Ann Thorac Surg 2005;79:2132-2134
© 2005 The Society of Thoracic Surgeons
Case report
Bacterial Pericarditis Due to Group F Streptococci as a Complication of Esophagomediastinal Fistula
Rajesh Kabra, MDa,
Karl F. Welke, MDb,
Kemp H. Kernstine, MD, PhDb,
F. Jeffrey Field, MDa,
Brad H. Thompson, MDc,
Hemender S. Vats, MDa,
Neal L. Weintraub, MDa,*
a Department of Internal Medicine, University of Iowa College of Medicine and VA Medical Center, Iowa City, Iowa, USA
b Department of Surgery, University of Iowa College of Medicine and VA Medical Center, Iowa City, Iowa, USA
c Department of Radiology, University of Iowa College of Medicine and VA Medical Center, Iowa City, Iowa
Accepted for publication December 10, 2003.
* Address reprint requests to Dr Weintraub, Division of Cardiovascular Diseases, Department of Internal Medicine, University of Iowa College of Medicine, 200 Hawkins Dr, E315-A1 GH, Iowa City, IA52242 (E-mail: neal-weintraub{at}uiowa.edu).
 |
Abstract
|
|---|
We report a case of group F streptococcal pericarditis, the source of which was found to be an esophagomediastinal fistula arising from a midesophageal diverticulum. The patient presented subacutely and had no preexisting symptoms of esophageal disease. Antibiotic therapy, surgical drainage, pericardiectomy, and esophageal myotomy led to a successful outcome.
|
Introduction
|
|---|
Bacterial pericarditis is a rare disease with a potentially lethal course requiring prompt treatment with pericardial drainage and antibiotics. It is equally important to identify and treat the source of infection. In most cases, the pericardium is secondarily infected by hematogenous spread, by direct spread from a contiguous focus of infection, or after thoracic surgery. However, in cases in which the primary source of infection cannot be identified, occult esophageal disease should be considered as a potential mechanism of entry of bacteria into the pericardium.
A 32-year-old woman previously healthy noted a gradual development of malaise, a low-grade fever, and upper respiratory symptoms. She was empirically treated as an outpatient with oral amoxicillin and later moxifloxacin. Ten days after the onset of symptoms, she developed pleuritic chest pain with dyspnea on exertion. On examination she was afebrile with stable vital signs. A three-component pericardial friction rub was detected. Her white blood cell count was elevated to 16,100/mL with a left shift. Initial chest roentgenogram showed an enlarged cardiac silhouette and a mild left-sided pleural effusion. The electrocardiogram was remarkable for diffuse ST segment elevation and PR segment depression consistent with acute pericarditis. Transthoracic echocardiogram showed a moderate pericardial effusion without evidence of tamponade. Antibiotics were discontinued, and the patient was started on indomethacin for treatment of acute pericarditis of an uncertain cause.
On day 3 of admission, the patient abruptly developed hypotension and tachycardia. Repeat transthoracic echocardiogram showed a large pericardial effusion with tamponade physiology. An emergent therapeutic pericardiocentesis was performed through a subxiphoid approach, and 650 mL of hemorrhagic fluid was drained, which resulted in immediate hemodynamic stabilization. A drainage catheter was left in place. The pericardial fluid analysis revealed 155,000/mL of white blood cells (primarily neutrophils) with gram-positive cocci in pairs and chains, and vancomycin and gentamicin were empirically started.
There was no obvious nidus for bacterial pericarditis. Computed tomographic scan of the chest showed a moderately large pericardial effusion with a pericardial catheter in place. There was extensive mediastinal lymphadenopathy and fibrosis with evidence of venous collaterals suggestive of chronic fibrosing mediastinitis (Fig 1A). Bilateral pleural effusions and fluid in the mediastinum were noted. The patient continued to exhibit tachycardia, fever, and leukocytosis. A repeat computed tomographic scan showed loculated pericardial and mediastinal fluids. Therefore, operative drainage of the pericardial sac was performed on day 5 of admission. Turbid fluid was drained from multiple loculations within the pericardial sac. A partial pericardiectomy was performed, and chest tubes were inserted into the pericardium and the left pleural space.
View larger version (91K):
[in this window]
[in a new window]
|
Fig 1. (A): Chest computed tomographic scan shows evidence of upper mediastinal lymphadenopathy (solid arrow) and fibrosis (dashed arrow). (B): Contrast esophagogram demonstrates esophagomediastinal fistula arising from the midesophageal diverticulum.
|
|
On day 7 of admission, group F streptococci grew from the pericardial fluid obtained during pericardiocentesis on day 3. The patient had no symptoms of oropharyngeal or esophageal disease. To search for a potential source of this organism, a contrast esophagram was done. It showed a midesophageal diverticulum with a fistulous tract into the mediastinum (Fig 1B). Endoscopy likewise demonstrated a pus-filled diverticulum in the middle esophagus (Fig 2).
View larger version (140K):
[in this window]
[in a new window]
|
Fig 2. Upper gastrointestinal endoscopy shows an opening in the midesophageal wall filled with pus (arrow).
|
|
The patient underwent a right thoracotomy that revealed a midesophageal diverticulum adherent to a calcified mediastinal lymph node. Its apex was perforated with communication into the mediastinum. The surrounding esophageal wall was inflamed, firm, and densely adherent to the surrounding tissues. Because of the suspected dysmotility of this area of the esophagus, a myotomy was performed extending 4 cm beyond the area of inflammation. The diverticulum was small, and the surrounding mucosa was healthy enough to allow suture placement. We chose to invaginate the diverticulum rather than perform a resection to reduce the likelihood for later leakage and failure without creating an obstructing intraluminal esophageal defect. The inversion technique took the tension off the repair of the mucosa hole. An intercostal muscle pedicle was placed to support the weakened inflamed esophageal wall. Tissue cultures were negative for yeast, fungi, and acid-fast bacilli. The patient was discharged home on postoperative day 15. She remained asymptomatic at her follow-up visit.
|
Comment
|
|---|
Several aspects of the present case underscore its uniqueness. First, the patient presented subacutely, probably because the infection had been partially treated with oral antibiotics. Hence, the diagnosis of bacterial pericarditis did not become evident until after antibiotics had been discontinued and the patient had abruptly deteriorated, which required emergent pericardiocentesis. Once the diagnosis of bacterial pericarditis was made, the source of infection was not readily apparent. The identification of an unusual pathogen led to the detection of an occult esophagomediastinal fistula as the source of infection.
Group F streptococci of the Streptococcus milleri group are normal resident flora of the oropharynx, gastrointestinal tract, and perineum. Streptococcus milleri is a rare cause of bacterial pericarditis with very few reported cases in the literature. One of these cases was attributed to transient bacteremia during vaginal delivery in a postpartum female, whereas in another case report esophageal-pericardial fistula secondary to esophageal carcinoma was found to be the source of this bacteria [1, 2]. In all other reported cases, the origin of infection or the mode of entry of the bacteria into the pericardium, or both, were not determined.
Remarkably the patient had no preexisting esophageal symptoms. We believe that the diverticulum resulted from both traction and pulsion forces on the esophageal wall. Although the wall consisted of mucosa only, without any muscularis component as would be expected in a pulsion diverticulum, its apex was adherent to a calcified lymph node, which suggested a traction diverticulum.The defect in the muscularis appears to have been caused by inflammation and local trauma, resulting in weakness and spasm. The pulsion defect then formed from intraluminal esophageal pressure on the weakened wall. As the inflammatory process progressed to a chronic, fibrosing stage, the diverticulum must have retracted into the inflamed lymphadenitis in the right paratracheal region and middle mediastinum. As a result, the wall became so inflamed, especially with the inspissated secretions that it eventually perforated into the mediastinum. The surgical repair was designed to address these components (ie, a myotomy to treat the pulsion-associated esophageal spasm and a vascular pedicle to support the weakened inflamed esophageal wall).
In endemic areas, histoplasmosis is the most likely cause of fibrosing mediastinitis. This was supported by the computed tomographic findings of fibrosing mediastinitis without other clinically apparent causes in our case. Pathogens are rarely cultured from the lymph nodes of fibrosing mediastinitis and require no antimicrobial treatment [3]. In a review of 95 patients seen at the Mayo Clinic with mediastinal granuloma, 10% had esophageal involvement in the form of compression, stricture, diverticulum, sinus tract formation, or tracheoesophageal fistula [4]. There are a few case reports of esophagomediastinal fistula in the literature secondary to tuberculous lymphadenitis, esophageal candidiasis, and Hodgkin’s disease [5–7]. However, none of the prior reported cases presented with acute bacterial pericarditis, as was observed in our case.
Our case highlights the importance of suspecting the diagnosis of bacterial pericarditis even in patients who present subacutely, particularly if they already had partial treatment with antibiotics. Identification of the offending organism(s) and investigations into the potential mechanisms of entry into the pericardium are critical to proper diagnosis and therapy.
|
Acknowledgments
|
|---|
Drs Kabra and Welke contributed equally to this work. Dr Weintraub is supported by grants (HL–070860 and HL–62948) from the National Institutes of Health.
|
References
|
|---|
- Snyder RW, Braun TI. Purulent pericarditis with tamponade in a postpartum patient due to group F streptococcus Chest 1999;115:1746-1748.[Abstract/Free Full Text]
- Kaufman J, Thongsuwan N, Stern E, Karmy-Jones R. Esophageal-pericardial fistula with purulent pericarditis secondary to esophageal carcinoma presenting with tamponade Ann Thorac Surg 2003;75:288-289.[Abstract/Free Full Text]
- Dines DE, Payne WS, Bernatz PE, et al. Mediastinal granuloma and fibrosing mediastinitis Chest 1979:320-324.
- Dukes RJ, Strimlan CV, Dines DE, et al. Esophageal involvement with mediastinal granuloma JAMA 1976;236:2313-2315.[Abstract/Free Full Text]
- Devarbhavi HC, Alvares JF, Radhikadevi M. Esophageal tuberculosis associated with esophagotracheal or esophagomediastinal fistulareport of 10 cases. Gastrointest Endosc 2003;57(4):588-592.[Medline]
- Kim BW, Cho SH, Rha SE, et al. Esophagomediastinal fistula and esophageal stricture as a complication of esophageal candidiasisa case report. Gastrointest Endosc 2000;52(6):772-775.[Medline]
- Papp JP, Penner JA. Esophagomediastinal fistula in Hodgkin’s disease Postgrad Med 1970;48(4):180-183.
Top
Abstract
Introduction
