Noniatrogenic Esophageal Trauma

Transcription

1 Noniatrogenic Esophageal Trauma Darryl S. Weiman, MD, William A. Walker, MD, Kathleen M. Brosnan, MSN, James W. Pate, MD, and Timothy C. Fabian, MD Department of Surgery, University of Tennessee College of Medicine, Memphis, Tennessee Few guidelines are available with which to facilitate treatment in patients with noniatrogenic injuries of the esophagus. Early diagnosis and proper management are essential if a good outcome is to be expected. In an effort to define better the treatment of patients with penetrating and blunt injuries of the esophagus, we report our recent 5-year experience at an urban trauma center. From July 1988 to June 1993, nineteen patients with esophageal perforations from penetrating (18) and blunt (1) trauma were identified by our trauma registry. There was no mortality in this group of patients and morbidity was mostly due to associated injuries. Eleven cervical esophageal injuries were repaired. One cervical injury was treated by stopping oral intake and giving intravenous antibiotics. The neck was not drained in 10 of the surgical cases. In I patient a tracheoesophageal fistula developed, which later was repaired with a pectoralis muscle flap. Seven perforations were identified in the thoracic (2) and abdominal (5) portions of the esophagus. All were due to gunshot wounds. In 4 cases, a fundal wrap was used to reinforce the repairs. Postoperative contrast studies confirmed that all repairs were intact. We conclude that penetrating and blunt tears of the esophagus can be repaired safely with minimal mortality. Morbidity is usually from associated injuries such as to the spinal cord and trachea. When identified early, cervical esophageal injuries do not need to be drained routinely. (Ann Thorac Surg 1995;59:845-50) E sophageal injuries can present uncommon technical problems. The complexity of the injury increases greatly if there is a concomitant injury of the trachea. If the trachea and esophagus are injured simultaneously, repair of the airway is jeopardized when it becomes infected by the salivary stream coming from an unrecognized esophageal injury or a leaking esophageal repair. Because the consequences of a failed tracheal repair are so devastating, the recognition of a simultaneous esophageal injury must be made and management must be successful. In an effort to better define how to deal with noniatrogenic injuries of the esophagus, we reviewed our most recent 5-year experience at a regional trauma center. Material and Methods Patients and Treatment From July 1988 to June 1993, nineteen patients with noniatrogenic esophageal injury were identified by the registry at the Elvis Presley Trauma Center at the University of Tennessee. Medical records were reviewed to assess mechanism of injury, types of repair, associated injuries, postoperative complications, and management. Eighteen of these injuries were from gunshot wounds and one was from blunt trauma when the neck struck a steering wheel in an automobile accident. All injuries from gunshot wounds were repaired urgently. The one blunt injury was managed by stopping oral intake and Presented at the Forty-first Annual Meeting of the Southern Thoracic Surgical Association, Marco Island, FL, Nov 10-12, Address reprint requests to Dr Weiman, Division of Cardiothoracic Surgery, Department of Surgery, University of Tennessee, 956 Court Ave, Suite G212, Memphis, TN administering antibiotics. All patients had nasogastric tubes placed. Ages ranged from 17 years to 55 years with a median of 22 years. There were 15 male and 4 female patients. Nine patients had combined tracheal and esophageal injury. All of these combined injuries were from gunshot wounds and all were repaired. Three of these patients with combined injury were explored through transverse cervical incisions (collar incisions) and 6 were explored through oblique cervical incisions (one bilateral). Tracheal repairs were done in a single layer with polypropylene or polyglycolic acid sutures. Esophageal repairs were done in two layers with polyglycolic acid on the mucosal layer and silk suture on the muscular layers. When feasible, viable muscle was placed between the two repairs (Fig 1). Drains were not used except in I case where the hypopharynx was injured. Five patients had tracheostomies. In 1 patient a tracheoesophageal fistula that developed later was repaired with a pectoralis muscle flap. Three cervical injuries did not involve the trachea. One of these patients (with blunt trauma) was treated nonoperatively with antibiotics and cessation of oral intake. The other 2 patients sustained gunshot wounds of the cervical esophagus, which were successfully closed in two layers. Seven esophageal perforations were in the thoracic (2) and abdominal (5) portions of the esophagus. These injuries were repaired and four were reinforced with a fundal wrap. Postoperative dye studies confirmed that all repairs were intact. Associated injuries were to the spinal cord (3), lung (7), thyroid gland (1), thumb (1), shoulder (1), brachial artery (1), larynx (1), stomach (4), diaphragm (4), spleen (2), liver 1995 by The Society of Thoracic Surgeons /95/$ (95)

2 846 WEIMAN ET AL Ann Thorac Surg ESOPHAGEAL TRAUMA 1995;59: i I ~i~f ~,4' f C D Fig 1. Repair of a combined tracheal and esophageal injury. Stenocleidomastoid muscle is placed between the repairs. (4), arm (2), aorta (1), ankle (1), small bowel (2), colon (1), kidney (1), carotid artery (1), and thoracic duct (1). Representative Case Histories TRACHEAL AND ESOPHAGEAL INJURY. There were 9 cases in which both the trachea and esophagus were injured (Table 1). Patient 1 was a 55-year-old man. He was shot in the neck and sustained an injury to the trachea, glottis, and esophagus. The esophagus was repaired in two layers and a tracheostomy was done. One month later, an esophageal stricture was successfully dilated. This patient was hospitalized for 26 days. Patient 5, a 19-year-old woman, had a gunshot wound of the esophagus and trachea. The trachea was repaired and a tracheostomy was done. The esophagus was repaired in two layers and a strap muscle was placed between the repairs. Two weeks later, a tracheoesophageal fistula measuring 3 cm in length was repaired by placing the sternal head of the sternocleidomastoid muscle between the trachea and esophagus. The fistula recurred. A tracheostomy was placed distal to the fistula and a mushroom catheter esophagostomy was placed. Two months later, a pectoralis major muscle flap with a split-thickness skin graft was used to close the fistula. This patient was paralyzed at the T2 level (from the gunshot wound). She finally was discharged to a nursing care facility 5 months after the initial injury. Patient 8, a 35-year-old woman, had a gunshot wound that injured the larynx and hypopharynx. A tracheostomy was placed and the larynx was repaired and stented. The hypopharynx was repaired in two layers and a drain was placed. A barium swallow showed no leak, but there was a delay in triggering a swallow. This patient

3 Ann Thorac Surg WEIMAN ET AL ;59: ESOPHAGEAL TRAUMA Table I. Nine Cases of Combined Tracheal and Esophageal Trauma Patient Age Hospital No. Sex (y) Other Injuries Complications Days 1 M 55 Destroyed Esophageal 26 glottis stricture 2 M Pneumonia 9 3 F 19 Spinal cord C5, Paralysis 34 hypopharynx 4 M Chronic pain 8 syndrome 5 F 19 Spinal cord T2 Paralysis, TE 137 fistula, ARDS, pressure sore 6 M 32 Hemothorax, thumb, thyroid 11 7 M 22 Spinal cord C F 35 Shoulder, larynx Paralysis, pneumonia, pressure sore Delay in triggering swallow 9 M 18 Brachial artery, Paralysis, 71 brachial pneumonia, plexus, sacral spinal cord decubitus T2, right arm ARDS - adult respiratory distress syndrome; TE 12 tracheoesophageal. learned to do a supraglottic swallow and was discharged on the twelfth postoperative day. ESOPHAGEAL NECK INJURY. Three patients had esophageal wounds in the neck that did not involve the trachea (Table 2). Patient 10, a 29-year-old woman, perforated her hypopharyngeal esophagus after striking her neck on the steering wheel during an automobile accident. She then was choked by the other driver. She presented with neck pain, dysphagia, and hoarseness. A barium swallow revealed the hypopharyngeal perforation. She was given intravenous fluids and antibiotics and was not allowed to take anything by mouth. Her injury healed, and she was discharged 9 days later when a repeat barium swallow was normal. Patient 11, a 17-year-old boy, was shot in the neck. Barium and Gastrografin (Squibb Diagnostic, Princeton, NJ) did not show a leak, but during exploration, bubbling was seen coming from the esophagus as air was injected Table 2. Three Cervical Esophageal Injuries Without Tracheal Involvement Patient Age Hospital No. Sex (y) Other Injuries Complications Days 10 F None 9 11 M 17 Left arm, colon, None 10 small bowel 12 M 32 Carotid artery, None 7 thoracic duct Table 3. Injuries of the Thoracic Esophagus Patient Age Hospital No. Sex (y) Other Injuries Complications Days 13 M 18 Aortic arch, Esophageal 15 ankle fracture leak 14 M Pneumonia 10 down the nasogastric tube. The defect was found and closed. A barium swallow done a week later was normal, and he was discharged on the tenth postoperative day. THORACIC ESOPHAGEAL INJURY. Two patients had esophageal wounds in the thoracic portion of the esophagus (Table 3). Patient 14, an 18-year-old man, was shot in the right side of the chest. The bullet passed through the mediastinum. An exploratory laparotomy was negative but methylene blue placed in the nasogastric tube was seen coming through the esophageal hiatus. The esophageal injury was found during a left thoracotomy. The esophagus was repaired in two layers and a pleural wrap was placed around the repair. An esophagram obtained on the ninth postoperative day showed the repair to be intact. ABDOMINAL ESOPHAGEAL INJURY. Five patients had injuries to the abdominal portion of the esophagus (Table 4). Patient 15, a 19-year-old man, was shot in the chest and abdomen. A through and through injury of the esophagus just proximal to the gastroesophageal junction was repaired in two layers. The fundus was used to wrap the repair. On the sixth postoperative day, a Gastrografin study showed the repair to be intact. The patient was discharged on the ninth postoperative day. Patient 18, a 31-year-old man shot in the epigastrium, had a through and through esophageal injury repaired in two layers. The fundus was mobilized and wrapped around the repair over a 50F bougie. Injuries to the liver and diaphragm also were repaired. Contrast studies Table 4. Injuries to the Abdominal Portion of the Esophagus Patient Age No. Sex (y) Other injuries Complications Days 15 M 19 Diaphragm, ulna, Dysphagia 10 stomach, pneumothorax, spleen 16 M 17 Diaphragm, Pleural effusion 14 stomach, liver, arm, hemothorax 17 M 26 Stomach, spleen, Pneumonia, bilateral 11 small bowel pleural effusions 18 M 31 Liver, diaphragm, Wound infection 12 pneurnothorax 19 M 35 Diaphragm, liver, Pneumonia 12 kidney, stomach, pneumothorax

4 848 WEIMAN ET AL Ann Thorac Surg ESOPHAGEAL TRAUMA 1995;59: (Gastrografin and barium) showed a narrowing at the gastroesophageal junction. A wound infection developed and this patient was treated as an outpatient. He went home on the twelfth hospital day. Results The 9 patients who had both tracheal and esophageal injuries had unusual problems because of the airway involvement. All 9 had an obvious airway injury at their initial physical examination. All 9 were explored and repaired on an emergent basis. All tracheal injuries were repaired with an interrupted single layer of either polypropylene or polyglycolic acid sutures. Five of these patients had a tracheostomy placed. In this group, all of the esophageal injuries were repaired in two layers. A tracheoesophageal fistula developed in 1 patient. The other patients healed with no leaks. Five of the patients with combined tracheal and esophageal injury had a muscle flap of either sternohyoid, sternothyroid, or sternocleidomastoid muscle placed between the repairs. The patient in whom a tracheoesophageal fistula developed had a muscle flap (sternohyoid) placed at the initial operation. This patient eventually was repaired with a pectoralis muscle flap after a previous attempt at closure (with sternocleidomastoid) had failed. In 1 patient with a tracheal and esophageal injury, an esophageal stricture developed, which was dilated successfully. Three patients had injuries to the cervical esophagus that did not involve the trachea. One patient was managed nonoperatively with antibiotics and cessation of oral intake. The other 2 patients in this group had the esophagus repaired in two layers. All of these patients did well. The repairs of the seven esophageal injuries in the chest and abdomen all healed. Comment Our series looks exclusively at patients with noniatrogenic esophageal injuries. Most reports dealing with esophageal injuries include patients with Boerhaave's syndrome and patients with iatrogenic perforations. These patients often have other medical problems as well as underlying esophageal pathology, and clearly are different from the usual trauma patient who is generally younger and otherwise in good condition. The course of the esophagus puts it in contact with several other organs that also may be injured. Associated injuries have a detrimental effect on the final outcome. The Neck Twelve of our patients had injuries to the cervical esophagus and in 9 of these the injury involved the trachea. Respiratory distress, obvious air leak from the cervical perforation, and subcutaneous emphysema are the findings commonly seen from a major tracheal injury. At our institution, neck injuries with violation of the platysma are not always explored. In stable patients being observed, bronchoscopy, contrast study of the esophagus (barium or Gastrografin), and evaluation of the neck vessels are done. There are false negatives with all of the above studies, so all patients must be observed closely and the studies should be repeated or the neck should be explored if the clinical course suggests a missed injury [11. In a combined tracheal and esophageal injury, the injured tracheal tissue should be debrided to viable tissue before the sutures are placed. The blood supply of the trachea is lateral and segmental, so care must be used in mobilizing and debriding so that the remaining trachea is not devascularized, which would result in tracheornalacia, tracheal stenosis, or breakdown of the repair [2]. A suture line leak of the cervical esophagus usually can be managed by opening the neck incision. However, with a tracheal repair, more aggressive treatment of an esophageal leak is warranted to protect the tracheal repair. In this situation, irrigation and repair of the esophagus with placement of a viable muscle flap may salvage the situation. If the esophageal leak is large and puts the tracheal repair in jeopardy, the salivary stream can be diverted with a "spit fistula" or a T tube. Reconstruction of the esophagus can be done after the tracheal repair has healed. The I patient in our series who had a breakdown of the tracheal repair (patient 5) had a large tracheoesophageal fistula that was not recognized until it was 3 cm in length. Findings that act as clues to the development of a tracheoesophageal fistula include (1) increasing volume loss from the respirator, (2) need to increase the volume of the cuff on the endotracheal tube, (3) presence of a nasogastric tube along with a cuffed tube in the trachea, and (4) eructation with each breath from the respirator. The presence of a fistula can be confirmed by bronchoscopy or esophagoscopy. The presence of a tracheoesophageal fistula mandates surgical repair. If an esophageal injury is suspected but not found, a nasogastric tube placed in the cervical esophagus can be used to insufflate air. Water placed in the wound will have bubbles coming through a small perforation. For extensive injuries, a stomach pull-up or a colon interposition or a jejeunal graft have been used to restore gastrointestinal continuity [3]. Other groups have described how difficult a combined injury to the trachea and esophagus can be. Feliciano and associates [1] reported 23 patients with combined tracheal and esophageal injury. Major complications occurred in 17 of these patients (74%) and there were four deaths. Shama and Odell [4] had 3 patients with combined tracheal and esophageal injury. In 1 an empyema developed from an esophageal leak. Thoracic Esophagus Two wounds of the thoracic esophagus were treated by primary closure and drainage. Cohn and colleagues [5] described 23 wounds of the thoracic esophagus, about half of which were iatrogenic.

5 Ann Thorac Surg WEIMAN ET AL ;59: ESOPHAGEAL TRAUMA Five of 10 patients treated with simple drainage died. These patients all had severe associated disease and were not considered to be suitable for an operation. Four of the 13 patients operated on died; all were more than 24 hours from the time of their injury. Noniatrogenic injuries of the thoracic esophagus must be recognized early. The wound edges can be debrided and the esophagus repaired in one or two layers. If possible, the repair should be reinforced with pleura, viable muscle, or serosa from a fundal wrap [6-16]. Nonoperative management can be done for some esophageal perforations [17], but this may lead to complications because it may be difficult to distinguish a contained injury from one that can cause extensive mediastinitis. Abdominal Esophagus Because most penetrating injuries of the abdomen are explored early, prompt recognition of an esophageal injury is the usual scenario. This allows for early repair and reinforcement with a fundal wrap. All 5 of our patients with this injury did well with this management. Conclusions Injuries to the esophagus should be repaired as soon as possible. Injuries of the thoracic and abdominal esophagus usually can be debrided and closed primarily. Placement of a pleural patch, a muscle flap, or a serosal patch can reinforce the repair. Patients who sustain a tracheal injury along with their esophageal wound are at risk for the development of major complications such as a tracheoesophageal fistula, breakdown of the tracheal or esophageal repairs, pneumonia, and mediastinal abscess. The risk of these complications can be lowered by early recognition of the injuries and appropriate repair of the trachea and esophagus with placement of a viable muscle between the repairs. Drainage is not always necessary. With esophageal injuries, associated injuries seem to increase the incidence of complications. These patients should be observed closely for signs and symptoms that would allow early intervention to minimize the damage of these complications. References 1. Feliciano DV, Bitondo CG, Mattox KL, et al. Combined tracheoesophageal injuries. Am J Surg 1985;150: Grillo HC, Zannini P, Michelassi F. Complications of tracheal reconstruction. J Thorac Cardiovasc Surg 1986;91: Orringer MB, Stirling MC. Esophagectomy for esophageal disruption. Ann Thorac Surg 1990;49: Shama DM, Odell J. Penetrating neck trauma with tracheal and esophageal injuries. Br J Surg 1984;71: Cohn HE, Hubbard A, Patton G. Management of esophageal injuries. Ann Thorac Surg 1989;48: Pate JW, Walker WA, Cole Jr FH. Spontaneous rupture of the esophagus: a 30-year experience. Ann Thorac Surg 1989;47: Nesbitt JC, Sawyers JL. Surgical management of esophageal perforation. Am Surg 1987;53: Skinner DB, Little AG, DeMeester TR. Management of esophageal perforation. Am J Surg 1980;139: Symbas PN, Hatcher Jr CR, Boehm GAW. Acute penetrating tracheal trauma. Ann Thorac Surg 1976;22: Campbell FC, Robbs JV. Penetrating injuries of the neck: a prospective study of 108 patients. Br J Surg 1980;67: Cheadle W, Richardson JD. Options in management of trauma to the esophagus. Surg Gynecol Obstet 1982;155: Popovsky J, Lee YC, Berk JL. Gunshot wounds of the esophagus. J Thorac Cardiovasc Surg 1976;72: Sheely II CH, Mattox KL, Beall AC, et al. Penetrating wounds of the cervical esophagus. Am J Surg 1975;130: Pate JW. Tracheobronchial and esophageal injuries. Surg Clin North Am 1989;69: Defore Jr WW, Mattox KL, Hansen HA, et al. Surgical management of penetrating injuries of the esophagus. Am J Surg 1977;134: Symbas PN, Tyras DH, Hatcher CR Jr, Perry B. Penetrating wounds of the esophagus. Ann Thorac Surg 1972;13: Cameron JL, Kieffer RF, Hendrix TR, et al. Selective nonoperative management of contained introthoracic esophageal disruptions. Ann Thorac Surg 1979;27: DISCUSSION DR SAFUH ATTAR (Baltimore, MD): In 1989 we presented to this Association in Scottsdale, Arizona, our experience with esophageal perforations. There were about 64. The second largest group after the iatrogenic perforations was 20 patients who had injury, 11 due to penetrating wounds, 10 with gunshot wounds, and I with a stab wound. In 3, the perforation resulted from blunt trauma; a motorcycle ran over the neck of a young man and produced a combined injury, and a child fell off a bicycle and perforated his abdominal esophagus. Since then we have seen two blunt ruptures of the esophagus; actually they were combined esophageal and tracheobronchial ruptures. The incidence of blunt tracheobronchial ruptures at the Shock Trauma Center was 11 per 17,000; the incidence of blunt esophageal rupture was 3 per 17,000, and the incidence of combined blunt, traumatic tracheoesophageal injury was 2 per 17,000. You had, I understand, 1 case of blunt rupture of the esophagus. My question is, how do you diagnose at the initial episode of trauma whether an esophageal injury is present or not? Most of these cases are missed; however, of the last 2 cases in the shock trauma, 1 was due to a vehicular accident and the second was in a young man, 16 years old, involved in a rapid motorcycle accident. The diagnosis was made accidentally in I of them. He had a nasogastric tube, and we noted that there was air coming from the nasogastric tube while he was on the ventilator. The nasogastric tube was put under water and one could see the bubbling, and this is how the diagnosis was made. So I would like to ask you, how can one suspect the presence of a combined injury and how do you treat it? DR WEIMAN: At the University of Tennessee we generally do not demand an exploration for neck trauma, but if the airway seems to be involved from some of the signs and symptoms that you mentioned, those patients will be explored. In patients with obvious respiratory distress, air coming out of the wound, or subcutaneous emphysema, even if they are going to be observed, the airway will be controlled and other studies will be done to evaluate for esophageal injury. These will include a barium study or Gastrografin; we prefer dilute barium. If during

6 850 WEIMAN ET AL Ann Thorac Surg ESOPHAGEAL TRAUMA 1995;59: the exploration for the airway injury you suspect an esophageal injury but you do not identify it, then what you mentioned is a really nice, slick way to identify it. You put a little water in the wound, pull your nasogastric tube back, and inject some air, and you will see bubbles when air is injected. There was 1 patient in our series in whom the diagnosis was made during the exploration. Even though the barium study was negative, we did see air coming from the esophagus, a small perforation was found, and that was repaired. So this trick is important to keep in mind. Even if you are observing the neck injuries and you do suspect--based on the physical examination and the history of the patient--that the patient has had an injury and your studies are negative, then either those studies need to be repeated or that patient needs to be explored. A high index of suspicion should not be ignored, because the results of missing an injury can be devastating. DR ATTAR: What about esophagoscopy? DR WEIMAN: Some people believe that an esophageal perforation is a relative contraindication for esophagoscopy. Also, if the injury is small, esophagoscopy will not pick it up, and I think barium probably still is the best way of diagnosing it if you are not exploring that patient. DR KAMAL A. MANSOUR (Atlanta, GA): While you are on that subject, how did you miss the perforation in the thoracic esophagus when you opened the abdomen and found nothing? Did you obtain any contrast studies before exploration? DR WEIMAN: Before the operation, that patient did not have a contrast study. Because of the mechanism of the injury, we were very, very concerned that there was an abdominal injury. The surgeons who were exploring him did not make the diagnosis until the abdomen was already opened, and they found nothing in the abdomen. DR MILTON GLATTERER, JR (Tampa, FL): I enjoyed the presentation, and certainly I think we all have to applaud your good results. I appreciate the difficulty you have with data collection in this fairly uncommon injury. While I was in training at the University of Texas in San Antonio, I was given the opportunity to review esophageal trauma. It was all traumatic, noniatrogenic injuries. We managed to come up with a few more patients, but it took us three times as long, about 15 years. We had 26 patients--21 with cervical injuries and 5 with thoracic injuries--and all injuries were penetrating. Unfortunately we also had more experience with complications in this group of patients. There were 3 early deaths, and 13 postoperative patients had complications. Four patients had esophageal leaks; in 2 they were healed by stopping oral intake and giving patients the hyperalimentation. One patient, again with cervical injuries, had a leak approximately the 6th postoperative week, which was treated by a diversion and drainage procedure. Unfortunately mediastinitis and sepsis developed, and he died. Our other death from a leak was in a patient who had an early repair of a cervical injury but a missed injury at the cervicothoracic level. This was repaired primarily about 2 weeks postoperatively. This patient also eventually died. Clearly, delayed recognition and repair of an injury leads to complications. All of your repairs fortunately were done early. I think the big question is, how does one decide between a primary repair versus a diversion and drainage procedure, especially in a thoracic patient? My other question is, given that 10% of your nondrained cervical injuries had a complication, and realizing that draining that I patient may not have changed the outcome, why would you not drain the cervical repair in all of your patients? DR WEIMAN: I will answer the second question first. The drain itself may be a problem because of a foreign body near a suture line. Doctors Tyras, Hatcher, and Symbas at Emory University did report a series of esophageal injuries, and they recommended very strongly Penrose drainage of injuries to the neck. Doctor Tyras was also one of my teachers, and I do not want to say that you do not necessarily have to listen to him, but clearly I do not think that you necessarily have to drain neck injuries. The second question on the delayed recognition and what repair you do when you do have that situation is a tough one to answer, and I think each individual thoracic surgeon has to answer that in his or her own way based on his or her experience. Remember, you can always get out of a really tight situation by proximal diversion and use of a gastrostomy tube and go back and reconstruct the mediastinal esophagus at a later time if you have to. My gut inclination is, unless the sutures are not holding or whatever, I will try to repair even with a delayed diagnosis of the esophageal injury and lay a viable muscle flap in there, intercostal or some other form of muscle flap, as my first step, Remember, if that does not work in 24 or 48 hours, you always can go back and exclude that esophagus. DR WATTS R. WEBB (New Orleans, LA): We have a continuing experience in this injury, and I think your emphasis on early operation on all of them is so important. I think this is the major difference in the good results we get in trauma and the poor results we so often see in iatrogenic injuries, where the endoscopists tend to refuse to accept that they have perforated the esophagus or the gastric surgeon to acknowledge that he or she put a finger through the esophagus while trying to dig out the phrenic nerves, or something of that nature. We have found that endoscopy is a better way of making a diagnosis than a contrast study. I think that may be because we still go back to the rigid esophagoscope, because the little flexible esophagoscope just does not do the job for you. We have picked up some injuries with the endoscope that have been missed by the barium swallow. I think as far as the neck drain is concerned, you are quite right. You do not need to drain if you get them early. If it is late and they have a phlegmon or infection or mediastinitis, then absolutely you have to, and not with a Penrose drain but with a sump drain. I also agree with your handling of the trachea early. Put in a tracheostomy, if they need it. If it is a massive injury and you need to reconstruct the trachea, do that later. That should not be done at that moment. DR WEIMAN: I want to emphasize Dr Webb's statement; you would be amazed at how many "esophageal diverticula" result after some sort of endoscopic procedure of the esophagus that we get called on to evaluate several hours later. DR WILLIAM A. COOK (North Andover, MA): In both the traumatic and iatrogenic cases of perforation, if the patient has had a flexible esophageal examination, there are two things that are very important: a preliminary film and a film taken afterward. If you take the films like that and look at them carefully, the very tiniest of perforations will allow air into the mediastinum. I have had several cases that were too small to be demonstrated with a swallow but did have air in the mediastihum after the perforation. So it can be helpful if you have those two films.