SEARCH

SEARCH BY CITATION

Introduction

  1. Top of page
  2. Introduction
  3. Calculation of posttest probabilities for the diagnosis of ReA
  4. Bacteria triggering ReA
  5. HLA-B27 and spondylarthropathy-specific features in the diagnosis of ReA
  6. Combination of tests for the identification of the triggering bacterium
  7. Posttest probability of ReA in the case of negative test results
  8. Future possibilities to achieve a higher posttest probability
  9. A need for diagnostic criteria in ReA
  10. Summary and conclusion
  11. REFERENCES

Reactive arthritis (ReA) occurs after a preceding infection of the urogenital tract with Chlamydia trachomatis or after a preceding infection of the gut with enterobacteriae. HLA-B27 is positive in ∼50% of these patients (1) and some of them have features such as enthesitis, sacroiliitis, or uveitis. Because of these characteristics, ReA is regarded as part of the spondylarthropathies (SpA) (2). The typical joint pattern is an asymmetric arthritis without radiographic changes that develops predominantly in the lower limbs and presents as an oligo- or monarthritis in the majority of patients (2). Since patients with such a joint pattern constitute up to 50% of the patients in early arthritis clinics (3–5), ReA is an important differential diagnosis in daily clinical practice. The mean disease duration has been reported to be between 3 months and 6 months; however, a chronic course occurs in up to 20% of patients whose duration of symptoms is longer than 12 months (6, 7).

For the diagnosis of ReA, there is no single diagnostic test. Early criteria for the diagnosis of ReA relied nearly exclusively on clinical indicators of a preceding symptomatic infection, such as urethritis/cervicitis, or on symptoms characteristic of the whole group of SpA, such as enthesitis (8–10). However, it is likely that in a substantial proportion of patients with ReA, the preceding infection is asymptomatic or associated with only minor symptoms. This form of arthritis is currently labeled as undifferentiated arthritis or undifferentiated oligoarthritis (4, 11, 12). Furthermore, none of the criteria based on clinical symptoms and laboratory tests has been unequivocally accepted by the clinical and scientific community (13, 14).

Laboratory tests for the identification of the triggering bacterium are available (15) and are increasingly used in a diagnostic evaluation for ReA (3, 6). Most patients with new onset of arthritis are first seen by primary care physicians. Consequently, laboratory tests, including tests for the diagnosis of ReA, are often ordered by these doctors before patients are referred to tertiary rheumatology clinics. For the physician, an estimation about the likelihood that a diagnosis can be made if a test result is positive (“posttest probability of having the disease”) is of great interest (16). To date, a posttest probability has not been calculated for any of these tests.

Calculation of posttest probabilities for the diagnosis of ReA

  1. Top of page
  2. Introduction
  3. Calculation of posttest probabilities for the diagnosis of ReA
  4. Bacteria triggering ReA
  5. HLA-B27 and spondylarthropathy-specific features in the diagnosis of ReA
  6. Combination of tests for the identification of the triggering bacterium
  7. Posttest probability of ReA in the case of negative test results
  8. Future possibilities to achieve a higher posttest probability
  9. A need for diagnostic criteria in ReA
  10. Summary and conclusion
  11. REFERENCES

The posttest probability of having the disease following a positive test result strongly depends on the pretest probability of the disease being evaluated (expected prevalence of the disease in a certain population before applying a test) as well as on the sensitivity (proportion of patients with the disease with a positive test result) and the specificity (proportion of patients without the disease with a negative test result) of the test used. Many of the data on the pretest probability and specificity and sensitivity of the tests applied and clinical symptoms evaluated are available for patients with ReA, and it is the purpose of this article to present calculations for the probability of a diagnosis of ReA using single tests or combinations of tests. Since the expected prevalence of ReA is strongly dependent on a preselection of patients, different calculations have been made after taking into account this variation in the expected prevalence of ReA.

For the calculation of the posttest probability of having the disease, we applied the following formula of Bayes' theorem (see refs. 16 and 17 and www.medizin.fu-berlin.de/rheuma/de/med_pre.htm). The probability of the presence of the disease in the case of a positiv test result is

  • equation image

and the probability of the presence of the disease in the case of a negative test result is

  • equation image

where ppd is the pretest probability of disease.

For most tests, their application is generally most useful if the pretest probability is relatively high, ideally between 40% and 60%. If the pretest probability is low, however, even a positive test result does not convincingly support the target diagnosis (16). The pretest probability can be increased by clinical preselection. Thus, in patients with a joint pattern that is noncharacteristic of ReA, such as arthralgias or a symmetric polyarthritis of the hands, the diagnosis of ReA is highly unlikely and the pretest probability is therefore low (assumed to be ≤1%) and any further test is useless. In contrast, if the clinical picture is compatible with or even strongly suggestive of ReA, the pretest probability is considerably increased. The pretest likelihood of ReA can be further increased by an active exclusion of other arthritides, such as Lyme arthritis, gout, early forms of rheumatoid arthritis, Löfgren syndrome, osteoarthritis, mechanical causes of joint inflammation, and other arthritic conditions; in up to 50% of patients with a clinical picture compatible with ReA, one of these diagnoses can be made (3). Importantly, a diagnosis of ReA irrespective of the triggering bacterium could be made in up to ∼40% of the remaining patients (3, 18, 19). Thus, for a patient with a characteristic joint pattern and after exclusion of other defined diagnoses, a pretest probability of 40% for the diagnosis of ReA can be assumed, in contrast to a probability of <10% for any early arthritis.

Bacteria triggering ReA

  1. Top of page
  2. Introduction
  3. Calculation of posttest probabilities for the diagnosis of ReA
  4. Bacteria triggering ReA
  5. HLA-B27 and spondylarthropathy-specific features in the diagnosis of ReA
  6. Combination of tests for the identification of the triggering bacterium
  7. Posttest probability of ReA in the case of negative test results
  8. Future possibilities to achieve a higher posttest probability
  9. A need for diagnostic criteria in ReA
  10. Summary and conclusion
  11. REFERENCES

Chlamydia trachomatis, Yersinia, and Salmonella are the most relevant pathogens triggering ReA in developed countries (3, 18, 19), and a preponderance of information is available for these bacteria. Therefore, only these pathogens will be discussed, but similar conclusions can be drawn for other bacteria if the respective data are available. The prevalence of infections with these bacteria certainly differs between countries and even between areas in the same country and might also change over time (20). Therefore, the examples given below have to be adjusted to the local conditions. For each of the 3 clinical scenarios with different pretest probabilities of ReA (arthritis suggestive of ReA plus a preceding symptomatic infection, arthritis suggestive of ReA without a preceding symptomatic infection, and any arthritis), we calculated posttest probabilities that would be obtained after the use of various tests that are commonly applied to detect Chlamydia or Yersinia/Salmonella infection.

Chlamydia trachomatis. A preceding symptomatic infection (sexually acquired) of the urogenital tract (urethritis/cervicitis) up to 4–6 weeks prior to the onset of arthritis increases the pretest probability for Chlamydia-induced arthritis. In such a case, it has been shown that Chlamydia trachomatis is the triggering pathogen for the arthritis in ∼50% of patients (18, 21, 22), resulting in a pretest probability for Chlamydia-induced arthritis of 50%. The pretest probability for Chlamydia-induced arthritis in patients presenting with a clinical picture compatible with ReA but without a symptomatic preceding infection has been estimated at ∼12% (18) and 22% (3) if the calculation is based on serologic findings or detection of Chlamydia in the urogenital tract, and at ∼30% (23–25) if based on polymerase chain reaction (PCR) methods to look for Chlamydia trachomatis in synovial fluid. Therefore, calculations are made for the different pretest probabilities of 12% and 30%. For comparisons, calculations are also given for a pretest probability of 1% that reflects a clinical picture not typical of ReA, such as arthralgias or a symmetric polyarthritis of the small joints.

The use of serology for the diagnosis of infections with Chlamydia trachomatis is hampered by a relatively high prevalence of positive antibody titers among controls and by a possible cross-reactivity with antibodies directed against Chlamydia pneumoniae. In addition, sensitivity seems to be limited, possibly because Chlamydia are obligatory intracellular pathogens that are fought predominantly by a cellular response and less, or not at all, by a humoral immune response. Antibodies of the IgG subclass alone do not sufficiently reflect a recent infection, because they can be elevated for months after an infection. Therefore, determination of IgG antibodies should be combined with tests for IgM or IgA antibodies, the latter 2 indicating an acute or persistent infection. However, these antibodies can also be positive in a control population with a high infection rate, resulting in a reduced specificity of this test. Antibody titers should be elevated at least 2 standard deviations above that of a control population (15). As antigens, whole Chlamydia trachomatis (26) or Chlamydia-specific major outer membrane protein and/or lipopolysaccharide (LPS) (27) have been used in the past. The specificity and sensitivity of currently available tests have been estimated to be not higher than 78% and 73% (26, 27), respectively, which will be used for further calculations (Table 1).

Table 1. Probability of reactive arthritis (ReA) following various tests for Chlamydia, Yersinia, and Salmonella
Clinical picture, laboratory testsPretest probability, %Sensitivity, % true positiveSpecificity, % true negative100 − specificity, % false positive*Posttest probability for a positive test result, %Posttest probability for a negative test result, %
  • *

    Percentage of healthy persons with positive results.

  • Arthritis suggestive of ReA refers to arthritis that is oligoarticular, asymmetric, and involves predominantly the lower limb. Furthermore, other diagnoses have to be excluded.

  • Calculations of 2 different pretest probabilities are given; 12% is based on diagnostic studies not applying polymerase chain reaction techniques, while 30% is based on studies looking for chlamydial DNA in the joint by polymerase chain reaction.

Arthritis suggestive of ReA plus preceding symptomatic infection
 Serology for Chlamydia507378227726
 Detection of Chlamydia in the urogenital tract50509649334
 Serology for Yersinia/Salmonella30909010805
Arthritis suggestive of ReA without preceding symptomatic infection
 Serology for Chlamydia12/3073782231/595/13
 Detection of Chlamydia in the urogenital tract12/305096463/847/18
 Serology for Yersinia/Salmonella12909010552
Any arthritis/arthralgia
 Serology for Chlamydia17378223<1
 Detection of Chlamydia in the urogenital tract15096411<1
 Serology for Yersinia/Salmonella19090108<1

In the case of a preceding symptomatic urethritis/cervicitis in a patient with an arthritis compatible with ReA, a positive serologic finding will result in a posttest likelihood of 77% compared with a posttest probability of 31–59% in a patient without a prior symptomatic urethritis/cervicitis. Thus, without a history of a urogenital tract infection, serology alone does not suffice to make a definite diagnosis of Chlamydia-induced ReA. In contrast, in a patient in whom the clinical picture is noncharacteristic of ReA, the estimated pretest probability of 1% increases only marginally to a posttest likelihood of 3% if serology yields positive results. Thus, serology is obviously of no use in patients with noncharacteristic symptoms (Table 1).

Another test to be considered for the diagnosis of Chlamydia-induced arthritis is the detection of Chlamydia trachomatis in the urogenital tract. The search for Chlamydia in the first portion of the morning urine by PCR or by ligand chain reaction (LCR) seems to be an acceptable and relatively easy diagnostic approach, with a result comparable with that demonstrated by urogenital swab (28, 29). The calculations presented are based on a test specificity of 96%, which probably applies to most populations. However, specificity can decrease to ∼90% if a control group of sexually active young adults is chosen, as has been recently reported for young female recruits in the US Army (29). Depending on the criteria used, a diagnostic sensitivity of ∼50% has been reported for the detection of Chlamydiain the urogenital tract (18, 19, 22) in Chlamydia-induced arthritis. Calculations of posttest probabilities based on various pretest probabilities are shown in Table 1. Compared with the results based on serology, the posttest probability for the detection of Chlamydia trachomatis in the urogenital tract is always higher because of a higher specificity of this test.

Chlamydial DNA and even RNA have been detected in the joints of patients with Chlamydia-induced ReA, indicating that live Chlamydia persist inside the joint. Only small amounts of synovial fluid or membrane are necessary for such a test. In patients with undifferentiated arthritis but with a joint pattern compatible with ReA, chlamydial DNA has been detected by PCR in ∼30% of cases in different countries (23–25). However, chlamydial DNA was also found in control groups such as patients with rheumatoid arthritis (25) or healthy subjects (30). Thus, for the PCR test, a specificity of 96% (24) is used for the calculations. The sensitivity is also limited. No chlamydial DNA has been found, especially in some early cases of typical Chlamydia-induced arthritis (23, 25), possibly because of an effective cellular immune response (25). Therefore, a sensitivity of only 80% is used. In the case of a 12% pretest probability, a positive finding on PCR in synovial fluid results in a 73% posttest probability of Chlamydia-induced arthritis but can be as high as 90% if a pretest probability of 30% is assumed. It has to be stressed that there is no agreement at the moment on the optimum technique to detect Chlamydia by PCR (31, 32). Furthermore, none of the commercially available tests is sensitive enough to detect chlamydial DNA reliably in synovial fluid when compared with the amplification methods used in these studies, such as nested PCR (33).

Yersinia and Salmonella. The prevalence of infections with Salmonella, and especially with Yersinia, differs considerably among countries. Yersinia infections are common in Scandinavia and Germany (34) but are rare in the UK and the US. For calculations of the posttest probability of a Yersinia-induced arthritis, a pretest probability of 30% in the case of preceding diarrhea and a probability of 12% in the case of no preceding symptoms of a gut infection are used, based on results from Germany (18). Similar values are used for Salmonella-induced arthritides, based on studies from Germany and Norway (3, 18). No good data are available on the pretest probability of Yersinia- or Salmonella-induced arthritis in the US. However, the prevalence of all Salmonella infections (excluding typhi and paratyphi) seems to be equal or higher in the US compared with Germany, whereas Yersinia infections are assumed to be only about one-fifth the prevalence found in Germany (35). Provided that the occurrence rate of ReA after infections with these bacteria is similar in different countries (which is not necessarily the case), the pretest probabilities given in Table 1 might be similar in the US for Salmonella-induced arthritis but likely to be substantially lower for Yersinia-induced arthritis.

Followup studies in patients with Yersinia- or Salmonella-induced arthritis have shown that Yersinia-specific IgA antibodies (36) and Salmonella-specific IgA or IgM antibodies (37) can be detected for months or even years in the serum of ReA patients. Again, a test result should be regarded as positive only if the levels of IgA or IgM, in addition to IgG, are positive by at least 2 standard deviations above that in a control population. But, even when this criterion is applied, it has to be kept in mind that in ∼10% of the healthy population, the test result can be positive (34, 38). Yersinia-specific LPS (by enzyme immunoassay [EIA]) or Yersinia outer membrane protein (by Western blot) are used as the antigen in Yersinia serology (34). In acute Yersinia-induced ReA, IgG plus IgA antibodies can be detected in close to 100% of patients and IgA is still detectable after 1 year in 84% of the patients (36). An LPS-based EIA (IgG plus IgA or IgM) is also used in Salmonella-induced arthritides (39). In acute Salmonella infections, this test has a sensitivity of ∼92%, compared with 64% for the Widal (hemagglutination) test, which detects IgM. Several months after infection, the EIA still yields positive results in ∼92% of patients, whereas with detection of IgM by the Widal test, this drops to 15% (37, 39). Thus, it can be assumed that these antibody tests have a relatively high sensitivity of ∼90% even in ReA patients with a longer disease duration. Combining IgG and IgA or IgM, a specificity of ∼90% can be assumed for Yersinia- and Salmonella-specific serologic tests (34, 38). Calculations of posttest probabilities for Yersinia- and Salmonella-induced ReA based on a test sensitivity of 90% and a test specificity of 90% are shown in Table 1 for different pretest probability assumptions.

Stool cultures for Yersinia or Salmonella have a low sensitivity of ≤1% in patients with undifferentiated arthritis (18), but a high specificity of ∼98% can be assumed. In such a scenario, the posttest probability would be 6.4%. A sensitivity as high as 4% has also been reported for Salmonella-positive stool cultures. However, in this study, it was not differentiated between patients with and without diarrhea (3). A sensitivity of ∼10% for positive stool cultures can be found (18) if there is previous diarrhea, resulting in a posttest probability of 68%.

Yersinia, and probably also Salmonella, can persist in vivo as indicated by elevated IgA and/or IgM titers (36, 37) and by the detection of these pathogens in the peripheral blood for months and even years (40). However, DNA of these bacteria is only rarely detected inside the joint, and it has been suggested that, instead, mucosa and lymph nodes might serve as a reservoir for Yersinia and Salmonella (for discussion, see ref. 20). Therefore, a Salmonella- or Yersinia-specific PCR for joint material does not currently play a role in the diagnosis of ReA.

HLA-B27 and spondylarthropathy-specific features in the diagnosis of ReA

  1. Top of page
  2. Introduction
  3. Calculation of posttest probabilities for the diagnosis of ReA
  4. Bacteria triggering ReA
  5. HLA-B27 and spondylarthropathy-specific features in the diagnosis of ReA
  6. Combination of tests for the identification of the triggering bacterium
  7. Posttest probability of ReA in the case of negative test results
  8. Future possibilities to achieve a higher posttest probability
  9. A need for diagnostic criteria in ReA
  10. Summary and conclusion
  11. REFERENCES

An association between ReA and HLA-B27 has been described previously in a report >25 years ago, in the same year as that for ankylosing spondylitis (41). However, former studies were mainly hospital based and therefore included mostly the more severe cases. In more recent studies based on epidemics and/or on community investigations, the HLA-B27 association is not higher than 50% (for review, see refs. 1 and 20), and therefore a sensitivity of 50% for HLA-B27 testing is used. The HLA-B27 positivity in control patients with other arthritides should be slightly above the background prevalence in the population, because B27-positive patients with an SpA other than ReA can be assumed to be among the controls (5). The HLA-B27 prevalence depends very much on the population being studied (42). For the calculations in this study, an HLA-B27 positivity of 15% is used, which is slightly higher than that found in the control population in Western and Central Europe, resulting in a specificity of 85% (5). Assuming a pretest probability of 40% in a patient with a typical clinical pattern and after exclusion of other diagnoses (see above), a posttest probability of 69% for ReA is achieved if HLA-B27 is positive (Table 2).

Table 2. HLA-B27 testing alone or in combination with other tests and probability of reactive arthritis (ReA)
Clinical picture in combination with laboratory testsPretest probability, %*Sensitivity, % true positiveSpecificity, % true negative100 − specificity, % false positivePosttest probability for a positive test result, %Posttest probability for a negative test result, %
  • *

    Pretest probabilities are based on a situation in which there is no prior information on laboratory tests, or on a situation in which one further test (serology for Yersinia/Salmonella, detection of Chlamydia in the urine, or serology for Chlamydia) has already yielded positive results (as indicated in Table 1).

  • Percentage of healthy persons with positive results.

  • Arthritis suggestive of ReA refers to arthritis that is oligoarticular, asymmetric, and involves predominantly the lower limb. Furthermore, other diagnoses have to be excluded.

  • §

    For a pretest probability of 12%, before applying Chlamydia-specific tests.

  • For a pretest probability of 30%, before applying Chlamydia-specific tests.

Arthritis suggestive of ReA and HLA-B27 testing405085156928
Arthritis suggestive of ReA plus positive serologic result for Yersinia or Salmonella and HLA-B27 testing555085158042
Arthritis suggestive of ReA plus positive urine test result for Chlamydia and HLA-B27 testing63§5085158550
Arthritis suggestive of ReA plus positive serologic result for Chlamydia and HLA-B27 testing
 Pretest probability of 12%§315085156021
 Pretest probability of 30%595085158346

The combination of HLA-B27 testing with either a positive serologic test result or the detection of Chlamydia trachomatis in the urogenital tract can further increase the probability for the diagnosis of ReA, as shown in Table 2.

SpA-specific features such as enthesitis have been suggested to be used for the diagnosis of ReA (2, 9). Enthesitis was found in ≤30% (sensitivity) of ReA patients (3, 43) and in ∼20% of control patients with no SpA in the study by the European Spondylarthropathy Study Group (2), resulting in a specificity of 80%. Using these numbers, the increase from a pretest probability of 40% to a posttest probability of 50% is small. Thus, arthritis plus enthesitis alone is not helpful in making a diagnosis of ReA.

Combination of tests for the identification of the triggering bacterium

  1. Top of page
  2. Introduction
  3. Calculation of posttest probabilities for the diagnosis of ReA
  4. Bacteria triggering ReA
  5. HLA-B27 and spondylarthropathy-specific features in the diagnosis of ReA
  6. Combination of tests for the identification of the triggering bacterium
  7. Posttest probability of ReA in the case of negative test results
  8. Future possibilities to achieve a higher posttest probability
  9. A need for diagnostic criteria in ReA
  10. Summary and conclusion
  11. REFERENCES

As discussed above, considering all patients with ReA irrespective of the triggering bacterium, a pretest probability of ∼40% can be assumed in the case of clinical preselection. If currently available tests designed for the identification of a single bacterium and having a reasonably high specificity of ≥90% and an acceptable sensitivity (serologic findings of Yersinia and Salmonella, and detection of Chlamydia trachomatis in the urogenital tract but not chlamydial serology) are applied, and if one of them produces a clearly positive result, a posttest probability of between 80% and 84% for the diagnosis of any ReA independent of the triggering bacterium can be achieved, depending on the sensitivity of the test used. The posttest probability can be increased if HLA-B27 is also positive (Table 3).

Table 3. Probability of reactive arthritis (ReA) in the case of positive serologic findings for Yersinia or for Salmonella or in the case of detection of Chlamydia trachomatis in the urogenital tract and in relation to HLA-B27 testing
Clinical picturePretest probability, %Sensitivity, % true positiveSpecificity, % true negative100 - specificity, % false positive*Posttest probability for a positive test result, %Posttest probability for a negative test result, %
  • *

    Percentage of positive results in healthy persons.

  • Arthritis suggestive of ReA refers to arthritis that is oligoarticular, asymmetric, and involves predominantly the lower limb. Furthermore, other diagnoses have to be excluded.

  • For additional HLA-B27 testing, the same specificity and sensitivity are used as in Table 2.

Arthritis suggestive of ReA408090108413
Arthritis suggestive of ReA plus HLA-B27 positive695085159523

Posttest probability of ReA in the case of negative test results

  1. Top of page
  2. Introduction
  3. Calculation of posttest probabilities for the diagnosis of ReA
  4. Bacteria triggering ReA
  5. HLA-B27 and spondylarthropathy-specific features in the diagnosis of ReA
  6. Combination of tests for the identification of the triggering bacterium
  7. Posttest probability of ReA in the case of negative test results
  8. Future possibilities to achieve a higher posttest probability
  9. A need for diagnostic criteria in ReA
  10. Summary and conclusion
  11. REFERENCES

The primary goal in daily clinical practice is to make a positive diagnosis. However, for the clinician, it is also important to know whether a diagnosis such as ReA can be excluded or is unlikely if a test gives a negative result. Therefore, the calculations for the posttest probability of having ReA following a negative test result are also given in Tables 1–3.

Future possibilities to achieve a higher posttest probability

  1. Top of page
  2. Introduction
  3. Calculation of posttest probabilities for the diagnosis of ReA
  4. Bacteria triggering ReA
  5. HLA-B27 and spondylarthropathy-specific features in the diagnosis of ReA
  6. Combination of tests for the identification of the triggering bacterium
  7. Posttest probability of ReA in the case of negative test results
  8. Future possibilities to achieve a higher posttest probability
  9. A need for diagnostic criteria in ReA
  10. Summary and conclusion
  11. REFERENCES

For Yersinia- and Salmonella-induced arthritides, it is doubtful whether the posttest probability can be substantially increased, because the pretest probability among patients with undifferentiated arthritis will always be relatively low and the sensitivity and specificity of the tests used in these studies can hardly be increased. However, not all of the tests that have shown a good sensitivity and specificity in the studies are commercially available. This should be changed in the near future.

For Chlamydia-induced arthritis, there might be a greater potential for diagnostic improvement. In many parts of the Western world, the prevalence (pretest probability) of Chlamydia-induced ReA is higher than the prevalence of enterobacteria-induced ReA. The currently used serologic methods have a relatively low sensitivity and specificity, which might be improvable (44). In contrast to serology, the testing for Chlamydia in the urogenital tract has an acceptable specificity and should be used more frequently in the evaluation of patients with undifferentiated arthritis. Although there is no established and uniformly accepted PCR test for the detection of Chlamydia in the joint, various studies have reported a high rate of positive results in patients with undifferentiated arthritis, suggesting that this will be an important diagnostic tool for the future. Depending on the local microbiology laboratory used, the results of all possible tests relevant for the diagnosis of ReA should be available within a few days, thus helping the physician to make a correct diagnosis.

A need for diagnostic criteria in ReA

  1. Top of page
  2. Introduction
  3. Calculation of posttest probabilities for the diagnosis of ReA
  4. Bacteria triggering ReA
  5. HLA-B27 and spondylarthropathy-specific features in the diagnosis of ReA
  6. Combination of tests for the identification of the triggering bacterium
  7. Posttest probability of ReA in the case of negative test results
  8. Future possibilities to achieve a higher posttest probability
  9. A need for diagnostic criteria in ReA
  10. Summary and conclusion
  11. REFERENCES

A major drawback for the diagnostic evaluation of a patient with suspected ReA is the fact that at the moment, no gold standard for the diagnosis of ReA is available (20, 45). We are aware that the numbers for the pretest probability applied in this study are mainly based on experts' opinion derived from the interpretation of clinical symptoms and tests in the various studies. Nevertheless, this seems currently to be the best approach. For the future, clear criteria for ReA are urgently needed because it seems to be a relatively frequent disease among patients with arthritis, and better-defined patients are needed for future studies on the pathogenesis and treatment of ReA. Currently, antibiotics might be effective in Chlamydia-induced ReA (6, 7) and sulfasalazine can be used in chronic ReA (46). However, there is clearly a need to get better information about the effectiveness of antibiotic treatment (47) and about therapy with disease-modifying drugs. Such new treatment strategies can only be studied in well-defined groups of patients. Thus, the criteria discussed herein should be evaluated in a prospective study. This, in turn, should result in a recommendation of a widely accepted set of clinical criteria and laboratory tests to be applied for the diagnosis of ReA. Such an international study is currently planned by the authors. Until this is accomplished, the considerations presented in this report should be helpful to estimate the value of a positive test result if used for the diagnosis of ReA.

Summary and conclusion

  1. Top of page
  2. Introduction
  3. Calculation of posttest probabilities for the diagnosis of ReA
  4. Bacteria triggering ReA
  5. HLA-B27 and spondylarthropathy-specific features in the diagnosis of ReA
  6. Combination of tests for the identification of the triggering bacterium
  7. Posttest probability of ReA in the case of negative test results
  8. Future possibilities to achieve a higher posttest probability
  9. A need for diagnostic criteria in ReA
  10. Summary and conclusion
  11. REFERENCES

Some important conclusions and considerations can be made from the data presented. First, none of the tests is useful if the clinical picture is not suggestive of ReA and if other diagnoses have not been excluded; in this case, the low pretest probability necessarily results in a low posttest probability. This means that tests used for the identification of triggering bacteria are only useful if applied by doctors familiar with the clinical picture of ReA and with the differential diagnosis of rheumatic diseases. Second, provided that a patient has been carefully selected, none of the tests or the clinical symptoms alone gives a posttest probability strong enough to make a definite diagnosis of ReA. This is especially true for serology and for HLA-B27 testing, but also for the SpA-specific symptoms such as enthesitis, which are of little value if used alone. An exception to this might be a positive PCR result for the detection of Chlamydia in the joint. Third, in order to make the diagnosis of ReA, a combination of tests or clinical symptoms or both has to be applied, such as clinical evidence of a preceding infection plus a positive result on serology or identification of a bacterium by serology or culture/PCR/LCR plus positivity for HLA-B27 (Figures 1 and 2 and Table 3).

thumbnail image

Figure 1. Diagnostic approach to the diagnosis of Chlamydia-induced arthritis. The posttest probability in the case of a positive test result (+) is given in parentheses. For details of the calculations, see the text and Tables 1 and 2. * = For the posttest probability in the case of an assumed pretest probability of 30%, see Table 1. SF/SM = synovial fluid/synovial membrane; PCR = polymerase chain reaction.

Download figure to PowerPoint

thumbnail image

Figure 2. Diagnostic approach to the diagnosis of Yersinia- or Salmonella-induced arthritis. The posttest probability in the case of a positive test result (+) is given in parentheses. For details of the calculations, see the text and Tables 1 and 2.

Download figure to PowerPoint

REFERENCES

  1. Top of page
  2. Introduction
  3. Calculation of posttest probabilities for the diagnosis of ReA
  4. Bacteria triggering ReA
  5. HLA-B27 and spondylarthropathy-specific features in the diagnosis of ReA
  6. Combination of tests for the identification of the triggering bacterium
  7. Posttest probability of ReA in the case of negative test results
  8. Future possibilities to achieve a higher posttest probability
  9. A need for diagnostic criteria in ReA
  10. Summary and conclusion
  11. REFERENCES
  • 1
    Sieper J, Kingsley GH, Märker-Hermann E. Aetiological agents and immune mechanisms in enterogenic reactive arthritis. Baillieres Clin Rheumatol 1996; 10: 10521.
  • 2
    Dougados M, van der Linden S, Juhlin R, Huitfeldt B, Amor B, Calin A, et al. The European Spondylarthropathy Study Group preliminary criteria for the classification of spondylarthropathy. Arthritis Rheum 1991; 34: 121827.
  • 3
    Kvien TK, Glennås A, Melby K, Granfors K, Andrup O, Karstensen B, et al. Reactive arthritis: incidence, triggering agents and clinical presentation. J Rheumatol 1994; 21: 11522.
  • 4
    Hülsemann JL, Zeidler H. Undifferentiated arthritis in an early synovitis out-patient clinic. Clin Exp Rheumatol 1995; 13: 3743.
  • 5
    El-Gabalawy HS, Goldbach-Mansky R, Smith D II, Arayssi T, Bale S, Gulko P, et al. Association of HLA alleles and clinical features in patients with synovitis of recent onset. Arthritis Rheum 1999; 42: 1696705.
  • 6
    Lauhio A, Leirisalo-Repo M, Lähdevirta J, Saikku P, Repo H. Double-blind, placebo-controlled study of three-month treatment with lymecycline in reactive arthritis, with special reference to Chlamydia arthritis. Arthritis Rheum 1991; 34: 614.
  • 7
    Sieper J, Fendler C, Laitko S, Sörensen H, Gripenberg-Lerche C, Hiepe F, et al. No benefit of long-term ciprofloxacin treatment in patients with reactive arthritis and undifferentiated oligoarthritis: a three-month, multicenter, double-blind, randomized, placebo-controlled study. Arthritis Rheum 1999; 42: 138696.
  • 8
    Willkens RF, Arnett FC, Bitter T, Calin A, Fisher L, Ford DK, et al. Reiter's syndrome: evaluation of preliminary criteria for definite disease. Arthritis Rheum 1981; 24: 8449.
  • 9
    Amor B. Reiter's syndrome and reactive arthritis. Clin Rheumatol 1984; 2: 3159.
  • 10
    Calin A. Reiter's syndrome. In: CalinA, editor. Spondyloarthropathies. Orlando (FL): Grune & Stratton; 1984. p. 11950.
  • 11
    Khan MA. An overview of clinical spectrum and heterogeneity of spondyloarthropathies. Rheum Dis Clin North Am 1992; 18: 110.
  • 12
    Sieper J, Braun J, Brandt J, Miksits K, Heesemann J, Laitko S, et al. Pathogenetic role of Chlamydia, Yersinia and Borrelia in undifferentiated oligoarthritis. J Rheumatol 1992; 19: 123642.
  • 13
    Pacheco-Tena C, Burgos-Vargas R, Vazquez-Mellado J, Cazarin J, Perez-Diaz JA. A proposal for the classification of patients for clinical and experimental studies on reactive arthritis. J Rheumatol 1999; 26: 133846.
  • 14
    Sieper J, Braun J. Problems and advances in the diagnosis of reactive arthritis [editorial]. J Rheumatol 1999; 26: 12224.
  • 15
    Kingsley G, Sieper J. Third International Workshop on Reactive Arthritis. 23–26 September 1995, Berlin, Germany: report and abstracts. Ann Rheum Dis 1996; 55:56484.
  • 16
    Sacket DL, Haynes RB, Guyatt GH, Tugwell P, editors. Clinical epidemiology: a basic science for clinical medicine. Boston: Little, Brown & Company; 1991. p. 69152.
  • 17
    Khan MA, Khan MK. Diagnostic value of HLA-B27 testing in ankylosing spondylitis and Reiter's syndrome. Ann Intern Med 1982; 96: 706.
  • 18
    Fendler C, Laitko S, Sörensen H, Gripenberg-Lerche C, Groh A, Uksila J, et al. Frequency of triggering bacteria in patients with reactive arthritis and undifferentiated oligoarthritis and the relative importance of the tests used for diagnosis. Ann Rheum Dis 2001; 60: 33743.
  • 19
    Hannu T, Puolakkainen M, Leirisalo-Repo M. Chlamydia pneumoniae as a triggering infection in reactive arthritis. Rheumatology (Oxford) 1999; 38: 4114.
  • 20
    Sieper J, Braun J, Kingsley GH. Report on the Fourth International Workshop of Reactive Arthritis. Arthritis Rheum 2000; 43: 72034.
  • 21
    Keat AC, Thomas BJ, Taylor-Robinson D, Pegrum GD, Maini RN, Scott JT. Evidence of Chlamydia trachomatis infection in sexually acquired reactive arthritis. Ann Rheum Dis 1980; 39: 4317.
  • 22
    Rich E, Hook EW, Alarcón GS, Moreland LW. Reactive arthritis in patients attending an urban sexually transmitted diseases clinic. Arthritis Rheum 1996; 39: 11727.
  • 23
    Bas S, Griffais R, Kvien TK, Glennås A, Melby K, Vischer TL. Amplification of plasmid and chromosome Chlamydia DNA in synovial fluid of patients with reactive arthritis and undifferentiated seronegative oligoarthropathies. Arthritis Rheum 1995; 38: 100513.
  • 24
    Branigan PJ, Gerard HC, Hudson AP, Schumacher HR Jr, Pando J. Comparison of synovial tissue and synovial fluid as the source of nucleic acids for detection of Chlamydia trachomatis by polymerase chain reaction. Arthritis Rheum 1996; 39: 17406.
  • 25
    Wilkinson NZ, Kingsley GH, Sieper J, Braun J, Ward ME. Lack of correlation between the detection of Chlamydia trachomatis DNA in synovial fluid from patients with a range of rheumatic diseases and the presence of an antichlamydial immune response. Arthritis Rheum 1998; 41: 84554.
  • 26
    Bas S, Cunningham T, Kvien TK, Glennås A, Melby K, Vischer TL. The value of isotype determination of serum antibodies against Chlamydia for the diagnosis of Chlamydia reactive arthritis. Br J Rheumatol 1996; 35: 5427.
  • 27
    Bas S, Vischer TL. Chlamydia trachomatis antibody detection and diagnosis of reactive arthritis. Br J Rheumatol 1998; 37: 10549.
  • 28
    Rabenau H, Berger A, Doerr HW, Weber B. Testing for Chlamydia trachomatis in urine. Lancet 1997; 349: 10245.
  • 29
    Gaydos CA, Howell MR, Pare B, Clark KL, Ellis DA, Hendrix RM, et al. Chlamydia trachomatis infections in female military recruits. N Engl J Med 1998; 339: 73944.
  • 30
    Schumacher HR Jr, Arayssi T, Crane M, Lee J, Gerard H, Hudson AP, et al. Chlamydia trachomatis nucleic acids can be found in the synovium of some asymptomatic subjects. Arthritis Rheum 1999; 42: 12814.
  • 31
    Kuipers JG, Nietfeld L, Dreses-Werringloer U, Koehler K, Wollenhaupt J, Zeidler H, et al. Optimised sample preparation of synovial fluid for detection of Chlamydia trachomatis DNA by polymerase chain reaction. Ann Rheum Dis 1999; 58: 1038.
  • 32
    Freise J, Gerard HC, Bunke T, Whittum-Hudson JA, Zeidler H, Kohler L, et al. Optimised sample DNA preparation for detection of Chlamydia trachomatis in synovial tissue by polymerase chain reaction and ligase chain reaction. Ann Rheum Dis 2001; 60: 1405.
  • 33
    Bas S, Ninet B, Delaspre O, Vischer TL. Evaluation of commercially available tests for Chlamydia nucleic acid detection in synovial fluid of patients. Br J Rheumatol 1997; 36: 198202.
  • 34
    Mäki-Ikola O, Heesemann J, Toivanen A, Granfors K. High frequency of Yersinia antibodies in healthy populations in Finland and Germany. Rheumatol Int 1997; 16: 2279.
  • 35
    Center for Disease Control and Prevention. Summary of notifiable diseases, United States, 1998. MMWR Morb Mortal Wkly Rep 1999; 47: 292.
  • 36
    Granfors K, Viljanen M, Tiilikainen A, Toivanen A. Persistence of IgM, IgG, and IgA antibodies to Yersinia in yersinia arthritis. J Infect Dis 1980; 141: 4249.
  • 37
    Mäki-Ikola O, Leirisalo-Repo M, Kantele A, Toivanen P, Granfors K. Salmonella-specific antibodies in reactive arthritis. J Infect Dis 1991; 164: 11418.
  • 38
    Mäki-Ikola O, Arvilommi H, Gaston JSH, Granfors K. Salmonella antibodies in healthy populations in Finland and the UK: how many Salmonella-triggered reactive arthritides are there? Br J Rheumatol 1993; 32: 2623.
  • 39
    Isomäki O, Vuento R, Granfors K. Serological diagnosis of salmonella infections by enzyme immunoassay. Lancet 1989; 14.1: 14114.
  • 40
    Granfors K, Merilahti-Palo R, Luukkainen R, Möttönen T, Lahesmaa R, Probst P, et al. Persistence of Yersinia antigens in peripheral blood cells from patients with Yersinia enterocolitica O:3 infection with or without reactive arthritis. Arthritis Rheum 1998; 41: 85562.
  • 41
    Brewerton DA, Caffrey M, Nicholls A, Walters D, Oates JK, James DC. Reiter's disease and HL-A 27. Lancet 1973; 3: 9968.
  • 42
    Khan MA. A worldwide overview: the epidemiology of HLA-B27 and associated spondyloarthritides. In: CalinA, TaurogJD, editors. The Spondylarthritides. Oxford (UK): Oxford University Press; 1998. p. 1726.
  • 43
    Mattila L, Leirisalo-Repo M, Koskimies S, Granfors K, Siitonen A. Reactive arthritis following an outbreak of Salmonella infection in Finland. Br J Rheumatol 1994; 33: 113641.
  • 44
    Bas S, Muzzin P, Ninet B, Bornand JE, Scieux C, Vischer TL. Chlamydial serology: comparative diagnostic value of immunoblotting, microimmunofluorescence test, and immunoassays using different recombinant proteins as antigens. J Clin Microbiol 2001; 39: 136877.
  • 45
    Braun J, van der Heijde D, Kingsley G, Sieper J. On the difficulties of establishing a consensus of and diagnostic investigations for reactive arthritis: results and discussion of a questionnaire prepared for the 4th International Workshop on Reactive Arthritis. J Rheumatol 2000; 27: 218592.
  • 46
    Clegg DO, Reda DJ, Abdellatif M. Comparison of sulfasalazine and placebo for the treatment of axial and peripheral articular manifestations of the seronegative spondylarthropathies: a Department of Veterans Affairs cooperative study. Arthritis Rheum 1999; 42: 23259.
  • 47
    Sieper J, Braun J. Treatment of reactive arthritis with antibiotics. Br J Rheumatol 1998; 37: 71720.