- Email: [email protected]
Endoscopic sclerotherapy plus propranolol versus propranolol alone in the primary prevention of bleeding in high risk cirrhotic patients with esophageal varices: A prospective multicenter randomized trial
Endoscopic sclerotherapy plus propranolol versus propranolol alone in the primary prevention of bleeding in high risk cirrhotic patients with esophageal varices: a prospective multicenter randomized trial Alec Avgerinos, MD, Anastasios Armonis, MD, Spilios Manolakopoulos, MD, George Rekoumis, MD, George Argirakis, MD, Nikolaos Viazis, MD, Jiannis Vlachogiannakos, MD, Adam Adamopoulos, MD, Themistoklis Kanaghinis, MD, Sotirios A. Raptis, MD Athens, Greece
Background: Analysis of primary prevention studies of the use of β-blockers has shown clear reductions in variceal bleeding in cirrhotic patients with varices. In contrast, the usefulness of prophylactic endoscopic sclerotherapy, alone or in combination with propranolol, in the management of these patients is still under investigation. The present study compared the efficacy of combined sclerotherapy and propranolol versus propranolol alone in the primary prevention of hemorrhage in cirrhotic patients with varices and high (greater than 18 mm Hg) intraesophageal variceal pressure. Methods: Patients were randomly assigned to propranolol (42 patients) or to propranolol plus sclerotherapy (44 patients). The mean duration of follow-up was 26.8 ± 7.7 and 24.6 ± 9.8 months, respectively. Results: During this period 23% of the patients in the combination group experienced at least 1 episode of bleeding due to varices or congestive gastropathy as compared with 14% in the propranolol group (not significant). Twenty-three patients (52%) in the combination group developed complications as compared with 8 (19%) in the propranolol group (p = 0.002). The mortality rate was similar in both groups (14% and 18%, respectively). The only independent factor predictive of survival was the level of serum albumin. Conclusions: Endoscopic sclerotherapy should not be used for the primary prevention of hemorrhage in cirrhotic patients at high risk of variceal bleeding who are undergoing treatment with propranolol. (Gastrointest Endosc 2000;51:652-8.)
Bleeding from esophageal varices is the major cause of death in patients with cirrhosis and portal hypertension. Approximately one third of patients with cirrhosis will experience at least 1 episode of variceal hemorrhage, usually within 1 year of the initial diagnosis of varices, and the initial episode will prove fatal in 30% to 50%.1-3 Consequently, the primary prevention of variceal hemorrhage is an important therapeutic goal. Various endoscopic and pharmacologic treatment options are available for prevention of the first Received April 30, 1999. For revision July 29, 1999. Accepted January 28, 2000. From the 2nd Department of Gastroenterology and 2nd Department of Internal Medicine, Propaedeutic, Athens University, Evangelismos Hospital and 3rd Department of Internal Medicine, Athens University, Athens, Greece. Supported by a grant (A. A.) from the Helenic Society of Gastroenterology. Reprint requests: Alec Avgerinos, MD, 2nd Department of Gastroenterology, Evangelismos Hospital, 10 Kaplanon St., 10680 Athens, Greece fax: 301-723-3671. Copyright © 2000 by the American Society for Gastrointestinal Endoscopy 0016-5107/2000/$12.00 + 0 37/1/105983 doi:10.1067/mge.2000.105983 652
GASTROINTESTINAL ENDOSCOPY
episode of bleeding from esophageal varices.3-14 The meta-analysis of primary prevention studies of the use of β-blockers has shown benefit with clinical and statistical homogeneity as regards efficacy.4,5 Nevertheless, only about one third of patients who are good candidates for treatment with β-blockers have a satisfactory hemodynamic response to the drug.5,13,14 Isosorbide-5-mononitrate (I5M) has been shown to be as effective as propranolol in preventing bleeding and the combination of the 2 drugs has been shown to decrease hepatic venous pressure gradient (HPVG) more than propranolol alone.12,15 However, unlike propranolol, I5M does not improve survival when used for primary prophylaxis.12 There is a striking heterogeneity of results with respect to the effect of sclerotherapy on both the risk of bleeding and death, with some trials reporting a beneficial and others a harmful effect.5,6 The results of sclerotherapy have tended to be favorable in trials in which the bleeding rate has been high in untreated control patients and unfavorable in those where the baseline bleeding risk in this group was low. A positive effect in high risk patients would be consistent with the known efficacy of sclerotherapy VOLUME 51, NO. 6, 2000
Propranolol with and without endoscopic sclerotherapy and esophageal varices
for prevention of recurrent bleeding, whereas the harmful effect in low risk patients probably results from the fact that side effects and complications exceed the potential advantage. Endoscopic band ligation appears to be more effective and safer than sclerotherapy in the management of variceal bleeding.3,7,8 However, its efficacy in the primary prevention of bleeding from esophageal varices is still under investigation.9-11 The addition of propranolol to sclerotherapy has been shown to be clearly beneficial compared with β-blockers or sclerotherapy alone for the secondary prevention of variceal bleeding.5,16 A possible effect of β-blockers might be a reduction in the risk of complications of sclerotherapy by allowing for use of a less intensive regimen. Evidence is accumulating to suggest that the height of portal and variceal pressure is a major predictive factor for variceal hemorrhage.17-21 Thus, there is a rationale for combining these 2 treatments in the primary prevention of variceal hemorrhage in cirrhotic patients with esophageal varices at high risk for bleeding. This study evaluated the efficacy of the combination of sclerotherapy and propranolol versus propranolol alone in the primary prevention of upper GI bleeding in cirrhotic patients with esophageal varices and high intravariceal pressure. PATIENTS AND METHODS The study design was randomized, prospective, unblinded, and multicenter. The protocol adhered to the principles of the Declaration of Helsinki and was approved by the ethics committee in each participating center. Patient enrollment From October 1991 to January 1995 all patients with cirrhosis who had never had upper GI bleeding ascribed to esophageal varices were consecutively offered upper endoscopy for assessment of esophageal varices. The diagnosis of cirrhosis was based on liver histology or the combination of clinical, biochemical and sonographic findings. As soon as possible after the screening endoscopy the patients were randomly allocated to 1 of 2 treatment groups if they met the following criteria: no previous treatment with β-blockers, endoscopic treatment or surgery for portal hypertension; no contraindication to the use of propranolol such as diabetes mellitus, heart failure or airway obstruction; endoscopic evidence of esophageal varices of medium or large size; intraesophageal variceal pressure (IVP) greater than 18 mm Hg; absence of hepatocellular carcinoma; written or verbal informed consent from the patient or next of kin. Intravariceal pressure measurement After premedication with midazolam, IVP was measured during esophagoscopy by direct puncture of the esophageal varix with a 25-gauge sclerotherapy needle VOLUME 51, NO. 6, 2000
A Avgerinos, A Armonis, S Manolakopoulos, et al.
connected to a pressure transducer. The transducer was attached to an amplifier and the pressure values were recorded by means of a computerized system (Gastrosoft; Synetics Inc, Stockholm, Sweden). Esophageal luminal pressure was recorded simultaneously via a second tube attached to the outer surface of the endoscope. IVP was defined as the gradient between intravariceal and intraesophageal pressure, the latter serving as the zero reference. The pressure was expressed in mm Hg and was the average of the values recorded at inspiration and expiration during a 30-second interval. Twelve patients developed oozing at the site of the puncture (5 were assigned to the group receiving propranolol only). In these 12 patients a maximum of 3 mL ethanolamine was injected immediately after withdrawal of the needle. In all but 1, bleeding stopped. In this patient (in the combination therapy group) 1 unit of blood was transfused. These bleeding episodes were considered complications of IVP measurement. Treatment and follow-up Patients fulfilling the above criteria were assigned to 1 of the 2 therapeutic groups, propranolol or sclerotherapy plus propranolol, by using opaque, serially numbered envelopes, within which the treatment assignment, made by means of a table of random numbers, was indicated. Propranolol was given orally at an initial dosage of 20 mg every 12 hours. Subsequently the dose was increased so as to reduce the resting pulse by approximately 25% but not below 55 beats/min or the systolic blood pressure below 90 mm Hg. Compliance was assessed at each visit by (1) measuring the resting pulse rate, (2) careful medication history taken from the patient and relatives, and (3) by checking for a discrepancy between the number of tablets prescribed and the patient’s request for further pills. Sclerotherapy was carried out with intravariceal injection of 5% ethanolamine oleate (2 to 3 mL per injection site, maximum 20 mL per session). The procedures were performed under midazolam sedation with the free hand technique at 2-week intervals until the varices disappeared or were too small for further injections. Endoscopic follow-up was performed every 3 months during the first year and every 6 months thereafter. Varices were treated as previously if they recurred. End point The primary end point was failure of therapy, defined as at least 1 bleeding episode during the follow-up period due to varices or congestive gastropathy. Death during this period was considered a secondary end point. Bleeding was defined as any episode of significant hematemesis and/or melena during the follow-up. Bleeding episodes were recorded by date, likely source and number of blood transfusions given. The bleeding episode was considered significant if at least 1 unit of blood transfusion was required to achieve hemodynamic stability. The diagnosis of variceal hemorrhage was based on endoscopic visualization of an GASTROINTESTINAL ENDOSCOPY
653
A Avgerinos, A Armonis, S Manolakopoulos, et al.
Propranolol with and without endoscopic sclerotherapy and esophageal varices
Table 1. Characteristics of therapeutic groups just before treatment Propranolol No. of patients Age (yr) Gender (m/f) Etiology of cirrhosis Posthepatic (hepatitis B virus–hepatitis C virus–hepatitis delta virus) Alcoholic Cryptogenic Primary biliary cirrhosis Child-Pugh grading A (5-8 points) B (9-11 points) C (12-15 points) Pugh score Variceal size Medium Large Red spots Fundus varices Ascites No Mild Severe Bilirubin Albumin Prothrombin time (sec) IVP (mm Hg)
actively bleeding varix, or varices with stigmata of recent bleeding and no other source of bleeding in the upper GI tract. In both groups, bleeding patients were treated with emergency sclerotherapy plus intravenous somatostatin. All bleeding patients were followed for 6 weeks to estimate bleeding-related mortality. Statistical analysis Statistical analysis was performed according to the intention to treat method. Comparisons between the 2 groups were carried out by using the Fisher exact test or chi-square test (with Yates correction) for qualitative data and Mann-Whitney test for quantitative variables. Bleeding and mortality were analyzed taking time into account, in relation to treatment group, gender, cause of cirrhosis, size of varices, red spots, ascites, age, variceal pressure, prothrombin time, albumin and bilirubin, first univariately and then multivariately by using Cox’s regression analysis. Kaplan-Meier estimates of bleeding and survival curves were computed. Termination of the study An interim analysis of the results of the study was performed as scheduled initially, by an independent reviewer, after a mean follow-up period of 2 years from the first enrollment. At this time, contrary to expectation, a tendency was noted for the propranolol group to fare better than the combination treatment group (p = 0.075), due to 654
GASTROINTESTINAL ENDOSCOPY
Propranolol plus sclerotherapy
42 56.3 ± 11.5 29/13
44 58.38 ± 12.5 34/10
22 15 4 1
19 22 3 —
25 11 6 7.8 ± 1.94
24 12 8 8.1 ± 2.11
29 13 14 7
26 18 13 5
23 12 7 2.81 ± 2.27 3.31 ± 0.66 17.05 ± 3.72 28.2 ± 6.2
16 15 13 2.8 ± 3.06 3.21 ± 0.44 16.33 ± 3.88 29.6 ± 6.9
a lower than expected bleeding rate in the propranolol group. A revised calculation of the sample size (80% power at p = 0.05), using bleeding rates similar to those observed, suggested that the trial should be continued until 29 events had been observed, according to the formula given by Parmar and Machin.22 However, a significant advantage for the propranolol group could only be demonstrated if all further events (beyond the 16 already observed) were in the propranolol group. It was concluded that it was highly improbable that the trial could demonstrate a statistically significant benefit of combination treatment. Consequently the decision was made to terminate the trial.
RESULTS From October 1991 to January 1995, 178 patients with liver cirrhosis and esophageal varices were examined; 92 were excluded for the following reasons: contraindication to propranolol (12 patients), IVP less than 18 mm Hg (35 patients), grade 1 esophageal varices (39 patients) and refusal to be randomized (6 patients). The remaining 86 patients were randomized to propranolol alone (42 patients) or propranolol in combination with sclerotherapy (44 patients). Patients in the 2 groups were comparable with regard to gender, cause of cirrhosis, Child-Pugh score, variceal size, variceal pressure, VOLUME 51, NO. 6, 2000
Propranolol with and without endoscopic sclerotherapy and esophageal varices
A Avgerinos, A Armonis, S Manolakopoulos, et al.
Figure 1. Patients free of bleeding versus time from randomization (p = 0.075); SCL, sclerotherapy.
Figure 2. Cumulative survival curves (p = 0.37); SCL, sclerotherapy.
presence of red spots and fundus varices, and biochemical parameters (Table 1).
Variceal bleeding
Variceal pressure Variceal pressure was significantly higher in patients with large varices than medium varices and in patients with red spots than without. No correlation was found between the height of IVP and the cause of cirrhosis or Child-Pugh score. The relationships between IVP and patient characteristics are presented in Table 2. Compliance and follow-up During the study period 4 patients (1 in the propranolol group) were lost from follow-up 1 to 4 months after randomization. Another 2 patients in the combination group refused further treatment with sclerotherapy but continued taking propranolol. The mean (± SD) overall duration of follow-up was 25 ± 9.3 months and was similar in both groups of patients (26.8 ± 7.7 and 24.6 ± 9.8 months in the sclerotherapy and combination groups, respectively). The mean dose of propranolol required was not significantly different between the sclerotherapy group (91.6 ± 31.1 mg) and combination group (88.1 ± 32.5 mg) (range 40 to 160 mg). Initial obliteration of varices was achieved in 38 of the 44 patients in 4.4 ± 2 sessions of sclerotherapy. Recurrence of esophageal varices after obliteration was observed in 17 patients; in 12 the recurrent varices were obliterated with a mean of 2 ± 1.8 further sessions of sclerotherapy. VOLUME 51, NO. 6, 2000
During the follow-up period 16 patients (6 in the propranolol group) had bleeding episodes requiring blood transfusions. The transfusion requirements during this period were similar in both groups (2.8 ± 2 units vs. 2.8 ± 1.4 units in the propranolol and combination therapy groups, respectively). Table 3 shows the sources of bleeding. The proportion of patients remaining free of bleeding over time is shown in Fig. 1. Cox regression analysis showed no significant differences with regard to bleeding between treatment groups. Univariate analysis of potential risk factors for bleeding are presented in Table 4; none were significant. Complications Eight patients (19%) in the propranolol group had treatment-related complications as compared with 23 (52%) in the combination group (p = 0.002) (Table 5). The most common complication due to sclerotherapy was retrosternal pain and fever (21 and 18, respectively). Complications directly attributable to propranolol were observed in 11 patients; in 6 the propranolol had to be discontinued. Survival Six patients (14%) in the propranolol group and eight (18%) in the combination group died during follow-up. The overall mortality was 16% and there were no differences in the causes of death (Table 6) or survival (Fig. 2) between the groups. Univariate GASTROINTESTINAL ENDOSCOPY
655
A Avgerinos, A Armonis, S Manolakopoulos, et al.
Propranolol with and without endoscopic sclerotherapy and esophageal varices
Table 2. Relationships between IVP and patient characteristics Patient characteristics
p Value
Age Gender Variceal size Red spots Fundus varices Ascites Bilirubin Albumin Prothrombin time
0.54 0.42 0.0015 0.002 0.59 0.25 0.63 0.92 0.61
Table 3. Sources of bleeding during follow-up
Esophageal varices Gastric varices Hypertensive gastropathy
Propranolol
Propranolol plus sclerotherapy
5 1 0
6 2 2
analysis of potential risk factors for death are shown in Table 4. Cox regression multivariate analysis showed that serum albumin was the only independent predictor of survival (p = 0.005). DISCUSSION This study has failed to show that primary endoscopic sclerotherapy reduces the incidence of the first bleeding episode in high risk cirrhotic patients with esophageal varices undergoing treatment with propranolol. In addition, combination treatment was associated with a significantly higher complication rate. The results of only 1 similar study23 have been published. This included 4 treatment groups (placebo, propranolol, sclerotherapy and combination therapy) and yielded evidence against the use of primary prophylaxis of variceal hemorrhage by endoscopic sclerotherapy alone, sclerotherapy in combination with propranolol, or with propranolol alone. However, the results of this trial have been questioned5,24 because it was performed in an unselected group of cirrhotic patients rather than high risk patients. Because only one third of patients with cirrhosis and varices are likely to bleed, invasive therapies or treatments with a high risk of serious complications in primary prophylaxis may be contraindicated in patients in whom the risk of bleeding is less than the risk of treatment-related complications. In the present study patients were selected who had a high risk for bleeding, i.e., patients with medium or large varices and variceal pressure greater than 18 mm Hg. Of note, 42% of the 656
Table 4. Univariate analysis of potential risk factors for bleeding and death
GASTROINTESTINAL ENDOSCOPY
Variables
Bleeding p value
Sclerotherapy Age Gender Variceal size Red spots Fundus varices Ascites Bilirubin Albumin Prothrombin time IVP
Death p value
0.075 0.61 0.42 0.32 0.51 0.41 0.11 0.13 0.21 0.29 0.61
0.37 0.49 0.33 1.0 0.31 0.88 0.23 0.014 0.0003 0.043 0.28
Table 5. Complications of treatment Complications Retrosternal pain Fever Esophageal ulcer Minor orthostatic symptoms Flapping tremor Cardiac failure No. of patients with complications
Propranolol
Propranolol plus sclerotherapy
— — — 4 2* 2* 8 (19%)
20 18 8 1 1* 1* 23 (52%)
*Patient stopped treatment because of complication.
Table 6. Causes of death during follow-up Propranolol Hepatic failure Postoperative death Related to bleeding Total
4 — 2 6 (14%)
Propranolol plus sclerotherapy 5 1 2 8 (18%)
patients considered for enrollment were excluded because they did not meet these criteria. The relationship between portal and variceal pressure, size of the varices, and risk of bleeding has been examined in several trials.17-21,25-28 Variceal pressure19,20 correlates with size and portal pressure (usually 2 mm Hg higher).29 In a well-designed study of bleeding varices by Ready et al.27 the lowest portal pressure (as estimated by the HVPG) associated with bleeding and recurrent bleeding was 16 mm Hg. At the time our study was initiated the only technique available in our departments for assessing portal pressure (indirectly) was measurement of IVP by direct puncture of the esophageal varix. We assumed that a high risk of variceal bleeding and benefit from sclerotherapy are directly related to the presence of a high variceal pressure VOLUME 51, NO. 6, 2000
Propranolol with and without endoscopic sclerotherapy and esophageal varices
(VP), i.e., greater than 18 mm Hg. This assumption has been indicated in 2 trials.17,24 In the first it was suggested that sclerotherapy reduced the incidence of first bleeding and prolonged survival in cirrhotic patients with large varices and HVPG greater than 16 mm Hg.24 In the other, in cirrhotic patients without a history of variceal bleeding, VP greater than 16 mm Hg was associated with the highest probability of bleeding during a follow-up period of 1 year.17 During a follow-up of 25 months, 14% of patients in the propranolol group experienced at least 1 episode of upper GI bleeding. This figure is compatible with other studies using β-blockers for the primary prevention of variceal bleeding.4,5 Although the mean dose of propranolol (90 mg per day) was lower than that reported previously by others,4,5 it was similar in both groups and similar to the dose reported in our previous trial.16 Compliance with treatment was also similar in both groups. It was necessary to inject a small volume of ethanolamine in 12 patients to control oozing from a varix at the site of needle puncture to measure IVP. It is unlikely that this affected outcome because a maximum of about 80 mL ethanolamine was required for the initial obliteration of varices in the combination therapy group. It seems paradoxical that more bleeding episodes occurred in the combination group (23%) as compared with the propranolol group (14%). This conflicts with the superior results of combination therapy for secondary prevention of esophageal variceal hemorrhage achieved in previous trials by our group21 and others.5 The reason for this discrepancy is unclear. Endoscopic treatment may have deleterious effects that counterbalance the benefit of reduced incidence of variceal bleeding, and in high risk patients chronic sclerotherapy may raise portal pressure,30 thereby increasing the incidence of bleeding from congestive gastropathy and gastric varices.31,32 Our follow-up data support this interpretation because 4 patients bled from gastric varices or hypertensive gastropathy in the combination treatment group as compared with only 1 in the propranolol group. Similar results were noted in our previous trial in which the addition of propranolol in patients undergoing sclerotherapy appeared mainly to reduce the incidence of recurrent bleeding from sources other than esophageal varices.16 Effective treatment with propranolol requires reduction of the portal pressure gradient by more than 20% of the baseline value and preferably under 12 mm Hg.5,13,14,28,33,34 Endoscopic measurement of VP appears to be a reliable alternative to HVPG measurements but has technical difficulties and VOLUME 51, NO. 6, 2000
A Avgerinos, A Armonis, S Manolakopoulos, et al.
limitations as regards interpretation in patients undergoing endoscopic treatment.33 Unfortunately, we did not measure the HVPG or IVP after treatment with propranolol. Although a smaller reduction in portal pressure in the combination treatment group is possible, the fact that the mean dose of propranolol was similar in both groups argues against this. Perhaps larger doses of propranolol are needed in the combination therapy group to counterbalance the rise in portal pressure induced by sclerotherapy. Kaplan-Meier estimates showed no difference in survival between the 2 treatment groups. The leading cause of death was hepatic failure. Mortality was found to be significantly associated with serum albumin,8,16,35 suggesting that hepatic failure is a more important prognostic indicator of survival than the type of treatment. In conclusion, the results of this trial do not support the use of endoscopic sclerotherapy for the primary prevention of variceal bleeding in high risk cirrhotic patients undergoing treatment with propranolol. Recent evidence suggests that endoscopic variceal ligation has advantages over sclerotherapy in the primary and secondary prevention of variceal bleeding.7-11 Whether endoscopic ligation proves to be of benefit in the primary prophylaxis of variceal bleeding in high risk cirrhotic patients undergoing treatment with propranolol is a matter for future research. ACKNOWLEDGMENT We thank Clive Richarson, PhD, for assistance with the statistical analysis of the data. REFERENCES 1. Goff JS. Esophageal varices. Gastrointest Endosc Clin N Am 1994;4:747-71. 2. Pinto HC, Abrantes A, Esteves AV, Almeida H, Correia JP. Long term prognosis of patients with cirrhosis of the liver and upper gastrointestinal bleeding. Am J Gastroenterol 1989;84: 1239-43. 3. Grace ND, Groszmann R, Garcia-Tsao G, Burroughs AK, Pagliaro L, Makuch RW, et al. Portal hypertension and variceal bleeding: an AASLD single topic symposium. Hepatology 1998;28:868-80. 4. Poynard T, Cales P, Pasta L, Ideo G, Pascal J-P, Pagliaro L, et al. Beta-adrenergic antagonist drugs in the prevention of gastrointestinal bleeding in patients with cirrhosis and esophageal varices: an analysis of data and prognostic factors in 589 patients from randomized clinical trials. N Engl J Med 1991;324:1532-8. 5. D’Amico G, Pagliaro L, Bosch J. The treatment of portal hypertension: a meta-analytic review. Hepatology 1995;22:332-54. 6. Fardy JM, Laupacis A. A meta-analysis of prophylactic endoscopic sclerotherapy for esophageal varices. Am J Gastroenterol 1994;89:1938-48. 7. Laine L, Cook D. Endoscopic ligation compared with sclerotherapy for treatment of esophageal variceal bleeding: a meta-analysis. Ann Intern Med 1995;123:280-7. GASTROINTESTINAL ENDOSCOPY
657
A Avgerinos, A Armonis, S Manolakopoulos, et al.
Propranolol with and without endoscopic sclerotherapy and esophageal varices
8. Avgerinos A, Armonis A, Manolakopoulos S, Poulianos G, Rekoumis G, Sgourou A, et al. Endoscopic sclerotherapy versus variceal ligation in the long term management of patients with cirrhosis after variceal bleeding: a prospective randomized trial. J Hepatol 1997;26:1034-41. 9. Lay CS, Tsai YT, Teg CY, Shyu WS, Guo WS, Wu KL, et al. Endoscopic variceal ligation in prophylaxis of first variceal bleeding in cirrhotic patients with high risk esophageal varices. Hepatology 1997;25:1346-50. 10. Sarin SK, Gurwant LS, Kumar M, Misra A, Nandagudi SM. Comparison of endoscopic ligation and propranolol for the primary prevention of variceal bleeding. N Engl J Med 1999;340:988-93. 11. Burroughs AK, Patch D. Primary prevention of bleeding from esophageal varices. N Engl J Med 1999;340:1033-5. 12. Angelico M, Carli L, Plat C, Capocaccia L. Effects of isosorbide-5-mononitrate compared with propranolol on first bleeding and long term survival in cirrhosis. Gastroenterology 1997;113:1632-9. 13. Garcia-Tsao G, Grace ND, Groszmann RJ, Conn HO, Bermann MM, Patrick MJC, et al. Short term effects of propranolol on portal venous pressure. Hepatology 1986;6:101-6. 14. Bosch J, Mastai R, Kravetz D, Bruix J, Gava I, Rigua I, et al. Effects of propranolol on azygos blood flow and hepatic systemic hemodynamics in cirrhosis. Hepatology 1984;4:1200-8. 15. Garcia-Pagan JC, Feu F, Bosch J, Rodes J. Propranolol compared with propranolol plus isosorbide-5-mononitrate for portal hypertension in cirrhosis: a randomized controlled study. Ann Intern Med 1991;114:869-873. 16. Avgerinos A, Rekoumis G, Klonis C, Papadimitriou N, Gouma P, Poumaras S, et al. Propranolol in the prevention of recurrent upper gastrointestinal bleeding in patients with cirrhosis undergoing endoscopic sclerotherapy: a randomized controlled trial. J Hepatol 1993;19:301-11. 17. Nevens F, Bustami R, Scheys I, Lesaffre E, Fevery J. Variceal pressure is a factor predicting the risk of a first variceal bleeding: a prospective cohort study in cirrhotic patients. Hepatology 1998;27:15-9. 18. Rigau J, Bosch J, Bordas JM, Navasa M, Mastai R, Kravetz MD, et al. Endoscopic measurement of variceal pressure in cirrhosis: correlation with portal pressure and variceal hemorrhage. Gastroenterology 1989;96:873-80. 19. Sarin SK, Sundaram KR, Ahuja PK. Predictors of variceal bleeding: an analysis of clinical, endoscopic, and haemodynamic variables, with special reference to intravariceal pressure. Gut 1989;30:1757-64. 20. Bandoh T, Mitarai Y, Kitani S, Kitano S, Yoshida T, Kobayashi M. Clinical signficance of esophageal variceal pressure in patients with esophageal varices. J Hepatol 1994;21:326-31. 21. Armonis A, Patch D, Burroughs A. Variceal and portal pressure measurement: techniques and applications. Ital J Gastroenterol 1996;28:272-9.
658
GASTROINTESTINAL ENDOSCOPY
22. Parmar MKB, Machin D. Survival analysis: a practical approach. Chichester: Wiley; 1995. p. 1-268. 23. The PROVA Study Group. Prophylaxis of first hemorrhage from esophageal varices by sclerotherapy, propranolol or both in cirrhotic patients: a randomized multicenter trial. Hepatology 1991;14:1016-24. 24. Paquet KJ, Kalk JF, Klein CP, Gad HA. Prophylactic sclerotherapy for esophageal varices in high-risk cirrhotic patients selected by endoscopic and hemodynamic criteria: a randomized, single-center controlled trial. Endoscopy 1994; 26:734-40. 25. The North Italian Endoscopic Club for the study and treatment of esophageal varices. Prediction of the first variceal hemorrhage in patients with cirrhosis of the liver and esophageal varices. N Engl J Med 1988;319:983-9. 26. Garcia-Tsao G, Groszmann RJ, Fisher RL, Conn HO, Atterbury CE, Glickman M. Portal pressure, presence of gastroesophageal varices and variceal bleeding. Hepatology 1985;5:419-24. 27. Ready JB, Robertson AD, Gow JS, Rector VVG. Assessment of the risk of bleeding from esophageal varices by continuous monitoring of portal pressure. Gastroenterology 1991;100: 1403-10. 28. Feu F, Garcia-Pagan JC, Bosch J, Luca A, Teres J, Escorsell A, et al. Relation between portal pressure response to pharmacotherapy and risk of recurrent variceal haemorrhage in patients with cirrhosis. Lancet 1995;346:1056-9. 29. Bosch J, Bordas JM, Mastai R, Kravetz D, Navasa M, Chesta J, et al. Effects of vasopressin on the intravariceal pressure in patients with cirrhosis: a comparison with the effects on portal pressure. Hepatology 1988;8:861-5. 30. Korula J, Rails P. The effects of chronic endoscopic variceal sclerotherapy on portal pressure in cirrhotics. Gastroenterology 1991;101:800-5. 31. Warren WD, Henderson JM, Millikan WJ, Galambos JT, Brooks WS, Riepe SP, et al. Distal splenorenal shunt versus endoscopic sclerotherapy for long-term management of variceal bleeding: preliminary report of a prospective, randomized trial. Ann Surg 1986;203:454-62. 32. Burroughs AK, McCormick PA. Variceal bleeding: acute and long-term management. Baillieres Clin Gastroenterol 1989;3: 131-63. 33. Bosch J. Medical treatment of portal hypertension. Digestion 1998;59:547-55. 34. Groszmann RJ, Bosh J, Grace N, Conn HO, Garcia-Tsao G, Navasa M, et al. Hemodynamic events in a prospective randomized trial of propranolol vs placebo in the prevention of the first variceal hemorrhage. Gastroenterology 1990;99: 1401-7. 35. Villanueva C, Balanzo J, Novella MT, Soriano G, Sainz S, Torras X, et al. Nadolol plus isosorbide mononitrate compared with sclerotherapy for the prevention of variceal rebleeding. N Engl J Med 1996;334:1624-9.
VOLUME 51, NO. 6, 2000
Background: Analysis of primary prevention studies of the use of β-blockers has shown clear reductions in variceal bleeding in cirrhotic patients with varices. In contrast, the usefulness of prophylactic endoscopic sclerotherapy, alone or in combination with propranolol, in the management of these patients is still under investigation. The present study compared the efficacy of combined sclerotherapy and propranolol versus propranolol alone in the primary prevention of hemorrhage in cirrhotic patients with varices and high (greater than 18 mm Hg) intraesophageal variceal pressure. Methods: Patients were randomly assigned to propranolol (42 patients) or to propranolol plus sclerotherapy (44 patients). The mean duration of follow-up was 26.8 ± 7.7 and 24.6 ± 9.8 months, respectively. Results: During this period 23% of the patients in the combination group experienced at least 1 episode of bleeding due to varices or congestive gastropathy as compared with 14% in the propranolol group (not significant). Twenty-three patients (52%) in the combination group developed complications as compared with 8 (19%) in the propranolol group (p = 0.002). The mortality rate was similar in both groups (14% and 18%, respectively). The only independent factor predictive of survival was the level of serum albumin. Conclusions: Endoscopic sclerotherapy should not be used for the primary prevention of hemorrhage in cirrhotic patients at high risk of variceal bleeding who are undergoing treatment with propranolol. (Gastrointest Endosc 2000;51:652-8.)
Bleeding from esophageal varices is the major cause of death in patients with cirrhosis and portal hypertension. Approximately one third of patients with cirrhosis will experience at least 1 episode of variceal hemorrhage, usually within 1 year of the initial diagnosis of varices, and the initial episode will prove fatal in 30% to 50%.1-3 Consequently, the primary prevention of variceal hemorrhage is an important therapeutic goal. Various endoscopic and pharmacologic treatment options are available for prevention of the first Received April 30, 1999. For revision July 29, 1999. Accepted January 28, 2000. From the 2nd Department of Gastroenterology and 2nd Department of Internal Medicine, Propaedeutic, Athens University, Evangelismos Hospital and 3rd Department of Internal Medicine, Athens University, Athens, Greece. Supported by a grant (A. A.) from the Helenic Society of Gastroenterology. Reprint requests: Alec Avgerinos, MD, 2nd Department of Gastroenterology, Evangelismos Hospital, 10 Kaplanon St., 10680 Athens, Greece fax: 301-723-3671. Copyright © 2000 by the American Society for Gastrointestinal Endoscopy 0016-5107/2000/$12.00 + 0 37/1/105983 doi:10.1067/mge.2000.105983 652
GASTROINTESTINAL ENDOSCOPY
episode of bleeding from esophageal varices.3-14 The meta-analysis of primary prevention studies of the use of β-blockers has shown benefit with clinical and statistical homogeneity as regards efficacy.4,5 Nevertheless, only about one third of patients who are good candidates for treatment with β-blockers have a satisfactory hemodynamic response to the drug.5,13,14 Isosorbide-5-mononitrate (I5M) has been shown to be as effective as propranolol in preventing bleeding and the combination of the 2 drugs has been shown to decrease hepatic venous pressure gradient (HPVG) more than propranolol alone.12,15 However, unlike propranolol, I5M does not improve survival when used for primary prophylaxis.12 There is a striking heterogeneity of results with respect to the effect of sclerotherapy on both the risk of bleeding and death, with some trials reporting a beneficial and others a harmful effect.5,6 The results of sclerotherapy have tended to be favorable in trials in which the bleeding rate has been high in untreated control patients and unfavorable in those where the baseline bleeding risk in this group was low. A positive effect in high risk patients would be consistent with the known efficacy of sclerotherapy VOLUME 51, NO. 6, 2000
Propranolol with and without endoscopic sclerotherapy and esophageal varices
for prevention of recurrent bleeding, whereas the harmful effect in low risk patients probably results from the fact that side effects and complications exceed the potential advantage. Endoscopic band ligation appears to be more effective and safer than sclerotherapy in the management of variceal bleeding.3,7,8 However, its efficacy in the primary prevention of bleeding from esophageal varices is still under investigation.9-11 The addition of propranolol to sclerotherapy has been shown to be clearly beneficial compared with β-blockers or sclerotherapy alone for the secondary prevention of variceal bleeding.5,16 A possible effect of β-blockers might be a reduction in the risk of complications of sclerotherapy by allowing for use of a less intensive regimen. Evidence is accumulating to suggest that the height of portal and variceal pressure is a major predictive factor for variceal hemorrhage.17-21 Thus, there is a rationale for combining these 2 treatments in the primary prevention of variceal hemorrhage in cirrhotic patients with esophageal varices at high risk for bleeding. This study evaluated the efficacy of the combination of sclerotherapy and propranolol versus propranolol alone in the primary prevention of upper GI bleeding in cirrhotic patients with esophageal varices and high intravariceal pressure. PATIENTS AND METHODS The study design was randomized, prospective, unblinded, and multicenter. The protocol adhered to the principles of the Declaration of Helsinki and was approved by the ethics committee in each participating center. Patient enrollment From October 1991 to January 1995 all patients with cirrhosis who had never had upper GI bleeding ascribed to esophageal varices were consecutively offered upper endoscopy for assessment of esophageal varices. The diagnosis of cirrhosis was based on liver histology or the combination of clinical, biochemical and sonographic findings. As soon as possible after the screening endoscopy the patients were randomly allocated to 1 of 2 treatment groups if they met the following criteria: no previous treatment with β-blockers, endoscopic treatment or surgery for portal hypertension; no contraindication to the use of propranolol such as diabetes mellitus, heart failure or airway obstruction; endoscopic evidence of esophageal varices of medium or large size; intraesophageal variceal pressure (IVP) greater than 18 mm Hg; absence of hepatocellular carcinoma; written or verbal informed consent from the patient or next of kin. Intravariceal pressure measurement After premedication with midazolam, IVP was measured during esophagoscopy by direct puncture of the esophageal varix with a 25-gauge sclerotherapy needle VOLUME 51, NO. 6, 2000
A Avgerinos, A Armonis, S Manolakopoulos, et al.
connected to a pressure transducer. The transducer was attached to an amplifier and the pressure values were recorded by means of a computerized system (Gastrosoft; Synetics Inc, Stockholm, Sweden). Esophageal luminal pressure was recorded simultaneously via a second tube attached to the outer surface of the endoscope. IVP was defined as the gradient between intravariceal and intraesophageal pressure, the latter serving as the zero reference. The pressure was expressed in mm Hg and was the average of the values recorded at inspiration and expiration during a 30-second interval. Twelve patients developed oozing at the site of the puncture (5 were assigned to the group receiving propranolol only). In these 12 patients a maximum of 3 mL ethanolamine was injected immediately after withdrawal of the needle. In all but 1, bleeding stopped. In this patient (in the combination therapy group) 1 unit of blood was transfused. These bleeding episodes were considered complications of IVP measurement. Treatment and follow-up Patients fulfilling the above criteria were assigned to 1 of the 2 therapeutic groups, propranolol or sclerotherapy plus propranolol, by using opaque, serially numbered envelopes, within which the treatment assignment, made by means of a table of random numbers, was indicated. Propranolol was given orally at an initial dosage of 20 mg every 12 hours. Subsequently the dose was increased so as to reduce the resting pulse by approximately 25% but not below 55 beats/min or the systolic blood pressure below 90 mm Hg. Compliance was assessed at each visit by (1) measuring the resting pulse rate, (2) careful medication history taken from the patient and relatives, and (3) by checking for a discrepancy between the number of tablets prescribed and the patient’s request for further pills. Sclerotherapy was carried out with intravariceal injection of 5% ethanolamine oleate (2 to 3 mL per injection site, maximum 20 mL per session). The procedures were performed under midazolam sedation with the free hand technique at 2-week intervals until the varices disappeared or were too small for further injections. Endoscopic follow-up was performed every 3 months during the first year and every 6 months thereafter. Varices were treated as previously if they recurred. End point The primary end point was failure of therapy, defined as at least 1 bleeding episode during the follow-up period due to varices or congestive gastropathy. Death during this period was considered a secondary end point. Bleeding was defined as any episode of significant hematemesis and/or melena during the follow-up. Bleeding episodes were recorded by date, likely source and number of blood transfusions given. The bleeding episode was considered significant if at least 1 unit of blood transfusion was required to achieve hemodynamic stability. The diagnosis of variceal hemorrhage was based on endoscopic visualization of an GASTROINTESTINAL ENDOSCOPY
653
A Avgerinos, A Armonis, S Manolakopoulos, et al.
Propranolol with and without endoscopic sclerotherapy and esophageal varices
Table 1. Characteristics of therapeutic groups just before treatment Propranolol No. of patients Age (yr) Gender (m/f) Etiology of cirrhosis Posthepatic (hepatitis B virus–hepatitis C virus–hepatitis delta virus) Alcoholic Cryptogenic Primary biliary cirrhosis Child-Pugh grading A (5-8 points) B (9-11 points) C (12-15 points) Pugh score Variceal size Medium Large Red spots Fundus varices Ascites No Mild Severe Bilirubin Albumin Prothrombin time (sec) IVP (mm Hg)
actively bleeding varix, or varices with stigmata of recent bleeding and no other source of bleeding in the upper GI tract. In both groups, bleeding patients were treated with emergency sclerotherapy plus intravenous somatostatin. All bleeding patients were followed for 6 weeks to estimate bleeding-related mortality. Statistical analysis Statistical analysis was performed according to the intention to treat method. Comparisons between the 2 groups were carried out by using the Fisher exact test or chi-square test (with Yates correction) for qualitative data and Mann-Whitney test for quantitative variables. Bleeding and mortality were analyzed taking time into account, in relation to treatment group, gender, cause of cirrhosis, size of varices, red spots, ascites, age, variceal pressure, prothrombin time, albumin and bilirubin, first univariately and then multivariately by using Cox’s regression analysis. Kaplan-Meier estimates of bleeding and survival curves were computed. Termination of the study An interim analysis of the results of the study was performed as scheduled initially, by an independent reviewer, after a mean follow-up period of 2 years from the first enrollment. At this time, contrary to expectation, a tendency was noted for the propranolol group to fare better than the combination treatment group (p = 0.075), due to 654
GASTROINTESTINAL ENDOSCOPY
Propranolol plus sclerotherapy
42 56.3 ± 11.5 29/13
44 58.38 ± 12.5 34/10
22 15 4 1
19 22 3 —
25 11 6 7.8 ± 1.94
24 12 8 8.1 ± 2.11
29 13 14 7
26 18 13 5
23 12 7 2.81 ± 2.27 3.31 ± 0.66 17.05 ± 3.72 28.2 ± 6.2
16 15 13 2.8 ± 3.06 3.21 ± 0.44 16.33 ± 3.88 29.6 ± 6.9
a lower than expected bleeding rate in the propranolol group. A revised calculation of the sample size (80% power at p = 0.05), using bleeding rates similar to those observed, suggested that the trial should be continued until 29 events had been observed, according to the formula given by Parmar and Machin.22 However, a significant advantage for the propranolol group could only be demonstrated if all further events (beyond the 16 already observed) were in the propranolol group. It was concluded that it was highly improbable that the trial could demonstrate a statistically significant benefit of combination treatment. Consequently the decision was made to terminate the trial.
RESULTS From October 1991 to January 1995, 178 patients with liver cirrhosis and esophageal varices were examined; 92 were excluded for the following reasons: contraindication to propranolol (12 patients), IVP less than 18 mm Hg (35 patients), grade 1 esophageal varices (39 patients) and refusal to be randomized (6 patients). The remaining 86 patients were randomized to propranolol alone (42 patients) or propranolol in combination with sclerotherapy (44 patients). Patients in the 2 groups were comparable with regard to gender, cause of cirrhosis, Child-Pugh score, variceal size, variceal pressure, VOLUME 51, NO. 6, 2000
Propranolol with and without endoscopic sclerotherapy and esophageal varices
A Avgerinos, A Armonis, S Manolakopoulos, et al.
Figure 1. Patients free of bleeding versus time from randomization (p = 0.075); SCL, sclerotherapy.
Figure 2. Cumulative survival curves (p = 0.37); SCL, sclerotherapy.
presence of red spots and fundus varices, and biochemical parameters (Table 1).
Variceal bleeding
Variceal pressure Variceal pressure was significantly higher in patients with large varices than medium varices and in patients with red spots than without. No correlation was found between the height of IVP and the cause of cirrhosis or Child-Pugh score. The relationships between IVP and patient characteristics are presented in Table 2. Compliance and follow-up During the study period 4 patients (1 in the propranolol group) were lost from follow-up 1 to 4 months after randomization. Another 2 patients in the combination group refused further treatment with sclerotherapy but continued taking propranolol. The mean (± SD) overall duration of follow-up was 25 ± 9.3 months and was similar in both groups of patients (26.8 ± 7.7 and 24.6 ± 9.8 months in the sclerotherapy and combination groups, respectively). The mean dose of propranolol required was not significantly different between the sclerotherapy group (91.6 ± 31.1 mg) and combination group (88.1 ± 32.5 mg) (range 40 to 160 mg). Initial obliteration of varices was achieved in 38 of the 44 patients in 4.4 ± 2 sessions of sclerotherapy. Recurrence of esophageal varices after obliteration was observed in 17 patients; in 12 the recurrent varices were obliterated with a mean of 2 ± 1.8 further sessions of sclerotherapy. VOLUME 51, NO. 6, 2000
During the follow-up period 16 patients (6 in the propranolol group) had bleeding episodes requiring blood transfusions. The transfusion requirements during this period were similar in both groups (2.8 ± 2 units vs. 2.8 ± 1.4 units in the propranolol and combination therapy groups, respectively). Table 3 shows the sources of bleeding. The proportion of patients remaining free of bleeding over time is shown in Fig. 1. Cox regression analysis showed no significant differences with regard to bleeding between treatment groups. Univariate analysis of potential risk factors for bleeding are presented in Table 4; none were significant. Complications Eight patients (19%) in the propranolol group had treatment-related complications as compared with 23 (52%) in the combination group (p = 0.002) (Table 5). The most common complication due to sclerotherapy was retrosternal pain and fever (21 and 18, respectively). Complications directly attributable to propranolol were observed in 11 patients; in 6 the propranolol had to be discontinued. Survival Six patients (14%) in the propranolol group and eight (18%) in the combination group died during follow-up. The overall mortality was 16% and there were no differences in the causes of death (Table 6) or survival (Fig. 2) between the groups. Univariate GASTROINTESTINAL ENDOSCOPY
655
A Avgerinos, A Armonis, S Manolakopoulos, et al.
Propranolol with and without endoscopic sclerotherapy and esophageal varices
Table 2. Relationships between IVP and patient characteristics Patient characteristics
p Value
Age Gender Variceal size Red spots Fundus varices Ascites Bilirubin Albumin Prothrombin time
0.54 0.42 0.0015 0.002 0.59 0.25 0.63 0.92 0.61
Table 3. Sources of bleeding during follow-up
Esophageal varices Gastric varices Hypertensive gastropathy
Propranolol
Propranolol plus sclerotherapy
5 1 0
6 2 2
analysis of potential risk factors for death are shown in Table 4. Cox regression multivariate analysis showed that serum albumin was the only independent predictor of survival (p = 0.005). DISCUSSION This study has failed to show that primary endoscopic sclerotherapy reduces the incidence of the first bleeding episode in high risk cirrhotic patients with esophageal varices undergoing treatment with propranolol. In addition, combination treatment was associated with a significantly higher complication rate. The results of only 1 similar study23 have been published. This included 4 treatment groups (placebo, propranolol, sclerotherapy and combination therapy) and yielded evidence against the use of primary prophylaxis of variceal hemorrhage by endoscopic sclerotherapy alone, sclerotherapy in combination with propranolol, or with propranolol alone. However, the results of this trial have been questioned5,24 because it was performed in an unselected group of cirrhotic patients rather than high risk patients. Because only one third of patients with cirrhosis and varices are likely to bleed, invasive therapies or treatments with a high risk of serious complications in primary prophylaxis may be contraindicated in patients in whom the risk of bleeding is less than the risk of treatment-related complications. In the present study patients were selected who had a high risk for bleeding, i.e., patients with medium or large varices and variceal pressure greater than 18 mm Hg. Of note, 42% of the 656
Table 4. Univariate analysis of potential risk factors for bleeding and death
GASTROINTESTINAL ENDOSCOPY
Variables
Bleeding p value
Sclerotherapy Age Gender Variceal size Red spots Fundus varices Ascites Bilirubin Albumin Prothrombin time IVP
Death p value
0.075 0.61 0.42 0.32 0.51 0.41 0.11 0.13 0.21 0.29 0.61
0.37 0.49 0.33 1.0 0.31 0.88 0.23 0.014 0.0003 0.043 0.28
Table 5. Complications of treatment Complications Retrosternal pain Fever Esophageal ulcer Minor orthostatic symptoms Flapping tremor Cardiac failure No. of patients with complications
Propranolol
Propranolol plus sclerotherapy
— — — 4 2* 2* 8 (19%)
20 18 8 1 1* 1* 23 (52%)
*Patient stopped treatment because of complication.
Table 6. Causes of death during follow-up Propranolol Hepatic failure Postoperative death Related to bleeding Total
4 — 2 6 (14%)
Propranolol plus sclerotherapy 5 1 2 8 (18%)
patients considered for enrollment were excluded because they did not meet these criteria. The relationship between portal and variceal pressure, size of the varices, and risk of bleeding has been examined in several trials.17-21,25-28 Variceal pressure19,20 correlates with size and portal pressure (usually 2 mm Hg higher).29 In a well-designed study of bleeding varices by Ready et al.27 the lowest portal pressure (as estimated by the HVPG) associated with bleeding and recurrent bleeding was 16 mm Hg. At the time our study was initiated the only technique available in our departments for assessing portal pressure (indirectly) was measurement of IVP by direct puncture of the esophageal varix. We assumed that a high risk of variceal bleeding and benefit from sclerotherapy are directly related to the presence of a high variceal pressure VOLUME 51, NO. 6, 2000
Propranolol with and without endoscopic sclerotherapy and esophageal varices
(VP), i.e., greater than 18 mm Hg. This assumption has been indicated in 2 trials.17,24 In the first it was suggested that sclerotherapy reduced the incidence of first bleeding and prolonged survival in cirrhotic patients with large varices and HVPG greater than 16 mm Hg.24 In the other, in cirrhotic patients without a history of variceal bleeding, VP greater than 16 mm Hg was associated with the highest probability of bleeding during a follow-up period of 1 year.17 During a follow-up of 25 months, 14% of patients in the propranolol group experienced at least 1 episode of upper GI bleeding. This figure is compatible with other studies using β-blockers for the primary prevention of variceal bleeding.4,5 Although the mean dose of propranolol (90 mg per day) was lower than that reported previously by others,4,5 it was similar in both groups and similar to the dose reported in our previous trial.16 Compliance with treatment was also similar in both groups. It was necessary to inject a small volume of ethanolamine in 12 patients to control oozing from a varix at the site of needle puncture to measure IVP. It is unlikely that this affected outcome because a maximum of about 80 mL ethanolamine was required for the initial obliteration of varices in the combination therapy group. It seems paradoxical that more bleeding episodes occurred in the combination group (23%) as compared with the propranolol group (14%). This conflicts with the superior results of combination therapy for secondary prevention of esophageal variceal hemorrhage achieved in previous trials by our group21 and others.5 The reason for this discrepancy is unclear. Endoscopic treatment may have deleterious effects that counterbalance the benefit of reduced incidence of variceal bleeding, and in high risk patients chronic sclerotherapy may raise portal pressure,30 thereby increasing the incidence of bleeding from congestive gastropathy and gastric varices.31,32 Our follow-up data support this interpretation because 4 patients bled from gastric varices or hypertensive gastropathy in the combination treatment group as compared with only 1 in the propranolol group. Similar results were noted in our previous trial in which the addition of propranolol in patients undergoing sclerotherapy appeared mainly to reduce the incidence of recurrent bleeding from sources other than esophageal varices.16 Effective treatment with propranolol requires reduction of the portal pressure gradient by more than 20% of the baseline value and preferably under 12 mm Hg.5,13,14,28,33,34 Endoscopic measurement of VP appears to be a reliable alternative to HVPG measurements but has technical difficulties and VOLUME 51, NO. 6, 2000
A Avgerinos, A Armonis, S Manolakopoulos, et al.
limitations as regards interpretation in patients undergoing endoscopic treatment.33 Unfortunately, we did not measure the HVPG or IVP after treatment with propranolol. Although a smaller reduction in portal pressure in the combination treatment group is possible, the fact that the mean dose of propranolol was similar in both groups argues against this. Perhaps larger doses of propranolol are needed in the combination therapy group to counterbalance the rise in portal pressure induced by sclerotherapy. Kaplan-Meier estimates showed no difference in survival between the 2 treatment groups. The leading cause of death was hepatic failure. Mortality was found to be significantly associated with serum albumin,8,16,35 suggesting that hepatic failure is a more important prognostic indicator of survival than the type of treatment. In conclusion, the results of this trial do not support the use of endoscopic sclerotherapy for the primary prevention of variceal bleeding in high risk cirrhotic patients undergoing treatment with propranolol. Recent evidence suggests that endoscopic variceal ligation has advantages over sclerotherapy in the primary and secondary prevention of variceal bleeding.7-11 Whether endoscopic ligation proves to be of benefit in the primary prophylaxis of variceal bleeding in high risk cirrhotic patients undergoing treatment with propranolol is a matter for future research. ACKNOWLEDGMENT We thank Clive Richarson, PhD, for assistance with the statistical analysis of the data. REFERENCES 1. Goff JS. Esophageal varices. Gastrointest Endosc Clin N Am 1994;4:747-71. 2. Pinto HC, Abrantes A, Esteves AV, Almeida H, Correia JP. Long term prognosis of patients with cirrhosis of the liver and upper gastrointestinal bleeding. Am J Gastroenterol 1989;84: 1239-43. 3. Grace ND, Groszmann R, Garcia-Tsao G, Burroughs AK, Pagliaro L, Makuch RW, et al. Portal hypertension and variceal bleeding: an AASLD single topic symposium. Hepatology 1998;28:868-80. 4. Poynard T, Cales P, Pasta L, Ideo G, Pascal J-P, Pagliaro L, et al. Beta-adrenergic antagonist drugs in the prevention of gastrointestinal bleeding in patients with cirrhosis and esophageal varices: an analysis of data and prognostic factors in 589 patients from randomized clinical trials. N Engl J Med 1991;324:1532-8. 5. D’Amico G, Pagliaro L, Bosch J. The treatment of portal hypertension: a meta-analytic review. Hepatology 1995;22:332-54. 6. Fardy JM, Laupacis A. A meta-analysis of prophylactic endoscopic sclerotherapy for esophageal varices. Am J Gastroenterol 1994;89:1938-48. 7. Laine L, Cook D. Endoscopic ligation compared with sclerotherapy for treatment of esophageal variceal bleeding: a meta-analysis. Ann Intern Med 1995;123:280-7. GASTROINTESTINAL ENDOSCOPY
657
A Avgerinos, A Armonis, S Manolakopoulos, et al.
Propranolol with and without endoscopic sclerotherapy and esophageal varices
8. Avgerinos A, Armonis A, Manolakopoulos S, Poulianos G, Rekoumis G, Sgourou A, et al. Endoscopic sclerotherapy versus variceal ligation in the long term management of patients with cirrhosis after variceal bleeding: a prospective randomized trial. J Hepatol 1997;26:1034-41. 9. Lay CS, Tsai YT, Teg CY, Shyu WS, Guo WS, Wu KL, et al. Endoscopic variceal ligation in prophylaxis of first variceal bleeding in cirrhotic patients with high risk esophageal varices. Hepatology 1997;25:1346-50. 10. Sarin SK, Gurwant LS, Kumar M, Misra A, Nandagudi SM. Comparison of endoscopic ligation and propranolol for the primary prevention of variceal bleeding. N Engl J Med 1999;340:988-93. 11. Burroughs AK, Patch D. Primary prevention of bleeding from esophageal varices. N Engl J Med 1999;340:1033-5. 12. Angelico M, Carli L, Plat C, Capocaccia L. Effects of isosorbide-5-mononitrate compared with propranolol on first bleeding and long term survival in cirrhosis. Gastroenterology 1997;113:1632-9. 13. Garcia-Tsao G, Grace ND, Groszmann RJ, Conn HO, Bermann MM, Patrick MJC, et al. Short term effects of propranolol on portal venous pressure. Hepatology 1986;6:101-6. 14. Bosch J, Mastai R, Kravetz D, Bruix J, Gava I, Rigua I, et al. Effects of propranolol on azygos blood flow and hepatic systemic hemodynamics in cirrhosis. Hepatology 1984;4:1200-8. 15. Garcia-Pagan JC, Feu F, Bosch J, Rodes J. Propranolol compared with propranolol plus isosorbide-5-mononitrate for portal hypertension in cirrhosis: a randomized controlled study. Ann Intern Med 1991;114:869-873. 16. Avgerinos A, Rekoumis G, Klonis C, Papadimitriou N, Gouma P, Poumaras S, et al. Propranolol in the prevention of recurrent upper gastrointestinal bleeding in patients with cirrhosis undergoing endoscopic sclerotherapy: a randomized controlled trial. J Hepatol 1993;19:301-11. 17. Nevens F, Bustami R, Scheys I, Lesaffre E, Fevery J. Variceal pressure is a factor predicting the risk of a first variceal bleeding: a prospective cohort study in cirrhotic patients. Hepatology 1998;27:15-9. 18. Rigau J, Bosch J, Bordas JM, Navasa M, Mastai R, Kravetz MD, et al. Endoscopic measurement of variceal pressure in cirrhosis: correlation with portal pressure and variceal hemorrhage. Gastroenterology 1989;96:873-80. 19. Sarin SK, Sundaram KR, Ahuja PK. Predictors of variceal bleeding: an analysis of clinical, endoscopic, and haemodynamic variables, with special reference to intravariceal pressure. Gut 1989;30:1757-64. 20. Bandoh T, Mitarai Y, Kitani S, Kitano S, Yoshida T, Kobayashi M. Clinical signficance of esophageal variceal pressure in patients with esophageal varices. J Hepatol 1994;21:326-31. 21. Armonis A, Patch D, Burroughs A. Variceal and portal pressure measurement: techniques and applications. Ital J Gastroenterol 1996;28:272-9.
658
GASTROINTESTINAL ENDOSCOPY
22. Parmar MKB, Machin D. Survival analysis: a practical approach. Chichester: Wiley; 1995. p. 1-268. 23. The PROVA Study Group. Prophylaxis of first hemorrhage from esophageal varices by sclerotherapy, propranolol or both in cirrhotic patients: a randomized multicenter trial. Hepatology 1991;14:1016-24. 24. Paquet KJ, Kalk JF, Klein CP, Gad HA. Prophylactic sclerotherapy for esophageal varices in high-risk cirrhotic patients selected by endoscopic and hemodynamic criteria: a randomized, single-center controlled trial. Endoscopy 1994; 26:734-40. 25. The North Italian Endoscopic Club for the study and treatment of esophageal varices. Prediction of the first variceal hemorrhage in patients with cirrhosis of the liver and esophageal varices. N Engl J Med 1988;319:983-9. 26. Garcia-Tsao G, Groszmann RJ, Fisher RL, Conn HO, Atterbury CE, Glickman M. Portal pressure, presence of gastroesophageal varices and variceal bleeding. Hepatology 1985;5:419-24. 27. Ready JB, Robertson AD, Gow JS, Rector VVG. Assessment of the risk of bleeding from esophageal varices by continuous monitoring of portal pressure. Gastroenterology 1991;100: 1403-10. 28. Feu F, Garcia-Pagan JC, Bosch J, Luca A, Teres J, Escorsell A, et al. Relation between portal pressure response to pharmacotherapy and risk of recurrent variceal haemorrhage in patients with cirrhosis. Lancet 1995;346:1056-9. 29. Bosch J, Bordas JM, Mastai R, Kravetz D, Navasa M, Chesta J, et al. Effects of vasopressin on the intravariceal pressure in patients with cirrhosis: a comparison with the effects on portal pressure. Hepatology 1988;8:861-5. 30. Korula J, Rails P. The effects of chronic endoscopic variceal sclerotherapy on portal pressure in cirrhotics. Gastroenterology 1991;101:800-5. 31. Warren WD, Henderson JM, Millikan WJ, Galambos JT, Brooks WS, Riepe SP, et al. Distal splenorenal shunt versus endoscopic sclerotherapy for long-term management of variceal bleeding: preliminary report of a prospective, randomized trial. Ann Surg 1986;203:454-62. 32. Burroughs AK, McCormick PA. Variceal bleeding: acute and long-term management. Baillieres Clin Gastroenterol 1989;3: 131-63. 33. Bosch J. Medical treatment of portal hypertension. Digestion 1998;59:547-55. 34. Groszmann RJ, Bosh J, Grace N, Conn HO, Garcia-Tsao G, Navasa M, et al. Hemodynamic events in a prospective randomized trial of propranolol vs placebo in the prevention of the first variceal hemorrhage. Gastroenterology 1990;99: 1401-7. 35. Villanueva C, Balanzo J, Novella MT, Soriano G, Sainz S, Torras X, et al. Nadolol plus isosorbide mononitrate compared with sclerotherapy for the prevention of variceal rebleeding. N Engl J Med 1996;334:1624-9.
VOLUME 51, NO. 6, 2000