Ruptures and Perforations of the Esophagus: The Case for Conservative Supportive Management William S. Lyons, M.D., Michael G. Seremetis, M.D., Vicente C. deguzman, M.D., and Joseph W. Peabody, Jr., M.D. ABSTRACT A series of 3 patients treated for ruptures and perforations of the intrathoracic esophagus is reviewed. Eighteen of these patients underwent major thoracotomy; were treated with minor procedures. Two died before treatment could be implemented. Of the 8 undergoing major operations, 7 died; among the managed conservatively there was only death. Based on this experience, we conclude that major surgical repair for esophageal perforation is often unnecessary. It has the additional drawback of sometimes resulting in equally serious secondary procedures. Ruptures and perforations of the intrathoracic esophagus are relatively uncommon but are dangerous. Without treatment they are nearly 00% fatal. Surgical intervention, although representing substantial progress, has often been unsatisfactory when universally applied. We reviewed 3 patients with ruptured esophagus. Eleven were managed conservatively, that is, without operative treatment except for limited thoracotomy for accurate placement of drainage tubes, gastrostomy, and jejunostomy. Eighteen underwent major thoracotomy. We found that the patients who were treated without operation had a much lower mortality rate. In this series we observed that the esophagus had a strong tendency to heal itself, and we believe that in many cases major surgical procedures are not necessary to cure this lesion. From the Georgetown University School of Medicine and the surgical services of Greater Southeast Community Hospital, Sibley Memorial Hospital, and Washington Hospital Center, Washington, DC. Presented at the Twenty-third Annual Meeting of the Southem Thoracic Surgical Association, Nov 4-6,976, Acapulco, Mexico. Address reprint requests to Dr. Lyons, The Atrium Building, Duke and Washington Sts, Alexandria, VA 2234. 346 0003-4975/78/0025-040$.25 @ 978 by William S. Lyons Materials All types of intrathoracic esophageal perforations are represented in our series with the exception of those due to inflatable bags used for esophageal tamponade, such as the Sengstaken-Blakemore tube. Iatrogenic ruptures due to endoscopy and dilations were very common and equal in incidence to those attributable to emetic injury (Boerhaave s syndrome). The causes of esophageal perforation may be conveniently grouped into iatrogenic and noniatrogenic: I. Iatrogenic A. Endoscopy B. Esophageal dilations C. Passage of tubes D. Major thoracotomy. Noniatrogenic A. B. C. D. E. Missiles Foreign bodies Spontaneous rupture (Boerhaave s syndrome) Blunt trauma Caustic injury The nature and distribution of the 3 intrathoracic perforations are shown in Table, with the general mode of treatment and the mortality rate for each group. Three perforations were in the middle third and 28 in the lower third of the esophagus. There were 3 spontaneous perforations due to emetic injury (Boerhaave s syndrome), 3 due to esophagoscopy with or without dilation but not associated with foreign bodies, perforation by a bullet, due to caustic ingestion, and due to severe blunt trauma from an automobile accident. There was injury during a pulmonary operation and foreign body perforation. There were 0 deaths in the total group of 3 patients.
~ ~ ~~~~~ ~ ~ ~~ ~ ~ ~~ 347 Lyons et al: Ruptures and Perforations of the Esophagus Table. Cause of Injury and Treatment Employed in 3 Patients with Esophageal Perforation Spontaneous Instrumental Caustic Surgical Missile Foreign body Blunt trauma Total 3= 3 lb 3 Treatment and Results Major Operation Minor Procedures Type of No. of Perforation Patients Alive Dead Alive Dead - 5 4 5 2 "One died after wrong operation (celiotomy), and died without treatment. bdied without treatment. Treatment The treatment of our 3 patients may be characterized as reparative or supportive. Repair in this series included at least major thoracotomy and suture of the esophageal rent. Supportive therapy involved solely the use of minor procedures, with major thoracotomy avoided. Thus, of the 3 patients, 8 underwent operations intended to be reparative, with 7 deaths, and received supportive care, with death (see Table ). Supportive care included minor surgical and other procedures such as closed thoracostomy (3 instances), insertion of a nasogastric tube (6 instances), gastrostomy (3 instances), and in instance a limited (hand's width) thoracotomy for placement of a chest tube (Table 2). Although this was not a matter of our policy, it so happened that gastrostomy was not employed with the reparative approach in this series. Table 2. Spontaneous and Instrumental Endoscopic Perforations Treated Conservatively a Type of Injury G CT NG S A Spontaneous 3 3 3 3 (N = 3) Instrumental 2 3 3 5 (N = 6) "Excludes caustic perforation treated conservatively. There were no deaths in this group. G = gastrostomy; CT = chest tube; NG = nasogastric tube; S = steroids; A = antibiotics, 7 Results For the spontaneous and instrumental perforations of the lower third of the esophagus, 5 patients had repair by direct suture, with 5 deaths (Table 3). Both mortality and dehiscence of the repair are thought to be related to delays in treatment, and our experience is consistent with Table 3. Outcome According to Time of Operation in 5 Patients with Spontaneous or lnstrumental Rupture in Lower Third of Esophagus Treated by Thoracotomy and Direct Suture Time between Rupture and Operation (hr) Reopened Died 4 7 2 5 20 2 24 24 30 36 72 20? 3 6 0 6 patients 5 patients (33%)..
3 The Annals of Thoracic Surgery Vol 25 No 4 April 978 Table 4. Results of Treatment for Boerhaave s Syndrome No. of Treatment Patients Lived Died Major surgery 8 5 3 (37%) (direct closure) Minor surgery, e.g., 3 3 thoracostomy, gas trostomy No treatment Wrong operation Total 3 8 5 other reported series in this respect. Five of the 6 dehiscences and 4 of the 5 deaths were in patients operated upon more than 20 hours after perforation occurred. Dehiscence of the repair, always a serious occurrence, was associated with fatality in 3 of 6 instances. Breaking this down another way, there were 3 instances of full-thickness emetic injury, or Boerhaave s syndrome (Table 4). In 8 of these, direct suture was carried out with 3 deaths. A ninth patient was mistakenly subjected to celiotomy for a perforated duodenal ulcer and died about 24 hours later, shortly after the correct diagnosis was made. Another death, following spontaneous rupture, occurred in a young woman with a chronic problem due to head injury who was unable to communicate very well. She was diagnosed very late and received no definitive therapy, either conservative or operative. Three patients were treated with minor procedures only, and all lived. There were 3 instrumental perforations. Seven of these patients had direct suture; 2 of them died. Six were managed conservatively; all 6 recovered. Five of the 3 patients with instrumental perforation showed evidence of mediastinitis and pleuritis similar to full-blown Boerhaave s syndrome. Four of these had thoracotomy and direct repair; 2 lived and 2 died. The fifth patient almost died but survived with gastrostomy and closed thoracostomy. The patient who had perforation from a bullet also had perforations of the stomach. Both the esophageal and gastric perforations were closed by direct suture within a few hours of injury. The patient lived and did well. The caustic perforation which revealed itself as a tracheoesophageal fistula in the acute phase of injury was managed entirely by hyperalimentation and antibiotics. Later, of course, esophageal reconstruction was required. One death occurred following a minor procedure employed after a foreign body perforation. An elderly patient reported to a local emergency room complaining of a bone caught in her gullet. She was sent home without esophagoscopy only to return two weeks later desperately ill with advanced mediastinitis and pyopericardium. Esophagoscopy and removal of the bone, cervical mediastinotomy, and drainage of the pericardium through the fifth costal cartilage were done. The addition of steroids and massive antibiotic therapy was to no avail; the patient died within 36 hours. The instance of rupture due to blunt trauma from an automobile accident was diagnosed 72 hours after injury, shortly before the patient died from advanced mediastinitis. Sadly, multiple rib fractures and pulmonary contusions were the only associated injuries. In the combined group of both spontaneous and instrumental perforations treated conservatively, a total of 9 patients (see Table 2), there were no deaths. The person with caustic perforation of the esophagus was also treated conservatively, for that injury and for 2 separate instances of instrumental perforation of his colon transplant later in the hospitalization. Comment Boerhaave s syndrome is said to have been 00% fatal prior to 947, when Mr. N. R. Barrett reported successful surgical attack on the problem [2]. In that patient the esophageal repair dehisced, convalescence was prolonged, and in all probability survival occurred only because of the drainage provided first by the chest tube and later by the costotransversectomy. Even with the most vigorous surgical approach today, mortality in Boerhaave s syndrome remains in the area of 30 to 40% [l, 8,0,4]. As our experience and that of others shows, instrumental perforation can be equally dangerous if diagnosis is delayed and mediastinitis and pleuritis have a chance to develop [2.
349 Lyons et al: Ruptures and Perforations of the Esophagus In an excellent comprehensive paper on spontaneous perforations, Abbott and associates [ll noted the high rate of repair dehiscence and fatality in patients operated on 2 to 8 hours after perforation. Generally speaking, 24 hours after spontaneous rupture the surgeon is looking at a patient with perhaps a 5O0/o chance of survival. This fact alone should tell us there is something wrong with the routine surgical approach to this group of patients. Surgical indications and treatment in patients with intrathoracic esophageal perforation are fairly well standardized. Loop and Groves [9] summarized well the prevalent bias in favor of surgical management by emphasizing the old axiom the patient is too sick not to operate. For authority, they turned to Barrett: Conservative therapy is likely to fail in that if the emergency is survived, the patient is apt to develop essentially untreatable mediastinal and pleural complications [9. The most elementary logic is invoked: the gut is leaking into the interior of the body; death will result unless the rent is closed. Therefore, thoracotomy and repair are the only solution, even though the esophagus is frequently not in suitable condition for suture or the patient is too ill for an operation of this magnitude. Abbott and co-workers [l] also favored primary surgical repair of the esophagus, although the high rate of repair breakdown (9 of 3) and associated mortality (8 of 3) they reported would seem to prompt a less certain attitude. As architects of the vigorous surgical approach, Loop and Groves [9 recommend that ancillary procedures be carried out as indicated at the time of thoracotomy and repair of the rent. For example, Heller procedures, hiatus hernia repair, even esophagogastrectomy are advocated. In 965 Foster and associates [5] strongly advocated the definitive approach of thoracotomy and direct suture of the perforation. By 970, however, Foster s stance had changed considerably, and he pointed out that drainage of the mediastinum and pleura was the really important thing [4. When a condition such as Boerhaave s syndrome exists, with a mortality risk rising by approximately 2% an hour, such general recommendations represent questionable surgical judgment. There were 3 patients with instrumental perforations in our series in whom ancillary procedures were carried out; 2 died. A hiatus hernia was repaired in and a Heller procedure and hernia repair were carried out in the other. To us it appears that improvement in management is necessary if the death rate in spontaneous and iatrogenic perforations of the lower third of the esophagus is to be reduced. The key to a successful outcome is prompt recognition. There is really no excuse for late diagnosis. Second, all oral intake must be stopped from the moment the diagnosis is suspected. Third, the stomach must be kept empty of acid juices and bile. Fourth, the pleural space, if involved, must be adequately drained, and the lung, if collapsed, promptly reexpanded. Fifth, antibiotics must be given in all cases. And sixth, steroids must be administered in patients with toxicity and evidence of extensive internal soilage. The role of thoracotomy and closure of the perforation is not easy to assess. In patients diagnosed early and showing little or no toxicity, good results will probably follow either minor or major surgical procedures [lo]. Nevertheless, thoracotomy and suture of the esophageal rent may be the procedures of choice. When primary healing occurs, certainly the morbidity is minimized. Presumably, even if the repair reopens, thoracotomy should have afforded the best possible placement of drainage tubes in the pleural space and mediastinum. Mortality rates in patients with Boerhaave s syndrome, for example, operated on before 2 to 8 hours have elapsed, are generally reported in the 0 to 20% range, although admittedly this is still too high. Conversely, direct suture after 24 hours is associated with a high rate of dehiscence and a mortality rate of 50% or more. Numerous reports attest to this [3,5,6,7,2,4, 6. Surely the axiom the patient is too sick not to operate is not the complete answer. The burden of thoracotomy in the desperately ill can only contribute to an unfavorable outcome. In this series, 4 of 0 patients who were operated upon 20 hours or more after their perforation died (see Table 3). On the other hand, consider the reported results on reasonably large groups of patients with instrumental perforation, both early and late,
350 The Annals of Thoracic Surgery Vol 25 No 4 April 978 treated conservatively. Mengoli and Klassen [ll] reported death among 8 patients, 4 of whom had perforations in the lower third of the esophagus. There is a report from Sinai Hospital of Detroit on 24 patients with instrumental perforation of the esophagus treated conservatively without a death [5]. However, the distribution of these patients according to the level of perforation is not given. Of no little significance are the important reports of Nealon [2], Matthewson [lo], Abbott [l], and Samson [3] and their colleagues showing comparable or better results with conservative management, a fact which did not seem to be fully recognized by the authors of these papers. In the experience reported here, among 9 patients with spontaneous and instrumental perforation of the lower third of the esophagus treated conservatively, there were no deaths. Four of these perforations were handled "late," in that they were "toxic" with mediastinitis and pleural effusion. In recent years we have been inclined to treat the sickest patients conservatively, and we think the reported results tend to favor this approach. We believe our data support the notion that if the mortality rate is to be decreased, alternative methods to major operation have to be more frequently employed for the ruptured or perforated intrathoracic esophagus, especially when diagnosis is delayed and the patient is very ill. While we have had no experience with major defunctionalizing procedures in either the acute situation or the chronic fistula stage, it is our belief, based on the results with the conservative approach reported here and by others, that such operations are rarely necessary. Additionally, of course, they have the great disadvantage of requiring major secondary procedures. Ancillary procedures such as the Heller operation and hernia repairs should not be carried out when the patient is acutely ill unless conditions are absolutely ideal. The objective is survival in this dangerous situation; hernias and other problems can be corrected later. References. Abbott OA, Mansour KA, Logan WD, et al: Atraumatic, so-called spontaneous rupture of the esophagus. J Thorac Cardiovasc Surg 59:67, 970 2. Barrett NR: Report of a case of spontaneous perforation of the esophagus successfully treated by operation. Br J Surg 35:26, 947 3. Campbell TC, Andrews JL, Neptune WB: Spontaneous rupture of the esophagus. JAMA 235: 526, 976 4. Foster JH: Esophageal perforation. Modern Treatment 7:284, 970 5. Foster JH, Jolly PC, Sawyers JL, et al: Esophageal perforation. Ann Surg 6:70, 965 6. Janssen CW Jr: Perforation of the intrathoracic esophagus: a report of nineteen cases. Scand J Thorac Cardiovasc Surg 0:89, 976 7. Jemerin E: Results of treatment of perforation of the esophagus. Ann Surg 28:97, 9 8. Lawson RAM, Butchart EG, Soriano A, et al: Spontaneous ruptures of the esophagus. J R Coll Surg 9:363, 974 9. Loop FD, Groves LK: Esophageal perforations. Ann Thorac Surg 0:57, 970 0. Matthewson C Jr, Dozier WE, Hamill JP, et al: Clinical experiences with perforation of the esophagus. Am J Surg 04:256, 962. Mengoli LR, Klassen KP: Conservative management of esophageal perforation. Arch Surg 9:238, 965 2. Nealon TF, Templeton JY, Cuddy VD, et al: Instrumental perforation of the esophagus. J Thorac Cardiovasc Surg 4:75, 96 3. Samson PC: Discussion of Nealon et a [2] 4. Sawyers JL, Lance CE, Foster JH, et al: Esophageal perforation, an increasing challenge. Ann Thorac Surg 9:233, 975 5. Shapiro R: Quoted by R Rappaport in discussion of Nealon et a [2] 6. Wychulis AR, Fontana RS, Payne WS: Noninstrumental perforation of the esophagus. Dis Chest 55:90, 969