Thoracic esophageal perforations. (Original Article).

Background: Recognition of the importance of early diagnosis and aggressive, definitive surgical intervention has brought about a dramatic decline in mortality related to distal esophageal perforation.

Methods: We retrospectively analyzed all cases of thoracic esophageal perforation diagnosed at our hospital from September 1, 1979, through April 1, 2001. The study group consisted of 62 patients (43 men) with a mean age of 58.8 years (range, 20-92 yr).

Results: In the group of 39 patients with early diagnosis ([less than or equal to]24 h), hospital survival was 87%, which increased to 93% when early diagnosis was combined with aggressive surgical treatment. Among the 23 patients with late diagnosis (>24 h), survival approached 70%. Yet, in patients who were treated aggressively with surgery, survival was almost 90% despite delayed diagnosis.

Conclusion: We recommend aggressive, definitive surgery for thoracic esophageal perforations, whether diagnosed early or late. A variety of options are discussed with regard to complicated presentations.

**********

Thoracic esophageal perforation, like any malady encountered infrequently, may be managed badly. We think that our experience and the lessons we have learned during the past 22 years may help others to improve the timely diagnosis and appropriate treatment of patients with this condition.

Patients and Methods

We retrospectively analyzed all cases of thoracic esophageal perforation diagnosed at the Inova Fairfax Hospital (Annandale, VA) from September 1, 1979, through September 1, 1990. (1) After performing that analysis, we prospectively studied all patients with thoracic esophageal perforation who were admitted through April 1, 2001.

The study group consisted of 62 patients (43 men) with a mean age of 58.8 years (range, 20-92 yr). Most commonly, patients presented with abdominal pain and associated nausea and/or vomiting. Nine patients were known to have either achalasia or benign esophageal stricture, 12 others had esophageal cancer, and 2 others had undergone resection for gastric carcinoma. One patient had a history of pneumonectomy for non-small cell carcinoma of the lung and thereafter had mediastinal recurrence, related radiotherapy, and consequent esophageal perforation with the formation of an esophagopleural fistula.

The diagnosis of perforation was confirmed preoperatively in 58 (94%) of the 62 patients. Esophagograms were obtained in 46 patients, computed tomography (CT) with contrast medium was performed in 9 patients, and both diagnostic methods were used in 2 patients. Methylene blue swallow resulted in dye recognition in a chest tube in one patient. In four patients, the diagnosis was determined intraoperatively; two of these patients underwent nondiagnostic, preoperative esophagography.

Perforations resulted from dilation or endoscopic false passage in 32 patients, including 6 who had dilation for obstructive esophageal carcinoma. Two others had perforations as a result of radiotherapy for carcinoma, and two more incurred delayed perforation because of endoscopic esophageal sclerotherapy for bleeding esophageal varices. One patient had an esophageal perforation caused by a gunshot wound. The remainder (25 patients) had spontaneous perforations (Boerhaave's syndrome), including one perforation caused by tuberculosis that had been present for 6 months before therapy.

Thirty-nine (63%) of the 62 patients were admitted and began treatment within the first 24 hours after perforation, and 23 (3 7%) were admitted more than 24 hours after perforation, with 12 of the latter group being admitted 10 or more days after the onset of symptoms. Seventeen patients were treated medically, and seven others were treated with conservative (ie, expectant) initial surgery, with treatment thereafter being related to complications. Two of these seven patients were treated initially with chest tube drainage alone before eventually requiring surgery. Thirty-eight patients were treated with early, aggressive surgical intervention.

Results

Hospital mortality (Table 1) in the group of 39 patients who were diagnosed early ([less than or equal to]24 h after perforation) was 13%, compared with 30% in the group of 23 patients who received a late diagnosis (>24 h). The mortality rates for iatrogenic and spontaneous perforations were 15 and 29% (Table 2). Both patients with perforations as a result of radiotherapy for squamous cell carcinoma died. Because the cause of perforation (iatrogenic or spontaneous) for these two patients was unknown, they were excluded from some calculations, as specifically stated hereafter. Of the 24 patients who were treated conservatively, 9 (3 8%) died. The mortality rates for conservative medical therapy (12 patients), conservative surgical therapy (7 patients), and aggressive endoscopic placement of covered stents (5 patients) were 33, 43, and 40%, respectively (Table 1). Of the 38 patients who underwent aggressive, definitive surgery, only 3 (8%) died (Table 1). In this group, 27 (93%) of 29 patients who were diagnose d early survived. The only deaths that occurred in this latter group were two patients in their 90s. One had a spontaneous distal esophageal perforation with complicating cardiopulmonary arrest and severe aspiration. This patient (who should not have undergone surgery) died within 48 hours of surgery without regaining any appreciable systemic pressure. The second was a 92-year-old patient transferred from a nursing home for evaluation of gastrointestinal bleeding. Both cervical and thoracic esophageal perforations occurred during diagnostic endoscopy and were promptly repaired primarily. Both sites healed well, as confirmed by postoperative esophagography. Nevertheless, the patient's convalescence was complicated by renal failure caused by the original bleeding and consequent hypotension. At the family's request, all supportive care was discontinued, and the patient died as a result of renal failure.

Treatment Provided

Medical Management (17 Patients). Of the 17 patients treated medically, 11(65%) survived. In five of these survivors, the diagnosis was made on the basis of a classic esophageal rupture history and confirmed by significant mediastinal free air on plain films and/or computed tomographic scans, with esophageal injury suggested by sinus tract at esophagography, yet, most important, without free flow of contrast medium into the mediastinal or pleural cavities. Excluding these five survivors from the study would have reduced the medical management survival rate from 17 to 12 patients with esophageal rupture and extravasation and survival from 65 to 50%.

Five patients were treated with endoscopic placement of stents after esophageal rupture, and extravasation was diagnosed as many as 10 days after perforation. Three of these patients survived aggressive medical management with the insertion of a covered Wallstent (Boston Scientific/Medi-tech, Natick, MA) and were discharged from the hospital. Two of these did not require further intervention. One of the three died 4 years later as a result of massive exsanguination caused by stent erosion of the esophagus with esophagoaortic fistula. The two patients who did not survive stent placement could not be resuscitated further. Both had stent migration and underwent futile further attempts at stent replacement, with subsequent perforation, sepsis, and death. Hospital mortality associated with stent placement alone was 40%. Hospital mortality for the remaining seven patients treated medically with or without drainage was 57%.

Conservative Surgery (7 Patients). Seven patients were managed with conservative surgery, as previously defined in this article. Four patients underwent surgery with exploration and drainage alone. Two of them survived after hospitalizations of 11/2 and 31/2 months, and the other two died without further intervention. Initial chest tube drainage and observation failed in three patients. One of these patients underwent transhiatal resection with primary extrathoracic anastomosis and ultimately died as a result of refractory mediastinal sepsis. The original diagnosis of esophageal perforation was delayed for 3 weeks because of missed recognition of an abnormal esophagogram after delivery of radio-and chemotherapy for a midesophageal squamous cell carcinoma. The remaining two patients recovered; one after delayed transhiatal resection and reconstruction and the other after a cervical diversion with exclusion of the thoracic perforation, gastrostomy and jejunostomy, and delayed restoration of esophageal function. Hospital mortality was 43% in this group managed with conservative surgery.

Aggressive Surgery (38 Patients). Of the 38 surgical patients managed aggressively, 14 had primary closure and bolstering tissue wrap (14% mortality); 4 had esophageal diversion or resection, with thoracic esophageal exclusion and later alimentary reconstruction (0% mortality); 15 had transhiatal resection with primary anastomosis (0% mortality), and 5 had thoracoabdominal resection (Ivor-Lewis procedure) and simultaneous reconstruction (20% mortality). Mortality for the entire group of patients treated aggressively was 8% (3 of 38 patients).

Hospital Mortality

Hospital mortality is summarized in Tables 2 and 3. Table 2 shows mortality by demographic factors as well as by history, cause, anatomy, and diagnosis. Mortality seemed to be increased by malignancy, advanced age, spontaneous etiology, and delayed diagnosis (>24 h). The two patients who had esophageal perforation as a result of radiotherapy (both of whom ultimately died in the hospital) were included in the category "other," along with those who had perforations caused by gunshots and tuberculosis. Table 3 shows mortality and length of stay according to surgery type. If the two patients who underwent radiotherapy had been excluded, mortality associated with transhiatal repair and nonoperative therapy would have decreased to 0 and 25%, respectively. Neither of the radiation-induced cases was included in the length-of-stay computations, because the one nonoperative case was managed on an outpatient basis.

Discussion

Recognition of the importance of early diagnosis and aggressive, definitive surgical intervention has brought about a dramatic decline in mortality related to distal esophageal perforation. (1-9) Of the 39 patients in this study who had diagnosis within 24 hours of perforation, 34 (87%) survived. Of the 29 patients who had early diagnosis and aggressive surgical treatment, 27 (93%) survived. Even among patients whose diagnosis was delayed, aggressive surgical treatment resulted in an 89% survival rate. In those patients treated with conservative surgery (eg, expectantly, responding to complications, whether diagnosis was early or late), survival was only 57%. In patients with malignancy of the esophagus, all who had early diagnosis and aggressive surgical treatment survived, whereas all who had late diagnosis and/or conservative medical treatment died.

One elderly woman presented a challenge, despite prompt diagnosis of iatrogenic injury. Twelve months before perforation, she had undergone a near-total gastrectomy for cancer with proximal Roux-en-y gastrojejunostomy, and a distal esophageal stricture subsequently developed because of alkaline reflux. To avoid a two-stage repair in an elderly patient at risk for recurrent cancer, primary repair of the perforation was performed with a rotated diaphragmatic flap, (10) based on a well-vascularized pericardiophrenic pedicle (Fig. 1). Recuperation was uncomplicated, and the patient was discharged to home on postoperative Day 10.

Reports from major institutions with significant surgical experience have noted mortality rates approaching 59 to 68%. (4,11) Furthermore, the degree of appropriateness of conservative medical management remains a topic of considerable controversy. (12-17) In our experience, all five patients with "microperforations" survived. Four of these patients were diagnosed within 24 hours, and all were treated with observation, NPO status, and broad-spectrum intravenous antibiotics.

Two other patients with perforations diagnosed more than 24 hours after occurrence and showing continuing mediastinal extravasation and pleural contamination were treated conservatively and with endoesophageal placement of self-expanding covered Wallstents. Both patients survived, demonstrating a role for appropriately selected medical management, including endoesophageal stents. (18,19) In fact, the recovery of one of these patients bordered on the miraculous. A 38-year-old white woman admitted for liver transplantation for alcohol-induced hepatic failure was hospitalized for bleeding esophageal varices with significant coagulopathy and underwent transjugular intrahepatic portosystemic shunting (TIPS procedure (20)) and endoesophageal sclerotherapy. Right-sided esophageal necrosis and perforation occurred, with consequent right pleural empyema. Ipsilateral tube thoracostomy, systemic antibiotics, and intravenous hyperalimentation (subsequently changed to enteral [ie, jejunal] feedings) were instituted. There after a covered stent was placed, but it migrated into the stomach. Another Wallstent was placed, and the fistula disappeared during the ensuing 7 to 14 days. Neither surgical intervention for esophageal repair or replacement nor decortication was ultimately necessary, and the patient was discharged to home 69 days after stent placement.

Of the remaining 10 patients, 7 received conservative medical treatment and 3 received stents. Four (57%) of the 7 conservatively treated patients died. Two of the three who received stents died. The stent recipient who initially survived died 2 years later as a result of massive exsanguination because of stent erosion that resulted in aortoesophageal fistula. Thus, stent placement alone was associated with early (hospital) and late (long-term) mortality of 40 and 60%.

We have no experience with T-tube management or absorbable staples. (21-23) Yet, in our experience and that of others, (3) diversion with temporary thoracic esophageal exclusion has been as effective as resection without immediate reconstruction. Moreover, diversion offers significantly simpler prospects with regard to reconstruction. Chromic catgut ligatures, which are used to ligate the cervical esophagus distal to diverting cervical esophagostomy and the distal esophagus proximal to gastrostomy, generally dissolve within 30 days. Once a contrast esophagogram shows restored esophageal patency and the perforation has healed (uniformly in our experience), the cervical (le, side-to-skin) esophagostomy is taken down and closed. Thus, alimentary continuity is restored without the need for more complicated reconstructive efforts. In our limited experience (three patients), patients treated in this manner to date have not had esophageal strictures at the site of either ligation or perforation.

The University of Michigan group has achieved excellent results in using the option of primary repair with pleural wrap in cases diagnosed early and without esophageal obstruction. (24,25) So, too, our preferred treatments also include primary repair and transhiatal resection with cervical reconstruction. Transhiatal resection and reconstruction may be particularly helpful in patients in whom distal esophageal obstruction exists (21 patients in our experience). Moreover, extrathoracic reconstruction seems desirable in cases diagnosed more than 24 hours after perforation or in cases with heavy contamination in the chest or the abdomen.

Conclusions

We continue to support aggressive, definitive surgery in patients with thoracic esophageal perforations that are diagnosed early and with extravasation of contrast medium. Moreover, we unanimously support conservative therapy for microperforations (or sinuses) or for rare cases such as fistulae caused by tuberculosis (26) without associated continuing contamination of mediastinal or pleural cavities. Although stent placement in patients with benign and malignant disease remains controversial, (18,27,28) rigid application of aggressive resuscitation is not always appropriate and, when it is not, may be replaced by covered stent placement and/or compassionate support. Our own experience, as well as that of others, (18,27,28) suggests that progressive stent expansion may be a potential cause of full-thickness esophageal necrosis and fistulae, and close clinical follow-up is important.

On the basis of our experience with six patients with perforation because of dilation of malignant strictures, we espouse a treatment protocol (ie, prophylaxis) for malignant strictures devoid of esophageal dilation with the use of jejunostomy for nutritional delivery, followed first by neoadjuvant therapy and then by surgical resection (PD Kieman, MJ Sheridan, E Elster, et al, manuscript submitted for publication). When perforation occurred as a complication of esophageal malignancy, early diagnosis followed by aggressive surgical intervention resulted in a 100% survival rate in our series. Late recognition in such circumstances and/or conservative medical measures yielded no survivors.

Table 1

Hospital mortality influenced by delay in diagnosis and by therapeutic
approach (a)

                                      1979-1989, no. (%)

Treatment                  [less than or equal to]24 h    >24 h

Medical (n = 17)
Nonsurgical                         0 S, 1 D            0 S, 2 D
Stenting                            0 S, 0 D            0 S, 0 D
Surgical (n = 45)
Aggressive                          6 S, 0 D            0 S, 0 D
Conservative                        1 S, 1 D            1 S, 1 D
Total no. of patients (%)           9 (22.2%)           4 (75.0%)
Total of all patients (%)          13 (38.5%)

                                     1990-Current, no. (%)

Treatment                  [less than or equal to]24 h    >24 h

Medical (n = 17)
Nonsurgical                         5 S, 0 D            3 S, 1 D
Stenting                            1 S, 1 D            2 S, 1 D
Surgical (n = 45)
Aggressive                         21 S, 2 D            8 S, 1 D
Conservative                        0 S, 0 D            2 S, 1 D
Total no. of patients (%)          30 (10.0%)           19 (21.1%)
Total of all patients (%)          49 (14.3%)


                           Total no. of    No. of
Treatment                    patients    deaths (%)

Medical (n = 17)
Nonsurgical                     12        4 (33.3%)
Stenting                         5        2 (40.0%)
Surgical (n = 45)
Aggressive                      38        3 (7.9%)
Conservative                     7        3 (42.9%)
Total no. of patients (%)       62       12 (19.4%)
Total of all patients (%)

(a)S, survived; D, died in hospital.

Table 2

Demographics and factors influencing

                                      Associated
Factors                               mortality (%)

Sex
  Male (n = 42)                       16.7%
  Female (n = 20)                     25.0% (P = 0.73)
History
  Benign (n = 50)                     14.3%
  Malignant (n = 12)                  41.7% (P = 0.04)
Age (yr)
  [less than or equal to]59 (n = 28)  14.3%
  60-79 (n = 31)                      19.4%
  [greater than or equal to]80        66.7% (P = 0.10)
   (n = 3)
Etiology
  Iatrogenic (n = 34)                 14.7%
  Spontaneous (n = 24)                29.2%
  Other (n = 4)                       50% (P = 0.14)
Anatomy
  With obstruction (n = 21)           14.3%
  Without obstructin (n = 41)         22% (P = 0.73)
Diagnosis
  [less than or equal to]24 h         12.8%
  (n = 39)
  >24 h (n = 23)                      30.4% (P = 0.11)

Table 3

Mortality and length of stay


                                    No. of    No. of
Surgery type                       patients  deaths (%)

Esophageal diversion                  3       0 (0%)
Exploratory laparotomy with chest     4       2 (50%)
 tube drainage
Thoracoabdominal resection and        5       1 (20%)
 reconstruction
Nonoperative (chest tube drain)      12       4 (33%)
Nonoperative (stent)                  5       2 (40%)
Primary closure                      14       2 (14%)
Resection without reconstruction      2       0 (0%)
Transhiatal resection with           17       1 (6%)
 reconstruction

Total                                62      12 (19.4%)

                                             Length of stay (d)

Surgery type                       Mean              Min       Max

Esophageal diversion                43               22         71
Exploratory laparotomy with chest  104               52        155
 tube drainage
Thoracoabdominal resection and      32               11         61
 reconstruction
Nonoperative (chest tube drain)     14                1         62
Nonoperative (stent)                33                2         69
Primary closure                     36                2        139
Resection without reconstruction    60               28         91
Transhiatal resection with          24                9         66
 reconstruction

Total

Accepted May 15, 2002.

References

(1.) Kieman PD. Conte JV Jr, Petri R, O'Hara K, Byrne WD, Fulcher TM, et al. Thoracic esophageal perforations at a Virginia hospital 1979- 1990. Va Med Q 1992;119:102-104.

(2.) Skinner DB, Little AG, DeMeester TR. Management of esophageal perforation. Am J Surg 1980;139:760-764.

(3.) Urschel HC Jr, Razzuk MA, Wood RE, Galbraith N, Pockey M, Paulson DL. Improved management of esophageal perforation: Exclusion and diversion in continuity. Ann Surg 1974;179:587-591.

(4.) Bladergroen MR, Lowe JE, Postlethwait RW. Diagnosis and recommended management of esophageal perforation and rupture. Ann Thorac Surg 1986;42:235-239.

(5.) Attar S, Hankins JR, Suter CM, Coughlin TR, Sequeira A, McLaughlin JS. Esophageal perforation: A therapeutic challenge. Ann Thorac Surg 1990;50:45-51.

(6.) Reeder LB, DeFilippi VJ, Ferguson MK. Current results of therapy for esophageal perforation. Am J Surg 1995;169:615-617.

(7.) Sawyer R, Phillips C, Vakil N. Short- and long-term outcome of esophageal perforation. Gastrointest Endosc 1995;41:130-134.

(8.) Wright CD, Mathisen DJ, Wain JC, Moncure AC, Hilgenberg AD, Grillo HC. Reinforced primary repair of thoracic esophageal perforation. Ann Thorac Surg 1995;60:245-249.

(9.) White RK, Morris DM. Diagnosis and management of esophageal perforations. Am Surg 1992;58:112-119.

(10.) Richardson JD, Tobin GR. Closure of esophageal defects with muscle flaps. Arch Surg 1994;129:541-548.

(11.) Salo JA, Isolauri JO, Heikkila LJ, Markkula HT, Heikkinen LO, Kivilaakso EO, et al. Management of delayed esophageal perforation with mediastinal sepsis: Esophagectomy or primary repair? J Thorac Cardiovasc Surg 1993;106:1088-1091.

(12.) 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;25:346-350.

(13.) Ivey TD, Simonowitz DA, Dillard DH, Miller DW Jr. Boerhaave syndrome: Successful conservative management in three patients with late presentation. Am J Surg 1981;141:531-533.

(14.) Swedlund A, Traube M, Siskind BN, McCallum RW. Nonsurgical management of esophageal perforation from pneumatic dilatation in achalasia. Dig Dis Sci 1989;34:379-384.

(15.) Cameron JL, Kieffer RF, Hendrix TR, Mehigan DG, Baker RR. Selective nonoperative management of contained intrathoracic esophageal disruptions. Ann Thorac Surg 1979;27:404-408.

(16.) Shaffer HA Jr, Valenzuela G, Mittal RK. Esophageal perforation: A reassessment of the criteria for choosing medical or surgical therapy. Arch Intern Med 1992;152:757-761.

(17.) Molina EG, Stollman N, Grauer L, Reiner DK, Barkin JS. Conservative management of esophageal nontransmural tears after pneumatic dilation for achalasia. An: J Gastroenterol 1996;91:15-18.

(18.) Adam A, Watkinson AF, Dussek J. Boerhaave syndrome: To treat or not to treat by means of insertion of a metallic stent. J Vasc Interv Radiol 1995;6:741-746.

(19.) Clouse TL, Takriti M, Gordon WH, Just-Viera JO. Surgical repair of esophageal perforation in cirrhotic patients with varices. Chest 1994; 105:1896-1898.

(20.) Zemel G, Katzen BT, Beeker GJ, Benenati JF, Sallee DS. Percutaneous transjugular portosystemic shunt, JAMA 1991;266:390-393.

(21.) Paramesh V, Rumisek JD, Chang FC. Spontaneous recanalization of the esophagus after exclusion using nonabsorbable staples. Ann Thorac Surg 1995;59:1214-1216.

(22.) Chang CH, Lin PJ, Chang JP, Hsieh MJ, Lee MC, Chu JJ. One-stage operation for treatment after delayed diagnosis of thoracic esophageal perforation. Ann Thorac Surg 1992;53:617-620.

(23.) Bufkin BL, Miller JI Jr, Mansour KA. Esophageal perforation: emphasis on management. Ann Thorac Surg 1996;61:1447-1452.

(24.) Whyte RI, Iannettoni MD, Orringer MB. Intrathoracic esophageal perforation: The merit of primary repair. J Thorac Cardiovasc Surg 1995;109:140-146.

(25.) Iannettoni MD, Vlessis AA, Whyte RI, Orringer MB. Functional outcome after surgical treatment of esophageal perforation. Ann Thorac Surg 1997;64:1606-1610.

(26.) Grubbs BC, Baldwin DR, Trenkner SW, McCabe RP Jr, Maddaus MA. Distal esophageal perforation caused by tuberculosis. J Thorac Cardiovasc Surg 2001;121:1003-1004.

(27.) Maier A, Pinter H, Friehs GB, Renner H, Smolle-Juttner FM. Self-expandable coated stent after intraluminal treatment of esophageal cancer: A risky procedure? Ann Thorac Surg 1999;67:781-784.

(28.) Christie NA, Buenaventura PO, Femando HC, Nguyen NT, Weigel TL, Ferson PF, et al. Results of expandable metal stents for malignant esophageal obstruction in 100 patients: Short-term and long-term follow-up. Ann Thorac Surg 2001;71:1797-1802.

RELATED ARTICLE: Key Points

* A retrospective analysis of patients with thoracic esophageal perforation is reported.

* Optimal survival was achieved in those patients who were treated early and aggressively with surgery.

* Esophageal exclusion and proximal diversion is suited to patients who are diagnosed late or too precarious to undergo more extensive resuscitation, and ligation is transient in effect, allowing reconstitution of anatomic continuity with takedown of the proximal esophageal diversion.

* Conservative medical therapy can reasonably be applied in patients with "microperforations" with no continuing leak.

From the Section of Thoracic and Cardiovascular Surgery, the Department of Medicine, and the Section of Thoracic and Cardiovascular Operating Room Nursing, Inova Fairfax Hospital, Inova Health System, Annandale, VA. (Dr. Elster is now with the Department of Surgery, Bethesda Naval Hospital, Bethesda, MD.)

Reprint requests to Paul D. Kiernan, MD, Cardiovascular and Thoracic Surgery Associates, PC, 3301 woodburn Road, Suite 301, Annandale, VA 22003.

Copyright [c] 2003 by The Southern Medical Association 0038-4348/03/9602-0158