Top Previous Next References

 
Although radiographic findings are not essential for diagnosis, they are important in the evaluation of patients with ReA. Early in the disease, radiographs may be entirely normal. Acute attacks of arthritis may be accompanied by soft tissue swelling. With repeated episodes of arthritis, permanent radiographic abnormalities may appear in up to 20% of patients (97).

The most characteristic sites of involvement are the small joints of the foot, calcaneus, ankle, and knee. In the axial skeleton, the sacroiliac joints, spine, symphysis pubis, and manubriosternal joints are frequent target areas. Enthesopathic lesions can be visualized at an early stage by focal uptake of 99 mTc-methylene diphosphonate and by magnetic resonance imaging.

Asymmetric reactive bone proliferation at the sites of inflammation is a helpful radiographic feature for diagnosis. Later fluffy periosteal reaction and erosions may be seen at the sites of ligamentous and tendon insertions, especially the Achilles tendon, the plantar tendons, and the fascia. Although magnetic resonance imaging has been used for diagnosing enthesitis, ultrasonography is the preferred technique, being both noninvasive and sensitive (148). Ultrasonography is also useful in demonstrating plantar fasciitis.

Erosive joint damage especially affects the small joints of the feet, with 12% of patients exhibiting foot deformities (144). Subchondral sclerosis, eburnation, and adjacent periostitis, as well as intra-articular bony ankylosis in the small joints, have also been observed.

Asymmetric sacroiliitis can be demonstrated radiologically in over one-third of patients with chronic ReA (133, 157, 163). In the early stages of sacroiliitis when radiographs may be negative, computed tomography may be useful for demonstrating early erosions and sclerosis. Magnetic resonance imaging is more sensitive than computed tomography in assessing cartilage changes, however. In the spine, asymmetric paravertebral ossification involving the lower three thoracic and upper three lumbar vertebrae is characteristic.

DIAGNOSIS

Top Previous Next References

 
The diagnosis of ReA is still a problem because the preceding bacterial infection is often asymptomatic, identification of the etiologic agent is difficult at the time when the arthritis occurs, and no validated criteria for the diagnosis of ReA are currently available (149, 179). Criteria from the American College of Rheumatology (ACR) require that the diagnosis of ReA be based on a documented urogenital C. trachomatis infection or a gastrointestinal infection by a relevant enteric pathogen. Patients meeting all ACR criteria except that specifying a previous infection usually are classified as having undifferentiated oligo- or monoarthritis (84). On the other hand, as screening methods for identification of Chlamydia have improved, patients who have chlamydial antigens or nucleic acids in the joint but who do not meet ACR criteria for ReA are being recognized. Some authors designate these patients as simply having Chlamydia-associated arthritis (84, 221).

There is no single category of laboratory tests that can be used to diagnose ReA and to determine the responsible organism (28). Currently, the best diagnostic results come from combining laboratory and clinical information (56). The diagnostic criteria for ReA shown in Table 6 were used in a recent study by Fendler et al. (50) and resemble those used in other studies (71, 107, 113). It has been reported that elevated C-reactive protein levels, genitourinary symptoms, metatarsophalangeal joint involvement, and HLA-B27 could predict the diagnosis of ReA with a sensitivity of 69.2% and a specificity of 93.5% (106).

The microimmunofluorescence test has been criticized for its requirement for considerable expertise, its subjective interpretation, and its unsuitability for testing large numbers of specimens (15). Chlamydia culture has been considered the gold standard for diagnosis (198); however, its sensitivity has been estimated at only 70 to 80% or less for women with cervical infection. This has led to the suspicion of "inapparent-nonculturable" chlamydial infections in humans and the need for better diagnostic tools.

The search for Chlamydia in the first portion of the morning urine by PCR or by ligase chain reaction seems to be an acceptable and relatively easy diagnostic approach with a result comparable to that demonstrated by urogenital swab (154). Antigen-detection tests (direct fluorescent-antibody assay, enzyme immunoassay [EIA], and immunoblotting), and nonamplified nucleic acid hybridization, as well as newer technologies based on amplified DNA assays (PCR, ligase chain reaction, strand displacement assay, hybrid capture system, and transcription-mediated amplification of RNA) may provide improved sensitivity, lower expense, availability, or timeliness of results over culture (15, 145). The sensitivities and specificities of nucleic acid amplification tests are all high, ranging from 82 to 100%. The sensitivity of antigen detection tests (EIA and direct fluorescent-antibody test) is slightly lower (70 to 80%), but the specificity remains high (96 to 100%) (144).

Differentiation between the Chlamydia species is crucial for diagnosis because of a high degree of cross-reactivity and a high prevalence of anti-C. pneumoniae antibodies. A substantial proportion of patients with Chlamydia-induced ReA can be antibody negative, resulting in low sensitivity. A peptide-based EIA detecting IgA and IgG antibodies against C. trachomatis in serum samples from patients with Chlamydia-triggered ReA has a sensitivity of 74% and a specificity of 84% (146). Antibodies of the IgG subclass alone do not sufficiently reflect a recent infection, because their levels can be elevated for months after an infection; therefore, determination of IgG antibodies should be combined with tests for IgM or IgA antibodies, with the last two indicating an acute or persistent infection.

As antigens, whole C. trachomatis or Chlamydia specific MOMP and/or LPS (13, 17) have been used. The specificity and sensitivity of currently available tests is not higher than 78 and 73%, respectively (182). Enzyme-linked immunosorbent assays (ELISAs) using synthetic peptides derived from species-specific epitopes in the variable domain IV of the MOMP of C. trachomatis and not homologous to C. pneumoniae have been developed (16). The determination of synovial fluid IgG anti-MOMP antibody could be particularly useful since a sensitivity and specificity of roughly 80% are observed (17). The specificity reaches 90% with the determination of synovial fluid IgA anti-MOMP antibody. A major disadvantage of these tests is that they are only qualitative (17). Another ELISA is based on an exclusively Chlamydia-specific recombinant fragment of the total LPS. This genus-specific antigen allows the detection of antibodies against C. trachomatis, C. pneumoniae, and C. psittaci (17). The reported sensitivity, specificity, and predictive values of these tests are showed in Table 7.

Salmonella and Yersinia antibodies can be studied by the agglutination test (Widal) and ELISA. The Widal agglutination test detects IgM antibodies and can therefore be used as evidence of recent bacterial infection, but, especially for Salmonella, its sensitivity is relatively low (64%), even in acute infections (178). Agglutination when only IgA antibodies are present may give false-negative results. ELISA distinguishing between antibodies to different immunoglobulin classes appears preferable to the agglutination technique. Combining IgG and IgA or IgM, a specificity of 90% can be assumed for Yersinia- and Salmonella-specific serologic tests.

For Yersinia infections, most data have been obtained by using an EIA with a sodium dodecyl sulfate extract of LPS from Yersinia or by using an immunoblot analysis with Yersinia OMP as the antigen (182). In acute Yersinia-induced ReA, IgG plus IgA antibodies can be detected in close to 100% of patients. A persistence of IgA antibodies in the sera of patients with Yersinia arthritis for 14 to 16 months after onset of the infection is found in 84% of the patients (67), and peak levels of IgA correlate directly with the severity of arthritis (56, 67). In patients without arthritic symptoms, these antibodies disappear in 5 months. IgG antibodies to Yersinia also persist longer in arthritic patients than in patients without arthritis, but not as consistently as IgA antibodies. IgM antibodies persist for only 1 to 3 months after the onset of infection (67). It is recommended to test for IgG, IgM, and IgA isotypes in patients with acute ReA and for IgG and IgA isotypes in patients with chronic ReA (56, 84).

ELISA is by far the most sensitive method to detect anti-Salmonella antibodies acute and late infections), it has a sensitivity of approximately 92% (125). These antibodies levels tend to persist longer among patients with ReA (9 to 14 months) than in those with uncomplicated enterocolitis (4 months) (129). In salmonellosis, a response in all immunoglobulin classes is seen, in contrast to Yersinia-triggered ReA, where IgA antibodies and particularly those of the secretory IgA class and the IgA2 subclass persist (125). No good serologic test is available for the diagnosis of Shigella infections (50).

Finally, because simultaneous multiple sexually-transmitted infections are common, serologic tests for HIV, syphilis, and gonococcal cultures should be done in patients with sexually transmitted ReA (144).

Chlamydial DNA and mRNA has been found in synovial membrane biopsy specimens, and C. trachomatis is found by PCR in peripheral blood cells but not in serum of patients with early Chlamydia-induced ReA (96, 179). Positivity for Chlamydia is found by PCR in 65% of patients with Reiter's syndrome and 42% of patients with other types of ReA. In patients with undifferentiated arthritis but with a joint pattern compatible with ReA, chlamydial DNA has been detected by PCR in approximately 30% of cases. Chlamydial DNA was also found in control groups such as patients with RA (21%), patients with OA (35%), and healthy subjects (172, 220). The above results, along with data showing the lack of correlation between the presence of chlamydial DNA in the joint and a specific cellular or humoral immune response to the organism (220), limit the usefulness of PCR as a diagnostic test (38, 86). There is no agreement at the moment about the optimum technique to detect Chlamydia by PCR (182). Salmonella- or Yersinia-specific PCR for joint material does not currently play a role in the diagnosis of ReA.