- Email: [email protected]
Injection sclerotherapy for variceal bleeding in patients with hepatocellular carcinoma: cyanoacrylate versus sodium tetradecyl sulphate
Injection sclerotherapy for variceal bleeding in patients with hepatocellular carcinoma: cyanoacrylate versus sodium tetradecyl sulphate Joseph J. Y. Sung, MD, Winnie Yeo, MD, Roamy Suen, RN, Yuk Tong Lee, MD, S. C. Sydney Chung, MD, Francis K. L. Chan, Philip J. Johnson, MD Background: Patients with hepatocellular carcinoma complicated by variceal bleeding have a very limited life span. Recurrent bleeding after endoscopic injection sclerotherapy is common. Our aim was to compare the efficacy of endoscopic injection of cyanoacrylate versus sodium tetradecyl sulphate in the control of variceal bleeding in patients with hepatocellular carcinoma. Methods: Patients known to be suffering from inoperable hepatocellular carcinoma who presented with upper gastrointestinal bleeding underwent endoscopy within 24 hours of admission. After bleeding from esophageal varices was confirmed, they were randomized to receive endoscopic injections of either cyanoacrylate (1:1 mixture with Lipoidol) or sodium tetradecyl sulphate (1.5%). Injection were given intravariceally into each visible column for up to four injections for cyanoacrylate and up to 30 mL for sodium tetradecyl sulphate. Results: A total of 50 patients were recruited for this study with 25 cases randomized to each endoscopic treatment group. Control of acute bleeding failed in four patients (16%) in both treatment groups, and two patients in each group died during the index episode of bleeding. Six patients (24%) in the cyanoacrylate group and four patients (16%) in the sodium tetradecyl sulphate group developed recurrent bleeding during their hospital stay (p 5 0.48). Recurrent bleeding within 30 days after the index episode of bleeding was documented in seven patients (28%) who received cyanoacrylate injection and five patients (20%) who received sodium tetradecyl sulphate injection (p 5 0.51). Median survival in the cyanoacrylate group was 16 days (range 1 to 485 days) and that of the sodium tetradecyl sulphate group was 13 days (range 1 to 407 days). There was no difference in cumulative survival between the two groups as analyzed by the Kaplan-Meier method. Patients with portal vein thrombosis had a higher risk of recurrent hemorrhage. Patients with Child’s C liver disease had a significantly higher mortality. Conclusions: Cyanoacrylate did not improve the outcome of hepatocellular carcinoma patients with variceal hemorrhage. (Gastrointest Endosc 1998;47:235-9.)
Gastrointestinal bleeding is frequently a terminal event in patients with hepatocellular carcinoma (HCC). Massive bleeding from gastroesophageal Received May 7, 1997. For revision September 22, 1997. Accepted November 4, 1997. From the Departments of Medicine, Clinical Oncology, and Surgery, The Chinese University of Hong Kong. Reprint requests: Joseph Sung, MD, Department of Medicine, Prince of Wales Hospital, Shatin, Hong Kong, China. Copyright © 1998 by the American Society for Gastrointestinal Endoscopy 0016-5107/98/$5.00 1 0 37/1/87366 VOLUME 47, NO. 3, 1998
varices has been reported as the immediate cause of death in about 20% of patients,1 and gastroesophageal varices account for half of the bleeding episodes.2 Life expectancy after bleeding was very limited, even with endoscopic sclerotherapy2-4 or variceal ligation.5 Compared with conservative treatment, endoscopic injection sclerotherapy has not been shown to prolong patient survival.3,4 Tumor invasion of the portal venous system is known to be associated with a higher risk of recurrent bleeding and mortality.2-4 Hepatic reserve is another prognostic factor determining the survival of patients.2,4 GASTROINTESTINAL ENDOSCOPY 235
J Sung, W Yeo, R Suen, et al.
Injection sclerotherapy for variceal bleeding with hepatocellular carcinoma
In view of the short life expectancy of HCC patients after an episode of variceal hemorrhage, we attempted to identify an endoscopic treatment with high success of initial hemostasis and low recurrent bleeding rate achieved by a single treatment to minimize discomfort and reduce hospital stay. N-butyl-2-cyanoacrylate (Histoacryl) was first used by Soehendra et al.6 in the treatment of large esophageal varices. This liquid tissue adhesive solidifies almost instantaneously on contact with blood. Recent randomized studies showed that nonHCC patients treated initially with cyanoacrylate injection during active bleeding and subsequently with conventional sclerosant have fewer postinjection episodes of bleeding7,8 and a better chance of survival.7 In this prospective randomized study, we compared the effects of endoscopic injection of cyanoacrylate against sodium tetradecyl sulphate (STD), after confirming bleeding from esophageal varices, in HCC patients whose course is complicated by variceal bleeding. PATIENTS AND METHODS Patients known to be suffering from unresectable HCC presenting with upper gastrointestinal bleeding were recruited for this study. All patients underwent endoscopic examination within 24 hours after admission. Patients with confirmed esophageal varices and no other possible source of bleeding were randomized to receive either cyanoacrylate or 1.5% STD injections. Signs of recent hemorrhage from varices (red color sign, fibrin clot) were documented during endoscopy. Exclusion criteria included (1) patients who were admitted in hepatic coma or moribund condition, (2) patients who had other possible source of bleeding in the upper gastrointestinal tract such as peptic ulcer or tumor invasion into stomach or duodenum, and (3) patients who had received endoscopic treatment or surgery for variceal bleeding in the previous 6 months. Endoscopic injection of the varices with cyanoacrylate or 1.5% STD was performed using a twin-channel therapeutic gastroscope (Olympus 2T-200; Olympus, Tokyo, Japan) and a flexible injection probe (Marcon-Harber injector; Wilson-Cook, Winston-Salem, N.C.). Cyanoacrylate was mixed in equal volume with Lipiodol (0.5 mL cyanoacrylate in 0.5 mL Lipiodol) before injection to prevent hardening of the glue in the injector and to allow x-ray monitoring of the injection. Both cyanoacrylate and STD were injected intravariceally into all identifiable columns of esophageal varices. Single injections of the cyanoacrylate mixture (1 mL per injection) were given into each variceal columns. STD injections (aliquots of 2 mL per injection) were injected into each column at multiple levels. Patients were carefully monitored for changes in blood pressure and pulse after undergoing endoscopic hemostasis. An acute bleeding episode was defined according to criteria set forth at the Second Baveno Consensus workshop as an interval of 48 hours from the time of admis236 GASTROINTESTINAL ENDOSCOPY
sion.9 Recurrent hematemesis after endoscopy, persistent tachycardia (pulse . 100 beats/min) and/or hypotension (systolic blood pressure , than 90 mm Hg) for 2 consecutive hours, and transfusion for more than 4 units of blood within 6 hours to maintain blood pressure were regarded as signs of failure of primary hemostasis. Octreotide infusion and/or balloon tamponade were offered to control bleeding. Endoscopic therapy was repeated, using the same sclerosing agent, if the patient’s condition could be stabilized. Evidence of bleeding after the first 48 hours was classified as a recurrent bleeding episodes.8 Patients with recurrent bleeding were treated by injection therapy with the same sclerosant; otherwise they were treated conservatively. During the period of hospital stay, abdominal ultrasonography was offered, once a patient’s condition was stabilized, for the assessment of portal vein invasion by tumor. Complications related to endoscopic therapy were documented. All patients received sucralfate after endoscopic therapy. The study was approved by the Ethics Committee of the Chinese University of Hong Kong. Informed consent was obtained before recruitment for the study. Initial hemostasis, recurrent bleeding and mortality rates between the two treatment groups were compared by Pearson chi-squared test. Blood transfusion was compared by Mann-Whitney U test. Survival of patients was also analyzed by Kaplan-Meier method and compared by the log-rank test.
RESULTS Fifty patients were recruited in this study with 25 randomized to receive cyanoacrylate injection and 25 to receive STD injection. The two treatment groups were comparable with regard to sex; age; hemoglobin, serum bilirubin, and albumin levels; prothrombin time; Child-Pugh’s grading of liver disease; and size and appearances of varices (Table 1). The incidence of active variceal bleeding and of portal vein thrombosis in the two groups were not significantly different. Two patients in the cyanoacrylate group died shortly after admission before abdominal ultrasonography could be performed. In the cyanoacrylate group, a median of two doses (range one to four doses) of cyanoacrylate mixture were given. In the STD group, the median volume of injected sclerosant was 11 mL (range 4 to 25 mL). There were four patients (16%) in each group in whom treatment failed to achieve primary hemostasis (p 5 1.0). Two patients in the cyanoacrylate group and two in the STD group died during the index episode of bleeding. Blood transfusion given to the two groups of patients were comparable. Six patients (24%) in the cyanoacrylate group and four patients (16%) in the STD group developed recurrent bleeding while in the hospital (p 5 0.48). Within 30 days after the initial bleeding episode, recurrent hemorrhage was documented in seven VOLUME 47, NO. 3, 1998
Injection sclerotherapy for variceal bleeding with hepatocellular carcinoma
J Sung, W Yeo, R Suen, et al.
Table 1. Patient characteristics No. of patients Age (mean 6 SE, yr) M:F Underlying cause (%) HBV HCV Alcohol Hypovolemic shock (%) Haemoglobin (mean 6 SE, mg/dL) Bilirubin (mean 6 SE, umol/L) Albumin (mean 6 SE, gm/L) Prothrombin time (mean 6 SE, seconds) Child-Pugh grading (%) A B C Size of varices (%) Grade 1 Grade 2 Grade 3 Actively bleeding varices (%) Varices with SRH (%) Portal vein thrombosis (%) Positive Negative Not sure
Cyanoacrylate
STD
25 49.8 6 17.0 22:3
25 55.7 6 13.2 24:1
17 (68) 0 (0) 8 (32) 2 (8) 8.1 (0.5) 62.8 (18.7) 22.8 (1.2) 15.6 (0.8)
18 (72) 2 (8) 5 (10) 2 (8) 9.7 (0.4) 110.0 (24.6) 23.0 (1.1) 17.0 (1.0)
0 (0) 16 (64) 9 (36)
0 (0) 10 (40) 15 (60)
4 (16) 14 (56) 7 (28) 13 (52) 12 (48)
2 (8) 14 (56) 9 (36) 8 (32) 17 (68)
15 (60) 8 (32) 2 (8)
15 (60) 10 (40) 0 (0)
patients (28%) who received cyanoacrylate injection and five patients (20%) who received STD injection (p 5 0.50). All patients experienced chest pain after injection sclerotherapy. Fever was documented in three patients (12%) in the STD group but in none of those in the cyanoacrylate group. In this study, the median (range) survival in the cyanoacrylate group and the STD group were 16 days (1 to 485 days) and 13 days (1 to 407 days), respectively (p 5 0.62). In-hospital mortality was lower in the cyanoacrylate group (7 patients, 28%) than in the STD group (10 patients, 40%), but the difference was not statistically significant (p 5 0.37). The 30-day mortality rates were also similar (Table 2). Hepatic failure and uncontrolled bleeding were the major causes of death. The cumulative survival curves for the two treatment groups were not significantly different (Fig. 1) (log-rank test, p 5 0.92). Cumulative survival of patients was also plotted according to Child-Pugh’s grading (Fig. 2). Although there was no between-group difference in survival for patients with similar hepatic reserve (log-rank test, Child’s B: p 5 0.26; Child’s C: p 5 0.90), patients with Child-Pugh’s B disease had significantly better survival than those with Child-Pugh’s C within the same treatment group (p 5 0.001). Most patients with failure of primary hemostasis VOLUME 47, NO. 3, 1998
Figure 1. Cumulative survival of all patients treated by cyanoacrylate (CYA) or sodium tetradecyl sulphate (STD) injection.
Table 2. Outcome of patients treated with cyanoacrylate or STD injection Cyanoacrylate STD (n 5 25) (n 5 25) p Value Control of bleeding Initial hemostasis (%) Recurrent bleeding in hospital (%) Recurrent bleeding within 30-days (%) Blood transfusion (median units [range]) Mortality In-hospital mortality (%) 30-day mortality (%) Cause of death Hepatic decompensation Uncontrolled/recurrent bleeding Rupture of HCC Sepsis
21 (84) 6 (24)
21 (84) 4 (16)
1.00 0.48
7 (28)
5 (20)
0.50
3 (0-10)
2 (0-8)
0.10
7 (28) 15 (60)
10 (40) 14 (56)
9 5
9 4
0 1
1 0
0.37 0.77
and all patients who developed recurrent bleeding had portal vein invasion by the tumor (Table 3). DISCUSSION The prognosis of patients with advanced HCC complicated by variceal bleeding is dismal.10,11 Treatment of variceal hemorrhage in these patients, who usually have advanced liver disease, is very difficult. Transjugular intrahepatic portal-systemic shunt (TIPS) carries the potential risk of tumor embolization and surgical shunting is seldom justified in patients with unresectable tumors. Despite GASTROINTESTINAL ENDOSCOPY 237
J Sung, W Yeo, R Suen, et al.
Injection sclerotherapy for variceal bleeding with hepatocellular carcinoma
Table 3. Portal vein thrombosis (PVT) and recurrent bleeding Cyanoacrylate*
Total No. of patients Failure of primary hemostasis Recurrent bleeding during hospital stay Recurrent bleeding within 30 days
STD
PVT1
PVT2
PVT1
PVT2
15 2
8 2
15 3
10 1
6
0
4
0
7
0
5
0
*Two patients in the cyanoacrylate group did not have ultrasonography for the detection of portal vein thrombosis.
Figure 2. Cumulative survival of patients treated by cyanoacrylate (CYA) or sodium tetradecyl sulphate (STD) subclassified according to Child-Pugh grading.
the enormous amount of data on endoscopic therapy for variceal bleeding in cirrhotic patients, data on the treatment of variceal hemorrhage in patients with HCC are few. The mean survival of patients after variceal bleeding was reported as only 3 to 6 weeks.2-4 Patients usually died from uncontrolled hemorrhage, cachexia, and decompensated hepatic function. The success of primary hemostasis after endoscopy ranged from 27% to 91%,3-5 but recurrent bleeding is common.5 Eradication of esophageal varices often requires 2 to 4 weekly sessions of endoscopic therapy during which the patients must be hospitalized. Endoscopy-related complications such as chest pain, fever, esophageal ulceration, and aspiration are common. As the remaining life-span of these patients is very limited, by the time obliteration of varices is achieved, most of them would not enjoy the success of the therapy. In view of the disappointing results of endoscopic treatment in prolonging survival of patients, we attempted to use endoscopic sclerotherapy as a “oneshot” therapy for the control acute variceal bleeding to minimize patient discomfort and shorten hospital stay. Cyanoacrylate was chosen in this study as it appears to offer better initial hemostatic control of bleeding from gastroesophageal varices when compared with standard sclerosants.7,8,12 This was an open prospective randomized study in which the gastroenterologists/endoscopists were not blinded as to the treatment given because the same sclerosing agent was used if a patient required further endoscopic treatment. We have adopted the consensus criteria of initial hemostasis and recurrent bleeding 238 GASTROINTESTINAL ENDOSCOPY
proposed by the Second Baveno International Consensus Workshop so that our data can be compared with that of other studies in the future.9 All patients had unresectable HCC, and most had advanced liver disease (Child-Pugh grade B or above), large varices (grade 2 or above), and portal vein invasion by tumor. Success of initial hemostasis was achieved in 84% in both treatment groups in our study. This result compares favorably with that of Lo et al.4 It was slightly inferior to the results of Chen et al.,5 but, in their study, control of bleeding within 12 hours after endoscopic therapy was used as the definition of initial hemostasis. Cyanoacrylate injection did not achieve better control of acute bleeding than STD injection in our study. In fact, recurrent bleeding during hospital stay and within the first 30 days after the index bleeding episode was slightly more common in the cyanoacrylate group than in the STD group. This is probably because cyanoacrylate could only occlude the bleeding vessel(s) without inducing the inflammation and fibrosis in the esophageal wall that prevents further bleeding. The mortality rates for the two treatment groups were not significantly different from the other. The median survival of our patients was comparable with that in previous studies.2-4 This study has confirmed the importance of two prognostic factors: portal vein thrombosis and hepatic reserve. More than half of cases in which there was a failure to achieve primary hemostasis had portal vein invasion by tumor. Furthermore, all recurrent bleeding episodes, both in cyanoacrylate group and STD group, were reported in patients with portal vein thrombosis. On the other hand, although there was no difference in cumulative survival between the two treatment groups, patients with Child’s B liver disease had significantly better survival than patients with Child’s C liver disease. VOLUME 47, NO. 3, 1998
Injection sclerotherapy for variceal bleeding with hepatocellular carcinoma
The importance of liver reserve in relation to survival of patients with variceal hemorrhage has also been demonstrated in non-HCC patients with cirrhosis.13-14 In conclusion, cyanoacrylate did not improve the outcome of HCC patients with variceal hemorrhage compared with standard sclerotherapy. Patients with portal vein thrombosis and advanced liver disease have a higher risk of recurrent bleeding. Single-injection sclerotherapy might be sufficient for the palliation of variceal bleeding in patients suffering from unresectable HCC. Eradication of varices should only be offered to selected patients with compensated liver disease and no evidence of portal vein thrombosis.
6. 7.
8.
9.
REFERENCES 1. Ho J, Wu PC, Kung TM. An autopsy of hepatocellular carcinoma in Hong Kong. Pathology 1981;13:409-16. 2. Yeo W, Sung JY, Ward SC, Chung SCS, Lee WY, Li AKC, et al. A prospective study of upper gastrointestinal hemorrhage in patients with hepatocellular carcinoma. Dig Dis Sci 1995; 40:2516-21. 3. Ng WD, Chan YT, Ho KK, Kong CK. Injection sclerotherapy for bleeding esophageal varices in cirrhotic patients with hepatocellular carcinoma. Gastrointest Endosc 1989;35:6970. 4. Lo GH, Lin CY, Lai KH, Malik U, Ng WW, Lee FY, et al. Endoscopic injection sclerotherapy versus conservative treatment for patients with unresectable hepatocellular carcinoma and bleeding esophageal varices. Gastrointest Endosc 1991; 37:161-4. 5. Chen CY, Chang TT, Lin CY, Shin JS, Chem CY, Chi CH, et
10. 11. 12.
13. 14.
J Sung, W Yeo, R Suen, et al. al. Endoscopic variceal ligation versus conservative treatment for patients with hepatocellular carcinoma and bleeding esophageal varices. Gastrointest Endosc 1995;42:535-9. Soehendra N, Grimm H, Nam VC. N-butyl-2-cyanoacrylate: a supplement to endoscopic sclerotherapy. Endoscopy 1987;19: 221-4. Feretis C, Dimopoulos C, Benakis P, Kalliakmanis B, Apostolidis N. N-butyl-2-cyanoacrylate plus sclerotherapy versus sclerotherapy alone in the treatment of bleeding esophageal varices: a randomized prospective study. Endoscopy 1995;27: 355-7. Thakeb F, Salama Z, Salama H, Abdel Raouf T, Abdel Kader S, Abdel Hamid H. The value of combined use of N-butyl-2cyanoacrylate and ethanolamine oleate in the management of bleeding esophago-gastric varices. Endoscopy 1995;27:35864. Burroughs AK, Alexandrino P, Cales P, Fleig W, Grace N, Minoli G, et al. Sore points. A review of the points where there was disagreement at Baveno I and an attempt to reach consensus. In: De Franchis R, editor. Portal hypertension II. Proceeding of Second Baveno International Consensus Workshop on Definitions, Methodology and Therapeutic Strategies. London: Blackwell Science; 1966. Lai CL, Lam KC, Wong KP, Wu PC, Todd D. Clinical features of hepatocellular carcinoma: review of 211 patients in Hong Kong. Cancer 1981;47:2746-55. Shiu W, Dewar G, Leung N, Leung WT, Chan M, Tao M, et al. Hepatocellular carcinoma in Hong Kong: clinical study on 340 cases. Oncology 1990;47:241-5. Oho K, Iwao T, Suminio M, Toyonaga A, Tanikawa K. Ethanolamine oleate versus butyl cyanoacrylate for bleeding gastric varices: a non-randomized study. Endoscopy 1995;27: 349-54. Graham DY, Smith JL. The course of patients after variceal hemorrhage. Gastroenterology 1981;80:800-9. Sung JJY, Chung SCS, Lai CW, Chan FKL, Leung JWC, Yung MY, et al. Octreotide infusion or emergency sclerotherapy for variceal hemorrhage. Lancet 1993;342:637-41.
Submission of FINAL manuscript on diskette Gastrointestinal Endoscopy strongly encourages the submission of final manuscripts on disk. Although files created with WordPerfect are preferred, please send your final manuscript in any electronic format. On your disk, please indicate computer system (e.g., IBM, MacIntosh) and word processing software used (e.g., WordPerfect 6.1).
VOLUME 47, NO. 3, 1998
GASTROINTESTINAL ENDOSCOPY 239
Gastrointestinal bleeding is frequently a terminal event in patients with hepatocellular carcinoma (HCC). Massive bleeding from gastroesophageal Received May 7, 1997. For revision September 22, 1997. Accepted November 4, 1997. From the Departments of Medicine, Clinical Oncology, and Surgery, The Chinese University of Hong Kong. Reprint requests: Joseph Sung, MD, Department of Medicine, Prince of Wales Hospital, Shatin, Hong Kong, China. Copyright © 1998 by the American Society for Gastrointestinal Endoscopy 0016-5107/98/$5.00 1 0 37/1/87366 VOLUME 47, NO. 3, 1998
varices has been reported as the immediate cause of death in about 20% of patients,1 and gastroesophageal varices account for half of the bleeding episodes.2 Life expectancy after bleeding was very limited, even with endoscopic sclerotherapy2-4 or variceal ligation.5 Compared with conservative treatment, endoscopic injection sclerotherapy has not been shown to prolong patient survival.3,4 Tumor invasion of the portal venous system is known to be associated with a higher risk of recurrent bleeding and mortality.2-4 Hepatic reserve is another prognostic factor determining the survival of patients.2,4 GASTROINTESTINAL ENDOSCOPY 235
J Sung, W Yeo, R Suen, et al.
Injection sclerotherapy for variceal bleeding with hepatocellular carcinoma
In view of the short life expectancy of HCC patients after an episode of variceal hemorrhage, we attempted to identify an endoscopic treatment with high success of initial hemostasis and low recurrent bleeding rate achieved by a single treatment to minimize discomfort and reduce hospital stay. N-butyl-2-cyanoacrylate (Histoacryl) was first used by Soehendra et al.6 in the treatment of large esophageal varices. This liquid tissue adhesive solidifies almost instantaneously on contact with blood. Recent randomized studies showed that nonHCC patients treated initially with cyanoacrylate injection during active bleeding and subsequently with conventional sclerosant have fewer postinjection episodes of bleeding7,8 and a better chance of survival.7 In this prospective randomized study, we compared the effects of endoscopic injection of cyanoacrylate against sodium tetradecyl sulphate (STD), after confirming bleeding from esophageal varices, in HCC patients whose course is complicated by variceal bleeding. PATIENTS AND METHODS Patients known to be suffering from unresectable HCC presenting with upper gastrointestinal bleeding were recruited for this study. All patients underwent endoscopic examination within 24 hours after admission. Patients with confirmed esophageal varices and no other possible source of bleeding were randomized to receive either cyanoacrylate or 1.5% STD injections. Signs of recent hemorrhage from varices (red color sign, fibrin clot) were documented during endoscopy. Exclusion criteria included (1) patients who were admitted in hepatic coma or moribund condition, (2) patients who had other possible source of bleeding in the upper gastrointestinal tract such as peptic ulcer or tumor invasion into stomach or duodenum, and (3) patients who had received endoscopic treatment or surgery for variceal bleeding in the previous 6 months. Endoscopic injection of the varices with cyanoacrylate or 1.5% STD was performed using a twin-channel therapeutic gastroscope (Olympus 2T-200; Olympus, Tokyo, Japan) and a flexible injection probe (Marcon-Harber injector; Wilson-Cook, Winston-Salem, N.C.). Cyanoacrylate was mixed in equal volume with Lipiodol (0.5 mL cyanoacrylate in 0.5 mL Lipiodol) before injection to prevent hardening of the glue in the injector and to allow x-ray monitoring of the injection. Both cyanoacrylate and STD were injected intravariceally into all identifiable columns of esophageal varices. Single injections of the cyanoacrylate mixture (1 mL per injection) were given into each variceal columns. STD injections (aliquots of 2 mL per injection) were injected into each column at multiple levels. Patients were carefully monitored for changes in blood pressure and pulse after undergoing endoscopic hemostasis. An acute bleeding episode was defined according to criteria set forth at the Second Baveno Consensus workshop as an interval of 48 hours from the time of admis236 GASTROINTESTINAL ENDOSCOPY
sion.9 Recurrent hematemesis after endoscopy, persistent tachycardia (pulse . 100 beats/min) and/or hypotension (systolic blood pressure , than 90 mm Hg) for 2 consecutive hours, and transfusion for more than 4 units of blood within 6 hours to maintain blood pressure were regarded as signs of failure of primary hemostasis. Octreotide infusion and/or balloon tamponade were offered to control bleeding. Endoscopic therapy was repeated, using the same sclerosing agent, if the patient’s condition could be stabilized. Evidence of bleeding after the first 48 hours was classified as a recurrent bleeding episodes.8 Patients with recurrent bleeding were treated by injection therapy with the same sclerosant; otherwise they were treated conservatively. During the period of hospital stay, abdominal ultrasonography was offered, once a patient’s condition was stabilized, for the assessment of portal vein invasion by tumor. Complications related to endoscopic therapy were documented. All patients received sucralfate after endoscopic therapy. The study was approved by the Ethics Committee of the Chinese University of Hong Kong. Informed consent was obtained before recruitment for the study. Initial hemostasis, recurrent bleeding and mortality rates between the two treatment groups were compared by Pearson chi-squared test. Blood transfusion was compared by Mann-Whitney U test. Survival of patients was also analyzed by Kaplan-Meier method and compared by the log-rank test.
RESULTS Fifty patients were recruited in this study with 25 randomized to receive cyanoacrylate injection and 25 to receive STD injection. The two treatment groups were comparable with regard to sex; age; hemoglobin, serum bilirubin, and albumin levels; prothrombin time; Child-Pugh’s grading of liver disease; and size and appearances of varices (Table 1). The incidence of active variceal bleeding and of portal vein thrombosis in the two groups were not significantly different. Two patients in the cyanoacrylate group died shortly after admission before abdominal ultrasonography could be performed. In the cyanoacrylate group, a median of two doses (range one to four doses) of cyanoacrylate mixture were given. In the STD group, the median volume of injected sclerosant was 11 mL (range 4 to 25 mL). There were four patients (16%) in each group in whom treatment failed to achieve primary hemostasis (p 5 1.0). Two patients in the cyanoacrylate group and two in the STD group died during the index episode of bleeding. Blood transfusion given to the two groups of patients were comparable. Six patients (24%) in the cyanoacrylate group and four patients (16%) in the STD group developed recurrent bleeding while in the hospital (p 5 0.48). Within 30 days after the initial bleeding episode, recurrent hemorrhage was documented in seven VOLUME 47, NO. 3, 1998
Injection sclerotherapy for variceal bleeding with hepatocellular carcinoma
J Sung, W Yeo, R Suen, et al.
Table 1. Patient characteristics No. of patients Age (mean 6 SE, yr) M:F Underlying cause (%) HBV HCV Alcohol Hypovolemic shock (%) Haemoglobin (mean 6 SE, mg/dL) Bilirubin (mean 6 SE, umol/L) Albumin (mean 6 SE, gm/L) Prothrombin time (mean 6 SE, seconds) Child-Pugh grading (%) A B C Size of varices (%) Grade 1 Grade 2 Grade 3 Actively bleeding varices (%) Varices with SRH (%) Portal vein thrombosis (%) Positive Negative Not sure
Cyanoacrylate
STD
25 49.8 6 17.0 22:3
25 55.7 6 13.2 24:1
17 (68) 0 (0) 8 (32) 2 (8) 8.1 (0.5) 62.8 (18.7) 22.8 (1.2) 15.6 (0.8)
18 (72) 2 (8) 5 (10) 2 (8) 9.7 (0.4) 110.0 (24.6) 23.0 (1.1) 17.0 (1.0)
0 (0) 16 (64) 9 (36)
0 (0) 10 (40) 15 (60)
4 (16) 14 (56) 7 (28) 13 (52) 12 (48)
2 (8) 14 (56) 9 (36) 8 (32) 17 (68)
15 (60) 8 (32) 2 (8)
15 (60) 10 (40) 0 (0)
patients (28%) who received cyanoacrylate injection and five patients (20%) who received STD injection (p 5 0.50). All patients experienced chest pain after injection sclerotherapy. Fever was documented in three patients (12%) in the STD group but in none of those in the cyanoacrylate group. In this study, the median (range) survival in the cyanoacrylate group and the STD group were 16 days (1 to 485 days) and 13 days (1 to 407 days), respectively (p 5 0.62). In-hospital mortality was lower in the cyanoacrylate group (7 patients, 28%) than in the STD group (10 patients, 40%), but the difference was not statistically significant (p 5 0.37). The 30-day mortality rates were also similar (Table 2). Hepatic failure and uncontrolled bleeding were the major causes of death. The cumulative survival curves for the two treatment groups were not significantly different (Fig. 1) (log-rank test, p 5 0.92). Cumulative survival of patients was also plotted according to Child-Pugh’s grading (Fig. 2). Although there was no between-group difference in survival for patients with similar hepatic reserve (log-rank test, Child’s B: p 5 0.26; Child’s C: p 5 0.90), patients with Child-Pugh’s B disease had significantly better survival than those with Child-Pugh’s C within the same treatment group (p 5 0.001). Most patients with failure of primary hemostasis VOLUME 47, NO. 3, 1998
Figure 1. Cumulative survival of all patients treated by cyanoacrylate (CYA) or sodium tetradecyl sulphate (STD) injection.
Table 2. Outcome of patients treated with cyanoacrylate or STD injection Cyanoacrylate STD (n 5 25) (n 5 25) p Value Control of bleeding Initial hemostasis (%) Recurrent bleeding in hospital (%) Recurrent bleeding within 30-days (%) Blood transfusion (median units [range]) Mortality In-hospital mortality (%) 30-day mortality (%) Cause of death Hepatic decompensation Uncontrolled/recurrent bleeding Rupture of HCC Sepsis
21 (84) 6 (24)
21 (84) 4 (16)
1.00 0.48
7 (28)
5 (20)
0.50
3 (0-10)
2 (0-8)
0.10
7 (28) 15 (60)
10 (40) 14 (56)
9 5
9 4
0 1
1 0
0.37 0.77
and all patients who developed recurrent bleeding had portal vein invasion by the tumor (Table 3). DISCUSSION The prognosis of patients with advanced HCC complicated by variceal bleeding is dismal.10,11 Treatment of variceal hemorrhage in these patients, who usually have advanced liver disease, is very difficult. Transjugular intrahepatic portal-systemic shunt (TIPS) carries the potential risk of tumor embolization and surgical shunting is seldom justified in patients with unresectable tumors. Despite GASTROINTESTINAL ENDOSCOPY 237
J Sung, W Yeo, R Suen, et al.
Injection sclerotherapy for variceal bleeding with hepatocellular carcinoma
Table 3. Portal vein thrombosis (PVT) and recurrent bleeding Cyanoacrylate*
Total No. of patients Failure of primary hemostasis Recurrent bleeding during hospital stay Recurrent bleeding within 30 days
STD
PVT1
PVT2
PVT1
PVT2
15 2
8 2
15 3
10 1
6
0
4
0
7
0
5
0
*Two patients in the cyanoacrylate group did not have ultrasonography for the detection of portal vein thrombosis.
Figure 2. Cumulative survival of patients treated by cyanoacrylate (CYA) or sodium tetradecyl sulphate (STD) subclassified according to Child-Pugh grading.
the enormous amount of data on endoscopic therapy for variceal bleeding in cirrhotic patients, data on the treatment of variceal hemorrhage in patients with HCC are few. The mean survival of patients after variceal bleeding was reported as only 3 to 6 weeks.2-4 Patients usually died from uncontrolled hemorrhage, cachexia, and decompensated hepatic function. The success of primary hemostasis after endoscopy ranged from 27% to 91%,3-5 but recurrent bleeding is common.5 Eradication of esophageal varices often requires 2 to 4 weekly sessions of endoscopic therapy during which the patients must be hospitalized. Endoscopy-related complications such as chest pain, fever, esophageal ulceration, and aspiration are common. As the remaining life-span of these patients is very limited, by the time obliteration of varices is achieved, most of them would not enjoy the success of the therapy. In view of the disappointing results of endoscopic treatment in prolonging survival of patients, we attempted to use endoscopic sclerotherapy as a “oneshot” therapy for the control acute variceal bleeding to minimize patient discomfort and shorten hospital stay. Cyanoacrylate was chosen in this study as it appears to offer better initial hemostatic control of bleeding from gastroesophageal varices when compared with standard sclerosants.7,8,12 This was an open prospective randomized study in which the gastroenterologists/endoscopists were not blinded as to the treatment given because the same sclerosing agent was used if a patient required further endoscopic treatment. We have adopted the consensus criteria of initial hemostasis and recurrent bleeding 238 GASTROINTESTINAL ENDOSCOPY
proposed by the Second Baveno International Consensus Workshop so that our data can be compared with that of other studies in the future.9 All patients had unresectable HCC, and most had advanced liver disease (Child-Pugh grade B or above), large varices (grade 2 or above), and portal vein invasion by tumor. Success of initial hemostasis was achieved in 84% in both treatment groups in our study. This result compares favorably with that of Lo et al.4 It was slightly inferior to the results of Chen et al.,5 but, in their study, control of bleeding within 12 hours after endoscopic therapy was used as the definition of initial hemostasis. Cyanoacrylate injection did not achieve better control of acute bleeding than STD injection in our study. In fact, recurrent bleeding during hospital stay and within the first 30 days after the index bleeding episode was slightly more common in the cyanoacrylate group than in the STD group. This is probably because cyanoacrylate could only occlude the bleeding vessel(s) without inducing the inflammation and fibrosis in the esophageal wall that prevents further bleeding. The mortality rates for the two treatment groups were not significantly different from the other. The median survival of our patients was comparable with that in previous studies.2-4 This study has confirmed the importance of two prognostic factors: portal vein thrombosis and hepatic reserve. More than half of cases in which there was a failure to achieve primary hemostasis had portal vein invasion by tumor. Furthermore, all recurrent bleeding episodes, both in cyanoacrylate group and STD group, were reported in patients with portal vein thrombosis. On the other hand, although there was no difference in cumulative survival between the two treatment groups, patients with Child’s B liver disease had significantly better survival than patients with Child’s C liver disease. VOLUME 47, NO. 3, 1998
Injection sclerotherapy for variceal bleeding with hepatocellular carcinoma
The importance of liver reserve in relation to survival of patients with variceal hemorrhage has also been demonstrated in non-HCC patients with cirrhosis.13-14 In conclusion, cyanoacrylate did not improve the outcome of HCC patients with variceal hemorrhage compared with standard sclerotherapy. Patients with portal vein thrombosis and advanced liver disease have a higher risk of recurrent bleeding. Single-injection sclerotherapy might be sufficient for the palliation of variceal bleeding in patients suffering from unresectable HCC. Eradication of varices should only be offered to selected patients with compensated liver disease and no evidence of portal vein thrombosis.
6. 7.
8.
9.
REFERENCES 1. Ho J, Wu PC, Kung TM. An autopsy of hepatocellular carcinoma in Hong Kong. Pathology 1981;13:409-16. 2. Yeo W, Sung JY, Ward SC, Chung SCS, Lee WY, Li AKC, et al. A prospective study of upper gastrointestinal hemorrhage in patients with hepatocellular carcinoma. Dig Dis Sci 1995; 40:2516-21. 3. Ng WD, Chan YT, Ho KK, Kong CK. Injection sclerotherapy for bleeding esophageal varices in cirrhotic patients with hepatocellular carcinoma. Gastrointest Endosc 1989;35:6970. 4. Lo GH, Lin CY, Lai KH, Malik U, Ng WW, Lee FY, et al. Endoscopic injection sclerotherapy versus conservative treatment for patients with unresectable hepatocellular carcinoma and bleeding esophageal varices. Gastrointest Endosc 1991; 37:161-4. 5. Chen CY, Chang TT, Lin CY, Shin JS, Chem CY, Chi CH, et
10. 11. 12.
13. 14.
J Sung, W Yeo, R Suen, et al. al. Endoscopic variceal ligation versus conservative treatment for patients with hepatocellular carcinoma and bleeding esophageal varices. Gastrointest Endosc 1995;42:535-9. Soehendra N, Grimm H, Nam VC. N-butyl-2-cyanoacrylate: a supplement to endoscopic sclerotherapy. Endoscopy 1987;19: 221-4. Feretis C, Dimopoulos C, Benakis P, Kalliakmanis B, Apostolidis N. N-butyl-2-cyanoacrylate plus sclerotherapy versus sclerotherapy alone in the treatment of bleeding esophageal varices: a randomized prospective study. Endoscopy 1995;27: 355-7. Thakeb F, Salama Z, Salama H, Abdel Raouf T, Abdel Kader S, Abdel Hamid H. The value of combined use of N-butyl-2cyanoacrylate and ethanolamine oleate in the management of bleeding esophago-gastric varices. Endoscopy 1995;27:35864. Burroughs AK, Alexandrino P, Cales P, Fleig W, Grace N, Minoli G, et al. Sore points. A review of the points where there was disagreement at Baveno I and an attempt to reach consensus. In: De Franchis R, editor. Portal hypertension II. Proceeding of Second Baveno International Consensus Workshop on Definitions, Methodology and Therapeutic Strategies. London: Blackwell Science; 1966. Lai CL, Lam KC, Wong KP, Wu PC, Todd D. Clinical features of hepatocellular carcinoma: review of 211 patients in Hong Kong. Cancer 1981;47:2746-55. Shiu W, Dewar G, Leung N, Leung WT, Chan M, Tao M, et al. Hepatocellular carcinoma in Hong Kong: clinical study on 340 cases. Oncology 1990;47:241-5. Oho K, Iwao T, Suminio M, Toyonaga A, Tanikawa K. Ethanolamine oleate versus butyl cyanoacrylate for bleeding gastric varices: a non-randomized study. Endoscopy 1995;27: 349-54. Graham DY, Smith JL. The course of patients after variceal hemorrhage. Gastroenterology 1981;80:800-9. Sung JJY, Chung SCS, Lai CW, Chan FKL, Leung JWC, Yung MY, et al. Octreotide infusion or emergency sclerotherapy for variceal hemorrhage. Lancet 1993;342:637-41.
Submission of FINAL manuscript on diskette Gastrointestinal Endoscopy strongly encourages the submission of final manuscripts on disk. Although files created with WordPerfect are preferred, please send your final manuscript in any electronic format. On your disk, please indicate computer system (e.g., IBM, MacIntosh) and word processing software used (e.g., WordPerfect 6.1).
VOLUME 47, NO. 3, 1998
GASTROINTESTINAL ENDOSCOPY 239