Abstract
Gastrointestinal tract fistulas are relatively uncommon and can be congenital or acquired. Acquired fistulas which communicate with the cardiovascular system are of particular concern, as they may be associated with significant morbidity and mortality. When found at autopsy, gastrointestinal fistulas should be thoroughly documented. Sudden and unexpected deaths with a history of gastrointestinal or cardiovascular surgery and a recent history of gastrointestinal bleeding or fever of an undetermined etiology should be approached with a high index of suspicion by medical examiners and signal a need for more careful dissection and documentation. This study reports on three deaths related to acquired, postsurgical fistula formation involving the gastrointestinal and cardiovascular systems.
INTRODUCTION
Gastrointestinal tract fistulas are relatively uncommon and can be congenital or acquired. Fistulas may communicate with a variety of anatomic locations, including body cavities, the skin, and other organ systems. Aortoenteric fistulas (AEF), having a direct connection between the aorta and the gastrointestinal tract, and gastrocardiac fistulas, characterized by a direct connection between the chambers of the heart and the stomach, are of particular concern since a delayed or missed diagnosis carries high morbidity and mortality. The most common sign is upper gastrointestinal bleeding (UGIB), which can present in several forms, including hematemesis, “coffee-ground” emesis, melena, and hematochezia, and range in severity from a minor “herald” bleed to an exsanguinating hemorrhage ((1–2)). Sudden and unexpected deaths involving gastrointestinal bleeding of undetermined etiology, especially when the decedent has a prior history of thoracic or abdominal surgery, should be considered with a high index of suspicion by medical examiners and be referred for autopsy. This study reports on three cases of acquired, postsurgical fistula formation involving the gastrointestinal and cardiovascular systems.
METHODS
The cases were selected from the files of one of the authors (JP). In each case, a full autopsy with histology and toxicology was performed.
CASE REPORTS
Case 1
A middle-aged male patient experienced sudden cardiac death shortly after having a permanent pacemaker placed for atrial fibrillation. He had a diaphragmatic hernia repair several weeks prior to death. Approximately two years earlier, he underwent surgical resection of his esophagus for cancer. He also had hypertensive and atherosclerotic cardiovascular disease and emphysema. At autopsy, the cause of death (COD) was determined to be upper gastrointestinal tract hemorrhage due to a gastro-cardiac (left atrial) fistula which developed as a complication following the esophagectomy for esophageal cancer (
Gross photo from autopsy in case 1, with probe extending through pulmonary vein (left), left atrium, and fistula, into opened stomach.
Fistula in case 1 viewed from the gastric side.
Fistula in case 1 viewed from left atrial side.
Opened fistula from case 1 (arrows indicate the edges which have been cut); LA = left atrium; G = gastric mucosa.
Case 2
An elderly female presented to the hospital with massive hematemesis with associated breathing difficulties and subsequent death. She had a history of severe chronic obstructive pulmonary disease (COPD), hypertensive and atherosclerotic cardiovascular disease (HASCVD), and a thoracic aortic aneurysm for which she had undergone aortic reconstruction with graft placement several years previously. Prior to dying, an upper endoscopy revealed a probable esophageal fistula thought to be communicating with the respiratory system. At autopsy, the esophageal fistula was found to be communicating with the aorta (
Opened esophagus at autopsy in case 2. Note the presence of blood clot within the fistula. Also note the wire mesh aortic graft within the opened aorta (right), adjacent to the opened esophagus.
After removing the adherent blood clot from the esophagus in case 2, a clear connection is evident between the esophagus (left) and aorta (right).
The fistula as viewed from the aortic side in case 2.
The fistula as viewed from the esophageal side in case 2.
A polarized microscopic image of a blood clot contained within the fistula in case 2. Note the presence of numerous birefringent particles, consistent with ingested material (hematoxylin & eosin; polarized light; original magnification 100×).
Case 3
A middle-aged woman was found unresponsive at home. All resuscitative efforts were unsuccessful. The case was referred for medicolegal autopsy. She had a past history of COPD, gastric bypass surgery, and relatively recent bacterial pericarditis, which required drainage and antibiotic therapy. Autopsy revealed the presence of a gastric fistula involving and traversing the left hemidiaphragm to involve the posterior wall of the left ventricle of the heart. Approximately 1000 ml of dark blood intermixed with gastrointestinal contents was present within the stomach and proximal small intestines, but no fresh-appearing hemorrhage was evident (
Gross image at autopsy in case 3, after initial incision, showing markedly distended loops of small intestine, filled with blood.
In situ view of the gastric ulceration in case 3, after removing much of the surrounding stomach.
External view of the explanted heart from case 3, showing the large ulceration. The probe is within the fistula.
Cross-section of the heart in case 3, showing the external ulceration (arrowhead) and the fistula (arrow) extending into the right ventricular cavity.
DISCUSSION
The rarity of gastrointestinal tract fistulas, especially when combined with a recent history of gastrointestinal bleeding of an undetermined etiology or a missing or incomplete patient history at the time of autopsy, can present several challenges to the medical examiner. Deaths related to fistulas involving the gastrointestinal and cardiovascular systems may fall under medicolegal jurisdiction either because of their unexplained and unexpected nature or when initial examination of the scene and circumstances seem suspicious. When determining medicolegal jurisdiction and the need for an autopsy in such cases, careful attention should be paid to the decedent's history, the individual's risk factors for that particular disease, and the possibility of a previous nonnatural, injurious event.
Sudden and unexpected deaths with a history of gastrointestinal or cardiovascular surgery and a recent history of gastrointestinal bleeding or fever of an undetermined etiology should be approached with a high index of suspicion by medical examiners and signal a need for more careful examination, dissection, and documentation. Postmortem chest radiographs and postmortem computed tomography (PMCT) have been shown to be valuable aids for both dissection planning and, in the case of PMCT, identifying the COD ((3–9)). While such radiologic techniques were not instrumental in the presented cases, the use of radiography in similar cases would likely be very valuable. It is well-known that careful dissection at autopsy helps preserve the spatial relationships between structures ((10)). This may be especially true in cases involving fistulas, where hemorrhage, scar tissue, inflammation, and possible necrosis may obscure anatomic structures. As evidenced by the three presented cases, extensive photography of the fistula from both sides with surrounding tissue improves clarity and allows for appropriate documentation of findings ((11–13)). Histologic examination can be very important in confirming gross autopsy findings ((14–16)). Histology samples from the fistula (and blood clot if present) can further support the findings with evidence of inflammation and debris, and to demonstrate the fistulas themselves. As demonstrated by the presented cases, in order for a cardiovascular-gastrointestinal fistula to be implicated as the COD, a reasonable mechanism of death must be present and associated with the fistula. In each of the cases presented, the mechanism of death involved exsanguination of blood from within the cardiovascular system into the gastrointestinal system.
Once discovered and documented, a cardiovascular-gastrointestinal fistula with associated extravasation of blood can usually be confidently ruled as the COD. Unfortunately, determining the most appropriate MOD in such deaths is not necessarily so easy. As shown by the three cases presented, a prior surgery is commonly involved in the formation of acquired fistulas connecting the cardiovascular and gastrointestinal systems. As such surgeries are typically related to underlying natural disease processes, and fistula formation is a known, albeit rare, postoperative complication, it is reasonable to rule the MOD in such cases as “natural” deaths, as occurred in cases 1 and 3. Depending on the pathologist and office policy, however, as well as the timing of the death in relation to the surgical procedure, some might consider these rare deaths to be unforeseen events and consequently rule the MOD as “accident.” Still others might consider such a case to be in the “grey zone” and rule the MOD as “undetermined,” as occurred in case 2. In certain jurisdictions, a MOD of “therapeutic complication” might be appropriate ((17)). In other cases, where a traumatic physical or caustic injury preceded and played a role in fistula formation, the MOD may be homicide ((18,19)), accident ((20–22)), or suicide ((23)).
CONCLUSION
Gastrointestinal fistulas communicating with the cardiovascular system are rare, but a predictable complication from abdominal and thoracic surgeries. These fistulas may form quickly or take years to develop, so in the absence of a complete medical history, the medical examiner must have a high index of suspicion when patient history or external examination shows evidence of previous surgeries and there is a recent history of gastrointestinal bleeding of undetermined source. These cases require more time and effort due to their complexity, but by employing postmortem radiology (as available), careful dissection, extensive photography, and histological sampling, the medical examiner can ultimately save time and produce better documentation of the process to ensure appropriate death certification.
Footnotes
DECLARATION OF CONFLICTING INTERESTS
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
ETHICAL APPROVAL
The institutional review board (WMU) does not require ethical approval for reporting individual cases or small case series involving autopsied decedents.
INFORMED CONSENT
Informed consent for information published in this article was not obtained because no identifying features are presented.
The authors received no financial support for the research, authorship, and/or publication of this article.
NOTE
This paper was presented as a poster at the 2017 Annual NAME Meeting and won the award for the John Pless Best Student Poster Presentation.
Authors
Allan J. Medwick, EdD, Western Michigan University Homer Stryker MD School of Medicine, Kalamazoo, MI; Ursidae Analytics, LLC, Clinton Township, MI
Joseph A. Prahlow, MD, Department of Pathology, St. Louis University School of Medicine, St. Louis, MO; Office of the Medical Examiner - City of St. Louis, St. Louis, MO