- Email: [email protected]
Esophagogastric anastomoses: The value of fibrin glue in preventing leakage
J THORAC CARDIOVASC SURG 1987;93:234-9
Esophagogastric anastomoses: The value of fibrin glue in preventing leakage Disruption of an esophagogastric anastomosis can result in a high mortality despite aggressive treatment. The efficacy of fibrin "glue" to seal esophagogastric anastomoses was evaluated as a means of preventing this complication. A left thoracotomy was performed in 25 adult mongrel dogs. After esophagogastric resection, a standardized esophagogastrostomy was performed and eight interrupted sutures were used to completely close the posterior waD. The anterior waD was approximated with only three sutures, leaving four large holes between sutures. The dogs were then randomized into the control group (n = 14; no attempt to seal the leaks) or into the fibrin glue-treated group (n = 11). An average of 3.3 ml of glue was applied to the anterior waD of the anastomosis in the treated group. In the control group, 13 of 14 dogs (92.9 %) died of anastomotic leak a median of 3 days after operation. In the fibrin glue-treated group, only four of 11 dogs (36.4%) died of anastomotic leaks (p < 0.01). Dogs that survived were put to death at 14 days. Postmortem examination in aU dogs revealed no deleterious effects or complications related to the glue. Postmortem examination of the one surviving control dog and the seven fibrin glue-treated dogs that did not die of sepsis revealed a healed anastomosis without abscess formation. We conclude that fibrin glue is effective in lessening the incidence of esophagogastric anastomotic leaks as employed in this experimental model.
Patrick M. McCarthy, M.D., Victor F. Trastek, M.D., Hartzell V. Schaff, M.D., Louis H. Weiland, M.D., Philip E. Bematz, M.D., W. Spencer Payne, M.D., and Peter C. Pairolero, M.D., Rochester, Minn.
Intrathoracic leak after esophagogastric anastomosis is a disastrous complication associated with a high mortality.!? "Leakage is perhaps the most serious sequela of esophagotomy.... The first prerequisite, however, in the prevention of this complication is adequate and precise watertight layer closure of esophageal incisions.... "6 Fibrin glue, a mixture of fibrinogen and thrombin, has been used extensively as a hemostatic agent, and excellent results have been achieved.?" However, studies evaluating its effectiveness as an adhesive to seal gastrointestinal anastomoses":" are less common, and the results are not as uniform. The purpose of this study was to test the efficacy of fibrin glue to seal esophagogastric anastomoses in dogs.
From the Section of Thoracic and Cardiovascular Surgery and the Section of Surgical Pathology, Mayo Clinic and Mayo Foundation, Rochester, Minn. Read at the Twelfth Annual Meeting of The Western Thoracic Surgical Association, Napa, Calif., June 25-29, 1986. Address for reprints: V. F. Trastek, M.D., Mayo Clinic, 200 First St. SW, Rochester, Minn. 55905.
234
Materials and methods Twenty-five fasting adult mongrel dogs, weighing between 16 and 28 kg, were anesthetized with pentobarbital (0.5 mljkg). Gentamicin (80 mg intravenously), clindamycin (300 mg intravenously), and benzathine pencillin G (1.2 million units intramuscularly) were given before operation. Each dog was endotracheally intubated and ventilated with a positive-pressure ventilator. Two liters of crystalloid solution were given intravenously during the operation. Under sterile conditions, a left thoracotomy was performed and the chest was entered through the seventh intercostal space. The inferior pulmonary ligament was divided, and the distal esophagus was mobilized up to the level of the inferior pulmonary vein. The diaphragm was divided at the esophageal hiatus, and the proximal stomach was delivered into the chest. The short gastric and the left gastric arteries were divided and ligated. The proximal stomach was divided using a TA-90 stapling device (AutoSuture Company, Norwalk, Conn.). A 3-0 silk suture was used to oversew the staple line. The esophagus was divided 8 em above the gastroesophageal junction. An esophago-
Volume 93 Number 2 February 1987
Esophagogastric anastomoses
Fig. 1. Technique of esophagogastric anastomosis. Posterior row completely closed, with four holes anteriorly.
gastrostomy was fashioned end to side to the anterior portion of the stomach. The anastomosis in each dog was standardized. The posterior wall of the anastomosis was closed with eight 3-0 polypropylene (Prolene), interrupted, full-thickness sutures. Only three single sutures were used to close the anterior wall. This preparation left four large holes in the anterior half of the anastomosis (Fig. 1). The stomach was then secured to the diaphragm with running suture placed several centimeters below the anastomosis. After completion of the anastomosis, the dogs were randomized into two groups. The control group consisted of dogs whose chests were irrigated with saline; after placement of a chest tube, the chest was closed in layers. In the experimental group, 3 to 4 m1 (average 3.3 rnI) of fibrin glue (Tisseel; Immuno AG; Vienna, Austria) were used to seal the anterior wall of the anastomosis (Fig. 2). The glue had been prewarmed and consisted of two components: thrombin (500 IUjmI) with calcium chloride (40 mmoljL) and fibrinogen (Sealer Protein Concentrate, Human, approximately 100 mg/rnl) with fibrinolysis inhibitor (Aprotinin 3,000 kIDjrnl). After the glue was allowed to harden for several minutes, the chest was irrigated and closed with a chest tube in place, similar to that done in the control dogs. After the operation, the control and experimental dogs were treated the same; water was given ad libitum, and standard feedings were permitted the following
235
Fig. 2. Fibrin glue applied to anastomosis.
morning. Clinical condition and leukocyte count were determined daily. All dogs received humane care in compliance with "Principles of Laboratory Animal Care," formulated by the National Society for Medical Research, and the "Guide for the Care and Use of Laboratory Animals," prepared by the National Academy of Sciences and published by the National Institutes of Health (NIH publication No. 80-23). Necropsy was performed on all dogs after death or sacrifice. Surviving dogs were put to death with sodium pentobarbital/isopropyl alcohol/propylene glycol (Sleepaway) (0.5 nil/kg) at 14 days, with the exception of one glue-treated dog, which was maintained for 28 days. At autopsy, the area of the anastomosis and mediastinum were evaluated grossly. If a leak was not evident, water was infused into the esophageal lumen to observe for leakage. Findings that confirmed anastomotic leak included either the extravasation of water, food particles, or gastric juices in the pleural cavity or the finding or purulent mediastinal secretions. Histologic sections of the anastomosis were prepared with hematoxylin and eosin stain. All glue-treated anastomoses were evaluated for inflammatory reaction and resorption of the fibrin glue. Statistical analysis was performed with the two-sample t test and Fisher's exact test. Results Leak of the esophageal anastomosis was demonstrated in 13 of the 14 control dogs (92.9%), and all 13 dogs eventually died of the leak (median, day 3; range, days 3
236 McCarthy et al.
The Journal of Thoracic and Cardiovascular Surgery
Fig. 3. Lung adherent over anastomosis at autopsy. 1, Esophagus. 2, Left lower lobe of lung. 3. Fibrin glue at anastomosis, 4, Stomach,
Fig. 4. A, Fibrin glue at anastomosis on day 14.1, Fibrin glue. 2, Lung. 3, Gastric wall. (Hematoxylin and eosin; original magnification X4.) B, Anastomosis healed at day 28. Fibrin glue resorbed. 1, Gastric mucosa. 2, Anastomosis. 3, Esophageal mucosa. (Hematoxylin and eosin; original magnification X4.)
to 13). In the one surviving control dog, the lower lobe of the left lung was firmly adherent to the anastomosis. There was no evidence of abscess formation or pleural contamination. It was presumed that survival was due to the immediate sealing of the perforation by the lung. Of the 11 glue-treated dogs, only four (36.4%) had an
anastomotic leak (p < 0,01). The four dogs with leak died early (median, day 5; range, days 3 to 8). Cause of leak was failure of the glue to adhere to tissue in three dogs and a perforation through the glue in one dog. Three additional dogs died of unrelated causes before the 14 day planned death. Cause of death in two dogs
Volume 93 Number 2 February 1987
(days 8 and 12) was infarcted small bowel after the bowel had herniated through the diaphragmatic defect. One other dog died (day 3) of massive herniation of abdominal contents into the left side of the chest when the gastric anchoring stitch tore, with subsequent respiratory failure. The anastomoses in all three dogs that died of other causes, and in the surviving four gluetreated dogs, were sealed without evidence of leak. In most, the lower lobe of the left lung was densely adherent around the anastomosis (Fig. 3). The postoperative leukocyte counts were lower for the glue-treated dogs (mean, 23,300jmm 3) than for the control animals (mean, 27,700jmm 3) . However, this difference was not statistically significant. Histologicevaluation of the glue-treated dogs demonstrated partial or near-complete resorption of the glue by 14 days. An inflammatory response consisting primarily of histiocytes without phagocytes was present around the glue. There was no giant cell formation. In one dog studied at 28 days, the glue was gone and there was no residual inflammatory response. The anastomosis was well-healed, with good coaptation of gastric and esophageal mucosae (Fig. 4). Discussion
Despite early aggressive medical and surgical management, the mortality associated with the leaking esophageal anastomosis remains high.!:' Recent techniques to secure the anastomoses include the use of the EEA staplers,13-15 "tunneled" esophagogastrostomy,16 omental wrap," and reinforcement of the suture line with adjacent stomach. 17. 18 Similarly, stomach used for esophageal replacement instead of colon or jejunum has provided a reliable conduit. Finally, anastomoses are more frequently being performed in the neck in an effort to prevent pleural contamination. Fibrin glue may provide an additional adjunct to this array of surgical techniques. In our model, fibrin glue was effective in preventing leakage despite being employed in a very difficult model with large anastomotic holes and early resumption of feeding. Unlike early reports describing synthetic glues,19.20 there was no foreign-body reaction to the fibrin glue, a physiologic substance, used in our study. In one dog studied at 28 days, the glue was completely resorbed without local tissue reaction, as others have also reported." Fibrin glue was also entirely nontoxic. Fibrin glue is rapidly becoming a useful tool in the surgical armamentarium. It has been very successful in controlling the bleeding in cardiothoracic operations. 7, 8. 22 Other uses in thoracic surgery include the sealing of tracheal anastomoses," the sealing of chyle
Esophagogastric anastomoses 2 3 7
leaks," and the sealing of pulmonary air leaks.2S-29* Other studies have suggested its effectiveness as an adjunct to seal bowel anastomoses.t'? Currently, commercially available fibrin glue (Immuno AG; personal communication) is undergoing evaluation by the Food and Drug Administration for possible release in the United States. In conclusion, fibrin glue was effective in decreasing the leakage from an experimental esophagogastric anastomosis. Healing was not complicated by foreign-body reaction. Fibrin glue providesan adhesive substance that may be useful clinically as an adjunct to seal esophageal anastomoses and prevent postoperative leaks.
'McCarthy PM, Trastek YF, Bell D, et al. The effectiveness offibrin glue sealant in reducing experimental pulmonary airleak. Submitted for publication. We would like to thank Duane M. Ilstrup for statistical assistance and Marilyn R. Oeltjen for technical assistance. REFERENCES Keagy BA, Murray GF, Starek PJK, Battaglini JW, Lores ME, Wilcox BR. Esophagogastrectomy as palliative treatment for esophageal carcinoma: results obtained in the setting of a thoracic surgery residency program. Ann Thorac Surg 1984;38:611-5. 2. Ellis FH Jr, Gibb SP, Watkins E Jr. Esophagogastrectomy: a safe, widely applicable, and expeditious form of palliation for patients with carcinoma of the esophagus and cardia. Ann Surg 1983;198:531-9. 3. Wilson SE, Stone R, Scully M, Ozeran L, Benfield JR. Modern management of anastomotic leakafter esophagogastrectomy. Am J Surg 1982;144:95-9. 4. Molina JE, Lawton BR, Myers WO, Humphrey EW. Esophagogastrectomy for adenocarcinoma of the cardia: ten years' experience and current approach. Ann Surg 1982; 195: 146-51. 5. Maillet P, Baulieux J, Boulez J, Benhaim R. Carcinoma of the thoracic esophagus: results of one-stage surgery (271 cases). Am J Surg 1982;143:629-34. 6. Payne WS, Leonardi HK, Ellis FH Jr. Complications of esophageal and diaphragmatic surgery. In: HardyJD, ed' Complications in surgery and their management, 4th ed. Philadelphia: WB Saunders, 1981 :330-66. 7. Borst HG, Haverich A, Walterbusch G, Maatz W. Fibrin adhesive: an important hemostatic adjunct in cardiovascular operations. J THoRAc CARDIOVASC SURG 1982;84:54852. 8. Borst HG, Haverich A. Clinical use of fibrin adhesive: summary. Thorac Cardiovasc Surg 1982;30:241. 9. Petrelli NJ, Cohen H, DeRisi D, Ambrus JL, Williams P, The application of tissue adhesives in small bowel anastomoses. J Surg Oneal 1982; 19:59-61. 10. Thorson GK, Perez-Brett R, Lillie DB, et al. The role of the tissue adhesive fibrin seal (FS) in esophageal anastomoses. J Surg Oncol 1983;24:221-3. I.
The Journal of Thoracic and Cardiovascular
2 3 8 McCarthy et al.
11. Oka H, Harrison RC, Burhenne HJ. Effect of a biologic glue on the leakage rate of experimental rectal anastomoses. Am J Surg 1982;143:561-4. 12. Harrison RC, Oka H. Rectal anastomosis: sutures vs staples and glue. Contemp Surg 1982;21:17-26. 13. Hopkins RA, Alexander JC, Postlethwait RW. Stapled esophagogastric anastomosis. Am J Surg 1984;147:2837. 14. Postlethwait RW. Complications and deaths after operations for esophageal carcinoma. J THoRAc CARDIOVASC SURG 1983;85:827-31. 15. Fekete F, Breil P, Ronsse BH, Tossen JC, Langonnet F. EEA stapler and omental graft in esophagogastrectomy: experience with 30 intrathoracic anastomoses for cancer. Ann Surg 1981;193:825-30. 16. Liu K, Zhang GC, Cai ZJ. Avoiding anastomotic leakage following esophagogastrostomy. J THoRAc CARDIOVASC SURG 1983;86:142-5. 17. Hoffmann TH, Kelley JR, Grover FL, Trinkle JK. Carcinoma of the esophagus: an aggressive one-stage palliative approach. J THoRAc CARDIOVASC SURG 1981;81:44-8. 18. Piccone VA, LeVeen HH, Ahmed N, Grosberg S. Reappraisal of esophagogastrectomy for esophageal malignancy. Am J Surg 1979;137:32-7. 19. Raekallio J, Seligman AM. Acute reaction to arterial adhesive in healing skin wounds. J Surg Res 1964;4:124-7. 20. Woodward SC, Herrmann JB, Leonard F. Histotoxicity of cyanoacrylate tissue adhesives (Abstract). Fed Proc 1964;23:495. 21. Redl H, Schlag G, Stanek G, HirschI A, Seelich T. In vitro properties of mixtures of fibrin seal and antibiotics. Biomaterials 1983;4:29-32. 22. Jakob H, Campbell CD, Qiu ZK, Pick R, Replogle RL. Evaluation of fibrin sealing for cardiovascular surgery. Circulation 1984;70(Pt 2):1138-46. 23. Kram HB, Shoemaker WC, Hino ST, Chiang HS, Harley DP, Fleming A W. Tracheal repair with fibrin glue. J THORAC CARDIOVASC SURG 1985;90:771-5. 24. Stenzl W, Rigler B, Tscheliessnigg KH, Beitzke A, Metzler H. Treatment of postsurgical chylothorax with fibrin glue. Thorac Cardiovasc Surg 1983;31:35-6. 25. DeRisi 0, Cohen H, Takita H, et al. Hemostasis in experimental pulmonary injury. J Surg Oncol 1982;19:208-1O. 26. J~rgensen A, M~ller IW, Johnsen A, Borgeskov S. Bronchopleural leakage treated with fibrin sealant and high-frequency positive-pressure ventilation. Scand J Thorae Cardiovasc Surg 1984;18:139-40. 27. Thetter O. Fibrin adhesive and its application in thoracic surgery. Thorac Cardiovasc Surg 1981;29:290-2. 28. Jessen C, Sharma P. Use of fibrin glue in thoracic surgery. Ann Thorac Surg 1985;39:521-4. 29. Tiirk R, Weidringer JW, Hartel W, Bliimel G. Closure of lung leaks by fibrin gluing: experimental investigations and clinical experience. Thorac Cardiovasc Surg 1983;31:185-6.
Surgery
Discussion DR. RICHARD M. PETERS San Diego, Calif
What was the reason for the failure of the glue to adhere in the animals in which a leak developed? DR. McCARTHY It is hard to determine at autopsy, but approximately 90% of the glue was adherent, and there were just small areas where the glue seemed to have lifted up from where it had been applied to the stomach, with leakage through those small areas. High intraluminal pressure may have pushed the glue off, or gastric enzymes may have broken down the fibrin matrix. DR. DONALD L. MORTON Los Angeles, Calif
This study clearly indicates the value of fibrin glue in sealing an inadequately performed anastomosis. The experimental model developed by Dr. McCarthy and his associates is an excellent one; 13 of 14 anastomoses in the control group leaked whereas only four of 11 anastomoses sealed with fibrin glue leaked-a highly significant difference and a clear demonstration of the effectiveness of fibrin glue in sealing the anastomosis. Most experience with fibrin glue has been in cardiac and vascular surgery, where it has been highly effective in controlling bleeding as a hemostatic. There has been little prior experience with it in gastrointestinal operations, where it might have been thought likely to be degraded by intestinal enzymes; this might be one of the explanations for the leaks that you did see. However, it would appear to have a definite advantage over other tissue adhesives, since it is a natural product and is completely absorbed by the body, unlike the tissue adhesive that we have used for bronchial anastomosis, which is not absorbed. What is the clinical significance of these observations? I must admit to some of the feelings of nostalgia expressed by Benson Roe regarding the passing of craftsmanship in surgery. My view has always been that the tragedy of anastomotic leak is so great that I teach my residents that if the anastomosis does not go well or look right, it should be taken down and redone. Furthermore, I test the anastomosis during the operation by instilling a methylene blue solution into the esophageal lumen to distend the anastomosis, allowing any leaks to be detected by the appearance of the blue dye. If these precepts are used, leakage of the esophagogastric anastomosis should be a phenomenon occurring rarely or not at all, as reported in the beautiful series by Dr. Mitchell, which reflects our own experience. Some of the major causes of anastomotic leaks are (1) ischemic necrosis of the anastomotic site because of inadequate blood supply to the esophagus or stomach, (2) excessive tension at the suture line, (3) sutures placed so tightly that they cut through the anastomosis, or (4) poorly placed sutures creating a gap in the suture line, as in this experimental model. It would clearly appear that the last two might be benefited by fibrin glue. However, I doubt if glue
Volume 93 Number 2 February 1987
will protect against ischemic necrosis or excessive tension. On the other hand, these are rarely a problem with esophagogastric anastomosis. This study clearly indicates that fibrin glue does provide an extra margin of safety to protect against sloppy craftsmanship in surgery and might be recommended for that purpose. DR. McCARTHY Thank you for your comments, Dr. Morton. Your point about precise anastomotic technique is a good one, and fibrin glue is intended to supplement excellent technique, not replace it. We hope that as a "backup" it will not be needed very often. DR. ROBERT W. JAMPLIS Palo Alto, Calif
This is very important paper. All of us know that an acute mediastinitis develops as a result of an esophageal leak; it is a death warrant diagnosis very often. I do not think all of us are as good surgeons as Dr. Morton or Dr. Mitchell, who reported yesterday on 40 consecutive esophageal resections without a leak. Even those of us who use staplers are going to get leaks. Therefore, I think it is very important, not so much that two of the three of the authors' dogs survived even with large holes,
Esophagogastric anastomoses
239
because that happens occasionally, but because there were no complications and no deleterious effects. Since I always cover the bronchus with pleura or whatever, you might say I am a big belt and suspenders man. Therefore, if there are no complications, I would want to use it. I would like to ask whether you are doing it in the operating room now, and if not, why not? DR. McCARTHY Right now, the fibrin glue that we used, the commercial product from Austria, is not available in the United States. It is currently undergoing phase 2 clinical trials in a multicenter study for bleeding with reoperations at cardiac surgery. The risk with using the commercial fibrin glue is that it theoretically can transmit hepatitis and acquired immune deficiency syndrome. However, the company reports that in one million applications in Europe, they have yet to have a reported causal association of hepatitis or AIDS. They go to extensive measures trying to screen the donors, and they also heat treat the product before it is released, to kill the virus. The Food and Drug Administration clinical trials will be completed soon and the commercial product may be made available in the United States, as it was in Canada a few months ago.
Esophagogastric anastomoses: The value of fibrin glue in preventing leakage Disruption of an esophagogastric anastomosis can result in a high mortality despite aggressive treatment. The efficacy of fibrin "glue" to seal esophagogastric anastomoses was evaluated as a means of preventing this complication. A left thoracotomy was performed in 25 adult mongrel dogs. After esophagogastric resection, a standardized esophagogastrostomy was performed and eight interrupted sutures were used to completely close the posterior waD. The anterior waD was approximated with only three sutures, leaving four large holes between sutures. The dogs were then randomized into the control group (n = 14; no attempt to seal the leaks) or into the fibrin glue-treated group (n = 11). An average of 3.3 ml of glue was applied to the anterior waD of the anastomosis in the treated group. In the control group, 13 of 14 dogs (92.9 %) died of anastomotic leak a median of 3 days after operation. In the fibrin glue-treated group, only four of 11 dogs (36.4%) died of anastomotic leaks (p < 0.01). Dogs that survived were put to death at 14 days. Postmortem examination in aU dogs revealed no deleterious effects or complications related to the glue. Postmortem examination of the one surviving control dog and the seven fibrin glue-treated dogs that did not die of sepsis revealed a healed anastomosis without abscess formation. We conclude that fibrin glue is effective in lessening the incidence of esophagogastric anastomotic leaks as employed in this experimental model.
Patrick M. McCarthy, M.D., Victor F. Trastek, M.D., Hartzell V. Schaff, M.D., Louis H. Weiland, M.D., Philip E. Bematz, M.D., W. Spencer Payne, M.D., and Peter C. Pairolero, M.D., Rochester, Minn.
Intrathoracic leak after esophagogastric anastomosis is a disastrous complication associated with a high mortality.!? "Leakage is perhaps the most serious sequela of esophagotomy.... The first prerequisite, however, in the prevention of this complication is adequate and precise watertight layer closure of esophageal incisions.... "6 Fibrin glue, a mixture of fibrinogen and thrombin, has been used extensively as a hemostatic agent, and excellent results have been achieved.?" However, studies evaluating its effectiveness as an adhesive to seal gastrointestinal anastomoses":" are less common, and the results are not as uniform. The purpose of this study was to test the efficacy of fibrin glue to seal esophagogastric anastomoses in dogs.
From the Section of Thoracic and Cardiovascular Surgery and the Section of Surgical Pathology, Mayo Clinic and Mayo Foundation, Rochester, Minn. Read at the Twelfth Annual Meeting of The Western Thoracic Surgical Association, Napa, Calif., June 25-29, 1986. Address for reprints: V. F. Trastek, M.D., Mayo Clinic, 200 First St. SW, Rochester, Minn. 55905.
234
Materials and methods Twenty-five fasting adult mongrel dogs, weighing between 16 and 28 kg, were anesthetized with pentobarbital (0.5 mljkg). Gentamicin (80 mg intravenously), clindamycin (300 mg intravenously), and benzathine pencillin G (1.2 million units intramuscularly) were given before operation. Each dog was endotracheally intubated and ventilated with a positive-pressure ventilator. Two liters of crystalloid solution were given intravenously during the operation. Under sterile conditions, a left thoracotomy was performed and the chest was entered through the seventh intercostal space. The inferior pulmonary ligament was divided, and the distal esophagus was mobilized up to the level of the inferior pulmonary vein. The diaphragm was divided at the esophageal hiatus, and the proximal stomach was delivered into the chest. The short gastric and the left gastric arteries were divided and ligated. The proximal stomach was divided using a TA-90 stapling device (AutoSuture Company, Norwalk, Conn.). A 3-0 silk suture was used to oversew the staple line. The esophagus was divided 8 em above the gastroesophageal junction. An esophago-
Volume 93 Number 2 February 1987
Esophagogastric anastomoses
Fig. 1. Technique of esophagogastric anastomosis. Posterior row completely closed, with four holes anteriorly.
gastrostomy was fashioned end to side to the anterior portion of the stomach. The anastomosis in each dog was standardized. The posterior wall of the anastomosis was closed with eight 3-0 polypropylene (Prolene), interrupted, full-thickness sutures. Only three single sutures were used to close the anterior wall. This preparation left four large holes in the anterior half of the anastomosis (Fig. 1). The stomach was then secured to the diaphragm with running suture placed several centimeters below the anastomosis. After completion of the anastomosis, the dogs were randomized into two groups. The control group consisted of dogs whose chests were irrigated with saline; after placement of a chest tube, the chest was closed in layers. In the experimental group, 3 to 4 m1 (average 3.3 rnI) of fibrin glue (Tisseel; Immuno AG; Vienna, Austria) were used to seal the anterior wall of the anastomosis (Fig. 2). The glue had been prewarmed and consisted of two components: thrombin (500 IUjmI) with calcium chloride (40 mmoljL) and fibrinogen (Sealer Protein Concentrate, Human, approximately 100 mg/rnl) with fibrinolysis inhibitor (Aprotinin 3,000 kIDjrnl). After the glue was allowed to harden for several minutes, the chest was irrigated and closed with a chest tube in place, similar to that done in the control dogs. After the operation, the control and experimental dogs were treated the same; water was given ad libitum, and standard feedings were permitted the following
235
Fig. 2. Fibrin glue applied to anastomosis.
morning. Clinical condition and leukocyte count were determined daily. All dogs received humane care in compliance with "Principles of Laboratory Animal Care," formulated by the National Society for Medical Research, and the "Guide for the Care and Use of Laboratory Animals," prepared by the National Academy of Sciences and published by the National Institutes of Health (NIH publication No. 80-23). Necropsy was performed on all dogs after death or sacrifice. Surviving dogs were put to death with sodium pentobarbital/isopropyl alcohol/propylene glycol (Sleepaway) (0.5 nil/kg) at 14 days, with the exception of one glue-treated dog, which was maintained for 28 days. At autopsy, the area of the anastomosis and mediastinum were evaluated grossly. If a leak was not evident, water was infused into the esophageal lumen to observe for leakage. Findings that confirmed anastomotic leak included either the extravasation of water, food particles, or gastric juices in the pleural cavity or the finding or purulent mediastinal secretions. Histologic sections of the anastomosis were prepared with hematoxylin and eosin stain. All glue-treated anastomoses were evaluated for inflammatory reaction and resorption of the fibrin glue. Statistical analysis was performed with the two-sample t test and Fisher's exact test. Results Leak of the esophageal anastomosis was demonstrated in 13 of the 14 control dogs (92.9%), and all 13 dogs eventually died of the leak (median, day 3; range, days 3
236 McCarthy et al.
The Journal of Thoracic and Cardiovascular Surgery
Fig. 3. Lung adherent over anastomosis at autopsy. 1, Esophagus. 2, Left lower lobe of lung. 3. Fibrin glue at anastomosis, 4, Stomach,
Fig. 4. A, Fibrin glue at anastomosis on day 14.1, Fibrin glue. 2, Lung. 3, Gastric wall. (Hematoxylin and eosin; original magnification X4.) B, Anastomosis healed at day 28. Fibrin glue resorbed. 1, Gastric mucosa. 2, Anastomosis. 3, Esophageal mucosa. (Hematoxylin and eosin; original magnification X4.)
to 13). In the one surviving control dog, the lower lobe of the left lung was firmly adherent to the anastomosis. There was no evidence of abscess formation or pleural contamination. It was presumed that survival was due to the immediate sealing of the perforation by the lung. Of the 11 glue-treated dogs, only four (36.4%) had an
anastomotic leak (p < 0,01). The four dogs with leak died early (median, day 5; range, days 3 to 8). Cause of leak was failure of the glue to adhere to tissue in three dogs and a perforation through the glue in one dog. Three additional dogs died of unrelated causes before the 14 day planned death. Cause of death in two dogs
Volume 93 Number 2 February 1987
(days 8 and 12) was infarcted small bowel after the bowel had herniated through the diaphragmatic defect. One other dog died (day 3) of massive herniation of abdominal contents into the left side of the chest when the gastric anchoring stitch tore, with subsequent respiratory failure. The anastomoses in all three dogs that died of other causes, and in the surviving four gluetreated dogs, were sealed without evidence of leak. In most, the lower lobe of the left lung was densely adherent around the anastomosis (Fig. 3). The postoperative leukocyte counts were lower for the glue-treated dogs (mean, 23,300jmm 3) than for the control animals (mean, 27,700jmm 3) . However, this difference was not statistically significant. Histologicevaluation of the glue-treated dogs demonstrated partial or near-complete resorption of the glue by 14 days. An inflammatory response consisting primarily of histiocytes without phagocytes was present around the glue. There was no giant cell formation. In one dog studied at 28 days, the glue was gone and there was no residual inflammatory response. The anastomosis was well-healed, with good coaptation of gastric and esophageal mucosae (Fig. 4). Discussion
Despite early aggressive medical and surgical management, the mortality associated with the leaking esophageal anastomosis remains high.!:' Recent techniques to secure the anastomoses include the use of the EEA staplers,13-15 "tunneled" esophagogastrostomy,16 omental wrap," and reinforcement of the suture line with adjacent stomach. 17. 18 Similarly, stomach used for esophageal replacement instead of colon or jejunum has provided a reliable conduit. Finally, anastomoses are more frequently being performed in the neck in an effort to prevent pleural contamination. Fibrin glue may provide an additional adjunct to this array of surgical techniques. In our model, fibrin glue was effective in preventing leakage despite being employed in a very difficult model with large anastomotic holes and early resumption of feeding. Unlike early reports describing synthetic glues,19.20 there was no foreign-body reaction to the fibrin glue, a physiologic substance, used in our study. In one dog studied at 28 days, the glue was completely resorbed without local tissue reaction, as others have also reported." Fibrin glue was also entirely nontoxic. Fibrin glue is rapidly becoming a useful tool in the surgical armamentarium. It has been very successful in controlling the bleeding in cardiothoracic operations. 7, 8. 22 Other uses in thoracic surgery include the sealing of tracheal anastomoses," the sealing of chyle
Esophagogastric anastomoses 2 3 7
leaks," and the sealing of pulmonary air leaks.2S-29* Other studies have suggested its effectiveness as an adjunct to seal bowel anastomoses.t'? Currently, commercially available fibrin glue (Immuno AG; personal communication) is undergoing evaluation by the Food and Drug Administration for possible release in the United States. In conclusion, fibrin glue was effective in decreasing the leakage from an experimental esophagogastric anastomosis. Healing was not complicated by foreign-body reaction. Fibrin glue providesan adhesive substance that may be useful clinically as an adjunct to seal esophageal anastomoses and prevent postoperative leaks.
'McCarthy PM, Trastek YF, Bell D, et al. The effectiveness offibrin glue sealant in reducing experimental pulmonary airleak. Submitted for publication. We would like to thank Duane M. Ilstrup for statistical assistance and Marilyn R. Oeltjen for technical assistance. REFERENCES Keagy BA, Murray GF, Starek PJK, Battaglini JW, Lores ME, Wilcox BR. Esophagogastrectomy as palliative treatment for esophageal carcinoma: results obtained in the setting of a thoracic surgery residency program. Ann Thorac Surg 1984;38:611-5. 2. Ellis FH Jr, Gibb SP, Watkins E Jr. Esophagogastrectomy: a safe, widely applicable, and expeditious form of palliation for patients with carcinoma of the esophagus and cardia. Ann Surg 1983;198:531-9. 3. Wilson SE, Stone R, Scully M, Ozeran L, Benfield JR. Modern management of anastomotic leakafter esophagogastrectomy. Am J Surg 1982;144:95-9. 4. Molina JE, Lawton BR, Myers WO, Humphrey EW. Esophagogastrectomy for adenocarcinoma of the cardia: ten years' experience and current approach. Ann Surg 1982; 195: 146-51. 5. Maillet P, Baulieux J, Boulez J, Benhaim R. Carcinoma of the thoracic esophagus: results of one-stage surgery (271 cases). Am J Surg 1982;143:629-34. 6. Payne WS, Leonardi HK, Ellis FH Jr. Complications of esophageal and diaphragmatic surgery. In: HardyJD, ed' Complications in surgery and their management, 4th ed. Philadelphia: WB Saunders, 1981 :330-66. 7. Borst HG, Haverich A, Walterbusch G, Maatz W. Fibrin adhesive: an important hemostatic adjunct in cardiovascular operations. J THoRAc CARDIOVASC SURG 1982;84:54852. 8. Borst HG, Haverich A. Clinical use of fibrin adhesive: summary. Thorac Cardiovasc Surg 1982;30:241. 9. Petrelli NJ, Cohen H, DeRisi D, Ambrus JL, Williams P, The application of tissue adhesives in small bowel anastomoses. J Surg Oneal 1982; 19:59-61. 10. Thorson GK, Perez-Brett R, Lillie DB, et al. The role of the tissue adhesive fibrin seal (FS) in esophageal anastomoses. J Surg Oncol 1983;24:221-3. I.
The Journal of Thoracic and Cardiovascular
2 3 8 McCarthy et al.
11. Oka H, Harrison RC, Burhenne HJ. Effect of a biologic glue on the leakage rate of experimental rectal anastomoses. Am J Surg 1982;143:561-4. 12. Harrison RC, Oka H. Rectal anastomosis: sutures vs staples and glue. Contemp Surg 1982;21:17-26. 13. Hopkins RA, Alexander JC, Postlethwait RW. Stapled esophagogastric anastomosis. Am J Surg 1984;147:2837. 14. Postlethwait RW. Complications and deaths after operations for esophageal carcinoma. J THoRAc CARDIOVASC SURG 1983;85:827-31. 15. Fekete F, Breil P, Ronsse BH, Tossen JC, Langonnet F. EEA stapler and omental graft in esophagogastrectomy: experience with 30 intrathoracic anastomoses for cancer. Ann Surg 1981;193:825-30. 16. Liu K, Zhang GC, Cai ZJ. Avoiding anastomotic leakage following esophagogastrostomy. J THoRAc CARDIOVASC SURG 1983;86:142-5. 17. Hoffmann TH, Kelley JR, Grover FL, Trinkle JK. Carcinoma of the esophagus: an aggressive one-stage palliative approach. J THoRAc CARDIOVASC SURG 1981;81:44-8. 18. Piccone VA, LeVeen HH, Ahmed N, Grosberg S. Reappraisal of esophagogastrectomy for esophageal malignancy. Am J Surg 1979;137:32-7. 19. Raekallio J, Seligman AM. Acute reaction to arterial adhesive in healing skin wounds. J Surg Res 1964;4:124-7. 20. Woodward SC, Herrmann JB, Leonard F. Histotoxicity of cyanoacrylate tissue adhesives (Abstract). Fed Proc 1964;23:495. 21. Redl H, Schlag G, Stanek G, HirschI A, Seelich T. In vitro properties of mixtures of fibrin seal and antibiotics. Biomaterials 1983;4:29-32. 22. Jakob H, Campbell CD, Qiu ZK, Pick R, Replogle RL. Evaluation of fibrin sealing for cardiovascular surgery. Circulation 1984;70(Pt 2):1138-46. 23. Kram HB, Shoemaker WC, Hino ST, Chiang HS, Harley DP, Fleming A W. Tracheal repair with fibrin glue. J THORAC CARDIOVASC SURG 1985;90:771-5. 24. Stenzl W, Rigler B, Tscheliessnigg KH, Beitzke A, Metzler H. Treatment of postsurgical chylothorax with fibrin glue. Thorac Cardiovasc Surg 1983;31:35-6. 25. DeRisi 0, Cohen H, Takita H, et al. Hemostasis in experimental pulmonary injury. J Surg Oncol 1982;19:208-1O. 26. J~rgensen A, M~ller IW, Johnsen A, Borgeskov S. Bronchopleural leakage treated with fibrin sealant and high-frequency positive-pressure ventilation. Scand J Thorae Cardiovasc Surg 1984;18:139-40. 27. Thetter O. Fibrin adhesive and its application in thoracic surgery. Thorac Cardiovasc Surg 1981;29:290-2. 28. Jessen C, Sharma P. Use of fibrin glue in thoracic surgery. Ann Thorac Surg 1985;39:521-4. 29. Tiirk R, Weidringer JW, Hartel W, Bliimel G. Closure of lung leaks by fibrin gluing: experimental investigations and clinical experience. Thorac Cardiovasc Surg 1983;31:185-6.
Surgery
Discussion DR. RICHARD M. PETERS San Diego, Calif
What was the reason for the failure of the glue to adhere in the animals in which a leak developed? DR. McCARTHY It is hard to determine at autopsy, but approximately 90% of the glue was adherent, and there were just small areas where the glue seemed to have lifted up from where it had been applied to the stomach, with leakage through those small areas. High intraluminal pressure may have pushed the glue off, or gastric enzymes may have broken down the fibrin matrix. DR. DONALD L. MORTON Los Angeles, Calif
This study clearly indicates the value of fibrin glue in sealing an inadequately performed anastomosis. The experimental model developed by Dr. McCarthy and his associates is an excellent one; 13 of 14 anastomoses in the control group leaked whereas only four of 11 anastomoses sealed with fibrin glue leaked-a highly significant difference and a clear demonstration of the effectiveness of fibrin glue in sealing the anastomosis. Most experience with fibrin glue has been in cardiac and vascular surgery, where it has been highly effective in controlling bleeding as a hemostatic. There has been little prior experience with it in gastrointestinal operations, where it might have been thought likely to be degraded by intestinal enzymes; this might be one of the explanations for the leaks that you did see. However, it would appear to have a definite advantage over other tissue adhesives, since it is a natural product and is completely absorbed by the body, unlike the tissue adhesive that we have used for bronchial anastomosis, which is not absorbed. What is the clinical significance of these observations? I must admit to some of the feelings of nostalgia expressed by Benson Roe regarding the passing of craftsmanship in surgery. My view has always been that the tragedy of anastomotic leak is so great that I teach my residents that if the anastomosis does not go well or look right, it should be taken down and redone. Furthermore, I test the anastomosis during the operation by instilling a methylene blue solution into the esophageal lumen to distend the anastomosis, allowing any leaks to be detected by the appearance of the blue dye. If these precepts are used, leakage of the esophagogastric anastomosis should be a phenomenon occurring rarely or not at all, as reported in the beautiful series by Dr. Mitchell, which reflects our own experience. Some of the major causes of anastomotic leaks are (1) ischemic necrosis of the anastomotic site because of inadequate blood supply to the esophagus or stomach, (2) excessive tension at the suture line, (3) sutures placed so tightly that they cut through the anastomosis, or (4) poorly placed sutures creating a gap in the suture line, as in this experimental model. It would clearly appear that the last two might be benefited by fibrin glue. However, I doubt if glue
Volume 93 Number 2 February 1987
will protect against ischemic necrosis or excessive tension. On the other hand, these are rarely a problem with esophagogastric anastomosis. This study clearly indicates that fibrin glue does provide an extra margin of safety to protect against sloppy craftsmanship in surgery and might be recommended for that purpose. DR. McCARTHY Thank you for your comments, Dr. Morton. Your point about precise anastomotic technique is a good one, and fibrin glue is intended to supplement excellent technique, not replace it. We hope that as a "backup" it will not be needed very often. DR. ROBERT W. JAMPLIS Palo Alto, Calif
This is very important paper. All of us know that an acute mediastinitis develops as a result of an esophageal leak; it is a death warrant diagnosis very often. I do not think all of us are as good surgeons as Dr. Morton or Dr. Mitchell, who reported yesterday on 40 consecutive esophageal resections without a leak. Even those of us who use staplers are going to get leaks. Therefore, I think it is very important, not so much that two of the three of the authors' dogs survived even with large holes,
Esophagogastric anastomoses
239
because that happens occasionally, but because there were no complications and no deleterious effects. Since I always cover the bronchus with pleura or whatever, you might say I am a big belt and suspenders man. Therefore, if there are no complications, I would want to use it. I would like to ask whether you are doing it in the operating room now, and if not, why not? DR. McCARTHY Right now, the fibrin glue that we used, the commercial product from Austria, is not available in the United States. It is currently undergoing phase 2 clinical trials in a multicenter study for bleeding with reoperations at cardiac surgery. The risk with using the commercial fibrin glue is that it theoretically can transmit hepatitis and acquired immune deficiency syndrome. However, the company reports that in one million applications in Europe, they have yet to have a reported causal association of hepatitis or AIDS. They go to extensive measures trying to screen the donors, and they also heat treat the product before it is released, to kill the virus. The Food and Drug Administration clinical trials will be completed soon and the commercial product may be made available in the United States, as it was in Canada a few months ago.