834 Brief Reports cm 1996;23 (October)

agglutinating sera, Murex, Liverpool, England). The patient began receiving treatment with doxycycline, rifampin, and co-trimoxa­zole. His weakness gradually abated, and 4 weeks after treatment was started he was able to walk unaided.

In September 1994, nerve-conduction studies showed the reap­pearance of F-waves and of sensory-action potentials in the upper extremities.

Our patient had chronic polyradiculoneuropathy with CSF albu­minocytological dissociation, a finding typically observed in in­flammatory demyelinating polyneuropathies. The findings of nerve-conduction studies, however, did not fulfill the electrodiag­nostic criteria established for diagnosis of chronic inflammatory demyelinating polyneuropathy [3]. The nerve-conduction data, as well as the patient's occupation, led us to further investigate the possibly infectious etiology of the disease. The LAL test indicated the presence of gram-negative endotoxin in the CSF, and CSF culture then yielded colonies of Brucella.

The laboratory picture observed was quite different from the typical neurobrucellosis findings previously reported [1, 4]. Our patient repeatedly had negative serum and CSF serologies for Bru­cella. Although it has been shown that in cases of systemic brucel­losis agglutination tests may become negative during the course of the disease [5], this observation has been rarely reported in association with neurobrucellosis [6]. In the latter, CSF albumino­cytological dissociation is an unusual finding as well. In their report of a large series of 41 patients with neurobrucellosis, Pedro­Pons et al. [7] described six patients with CSF albuminocytological dissociation, but only one presented with meningoradiculitis.

Spontaneous Bacterial Empyema in a Patient with Hepatitis C Virus Cirrhosis and Sterile Ascitic Fluid

Spontaneous bacterial peritonitis is a complication that occurs in 10%-20% of hospitalized patients with cirrhosis [1]. It is thought that portal hypertension increases migration of enteric or­ganisms across the bowel wall and causes bacteremia with second­ary seeding of the ascitic fluid [1]. Approximately 10% of cirrhotic patients with ascites have an associated pleural effusion [2], which is presumed to be due to passage of ascitic fluid through defects in the diaphragm [3]. However, spontaneous bacterial infection of pleural fluid with, or especially without, infection of the ascitic fluid is rare.

A 59-year-old female nonsmoker was admitted to the hospital because of a 2-day history of fever and dyspnea. Her condition had been diagnosed as hepatitis C virus cirrhosis 3 years previously, and for 6 months before admission she had had both ascites and a right­sided pleural effusion that had never been infected. Two days before admission she had complained of steadily increasing dyspnea; at that time she did not have a cough, and her temperature was 39°C. On physical examination she appeared ill and dyspneic (30 breaths/min); her temperature was 38.3°C, and her pulse was 90/min. There were

Reprints or correspondence: Dr. Stephen D. H. Malnick, Department of Internal Medicine 'C', Kaplan Medical Center, 76100 Rehovot, Israel.

Clinical Infectious Diseases 1996;23:834-5 © 1996 by The University of Chicago. All rights reserved. 1058--4838/96/2304-0031 $02.00

We believe that neurobrucellosis must be ruled out when pa­tients present with chronic polyradiculoneuropathy, despite having normal serological and CSF agglutination titers of antibody to Brucella species and even CSF albuminocytological dissociation.

Rosario Luciano Oliveri, Giovanni Matera, Alfredo Foes, Mario Zappia, Umberto Aguglia, and Aldo Quattrone Institute of Neurology and Institute of Microbiology, Faculty of

Medicine, Catanzaro, Italy

References

I. AI Deeb SM, Yaqub BA, SharifHS, Phadke JG. Neurobrucellosis: clinical characteristics, diagnosis, and outcome. Neurology 1989;39:498-501.

2. Brandtzaeg P, Ovsteboo R, KierulfP. Compartmentalization oflipopolysac­charide production correlates with clinical presentation in meningococcal disease. J Infect Dis 1992; 166:650-2.

3. Ad Hoc Subcommittee of the American Academy of Neurology AIDS Task Force. Research criteria for diagnosis of chronic inflammatory demyelin­ating polyneuropathy (CIDP). Neurology 1991;41:617-8.

4. McLean DR, Russell N, Khan MY. Neurobrucellosis: clinical and therapeu­tic features. Clin Infect Dis 1992; 15:582-90.

5. Young EJ. Serologic diagnosis of human brucellosis: analysis of 214 cases by agglutination tests and review of the literature. Rev Infect Dis 1991; 13:359-72.

6. Dupont B, Bletry 0, Tucat G, Veyssier P, Philippon AM, Godeau P. Neuro­brucelloses chroniques avec serodiagnostic de Wright negatif: 2 observa­tions. Nouv Presse Med 1980;9:2721-4.

7. Pedro-Pons A, Foz M, Codina A, Rey C. Les neurobrucelloses (etUde de 41 cas) Cab Med 1972; 13:855-62.

signs of a large right pleural effusion, ascites, hepatosplenomegaly, and pitting edema of the ankles.

Laboratory investigations showed a hemoglobin level of 11 g/dL, a WBC count of 5,900Ip,L (normal differential count), and thrombocytopenia (platelets, 40,000Ip,L). The prothrombin time INR (international normalized ratio) was 2.1, the serum glucose level was 409 mg/dL, the urea level was 54 mg/dL, the sodium level was 129 mEqIL, and the albumin level was 2.1 g/dL. The results of all the other routine laboratory investigations were nor­mal, as were the levels of transarninases.

A chest roentgenogram confirmed the presence of a large right pleural effusion, and pleurocentesis revealed an empyema. Exam­ination of pleural fluid revealed a pH of 7.06, a WBC count of 13,OOOIJLL, a total protein level of 2.2 g/dL, a glucose level of 261 mg/dL, and a lactate dehydrogenase (LDH) level of 1,252 U/L (the level ofLDH in the blood was 354 U/L). Culture of this fluid yielded Escherichia coli. Simultaneous examination of the ascitic fluid re­vealed a polymorphonuclear leukocyte count of 120/ JLL and a total protein level of 0.5 g/dL; culture of this fluid was sterile.

A repeated chest roentgenogram did not reveal pneumonia. The patient was treated with ceftriaxone (1 g per day), and a chest drainage tube was inserted. The pleural fluid became sterile, and the patient made an uneventful recovery.

We describe herein a rare case of a cirrhotic patient with sponta­neous bacterial empyema (SBE) in association with sterile ascitic fluid. We believe that this patient had SBE and not parapneumonic empyema. First, there was no evidence of pneumonia on chest roentgenograms, including those taken after insertion of a chest drainage tube. Second, the pleural effusion was a transudate, not

Dow

nloaded from https://academ

ic.oup.com/cid/article/23/4/834/290311 by guest on 16 February 2022

CID 1996;23 (October) Brief Reports 835

an exudate. Third, culture of the pleural fluid yielded E. coli, which is also a common infectious agent in spontaneous bacterial peritonitis [4].

Although 10% of cirrhotic patients with ascites have an associ­ated pleural effusion [2], SBE is rare; it has been reported pre­viously in only 21 patients [5], most of whom presented concomi­tantly with spontaneous bacterial peritonitis. SBE occurred in association with sterile ascitic fluid in only three cases [5].

The pathogenesis of SBE is unclear. Infected ascitic fluid may penetrate into the pleural cavity through the diaphragm, which would explain the simultaneous infection of both ascitic fluid and pleural fluid. In our case it appears that bacteremia, which com­monly occurs in portal hypertension, caused seeding of the E. coli in the pleural fluid. However, it is unclear why bacterial seeding occurs in the pleural fluid but not in the ascitic fluid.

Since spontaneous bacterial peritonitis can cause clinical deteri­oration in cirrhotic patients who do not have signs of peritonitis, patients in whom clinical deterioration of unknown cause occurs should be evaluated (by pleurocentesis) for the presence of SBE even when examination of ascitic fluid does not reveal infection.

Prolonged Fecal Shedding of Escherichia coli 0157:H7 During an Outbreak at a Day Care Center

It has been reported that Escherichia coli 0157:H7 is cleared from the gastrointestinal tract rapidly after the onset of hemor­rhagic colitis [1]; recovery of the organism in stool is unlikely> 1 week after the onset of symptoms [2, 3]. However, the duration of excretion of E. coli 0157:H7 has been found to vary inversely with age [4]. Pai et al. [5] found that 53% of children <5 years of age had stool cultures positive for E. coli 0157:H7 3 weeks after the onset of symptoms. In a study of children <5 years of age in Minnesota child care centers, the median duration of shedding of E. coli 0157:H7 in a cohort of culture-positive children was esti­mated to be 17 days, with a range of 2-62 days; 38% of these children continued to shed the organisms for ~3 weeks [6]. The 16th edition of the Control of Communicable Diseases Manual reported a duration of excretion of up to 1 week in adults and up to 3 weeks in one-third of children [7].

An outbreak of hemorrhagic colitis due to E. coli 0157:H7 in a Colorado child care center in June 1995 provided the opportunity to study the duration of shedding of the organism. A case was defined as any child within the child care center who had a stool culture positive for E. coli 0157:H7 or who had diarrhea for ~2 days. Of the 141 children in the center, 24 met the case definition.

Grant support: This work was supported in part by a grant from the Division of Medicine ofthe Health Resources and Services Administration, U.S. Depart­ment of Health and Human Services (5 D33 AHI8002-06).

Reprints or correspondence: Dr. Richard Hoffman, 4300 Cherry Creek Drive South, Denver, Colorado 80222-1530.

Clinical Infections Diseases 1996;23:835-6 © 1996 by The University of Chicago. All rights reserved. 1058-4838/96/2304-0032$02.00

As this case demonstrates, the finding of sterile ascitic fluid does not exclude the presence of SBE in cirrhotic patients.

Stephen D. H. Malnick, Marina Somin, Oren Zimchoni, and Zev M. Sthoeger

Department of Internal Medicine 'C', Kaplan Medical Center, Rehovot, Israel

References

1. Gilbert JA, Kamath PS. Spontaneous bacterial peritonitis: an update. Mayo Clin Proc 1995;70:365-70.

2. Esteve M, Xiol X, Fernandez F, et al. Treatment and outcome of hydrothorax in liver cirrhosis. Journal of Clinical Nutrition and Gastroenterology 1986; 1:139-44.

3. Rubinstein D, McInnes IE, Dudley FJ. Hepatic hydrothorax in the absence of ascites: diagnosis and management. Gastroenterology 1985; 88: 188-91.

4. Runyon BA, Canawati HN, Akriviadis EA. Optimization of ascitic fluid culture technique. Gastroenterology 1988; 95: 1351.

5. Xiol X, Castellote J, Baliellas C, et al. Spontaneous bacterial empyema in cirrhotic patients: analysis of eleven cases. Hepatology 1990; 11: 365-70.

Twelve of the 24 cases had stool cultures positive for E. coli 0157:H7. The average time between the onset of symptoms and the first positive stool culture was 10.5 days for the 12 culture­positive children, whereas the average time between the onset of symptoms and the first negative stool culture was 22.5 days for the 12 culture-negative children.

Children with stool cultures positive for E. coli 0157:H7 were included with other cases at the center or excluded until they had two consecutive stool cultures negative for E. coli 0157:H7. Thus, the duration offecal shedding of E. coli 0157:H7 for symptomatic, culture-positive children could be estimated from the results of serial stool cultures. The duration of fecal shedding was defined as the interval from the onset of diarrhea to the first oftwo consecu­tive negative stool cultures. The median duration of shedding was 29 days, with a range of 11-57 days. The duration of shedding was ~3 weeks for 92% of the children. Three culture-positive cases were treated with antibiotics and had a mean (± SD) duration of excretion of 35.7 days (± 12.4) versus 30.1 days (± 13.0) for the culture-positive cases not treated with antibiotics.

The present study may overestimate the duration of excretion, since children who shed E. coli 0157:H7 for shorter periods were more likely to be culture-negative when they were first tested. We accounted for this possibility by arbitrarily assuming a conserva­tive shedding period of 7 days for the 12 culture-negative cases. The mean duration of shedding was then recalculated for all 24 cases. Even with this conservative estimate, the mean duration of shedding was 19.3 days.

In summary, the duration of excretion of E. coli 0157 :H7 among children in this outbreak was longer than has been previously reported. The prolonged shedding period had a significant impact on day care staff and parents because of the inconvenience of cohorting and/or excluding children (i.e., keeping them at home or in separate rooms) until stool cultures demonstrated that fecal shedding was no longer present. The results of this study are

Dow

nloaded from https://academ

ic.oup.com/cid/article/23/4/834/290311 by guest on 16 February 2022