Endoscopic Ligation of Esophageal Varices Greg Van Stiegmann, MD, John S. Goff, MD, John H. Sun, MD, Daniel Hruza, MD, R. Matthew Reveille, ~O, Denver,Colorado
One hundred consecutive patients with bleeding esophageal varices were treated with a new endoscopic ligating device that effects strangulation of varices using small elastic " O " rings. Treatments were continued after initial hospitalization t o achieve variceal eradication. Follow-up ranged from 6 t o 26 (mean: i 5 ) months. Bleeding was controlled until discharge from hospital or death in 18 of 21 patients who were actively bleeding at index endoscopy. Overall, 26 patients died during the study, 12 during the index hospitalization. Cause of death was organ failure in 21, exsanguination in 3, and cancer in 2. Forty-one of 88 initial survivors experienced 72 episodes of recurrent bleeding ( 1 t o 4 per patient). All but five rebleeds occurred before eradication. Sixty of 88 patients (68%) who survived index hospitalization had their varices eradicated. A median of 5 (1 to 12) treatments was required. Nine patients eventually had other forms of treatment for recurrent bleeding. Only 3 non-bleeding complications resulted from 462 endoscopic treatment sessions. We conclude that endoscopic ligation controls active variceal bleeding and eradicates varices with efficacy similar to that of sclerotherapy and with minimal risk of complications.
ndoscopic treatment of hemorrhage from esophageal E varices has gained wide acceptance. Sclerotherapy appears to confer a decrease in mortality to those so treated; however, significant complications associated with this therapy occur in up to 20% of patients [1,2]. We have developed a new form of endoscopic treatment for bleeding from esophageal varices that relies on mechanical ligation and strangulation of varices using small elastic "O" rings [3,4]. Endoscopic variceal ligation was developed to provide a safer alternative than sclerotherapy. The purpose of this study was to assess the shortand long-term efficacy of this new technique in the first 100 adult patients so treated.
MATERIAL AND METHODS One hundred consecutive patients with bleeding from esophageal varices had endoscopic variceal ligation as primary and definitive therapy during the 26-month period ending November 1~, 1988. This group constituted our entire clinical experience with endoscopic variceal ligation in adults and includes subsets of patients who have been previously described [5,6]. No patients had primary treatment by operation or sclerotherapy during this period with the exception of 10 good-risk (Child A and B) patients who were entered into another study. Cause of portal hypertension and modified ChildPugh class of the endoscopic variceal ligation cohort are shown in Table I [7]. The mean age was 52 years (range: 20 to 87), and 67 were men. Twenty-four patients had undergone from 1 to 7 (mean: 2) endoscopic sclerotherapy sessions elsewhere and were referred for treatment of recurrent (18 patients) or uncontrollable (6 patients) hemorrhage. Complete blood product data for the bleeding episode that initiated endoscopic variceal ligation treatment were available for 85 patients; the number of blood transfusions ranged from 0 to 30 (mean: 5) units per patient. All patients had pan-upper endoscopy after hemodynamic stabilization or within 12 hours of admission to the hospital. Endoscopic examinations (and treatments) were performed using intravenous sedation (midazolam, meperidine) with or without topical pharyngeal anesthesia. Bleeding from esophageal varices was determined if an actively bleeding varix was identified endoscopically, clot was present on a varix with no other bleeding source found, or large esophageal varices were present and no other source of bleeding was visualized. The number of discrete variceal changes was noted, and the size of varices was estimated and graded (grade I, 1 to 3 mm; grade Fromthe Departmentsof Surgery(Gastrointestinal/Tumor)and Med- II, 4 to 6 ram; grade III, 7 to 9 mm; grade IV, 10+ mm). Endoscopic variceal ligation was performed at the icine(Gastroenterology),Universityof Colorado,and DenverVeterans AdministrationHospitals,Denver,Colorado. initial endoscopic session immediately after confirmation Requestsfor reprintsshouldbe addressedto GregVan Stiegmann, of variceal hemorrhage. The technique has been deMD, Universityof ColoradoHealthScienceCenter,C-313 4200 East scribed in detail elsewhere [8]. Briefly, a specially deNinth Avenue,Denver,Colorado80262. Presentedat the 30thAnnualMeetingof the Societyfor Surgeryof signed 27-cm overtube was advanced over the endoscope the AlimentaryTract, Washington,D.C., May 16-17, 1989. into the proximal esophagus. The ligating device (Bard THE AMERICAN JOURNAL OF SURGERY VOLUME 159 JANUARY 1990 21
VAN STIEGMANN ET AL
TABLE I
Cause of Portal Hypertension and Modified Child-Pugh Class of 100 Patients with Bleeding Varlces Treated with Endoscopic Variceal Ligation n
Cause of portal hypertension Laennec's cirrhosis Other cirrhosis Chronic hepatitis Portal vein thi'ombosis Modified Child-Pughclass
68 16 12 4
A
3i
B C
38 31
Figure 1. The technique of endoscopic variceal ligation. Top lett, the endoscopist approaches a Varix in the distal esophagus and makes contact between the varix and the end of the endoscopic ligating device. Top right, suction (from the endoscope) draws the varix into the ,gating chamber. Bottom left, the inner cylinder is movedtoward the endoscope, by the oPerator who pulls a trip wire connected to it that runs retrograde through the endoscope's working channel. This movement ejects the elastic "O" ring from the inner cylinder, capturing the varix. Bottom right, the varix is ensnared by the elastic "O" ring. The procedure is repeated until all variceal tissue in the distal esophagus and proximal stomachs is ,gated.
22
THE AMERICAN JOURNAL OF SURGERY
Interventional Products, Billerica, MA) was attached to the endoscope (Pentax F G / F E 34-A, Olympus Q) and the endoscope was reintroduced through the overtube. The method of ligation is shown in Figure 1. Varices in the far distal esophagus or proximal 1 to 2 cm of the stomach were treated first. Treatments continued proximally until all variceal tissue in the distal 5 to 7 cm of the esophagus was ligated. The initial session usually consisted of 8 to 12 individual ligations per patient. Actively bleeding varices were treated by direct ligation of the bleeding site when possible. If the bleeding varix was not identified, ligation was initiated at or just below the gastroesophageal junction, and multiple ligations were performed until bleeding ceased. Patients treated for active bleeding received maintenance therapy with an oral sucralfate suspension or H2 antagonists, or both, while in the hospital. When feasible, a liquid diet was allowed the day of treatment and a full diet was resumed the next morning. Endoscopic variceal ligation was repeated at 7- to 14day intervals until such time as varices in the distal esophagus were eradicated or reduced to grade I or less in size. Second and third treatment sessions usually consisted of three to six individual ligations of residual variceal tissue. Fourth and subsequent treatments usually consisted of one to three individual ligations since there was progression toward eradication. Thereafter, endoscopy was repeated at three-month intervals, with retreatment for recurrent varices at the time of their discovery. Diagnostic endoscopy and repeat endoscopic variceal ligation were performed as needed for recurrent hemorrhage. The following parameters were assessed during the course of the study: Survival, ability to control active variceal hemorrhage, incidence of recurrent hemorrhage, ability to eradicate varices, recurrence of varices after eradication, and non-bleeding complications. Control of hemorrhage was defined as absence of clinically detectable Upper gastrointestinal bleeding after endoscopic variceal ligation was performed for active bleeding. Complications were defined as any untoward non-hemorrhagic event resulting from endoscopic treatment that required prolonged hospitalization or necessitated active medical or surgical treatment. Recurrent hemorrhage was defined as clinically apparent upper gastrointestinal bleeding that resulted in the patient's seeking medical attention regardless of the need for blood transfusion. Variceal eradication was defined as the absence of visible variceal channels in the distal 5 cm of the esophagus or reduction in size to venules o f diminutive grade I size. Statistical analysis of discrete variables was performed using chi-square with Yates' correction when applicable and Student's t test for unpaired variables. This study was approved by the investigational review board of both hospitals and all patients gave informed consent. RESULTS Bleeding had ceased in 79 patients (79%) before the first endoscopic variceai ligation treatment. Forty-eight of these 79 patients were admitted directly to our hospital
VOLUME 159
JANUARY 1990
ENDOSCOPIC LIGATION OF ESOPHAGEAL VARICES
because of upper gastrointestinal hemorrhage, and 31 were stabilized at other hospitals and transferred to our institution. Twenty-nine were Child-Pugh class A, 30 class B, and 20 class C. Seventy-six had grade III or grade IV varices, and three had grade II varices. Three (4%) of these 79 patients (1 Child B, 2 Child C) did not survive index hospitalization (Figure 2). Twenty-one patients (21%) were actively bleeding at the time of their initial endoscopic variceal ligation treatment. Two were Child-Pugh A, 8 B, and 11 C. All had grade III to IV varices. In 18 patients (86%), control of bleeding was achieved during the index admission or until the time of death. In 14 patients, bleeding was controlled with one endoscopic variceal ligation treatment, whereas four patients required two such treatments. Two of the three patients in whom control of hemorrhage was not achieved were bleeding from mucosal slough and ulceration induced by prior sclerotherapy. Nine (43%) patients with active bleeding (1 Child B, 8 Child C) did not survive their index hospitalization. Mortality in actively bleeding patients was significantly greater than in those in whom bleeding ceased spontaneously before endoscopic variceal ligation (p
One hundred consecutive patients with bleeding esophageal varices were treated with a new endoscopic ligating device that effects strangulation of varices using small elastic " O " rings. Treatments were continued after initial hospitalization t o achieve variceal eradication. Follow-up ranged from 6 t o 26 (mean: i 5 ) months. Bleeding was controlled until discharge from hospital or death in 18 of 21 patients who were actively bleeding at index endoscopy. Overall, 26 patients died during the study, 12 during the index hospitalization. Cause of death was organ failure in 21, exsanguination in 3, and cancer in 2. Forty-one of 88 initial survivors experienced 72 episodes of recurrent bleeding ( 1 t o 4 per patient). All but five rebleeds occurred before eradication. Sixty of 88 patients (68%) who survived index hospitalization had their varices eradicated. A median of 5 (1 to 12) treatments was required. Nine patients eventually had other forms of treatment for recurrent bleeding. Only 3 non-bleeding complications resulted from 462 endoscopic treatment sessions. We conclude that endoscopic ligation controls active variceal bleeding and eradicates varices with efficacy similar to that of sclerotherapy and with minimal risk of complications.
ndoscopic treatment of hemorrhage from esophageal E varices has gained wide acceptance. Sclerotherapy appears to confer a decrease in mortality to those so treated; however, significant complications associated with this therapy occur in up to 20% of patients [1,2]. We have developed a new form of endoscopic treatment for bleeding from esophageal varices that relies on mechanical ligation and strangulation of varices using small elastic "O" rings [3,4]. Endoscopic variceal ligation was developed to provide a safer alternative than sclerotherapy. The purpose of this study was to assess the shortand long-term efficacy of this new technique in the first 100 adult patients so treated.
MATERIAL AND METHODS One hundred consecutive patients with bleeding from esophageal varices had endoscopic variceal ligation as primary and definitive therapy during the 26-month period ending November 1~, 1988. This group constituted our entire clinical experience with endoscopic variceal ligation in adults and includes subsets of patients who have been previously described [5,6]. No patients had primary treatment by operation or sclerotherapy during this period with the exception of 10 good-risk (Child A and B) patients who were entered into another study. Cause of portal hypertension and modified ChildPugh class of the endoscopic variceal ligation cohort are shown in Table I [7]. The mean age was 52 years (range: 20 to 87), and 67 were men. Twenty-four patients had undergone from 1 to 7 (mean: 2) endoscopic sclerotherapy sessions elsewhere and were referred for treatment of recurrent (18 patients) or uncontrollable (6 patients) hemorrhage. Complete blood product data for the bleeding episode that initiated endoscopic variceal ligation treatment were available for 85 patients; the number of blood transfusions ranged from 0 to 30 (mean: 5) units per patient. All patients had pan-upper endoscopy after hemodynamic stabilization or within 12 hours of admission to the hospital. Endoscopic examinations (and treatments) were performed using intravenous sedation (midazolam, meperidine) with or without topical pharyngeal anesthesia. Bleeding from esophageal varices was determined if an actively bleeding varix was identified endoscopically, clot was present on a varix with no other bleeding source found, or large esophageal varices were present and no other source of bleeding was visualized. The number of discrete variceal changes was noted, and the size of varices was estimated and graded (grade I, 1 to 3 mm; grade Fromthe Departmentsof Surgery(Gastrointestinal/Tumor)and Med- II, 4 to 6 ram; grade III, 7 to 9 mm; grade IV, 10+ mm). Endoscopic variceal ligation was performed at the icine(Gastroenterology),Universityof Colorado,and DenverVeterans AdministrationHospitals,Denver,Colorado. initial endoscopic session immediately after confirmation Requestsfor reprintsshouldbe addressedto GregVan Stiegmann, of variceal hemorrhage. The technique has been deMD, Universityof ColoradoHealthScienceCenter,C-313 4200 East scribed in detail elsewhere [8]. Briefly, a specially deNinth Avenue,Denver,Colorado80262. Presentedat the 30thAnnualMeetingof the Societyfor Surgeryof signed 27-cm overtube was advanced over the endoscope the AlimentaryTract, Washington,D.C., May 16-17, 1989. into the proximal esophagus. The ligating device (Bard THE AMERICAN JOURNAL OF SURGERY VOLUME 159 JANUARY 1990 21
VAN STIEGMANN ET AL
TABLE I
Cause of Portal Hypertension and Modified Child-Pugh Class of 100 Patients with Bleeding Varlces Treated with Endoscopic Variceal Ligation n
Cause of portal hypertension Laennec's cirrhosis Other cirrhosis Chronic hepatitis Portal vein thi'ombosis Modified Child-Pughclass
68 16 12 4
A
3i
B C
38 31
Figure 1. The technique of endoscopic variceal ligation. Top lett, the endoscopist approaches a Varix in the distal esophagus and makes contact between the varix and the end of the endoscopic ligating device. Top right, suction (from the endoscope) draws the varix into the ,gating chamber. Bottom left, the inner cylinder is movedtoward the endoscope, by the oPerator who pulls a trip wire connected to it that runs retrograde through the endoscope's working channel. This movement ejects the elastic "O" ring from the inner cylinder, capturing the varix. Bottom right, the varix is ensnared by the elastic "O" ring. The procedure is repeated until all variceal tissue in the distal esophagus and proximal stomachs is ,gated.
22
THE AMERICAN JOURNAL OF SURGERY
Interventional Products, Billerica, MA) was attached to the endoscope (Pentax F G / F E 34-A, Olympus Q) and the endoscope was reintroduced through the overtube. The method of ligation is shown in Figure 1. Varices in the far distal esophagus or proximal 1 to 2 cm of the stomach were treated first. Treatments continued proximally until all variceal tissue in the distal 5 to 7 cm of the esophagus was ligated. The initial session usually consisted of 8 to 12 individual ligations per patient. Actively bleeding varices were treated by direct ligation of the bleeding site when possible. If the bleeding varix was not identified, ligation was initiated at or just below the gastroesophageal junction, and multiple ligations were performed until bleeding ceased. Patients treated for active bleeding received maintenance therapy with an oral sucralfate suspension or H2 antagonists, or both, while in the hospital. When feasible, a liquid diet was allowed the day of treatment and a full diet was resumed the next morning. Endoscopic variceal ligation was repeated at 7- to 14day intervals until such time as varices in the distal esophagus were eradicated or reduced to grade I or less in size. Second and third treatment sessions usually consisted of three to six individual ligations of residual variceal tissue. Fourth and subsequent treatments usually consisted of one to three individual ligations since there was progression toward eradication. Thereafter, endoscopy was repeated at three-month intervals, with retreatment for recurrent varices at the time of their discovery. Diagnostic endoscopy and repeat endoscopic variceal ligation were performed as needed for recurrent hemorrhage. The following parameters were assessed during the course of the study: Survival, ability to control active variceal hemorrhage, incidence of recurrent hemorrhage, ability to eradicate varices, recurrence of varices after eradication, and non-bleeding complications. Control of hemorrhage was defined as absence of clinically detectable Upper gastrointestinal bleeding after endoscopic variceal ligation was performed for active bleeding. Complications were defined as any untoward non-hemorrhagic event resulting from endoscopic treatment that required prolonged hospitalization or necessitated active medical or surgical treatment. Recurrent hemorrhage was defined as clinically apparent upper gastrointestinal bleeding that resulted in the patient's seeking medical attention regardless of the need for blood transfusion. Variceal eradication was defined as the absence of visible variceal channels in the distal 5 cm of the esophagus or reduction in size to venules o f diminutive grade I size. Statistical analysis of discrete variables was performed using chi-square with Yates' correction when applicable and Student's t test for unpaired variables. This study was approved by the investigational review board of both hospitals and all patients gave informed consent. RESULTS Bleeding had ceased in 79 patients (79%) before the first endoscopic variceai ligation treatment. Forty-eight of these 79 patients were admitted directly to our hospital
VOLUME 159
JANUARY 1990
ENDOSCOPIC LIGATION OF ESOPHAGEAL VARICES
because of upper gastrointestinal hemorrhage, and 31 were stabilized at other hospitals and transferred to our institution. Twenty-nine were Child-Pugh class A, 30 class B, and 20 class C. Seventy-six had grade III or grade IV varices, and three had grade II varices. Three (4%) of these 79 patients (1 Child B, 2 Child C) did not survive index hospitalization (Figure 2). Twenty-one patients (21%) were actively bleeding at the time of their initial endoscopic variceal ligation treatment. Two were Child-Pugh A, 8 B, and 11 C. All had grade III to IV varices. In 18 patients (86%), control of bleeding was achieved during the index admission or until the time of death. In 14 patients, bleeding was controlled with one endoscopic variceal ligation treatment, whereas four patients required two such treatments. Two of the three patients in whom control of hemorrhage was not achieved were bleeding from mucosal slough and ulceration induced by prior sclerotherapy. Nine (43%) patients with active bleeding (1 Child B, 8 Child C) did not survive their index hospitalization. Mortality in actively bleeding patients was significantly greater than in those in whom bleeding ceased spontaneously before endoscopic variceal ligation (p
