1 I90 BRIEF REPORT REACTIVE ARTHRITIS ASSOCIATED WITH SHIGELLA SONNEI INFECTION ANNELI LAUHIO, JUHANI LAHDEVIRTA, RITA JANES, SIRKKA KONTIAINEN, and HEIKKI REP0 Several studies have failed to show an association between Shigella sonnei dysentery and reactive arthritis. We describe 3 patients who had reactive arthritis and a recent or concurrent S sonnei infection. To our knowledge, this is only the second study to suggest this association. We propose that S sonnei should be considered as a triggering agent for reactive arthritis. It is well established that reactive arthritis, sterile joint inflammation triggered by infection elsewhere in the body, can follow enteritis due to Shigella flexneri (1-3). In 1947, Young and McEven (4) described 14 patients with bacillary dysentery who developed Reiter’s syndrome, a form of reactive arthritis associated with urethritis and conjunctivitis. Stool cultures from 1 patient yielded Shigella sonnei, and in 2 others, antibodies against S sonnei were found. However, an association between S sonnei enteritis and reactive arthritis was not evident in recent studies of 3 outbreaks of S sonnei dysentery (2,5,6). These epidemiologic findings suggested that, unlike SfIexneri serotypes 1, 2, and 2a, S sonnei is not arthritogenic (23. We recently treated a patient with Reiter’s syndrome that developed after S sonnei dysentery. This prompted us to retrospectively investigate From the Departments of Medicine and Microbiology, Aurora Hospital, and the Second Department of Medicine, Helsinki University Central Hospital, Helsinki, Finland. Anneli Lauhio, MD; Juhani Lahdevirta, MD; Rita Janes, MD; Sirkka Kontiainen, MD; Heikki Repo, MD. Address reprint requests to Anneli Lauhio, MD, Department of Medicine, Aurora Hospital, Nordenskioldinkatu 20, Helsinki SF-00250, Finland. Submitted for publication December 31, 1987; accepted in revised form March 22, 1988. Arthritis and Rheumatism, Vol. 31, No. 9 (September 1988) whether any of the patients who had been treated in our hospital during the past 10 years because of S sonnei infection had developed clinical symptoms of reactive arthritis: 2 additional patients were found. CASE REPORTS Patient 1. The patient, a 31-year-old man who had previously been in good health, presented with bloody diarrhea and fever. A stool culture yielded S sonnei. Eleven days after the onset of diarrhea, the patient developed a urethral discharge. Two days later, he had fever and pain in the right knee, and he was admitted to Aurora Hospital. The patient had effusions in both knees and conjunctivitis in the left eye. The Westergren erythrocyte sedimentation rate (ESR) was 87 mdhour, the C-reactive protein (CRP) level was 21 0 mg/liter (normal <lo), and the white blood cell (WBC) count was 14.5 x 109/liter (normal 4.0-10.0 x 109/liter). Because septicemia was suspected, the patient was treated for 11 days with oral ciprofloxacin hydrochloride, 750 mg twice a day. Results of urinalyses performed before and after the ciprofloxacin therapy repeatedly showed a leukocyte count of >200 x lo6/ liter (normal < I 3 X 106/liter), although urine cultures for bacteria were negative, as were urethral cultures for Neisseria gonorrhoeae and isolations of Chlamydia trachomatis. Repeated stool cultures for salmonellae, yersiniae, and campylobacteriae, and results of examinations for parasites were negative. Blood cultures were performed twice and both times the findings were negative. Test results for antistreptolysin 0 (ASO) and anti-chlamydia antibodies in serum were repeatedly negative. Results of tests for rheumatoid BRIEF REPORTS factor (RF; by Rose-Waaler and latex fixation tests) and antinuclear antibodies (ANA; by immunofluorescence) were negative. Radiographic findings on examination of the chest were normal. The patient was HLA-B27 positive. The patient was febrile (38-39°C) for 3 weeks. He developed pain in his right shoulder and an ulceration in the cornea of his left eye. Two months after admission, the CRP value was 25 mglliter and the ESR was 55 mmhour, and the patient was discharged. Twelve months later he still had pain in his right shoulder. His ESR was 15 mm/hour and the CRP level was <lo mg/liter. Retrospective study method. The Department of Microbiology of Aurora Hospital receives stool samples from the Helsinki area, which has approximately one-half million inhabitants. Stool samples for culturing shigellae were transported to the Department of Microbiology in Stuart transport tubes (Difco, Detroit, MI). Procedures for the culture and identification of shigellae were performed as previously described (6). For biochemical identification, API20E profiles (API System, Montalieu-Vercieu, France) were used. Antisera used in serologic identification of the Shigella isolates were obtained from Wellcome (Dartford, UK). The antisera used allowed confirmation of biochemical identification, as well as serologic grouping for S jlexneri and Shigella hoydii, which may be identical on biochemical tests. Review of the laboratory files revealed that during the past 10 years, stool or urine cultures from 371 patients grew shigellae: 183 (49%) grew S sonnei, 162 (44%) grew SfEexneri, 19 (5%) grew S boydii, and 7 (2%) grew Shigella dysenteriae. Of the 183 subjects with positive cultures for S sonnei, 48 had been treated at Aurora hospital. Medical records revealed that 2 of the patients had had a clinical picture that denoted reactive arthritis. These 2 patients are described below. Patient 2. Patient 2, a 35-year-old man with a history of bronchial asthma, had diarrhea for 1 week during a visit to Ethiopia in 1986. Two weeks later, he developed a urethral discharge, and 5 days after this, pain in the right knee and fever up to 38.4"C. On admission, his right knee was warm and swollen. His ESR was 79 mdhour, the CRP value was 168 mghter, the peripheral blood WBC count was 12.0 X 109/liter,and the hemoglobin level was 13.9 gmfliter. Synovial fluid samples contained 39,000 leukocytes/ mm3 (84% polymorphonuclear granulocytes). Lactic acid concentration was 3.99 mmoles/liter, which suggested inflammation rather than infection. Results of 1191 Gram stain and bacterial cultures of synovial fluid were negative. Repeated stool cultures for shigellae, campylobacteriae, yersiniae, or Clostridium dificile, as well as examinations for parasites, were negative. Urethral cultures for N gonovrhoeae and isolations of C trachomatis were negative. No serologic evidence of streptococcal (streptolysin 0, deoxyribonuclease B), staphylococcal (staphylolysin, teichoic acid), Carnpylohacter, Salmonella, or Yersinia infection was obtained. Results of tests for ANA and RF were negative. Urine cultures for bacteria, including Salmonella, were negative. Results of radiologic examination of the knees and the sacroiliac joints were normal. The patient was HLA-B27 positive. The patient was treated with nonsteroidal antiinflammatory drugs and physiotherapy. He also took oral doxycycline, 100 mg once a day, for a 12-day period; urinalysis before and after therapy revealed 100-200 x 106/literleukocytes. The patient became afebrile over a 2-week period. He developed pain in the left ankle and the first metatarsophalangeal joint of the right foot. After 3 weeks, he left the hospital. Two weeks later, in tests performed at the outpatient department, urinalysis showed 100 X 106/liter leukocytes and, unexpectedly, urine culture yielded S sonnei. The patient was then treated with oral amoxicillin, 375 mg 3 times a day, for 10 days, after which the subsequent 4 urine cultures failed to grow shigellae, and the urinalysis results were repeatedly normal. Eight months later the patient had recovered completely. Patient 3. The patient, a 25-year-old, previously healthy man, developed diarrhea and fever in 1981 during ajourney to Mexico. Four weeks later he developed severe pains in his wrists, knees, ankles, and neck. Results of a stool culture were positive for S sonnei and negative for Salmonella. His ESR was 3 mm/hour, WBC count 5.6 X 109/liter,and hemoglobin level 14.3 gm/liter. A urethral culture for N gonorrhoeae was negative. Findings of urinalysis and radiographic examination of the chest were normal. The patient was treated with a 2-week course of oral trimethoprim/sulfamethoxazole and indomethacin. Stool cultures for S sonnei became negative. However, he continued to have pain in his knees, ankles, and wrists. Results of ESR, CRP, WBC, ASO, Rose-Waaler, and latex fixation tests were repeatedly within the normal range during a 12-month followup period. In 1987, we contacted this patient for reexamination. During the past 6 years he had had, particularly BRIEF REPORTS 1192 early in the morning, pain and stiffness in the lower back, and arthralgias following respiratory infections. There was tenderness in the left sacroiliac joint. His scores on a chest expansion test, Schober’s test, and fingertips-to-floor test were 8 cm, 3 cm, and 30 cm, respectively. The ESR, CRP level, WBC count, hemoglobin level, and results of urinalysis, Rose-Waaler test, latex fixation test, and tests for ANA were normal or negative. Findings of radiologic examination of the sacroiliac joint were normal. He was not HLA-B27 positive but had HLA-Bw22. DISCUSSION All 3 patients described in the present study had clinical pictures that denoted reactive arthritis. Two of the patients were HLA-B27 positive and the third had HLA-Bw22, an antigen also belonging to the crossreactive group of HLA antigens centered on B7 (the B7 CREG). There was no evidence of enteric or urogenital infections known to trigger reactive arthritis. However, each patient had a recent or concurrent S sonnei infection. Previous failure to demonstrate an association between S sonnei dysentery and the development of reactive arthritis in epidemiologic studies could have been attributed to a sampling error, as has been suggested earlier (7). Furthermore, the frequency of HLA-B27, and thereby the genetic susceptibility to reactive arthritis, varies. In Finland, the frequency is 14%; therefore, the number of subjects susceptible to reactive arthritis is high (8). The etiology of the triggering episode remains unclear in about 26% of patients with typical reactive arthritis (9). This may be due to eradication of the microbe before the development of reactive symptoms, intermittent microbial excretion, fastidious growth characteristics of the microbe, or lack of appropriate serologic testing. Serologic testing is unreliable in the detection of Shigella infections because cross-reactions between Shigella and other enteric bacteria are common (10). Furthermore, repeated cultures of stool and urine may be needed, as shown in the case of our patient 2. The mechanism by which arthritogenic strains cause reactive arthritis is not known, although it has been suggested that structural similarities between patient cells and SJEexnrri or other microbes trigger reactive arthritis (1 1). HLA-B27-like antigenic epitopes have been detected on SJlexneri serotype 2a, but not on S sonnei (12). The presence of these epitopes did not, however, correlate with the occurrence of reactive arthritis. Furthermore, Welsh et a1 (13) have reported evidence of cross-reacting antigenic determinants between HLA-B27 positive cells and S sonnei. In a recent study, 53% of patients with Reiter’s syndrome were found to have serum antibodies against a 6-amino acid peptide shared by HLA-B27 antigen and Klebsiella pneumoniae nitrogenase (14); it remains to be seen whether such peptide structures are also components of the shigellae. Unfortunately, the strains of S sonnei isolated from the 3 patients described in the present study are no longer available for further evaluation. Although we cannot exclude with certainty the possibility that the reactive arthritis was triggered by a microorganism that we did not recognize, the findings in this report and those in an earlier one (4) support the view that S sonnei infection should be taken into account as a possible triggering episode of reactive arthritis. REFERENCES 1 . Paronen I: Reiter’s disease: a study of 344 cases observed in Finland. Acta Med Scand 131 (suppl 212):l144, 1948 2. Simon D, Kaslow R, Calin A, Kaye R: Studies of Reiter’s syndrome following epidemic shigellosis (abstract). Arthritis Rheum 22:659-660, 1979 3. Good AE, Schultz JS: Reiter’s syndrome following Shigella flexneri 2a: a sequel to traveller’s diarrhea: report of a case with hepatitis. Arthritis Rheum 20: 100104, 1977 4. Young RH, McEven EG: Bacillary dysentery as the cause of Reiter’s syndrome. JAMA 134:1456-1459, 1947 5. Kaslow RA, Ryder RW, Calin A: Search for Reiter’s syndrome after an outbreak of Shigella sonnei dysentery. J Rheumatol 6:562-566, 1979 6. Kelly MT, Brenner DJ, Farmer JJ: Enterobacteriaceae, Manual of Clinical Microbiology. Fourth edition. Edited by EH Lennette, A Balows, WJ Hausler Jr, HJ Shadomy. Washington, DC, American Society for Microbiology, 1985 7. Lewis RB: The absence of reactive arthritis after Shigella sonnei infection (letter). Arthritis Rheum 25: 1267, 1982 8. Aho K , Leirisalo-Rep0 M, Rep0 H: Reactive arthritis. Clin Rheum Dis 11:25-39, 1985 9. Valtonen VV, Leirisalo M, Pentikainen PJ, Rasanen T, Seppala I, Larinkari U, Ranki M, Koskimies S, Malkamaki M, Makela PH: Triggering infections in reactive arthritis. Ann Rheum Dis 44:39905, 1985 10. Lemeland JF: Les germes intestinaux incriminks dans BRIEF REPORTS 1193 les arthrites rkactionnelles. Rev Rhum Ma1 Osteoartic 50:7 19-722, 1983 11. Yu DTY, Ogasawara M, Hill JL, Kono DH: Study of Reiter’s syndrome, with special emphasis on Yersinia enterocolitica. Imrnunol Rev 86:27-45, 1985 12. Van Bohemen CG, Nabbe AJJM, Grumet FC, Landheer JE, Dinant HJ, Zanen HC: Lack of correlation between HLA B27 like antigenic epitopes on Shigella flexneri and the occurrence of reactive arthritis. Clin Exp Immunol 65:679-682, 1986 13. Welsh J , Avakian H , Cowling P, Ebringer A, Wooley P, Panayi G , Ebringer R: Ankylosing spondylitis, HLAB27 and Klebsiella. I. Cross-reactivity studies with rabbit antisera. Br J Exp Pathol 61:85-91, 1980 14. Schwimmbeck PL, Yu DTY, Oldstone MBA: Autoantibodies to HLA B27 in the sera of HLA B27 patients with ankylosing spondylitis and Reiter’s syndrome: molecular mimicry with Klebsiella pneumoniae as potential mechanism of autoimmune disease. J Exp Med 166:173181, 1987 Fax Correspondence to and from A & R Editorial Office Now Available Authors and reviewers can now transmit reviews, revisions, and other correspondence to and from the Arthritis & Rheumatism editorial office in Birmingham via Fax. The telephone number is (205) 934-4761.