- Email: [email protected]
Transdiaphragmatic Gastroventricular Fistula
Accepted Manuscript Trans-diaphragmatic gastro-ventricular fistula Paul A. Carroll, Aisling Kinsella, Gail Darling, Maral Ouzounian, Jonathan C. Yeung PII:
S0003-4975(18)31541-8
DOI:
https://doi.org/10.1016/j.athoracsur.2018.09.026
Reference:
ATS 32064
To appear in:
The Annals of Thoracic Surgery
Received Date: 5 September 2018 Accepted Date: 9 September 2018
Please cite this article as: Carroll PA, Kinsella A, Darling G, Ouzounian M, Yeung JC, Transdiaphragmatic gastro-ventricular fistula, The Annals of Thoracic Surgery (2018), doi: https:// doi.org/10.1016/j.athoracsur.2018.09.026. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Trans-diaphragmatic gastro-ventricular fistula
Paul A. Carroll1, Aisling Kinsella2, Gail Darling1, Maral Ouzounian2, Jonathan C. Yeung1
RI PT
1. Division of Thoracic Surgery, Toronto General Hospital, University of Toronto, Toronto, Canada
SC
2. Division of Cardiovascular Surgery, Toronto General Hospital, University of Toronto,
Corresponding author:
Jonathan Yeung, MD, PhD, FRCSC
200 Elizabeth St, 9N983
TE D
Toronto General Hospital,
M AN U
Toronto, Canada
Toronto, ON, M5G 2C4, Canada
AC C
EP
Email: [email protected]
ACCEPTED MANUSCRIPT Abstract Gastro-cardiac fistulae are an exceedingly rare complication following esophagectomy with gastric conduit pull up. Unsurprisingly, there is a significant ascribed mortality with the disease process often only identified post-mortem. We report a case of trans-diaphragmatic
RI PT
fistula between the intra-abdominal gastric conduit and the right ventricle and surgical
AC C
EP
TE D
M AN U
SC
management thereof.
ACCEPTED MANUSCRIPT Esophagectomy with gastric conduit is a common operation for esophageal cancer. While the early complications of this operation are well-documented, late complications are less well-known and require a high index of suspicion for diagnosis. We present a rare cause of
RI PT
GI bleeding following esophagectomy.
A 54-year-old gentleman with a medical history of well-controlled paranoid schizophrenia,
SC
systemic lupus erythematosus, and alcohol/illicit drug abuse underwent an uncomplicated transhiatal esophagectomy for a T1a N0 M0 adenocarcinoma arising from Barrett’s
M AN U
esophagus in May 2009. As standard, the conduit was fashioned with sequential firings of a linear stapler with 2-0 silk reinforcement of the individual staple line junctions.
This gentleman presented 9 years later to a city emergency department with melena and a
TE D
haemoglobin of 6.7 g/dL. Esophagogastroscopy at the time revealed a deep gastric ulcer with no active bleeding. Biopsies were taken due to concern for malignancy which was duly disproven. Further investigations revealed an Enterococcus faecalis bacteremia at this time
EP
with apparent vegetations on his mitral and aortic valves. Long-term parenteral antibiotic
AC C
treatment was commenced for endocarditis. He re-presented 3 months later with another major gastrointestinal bleed which precipitated resuscitation and endoscopy at which time the large ulcer was clearly re-identified at the level of the diaphragmatic indentation. He declined transfer to our thoracic surgery service and left the hospital against medical advice.
Ultimately, one week later, he presented to our emergency room with a hemoglobin of 3.0g/dL. He was alert, stable, and was resuscitated with 4 units of packed red cells. Based on his history, we suspected a gastro-cardiac fistula. Urgent CT angiogram was obtained but
ACCEPTED MANUSCRIPT was non-diagnostic.
Transthoracic echocardiogram was performed that demonstrated
persistent small vegetations associated with mild to moderate regurgitation of the aortic and mitral valves. There was no evidence of abscess formation or other mechanical complication of endocarditis and biventricular function was preserved. He was taken
RI PT
expeditiously to the operating room by our cardiac and thoracic surgical services.
SC
At time of surgery, gastroscopy demonstrated a large (5x5cm) ulcer crater at or below the diaphragmatic indent. The floor of the ulcer was clearly beating (Video 1). There was
M AN U
clotted blood within the stomach but no active bleeding was identified. A median sternotomy was performed to approach the heart/gastric conduit. The pericardium was opened to the right of the midline in order to create a flap. Cardiopulmonary bypass was instituted by ascending aortic/right atrial cannulation with two-stage cannula.
Dense
TE D
adhesions of the inferoposterior heart to the pericardium were sharply divided with subsequent identification of a large erosive defect in the diaphragm and free spillage of gastric content into the pericardium. The ulcer had eroded through the epicardium and
EP
partially through the myocardium (Figure 1). Given the location of the defect, we elected
AC C
not to arrest the heart. The gastric perforation was approached by opening the left dome of the diaphragm circumferentially around the ulcer to allow mobilisation of the stomach and defect below by sweeping it off the under surface of the diaphragm. The fibrous circumference of the ulcer was then resected. The stomach, now freely mobile, was closed in two layers using 2-0 polyglactin sutures and 0 silk. The diaphragmatic defect was closed using interrupted, plegetted, and mattressed 0 silk sutures. The partial thickness defect in the right ventricle was repaired with a large bovine pericardial patch secured with multiple interrupted 3-0 polypropylene sutures. Finally, the vascularized pericardial patch was
ACCEPTED MANUSCRIPT interposed between the diaphragmatic and pericardial repairs and secured in place (Figure 2).
Separation from cardio-pulmonary bypass was uneventful and he arrived in the
intensive care unit on minimal inotropic support. He was extubated on post-operative day 2 and transferred to the floor on day 3. The patient returned to oral diet on day 7 without
RI PT
issue and was subsequently discharged on day 11 once social issues had been addressed and a plan established for parenteral antibiotics. At 6 weeks post-surgery, patient was alive
SC
and continuing his parenteral antibiotics in the community.
M AN U
Comment
Gastro-cardiac fistulae are exceedingly rare, with an overall poor survival.1 Conduit fistulae have been described to both thoracic and abdominal viscera. Fistulae to the left atrium are reported most commonly, followed by the right ventricle in esophagectomy cases where the
TE D
conduit is placed in the native position.2 Other reports of fistulisation between the stomach and heart have usually been described in the displaced, usually substernal, stomach conduit position. Penetration of an ulcer through the diaphragm is exceedingly unusual, with a
EP
single case described post Roux-en-Y gastric bypass where the remnant stomach developed
AC C
a peptic ulcer which eroded into the left ventricle through the intervening diaphragm.3 This is to our knowledge the first description of a trans-diaphragmatic gastro-right ventricular fistula post esophagectomy.
This patient, similar to those published before, presented with gastrointestinal bleeding after a prolonged interval. The etiology of late fistulae are reported as peptic or malignant, whereas earlier occurrences are most often secondary to anastomotic staple line erosion or leaks.4 Given the trans-diaphragmatic nature of this fistula, it draws similarities to gastro-
ACCEPTED MANUSCRIPT cardiac fistulae that occur post-open fundoplication.5 Fundoplications that develop ulcers occur commonly on the proximal lesser curve with purported causes to be alkaline reflux exposure, ischemia, and mechanical irritation caused by movement against the right crus.6,7 Mechanical irritation may be a particular cause of this fistula given the location at the level
RI PT
of the impingement of the diaphragm.
SC
Radiologic investigation did not help guide us to the site of the fistula. We had high suspicion of a conduit to left atrial fistula as described in previous reports given that the
M AN U
ulcer was identified on the anterior ‘wall’ of the conduit at endoscopy.8 Though we did not ultimately need to arrest the heart to open a left-sided chamber, pre-operative echocardiographic investigations prepared us for the need for retrograde cardioplegia and provided important information about his valvular and cardiac function given his history of
TE D
medically treated endocarditis.
The diagnosis of gastro-cardiac fistula requires high clinical suspicion in someone who
EP
presents with major GI bleeding and a history of a previous esophagectomy and gastric
AC C
pullup akin to major GI bleeding in someone with a previous aortic graft. Despite multiple previous endoscopies and imaging, there was a delay in diagnosis and referral for surgical opinion. The positioning and appearance of the ulcer should similarly have triggered an earlier urgent referral to an esophageal surgeon for assessment given the complexity of these cases and decision making with respect to life and conduit preservation. While the finding of endocarditis was previously largely attributed to his illicit drug use history, E. faecalis as the offending organism should raise the concern for a gastrointestinal origin. Similarly, although rare, major bleeding with a prior esophagectomy should probably be
ACCEPTED MANUSCRIPT considered to be a gastro-vascular fistula until proven otherwise. Together, these clues point towards a cardiac fistula.
From a technical standpoint, the location of this ulcer was not readily apparent and
RI PT
identified only upon direct exploration of the mediastinum. The circumferential incision of the diaphragm around the fistula allowed entry into the abdominal cavity to mobilise the
SC
gastric conduit without compromising the gastroepiploic pedicle or the capacity of the conduit. Closure of the defect was assisted by endoscopy to ensure an adequate lumen.
M AN U
Interposing a pedicled pericardial flap onto the diaphragmatic suture line followed the principle of separating a fistula repair with vascularized tissue.
Gastro-cardiac fistulae are a rare entity but within the history and clinical findings, diagnosis
AC C
EP
TE D
is apparent. Prompt surgical management is the only realistic treatment option for survival.
ACCEPTED MANUSCRIPT Figure Legends Video 1: Endoscopic view of the gastric conduit ulcer Figure 1: Intraoperative view of the right ventricle defect (*), the diaphragm (D), and the gastric conduit (G).
RI PT
Figure 2: Intraoperative view of the pericardial patch repair of the right ventricle (P) and the
AC C
EP
TE D
M AN U
SC
vascularized pericardium overlying the repaired gastric conduit (R).
ACCEPTED MANUSCRIPT References
AC C
EP
TE D
M AN U
SC
RI PT
1. Rana ZA, Hosmane VR, Rana NR, Emery DL, Goldenberg EM, Gardner TJ. Gastro-right ventricular fistula: a deadly complication of a gastric pull-through. Ann Thorac Surg 2010;90:297-9. 2. West AB, Nolan N, O'Briain DS. Benign peptic ulcers penetrating pericardium and heart: clinicopathological features and factors favoring survival. Gastroenterology 1988;94:1478-87. 3. Rutkoski JD, Schrope BA, Lee BE. Survival Following Gastro-Left Ventricular Fistula in a Patient Post Roux-en-Y Gastric Bypass. Ann Thorac Surg 2017;103:e51-e3. 4. Brynjolfsson G, Kania R, Bekeris L. Gastroesophageal cardiac fistula due to perforation of an esophagogastric anastomotic ulcer into the left atrium. Hum Pathol 1980;11:677-9. 5. Hauters P, de Canniere L, Collard JM, Buysschaert M, Michel LA. Gastrodiaphragmatic fistula after transabdominal Nissen fundoplication. An unusual complication. J Clin Gastroenterol 1990;12:313-5. 6. Scobie BA. High gastric ulcer after Nissen fundoplication. Med J Aust 1979;1:409-10. 7. Agrez MV, Ferguson NW. Aorto-gastric fistula following Nissen fundoplication. Aust N Z J Surg 1990;60:823-5. 8. Panda N, Feins EN, Axtell A, Lui N, Melnitchouk SI, Donahue DM. A Novel and Successful Repair of a Left Atriogastric Fistula After Esophagectomy. Ann Thorac Surg 2017;104:e157-e9.
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
S0003-4975(18)31541-8
DOI:
https://doi.org/10.1016/j.athoracsur.2018.09.026
Reference:
ATS 32064
To appear in:
The Annals of Thoracic Surgery
Received Date: 5 September 2018 Accepted Date: 9 September 2018
Please cite this article as: Carroll PA, Kinsella A, Darling G, Ouzounian M, Yeung JC, Transdiaphragmatic gastro-ventricular fistula, The Annals of Thoracic Surgery (2018), doi: https:// doi.org/10.1016/j.athoracsur.2018.09.026. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Trans-diaphragmatic gastro-ventricular fistula
Paul A. Carroll1, Aisling Kinsella2, Gail Darling1, Maral Ouzounian2, Jonathan C. Yeung1
RI PT
1. Division of Thoracic Surgery, Toronto General Hospital, University of Toronto, Toronto, Canada
SC
2. Division of Cardiovascular Surgery, Toronto General Hospital, University of Toronto,
Corresponding author:
Jonathan Yeung, MD, PhD, FRCSC
200 Elizabeth St, 9N983
TE D
Toronto General Hospital,
M AN U
Toronto, Canada
Toronto, ON, M5G 2C4, Canada
AC C
EP
Email: [email protected]
ACCEPTED MANUSCRIPT Abstract Gastro-cardiac fistulae are an exceedingly rare complication following esophagectomy with gastric conduit pull up. Unsurprisingly, there is a significant ascribed mortality with the disease process often only identified post-mortem. We report a case of trans-diaphragmatic
RI PT
fistula between the intra-abdominal gastric conduit and the right ventricle and surgical
AC C
EP
TE D
M AN U
SC
management thereof.
ACCEPTED MANUSCRIPT Esophagectomy with gastric conduit is a common operation for esophageal cancer. While the early complications of this operation are well-documented, late complications are less well-known and require a high index of suspicion for diagnosis. We present a rare cause of
RI PT
GI bleeding following esophagectomy.
A 54-year-old gentleman with a medical history of well-controlled paranoid schizophrenia,
SC
systemic lupus erythematosus, and alcohol/illicit drug abuse underwent an uncomplicated transhiatal esophagectomy for a T1a N0 M0 adenocarcinoma arising from Barrett’s
M AN U
esophagus in May 2009. As standard, the conduit was fashioned with sequential firings of a linear stapler with 2-0 silk reinforcement of the individual staple line junctions.
This gentleman presented 9 years later to a city emergency department with melena and a
TE D
haemoglobin of 6.7 g/dL. Esophagogastroscopy at the time revealed a deep gastric ulcer with no active bleeding. Biopsies were taken due to concern for malignancy which was duly disproven. Further investigations revealed an Enterococcus faecalis bacteremia at this time
EP
with apparent vegetations on his mitral and aortic valves. Long-term parenteral antibiotic
AC C
treatment was commenced for endocarditis. He re-presented 3 months later with another major gastrointestinal bleed which precipitated resuscitation and endoscopy at which time the large ulcer was clearly re-identified at the level of the diaphragmatic indentation. He declined transfer to our thoracic surgery service and left the hospital against medical advice.
Ultimately, one week later, he presented to our emergency room with a hemoglobin of 3.0g/dL. He was alert, stable, and was resuscitated with 4 units of packed red cells. Based on his history, we suspected a gastro-cardiac fistula. Urgent CT angiogram was obtained but
ACCEPTED MANUSCRIPT was non-diagnostic.
Transthoracic echocardiogram was performed that demonstrated
persistent small vegetations associated with mild to moderate regurgitation of the aortic and mitral valves. There was no evidence of abscess formation or other mechanical complication of endocarditis and biventricular function was preserved. He was taken
RI PT
expeditiously to the operating room by our cardiac and thoracic surgical services.
SC
At time of surgery, gastroscopy demonstrated a large (5x5cm) ulcer crater at or below the diaphragmatic indent. The floor of the ulcer was clearly beating (Video 1). There was
M AN U
clotted blood within the stomach but no active bleeding was identified. A median sternotomy was performed to approach the heart/gastric conduit. The pericardium was opened to the right of the midline in order to create a flap. Cardiopulmonary bypass was instituted by ascending aortic/right atrial cannulation with two-stage cannula.
Dense
TE D
adhesions of the inferoposterior heart to the pericardium were sharply divided with subsequent identification of a large erosive defect in the diaphragm and free spillage of gastric content into the pericardium. The ulcer had eroded through the epicardium and
EP
partially through the myocardium (Figure 1). Given the location of the defect, we elected
AC C
not to arrest the heart. The gastric perforation was approached by opening the left dome of the diaphragm circumferentially around the ulcer to allow mobilisation of the stomach and defect below by sweeping it off the under surface of the diaphragm. The fibrous circumference of the ulcer was then resected. The stomach, now freely mobile, was closed in two layers using 2-0 polyglactin sutures and 0 silk. The diaphragmatic defect was closed using interrupted, plegetted, and mattressed 0 silk sutures. The partial thickness defect in the right ventricle was repaired with a large bovine pericardial patch secured with multiple interrupted 3-0 polypropylene sutures. Finally, the vascularized pericardial patch was
ACCEPTED MANUSCRIPT interposed between the diaphragmatic and pericardial repairs and secured in place (Figure 2).
Separation from cardio-pulmonary bypass was uneventful and he arrived in the
intensive care unit on minimal inotropic support. He was extubated on post-operative day 2 and transferred to the floor on day 3. The patient returned to oral diet on day 7 without
RI PT
issue and was subsequently discharged on day 11 once social issues had been addressed and a plan established for parenteral antibiotics. At 6 weeks post-surgery, patient was alive
SC
and continuing his parenteral antibiotics in the community.
M AN U
Comment
Gastro-cardiac fistulae are exceedingly rare, with an overall poor survival.1 Conduit fistulae have been described to both thoracic and abdominal viscera. Fistulae to the left atrium are reported most commonly, followed by the right ventricle in esophagectomy cases where the
TE D
conduit is placed in the native position.2 Other reports of fistulisation between the stomach and heart have usually been described in the displaced, usually substernal, stomach conduit position. Penetration of an ulcer through the diaphragm is exceedingly unusual, with a
EP
single case described post Roux-en-Y gastric bypass where the remnant stomach developed
AC C
a peptic ulcer which eroded into the left ventricle through the intervening diaphragm.3 This is to our knowledge the first description of a trans-diaphragmatic gastro-right ventricular fistula post esophagectomy.
This patient, similar to those published before, presented with gastrointestinal bleeding after a prolonged interval. The etiology of late fistulae are reported as peptic or malignant, whereas earlier occurrences are most often secondary to anastomotic staple line erosion or leaks.4 Given the trans-diaphragmatic nature of this fistula, it draws similarities to gastro-
ACCEPTED MANUSCRIPT cardiac fistulae that occur post-open fundoplication.5 Fundoplications that develop ulcers occur commonly on the proximal lesser curve with purported causes to be alkaline reflux exposure, ischemia, and mechanical irritation caused by movement against the right crus.6,7 Mechanical irritation may be a particular cause of this fistula given the location at the level
RI PT
of the impingement of the diaphragm.
SC
Radiologic investigation did not help guide us to the site of the fistula. We had high suspicion of a conduit to left atrial fistula as described in previous reports given that the
M AN U
ulcer was identified on the anterior ‘wall’ of the conduit at endoscopy.8 Though we did not ultimately need to arrest the heart to open a left-sided chamber, pre-operative echocardiographic investigations prepared us for the need for retrograde cardioplegia and provided important information about his valvular and cardiac function given his history of
TE D
medically treated endocarditis.
The diagnosis of gastro-cardiac fistula requires high clinical suspicion in someone who
EP
presents with major GI bleeding and a history of a previous esophagectomy and gastric
AC C
pullup akin to major GI bleeding in someone with a previous aortic graft. Despite multiple previous endoscopies and imaging, there was a delay in diagnosis and referral for surgical opinion. The positioning and appearance of the ulcer should similarly have triggered an earlier urgent referral to an esophageal surgeon for assessment given the complexity of these cases and decision making with respect to life and conduit preservation. While the finding of endocarditis was previously largely attributed to his illicit drug use history, E. faecalis as the offending organism should raise the concern for a gastrointestinal origin. Similarly, although rare, major bleeding with a prior esophagectomy should probably be
ACCEPTED MANUSCRIPT considered to be a gastro-vascular fistula until proven otherwise. Together, these clues point towards a cardiac fistula.
From a technical standpoint, the location of this ulcer was not readily apparent and
RI PT
identified only upon direct exploration of the mediastinum. The circumferential incision of the diaphragm around the fistula allowed entry into the abdominal cavity to mobilise the
SC
gastric conduit without compromising the gastroepiploic pedicle or the capacity of the conduit. Closure of the defect was assisted by endoscopy to ensure an adequate lumen.
M AN U
Interposing a pedicled pericardial flap onto the diaphragmatic suture line followed the principle of separating a fistula repair with vascularized tissue.
Gastro-cardiac fistulae are a rare entity but within the history and clinical findings, diagnosis
AC C
EP
TE D
is apparent. Prompt surgical management is the only realistic treatment option for survival.
ACCEPTED MANUSCRIPT Figure Legends Video 1: Endoscopic view of the gastric conduit ulcer Figure 1: Intraoperative view of the right ventricle defect (*), the diaphragm (D), and the gastric conduit (G).
RI PT
Figure 2: Intraoperative view of the pericardial patch repair of the right ventricle (P) and the
AC C
EP
TE D
M AN U
SC
vascularized pericardium overlying the repaired gastric conduit (R).
ACCEPTED MANUSCRIPT References
AC C
EP
TE D
M AN U
SC
RI PT
1. Rana ZA, Hosmane VR, Rana NR, Emery DL, Goldenberg EM, Gardner TJ. Gastro-right ventricular fistula: a deadly complication of a gastric pull-through. Ann Thorac Surg 2010;90:297-9. 2. West AB, Nolan N, O'Briain DS. Benign peptic ulcers penetrating pericardium and heart: clinicopathological features and factors favoring survival. Gastroenterology 1988;94:1478-87. 3. Rutkoski JD, Schrope BA, Lee BE. Survival Following Gastro-Left Ventricular Fistula in a Patient Post Roux-en-Y Gastric Bypass. Ann Thorac Surg 2017;103:e51-e3. 4. Brynjolfsson G, Kania R, Bekeris L. Gastroesophageal cardiac fistula due to perforation of an esophagogastric anastomotic ulcer into the left atrium. Hum Pathol 1980;11:677-9. 5. Hauters P, de Canniere L, Collard JM, Buysschaert M, Michel LA. Gastrodiaphragmatic fistula after transabdominal Nissen fundoplication. An unusual complication. J Clin Gastroenterol 1990;12:313-5. 6. Scobie BA. High gastric ulcer after Nissen fundoplication. Med J Aust 1979;1:409-10. 7. Agrez MV, Ferguson NW. Aorto-gastric fistula following Nissen fundoplication. Aust N Z J Surg 1990;60:823-5. 8. Panda N, Feins EN, Axtell A, Lui N, Melnitchouk SI, Donahue DM. A Novel and Successful Repair of a Left Atriogastric Fistula After Esophagectomy. Ann Thorac Surg 2017;104:e157-e9.
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT