Original Article Mortality Secondary to Esophageal Anastomotic Leak Khaled Alanezi, MD, and John D. Urschel, MD Background: Esophageal anastomotic leak is a potentially life threatening complication of esophagectomy and esophagogastrectomy. We reviewed our experience with this complication and tried to identify factors predictive of mortality after esophageal anastomotic leak. Methods: Records of patients undergoing esophagectomy and esophagogastrectomy for benign or malignant disease over a 1
of 5
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
  71Ann Thorac Cardiovasc Surg Vol. 10, No. 2 (2004) Original  Article Introduction Many patients undergoing esophagectomy andesophagogastrectomy suffer postoperative complications,and these postoperative complications are often life threat-ening in nature. 1-4) The anastomosis between the remain-ing esophagus and its replacement conduit is more likelyto leak than most other gastrointestinal anastomoses. Post-operative esophageal anastomotic leaks range in severityfrom asymptomatic and minor anastomotic defects thatare only apparent on contrast studies, to fulminant leakswith systemic sepsis and multiorgan failure. 2) Most esoph-ageal surgery units have noted a favorable trend towardsreduced incidence and morbidity of esophageal leaks overthe past two decades. 1) Nevertheless, esophageal anasto-motic leaks remain an important source of morbidity andmortality after surgery.We reviewed our experience with esophageal anasto-motic leaks. Our goal was to identify factors predictiveof mortality. By identifying these factors, we hoped tomodify our future management of leaks, and reduce mor-tality. Mortality Secondary to Esophageal Anastomotic Leak Khaled Alanezi, MD, and John D. Urschel, MD From Department of Surgery, McMaster University, Ontario,Canada Received April 16, 2003; accepted for publication June 27, 2003.Address reprint requests to John D. Urschel, MD: P.O. Box 612,Buffalo, NY 14209-0612, USA. Background: Esophageal anastomotic leak is a potentially life threatening complication of esophagectomy and esophagogastrectomy. We reviewed our experience with this complicationand tried to identify factors predictive of mortality after esophageal anastomotic leak.Methods: Records of patients undergoing esophagectomy and esophagogastrectomy for benignor malignant disease over a 10-year period (1989-1999), who developed esophageal anastomoticleaks, were reviewed.Results: Three-hundred and seven patients underwent esophagectomy or esophagogastrectomy.Twenty-three (7.5%) developed esophageal anastomotic leaks. Eight of these patients (35%)died. Four of 23 (17%) patients had seemingly normal postoperative contrast studies. Factorspotentially predictive of death included age (died, 72.8 ± 8.3 years; survived, 65.3 ± 8.8 years;p=0.063), location of anastomosis (cervical, 3/9 died; thoracic, 5/14 died; p=0.91), leak presenta-tion (clinical, 6/12 died; contrast study, 2/11 died; p=0.11), time of leak (<7 days, 3/5 died; ≥ 7days, 5/18 died; p=0.18), presence of gastric necrosis (necrosis, 3/3 died; no necrosis, 5/20 died;p=0.019), and treatment (surgical, 4/4 died; conservative, 4/19 died; p=0.005).Conclusions: Postoperative esophageal anastomotic leaks prove fatal in a significant number of cases. The lethal potential of cervical anastomotic leaks should not be underestimated. Gastricnecrosis is an important predictor of subsequent death. Advanced age, early postoperative (<7days) leakage, and clinically apparent signs of leakage may be predictive of death but thesefactors did not reach statistical significance in our study. Surgical treatment of esophageal anas-tomotic leaks is associated with subsequent death, but this relationship is unlikely to be causal; se-verely ill patients tend to be treated surgically. (Ann Thorac Cardiovasc Surg 2004; 10: 71–5)Key words: esophageal neoplasms/surgery, postoperative complications, esophagectomy/adverseeffects, anastomosis/surgical, reconstructive surgical procedures  72  Alanezi et al. Ann Thorac Cardiovasc Surg Vol. 10, No. 2 (2004) Patients and Methods St. Joseph’s Healthcare in Hamilton, Ontario, Canada isa regional thoracic surgical center, and a McMaster Uni-versity teaching hospital. Records of patients undergoingesophagectomy and esophagogastrectomy for benign ormalignant disease over a 10-year period (1989-1999), anddeveloping esophageal anastomotic leaks, were reviewed.Data was collected on age, sex, pathology, operative ap-proach, location and technique of anastomosis, time andpresentation of leak, gastric necrosis, treatment, and out-come. Our analysis was designed to identify factors pre-dictive of death.Continuous data are presented as mean values withstandard deviations (mean ± SD). Means were comparedwith the Mann-Whitney test. Categorical data was com-pared with a chi-squared test (with Yates correction) orFisher’s exact test. A p<0.05 was considered significant.Statistical analysis was done using Biostat software(Biostat, Englewood, NJ). Results Three-hundred and seven patients underwent eso-phagectomy or esophagogastrectomy. Twenty-three(7.5%) developed esophageal anastomotic leaks (diag-nosed clinically and/or radiographically), and these pa-tients form the basis of this analysis. Twenty-one patientshad undergone transthoracic (McKeown, Lewis, orthoracoabdominal) resections, and two had non-thoracicapproaches (transhiatal and laparoscopic). Summary dataon the patients is presented in Table 1. Four of 23 pa-tients (17%) had seemingly normal postoperative con-trast studies. Eight of 23 patients (35%) died.Factors potentially predictive of death included age(died, 72.8 ± 8.3 years; survived, 65.3 ± 8.8 years; p=0.063),location of anastomosis (cervical, 3/9 died; thoracic, 5/ 14 died; p=0.91), technique of anastomosis (stapled, 4/7died; hand-sewn, 4/16 died; p=0.14), leak presentation(clinical, 6/12 died; contrast study, 2/11 died; p=0.11),time of leak (<7 days, 3/5 died; ≥ 7 days, 5/18 died;p=0.18), presence of gastric necrosis (necrosis, 3/3 died;no necrosis, 5/20 died; p=0.019), and treatment (sur-gical, 4/4 died; conservative, 4/19 died; p=0.005)(Table 2). Discussion The etiology of esophageal anastomotic leaks afteresophagectomy and esophagogastrectomy is often mul- AgeSexPathologyTumor locationTechnique of anastomosisLocation of anastomosisPresentation of leak - clinical or radiographicPresentation of leak - timeGastric necrosisManagement of leak Length of hospital staySurvival67.9 ± 9.1 yearsMale 18Female 5Adenocarcinoma 16Squamous cancer 7Proximal esophagus 1Mid esophagus 5Lower esophagus 4Gastric cardia 13Stapled 7Hand-sewn 16Intrathoracic 14Cervical 9Clinical 12Radiographic 11 ≤ 7 days 11>7 days 12Necrosis 3No necrosis 20Surgical 4Conservative 1936.4 ± 19.4 daysSurvived 15Died 8 Table 1. Summary data on 23 patients suffering esophageal anastomotic leaksafter esophagectomy and esophagogastrecomy  73  Esophageal Leaks Ann Thorac Cardiovasc Surg Vol. 10, No. 2 (2004) tifactorial. 1,2) However, most leaks can be attributed toconduit ischemia, technical errors, or a combination of the two. 5,6) Some anastomotic leaks are clinically asymp-tomatic (radiographic finding) while others cause fulmi-nant systemic sepsis. Given this spectrum of severity it isnot surprising that there is no standard management of esophageal anastomotic leaks. Treatment approaches toleaks are loosely based on the severity, or perceived se-verity, of the leak. Although treatment results have gen-erally improved over time, esophageal anastomotic leaksremain a major problem after esophageal surgery. 1) Our study was designed to identify factors predic-tive of death secondary to anastomotic leak. Regreta-bly, our group of 23 patients with anastomotic leaks isa large series relative to many others in the publishedliterature. 5,7-10) Nevertheless, the numbers of patients andevents (death) were too small to draw statistically soundconclusions on many issues. That, along with the retro-spective nature of our data collection, is the major limita-tion of our study. We have therefore tried to interpret ourfindings within the context of existing published data, andthen cautiously draw conclusions.We found cervical esophageal anastomotic leaks to bemore lethal than previously appreciated. 2) Surprisingly,our cervical leaks were as morbid as those complicatingthoracic anastomoses. An anastomosis constructed in theneck does not necessarily leak into the cervical wound; itmay leak into the mediastinum or pleural space. 7,8) We noted a trend (not significant) toward greater le-thality with leaks that manifest within the first week of surgery, and those that presented with clinical indicationsof leakage, such as wound drainage or systemic sepsis.This is consistent with current concepts of anastomoticleak pathophysiology. 2,3) The earlier an anastomosis fails,the more severe it tends to be; surrounding tissues needtime to wall off, or contain, the anastomosis. Similarly,leaks that manifest clinically are less likely to be con-tained by soft tissues than leaks that are asymptomatic,and only detectable on contrast studies. We also noted anon-significant (p=0.063) trend towards increasing leak lethality with advancing age, a finding that requires noexplanation.Our study was not designed to analyze diagnostic ap-proaches to possible esophageal leaks, but we were struck by the occurrence (four of 23 patients) of seemingly nor-mal water-soluble contrast studies in patients that laterproved to have leaks. Others have reported similar find-ings. 2,9) Although this could be a reflection of the limita-tions of water-soluble contrast studies relative to bariumexaminations, 11) it could also highlight some fundamen-tal limitations of routine contrast studies after esophagealanastomoses. For example, a fragile and faltering anasto-mosis on postoperative day 7 may not be leaking whenthe patient swallows contrast, but its insufficiency mayquickly become apparent when the nasogastric tube isremoved and the stomach distends. 2) The timing of con-trast studies is important. The earlier the study is done,the less reassuring it is. Patients destined to develop con-duit necrosis, for example, may have an apparently satis-factory contrast radiograph before the necrotic processbecomes full-thickness in extent. 9) We found surgical management of esophageal leaks tobe predictive of subsequent death, but this relationship isprobably not causal; surgery is a marker for severity of disease. Our four surgically treated patients were ex-tremely ill, and their surgical intervention could be char-acterized as being “ too little too late. ” This is a recurrenttheme in the unsuccessful management of esophagealanastomotic leaks. Most patients do not need formal sur-gical intervention but those who do need it require earlyand aggressive operative approaches. 1,2,5,7,8) Survival be-comes less likely if the septic cascade is allowed toprogress to shock and multiple organ failure. 7) Gastric necrosis was the only unequivocal predictor of mortality secondary to esophageal leaks in our patients.Other surgeons have also emphasized the lethal nature of  AgeTechnique of anastomosis - stapled vs. hand-sewnLocation of anastomosis - thoracic vs. cervicalPresentation of leak - clinical vs. radiographicPresentation of leak - before vs. after 7 daysGastric necrosis - present vs. absentManagement - surgical vs. conservativeVariablep=0.063p=0.14p=0.91p=0.11p=0.18p=0.019p=0.005Significance Table 2. Variables potentially predictive of death following esophagealanastomotic leak  74  Alanezi et al. Ann Thorac Cardiovasc Surg Vol. 10, No. 2 (2004) this complication. 2,7-9,12) Conduit necrosis, at best, producesa large and gaping anastomotic defect that leaks freely.At worst, the entire anastomosis dehisces and the foregutlooses continuity. Early recognition of this complicationis critical. Otherwise, these patients receive inappropri-ately conservative treatment. Clinical suspicion and lib-eral use of flexible endoscopy are the keys to early diag-nosis. 2,9,12) Our experience with esophageal anastomotic leaks,together with the published experience of others, hasshaped our conceptual framework for leak management.No one standard management strategy is applicable to allsituations. Instead, the aggressiveness and invasivenessof treatment should parallel the severity and potential le-thality of the leak. We believe that many fatal outcomesare secondary to one of two common mistakes: a failureto adequately diagnose and assess the severity of the leak,and a failure to appropriately match aggressiveness of treatment to leak severity.Early diagnosis of esophageal anastomotic leaks re-quires a willingness to liberally investigate patients whoshow subtle clinical signs of septic illness. Contrast radi-ography is the diagnostic cornerstone of esophageal leak investigation, but a seemingly normal contrast study, inisolation, may not exclude a leak. Gentle flexibleesophagoscopy is critical for assessment of possible con-duit necrosis. Computed tomography provides visualiza-tion of perianastomotic fluid collections and abscesses.Contrast studies, esophagoscopy, and computed tomog-raphy should be considered complimentary in the assess-ment of esophageal anastomotic leaks. They permit thesurgeon to answer these essential questions: is the leak contained, is the conduit viable, and is there an abscess?Aggressiveness of treatment should parallel the sever-ity of the anastomotic leak. We now try to match treat-ment and leak pathology by considering three broad treat-ment goals: adequate drainage of infected fluid collec-tions, resection of necrotic tissue, and prevention of fur-ther soilage from the leak. The importance of adequatedrainage cannot be over stated. 2,5,7) Seemingly innocuousperianastomotic fluid collections can fistulize into theairway or aorta, with catastrophic consequences. 2,7) In thepast drainage often required an operative approach, 2) butpercutaneous drainage is now effective for most pa-tients. 1,13) The provision of adequate drainage is moreimportant than the specific method by which it is obtained.Surgical drainage is currently reserved for patients withwidespread mediastinal and pleural soilage, and those withanother indication for surgery, namely tissue resection.The presence of conduit necrosis should prompt an earlyand bold return to the operating theater for disconnectionof the foregut, resection of nonviable tissue, and creationof an end esophagostomy. 2,7) Patients fortunate enough tosurvive such an illness, and not succumb to cancer recur-rence, can be reconstructed secondarily at another time.The last general management principle, prevention of further soilage from the leak, can be problematic. Elimi-nating oral intake and decompressing the conduit with anasogastric tube are basic steps. Many surgeons have at-tempted secondary surgical repair of anastomotic leaksin an effort to eliminate further soilage. If vascularizedtissue is used to reinforce the anastomosis there is somehope for success, and any recurrent leak may be bettercontained. Secondary repair can therefore be a usefuladjunctive strategy if surgical drainage is being done, butsecondary repair is not effective enough to warrant surgi-cal intervention for this reason alone. Finally, coveredexpandable metal stents can occlude anastomotic leaks. 10) Our recent and admittedly limited experience with thesestents has been favorable. We foresee increased use of stents in the future, but of course, they only address theissue of prevention of further soilage; drainage is stillnecessary and the possibility of conduit necrosis must beconsidered. References 1.Lerut T, Coosemans W, Decker G, De Leyn P, NafteuxP, van Raemdonck D. Anastomotic complications af-ter esophagectomy.  Dig Surg 2002; 19 : 92 – 8.2.Urschel JD. Esophagogastrostomy anastomotic leakscomplicating esophagectomy: a review.  Am J Surg 1995; 169 : 634 – 40.3.Pierie JP, de Graaf PW, van Vroonhoven TJ, ObertopH. Healing of the cervical esophagogastrostomy.  J AmColl Surg 1999; 188 : 448 – 54.4.Urschel JD, Blewett CJ, Bennett WF, Miller JD, YoungJE. Handsewn or stapled esophagogastric anastomosesafter esophagectomy for cancer: meta-analysis of ran-domized controlled trials.  Dis Esophagus 2001; 14 :212 – 7.5.Whooley BP, Law S, Alexandrou A, Murthy SC, WongJ. Critical appraisal of the significance of intrathoracicanastomotic leakage after esophagectomy for cancer.  Am J Surg 2001; 181 : 198 – 203.6.Urschel JD. Esophagogastric anastomotic leaks: theimportance of gastric ischemia and therapeutic appli-cations of gastric conditioning.  J Invest Surg 1998; 11 :245 – 50.7.Iannettoni MD, Whyte RI, Orringer MB. Catastrophiccomplications of the cervical esophagogastric anasto-mosis.  J Thorac Cardiovasc Surg 1995; 110 : 1493 –
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks