Case 3

Contributed by Timothy Williams, MD, Dhaval R. Parikh, MD, Jeremy R. Hopkin, MD and Brent Harris, MD, PhD

Depts. of Pathology (TW, BH), Medicine (DP), and Radiology (JH) Dartmouth Medical School

 
   

DIAGNOSIS AND DISCUSSION:

Dx: Cerebral air emboli and infarctions secondary to atrial-esophageal fistula

Air emboli may cause cerebral infarctions which, as with other types of emboli, typically involve many vascular territories. Early CT findings, including poor grey-white matter differentiation and mass effects from surrounding edema, resemble those from other causes of ischemia. Initial MRI findings, secondary to cytotoxic edema, are increased signal on fast fluid-attenuated inversion recovery (FLAIR) imaging, and, more specifically, restricted diffusion on diffusion weighted magnetic resonance imaging (DWI). Differentiation between air emboli and septic emboli requires correlation between the age of the infarct and the presence of air. While rapid gas production in abscesses containing Klebsiella organisms has been reported (6), the development of air in an abscess from septic emboli generally occurs in a characteristic step-wise fashion after the initial embolic event. Findings of early ischemia and pockets of air, as were present in our patient, are more consistent with air emboli.

The source of air emboli in this patient was evident at autopsy upon dissecting the area around the patient's prior esophagogastrectomy. The anastomotic site was densely fibrotic and adhered to adjacent structures, including the left atrium. Corresponding transmural defects in both the esophagus and left atrium proved to be ostia of a fistulous tract between these two areas. The fistula contained chronic inflammation and granulation tissue, vital reactions indicating that the defect was present ante mortem. Along the tract were carbon-like granules with associated macrophages, likely representing precipitated material from the esophagus. Diffuse pericardial adhesions, corresponding to an episode of presumed viral pericarditis two years prior to the current presentation, had obliterated the pericardial space and likely facilitated fistula formation by bringing the left atrial epicardium and the esophagogastric anastomotic site into contiguous apposition. It is conceivable that this episode of pericarditis was caused by the initial erosion of the fistula from the esophageal lumen into the pericardial sac. Highly conspicuous in this patient's terminal presentation was the onset of symptoms shortly after esophageal dilatation for stricture. Quite possibly this procedure eventuated in the final opening of the fistulous tract through the sclerotic tissue between the esophagus and left atrium. An esophageal-atrial fistula would account for both ongoing multiorganism sepsis and thromboembolic events, including air emboli, as micro-organisms, air, and food material from the esophagus gained intermittent access directly into the arterial circulation.

In a review (4) spanning the past 124 years, Luthi et al identified 50 cases of GI-cardiac fistulae specifically associated with esophagogastric neoplasms. Like our patient, thirteen (32%) of these patients had prior surgery, and eight (19%) had radiation therapy. Only one patient in this review had undergone recent esophageal dilatation; however, in other reports (1, 7), esophageal dilatation anteceded clinical presentation. Some authors (8, 9) also emphasize the role of prior pericarditis in the process of GI-cardiac fistula formation.

To a large extent, clinical manifestations of particular fistulae follow from the pressure gradients between the areas involved. For example, pericardial pressure is normally lower than upper GI tract pressures; thus fistulae between these areas usually present as pericarditis caused by GI material entering the pericardial sac. Likewise, GI-left ventricular fistulae commonly present with massive and often fatal hematemesis. Particularly challenging are cases of GI-left atrial fistulae (10), as the pressure gradient between these two areas is relatively small, allowing movement of blood from the atrium into the esophagus as well as esophageal contents into the atrium. The latter may occur intermittently during episodes of increased esophageal pressure, such as vomiting (11). Thus, fistulae of this sort may present with any combination of acute focal neurological symptoms, embolic phenomena including cerebral air emboli, multiorganism sepsis, intermittent hematemsis, or simply anemia with heme-positive stool from chronic low-level blood loss. All of these signs and symptoms were present in our patient during his terminal course.

The presence of air within the parenchyma of the brain requires careful investigation to determine how air could enter the arterial system. In a patient with a history of esophageal cancer or surgery—particularly with a subsequent history of pericarditis or recent esophageal dilatation—and any combination of unexplained embolic events including air emboli, acute focal neurological disturbance, multi-organism sepsis, and hematemesis or anemia, a GI-cardiac fistula, while unlikely, should be considered.


References

1. Snyder RW, Dumas PR, Kolts BE. Esophagoatrial fistula with previous pericarditis complicating esophageal ulceration. report of two cases and a review of the literature. Chest 1990;98:679-681.
2. Gonzalez-Ojeda A, de Luna Vargas MA, Nazara-Cazorla Z, Anaya-Prado R, Radillo LG. Esophagopericardial fistula attributed to a barogenic rupture of the esophagus. report of a case and review of the literature. Hepatogastroenterology 2001;48:1375-1378.
3. Miller WL, Osborn MJ, Sinak LJ, Westbrook BM. Pyopneumopericardium attributed to an esophagopericardial fistula: Report of a survivor and review of the literature. Mayo Clin Proc 1991;66:1041-1045.
4. Luthi F, Groebli Y, Newton A, Kaeser P. Cardiac and pericardial fistulae associated with esophageal or gastric neoplasms: A literature review. Int Surg 2003;88:188-193.
5. Lambert DR, Llaneza PP, Gaglani RD, Lach RD, Beaver WL. Esophageal-atrial fistula. J Clin Gastroenterol 1987;9:345-349.
6. Liliang PC, Hung KS, Cheng CH, Chen HJ, Ohta I, Lui CC. Rapid gas-forming brain abscess due to klebsiella pneumoniae. case illustration. J Neurosurg 1999;91:1060.
7. Prolla JC, Taebel DW, Kirsner JB. Perforation of an esophagogastric anastomotic ulcer into the left atrium. case report and review of the literature. Gastroenterology 1967;52:871-874.
8. Mott LJ, Austin GE. Cerebral embolization resulting from esophageal-atrial fistula. Arch Intern Med 1976;136:718-720.
9. Murphy GF, Raymond AK, Scannell JG. Esophageal-atrial perforation due to recurrent esophagitis 18 years after esophageal bypass surgery. J Thorac Cardiovasc Surg 1979;78:181-184.
10. Reynolds P, Walker FO, Eades J, Smith JD, Lantz PE. Food embolus. J Neurol Sci 1997;149:185-190.
11. Klein WA, Parkman HP, Dempsey DT, Fisher RS. Sphincterlike thoracoabdominal high pressure zone after esophagogastrectomy. Gastroenterology 1993;105:1362-1369.



 

Return to Cases Home Page