- Email: [email protected]
Endoscopic sclerotherapy with fibrin glue as compared with polidocanol to prevent early esophageal variceal rebleeding
Journal of Hepatology 1998; 28: 292–297 Printed in Denmark ¡ All rights reserved Munksgaard ¡ Copenhagen
Copyright C European Association for the Study of the Liver 1998
Journal of Hepatology ISSN 0168-8278
Endoscopic sclerotherapy with fibrin glue as compared with polidocanol to prevent early esophageal variceal rebleeding Thomas Zimmer1, Frank Rucktäschel1, Ullrich Stölzel1, Ralf-Marco Liehr1, Detlef Schuppan1, Andreas Stallmach1, Martin Zeitz1, Ernst Weber2 and Ernst Otto Riecken1 1
Department of Gastroenterology, Universitätsklinikum Benjamin Franklin, Free University of Berlin 2and former Department of Biostatistics, Cancer Resarch, Heidelberg, Germany
Background/Aims: Endoscopic sclerotherapy is of proven benefit for patients after esophageal variceal bleeding, but is associated with substantial local and systemic complications. Since fibrin glue is a promising agent for endoscopic sclerotherapy of esophageal varices, we compared its safety and efficacy in patients after esophageal variceal bleeding. Patients and Methods: In a randomized, controlled trial, 36 patients with an acute episode of variceal bleeding were endoscopically treated with either polidocanol (18 patients) or fibrin glue (18 patients) by intravariceal injections within 12 h of admission. Tissue compatibility, incidence of various complications, episodes of rebleeding and overall survival rates were investigated. Results: Rebleeding, especially from enrollment to day 28, was less common in the fibrin group (pΩ0.046),
and all patients treated with fibrin glue survived for more than 28 days, whereas five patients treated with polidocanol died within this period. The incidence of sclerotherapy-induced ulcers was significantly lower in the fibrin group than in the polidocanol group (pΩ 0.001), and major complications such as perforation or ulcer bleeding were observed only in the polidocanol group. There were no complications in any group due to activation of systemic coagulation, fibrinolysis or clinically relevant pulmonary embolization. Conclusions: We conclude that fibrin glue is an efficient and safe agent for endoscopic sclerotherapy of bleeding esophageal varices, especially in the immediate posthemorrhagic period.
sclerotherapy is a treatment of proven benefit for acute esophageal variceal hemorrhage and for long-term prevention of recurrent bleeding (1– 5). Meta-analysis of several trials showed an improved long-term survival for patients subjected to endoscopic sclerotherapy when compared with pure medical management (6). However, the overall rate of significant complications ranges from 20 to 40% and treatmentrelated mortality may be as high as 12% (2,7). Most of these complications are due to the local tissue toxicity of the agent used for injection. This is illustrated by the high percentage of therapy-induced esophageal ul-
cers that occur in 56–87% of patients, depending on treatment modalities and the kind of sclerosant used (4–5, 8–11). Therefore, a sclerosant that is capable of inducing thrombosis and obliterating the varices without, or with significantly lower, tissue toxicity should be advantageous. TissucolA is a biological two-component fibrin sealant with high tissue compatibility, which is used to obtain effective hemostasis of bleeding gastric, duodenal and sclerotherapy-induced esophageal ulcers by endoscopic injection therapy (12,13). A preliminary study published in abstract form indicated that intravariceal injection of fibrin sealant may be a safe and efficient treatment of actively bleeding esophageal varices (14). However, the overall benefit of using conventional sclerotherapy compared with fibrin glue is not known. Here, we report a randomized controlled study comparing the tissue toxicity and the efficacy of intravariceal sclerotherapy with fibrin sealant compared to in-
E
Received 5 March; revised 25 August; accepted 2 September 1997
Correspondence: Thomas Zimmer, Department of Gastroenterology, Universitätsklinikum Benjamin Franklin, Free Universitity of Berlin, Hindenburgdamm 30, D-12200 Berlin, Germany. Tel: 0049-30-8445 2347. Fax: 0049-30-8445 4141.
292
Key words: Endoscopic sclerotherapy; Esophageal variceal bleeding; Fibrin glue; Polidocanol.
Endoscopic sclerotherapy with fibrin glue
travariceal sclerotherapy with polidocanol in patients with recent hemorrhage from esophageal varices referred to a single center.
Patients and Methods Selection and randomization of patients Patients with chronic liver disease and esophageal varices that were either bleeding actively or had bled acutely within the last 24 h were elegible for the study if they were 18 years or older. Patients were excluded if they had had endoscopic treatment for esophageal varices within the past 6 months, were unable or unwilling to sign an informed consent, had a contraindication to endoscopy (e.g. esophageal stenosis) or had another disease with death expected within the next 12 months. The study design was approved by the human research ethics committee of the study center (Benjamin Franklin Hospital). Informed consent was obtained from all patients or their nearest relatives before the initial endoscopy. Patients qualifying for the study were randomly assigned to receive sclerotherapy with polidocanol or fibrin glue at their initial endoscopy according to a computer-generated randomization sequence. Patients were randomized after endoscopically confirmed esophageal varices as the source of hemorrhage. Variceal hemorrhage was defined as bleeding from esophageal varices or as the presence of bleeding lesions on esophageal varices or simply as the presence of esophageal varices with blood in the upper gastrointestinal tract if other potential sources of bleeding were excluded. Of 68 patients screened from March 1991 to April 1994, 36 were enrolled in the study. In 30 patients, previous sclerotherapy (nΩ23) or reluctance to participate (nΩ7) were the reasons for exclusion. Two patients who met the inclusion criteria had to be excluded for the following reasons: death before the first endoscopic treatment was started or metastatic colon cancer. Eighteen patients were randomly assigned to treatment with polidocanol and 18 to treatment with fibrin glue. Treatment Sclerotherapy was performed with 1% polidocanol (AethoxysklerolA) or fibrin glue (TissucolA), injected with 22-gauge disposable needles. Intravariceal injections of 2–8 ml polidocanol (average 5.8 ml) or of 1–3 ml fibrin glue (average 2.1 ml) were applied to each varix depending on its size. Varices were graded from 1 to 4, according to the criteria used by Korula et al. (4). Injections were performed with proximal variceal blocking, using balloon scleroscopes (Olympus GIF-
SCL 2, Wolf RW-E 13). Before endoscopic sclerotherapy patients received premedication with 3–5 mg midazolam (DormicumA) and, if necessary, with 5–10 mg triflupromazin (PsyquilA) and 15–30 mg pentazocin (FortralA). Follow-up Patients underwent endoscopy, randomization and treatment within 12 h of admission. After enrollment, the time of randomization was taken as the starting point for evaluation. Treatment was initiated after spontaneous cessation of bleeding or after control of active bleeding with a Sengstaken-Blakemore tube. Treatment was repeated weekly until variceal obliteration was achieved. Variceal obliteration was assumed if endoscopic criteria such as induration or disappearance of varices were fulfilled. If extensive therapy-induced lesions, such as large ulcerations, were observed, treatment was postponed, with endoscopic follow-up being continued at weekly intervals. All patients with therapy-induced ulcers received omeprazole (AntraA) at a dose of 20 mg per day, up to 40 mg twice daily. After eradication of varices, patients underwent endoscopic control examinations after 3 weeks and then every 3 months, and for every episode of rebleeding. Recurrent varices were treated again with the originally assigned form of therapy. Monitoring included clinical parameters (pulse, blood pressure, temperature) and laboratory tests (peripheral blood and platelet counts, coagulation tests including plasma fibrinogen levels, partial thromboplastin time and thromboplastin time as well as tests of fibrinolysis, i.e. plasma levels of plasminogen, fibrin split products, thrombin-antithrombin complex and fibrin d-dimers). These parameters were determined before, 30 min, 60 min, 120 min and 24 h after each treatment session. For control of pulmonary embolic complications, lung perfusion scintigraphy and, if indicated by abnormal perfusion, ventilation scintigraphy were performed 30–60 min after sclerotherapy. Initial procedures were performed in an inpatient setting, but most subsequent treatments were provided on an outpatient basis. Patients were followed for 2 years. Only for survival analysis were patients followed for longer than 2 years. Rebleeding was defined as subsequent upper gastrointestinal hemorrhage that resulted in unscheduled endoscopy and a need for blood transfusions. Only bleeding episodes that occurred after the first 12 h after the initial bleeding were considered as rebleeding. Rebleedings were divided into those from esophageal varices or from other sources. Rebleeding from esophageal varices was defined as a visible bleeding from eso-
293
T. Zimmer et al.
phageal varices or as the presence of bleeding lesions on esophageal varices, or simply as the presence of esophageal varices with blood in the upper gastrointestinal tract if other potential sources of bleeding were excluded. Treatment failure was defined as two rebleeding episodes within the first 28 days after randomization, as death related to esophageal variceal bleeding, or to complications of treatment or to subsequent complications of bleeding. Any death that occurred within 28 days after a recurrent bleeding was defined as a bleeding-related death. Endpoints The primary endpoint of the study was the incidence of sclerotherapy-induced ulcers. This was chosen as the primary endpoint, because we did not anticipate substantial differences between the two treatments with regard to more traditional endpoints, such as rebleeding or survival. Secondary endpoints included the incidence of rebleeding, rebleeding within the time period from randomization to day 28 after randomization, the mortality, and the incidence of treatment-associated complications other than ulcers. Study termination and statistical analysis The size of the study sample was based on the primary end point, i.e. the incidence of sclerotherapy-induced ulcers. The assumed incidences of ulcers due to polidocanol and fibrin glue were 60% and 30%, respectively. The basic assumptions for the estimation of the sample size are alphaΩ5%, bΩ20% and kΩ30% with a onesided alternative. Based on this, a sample size of 40 patients for each group was calculated. Interim analyses were carried out in connection with the monitoring during progression of the trial that showed that the primary end-point among the two groups was significantly different at nΩ18 for each group with alphaΩ5%, bΩ5.38% (power of the testΩ 94.62%) and a reduction of the incidence of ulcers from 83% (polidocanol) to 28% (fibrin glue), yielding a k of 55%. For ethical reasons, we therefore decided to stop the trial in favor of fibrin glue treatment. Proportional data in a fourfold table were compared with the Fisher exact test (one-sided alternative). Proportional data in a 2¿ k-table were analyzed according to McCullagh (15). Quantitative data were described and compared using medians with 95% confidence interval. Kaplan-Meier life-time analysis was used to examine the time to death. The difference between the two groups was compared by log-rank test and by approximated acceleration. p-values less than 0.05 were considered significant.
294
Results The baseline characteristics of the two study groups showed no significant differences (Table 1). Active bleeding varices were present in four patients in the polidocanol group and in five patients in the fibrin group. Active bleeding in these patients was terminated by the application of a Sengstaken-Blakemore tube. For the 12 patients in the polidocanol group and the 17 patients in the fibrin group who remained in the study for more than 30 days, the mean duration of follow-up was 394∫304 and 460∫306 days, respectively. The probability of rebleeding from esophageal varices or therapy-induced ulcers was significantly lower in the group of patients treated with fibrin glue (pΩ0.046). Eleven of the patients who were treated with polidocanol showed a total of 19 episodes of rebleeding from esophageal varices or therapy-induced ulcers, compared with only five patients in the fibrin group, who showed seven episodes of rebleeding (Table 2). Four episodes of rebleeding occurred in three patients in the polidocanol group as a result of treatmentinduced ulcers, whereas bleeding due to iatrogenic ulcers was not observed in the fibrin group (Table 2). Most rebleedings in the polidocanol group occurred within 28 days after admission, whereas only one patient with two episodes of rebleeding was observed in the fibrin group. The chance of having at least one episode of rebleeding within 28 days after admission as expressed by odd’s ratio was 8 times greater for pa-
TABLE 1 Base-line characteristics of patients with esophageal variceal bleeding randomized to treatment by endoscopic sclerotherapy with polidocanol or fibrin glue Characteristic
Polidocanol (nΩ18)
Age (years) 59.4∫10.3 6/12 Sex (m/f). Alcoholic cirrhosis 12 Nonalcoholic cirrhosis 6 Child-Pugh class (no. of patients) A 8 B 7 C 3 Child-Pugh score 8.4∫3 Active bleeding (no. of patients) 4 Blood transfused for 3.4∫3.4 index episode (units) Grade of varices at index treatment (no. of patients) Grade 4 7 Grade 3 9 Grade 2 2 Red color signs (no. of patients) 10
Fibrin glue (nΩ18) 56.3∫9.9 10/8 12 6 9 6 3 7.6∫3.3 5 2.9∫4.2
6 10 2 9
Values are given as means∫SE. No differences between the two groups were statistically significant (. pΩ0.315).
Endoscopic sclerotherapy with fibrin glue TABLE 2 Rebleeding, tissue toxicity and minor complications in patients randomized to treatment by endoscopic sclerotherapy with polidocanol or fibrin glue Polidocanol (nΩ18) No. of patients with recurrent 11 bleeding from esophageal varices. No. of rebleedings 19 Esophageal varices 15 Treatment-induced ulcers 4 No. of patients with rebleeding 8 within 28 days.. No. of rebleedings within 28 days 12 No. of ulcerations/ 41/68 (60) no. of treatments (%)§ No. of patients with ulcerations/ 15/18 (83) no. of patients (%)§§ Blood/fresh frozen plasma 5.0∫2.1/ transfused for rebleedings (units) 4.4∫3.1 Minor complications (no. of complications/no. of treatments (%)) Fever 10/68 (15) Pleural effusion 2/68 (3) Dyspnea 1/68 (1.5) Dysphagia* 11/68 (16) Stricture ª Chest pain 1/68 (1.5)
Fibrin glue (nΩ18) 5 7 7 ª 1 2 6/112 (6) 5/18 (28) 3.9∫2.7/ 4.4∫2.9
11/112 3/112 2/112 1/112 ª 0/112
(10) (3) (2) (1) (0)
Values are given as means∫SE. . pΩ0.046, .. pΩ0.009, § p∞0.001; §§ pΩ0.001; *pΩ0.009 for the difference between groups. No other differences were statistically significant.
tients treated with polidocanol than for patients treated with fibrin glue (15). The severity of esophageal variceal rebleeding, as indicated by the mean number of units of blood transfused for each rebleeding, was similar in both groups (polidocanol 5.0∫2.1, fibrin 3.9∫2.7). However, due to the greater number of rebleeding episodes in the polidocanol group more units of blood and fresh frozen plasma had to be given (Table 2). The average volume per patient was 68 ml polidocanol and 43 ml fibrin glue. The extent of eradication of esophageal varices in patients remaining in the study for more than 28 days was not statistically different among the groups, but more treatments were necessary to obtain eradication in patients treated with fibrin (5.6∫2.8) as compared to patients treated with polidocanol (2.9∫1.6). We found a significant difference in the incidence of sclerotherapy-induced ulcers between the groups. Such ulcers occurred after 60% of treatments in the polidocanol group, compared to only 6% in the fibrin group. During sclerotherapy, 15 patients in the polidocanol group and five patients in the fibrin group developed at least one esophageal ulcer (pΩ0.001) (Table 2). Among major complications esophageal perforation was found in one patient and bleeding from therapyinduced ulcers in three patients in the polidocanol
group. The only major complication in the fibrin group was an abscess which formed in the esophageal wall in one patient, and which was treated successfully with antibiotics. Dysphagia lasting 1–5 days after sclerotherapy was found more frequently in the polidocanol group (pΩ 0.009) (Table 2). Other minor complications, such as fever, pleural effusion and dyspnea, were rarely observed in the two groups (Table 2). No relevant changes in laboratory parameters before and after sclerotherapy were found in the fibrin group or in the polidocanol group. Many patients in both groups already showed increased levels of thrombin-antithrombin complex and d-dimers before treatment. Lung perfusion scintigraphies were performed in all 18 patients treated with fibrin and in 15 of 18 patients treated with polidocanol. Lung perfusion scintigraphies were performed after 73 of 112 fibrin treatments and after 34 of 68 polidocanol treatments. After two treatment courses with fibrin, minor pulmonary embolizations with segmental perfusion defects but normal ventilation scans were found. There were, however, no signs of dyspnea, chest pain, hypotension or tachycardia, and scintigraphy showed a complete reperfusion after 24 and 48 h, respectively. In the polidocanol group, no specific perfusion disturbances were found. In the polidocanol group ten patients (55%) and in the fibrin group five patients (28%) died during an observation period of 394∫304 days and 460∫306 days, respectively This did not reach statistical significance, when Kaplan-Meier life-time analysis was performed. Half the deaths in the polidocanol group occurred within the first 28 days after entry into the study, whereas all patients in the fibrin group survived for more than 28 days after enrollment.
Discussion This is the first randomized, prospective trial comparing endoscopic sclerotherapy with fibrin glue with the widely-used sclerosant polidocanol in the prevention of rebleeding from esophageal varices. Two major findings merit discussion. First, recurrent bleeding, especially within the first 28 days after the first bleeding episode, was significantly less frequent in patients treated with fibrin glue when compared to patients treated with polidocanol. Second, fibrin-glue-induced ulcers were significantly less frequent compared with polidocanol-induced ulcers. The 61% incidence of recurrent bleeding in the polidocanol group is at the upper level of that reported in other studies, which showed rebleeding rates for sclero-
295
T. Zimmer et al.
therapy between 44 and 58% (2,9,10,16–19). Treatments were performed by four endoscopists who all had much experience. The small sample size and the relatively aggressive approach with short time intervals between treatments may explain this high rate of rebleeding. This may also be the reason for the relatively high number of major complications, with one perforation and three rebleedings from ulcers in the polidocanol group. As in other studies, more than 70% of all rebleeding episodes in our patients treated with polidocanol occurred within 30–40 days after the first bleeding (2,5,10,16–18). The lower risk of rebleeding for patients in the fibrin group cannot be explained by differences in predictive factors for variceal bleeding such as Child-Pugh score, variceal size or active bleeding at index endoscopy, which were comparable in the two study groups. We assume that the benefit of fibrin glue is due to immediate blockage of variceal blood flow in the distal part of the esophagus by intraluminal thrombosis. A significantly lower incidence of sclerotherapy-induced ulcers was observed in the fibrin group than in the polidocanol group. The few small ulcers found after injections of fibrin glue appeared to be caused by the elevated intramural pressure due to submucosal injection of fibrin glue. Healing of these ulcers occurred within a few days. Our data are consistent with the preliminary results of a non-randomized study published in abstract form which assessed the value of sclerotherapy with fibrin-glue to eradicate esophageal varices (14). In our study the high proportion of ulcers in the polidocanol group was comparable to that described in several previous reports (4,5,8,10,16,17,20,21). Ulcers after injection of polidocanol were larger and deeper than after injection of fibrin glue, and healed only after 3–12 weeks. In agreement with other investigators, we consider esophageal ulceration as an inevitable consequence of effective polidocanol sclerotherapy. Therefore most major and minor complications of sclerotherapy are related to local tissue injury and subsequent ulceration (1–3). Accordingly, as a result of the high tissue compatibility of fibrin glue, there was a clear trend to a reduction of major complications such as perforation or bleeding of sclerotherapy-induced ulcers in the fibrin group. Other complications of less significance, such as fever, pleural effusions, dyspnea and chest pain, were similar in both groups and comparable with those reported by other workers (4,8–10,18,22). Dysphagia has been reported to occur after 30–50% of treatments with polidocanol or related substances (17,20,21). In our study, dysphagia was rare after treatment with fibrin glue, but reached 16% in the polidocanol group. We assume that the lower rate of dysphagia is a good
296
clinical indicator of the high tissue compatibility of fibrin glue. Activation of coagulation and/or fibrinolysis as a consequence of intravariceal application of fibrin glue was particularly carefully explored, but did not play a role either clinically or in laboratory evaluations. Pulmonary embolization, another potential complication of intravariceal application of fibrin-glue that may be caused by fibrin clots entering the pulmonary circulation through the azygos venous system, was observed after two treatment episodes with fibrin glue. However, embolizations were subsegmental, self-limiting and of no clinical relevance. To our knowledge, these are the first controlled data on changes in coagulation, fibrinolysis and lung embolization after intravenous application of fibrin. Another risk of fibrin injections is the transmission of viral infections. TissucolA is produced from human plasma and every single donation has been tested for the presence of HBsAg, anti-HCV and anti-HIV1/2, and normal ALT values. In addition, every plasma pool used for fractionation is routinely screened for HIV-, HBV- and HCV-nucleic acids in quality-controlled gene amplification assay systems (IQ-PCR) (23). Thus state-of-the-art procedures for virus inactivation were performed (24). In the fibrin group more patients survived, compared to patients treated with polidocanol (72% vs. 45%). The risk of rebleeding and death is highest in the immediate posthemorrhagic period and decreases exponentially during the first few days after the initial hemorrhage (25). Deaths from exsanguination or related complications were more frequent in the polidocanol group. The lower risk for early rebleeding and the low tissue toxicity preventing major complications of sclerotherapy, such as bleeding from therapy-induced ulcers or perforation, are likely explanations for the higher rate of survival of the patients in the fibrin group within the first 28 days after index bleeding. Although our data reached statistical significance, these observations should be confirmed by further studies with more patients and a design primarily aimed to study early rebleeding rates and survival in the critical period of 1–4 weeks after initial hemorrhage. Endoscopic ligation of esophageal varices is a safer procedure than endoscopic sclerotherapy, and has replaced it for elective treatment in many centers (9,10,18). In future studies the advantages of fibrin injection have to be compared with those of endoscopic ligation. Thus a large randomized study comparing the two methods, fibrin injection and endoscopic ligation, in stopping esophageal variceal bleeding and preventing rebleeding is in preparation.
Endoscopic sclerotherapy with fibrin glue
We conclude that, because of high tissue compatibility and a low rate of major complications, sclerotherapy with fibrin is superior to sclerotherapy with tissue-damaging substances such as polidocanol. The trend to increased survival within the first few days after bleeding in patients treated with fibrin results from fewer complications and fewer recurrent bleedings. Therefore, fibrin glue is a promising and safe agent for sclerotherapy of bleeding esophageal varices, especially in the immediate posthemorrhagic period.
11.
12.
13.
14.
Acknowledgement The study was in part supported by Immuno GmbH (Heidelberg, FRG).
15.
16.
References 1. Westaby D, Hayes PC, Gimson AES, Polson RJ, Williams R. Controlled clinical trial of injection sclerotherapy for active variceal bleeding. Hepatology 1989; 9: 274–7. 2. The Copenhagen Esophageal Varices Sclerotherapy Project. Sclerotherapy after first variceal hemorrhage in cirrhosis. N Engl J Med 1984; 311: 1594–600. 3. Terblanche J, Kahn D, Campbell JAH, Bornman PC, Jonker MAT, Wright J. Failure of repeated injection sclerotherapy to improve long-term survival after oesophageal variceal bleeding. A 5-year prospective controlled clinical trial. Lancet 1983; ii: 1328–32. 4. Korula J, Balart LA, Radvan G, Zweiban BE, Larson AW, Kao HW, et al. A prospective, randomized controlled trial of chronic esophageal variceal sclerotherapy. Hepatology 1985; 5: 584–9. 5. Larson AW, Cohen H, Zweiban B, Chapman D, Gourdji M, Korula J, et al. Acute esophageal variceal sclerotherapy. Results of a prospective randomized controlled trial. JAMA 1986; 255: 497–500. 6. Infante-Rivard C, Esnaola S, Villeneuve J-P. Role of endoscopic variceal sclerotherapy in the long-term management of variceal bleeding: a meta-analysis. Gastroenterology 1989; 96: 1087–92. 7. Schumann BM, Beckmann JW, Tedesco FJ, Griffin JW, Assad RT. Complications of endoscopic injection sclerotherapy: a review. Am J Gastroenterol 1987; 82: 823–30. 8. Akriviadis E, Korula J, Gupta S, Ko Y, Yamada S. Frequent endoscopic variceal sclerotherapy increases risk of complications. Prospective randomized controlled study of two treatment schedules. Dig Dis Sci 1989; 34: 1068–74. 9. Gimson AES, Ramage JK, Panos MZ, Hayllar K, Harrison PM, Williams R, et al. Randomized trial of variceal banding ligation versus injection sclerotherapy for bleeding oesophageal varices. Lancet 1993; 342: 391–4. 10. Laine L, Hussein M E-N, Migikovsky B, Sloane R, Garcia
17.
18.
19.
20.
21.
22.
23.
24.
25.
F. Endoscopic ligation compared with sclerotherapy for the treatment of bleeding esophageal varices. Ann Intern Med 1993; 119: 1–7. Sarles HE, Sanowski RA, Talbert G. Course and complications of endoscopic variceal sclerotherapy: a prospective study of 50 patients. Am J Gastroenterol 1985; 80: 595–9. Berg PL, Barina W, Born P. Endoscopic injection of fibrin glue versus polidocanol in peptic ulcer hemorrhage: a pilot study. Endoscopy 1994; 26: 528–30. Prassler R, Reitz B, Barnert J, Richter G, Wienbeck M. Endoskopische Fibrintherapie bei blutenden Sklerosierungsulzera. Intensiv- und Notfallbehandlung 1994; 19: 25–7. Eimiller A, Berg P, Born P, Barina W, Paul F. Fibrinkleber als Sklerotherapeutikum (Abstract). Z Gastroenterol 1988; 26: 458. McCullagh P. Regressions models for ordinal data, generalized empirical logit transformation. J R Stat Soc B 1980; 42: 109–42. Macdougall BRD, Westaby D, Theodossi A, Dawson JL, Williams R. Increased long-term survival in variceal haemorrhage using injection sclerotherapy. Results of a controlled trial. Lancet 1982; ii: 124–7. Westaby D, Melia WM, MacDougall BRD, Hegarty JE, Williams R. Injection sclerotherapy for oesophageal varices: a prospective randomised trial of different treatment shedules. GUT 1984; 25: 129–32. Stiegmann GV, Goff JS, Michaletz-Onody PA, Korula J, Lieberman D, Saeed ZA, et al. Endoscopic sclerotherapy as compared with endoscopic ligation for bleeding esophageal varices. N Engl J Med 1992; 326: 1527–32. Soderlund C, Ihre T. Endoscopic sclerotherapy versus conservative management of bleeding esophageal varices. Acta Chir Scand 1985; 151: 449–56. Santangelo WC, Dueno MI, Estes BL, Krejs GJ. Prophylactic sclerotherapy of large esophageal varices. N Engl J Med 1988; 318: 814–18. The Veterans Affairs Cooperative Variceal Sclerotherapy Group. Prophylactic sclerotherapy for esophageal varices in men with alcoholic liver disease. A randomized, single-blind, multicenter clinical trial. N Engl J Med 1991; 324: 1779–84. Sauerbruch T, Wotzka R, Ko¨pcke W, Ha¨rlin M, Heldwein W, Bayerdo¨rffer E, et al. Prophylactic sclerotherapy before the first episode of variceal hemorrhage in patients with cirrhosis. N Engl J Med 1988; 319: 8–15. Ha¨mmerle T, Falkner F, Dorner F. A sensitive PCR assay system for the quantitation of viral genome equivalents: hepatitis C virus (HCV). Arch Virol 1996; 141: 2103–14. Rousou J, Gonzalez-Lavin L, Cosgrove D, Weldon C, Hess P, Joyce L, et al. Randomized clinical trial of fibrin sealant in patients undergoing resternotomy or reoperation after cardiac operations. J Thorac Cardiovasc Surg 1989; 97: 194– 203. Smith LJ, Graham DY. Variceal hemorrhage. A critical evaluation of survival analysis. Gastroenterology 1982; 82: 968–73.
297
Copyright C European Association for the Study of the Liver 1998
Journal of Hepatology ISSN 0168-8278
Endoscopic sclerotherapy with fibrin glue as compared with polidocanol to prevent early esophageal variceal rebleeding Thomas Zimmer1, Frank Rucktäschel1, Ullrich Stölzel1, Ralf-Marco Liehr1, Detlef Schuppan1, Andreas Stallmach1, Martin Zeitz1, Ernst Weber2 and Ernst Otto Riecken1 1
Department of Gastroenterology, Universitätsklinikum Benjamin Franklin, Free University of Berlin 2and former Department of Biostatistics, Cancer Resarch, Heidelberg, Germany
Background/Aims: Endoscopic sclerotherapy is of proven benefit for patients after esophageal variceal bleeding, but is associated with substantial local and systemic complications. Since fibrin glue is a promising agent for endoscopic sclerotherapy of esophageal varices, we compared its safety and efficacy in patients after esophageal variceal bleeding. Patients and Methods: In a randomized, controlled trial, 36 patients with an acute episode of variceal bleeding were endoscopically treated with either polidocanol (18 patients) or fibrin glue (18 patients) by intravariceal injections within 12 h of admission. Tissue compatibility, incidence of various complications, episodes of rebleeding and overall survival rates were investigated. Results: Rebleeding, especially from enrollment to day 28, was less common in the fibrin group (pΩ0.046),
and all patients treated with fibrin glue survived for more than 28 days, whereas five patients treated with polidocanol died within this period. The incidence of sclerotherapy-induced ulcers was significantly lower in the fibrin group than in the polidocanol group (pΩ 0.001), and major complications such as perforation or ulcer bleeding were observed only in the polidocanol group. There were no complications in any group due to activation of systemic coagulation, fibrinolysis or clinically relevant pulmonary embolization. Conclusions: We conclude that fibrin glue is an efficient and safe agent for endoscopic sclerotherapy of bleeding esophageal varices, especially in the immediate posthemorrhagic period.
sclerotherapy is a treatment of proven benefit for acute esophageal variceal hemorrhage and for long-term prevention of recurrent bleeding (1– 5). Meta-analysis of several trials showed an improved long-term survival for patients subjected to endoscopic sclerotherapy when compared with pure medical management (6). However, the overall rate of significant complications ranges from 20 to 40% and treatmentrelated mortality may be as high as 12% (2,7). Most of these complications are due to the local tissue toxicity of the agent used for injection. This is illustrated by the high percentage of therapy-induced esophageal ul-
cers that occur in 56–87% of patients, depending on treatment modalities and the kind of sclerosant used (4–5, 8–11). Therefore, a sclerosant that is capable of inducing thrombosis and obliterating the varices without, or with significantly lower, tissue toxicity should be advantageous. TissucolA is a biological two-component fibrin sealant with high tissue compatibility, which is used to obtain effective hemostasis of bleeding gastric, duodenal and sclerotherapy-induced esophageal ulcers by endoscopic injection therapy (12,13). A preliminary study published in abstract form indicated that intravariceal injection of fibrin sealant may be a safe and efficient treatment of actively bleeding esophageal varices (14). However, the overall benefit of using conventional sclerotherapy compared with fibrin glue is not known. Here, we report a randomized controlled study comparing the tissue toxicity and the efficacy of intravariceal sclerotherapy with fibrin sealant compared to in-
E
Received 5 March; revised 25 August; accepted 2 September 1997
Correspondence: Thomas Zimmer, Department of Gastroenterology, Universitätsklinikum Benjamin Franklin, Free Universitity of Berlin, Hindenburgdamm 30, D-12200 Berlin, Germany. Tel: 0049-30-8445 2347. Fax: 0049-30-8445 4141.
292
Key words: Endoscopic sclerotherapy; Esophageal variceal bleeding; Fibrin glue; Polidocanol.
Endoscopic sclerotherapy with fibrin glue
travariceal sclerotherapy with polidocanol in patients with recent hemorrhage from esophageal varices referred to a single center.
Patients and Methods Selection and randomization of patients Patients with chronic liver disease and esophageal varices that were either bleeding actively or had bled acutely within the last 24 h were elegible for the study if they were 18 years or older. Patients were excluded if they had had endoscopic treatment for esophageal varices within the past 6 months, were unable or unwilling to sign an informed consent, had a contraindication to endoscopy (e.g. esophageal stenosis) or had another disease with death expected within the next 12 months. The study design was approved by the human research ethics committee of the study center (Benjamin Franklin Hospital). Informed consent was obtained from all patients or their nearest relatives before the initial endoscopy. Patients qualifying for the study were randomly assigned to receive sclerotherapy with polidocanol or fibrin glue at their initial endoscopy according to a computer-generated randomization sequence. Patients were randomized after endoscopically confirmed esophageal varices as the source of hemorrhage. Variceal hemorrhage was defined as bleeding from esophageal varices or as the presence of bleeding lesions on esophageal varices or simply as the presence of esophageal varices with blood in the upper gastrointestinal tract if other potential sources of bleeding were excluded. Of 68 patients screened from March 1991 to April 1994, 36 were enrolled in the study. In 30 patients, previous sclerotherapy (nΩ23) or reluctance to participate (nΩ7) were the reasons for exclusion. Two patients who met the inclusion criteria had to be excluded for the following reasons: death before the first endoscopic treatment was started or metastatic colon cancer. Eighteen patients were randomly assigned to treatment with polidocanol and 18 to treatment with fibrin glue. Treatment Sclerotherapy was performed with 1% polidocanol (AethoxysklerolA) or fibrin glue (TissucolA), injected with 22-gauge disposable needles. Intravariceal injections of 2–8 ml polidocanol (average 5.8 ml) or of 1–3 ml fibrin glue (average 2.1 ml) were applied to each varix depending on its size. Varices were graded from 1 to 4, according to the criteria used by Korula et al. (4). Injections were performed with proximal variceal blocking, using balloon scleroscopes (Olympus GIF-
SCL 2, Wolf RW-E 13). Before endoscopic sclerotherapy patients received premedication with 3–5 mg midazolam (DormicumA) and, if necessary, with 5–10 mg triflupromazin (PsyquilA) and 15–30 mg pentazocin (FortralA). Follow-up Patients underwent endoscopy, randomization and treatment within 12 h of admission. After enrollment, the time of randomization was taken as the starting point for evaluation. Treatment was initiated after spontaneous cessation of bleeding or after control of active bleeding with a Sengstaken-Blakemore tube. Treatment was repeated weekly until variceal obliteration was achieved. Variceal obliteration was assumed if endoscopic criteria such as induration or disappearance of varices were fulfilled. If extensive therapy-induced lesions, such as large ulcerations, were observed, treatment was postponed, with endoscopic follow-up being continued at weekly intervals. All patients with therapy-induced ulcers received omeprazole (AntraA) at a dose of 20 mg per day, up to 40 mg twice daily. After eradication of varices, patients underwent endoscopic control examinations after 3 weeks and then every 3 months, and for every episode of rebleeding. Recurrent varices were treated again with the originally assigned form of therapy. Monitoring included clinical parameters (pulse, blood pressure, temperature) and laboratory tests (peripheral blood and platelet counts, coagulation tests including plasma fibrinogen levels, partial thromboplastin time and thromboplastin time as well as tests of fibrinolysis, i.e. plasma levels of plasminogen, fibrin split products, thrombin-antithrombin complex and fibrin d-dimers). These parameters were determined before, 30 min, 60 min, 120 min and 24 h after each treatment session. For control of pulmonary embolic complications, lung perfusion scintigraphy and, if indicated by abnormal perfusion, ventilation scintigraphy were performed 30–60 min after sclerotherapy. Initial procedures were performed in an inpatient setting, but most subsequent treatments were provided on an outpatient basis. Patients were followed for 2 years. Only for survival analysis were patients followed for longer than 2 years. Rebleeding was defined as subsequent upper gastrointestinal hemorrhage that resulted in unscheduled endoscopy and a need for blood transfusions. Only bleeding episodes that occurred after the first 12 h after the initial bleeding were considered as rebleeding. Rebleedings were divided into those from esophageal varices or from other sources. Rebleeding from esophageal varices was defined as a visible bleeding from eso-
293
T. Zimmer et al.
phageal varices or as the presence of bleeding lesions on esophageal varices, or simply as the presence of esophageal varices with blood in the upper gastrointestinal tract if other potential sources of bleeding were excluded. Treatment failure was defined as two rebleeding episodes within the first 28 days after randomization, as death related to esophageal variceal bleeding, or to complications of treatment or to subsequent complications of bleeding. Any death that occurred within 28 days after a recurrent bleeding was defined as a bleeding-related death. Endpoints The primary endpoint of the study was the incidence of sclerotherapy-induced ulcers. This was chosen as the primary endpoint, because we did not anticipate substantial differences between the two treatments with regard to more traditional endpoints, such as rebleeding or survival. Secondary endpoints included the incidence of rebleeding, rebleeding within the time period from randomization to day 28 after randomization, the mortality, and the incidence of treatment-associated complications other than ulcers. Study termination and statistical analysis The size of the study sample was based on the primary end point, i.e. the incidence of sclerotherapy-induced ulcers. The assumed incidences of ulcers due to polidocanol and fibrin glue were 60% and 30%, respectively. The basic assumptions for the estimation of the sample size are alphaΩ5%, bΩ20% and kΩ30% with a onesided alternative. Based on this, a sample size of 40 patients for each group was calculated. Interim analyses were carried out in connection with the monitoring during progression of the trial that showed that the primary end-point among the two groups was significantly different at nΩ18 for each group with alphaΩ5%, bΩ5.38% (power of the testΩ 94.62%) and a reduction of the incidence of ulcers from 83% (polidocanol) to 28% (fibrin glue), yielding a k of 55%. For ethical reasons, we therefore decided to stop the trial in favor of fibrin glue treatment. Proportional data in a fourfold table were compared with the Fisher exact test (one-sided alternative). Proportional data in a 2¿ k-table were analyzed according to McCullagh (15). Quantitative data were described and compared using medians with 95% confidence interval. Kaplan-Meier life-time analysis was used to examine the time to death. The difference between the two groups was compared by log-rank test and by approximated acceleration. p-values less than 0.05 were considered significant.
294
Results The baseline characteristics of the two study groups showed no significant differences (Table 1). Active bleeding varices were present in four patients in the polidocanol group and in five patients in the fibrin group. Active bleeding in these patients was terminated by the application of a Sengstaken-Blakemore tube. For the 12 patients in the polidocanol group and the 17 patients in the fibrin group who remained in the study for more than 30 days, the mean duration of follow-up was 394∫304 and 460∫306 days, respectively. The probability of rebleeding from esophageal varices or therapy-induced ulcers was significantly lower in the group of patients treated with fibrin glue (pΩ0.046). Eleven of the patients who were treated with polidocanol showed a total of 19 episodes of rebleeding from esophageal varices or therapy-induced ulcers, compared with only five patients in the fibrin group, who showed seven episodes of rebleeding (Table 2). Four episodes of rebleeding occurred in three patients in the polidocanol group as a result of treatmentinduced ulcers, whereas bleeding due to iatrogenic ulcers was not observed in the fibrin group (Table 2). Most rebleedings in the polidocanol group occurred within 28 days after admission, whereas only one patient with two episodes of rebleeding was observed in the fibrin group. The chance of having at least one episode of rebleeding within 28 days after admission as expressed by odd’s ratio was 8 times greater for pa-
TABLE 1 Base-line characteristics of patients with esophageal variceal bleeding randomized to treatment by endoscopic sclerotherapy with polidocanol or fibrin glue Characteristic
Polidocanol (nΩ18)
Age (years) 59.4∫10.3 6/12 Sex (m/f). Alcoholic cirrhosis 12 Nonalcoholic cirrhosis 6 Child-Pugh class (no. of patients) A 8 B 7 C 3 Child-Pugh score 8.4∫3 Active bleeding (no. of patients) 4 Blood transfused for 3.4∫3.4 index episode (units) Grade of varices at index treatment (no. of patients) Grade 4 7 Grade 3 9 Grade 2 2 Red color signs (no. of patients) 10
Fibrin glue (nΩ18) 56.3∫9.9 10/8 12 6 9 6 3 7.6∫3.3 5 2.9∫4.2
6 10 2 9
Values are given as means∫SE. No differences between the two groups were statistically significant (. pΩ0.315).
Endoscopic sclerotherapy with fibrin glue TABLE 2 Rebleeding, tissue toxicity and minor complications in patients randomized to treatment by endoscopic sclerotherapy with polidocanol or fibrin glue Polidocanol (nΩ18) No. of patients with recurrent 11 bleeding from esophageal varices. No. of rebleedings 19 Esophageal varices 15 Treatment-induced ulcers 4 No. of patients with rebleeding 8 within 28 days.. No. of rebleedings within 28 days 12 No. of ulcerations/ 41/68 (60) no. of treatments (%)§ No. of patients with ulcerations/ 15/18 (83) no. of patients (%)§§ Blood/fresh frozen plasma 5.0∫2.1/ transfused for rebleedings (units) 4.4∫3.1 Minor complications (no. of complications/no. of treatments (%)) Fever 10/68 (15) Pleural effusion 2/68 (3) Dyspnea 1/68 (1.5) Dysphagia* 11/68 (16) Stricture ª Chest pain 1/68 (1.5)
Fibrin glue (nΩ18) 5 7 7 ª 1 2 6/112 (6) 5/18 (28) 3.9∫2.7/ 4.4∫2.9
11/112 3/112 2/112 1/112 ª 0/112
(10) (3) (2) (1) (0)
Values are given as means∫SE. . pΩ0.046, .. pΩ0.009, § p∞0.001; §§ pΩ0.001; *pΩ0.009 for the difference between groups. No other differences were statistically significant.
tients treated with polidocanol than for patients treated with fibrin glue (15). The severity of esophageal variceal rebleeding, as indicated by the mean number of units of blood transfused for each rebleeding, was similar in both groups (polidocanol 5.0∫2.1, fibrin 3.9∫2.7). However, due to the greater number of rebleeding episodes in the polidocanol group more units of blood and fresh frozen plasma had to be given (Table 2). The average volume per patient was 68 ml polidocanol and 43 ml fibrin glue. The extent of eradication of esophageal varices in patients remaining in the study for more than 28 days was not statistically different among the groups, but more treatments were necessary to obtain eradication in patients treated with fibrin (5.6∫2.8) as compared to patients treated with polidocanol (2.9∫1.6). We found a significant difference in the incidence of sclerotherapy-induced ulcers between the groups. Such ulcers occurred after 60% of treatments in the polidocanol group, compared to only 6% in the fibrin group. During sclerotherapy, 15 patients in the polidocanol group and five patients in the fibrin group developed at least one esophageal ulcer (pΩ0.001) (Table 2). Among major complications esophageal perforation was found in one patient and bleeding from therapyinduced ulcers in three patients in the polidocanol
group. The only major complication in the fibrin group was an abscess which formed in the esophageal wall in one patient, and which was treated successfully with antibiotics. Dysphagia lasting 1–5 days after sclerotherapy was found more frequently in the polidocanol group (pΩ 0.009) (Table 2). Other minor complications, such as fever, pleural effusion and dyspnea, were rarely observed in the two groups (Table 2). No relevant changes in laboratory parameters before and after sclerotherapy were found in the fibrin group or in the polidocanol group. Many patients in both groups already showed increased levels of thrombin-antithrombin complex and d-dimers before treatment. Lung perfusion scintigraphies were performed in all 18 patients treated with fibrin and in 15 of 18 patients treated with polidocanol. Lung perfusion scintigraphies were performed after 73 of 112 fibrin treatments and after 34 of 68 polidocanol treatments. After two treatment courses with fibrin, minor pulmonary embolizations with segmental perfusion defects but normal ventilation scans were found. There were, however, no signs of dyspnea, chest pain, hypotension or tachycardia, and scintigraphy showed a complete reperfusion after 24 and 48 h, respectively. In the polidocanol group, no specific perfusion disturbances were found. In the polidocanol group ten patients (55%) and in the fibrin group five patients (28%) died during an observation period of 394∫304 days and 460∫306 days, respectively This did not reach statistical significance, when Kaplan-Meier life-time analysis was performed. Half the deaths in the polidocanol group occurred within the first 28 days after entry into the study, whereas all patients in the fibrin group survived for more than 28 days after enrollment.
Discussion This is the first randomized, prospective trial comparing endoscopic sclerotherapy with fibrin glue with the widely-used sclerosant polidocanol in the prevention of rebleeding from esophageal varices. Two major findings merit discussion. First, recurrent bleeding, especially within the first 28 days after the first bleeding episode, was significantly less frequent in patients treated with fibrin glue when compared to patients treated with polidocanol. Second, fibrin-glue-induced ulcers were significantly less frequent compared with polidocanol-induced ulcers. The 61% incidence of recurrent bleeding in the polidocanol group is at the upper level of that reported in other studies, which showed rebleeding rates for sclero-
295
T. Zimmer et al.
therapy between 44 and 58% (2,9,10,16–19). Treatments were performed by four endoscopists who all had much experience. The small sample size and the relatively aggressive approach with short time intervals between treatments may explain this high rate of rebleeding. This may also be the reason for the relatively high number of major complications, with one perforation and three rebleedings from ulcers in the polidocanol group. As in other studies, more than 70% of all rebleeding episodes in our patients treated with polidocanol occurred within 30–40 days after the first bleeding (2,5,10,16–18). The lower risk of rebleeding for patients in the fibrin group cannot be explained by differences in predictive factors for variceal bleeding such as Child-Pugh score, variceal size or active bleeding at index endoscopy, which were comparable in the two study groups. We assume that the benefit of fibrin glue is due to immediate blockage of variceal blood flow in the distal part of the esophagus by intraluminal thrombosis. A significantly lower incidence of sclerotherapy-induced ulcers was observed in the fibrin group than in the polidocanol group. The few small ulcers found after injections of fibrin glue appeared to be caused by the elevated intramural pressure due to submucosal injection of fibrin glue. Healing of these ulcers occurred within a few days. Our data are consistent with the preliminary results of a non-randomized study published in abstract form which assessed the value of sclerotherapy with fibrin-glue to eradicate esophageal varices (14). In our study the high proportion of ulcers in the polidocanol group was comparable to that described in several previous reports (4,5,8,10,16,17,20,21). Ulcers after injection of polidocanol were larger and deeper than after injection of fibrin glue, and healed only after 3–12 weeks. In agreement with other investigators, we consider esophageal ulceration as an inevitable consequence of effective polidocanol sclerotherapy. Therefore most major and minor complications of sclerotherapy are related to local tissue injury and subsequent ulceration (1–3). Accordingly, as a result of the high tissue compatibility of fibrin glue, there was a clear trend to a reduction of major complications such as perforation or bleeding of sclerotherapy-induced ulcers in the fibrin group. Other complications of less significance, such as fever, pleural effusions, dyspnea and chest pain, were similar in both groups and comparable with those reported by other workers (4,8–10,18,22). Dysphagia has been reported to occur after 30–50% of treatments with polidocanol or related substances (17,20,21). In our study, dysphagia was rare after treatment with fibrin glue, but reached 16% in the polidocanol group. We assume that the lower rate of dysphagia is a good
296
clinical indicator of the high tissue compatibility of fibrin glue. Activation of coagulation and/or fibrinolysis as a consequence of intravariceal application of fibrin glue was particularly carefully explored, but did not play a role either clinically or in laboratory evaluations. Pulmonary embolization, another potential complication of intravariceal application of fibrin-glue that may be caused by fibrin clots entering the pulmonary circulation through the azygos venous system, was observed after two treatment episodes with fibrin glue. However, embolizations were subsegmental, self-limiting and of no clinical relevance. To our knowledge, these are the first controlled data on changes in coagulation, fibrinolysis and lung embolization after intravenous application of fibrin. Another risk of fibrin injections is the transmission of viral infections. TissucolA is produced from human plasma and every single donation has been tested for the presence of HBsAg, anti-HCV and anti-HIV1/2, and normal ALT values. In addition, every plasma pool used for fractionation is routinely screened for HIV-, HBV- and HCV-nucleic acids in quality-controlled gene amplification assay systems (IQ-PCR) (23). Thus state-of-the-art procedures for virus inactivation were performed (24). In the fibrin group more patients survived, compared to patients treated with polidocanol (72% vs. 45%). The risk of rebleeding and death is highest in the immediate posthemorrhagic period and decreases exponentially during the first few days after the initial hemorrhage (25). Deaths from exsanguination or related complications were more frequent in the polidocanol group. The lower risk for early rebleeding and the low tissue toxicity preventing major complications of sclerotherapy, such as bleeding from therapy-induced ulcers or perforation, are likely explanations for the higher rate of survival of the patients in the fibrin group within the first 28 days after index bleeding. Although our data reached statistical significance, these observations should be confirmed by further studies with more patients and a design primarily aimed to study early rebleeding rates and survival in the critical period of 1–4 weeks after initial hemorrhage. Endoscopic ligation of esophageal varices is a safer procedure than endoscopic sclerotherapy, and has replaced it for elective treatment in many centers (9,10,18). In future studies the advantages of fibrin injection have to be compared with those of endoscopic ligation. Thus a large randomized study comparing the two methods, fibrin injection and endoscopic ligation, in stopping esophageal variceal bleeding and preventing rebleeding is in preparation.
Endoscopic sclerotherapy with fibrin glue
We conclude that, because of high tissue compatibility and a low rate of major complications, sclerotherapy with fibrin is superior to sclerotherapy with tissue-damaging substances such as polidocanol. The trend to increased survival within the first few days after bleeding in patients treated with fibrin results from fewer complications and fewer recurrent bleedings. Therefore, fibrin glue is a promising and safe agent for sclerotherapy of bleeding esophageal varices, especially in the immediate posthemorrhagic period.
11.
12.
13.
14.
Acknowledgement The study was in part supported by Immuno GmbH (Heidelberg, FRG).
15.
16.
References 1. Westaby D, Hayes PC, Gimson AES, Polson RJ, Williams R. Controlled clinical trial of injection sclerotherapy for active variceal bleeding. Hepatology 1989; 9: 274–7. 2. The Copenhagen Esophageal Varices Sclerotherapy Project. Sclerotherapy after first variceal hemorrhage in cirrhosis. N Engl J Med 1984; 311: 1594–600. 3. Terblanche J, Kahn D, Campbell JAH, Bornman PC, Jonker MAT, Wright J. Failure of repeated injection sclerotherapy to improve long-term survival after oesophageal variceal bleeding. A 5-year prospective controlled clinical trial. Lancet 1983; ii: 1328–32. 4. Korula J, Balart LA, Radvan G, Zweiban BE, Larson AW, Kao HW, et al. A prospective, randomized controlled trial of chronic esophageal variceal sclerotherapy. Hepatology 1985; 5: 584–9. 5. Larson AW, Cohen H, Zweiban B, Chapman D, Gourdji M, Korula J, et al. Acute esophageal variceal sclerotherapy. Results of a prospective randomized controlled trial. JAMA 1986; 255: 497–500. 6. Infante-Rivard C, Esnaola S, Villeneuve J-P. Role of endoscopic variceal sclerotherapy in the long-term management of variceal bleeding: a meta-analysis. Gastroenterology 1989; 96: 1087–92. 7. Schumann BM, Beckmann JW, Tedesco FJ, Griffin JW, Assad RT. Complications of endoscopic injection sclerotherapy: a review. Am J Gastroenterol 1987; 82: 823–30. 8. Akriviadis E, Korula J, Gupta S, Ko Y, Yamada S. Frequent endoscopic variceal sclerotherapy increases risk of complications. Prospective randomized controlled study of two treatment schedules. Dig Dis Sci 1989; 34: 1068–74. 9. Gimson AES, Ramage JK, Panos MZ, Hayllar K, Harrison PM, Williams R, et al. Randomized trial of variceal banding ligation versus injection sclerotherapy for bleeding oesophageal varices. Lancet 1993; 342: 391–4. 10. Laine L, Hussein M E-N, Migikovsky B, Sloane R, Garcia
17.
18.
19.
20.
21.
22.
23.
24.
25.
F. Endoscopic ligation compared with sclerotherapy for the treatment of bleeding esophageal varices. Ann Intern Med 1993; 119: 1–7. Sarles HE, Sanowski RA, Talbert G. Course and complications of endoscopic variceal sclerotherapy: a prospective study of 50 patients. Am J Gastroenterol 1985; 80: 595–9. Berg PL, Barina W, Born P. Endoscopic injection of fibrin glue versus polidocanol in peptic ulcer hemorrhage: a pilot study. Endoscopy 1994; 26: 528–30. Prassler R, Reitz B, Barnert J, Richter G, Wienbeck M. Endoskopische Fibrintherapie bei blutenden Sklerosierungsulzera. Intensiv- und Notfallbehandlung 1994; 19: 25–7. Eimiller A, Berg P, Born P, Barina W, Paul F. Fibrinkleber als Sklerotherapeutikum (Abstract). Z Gastroenterol 1988; 26: 458. McCullagh P. Regressions models for ordinal data, generalized empirical logit transformation. J R Stat Soc B 1980; 42: 109–42. Macdougall BRD, Westaby D, Theodossi A, Dawson JL, Williams R. Increased long-term survival in variceal haemorrhage using injection sclerotherapy. Results of a controlled trial. Lancet 1982; ii: 124–7. Westaby D, Melia WM, MacDougall BRD, Hegarty JE, Williams R. Injection sclerotherapy for oesophageal varices: a prospective randomised trial of different treatment shedules. GUT 1984; 25: 129–32. Stiegmann GV, Goff JS, Michaletz-Onody PA, Korula J, Lieberman D, Saeed ZA, et al. Endoscopic sclerotherapy as compared with endoscopic ligation for bleeding esophageal varices. N Engl J Med 1992; 326: 1527–32. Soderlund C, Ihre T. Endoscopic sclerotherapy versus conservative management of bleeding esophageal varices. Acta Chir Scand 1985; 151: 449–56. Santangelo WC, Dueno MI, Estes BL, Krejs GJ. Prophylactic sclerotherapy of large esophageal varices. N Engl J Med 1988; 318: 814–18. The Veterans Affairs Cooperative Variceal Sclerotherapy Group. Prophylactic sclerotherapy for esophageal varices in men with alcoholic liver disease. A randomized, single-blind, multicenter clinical trial. N Engl J Med 1991; 324: 1779–84. Sauerbruch T, Wotzka R, Ko¨pcke W, Ha¨rlin M, Heldwein W, Bayerdo¨rffer E, et al. Prophylactic sclerotherapy before the first episode of variceal hemorrhage in patients with cirrhosis. N Engl J Med 1988; 319: 8–15. Ha¨mmerle T, Falkner F, Dorner F. A sensitive PCR assay system for the quantitation of viral genome equivalents: hepatitis C virus (HCV). Arch Virol 1996; 141: 2103–14. Rousou J, Gonzalez-Lavin L, Cosgrove D, Weldon C, Hess P, Joyce L, et al. Randomized clinical trial of fibrin sealant in patients undergoing resternotomy or reoperation after cardiac operations. J Thorac Cardiovasc Surg 1989; 97: 194– 203. Smith LJ, Graham DY. Variceal hemorrhage. A critical evaluation of survival analysis. Gastroenterology 1982; 82: 968–73.
297