- Email: [email protected]
Management of esophageal perforations: the value of aggressive surgical treatment
The American Journal of Surgery 190 (2005) 161–165
Management of esophageal perforations: the value of aggressive surgical treatment J. David Richardson, M.D.* From the Department of Surgery, University of Louisville School of Medicine, Louisville, KY 40292, USA Manuscript received April 13, 2005; accepted manuscript April 15, 2005 Presented at the Kentucky Chapter of American College of Surgeons and Hiram C. Polk, Jr, Surgical Society, February 4, 2005, Louisville, KY
Abstract Background: The treatment of esophageal perforation remains controversial, particularly in terms of the type of operative therapy. This report analyzed results of an aggressive treatment protocol. Methods: Patients with esophageal perforations in a normal esophagus or those with a motor disorder were treated by operative closure. All defects were buttressed or closed by either muscle or pleura. Sternocleidomastoid muscle was used to buttress or primarily close the defects in the neck, and a flap of diaphragm was often used for thoracic perforation. Patients with perforated cancer or severe underlying disease had an esophagectomy. Results: Sixty-four patients had operation: 50 underwent preservation of the esophagus after closure of the perforation and 14 underwent resection. The leak rate was 17%, but all healed. One patient treated with primary closure died (1.5% mortality); only 1 patient required subsequent esophagectomy. Thirteen of 14 patients treated with esophagectomy had an excellent result. Conclusion: The aggressive approach to esophageal perforations with attempt at uniform closure or resection of severe disease produced excellent results with reduced morbidity and low mortality. © 2005 Excerpta Medica Inc. All rights reserved. Keywords: Esophageal perforation; Muscle flap
The treatment of esophageal perforations has received considerable attention in the surgical literature but remains a highly controversial topic. Areas of controversy include (1) the role of nonoperative treatment versus surgical therapy; (2) the management of patients with delayed presentation; and (3) the type of surgical therapy to be performed, particularly whether drainage alone is adequate treatment. In 1985, our unit [1] reported a group of patients managed by a variety of strategies; as a result of that review, we developed an aggressive protocol that attempted to close most perforations of the esophagus. This report outlines the 20-year personal experience of the author with this approach.
* Corresponding author. Tel.: ⫹1-502-852-5452; fax: ⫹1-502-8528915. E-mail address: [email protected]
Clinical Material Sixty-four patients with esophageal perforation were treated by operations performed by a single surgeon from 1985 to 2004. Five patients with suspected leaks had intramural contrast and few signs of inflammation; they were managed nonoperatively, did well, and will not be discussed further. Patients included in the operative group had perforations caused by trauma, iatrogenic injury, and by emetogenic rupture (Boerhaave’s syndrome). Patients with caustic injury and anastomotic leak were excluded. Patients ranged in age from 23 to 84 years, with an average age of 46 years. Fifty-one were men. The site of esophageal perforation was the cervical esophagus in 13 patients, none of whom had malignant lesions. The thoracic esophageal disruptions were divided into the upper and middle esophagus and those lesions within 5 cm of the gastroesophageal junction. Fifteen patients had perforation in the upper or middle esophagus; 9 were benign and 6 had underlying cancers. Thirtysix patients had lower esophageal perforation: 33 occurred
0002-9610/05/$ – see front matter © 2005 Excerpta Medica Inc. All rights reserved. doi:10.1016/j.amjsurg.2005.05.004
162
J.D. Richardson / The American Journal of Surgery 190 (2005) 161–165 Table 3 Causes of iatrogenic perforation
Table 1 Esophageal perforation: location and disease state
Cervical Thoracic Upper and middle G E junction
Benign
Malignant
13
0
9 33 55
6 3 9
in patients without cancers and 3 had underlying malignancies (Table 1). The etiology of the perforations is indicated in Table 2. Traumatic injuries accounted for only 9 cases. Iatrogenic perforations produced 34 of the lesions treated operatively. Eighteen patients had emetogenic rupture near the gastroesophageal junction. The causes of the iatrogenic perforations were varied, but nearly every instrument that passed into the aerodigestive tract was responsible for at least an occasional injury (Table 3). Fifteen patients were operated on within 12 hours of the occurrence of perforation, 28 between 12 and 24 hours, and 21 patients after 24 hours. Fourteen of these patients had operative treatment greater than 48 hours after perforation. Repair was attempted in all patients with a normal esophagus before the perforation. Patients with motor disorders were also treated with primary repair. Those with a diagnosis of achalasia had a concomitant myotomy performed in 4 of 5 cases. If the patient had an underlying malignancy or severely diseased esophagus (particularly with fibrotic strictures), resection was performed using the stomach as a conduit. Cervical esophagostomy for esophageal diversion was used only in those patients in whom mediastinal sepsis was ongoing and uncontrolled by other means. Exclusion and diversion in continuity [2] were not used in any patient.
Operative Management When possible, the esophageal defect was closed by suture repair; in early treatment, it was often possible to perform a separate 2-layer closure of mucosa and muscularis. With delayed operations, a 1-layer closure was attempted if feasible; often no direct closure was possible because of friability of the tissue. In some cases, the esophageal tear appeared partially sealed or may have healed without surgical repair. Nonetheless, closure was attempted using either suture repair or overlay flaps over the defects. Table 2 Etiology of perforation Cervical Iatrogenic Trauma
Thoracic 7 6
Iatrogenic Trauma Spontaneous (emetogenic)
27 3 18
Attempted endotracheal ERCP EGD Endoscopic ultrasound Nasogastric tube Gavage tube Dilation (stricture) Pneumatic dilation Lap Nissen Foreign body removal Stent placement
3 1 2 4 2 1 6 5 3 4 3
ERCP ⫽ endoscopic retrograde cholangiopancreatography; EGD ⫽ esophagogastroduodenoscopy.
Perforations untreated for several days were often difficult to visualize because of associated mediastinitis and empyema. A decortication was performed, and necrotic mediastinal tissue was débrided. If the esophagus was difficult to mobilize at the site of perforation, it was isolated at an area remote from the area of perforation. This was generally straightforward to accomplish; continued gentle dissection along the esophagus usually allowed the area of perforation to be visualized and mobilized. Necrotic tissue was débrided, and the mucosa was exposed and closed by suture or muscle flap. In patients thought to be at high risk for persistent leak, a tube gastrostomy for drainage and jejunostomy for feeding were added via a small left subcostal incision. Every esophageal hole that was treated operatively by primary repair had the defect covered with a buttress of either muscle or pleura. Muscle flaps were used preferentially because they provided a more reliable flap to close the defect. All wounds of the cervical esophagus were buttressed or closed primarily using a muscle flap of sternocleidomastoid muscle. Generally, a portion of the muscle was detached from the sternal end and maneuvered into position to cover the suture line. The sternocleidomastoid muscle was used as a primary closure in 1 case involving a gunshot wound to the cervical esophagus, which injured the posterior tracheal wall and anterior esophagus. Suture closure would have caused marked narrowing of the esophagus. A muscle flap was sutured over the defect with absorbable sutures for primary repair. Patients with thoracic perforations were repaired via thoracotomy. Proximal and midesophageal perforations of the thoracic esophagus were approached via a right thoracotomy, whereas distal lesions were exposed by a left-sided approach. In upper esophageal lesions, a rhomboid muscle flap [3] was used on 1 occasion to close a perforation primarily. The pleura was used to buttress the closure on 4 occasions. Perforations or injuries to the distal esophagus were either closed primarily by suture repair with a diaphragm buttress or directly by a diaphragm flap. The technique of diaphragm flap closure is illustrated in Figure 1 [4]. A
J.D. Richardson / The American Journal of Surgery 190 (2005) 161–165
Fig. 1. The diaphragm lies in close proximity to the distal esophagus. It is often necessary to mobilize the area of perforation to use the muscle flap properly.
posteriorly based flap is created of adequate length and width to close the area of perforation (Figs. 2 and 3). In many patients, these flaps can reach to near the level of the azygous vein. The diaphragm is then closed primarily with nonabsorbable sutures in either 1 or 2 layers (Fig. 4). Fifty-one patients suffered thoracic esophageal perforations. Thirty-seven underwent repair aimed at esophageal preservation: 25 underwent primary suture repair with a buttress closure, and 12 underwent closure of the defect with a muscle flap. Eleven patients had primary diaphragm closure and one patient with a proximal defect underwent rhomboid flap repair.
Results All of the 13 patients with cervical perforation survived. The patient treated with primary muscle flap closure had contrast extravasation that had closed by day 12. No patient required reoperation, and delayed cervical or mediastinal
Fig. 2. A full-thickness flap of diaphragm is mobilized. The sutures are placed in the defect before placing them through the muscular flap and tying them in place. It is necessary to suture the flap to the defect and not simply “cover” the defect.
163
Fig. 3. The flap is approximated over the defect using sutures.
abscess did not occur. Twelve patients had normal swallowing postoperatively. The patient with the muscle flap closure and leak required several dilations but was able to eat a regular diet. One death occurred in the 37 patients with thoracic perforations in whom esophageal preservation was attempted. This represented the only death in the 64 cases treated operatively (1.5% mortality). The death occurred in a 68-year-old alcoholic who was transferred to our care 48 hours after diagnosis of an emetogenic rupture and delirium tremens. A leak into the left chest was closed by a diaphragm flap and a decortication was performed, but the patient continued to be febrile. Numerous contrast studies failed to disclose subsequent extravasation from the esophagus; however, the patient continued with a septic course requiring ventilator assistance. Eventually, a cervical esophagostomy was performed, which did not change the course of his disease. He died 5 weeks after perforation. Autopsy disclosed bilateral pneumonia. No abscesses were found, and the esophageal perforation appeared well healed. Of the 36 remaining patients treated by nonresectional methods, 7 had evidence of contrast extravasation at greater than 2 weeks. All of these cases were complicated problems, and 5 had primary muscle flap closure. All residual
Fig. 4. The esophageal defect is closed in 1 or 2 layers.
164
J.D. Richardson / The American Journal of Surgery 190 (2005) 161–165
leaks had healed within 30 days. Oral intake was resumed by at least postoperative day 10 in 27 patients. All patients were eventually able to eat a regular diet. Four patients had postoperative dysphagia treated by esophageal dilation. Three had reasonable resolution of symptoms, but 1 patient was treated by esophagectomy 2 years after the initial treatment. This patient had a perforation at the time of stricture dilation, and the scarring never relented. The patient did well after resection. Three patients underwent drainage of recurrent intrathoracic abscesses: percutaneous drainage was required in 1 patient and repeat thoracotomy in 2 patients. No complications related to the diaphragm flap occurred, and no patient developed an intraabdominal infection. Fourteen patients were treated with esophageal resection. Nine patients had perforations of malignant lesions, and 5 were benign. A right thoracotomy was used in 8 patients and a left-sided approach in 6 cases. Thirteen patients underwent reconstruction using the stomach. In 1 complex case of a perforated stricture in a patient with a previous distal gastrectomy, the esophagus was resected with subsequent reconstruction via a colonic interposition. All patients survived, and none had an esophageal leak. One patient with a benign stricture that was perforated by dilation had a resection via a left thoracotomy. The gastric conduit was unable to be well mobilized, and the proximal stomach was used for reconstruction. The patient was troubled by persistent reflux and had a very poor postoperative result. He refused further therapy before succumbing to a myocardial infarction 9 months postoperatively. Three patients were treated by concomitant cervical esophagostomy, including the aforementioned single death in the series. The 2 additional patients had loop ostomies; both patients recovered, and their ostomies were primarily closed 3 and 4 months after their creation.
Comments Several clinical and experimental studies from this department in the early 1980s prompted a reevaluation of our clinical care. Cheadle and Richardson [5] reported excellent results in a series of esophageal injuries and confirmed the value of the buttress technique using either muscle or pleura. The rhomboid flap was used with success to cover defects of the proximal esophagus caused by delayed recognition of blunt injuries in 2 patients [3]. The excellent results achieved in this highly unusual injury by placing muscle over an esophageal defect led to a series of laboratory studies [6] using a feline model to evaluate the efficacy of muscle flaps to close defects in the esophagus. These studies confirmed that defects up to 60% of esophageal lumen could be closed with muscle flaps without stricture formation. The clinical care of large series of esophageal perforations of diverse etiologies leads to several unifying concepts
regardless of the mechanism of perforation [1]. Several patients had previously been treated by exclusion and diversion, which was described in 1974 [2]. The results with this approach were not promising in our experience, and it was not used further. Likewise, several patients treated with a cervical esophagostomy had life-altering problems with that approach even if they survived the perforation. Thus, we used diversion only in patients with severe ongoing sepsis. Although many patients spontaneously healed esophageal perforations after the leak was well drained, some will not. Even if healing did occur, it was often accompanied by stricture formation and required extremely long hospital stays. As a result of these observations, most patients subsequently had aggressive treatment with an attempt to close the perforation. Because of the unreliability of esophageal healing in the face of severe inflammation, a practice of uniform reinforcement of suture lines was adopted. When suture closure was not feasible or would narrow the esophageal lumen, direct closure with a muscle flap was performed. A number of local tissues near the site of perforation have been used for buttress or even primary repair. Omental flaps have been used on numerous occasions [7,8]. Pleural flaps have been advocated by many [9 –12] but were often found to be inadequate, especially in cases of early operative treatment [8]. Intercostal muscle had been used experimentally [13,14] and in several clinical studies [15,16]. In my experience, the intercostal muscle pedicle is often inadequate to cover large lacerations and extensive areas of disruption. Pericardial flaps have been described [17], but their use seemed unwise in view of the need to open the pericardium and expose this sac to contamination. Recently, the use of absorbable mesh with fibrin glue has been reported [18]. The diaphragm flap was described as a means of buttressing the suture lines in elective resections of the fundus [19]. Its thickness with pliability and resemblance to esophageal tissue was noted from the inception of its use [19]. Jara [20] described the use of this technique in a single patient as a buttress to a compromised suture line. Rao et al [21] performed primary closure of a defect in a single patient using a diaphragm flap. Since our report of a number of cases using this technique, several other surgeons have successfully used it on occasion [15,22]. Regardless of the technique chosen, an increasing number of reports noted the use of buttress techniques with a decrease in leaks and mortality in these cases [10,11,22–26]. The use of esophagectomy for perforations was initially reported in the 1950s [27]. Since that time, there have been numerous reports on the use of this technique for several clinical situations. Patients with perforated cancers are best treated by resection [28 –30]. Likewise, the use of resection for patients with severe strictures has been advocated [29]. Our results with resection for perforated cancer and disruption of severe strictures confirmed the safety of this approach. One of the most controversial areas of the management
J.D. Richardson / The American Journal of Surgery 190 (2005) 161–165
of esophageal perforation relates to the central question of whether drainage alone (accomplished either open or closed) will suffice for treatment of this problem or whether a more aggressive surgical approach should be used [31]. Although there are patients who meet the established guidelines [32] for nonaggressive treatment, the trend in the literature seems to favor an aggressive approach. A collective review [25] of 13 series published between 1980 and 1990 disclosed an overall mortality of 22% from esophageal perforation. In those treated by drainage, only the mortality was 34%, whereas suture closure had a 15% death rate. Port et al [26] recently noted that primary repair could be performed in most patients, even when treated late after perforation. He operated on 26 patients with a 3.6% mortality. Kiernan et al [22] noted a mortality of 33% in patients treated by drainage but only 7.9% in those aggressively closed. Kotsis et al [16] noted no patient with a “reinforced repair” of diaphragm, intercostal muscle, or pleura had a leak contributing to mortality. Sung et al [15] noted only death (5% mortality) in patients treated with aggressive reinforced closure. Although there are clear indications for nonoperative treatment and clearly some leaks will heal with drainage alone, the improved results with attempts at reinforced closure or resection will undoubtedly continue to argue for aggressive operative approach.
References [1] Richardson JD, Martin LF, Borzotta AP, et al. Unifying concepts in treatment of esophageal leaks. Am J Surg 1985;149:157– 62. [2] Urschel HC, Razzuk MA, Wood RE, et al. Improved management of esophageal perforation: exclusion and diversion in continuity. Ann Surg 1974;179:587–91. [3] Lucas AE, Snow N, Tobin GR, et al. Use of the rhomboid major muscle flap for esophageal repair. Ann Thorac Surg 1982;33:619 –23. [4] Richardson JD, Tobin GR. Closure of esophageal defects with muscle flaps. Arch Surg 1994;129:541– 8. [5] Cheadle W, Richardson JD. Options in the management of trauma to the esophagus. Surg Gynecol Obstet 1982;187:634 – 46. [6] Tobin GR, Netscher DT, Williams RA, et al. Epithelial lining methods in esophageal repair: a comparative study using pedicle flaps in cats. Surgery 1985;98:158 – 65. [7] Dicks JR, Majeed AW, Stoddard CJ. Omental wrapping of perforated esophagus. Dis Esophagus 1998;11:276 – 8. [8] Wright CD, Mathisen DJ, Wain JC, et al. Reinforced primary repair of thoracic esophageal perforation. Ann Thorac Surg 1995;60:239 – 44. [9] Grillo HC, Wilkins EW Jr. Esophageal repair following late diagnosis of intrathoracic perforation. Ann Thorac Surg 1975;20:387–99.
165
[10] Gouge TH, Depan HJ, Spencer FC. Experience with the Grillo pleural wrap procedure in 18 patients with perforation of the thoracic esophagus. Am Surg 1989;209:612–9. [11] Zumbro CL, Anstadt MP, Mawulawde K, et al. Surgical management of esophageal perforation: role of esophageal conservation in delayed perforation. Am Surg 2002;68:36 – 40. [12] Attar S, Hankins JR, Suter CM, et al. Esophageal perforation: a therapeutic challenge. Ann Thorac Surg 1990;50:45–51. [13] Dooling JA, Zick HR. Closure of an esophagopleural fistula using onlay intercostal pedicle graft. Ann Thorac Surg 1967;3:553–7. [14] Bryant LR, Eiseman, BE. Experimental evaluation of intercostal pedicle grafts in oesophageal repair. J Thorac Cardiovasc Surg 1965; 50:626 –9. [15] Sung SW, Park JJ, Kim YT, et al. Surgery in thoracic esophageal perforation: primary repair is feasible. Dis Esophagus 2002;15: 204 –9. [16] Kotsis L, Kostic S, Zubovits K. Multimodality treatment of esophageal disruptions. Chest 1997;112:1304 –9. [17] Millard AH. Spontaneous perforation of the esophagus treated by utilization of a pericardial flap. Br J Surg 1971;58:70 –2. [18] Bardaxoglou E, Mangawas D, Meunier B, et al. New approach to surgical management of early esophageal thoracic perforation: primary suture repair reinforced with absorbable mesh and fibrin glue. World J Surg 1997;21:618 –21. [19] Petrovsky BV. The use of diaphragm grafts for plastic operations in thoracic surgery. J Thorac Cardiovasc Surg 1961;41:348 –55. [20] Jara FM. Diaphragmatic pedicle flap for the treatment of Boerhaave’s syndrome. J Thorac Cardiovasc Surg 1979;78:931–3. [21] Rao KV, Mir M, Cogbill CL. Management of perforations of the thoracic esophagus. A new technique utilizing a pedicle flap of diaphragm. Am J Surg 1974;127:609 –12. [22] Kiernan PD, Sheridan MJ, Elster E, et al. Thoracic esophageal perforations. South Med J 2003;96:158 – 63. [23] Michel L, Grollo HC, Malt RC. Operative and nonoperative management of esophageal perforations. Ann Surg 1980;194:57– 63. [24] Bufkin BL, Miller JI, Mansour KA. Esophageal perforation: emphasis on management. Ann Thorac Surg 1996;61:1447–52. [25] Jones WGII, Ginsberg RJ. Esophageal perforation: a continuing challenge. Ann Thorac Surg 1992;53:534 – 43. [26] Port JL, Kent MS, Korst RJ, et al. Thoracic esophageal perforations: a decade of experience. Ann Thorac Surg 2003;75:1071– 4. [27] Johnson J, Schwegman CW, MacVaugh H III. Early esophagogastrostomy in the treatment of iatrogenic perforation of the distal esophagostomy. J Thorac Cardiovasc Surg 1956;32:827–31. [28] Altorjay A, Kiss J, Voros A, et al. The role of esophagectomy in the management of esophageal perforation. Ann Thorac Surg 1998;65: 1433– 6. [29] Iannettoni MD, Vlessis AA, Whyte RI, et al. Outcome after surgical treatment of esophageal perforation. Ann Thorac Surg 1997;64: 1606 –10. [30] Yeo CJ, Killemoe KD, Klein AS, et al. Treatment of instrument perforation of esophageal malignancy by transhiatal esophagectomy. Arch Surg 1998;123:1016 – 8. [31] Altorjay A, Kiss J, Voros A, et al. Nonoperative management of esophageal perforations—is it justified? Ann Surg 1997;225:415–21. [32] Cameron JL, Kieffer RF. Selective nonoperative management of contained intrathoracic esophageal disruptions. Ann Thorac Surg 1979;27:404 – 8.
Management of esophageal perforations: the value of aggressive surgical treatment J. David Richardson, M.D.* From the Department of Surgery, University of Louisville School of Medicine, Louisville, KY 40292, USA Manuscript received April 13, 2005; accepted manuscript April 15, 2005 Presented at the Kentucky Chapter of American College of Surgeons and Hiram C. Polk, Jr, Surgical Society, February 4, 2005, Louisville, KY
Abstract Background: The treatment of esophageal perforation remains controversial, particularly in terms of the type of operative therapy. This report analyzed results of an aggressive treatment protocol. Methods: Patients with esophageal perforations in a normal esophagus or those with a motor disorder were treated by operative closure. All defects were buttressed or closed by either muscle or pleura. Sternocleidomastoid muscle was used to buttress or primarily close the defects in the neck, and a flap of diaphragm was often used for thoracic perforation. Patients with perforated cancer or severe underlying disease had an esophagectomy. Results: Sixty-four patients had operation: 50 underwent preservation of the esophagus after closure of the perforation and 14 underwent resection. The leak rate was 17%, but all healed. One patient treated with primary closure died (1.5% mortality); only 1 patient required subsequent esophagectomy. Thirteen of 14 patients treated with esophagectomy had an excellent result. Conclusion: The aggressive approach to esophageal perforations with attempt at uniform closure or resection of severe disease produced excellent results with reduced morbidity and low mortality. © 2005 Excerpta Medica Inc. All rights reserved. Keywords: Esophageal perforation; Muscle flap
The treatment of esophageal perforations has received considerable attention in the surgical literature but remains a highly controversial topic. Areas of controversy include (1) the role of nonoperative treatment versus surgical therapy; (2) the management of patients with delayed presentation; and (3) the type of surgical therapy to be performed, particularly whether drainage alone is adequate treatment. In 1985, our unit [1] reported a group of patients managed by a variety of strategies; as a result of that review, we developed an aggressive protocol that attempted to close most perforations of the esophagus. This report outlines the 20-year personal experience of the author with this approach.
* Corresponding author. Tel.: ⫹1-502-852-5452; fax: ⫹1-502-8528915. E-mail address: [email protected]
Clinical Material Sixty-four patients with esophageal perforation were treated by operations performed by a single surgeon from 1985 to 2004. Five patients with suspected leaks had intramural contrast and few signs of inflammation; they were managed nonoperatively, did well, and will not be discussed further. Patients included in the operative group had perforations caused by trauma, iatrogenic injury, and by emetogenic rupture (Boerhaave’s syndrome). Patients with caustic injury and anastomotic leak were excluded. Patients ranged in age from 23 to 84 years, with an average age of 46 years. Fifty-one were men. The site of esophageal perforation was the cervical esophagus in 13 patients, none of whom had malignant lesions. The thoracic esophageal disruptions were divided into the upper and middle esophagus and those lesions within 5 cm of the gastroesophageal junction. Fifteen patients had perforation in the upper or middle esophagus; 9 were benign and 6 had underlying cancers. Thirtysix patients had lower esophageal perforation: 33 occurred
0002-9610/05/$ – see front matter © 2005 Excerpta Medica Inc. All rights reserved. doi:10.1016/j.amjsurg.2005.05.004
162
J.D. Richardson / The American Journal of Surgery 190 (2005) 161–165 Table 3 Causes of iatrogenic perforation
Table 1 Esophageal perforation: location and disease state
Cervical Thoracic Upper and middle G E junction
Benign
Malignant
13
0
9 33 55
6 3 9
in patients without cancers and 3 had underlying malignancies (Table 1). The etiology of the perforations is indicated in Table 2. Traumatic injuries accounted for only 9 cases. Iatrogenic perforations produced 34 of the lesions treated operatively. Eighteen patients had emetogenic rupture near the gastroesophageal junction. The causes of the iatrogenic perforations were varied, but nearly every instrument that passed into the aerodigestive tract was responsible for at least an occasional injury (Table 3). Fifteen patients were operated on within 12 hours of the occurrence of perforation, 28 between 12 and 24 hours, and 21 patients after 24 hours. Fourteen of these patients had operative treatment greater than 48 hours after perforation. Repair was attempted in all patients with a normal esophagus before the perforation. Patients with motor disorders were also treated with primary repair. Those with a diagnosis of achalasia had a concomitant myotomy performed in 4 of 5 cases. If the patient had an underlying malignancy or severely diseased esophagus (particularly with fibrotic strictures), resection was performed using the stomach as a conduit. Cervical esophagostomy for esophageal diversion was used only in those patients in whom mediastinal sepsis was ongoing and uncontrolled by other means. Exclusion and diversion in continuity [2] were not used in any patient.
Operative Management When possible, the esophageal defect was closed by suture repair; in early treatment, it was often possible to perform a separate 2-layer closure of mucosa and muscularis. With delayed operations, a 1-layer closure was attempted if feasible; often no direct closure was possible because of friability of the tissue. In some cases, the esophageal tear appeared partially sealed or may have healed without surgical repair. Nonetheless, closure was attempted using either suture repair or overlay flaps over the defects. Table 2 Etiology of perforation Cervical Iatrogenic Trauma
Thoracic 7 6
Iatrogenic Trauma Spontaneous (emetogenic)
27 3 18
Attempted endotracheal ERCP EGD Endoscopic ultrasound Nasogastric tube Gavage tube Dilation (stricture) Pneumatic dilation Lap Nissen Foreign body removal Stent placement
3 1 2 4 2 1 6 5 3 4 3
ERCP ⫽ endoscopic retrograde cholangiopancreatography; EGD ⫽ esophagogastroduodenoscopy.
Perforations untreated for several days were often difficult to visualize because of associated mediastinitis and empyema. A decortication was performed, and necrotic mediastinal tissue was débrided. If the esophagus was difficult to mobilize at the site of perforation, it was isolated at an area remote from the area of perforation. This was generally straightforward to accomplish; continued gentle dissection along the esophagus usually allowed the area of perforation to be visualized and mobilized. Necrotic tissue was débrided, and the mucosa was exposed and closed by suture or muscle flap. In patients thought to be at high risk for persistent leak, a tube gastrostomy for drainage and jejunostomy for feeding were added via a small left subcostal incision. Every esophageal hole that was treated operatively by primary repair had the defect covered with a buttress of either muscle or pleura. Muscle flaps were used preferentially because they provided a more reliable flap to close the defect. All wounds of the cervical esophagus were buttressed or closed primarily using a muscle flap of sternocleidomastoid muscle. Generally, a portion of the muscle was detached from the sternal end and maneuvered into position to cover the suture line. The sternocleidomastoid muscle was used as a primary closure in 1 case involving a gunshot wound to the cervical esophagus, which injured the posterior tracheal wall and anterior esophagus. Suture closure would have caused marked narrowing of the esophagus. A muscle flap was sutured over the defect with absorbable sutures for primary repair. Patients with thoracic perforations were repaired via thoracotomy. Proximal and midesophageal perforations of the thoracic esophagus were approached via a right thoracotomy, whereas distal lesions were exposed by a left-sided approach. In upper esophageal lesions, a rhomboid muscle flap [3] was used on 1 occasion to close a perforation primarily. The pleura was used to buttress the closure on 4 occasions. Perforations or injuries to the distal esophagus were either closed primarily by suture repair with a diaphragm buttress or directly by a diaphragm flap. The technique of diaphragm flap closure is illustrated in Figure 1 [4]. A
J.D. Richardson / The American Journal of Surgery 190 (2005) 161–165
Fig. 1. The diaphragm lies in close proximity to the distal esophagus. It is often necessary to mobilize the area of perforation to use the muscle flap properly.
posteriorly based flap is created of adequate length and width to close the area of perforation (Figs. 2 and 3). In many patients, these flaps can reach to near the level of the azygous vein. The diaphragm is then closed primarily with nonabsorbable sutures in either 1 or 2 layers (Fig. 4). Fifty-one patients suffered thoracic esophageal perforations. Thirty-seven underwent repair aimed at esophageal preservation: 25 underwent primary suture repair with a buttress closure, and 12 underwent closure of the defect with a muscle flap. Eleven patients had primary diaphragm closure and one patient with a proximal defect underwent rhomboid flap repair.
Results All of the 13 patients with cervical perforation survived. The patient treated with primary muscle flap closure had contrast extravasation that had closed by day 12. No patient required reoperation, and delayed cervical or mediastinal
Fig. 2. A full-thickness flap of diaphragm is mobilized. The sutures are placed in the defect before placing them through the muscular flap and tying them in place. It is necessary to suture the flap to the defect and not simply “cover” the defect.
163
Fig. 3. The flap is approximated over the defect using sutures.
abscess did not occur. Twelve patients had normal swallowing postoperatively. The patient with the muscle flap closure and leak required several dilations but was able to eat a regular diet. One death occurred in the 37 patients with thoracic perforations in whom esophageal preservation was attempted. This represented the only death in the 64 cases treated operatively (1.5% mortality). The death occurred in a 68-year-old alcoholic who was transferred to our care 48 hours after diagnosis of an emetogenic rupture and delirium tremens. A leak into the left chest was closed by a diaphragm flap and a decortication was performed, but the patient continued to be febrile. Numerous contrast studies failed to disclose subsequent extravasation from the esophagus; however, the patient continued with a septic course requiring ventilator assistance. Eventually, a cervical esophagostomy was performed, which did not change the course of his disease. He died 5 weeks after perforation. Autopsy disclosed bilateral pneumonia. No abscesses were found, and the esophageal perforation appeared well healed. Of the 36 remaining patients treated by nonresectional methods, 7 had evidence of contrast extravasation at greater than 2 weeks. All of these cases were complicated problems, and 5 had primary muscle flap closure. All residual
Fig. 4. The esophageal defect is closed in 1 or 2 layers.
164
J.D. Richardson / The American Journal of Surgery 190 (2005) 161–165
leaks had healed within 30 days. Oral intake was resumed by at least postoperative day 10 in 27 patients. All patients were eventually able to eat a regular diet. Four patients had postoperative dysphagia treated by esophageal dilation. Three had reasonable resolution of symptoms, but 1 patient was treated by esophagectomy 2 years after the initial treatment. This patient had a perforation at the time of stricture dilation, and the scarring never relented. The patient did well after resection. Three patients underwent drainage of recurrent intrathoracic abscesses: percutaneous drainage was required in 1 patient and repeat thoracotomy in 2 patients. No complications related to the diaphragm flap occurred, and no patient developed an intraabdominal infection. Fourteen patients were treated with esophageal resection. Nine patients had perforations of malignant lesions, and 5 were benign. A right thoracotomy was used in 8 patients and a left-sided approach in 6 cases. Thirteen patients underwent reconstruction using the stomach. In 1 complex case of a perforated stricture in a patient with a previous distal gastrectomy, the esophagus was resected with subsequent reconstruction via a colonic interposition. All patients survived, and none had an esophageal leak. One patient with a benign stricture that was perforated by dilation had a resection via a left thoracotomy. The gastric conduit was unable to be well mobilized, and the proximal stomach was used for reconstruction. The patient was troubled by persistent reflux and had a very poor postoperative result. He refused further therapy before succumbing to a myocardial infarction 9 months postoperatively. Three patients were treated by concomitant cervical esophagostomy, including the aforementioned single death in the series. The 2 additional patients had loop ostomies; both patients recovered, and their ostomies were primarily closed 3 and 4 months after their creation.
Comments Several clinical and experimental studies from this department in the early 1980s prompted a reevaluation of our clinical care. Cheadle and Richardson [5] reported excellent results in a series of esophageal injuries and confirmed the value of the buttress technique using either muscle or pleura. The rhomboid flap was used with success to cover defects of the proximal esophagus caused by delayed recognition of blunt injuries in 2 patients [3]. The excellent results achieved in this highly unusual injury by placing muscle over an esophageal defect led to a series of laboratory studies [6] using a feline model to evaluate the efficacy of muscle flaps to close defects in the esophagus. These studies confirmed that defects up to 60% of esophageal lumen could be closed with muscle flaps without stricture formation. The clinical care of large series of esophageal perforations of diverse etiologies leads to several unifying concepts
regardless of the mechanism of perforation [1]. Several patients had previously been treated by exclusion and diversion, which was described in 1974 [2]. The results with this approach were not promising in our experience, and it was not used further. Likewise, several patients treated with a cervical esophagostomy had life-altering problems with that approach even if they survived the perforation. Thus, we used diversion only in patients with severe ongoing sepsis. Although many patients spontaneously healed esophageal perforations after the leak was well drained, some will not. Even if healing did occur, it was often accompanied by stricture formation and required extremely long hospital stays. As a result of these observations, most patients subsequently had aggressive treatment with an attempt to close the perforation. Because of the unreliability of esophageal healing in the face of severe inflammation, a practice of uniform reinforcement of suture lines was adopted. When suture closure was not feasible or would narrow the esophageal lumen, direct closure with a muscle flap was performed. A number of local tissues near the site of perforation have been used for buttress or even primary repair. Omental flaps have been used on numerous occasions [7,8]. Pleural flaps have been advocated by many [9 –12] but were often found to be inadequate, especially in cases of early operative treatment [8]. Intercostal muscle had been used experimentally [13,14] and in several clinical studies [15,16]. In my experience, the intercostal muscle pedicle is often inadequate to cover large lacerations and extensive areas of disruption. Pericardial flaps have been described [17], but their use seemed unwise in view of the need to open the pericardium and expose this sac to contamination. Recently, the use of absorbable mesh with fibrin glue has been reported [18]. The diaphragm flap was described as a means of buttressing the suture lines in elective resections of the fundus [19]. Its thickness with pliability and resemblance to esophageal tissue was noted from the inception of its use [19]. Jara [20] described the use of this technique in a single patient as a buttress to a compromised suture line. Rao et al [21] performed primary closure of a defect in a single patient using a diaphragm flap. Since our report of a number of cases using this technique, several other surgeons have successfully used it on occasion [15,22]. Regardless of the technique chosen, an increasing number of reports noted the use of buttress techniques with a decrease in leaks and mortality in these cases [10,11,22–26]. The use of esophagectomy for perforations was initially reported in the 1950s [27]. Since that time, there have been numerous reports on the use of this technique for several clinical situations. Patients with perforated cancers are best treated by resection [28 –30]. Likewise, the use of resection for patients with severe strictures has been advocated [29]. Our results with resection for perforated cancer and disruption of severe strictures confirmed the safety of this approach. One of the most controversial areas of the management
J.D. Richardson / The American Journal of Surgery 190 (2005) 161–165
of esophageal perforation relates to the central question of whether drainage alone (accomplished either open or closed) will suffice for treatment of this problem or whether a more aggressive surgical approach should be used [31]. Although there are patients who meet the established guidelines [32] for nonaggressive treatment, the trend in the literature seems to favor an aggressive approach. A collective review [25] of 13 series published between 1980 and 1990 disclosed an overall mortality of 22% from esophageal perforation. In those treated by drainage, only the mortality was 34%, whereas suture closure had a 15% death rate. Port et al [26] recently noted that primary repair could be performed in most patients, even when treated late after perforation. He operated on 26 patients with a 3.6% mortality. Kiernan et al [22] noted a mortality of 33% in patients treated by drainage but only 7.9% in those aggressively closed. Kotsis et al [16] noted no patient with a “reinforced repair” of diaphragm, intercostal muscle, or pleura had a leak contributing to mortality. Sung et al [15] noted only death (5% mortality) in patients treated with aggressive reinforced closure. Although there are clear indications for nonoperative treatment and clearly some leaks will heal with drainage alone, the improved results with attempts at reinforced closure or resection will undoubtedly continue to argue for aggressive operative approach.
References [1] Richardson JD, Martin LF, Borzotta AP, et al. Unifying concepts in treatment of esophageal leaks. Am J Surg 1985;149:157– 62. [2] Urschel HC, Razzuk MA, Wood RE, et al. Improved management of esophageal perforation: exclusion and diversion in continuity. Ann Surg 1974;179:587–91. [3] Lucas AE, Snow N, Tobin GR, et al. Use of the rhomboid major muscle flap for esophageal repair. Ann Thorac Surg 1982;33:619 –23. [4] Richardson JD, Tobin GR. Closure of esophageal defects with muscle flaps. Arch Surg 1994;129:541– 8. [5] Cheadle W, Richardson JD. Options in the management of trauma to the esophagus. Surg Gynecol Obstet 1982;187:634 – 46. [6] Tobin GR, Netscher DT, Williams RA, et al. Epithelial lining methods in esophageal repair: a comparative study using pedicle flaps in cats. Surgery 1985;98:158 – 65. [7] Dicks JR, Majeed AW, Stoddard CJ. Omental wrapping of perforated esophagus. Dis Esophagus 1998;11:276 – 8. [8] Wright CD, Mathisen DJ, Wain JC, et al. Reinforced primary repair of thoracic esophageal perforation. Ann Thorac Surg 1995;60:239 – 44. [9] Grillo HC, Wilkins EW Jr. Esophageal repair following late diagnosis of intrathoracic perforation. Ann Thorac Surg 1975;20:387–99.
165
[10] Gouge TH, Depan HJ, Spencer FC. Experience with the Grillo pleural wrap procedure in 18 patients with perforation of the thoracic esophagus. Am Surg 1989;209:612–9. [11] Zumbro CL, Anstadt MP, Mawulawde K, et al. Surgical management of esophageal perforation: role of esophageal conservation in delayed perforation. Am Surg 2002;68:36 – 40. [12] Attar S, Hankins JR, Suter CM, et al. Esophageal perforation: a therapeutic challenge. Ann Thorac Surg 1990;50:45–51. [13] Dooling JA, Zick HR. Closure of an esophagopleural fistula using onlay intercostal pedicle graft. Ann Thorac Surg 1967;3:553–7. [14] Bryant LR, Eiseman, BE. Experimental evaluation of intercostal pedicle grafts in oesophageal repair. J Thorac Cardiovasc Surg 1965; 50:626 –9. [15] Sung SW, Park JJ, Kim YT, et al. Surgery in thoracic esophageal perforation: primary repair is feasible. Dis Esophagus 2002;15: 204 –9. [16] Kotsis L, Kostic S, Zubovits K. Multimodality treatment of esophageal disruptions. Chest 1997;112:1304 –9. [17] Millard AH. Spontaneous perforation of the esophagus treated by utilization of a pericardial flap. Br J Surg 1971;58:70 –2. [18] Bardaxoglou E, Mangawas D, Meunier B, et al. New approach to surgical management of early esophageal thoracic perforation: primary suture repair reinforced with absorbable mesh and fibrin glue. World J Surg 1997;21:618 –21. [19] Petrovsky BV. The use of diaphragm grafts for plastic operations in thoracic surgery. J Thorac Cardiovasc Surg 1961;41:348 –55. [20] Jara FM. Diaphragmatic pedicle flap for the treatment of Boerhaave’s syndrome. J Thorac Cardiovasc Surg 1979;78:931–3. [21] Rao KV, Mir M, Cogbill CL. Management of perforations of the thoracic esophagus. A new technique utilizing a pedicle flap of diaphragm. Am J Surg 1974;127:609 –12. [22] Kiernan PD, Sheridan MJ, Elster E, et al. Thoracic esophageal perforations. South Med J 2003;96:158 – 63. [23] Michel L, Grollo HC, Malt RC. Operative and nonoperative management of esophageal perforations. Ann Surg 1980;194:57– 63. [24] Bufkin BL, Miller JI, Mansour KA. Esophageal perforation: emphasis on management. Ann Thorac Surg 1996;61:1447–52. [25] Jones WGII, Ginsberg RJ. Esophageal perforation: a continuing challenge. Ann Thorac Surg 1992;53:534 – 43. [26] Port JL, Kent MS, Korst RJ, et al. Thoracic esophageal perforations: a decade of experience. Ann Thorac Surg 2003;75:1071– 4. [27] Johnson J, Schwegman CW, MacVaugh H III. Early esophagogastrostomy in the treatment of iatrogenic perforation of the distal esophagostomy. J Thorac Cardiovasc Surg 1956;32:827–31. [28] Altorjay A, Kiss J, Voros A, et al. The role of esophagectomy in the management of esophageal perforation. Ann Thorac Surg 1998;65: 1433– 6. [29] Iannettoni MD, Vlessis AA, Whyte RI, et al. Outcome after surgical treatment of esophageal perforation. Ann Thorac Surg 1997;64: 1606 –10. [30] Yeo CJ, Killemoe KD, Klein AS, et al. Treatment of instrument perforation of esophageal malignancy by transhiatal esophagectomy. Arch Surg 1998;123:1016 – 8. [31] Altorjay A, Kiss J, Voros A, et al. Nonoperative management of esophageal perforations—is it justified? Ann Surg 1997;225:415–21. [32] Cameron JL, Kieffer RF. Selective nonoperative management of contained intrathoracic esophageal disruptions. Ann Thorac Surg 1979;27:404 – 8.