Esophageal Perforation Management Using a Multidisciplinary Minimally Invasive Treatment Algorithm Kfir Ben-David, MD, FACS, Kevin Behrns, MD, FACS, Steven Hochwald, MD, FACS, Georgios Rossidis, MD, Angel Caban, MD, Cristina Crippen, RN, Thomas Caranasos, MD, Steven Hughes, MD, FACS, Peter Draganov, MD, Christopher Forsmark, MD, Shailendra Chauhan, MD, Mihir S Wagh, MD, George Sarosi, MD, FACS The surgical management of esophageal perforation (EP) often results in mortality and significant morbidity. Recent less invasive approaches to EP management include endoscopic luminal stenting and minimally invasive surgical therapies. We wished to establish therapeutic efficacy of minimally invasive therapies in a consecutive series of patients. STUDY DESIGN: An IRB-approved retrospective review of all acute EPs between 2007 and 2013 at a single institution was performed. Patient demographic, clinical outcomes data, and hospital charges were collected. RESULTS: We reviewed 76 consecutive patients with acute EP presenting to our tertiary care center. Median age was 64 16 years (range 25 to 87 years), with 50 men and 26 women. Ninety percent of EPs were in the distal esophagus, with 67% of iatrogenic perforations occurring within 4 cm of the gastroesophageal junction. All patients were treated within 24 hours of initial presentation with a removable covered esophageal stent. Leak occlusion was confirmed within 48 hours of esophageal stent placement in 68 patients. Median lengths of ICU and hospital stay were 3 and 10 days, respectively (range 1 to 86 days). One-third of the patients were noted to have prolonged intubation (>7 days) and pneumonia that required a tracheostomy. One in-hospital (1.3%) mortality occurred within 30 days. Median total hospital charges for EP were $85,945. CONCLUSIONS: Endoscopically placed removable esophageal stents with minimally invasive repair of the perforation and feeding access is an effective treatment method for patients with EP. This multidisciplinary method enabled us to care for severely ill patients while minimizing morbidity and mortality and avoiding open esophageal surgery. (J Am Coll Surg 2014; 218:768e775. 2014 by the American College of Surgeons)
BACKGROUND:
either diagnostic or therapeutic endoscopic or surgical procedures, or spontaneous, such as Boerhaave’s syndrome. The majority of perforations are iatrogenic in origin and are located in the thoracic rather than the cervical or intra-abdominal esophagus.1 Despite aggressive therapy, the mortality rate of thoracic and abdominal perforations remains as high as 20% in recent published reports.2,3 Traditional management of thoracic and abdominal EP has been prompt surgery, with primary repair of the perforation the preferred approach in a patient without underlying esophageal pathology or prolonged inflammation.4 Despite data suggesting that early surgical repair of perforations is the preferred approach, up to 30% of repairs demonstrate a persistent leak and may require additional esophageal procedures.5 Although many authors
Esophageal perforation is a rare but potentially lethal condition that requires prompt recognition and management to achieve a favorable outcome. Esophageal perforations (EP) are typically classified as iatrogenic, associated with Disclosure Information: Dr Ben-David receives pay as a consultant and received a research grant from Ethicon; Dr Forsmark is paid as a consultant by Applied Clinical Intelligence; and Dr Hochwald is paid as a consultant for Ethicon Endosurgery. All other authors have nothing to declare. Presented at the Southern Surgical Association 125th Annual Meeting, Hot Springs, VA, December 2013. Received December 14, 2013; Accepted December 17, 2013. From the Departments of Surgery (Ben-David, Behrns, Rossidis, Caban, Crippen, Caranasos, Hughes, Sarosi) and Medicine (Dragonov, Forsmark, Chauhan, Wagh), University of Florida, Gainesville, FL; and Roswell Park Cancer Center, Buffalo, NY (Hochwald). Correspondence address: Kfir Ben-David, MD, FACS, Division of General Surgery, 1600 SW Archer Rd, Room 6164, P.O. Box 100109, Gainesville, FL 32610. email: [email protected]
ª 2014 by the American College of Surgeons Published by Elsevier Inc.
768
ISSN 1072-7515/14/$36.00 http://dx.doi.org/10.1016/j.jamcollsurg.2013.12.033
Vol. 218, No. 4, April 2014
Ben-David et al
advocate for open, operative management of esophageal perforations, others have demonstrated excellent outcomes with nonoperative treatment of perforations and percutaneous control of mediastinal sepsis. Vogel and colleagues6 reported no deaths and 96% esophageal healing in a group of 28 patients with thoracic perforations treated nonoperatively. With the development of selfexpanding covered stents used for treating both benign perforations and anastomotic leaks in the esophagus,7 we reasoned that stent occlusion of an EP could be combined with laparoscopic or thoracoscopic surgical drainage and/or repair with enteral nutritional access to effectively treat esophageal perforations. Esophageal stenting and nonoperative management was developed and implemented using a multidisciplinary minimally invasive protocol for the treatment of EP. We previously reported our pilot series of our first 11 patients.8 Herein, the results of all patients with EP treated under this protocol since 2007 are reported. The results support the effectiveness of a stent-first strategy in treating esophageal perforation at a high volume esophageal disease treatment center.
METHODS An IRB-approved retrospective review of acute EP between 2007 and 2013 was performed. Patient demographic, clinical outcomes data, and hospital charges were collected. We reviewed our series of 76 consecutive EP patients, who presented or were transferred to our tertiary care center between July 2007 and January 2013. All patients were managed according to our treatment algorithm for acute EP (Fig. 1). Diagnostic management included a water-soluble contrast study performed as an esophagram or CT of the chest using Gastrografin (Bracco Diagnostics Inc). Patients who were noted to have a contained leak within the mediastinum were treated with intravenous antibiotics and chest-tube drainage, if appropriate. Patients noted to have uncontained esophageal contrast extravasation were treated with an endoluminally placed covered esophageal stent and laparoscopic and/or video assisted thoracoscopic surgery with drainage of the posterior mediastinum. These patients also received laparoscopic enteral feeding access, usually in the form of a feeding jejunostomy. Patients who had an uncontained leak within the abdominal cavity were managed based on time elapsed from initial presentation to radiographic study. Patients who presented within 24 hours and had an uncontained abdominal leak were taken to the operating room for primary EP repair and a 360-degree loose fundoplication in order to cover the primary repair. If the esophageal perforation
Esophageal Perforation Treatment Algorithm
769
was greater than 2 cm, an esophageal stent (Alveolus Alimaxx, Merit Medical Endotek) was also used to aid with appropriate coverage and sealing of the esophageal perforation. The size of perforation was measured endoscopically at the time of initial stent placement. For patients who presented beyond 24 hours from the injury, the treatment protocol involved esophageal stenting, drainage, and distal enteral feeding access. In all cases after stent placement, occlusion of the esophageal perforation with stent was confirmed within 24 to 48 hours with either an esophagogram or CT of the chest. If leak occlusion was verified, nonintubated patients were started on a clear liquid diet and advanced to full liquid diet until their outpatient follow-up appointments. The patients were also discharged home on nocturnal cycled enteral feeds for nutritional support. The patients returned 4weeks post-stent placement for outpatient stent removal and follow-up esophagogram. All patients were assessed in clinic for symptoms of dysphagia or other esophageal disorders after their esophageal stent was removed. If the leak was not occluded with the initial stent placement, repeat endoscopy and repositioning of the stent were undertaken. Patients with successful stenting were evaluated on a daily basis with chest x-rays to detect evidence of stent migration. If stent migration occurred, or the clinical condition deteriorated, the patient was reevaluated with an esophagogram, and endoscopic repositioning of the esophageal stent was carried out. This management algorithm was repeated until leak occlusion was confirmed and maintained. If a leak occurred at the proximal end of the esophageal stent, an additional stent was deployed, creating a stentwithin-a-stent to ensure occlusion.
RESULTS We reviewed 76 consecutive patients with acute esophageal perforation treated at our tertiary-care center. Forty-nine patients were hospital-to-hospital transfers, and 27 patients presented primarily to our center. The demographics of our patient population are shown in Table 1. The median and mean ages were 64 and 61, respectively (range 25 to 87 years; SD 16 years). There were 11 patients who were 80 years or older and 34% were older than 70 years. There was a male predominance, with 50 men and 26 women within our group. Sixty-four patients were Caucasian, 9 were AfricanAmerican and 3 were Hispanic (Table 1). The majority of our EP (n ¼ 69) were located in the distal esophagus within approximately 4 cm of the gastroesophageal junction. The remaining esophageal perforations were in the mid-thoracic (n ¼ 5) and cervical esophagus (n ¼ 2).
770
Ben-David et al
Esophageal Perforation Treatment Algorithm
J Am Coll Surg
Figure 1. Esophageal perforation treatment algorithm.
The mean size of esophageal perforation was 3.4 cm, ranging from 1 cm to 11 cm (SD 1.67 cm). Sixty-eight (90%) of our esophageal perforation patients were noted to have uncontained extravasation of oral contrast into the mediastinum, pleural space, or Table 1.
Patient Demographics
Demographic
Age, y Mean Median Range Sex, n (%) Male Female Race, n (%) White Black Hispanic
Data
61 64 25e87 50 (66) 26 (34) 64 (84) 9 (11) 3 (4)
peritoneum. Ten of these patients (15%) had contrast extravasation into the mediastinum and pleural space, with contamination of the thoracic cavity. Eight of these were the result of esophageal instrumentation and 2 perforations were spontaneous (Boerhaave’s syndrome). Fifty-eight of 68 patients (85%) had extravasation into the distal posterior mediastinum with contamination into the abdominal peritoneal cavity. Fifty-one of these patients had an iatrogenic esophageal perforation due to esophageal instrumentation for various esophageal diseases. The remaining 7 were the result of spontaneous esophageal perforation after aggressive emesis. Ten percent (8 of 76) of patients presented with a contained esophageal perforation: 2 cervical, 4 thoracic, and 2 contained distal esophageal leaks (Table 2). Based on the patients’ clinical status, they were admitted to general care or the ICU. Twenty-seven patients who did not require ICU admission were able to start an oral liquid diet immediately after their swallow study post stent placement. Forty-nine patients required ICU admission, with
Vol. 218, No. 4, April 2014
Table 2.
Ben-David et al
Esophageal Perforation by Location and Etiology
Leak type and location
Contained Cervical Mid-thoracic Distal Contained Thoracic, mediastinal, and pleural Uncontained Thoracic, mediastinal, and pleural
Etiology
n
%
Iatrogenic Iatrogenic Iatrogenic Spontaneous
2 4 1 1
2.6 5.3 1.3 1.3
Iatrogenic Spontaneous
8 2
10.5 2.6
Iatrogenic Spontaneous
51 7
67.0 9.9
median and mean lengths of ICU stay of 3 and 10 days, respectively (range 1 to 46 days). The overall median and mean hospital stays were 10 and 18 days, respectively, with a range of 1 to 86 days. Twenty-five (33%) of the patients were noted to have prolonged intubation (>7 days) complicated by pneumonia and required a tracheostomy (Table 3). In this patient group the median and mean ICU stays were 25 and 30 days, respectively. These patients also had longer median and mean overall hospital stays of 34 and 38.5 days, respectively, and were all discharged to a rehabilitation facility. The remaining 51 patients were discharged to their residence. Twenty-six patients experienced stent migration requiring repositioning within the first week of stent Table 3.
Postoperative Data
Variable
Mortality, n (%) Overall length of stay, d Mean Median Range ICU admission, n (%) ICU length of stay, d Mean Median Range ICU patients with prolonged intubation (>7 days), n (%) Patients able to resume oral intake within 48 hours, n (%) Patients with stent migration requiring intervention, n (%) Mean stent interventions, n
Data
1 (1.3) 18 10 (1e86) 49 (64.5) 10 3 (1e46) 25 (51) 27 (35.5) 26 (34.2) 1.7
Esophageal Perforation Treatment Algorithm
771
placement. Nineteen of the 26 patients with stent migration had to have their stent repositioned 1 to 3 times, with mean and median of 1.7 and 2 stent adjustments, respectively. For patients who had a prolonged hospital course, their esophageal stent was exchanged every 4 to 5 weeks after initial placement. One in-hospital mortality occurred within 30 days as a result of hemorrhagic shock due to an atria-esophageal fistula related to catheterdirected atrial fibrillation ablation thermal injury. Median and mean total hospital charges for EP at our institution are $85,945 and $161,698, respectively. The median and mean lengths of time to final stent removal with confirmation of healing of the perforation were 28 and 36 days, respectively, demonstrating that this modality allows for rapid healing of the esophageal injury. None of the patients in our series required any additional esophageal surgical procedure for their EP other than what was initially performed per our treatment algorithm. The esophageal salvage rate in our series was 100%, with no conversion to an open esophageal repair or esophagectomy due to their EP.
DISCUSSION Although there are a multitude of etiologies, the most common cause of esophageal perforation is iatrogenic, related to esophageal instrumentation. This acute esophageal event is a dreaded life-threatening surgical condition diagnosed by clinical suspicion, diagnostic radiographic studies, or endoscopy. With historically high morbidity and mortality rates, the need for rapid detection and initiation of effective treatment is paramount.9-11 Multiple factors have been associated with successful outcomes of esophageal perforation management including the time elapsed between rupture and diagnosis, the size and location of the rupture, degree of contamination, and the overall health of the patient. In fact, time from perforation to diagnosis is considered the most crucial factor influencing the outcome of esophageal perforation.12,13 Despite the morbidity of the perforation, the optimal treatment remains uncertain because there is no single strategy that is sufficient to deal with the majority of these situations. Although many believe that open surgical repair of the perforation is the most effective treatment for esophageal perforations,9-11 our institution has reported that aggressive conservative therapy in patients with esophageal perforations results in lower morbidity and mortality while avoiding major open surgery and its associated morbidities.6,8 In addition, with the advancements in both endoscopic and minimally invasive techniques, esophageal perforation treatments have gradually shifted toward self-expanding endoluminal stents, endoscopic intraluminal clipping,
772
Ben-David et al
Esophageal Perforation Treatment Algorithm
and application of various sealant products.9,11,13-15 With our institution’s significant clinical experience in minimally invasive treatment of benign and malignant esophageal diseases, our approach to treatment of acute esophageal perforation has been to use an endoscopically placed, self-expanding esophageal stent to cover the perforation, occluding the leak.8 We strongly believe that the addition of an endoscopically placed removable covered stent results in rapid leak occlusion and sepsis control, resulting in decreasing contamination, morbidity, and hospital stay. This modality also enables treatment of esophageal perforations in patients with significant comorbidities, and/or elderly patients, who are often poor candidates for thoracotomy. Our observation that 26% of patients who presented with an acute esophageal perforation were 70 years old or older demonstrates the high risk population this condition afflicts. Early reports, including our own institution’s previous description of successful nonoperative treatment of esophageal perforations, have demonstrated success with this approach, especially if the perforation is confined to the mediastinum, a contained leak that easily drains back into the esophagus and is associated with minimal signs of clinical sepsis.16-18 This conservative treatment is centered on broad-spectrum antibiotics, wide drainage, and nothing by mouth. Most authors have advocated that nonoperative esophageal perforation treatment should be reserved for patients with micro-perforations and no evidence of ongoing leaks.19 However, perforations that did not meet these criteria have undergone surgical intervention to minimize mediastinal and/or peritoneal contamination. Despite well-documented results with aggressive nonoperative management, soilage of the mediastinum is of great concern.20 Therefore, this type of nonoperative treatment has not been adopted in all esophageal perforation patients because of the high mortality rate associated with contamination of the mediastinum and peritoneal cavity, resulting in significant clinical sepsis.21,22 Evidence suggests that mortality among patients treated within 24 hours of sustaining an esophageal injury is decreased by 10-fold compared with mortality in those for whom diagnosis and treatment are delayed.4 The success of early treatment strengthens our current notion that sealing the esophageal perforation is of critical importance to prevent thoracic or abdominal soilage. Therefore, all of our patients were treated with stenting within 24 hours of their initial presentation to our institution, and leak occlusion was confirmed within 48 hours of esophageal stent placement. Because of the initial success with intraluminal esophageal stent placement, this approach has become our first line treatment of choice when managing
J Am Coll Surg
patients with acute esophageal perforation. This has evolved into our current treatment algorithm (Fig. 1). Our protocol incorporates a combination of aggressive medical management with the wide use of intraluminal esophageal stent to control and seal esophageal leaks without a thoracotomy or a laparotomy and their associated surgical morbidities. In line with other studies,23 we have demonstrated that endoscopically placed esophageal stents and distal feeding access have resulted in lower rates of reoperation, open surgical intervention, esophagectomies, morbidity, and mortality than those that have been reported with standard surgical treatment.24,25 We had only 1 in-hospital mortality due to hemorrhagic shock from an iatrogenic atrial-esophageal fistula that resulted from catheter-based atrial ablation for treatment of atrial fibrillation. Additionally, none of the patients in our series required conversion to open esophageal surgical procedures after initial management with stenting. Although we had 26 patients who experienced stent migration requiring repositioning within the first week of stent placement, this was done at the bedside in the ICU with no adverse events. These patients had longer ICU stays, hospital courses, stent intervals, and higher hospital charges and costs. Unfortunately, we were not able to identify predictors that could prospectively identify the patients or esophageal perforations that result in the longer hospital course as a result of stent migration. To try to prevent this event, we have started to suture all of the stents with an absorbable suture, such as 2-0 chromic, laparoscopically to help anchor the stent in place at the time of initial placement and theoretically prevent migration. The use of esophageal stents allowed us to start early oral intake in 27 patients who were started on an oral liquid diet immediately after their post-stent placement swallow study. This treatment algorithm has also decreased our hospital stays and resulted in less overall morbidity for the patients, theoretically reducing medical cost and patient convalescence. In fact, our median lengths of ICU and hospital stay were 3 and 10 days, respectively, and only 33% of our patients were noted to have prolonged intubation (>7 days), pneumonia and required a tracheostomy. Esophageal stents need to be exchanged or removed because the complications associated with prolonged intubation of the esophagus are well known,20 but the exact time course for stent removal is still unknown. Most studies recommend a period of 10 days for small esophageal perforations and as long as 8 weeks for extended esophageal wall defects.20,26,27 We used fully covered metallic stents to allow ease of removal and to minimize tissue in-growth.
Vol. 218, No. 4, April 2014
Ben-David et al
It is important to emphasize that although this is a retrospective series of EP, it represents a treatment algorithm for the most common forms of esophageal perforations, iatrogenic and spontaneous, in all of the patients who presented to our institution from 2007 to 2013. Our protocol allowed us to be consistent in providing the same care to various types of esophageal perforations and clinical presentation at our institution regardless of physicians involved. We were also able to use the expertise of multiple medical services, allowing for a multidisciplinary care approach for these complex patients. This multidisciplinary method enabled us to care for severely ill patients while minimizing the morbidity and mortality and avoiding major open esophageal surgery.
CONCLUSIONS We have been able to demonstrate that endoscopically placed removable esophageal stents with aggressive conservative therapy is an effective method of treatment for patients with esophageal perforation at our tertiary care institution. This treatment protocol enables us to care for severely ill patients, minimizing the morbidity and mortality associated with major open esophageal surgery. We strongly believe that this multidisciplinary approach can be easily implemented at other institutions caring for esophageal perforation patients. Author Contributions Study conception and design: Ben-David, Hochwald, Rossidis, Caban, Draganov, Forsmark, Chauhan, Wagh, Sarosi Acquisition of data: Ben-David, Rossidis, Crippen, Caranasos Analysis and interpretation of data: Ben-David, Hochwald, Rossidis, Caban, Crippen, Caranasos, Sarosi Drafting of manuscript: Ben-David, Behrns, Hochwald, Rossidis, Hughes, Sarosi Critical revision: Ben-David, Behrns, Hochwald, Sarosi REFERENCES 1. Hasimoto CN, Cataneo C, Eldib R, et al. Efficacy of surgical versus conservative treatment in esophageal perforation: a systematic review of case series studies. Acta cirurgica brasileira / Sociedade Brasileira para Desenvolvimento Pesquisa em Cirurgia 2013;28:266e271. 2. Bhatia P, Fortin D, Inculet RI, Malthaner RA. Current concepts in the management of esophageal perforations: a twentyseven year Canadian experience. Ann Thorac Surg 2011;92: 209e215. 3. Ryom P, Ravn JB, Penninga L, et al. Aetiology, treatment and mortality after oesophageal perforation in Denmark. Danish Med Bull 2011;58:A4267.
Esophageal Perforation Treatment Algorithm
773
4. Brinster CJ, Singhal S, Lee L, et al. Evolving options in the management of esophageal perforation. Ann Thorac Surg 2004;77:1475e1483. 5. Wright CD, Mathisen DJ, Wain JC, et al. Reinforced primary repair of thoracic esophageal perforation. Ann Thorac Surg 1995;60:245e248; discussion 248e249. 6. Vogel SB, Rout WR, Martin TD, Abbitt PL. Esophageal perforation in adults: aggressive, conservative treatment lowers morbidity and mortality. Ann Surg 2005;241:1016e1021; discussion 1021e1023. 7. van Boeckel PG, Sijbring A, Vleggaar FP, Siersema PD. Systematic review: temporary stent placement for benign rupture or anastomotic leak of the oesophagus. Alimentary Pharma Therapeu 2011;33:1292e1301. 8. Ben-David K, Lopes J, Hochwald S, et al. Minimally invasive treatment of esophageal perforation using a multidisciplinary treatment algorithm: a case series. Endoscopy 2011;43: 160e162. 9. Biancari F, D’Andrea V, Paone R, et al. Current treatment and outcome of esophageal perforations in adults: systematic review and meta-analysis of 75 studies. World J Surg 2013;37: 1051e1059. 10. Huber-Lang M, Henne-Bruns D, Schmitz B, Wuerl P. Esophageal perforation: principles of diagnosis and surgical management. Surg Today 2006;36:332e340. 11. Schweigert M, Beattie R, Solymosi N, et al. Endoscopic stent insertion versus primary operative management for spontaneous rupture of the esophagus (Boerhaave syndrome): an international study comparing the outcome. Am Surg 2013;79:634e640. 12. Martinek J, Kovacova S, Nosek V, et al. Successful endoscopic treatment (clipping) of esophageal perforation during balloon dilatation in a patient with achalasia. Endoscopy 2008;40(Suppl 2):E61e62. 13. Sung HY, Kim JI, Cheung DY, et al. Successful endoscopic hemoclipping of an esophageal perforation. Dis Esophagus: official journal of the International Society for Diseases of the Esophagus / ISDE 2007;20:449e452. 14. Gomez-Esquivel R, Raju GS. Endoscopic closure of acute esophageal perforations. Curr Gastroenterol Rep 2013;15:321. 15. Kowalski C, Kastuar S, Mehta V, Brolin RE. Endoscopic injection of fibrin sealant in repair of gastrojejunostomy leak after laparoscopic Roux-en-Y gastric bypass. Surg Obes Relat Dis: official journal of the American Society for Bariatric Surgery 2007;3:438e442. 16. Cameron JL, Kieffer RF, Hendrix TR, et al. Selective nonoperative management of contained intrathoracic esophageal disruptions. Ann Thorac Surg 1979;27:404e408. 17. Lyons WS, Seremetis MG, deGuzman VC, Peabody JW Jr. Ruptures and perforations of the esophagus: the case for conservative supportive management. Ann Thorac Surg 1978;25: 346e350. 18. Shirkey HC. Conservative management of esophageal perforation. JAMA 1965;193:537. 19. Kiernan PD, Sheridan MJ, Elster E, et al. Thoracic esophageal perforations. Southern Med J 2003;96:158e163. 20. El Hajj II, Imperiale TF, Rex DK, et al. Treatment of esophageal leaks, fistulae, and perforations with temporary stents: evaluation of efficacy, adverse events, and factors associated with successful outcomes. Gastrointest Endosc 2013; Oct 11. [Epub ahead of print]. 21. Okten I, Cangir AK, Ozdemir N, et al. Management of esophageal perforation. Surg Today 2001;31:36e39.
774
Ben-David et al
Discussion
22. Salo JA, Isolauri JO, Heikkila LJ, et al. Management of delayed esophageal perforation with mediastinal sepsis. Esophagectomy or primary repair? J Thorac Cardiovasc Surg 1993; 106:1088e1091. 23. Freeman RK, Van Woerkom JM, Vyverberg A, Ascioti AJ. Esophageal stent placement for the treatment of spontaneous esophageal perforations. Ann Thorac Surg 2009;88:194e198. 24. Orringer MB, Stirling MC. Esophagectomy for esophageal disruption. Ann Thorac Surg 1990;49:35e42; discussion 42e43. 25. Sung SW, Park JJ, Kim YT, Kim JH. Surgery in thoracic esophageal perforation: primary repair is feasible. Dis Esophagus 2002;15:204e209. 26. Hourneaux de Moura EG, Toma K, Goh KL, et al. Stents for benign and malignant esophageal strictures. Ann NY Acad Sci 2013;1300:119e143. 27. Wilson JL, Louie BE, Farivar AS, et al. Fully covered self-expanding metal stents are effective for benign esophagogastric disruptions and strictures. J Gastrointest Surg 2013;17:2045e2050.
Discussion DR KEN SHARP (Nashville, TN): This is not your usual retrospective review, in which you cherry-pick your good patients and ignore the patients who went to surgery and did even better. Instead, this is a retrospective review of a change in practice. It’s important to note that all patients the authors saw after 2007 were treated with this protocol, and no single patient required salvage or rescue operations such as a thoracotomy to repair a perforation that did not respond to the stent placement. So it’s a little different than your usual retrospective review. In the manuscript, it was not clear that the abdominal perforations were in this algorithm, but that’s been clarified by the presentation. Please clarify how many patients had abdominal perforation, and were their results better or worse than the thoracic perforations as far as morbidity and mortality? You mentioned that it’s difficult to find predictors of treatment failure, length of stay, and morbidity, but the data presented say that all patients were treated within 24 hours of presentation to the University of Florida. There are a large number of patients in this study who came from outside institutions. And we don’t know the duration of their perforation before they came to our attention. It’s intuitive that the person who comes in 72 hours after perforation will do worse than somebody who has a fresh perforation. Do you have any data in your data set that would support this? And do you have any opinions as to whether or not a patient having a late presentation would do better or worse with an operative intervention? Finally, I have 2 questions about disease-specific entities. First, no patients were recorded as having esophageal cancer and a perforation. Were any other patients during this time frame seen with a perforated esophageal cancer, and how were they treated? What would your recommendations be for a patient who has a perforation in the setting of an obstructing esophageal carcinoma? Second, there are patients in this series with achalasia. Stenting a nonrelaxing lower esophageal sphincter may allow this perforation to heal, but it doesn’t treat the underlying disease. Did these patients do well after stent removals because presumably their
J Am Coll Surg
achalasia remained untreated? This seems to violate the tenet that you should treat the disease, not just the complication. DR MARY T HAWN (Birmingham, AL): The group from Florida reported their experience in managing 76 patients with esophageal perforation over a 6-year period. Their outcomes are excellent and commendable. And this is a very important algorithm contribution to the literature. My main questions relate to their algorithm. I think they didn’t clearly state who placed the stents. Was it a surgeon or was it a gastroenterologist? How did you coordinate and sequence that care? Did you have somebody with that expertise available to you 24/7? If not, how did you get these patients to the operating room so quickly to manage them? Were the stents always performed in the operating room? Did that require any adaptation of the operating room suite to perform this procedure? Finally, it appears that a lot of the patients, almost 40%, had stent migration and required some additional manipulation. Did you look at risk factors for stent migration, and did you test for stent occlusion at the time of the original placement? Are there patients in whom stent placement is unlikely to be successful and an alternative first therapy might be more efficient? DR DOUGLAS J MINNICH (Birmingham, AL): Patients with esophageal perforation are challenging, because of the clinical severity of the patient problem as well as the variety of patient presentations. Two of the primary issues when evaluating these patients relate to delineating the time from injury to presentation, as well as determining whether the perforation is contained within the mediastinum. Time from injury is often known with an iatrogenic perforation, but can sometimes be challenging with a spontaneous perforation. The overall components of therapy include intravenous antibiotics, control of sepsis, and nutritional support. The variability in managing these patients comes from exactly how sepsis is controlled. Your group should be commended on the low morbidity and mortality in the series. I have 4 questions for further discussion. For the intubated ICU patient, do you think that patient is getting as much benefit from the stent being in place since they are unable to tolerate oral intake at that time? For these patients, would you consider gastric decompression and portal drainage as an initial treatment, and reserve the stent for persistent extravasation once the patient is able to tolerate oral intake and have a formal esophagram? For early patients with peritoneal involvement, your operative algorithm was to perform a 360-degree fundoplication in addition to the repair. Most of these patients in your series were iatrogenic, so, by definition, there was some level of underlying esophageal pathology that prompted the procedure. Although I agree that a loosely performed 360-degree fundoplication has a low chance of being obstructive, have you considered any noncircumferential wraps that could still cover the area of repair? Was leak occlusion obtained in 100% of the patients in whom stents were placed? I have often found it difficult to obtain a proximal seal of the stent. This is worse when cases of the underlying disease delineate some level of esophageal dilatation that can
BACKGROUND:
either diagnostic or therapeutic endoscopic or surgical procedures, or spontaneous, such as Boerhaave’s syndrome. The majority of perforations are iatrogenic in origin and are located in the thoracic rather than the cervical or intra-abdominal esophagus.1 Despite aggressive therapy, the mortality rate of thoracic and abdominal perforations remains as high as 20% in recent published reports.2,3 Traditional management of thoracic and abdominal EP has been prompt surgery, with primary repair of the perforation the preferred approach in a patient without underlying esophageal pathology or prolonged inflammation.4 Despite data suggesting that early surgical repair of perforations is the preferred approach, up to 30% of repairs demonstrate a persistent leak and may require additional esophageal procedures.5 Although many authors
Esophageal perforation is a rare but potentially lethal condition that requires prompt recognition and management to achieve a favorable outcome. Esophageal perforations (EP) are typically classified as iatrogenic, associated with Disclosure Information: Dr Ben-David receives pay as a consultant and received a research grant from Ethicon; Dr Forsmark is paid as a consultant by Applied Clinical Intelligence; and Dr Hochwald is paid as a consultant for Ethicon Endosurgery. All other authors have nothing to declare. Presented at the Southern Surgical Association 125th Annual Meeting, Hot Springs, VA, December 2013. Received December 14, 2013; Accepted December 17, 2013. From the Departments of Surgery (Ben-David, Behrns, Rossidis, Caban, Crippen, Caranasos, Hughes, Sarosi) and Medicine (Dragonov, Forsmark, Chauhan, Wagh), University of Florida, Gainesville, FL; and Roswell Park Cancer Center, Buffalo, NY (Hochwald). Correspondence address: Kfir Ben-David, MD, FACS, Division of General Surgery, 1600 SW Archer Rd, Room 6164, P.O. Box 100109, Gainesville, FL 32610. email: [email protected]
ª 2014 by the American College of Surgeons Published by Elsevier Inc.
768
ISSN 1072-7515/14/$36.00 http://dx.doi.org/10.1016/j.jamcollsurg.2013.12.033
Vol. 218, No. 4, April 2014
Ben-David et al
advocate for open, operative management of esophageal perforations, others have demonstrated excellent outcomes with nonoperative treatment of perforations and percutaneous control of mediastinal sepsis. Vogel and colleagues6 reported no deaths and 96% esophageal healing in a group of 28 patients with thoracic perforations treated nonoperatively. With the development of selfexpanding covered stents used for treating both benign perforations and anastomotic leaks in the esophagus,7 we reasoned that stent occlusion of an EP could be combined with laparoscopic or thoracoscopic surgical drainage and/or repair with enteral nutritional access to effectively treat esophageal perforations. Esophageal stenting and nonoperative management was developed and implemented using a multidisciplinary minimally invasive protocol for the treatment of EP. We previously reported our pilot series of our first 11 patients.8 Herein, the results of all patients with EP treated under this protocol since 2007 are reported. The results support the effectiveness of a stent-first strategy in treating esophageal perforation at a high volume esophageal disease treatment center.
METHODS An IRB-approved retrospective review of acute EP between 2007 and 2013 was performed. Patient demographic, clinical outcomes data, and hospital charges were collected. We reviewed our series of 76 consecutive EP patients, who presented or were transferred to our tertiary care center between July 2007 and January 2013. All patients were managed according to our treatment algorithm for acute EP (Fig. 1). Diagnostic management included a water-soluble contrast study performed as an esophagram or CT of the chest using Gastrografin (Bracco Diagnostics Inc). Patients who were noted to have a contained leak within the mediastinum were treated with intravenous antibiotics and chest-tube drainage, if appropriate. Patients noted to have uncontained esophageal contrast extravasation were treated with an endoluminally placed covered esophageal stent and laparoscopic and/or video assisted thoracoscopic surgery with drainage of the posterior mediastinum. These patients also received laparoscopic enteral feeding access, usually in the form of a feeding jejunostomy. Patients who had an uncontained leak within the abdominal cavity were managed based on time elapsed from initial presentation to radiographic study. Patients who presented within 24 hours and had an uncontained abdominal leak were taken to the operating room for primary EP repair and a 360-degree loose fundoplication in order to cover the primary repair. If the esophageal perforation
Esophageal Perforation Treatment Algorithm
769
was greater than 2 cm, an esophageal stent (Alveolus Alimaxx, Merit Medical Endotek) was also used to aid with appropriate coverage and sealing of the esophageal perforation. The size of perforation was measured endoscopically at the time of initial stent placement. For patients who presented beyond 24 hours from the injury, the treatment protocol involved esophageal stenting, drainage, and distal enteral feeding access. In all cases after stent placement, occlusion of the esophageal perforation with stent was confirmed within 24 to 48 hours with either an esophagogram or CT of the chest. If leak occlusion was verified, nonintubated patients were started on a clear liquid diet and advanced to full liquid diet until their outpatient follow-up appointments. The patients were also discharged home on nocturnal cycled enteral feeds for nutritional support. The patients returned 4weeks post-stent placement for outpatient stent removal and follow-up esophagogram. All patients were assessed in clinic for symptoms of dysphagia or other esophageal disorders after their esophageal stent was removed. If the leak was not occluded with the initial stent placement, repeat endoscopy and repositioning of the stent were undertaken. Patients with successful stenting were evaluated on a daily basis with chest x-rays to detect evidence of stent migration. If stent migration occurred, or the clinical condition deteriorated, the patient was reevaluated with an esophagogram, and endoscopic repositioning of the esophageal stent was carried out. This management algorithm was repeated until leak occlusion was confirmed and maintained. If a leak occurred at the proximal end of the esophageal stent, an additional stent was deployed, creating a stentwithin-a-stent to ensure occlusion.
RESULTS We reviewed 76 consecutive patients with acute esophageal perforation treated at our tertiary-care center. Forty-nine patients were hospital-to-hospital transfers, and 27 patients presented primarily to our center. The demographics of our patient population are shown in Table 1. The median and mean ages were 64 and 61, respectively (range 25 to 87 years; SD 16 years). There were 11 patients who were 80 years or older and 34% were older than 70 years. There was a male predominance, with 50 men and 26 women within our group. Sixty-four patients were Caucasian, 9 were AfricanAmerican and 3 were Hispanic (Table 1). The majority of our EP (n ¼ 69) were located in the distal esophagus within approximately 4 cm of the gastroesophageal junction. The remaining esophageal perforations were in the mid-thoracic (n ¼ 5) and cervical esophagus (n ¼ 2).
770
Ben-David et al
Esophageal Perforation Treatment Algorithm
J Am Coll Surg
Figure 1. Esophageal perforation treatment algorithm.
The mean size of esophageal perforation was 3.4 cm, ranging from 1 cm to 11 cm (SD 1.67 cm). Sixty-eight (90%) of our esophageal perforation patients were noted to have uncontained extravasation of oral contrast into the mediastinum, pleural space, or Table 1.
Patient Demographics
Demographic
Age, y Mean Median Range Sex, n (%) Male Female Race, n (%) White Black Hispanic
Data
61 64 25e87 50 (66) 26 (34) 64 (84) 9 (11) 3 (4)
peritoneum. Ten of these patients (15%) had contrast extravasation into the mediastinum and pleural space, with contamination of the thoracic cavity. Eight of these were the result of esophageal instrumentation and 2 perforations were spontaneous (Boerhaave’s syndrome). Fifty-eight of 68 patients (85%) had extravasation into the distal posterior mediastinum with contamination into the abdominal peritoneal cavity. Fifty-one of these patients had an iatrogenic esophageal perforation due to esophageal instrumentation for various esophageal diseases. The remaining 7 were the result of spontaneous esophageal perforation after aggressive emesis. Ten percent (8 of 76) of patients presented with a contained esophageal perforation: 2 cervical, 4 thoracic, and 2 contained distal esophageal leaks (Table 2). Based on the patients’ clinical status, they were admitted to general care or the ICU. Twenty-seven patients who did not require ICU admission were able to start an oral liquid diet immediately after their swallow study post stent placement. Forty-nine patients required ICU admission, with
Vol. 218, No. 4, April 2014
Table 2.
Ben-David et al
Esophageal Perforation by Location and Etiology
Leak type and location
Contained Cervical Mid-thoracic Distal Contained Thoracic, mediastinal, and pleural Uncontained Thoracic, mediastinal, and pleural
Etiology
n
%
Iatrogenic Iatrogenic Iatrogenic Spontaneous
2 4 1 1
2.6 5.3 1.3 1.3
Iatrogenic Spontaneous
8 2
10.5 2.6
Iatrogenic Spontaneous
51 7
67.0 9.9
median and mean lengths of ICU stay of 3 and 10 days, respectively (range 1 to 46 days). The overall median and mean hospital stays were 10 and 18 days, respectively, with a range of 1 to 86 days. Twenty-five (33%) of the patients were noted to have prolonged intubation (>7 days) complicated by pneumonia and required a tracheostomy (Table 3). In this patient group the median and mean ICU stays were 25 and 30 days, respectively. These patients also had longer median and mean overall hospital stays of 34 and 38.5 days, respectively, and were all discharged to a rehabilitation facility. The remaining 51 patients were discharged to their residence. Twenty-six patients experienced stent migration requiring repositioning within the first week of stent Table 3.
Postoperative Data
Variable
Mortality, n (%) Overall length of stay, d Mean Median Range ICU admission, n (%) ICU length of stay, d Mean Median Range ICU patients with prolonged intubation (>7 days), n (%) Patients able to resume oral intake within 48 hours, n (%) Patients with stent migration requiring intervention, n (%) Mean stent interventions, n
Data
1 (1.3) 18 10 (1e86) 49 (64.5) 10 3 (1e46) 25 (51) 27 (35.5) 26 (34.2) 1.7
Esophageal Perforation Treatment Algorithm
771
placement. Nineteen of the 26 patients with stent migration had to have their stent repositioned 1 to 3 times, with mean and median of 1.7 and 2 stent adjustments, respectively. For patients who had a prolonged hospital course, their esophageal stent was exchanged every 4 to 5 weeks after initial placement. One in-hospital mortality occurred within 30 days as a result of hemorrhagic shock due to an atria-esophageal fistula related to catheterdirected atrial fibrillation ablation thermal injury. Median and mean total hospital charges for EP at our institution are $85,945 and $161,698, respectively. The median and mean lengths of time to final stent removal with confirmation of healing of the perforation were 28 and 36 days, respectively, demonstrating that this modality allows for rapid healing of the esophageal injury. None of the patients in our series required any additional esophageal surgical procedure for their EP other than what was initially performed per our treatment algorithm. The esophageal salvage rate in our series was 100%, with no conversion to an open esophageal repair or esophagectomy due to their EP.
DISCUSSION Although there are a multitude of etiologies, the most common cause of esophageal perforation is iatrogenic, related to esophageal instrumentation. This acute esophageal event is a dreaded life-threatening surgical condition diagnosed by clinical suspicion, diagnostic radiographic studies, or endoscopy. With historically high morbidity and mortality rates, the need for rapid detection and initiation of effective treatment is paramount.9-11 Multiple factors have been associated with successful outcomes of esophageal perforation management including the time elapsed between rupture and diagnosis, the size and location of the rupture, degree of contamination, and the overall health of the patient. In fact, time from perforation to diagnosis is considered the most crucial factor influencing the outcome of esophageal perforation.12,13 Despite the morbidity of the perforation, the optimal treatment remains uncertain because there is no single strategy that is sufficient to deal with the majority of these situations. Although many believe that open surgical repair of the perforation is the most effective treatment for esophageal perforations,9-11 our institution has reported that aggressive conservative therapy in patients with esophageal perforations results in lower morbidity and mortality while avoiding major open surgery and its associated morbidities.6,8 In addition, with the advancements in both endoscopic and minimally invasive techniques, esophageal perforation treatments have gradually shifted toward self-expanding endoluminal stents, endoscopic intraluminal clipping,
772
Ben-David et al
Esophageal Perforation Treatment Algorithm
and application of various sealant products.9,11,13-15 With our institution’s significant clinical experience in minimally invasive treatment of benign and malignant esophageal diseases, our approach to treatment of acute esophageal perforation has been to use an endoscopically placed, self-expanding esophageal stent to cover the perforation, occluding the leak.8 We strongly believe that the addition of an endoscopically placed removable covered stent results in rapid leak occlusion and sepsis control, resulting in decreasing contamination, morbidity, and hospital stay. This modality also enables treatment of esophageal perforations in patients with significant comorbidities, and/or elderly patients, who are often poor candidates for thoracotomy. Our observation that 26% of patients who presented with an acute esophageal perforation were 70 years old or older demonstrates the high risk population this condition afflicts. Early reports, including our own institution’s previous description of successful nonoperative treatment of esophageal perforations, have demonstrated success with this approach, especially if the perforation is confined to the mediastinum, a contained leak that easily drains back into the esophagus and is associated with minimal signs of clinical sepsis.16-18 This conservative treatment is centered on broad-spectrum antibiotics, wide drainage, and nothing by mouth. Most authors have advocated that nonoperative esophageal perforation treatment should be reserved for patients with micro-perforations and no evidence of ongoing leaks.19 However, perforations that did not meet these criteria have undergone surgical intervention to minimize mediastinal and/or peritoneal contamination. Despite well-documented results with aggressive nonoperative management, soilage of the mediastinum is of great concern.20 Therefore, this type of nonoperative treatment has not been adopted in all esophageal perforation patients because of the high mortality rate associated with contamination of the mediastinum and peritoneal cavity, resulting in significant clinical sepsis.21,22 Evidence suggests that mortality among patients treated within 24 hours of sustaining an esophageal injury is decreased by 10-fold compared with mortality in those for whom diagnosis and treatment are delayed.4 The success of early treatment strengthens our current notion that sealing the esophageal perforation is of critical importance to prevent thoracic or abdominal soilage. Therefore, all of our patients were treated with stenting within 24 hours of their initial presentation to our institution, and leak occlusion was confirmed within 48 hours of esophageal stent placement. Because of the initial success with intraluminal esophageal stent placement, this approach has become our first line treatment of choice when managing
J Am Coll Surg
patients with acute esophageal perforation. This has evolved into our current treatment algorithm (Fig. 1). Our protocol incorporates a combination of aggressive medical management with the wide use of intraluminal esophageal stent to control and seal esophageal leaks without a thoracotomy or a laparotomy and their associated surgical morbidities. In line with other studies,23 we have demonstrated that endoscopically placed esophageal stents and distal feeding access have resulted in lower rates of reoperation, open surgical intervention, esophagectomies, morbidity, and mortality than those that have been reported with standard surgical treatment.24,25 We had only 1 in-hospital mortality due to hemorrhagic shock from an iatrogenic atrial-esophageal fistula that resulted from catheter-based atrial ablation for treatment of atrial fibrillation. Additionally, none of the patients in our series required conversion to open esophageal surgical procedures after initial management with stenting. Although we had 26 patients who experienced stent migration requiring repositioning within the first week of stent placement, this was done at the bedside in the ICU with no adverse events. These patients had longer ICU stays, hospital courses, stent intervals, and higher hospital charges and costs. Unfortunately, we were not able to identify predictors that could prospectively identify the patients or esophageal perforations that result in the longer hospital course as a result of stent migration. To try to prevent this event, we have started to suture all of the stents with an absorbable suture, such as 2-0 chromic, laparoscopically to help anchor the stent in place at the time of initial placement and theoretically prevent migration. The use of esophageal stents allowed us to start early oral intake in 27 patients who were started on an oral liquid diet immediately after their post-stent placement swallow study. This treatment algorithm has also decreased our hospital stays and resulted in less overall morbidity for the patients, theoretically reducing medical cost and patient convalescence. In fact, our median lengths of ICU and hospital stay were 3 and 10 days, respectively, and only 33% of our patients were noted to have prolonged intubation (>7 days), pneumonia and required a tracheostomy. Esophageal stents need to be exchanged or removed because the complications associated with prolonged intubation of the esophagus are well known,20 but the exact time course for stent removal is still unknown. Most studies recommend a period of 10 days for small esophageal perforations and as long as 8 weeks for extended esophageal wall defects.20,26,27 We used fully covered metallic stents to allow ease of removal and to minimize tissue in-growth.
Vol. 218, No. 4, April 2014
Ben-David et al
It is important to emphasize that although this is a retrospective series of EP, it represents a treatment algorithm for the most common forms of esophageal perforations, iatrogenic and spontaneous, in all of the patients who presented to our institution from 2007 to 2013. Our protocol allowed us to be consistent in providing the same care to various types of esophageal perforations and clinical presentation at our institution regardless of physicians involved. We were also able to use the expertise of multiple medical services, allowing for a multidisciplinary care approach for these complex patients. This multidisciplinary method enabled us to care for severely ill patients while minimizing the morbidity and mortality and avoiding major open esophageal surgery.
CONCLUSIONS We have been able to demonstrate that endoscopically placed removable esophageal stents with aggressive conservative therapy is an effective method of treatment for patients with esophageal perforation at our tertiary care institution. This treatment protocol enables us to care for severely ill patients, minimizing the morbidity and mortality associated with major open esophageal surgery. We strongly believe that this multidisciplinary approach can be easily implemented at other institutions caring for esophageal perforation patients. Author Contributions Study conception and design: Ben-David, Hochwald, Rossidis, Caban, Draganov, Forsmark, Chauhan, Wagh, Sarosi Acquisition of data: Ben-David, Rossidis, Crippen, Caranasos Analysis and interpretation of data: Ben-David, Hochwald, Rossidis, Caban, Crippen, Caranasos, Sarosi Drafting of manuscript: Ben-David, Behrns, Hochwald, Rossidis, Hughes, Sarosi Critical revision: Ben-David, Behrns, Hochwald, Sarosi REFERENCES 1. Hasimoto CN, Cataneo C, Eldib R, et al. Efficacy of surgical versus conservative treatment in esophageal perforation: a systematic review of case series studies. Acta cirurgica brasileira / Sociedade Brasileira para Desenvolvimento Pesquisa em Cirurgia 2013;28:266e271. 2. Bhatia P, Fortin D, Inculet RI, Malthaner RA. Current concepts in the management of esophageal perforations: a twentyseven year Canadian experience. Ann Thorac Surg 2011;92: 209e215. 3. Ryom P, Ravn JB, Penninga L, et al. Aetiology, treatment and mortality after oesophageal perforation in Denmark. Danish Med Bull 2011;58:A4267.
Esophageal Perforation Treatment Algorithm
773
4. Brinster CJ, Singhal S, Lee L, et al. Evolving options in the management of esophageal perforation. Ann Thorac Surg 2004;77:1475e1483. 5. Wright CD, Mathisen DJ, Wain JC, et al. Reinforced primary repair of thoracic esophageal perforation. Ann Thorac Surg 1995;60:245e248; discussion 248e249. 6. Vogel SB, Rout WR, Martin TD, Abbitt PL. Esophageal perforation in adults: aggressive, conservative treatment lowers morbidity and mortality. Ann Surg 2005;241:1016e1021; discussion 1021e1023. 7. van Boeckel PG, Sijbring A, Vleggaar FP, Siersema PD. Systematic review: temporary stent placement for benign rupture or anastomotic leak of the oesophagus. Alimentary Pharma Therapeu 2011;33:1292e1301. 8. Ben-David K, Lopes J, Hochwald S, et al. Minimally invasive treatment of esophageal perforation using a multidisciplinary treatment algorithm: a case series. Endoscopy 2011;43: 160e162. 9. Biancari F, D’Andrea V, Paone R, et al. Current treatment and outcome of esophageal perforations in adults: systematic review and meta-analysis of 75 studies. World J Surg 2013;37: 1051e1059. 10. Huber-Lang M, Henne-Bruns D, Schmitz B, Wuerl P. Esophageal perforation: principles of diagnosis and surgical management. Surg Today 2006;36:332e340. 11. Schweigert M, Beattie R, Solymosi N, et al. Endoscopic stent insertion versus primary operative management for spontaneous rupture of the esophagus (Boerhaave syndrome): an international study comparing the outcome. Am Surg 2013;79:634e640. 12. Martinek J, Kovacova S, Nosek V, et al. Successful endoscopic treatment (clipping) of esophageal perforation during balloon dilatation in a patient with achalasia. Endoscopy 2008;40(Suppl 2):E61e62. 13. Sung HY, Kim JI, Cheung DY, et al. Successful endoscopic hemoclipping of an esophageal perforation. Dis Esophagus: official journal of the International Society for Diseases of the Esophagus / ISDE 2007;20:449e452. 14. Gomez-Esquivel R, Raju GS. Endoscopic closure of acute esophageal perforations. Curr Gastroenterol Rep 2013;15:321. 15. Kowalski C, Kastuar S, Mehta V, Brolin RE. Endoscopic injection of fibrin sealant in repair of gastrojejunostomy leak after laparoscopic Roux-en-Y gastric bypass. Surg Obes Relat Dis: official journal of the American Society for Bariatric Surgery 2007;3:438e442. 16. Cameron JL, Kieffer RF, Hendrix TR, et al. Selective nonoperative management of contained intrathoracic esophageal disruptions. Ann Thorac Surg 1979;27:404e408. 17. Lyons WS, Seremetis MG, deGuzman VC, Peabody JW Jr. Ruptures and perforations of the esophagus: the case for conservative supportive management. Ann Thorac Surg 1978;25: 346e350. 18. Shirkey HC. Conservative management of esophageal perforation. JAMA 1965;193:537. 19. Kiernan PD, Sheridan MJ, Elster E, et al. Thoracic esophageal perforations. Southern Med J 2003;96:158e163. 20. El Hajj II, Imperiale TF, Rex DK, et al. Treatment of esophageal leaks, fistulae, and perforations with temporary stents: evaluation of efficacy, adverse events, and factors associated with successful outcomes. Gastrointest Endosc 2013; Oct 11. [Epub ahead of print]. 21. Okten I, Cangir AK, Ozdemir N, et al. Management of esophageal perforation. Surg Today 2001;31:36e39.
774
Ben-David et al
Discussion
22. Salo JA, Isolauri JO, Heikkila LJ, et al. Management of delayed esophageal perforation with mediastinal sepsis. Esophagectomy or primary repair? J Thorac Cardiovasc Surg 1993; 106:1088e1091. 23. Freeman RK, Van Woerkom JM, Vyverberg A, Ascioti AJ. Esophageal stent placement for the treatment of spontaneous esophageal perforations. Ann Thorac Surg 2009;88:194e198. 24. Orringer MB, Stirling MC. Esophagectomy for esophageal disruption. Ann Thorac Surg 1990;49:35e42; discussion 42e43. 25. Sung SW, Park JJ, Kim YT, Kim JH. Surgery in thoracic esophageal perforation: primary repair is feasible. Dis Esophagus 2002;15:204e209. 26. Hourneaux de Moura EG, Toma K, Goh KL, et al. Stents for benign and malignant esophageal strictures. Ann NY Acad Sci 2013;1300:119e143. 27. Wilson JL, Louie BE, Farivar AS, et al. Fully covered self-expanding metal stents are effective for benign esophagogastric disruptions and strictures. J Gastrointest Surg 2013;17:2045e2050.
Discussion DR KEN SHARP (Nashville, TN): This is not your usual retrospective review, in which you cherry-pick your good patients and ignore the patients who went to surgery and did even better. Instead, this is a retrospective review of a change in practice. It’s important to note that all patients the authors saw after 2007 were treated with this protocol, and no single patient required salvage or rescue operations such as a thoracotomy to repair a perforation that did not respond to the stent placement. So it’s a little different than your usual retrospective review. In the manuscript, it was not clear that the abdominal perforations were in this algorithm, but that’s been clarified by the presentation. Please clarify how many patients had abdominal perforation, and were their results better or worse than the thoracic perforations as far as morbidity and mortality? You mentioned that it’s difficult to find predictors of treatment failure, length of stay, and morbidity, but the data presented say that all patients were treated within 24 hours of presentation to the University of Florida. There are a large number of patients in this study who came from outside institutions. And we don’t know the duration of their perforation before they came to our attention. It’s intuitive that the person who comes in 72 hours after perforation will do worse than somebody who has a fresh perforation. Do you have any data in your data set that would support this? And do you have any opinions as to whether or not a patient having a late presentation would do better or worse with an operative intervention? Finally, I have 2 questions about disease-specific entities. First, no patients were recorded as having esophageal cancer and a perforation. Were any other patients during this time frame seen with a perforated esophageal cancer, and how were they treated? What would your recommendations be for a patient who has a perforation in the setting of an obstructing esophageal carcinoma? Second, there are patients in this series with achalasia. Stenting a nonrelaxing lower esophageal sphincter may allow this perforation to heal, but it doesn’t treat the underlying disease. Did these patients do well after stent removals because presumably their
J Am Coll Surg
achalasia remained untreated? This seems to violate the tenet that you should treat the disease, not just the complication. DR MARY T HAWN (Birmingham, AL): The group from Florida reported their experience in managing 76 patients with esophageal perforation over a 6-year period. Their outcomes are excellent and commendable. And this is a very important algorithm contribution to the literature. My main questions relate to their algorithm. I think they didn’t clearly state who placed the stents. Was it a surgeon or was it a gastroenterologist? How did you coordinate and sequence that care? Did you have somebody with that expertise available to you 24/7? If not, how did you get these patients to the operating room so quickly to manage them? Were the stents always performed in the operating room? Did that require any adaptation of the operating room suite to perform this procedure? Finally, it appears that a lot of the patients, almost 40%, had stent migration and required some additional manipulation. Did you look at risk factors for stent migration, and did you test for stent occlusion at the time of the original placement? Are there patients in whom stent placement is unlikely to be successful and an alternative first therapy might be more efficient? DR DOUGLAS J MINNICH (Birmingham, AL): Patients with esophageal perforation are challenging, because of the clinical severity of the patient problem as well as the variety of patient presentations. Two of the primary issues when evaluating these patients relate to delineating the time from injury to presentation, as well as determining whether the perforation is contained within the mediastinum. Time from injury is often known with an iatrogenic perforation, but can sometimes be challenging with a spontaneous perforation. The overall components of therapy include intravenous antibiotics, control of sepsis, and nutritional support. The variability in managing these patients comes from exactly how sepsis is controlled. Your group should be commended on the low morbidity and mortality in the series. I have 4 questions for further discussion. For the intubated ICU patient, do you think that patient is getting as much benefit from the stent being in place since they are unable to tolerate oral intake at that time? For these patients, would you consider gastric decompression and portal drainage as an initial treatment, and reserve the stent for persistent extravasation once the patient is able to tolerate oral intake and have a formal esophagram? For early patients with peritoneal involvement, your operative algorithm was to perform a 360-degree fundoplication in addition to the repair. Most of these patients in your series were iatrogenic, so, by definition, there was some level of underlying esophageal pathology that prompted the procedure. Although I agree that a loosely performed 360-degree fundoplication has a low chance of being obstructive, have you considered any noncircumferential wraps that could still cover the area of repair? Was leak occlusion obtained in 100% of the patients in whom stents were placed? I have often found it difficult to obtain a proximal seal of the stent. This is worse when cases of the underlying disease delineate some level of esophageal dilatation that can
