Variceal bleeding is a major cause of morbidity in patients

Transcription

1 GASTROENTEROLOGY 2010;139: Equal Efficacy of Endoscopic Variceal Ligation and Propranolol in Preventing Variceal Bleeding in Patients With Noncirrhotic Portal Hypertension SHIV KUMAR SARIN,*,, NITIN GUPTA, SANJEEV KUMAR JHA, AMIT AGRAWAL, SMRUTI RANJAN MISHRA, BARJESH CHANDER SHARMA, and ASHISH KUMAR*,, *Department of Hepatology, Institute of Liver and Biliary Sciences, New Delhi; Department of Gastroenterology, G B Pant Hospital, New Delhi; and Special Center for Molecular Medicine, Jawaharlal Nehru University, New Delhi, India BACKGROUND & AIMS: Variceal bleeding increases morbidity and mortality among patients with noncirrhotic portal hypertension (NCPH). Blockers of -adrenergic receptor signaling and endoscopic variceal ligation (EVL) have been used to prevent recurrence of bleeding, based on data from cirrhotic patients. We compared the efficacy and safety of the -blocker propranolol with that of EVL in preventing the recurrence of variceal bleeding in patients with NCPH. METHODS: Consecutive patients with NCPH with a history of variceal bleeding in the past 6 weeks were assigned randomly to groups treated every 3 weeks with EVL (n 51) or propranolol (until they had a resting heart rate of 55 beats per minute or to a maximum of 320 mg/day; n 50). Primary end points were recurrence of variceal bleeding or death. Secondary end points were complications of EVL in patients given EVL, variceal eradication after EVL, variceal recurrence after EVL, or a decrease in variceal grade in patients given propranolol. RESULTS: After a median follow-up period of 23 months, rates of recurrence of bleeding were similar between the groups (EVL, 23.5%; propranolol, 18%; P.625). The actuarial probability of remaining free of bleeding recurrence was similar between the groups. No deaths occurred in either group. Of the patients given propranolol, 47% had a decrease in the grade of varices and none experienced bleeding. Adverse events were minor and comparable between groups (EVL, 12%; propranolol, 18%; P.635). CONCLUSIONS: EVL was not more effective than the -blocker propranolol for the secondary prophylaxis of variceal bleeding in patients with NCPH. Keywords: Extrahepatic Portal Vein Obstruction; Portal Vein Thrombosis; Noncirrhotic Portal Fibrosis; -Blockers. Variceal bleeding is a major cause of morbidity in patients with noncirrhotic portal hypertension (NCPH), especially in developing countries where it accounts for almost 40% of all variceal bleeds. 1 Endoscopic variceal ligation (EVL) or endoscopic sclerotherapy is the first-line treatment for the control of active bleeding from esophageal varices in such patients. 2,3 However, up to 80% of these patients rebleed in the absence of secondary prophylaxis. 4 6 Hence, secondary prophylaxis of variceal bleeding is recommended. 3 However, there are limited data on the secondary prophylaxis of variceal bleeding in NCPH patients. Sclerotherapy has been well studied in preventing rebleed in patients with NCPH. Initial studies showed that sclerotherapy significantly reduced the rebleeding rate in patients with NCPH. 7 9 Data from these studies suggest a rebleed rate of approximately 25% at 2 years and 35% at 5 years. EVL has been reported to be as effective or even better than sclerotherapy with fewer complications and is the preferred endoscopic modality at present blockers have been found to be quite effective in both primary as well as secondary prophylaxis of variceal bleeding in cirrhotic patients and are the accepted mode of treatment. 13,14 In contrast to cirrhosis, published data on the effect of -blocker therapy on NCPH are scanty. As in patients with cirrhosis, patients with NCPH also have hyperdynamic circulation characterized by high cardiac index, low systemic vascular resistance, and increased azygous blood flow. 15 Increased sympathetic activity in these patients is thought to play a role in the development of these hemodynamic abnormalities. Propranolol, a -adrenoreceptor blocker, reverses these hemodynamic abnormalities by inhibiting this increased sympathetic activity and thus has the potential to improve hemodynamics in NCPH patients, thus decreasing the rate of rebleed. In a murine model of chronic schistosomiasis, administration of propranolol led to a significant decrease in portal venous inflow and portal pressure and ameliorated the development of portal systemic shunting. 16 Human studies on hemodynamic parameters show that propranolol leads to a decrease in cardiac Abbreviations used in this paper: bpm, beats per minute; EHPVO, extrahepatic portal vein obstruction; EVL, endoscopic variceal ligation; HVPG, hepatic venous pressure gradient; NCPF, noncirrhotic portal fibrosis; NCPH, noncirrhotic portal hypertension by the AGA Institute /$36.00 doi: /j.gastro

2 October 2010 EVL VERSUS BETA-BLOCKERS FOR NCPH 1239 index and azygous blood flow in patients with NCPH. 15,17 In a recent study, administration of propranolol to 13 schistosomiasis patients led to a decrease in variceal pressure and wall tension, which are surrogate markers of portal pressure. 18 These data suggest that -blockers ameliorate the hyperdynamic state and decrease portal pressure in patients with NCPH. -blockers also have been studied in secondary prophylaxis of variceal bleeding in NCPH patients. There are 2 placebo-controlled trials of propranolol on secondary prophylaxis of variceal rebleeding in noncirrhotic patients. 6,19 Both studies showed the efficacy of propranolol in decreasing rebleed rate; one showed improvement in survival also. 19 However, both the studies were placebo-controlled, and did not include any endoscopic intervention. These studies and expert opinions 3 justify the need for further clinical trials investigating the beneficial effect of -blockers on gastrointestinal bleeding caused by NCPH. In a preliminary study from our center, EVL was compared with propranolol in prevention of variceal rebleeds in NCPH patients and found to be more effective. 20 However, it was a subgroup analysis conducted on studied portal hypertension patients, and the number of NCPH patients was small. The present randomized controlled trial was initiated to study the superiority, if any, of EVL in its efficacy and safety in comparison with pharmacologic treatment using propranolol in the prevention of variceal rebleeding in patients with NCPH. We also aimed to determine the predictors of variceal rebleed in NCPH patients. Materials and Methods Patients Between January 2005 and April 2009, patients with NCPH presenting to our Liver Clinic with a history of recent hematemesis and/or melena within the past 6 weeks and proven to have esophageal varices as the bleeding source on upper gastrointestinal endoscopy, were included in the study. Exclusion criteria were as follows: (1) patients who were already on a secondary prophylaxis protocol (eg, -blockers, sclerotherapy, EVL, or glue injection) before presenting to our hospital; (2) a history of surgery for portal hypertension; (3) patients bleeding from additional sources such as gastric varices or portal hypertensive gastropathy; (4) severe cardiopulmonary or renal disease; (5) a history of severe side effects or contraindications to -blockers such as bronchial asthma, diabetes mellitus, heart failure, peripheral vascular disease, prostatic hypertrophy, arterial hypotension (systolic blood pressure, 90 mm Hg), bradycardia (basal heart rate, 50 beats per minute [bpm]), or complete heart block; (6) failure of primary hemostasis during acute bleed, and (7) refusal to give informed written consent to participate in the trial. Informed consent was provided by all the patients. The ethical committee of the institution approved the study protocol. The trial is registered with ClinicalTrial.gov vide NCT Baseline Evaluation The size of esophageal varices was assessed according to Conn s 21 classification: grade I, visible only during one phase of respiration/performance of Valsalva maneuver; grade II, visible during both phases of respiration; grade III, 3 6 mm; and grade IV, greater than 6 mm. The size of the largest varix was assessed by comparison with the shaft of the biopsy forceps (3 mm) or with the distance between the open jaws of the biopsy forceps (6 mm) in the lumen of the lower 2 3 cm of the esophagus. The included patients were categorized into noncirrhotic portal fibrosis (NCPF) and extrahepatic portal vein obstruction (EHPVO). NCPF was diagnosed when varices were present and there was no evidence of thrombosis in the splenoportal axis on ultrasonography and no evidence of cirrhosis on clinical, biochemical, histologic, radiologic profile, or hepatic venous pressure gradient (HVPG) measurement. EHPVO was diagnosed when a portal cavernoma was detected by ultrasonography and there was no evidence of cirrhosis. Study Design Patients with NCPH presenting with an acute upper gastrointestinal bleed underwent emergency upper gastrointestinal endoscopy (Figure 1). Acute esophageal variceal bleed was diagnosed if active bleeding was seen from the varix, a white nipple or a clot was seen on the varix, or if there was blood in the stomach in a patient with an esophageal varix and no other potential bleeding source. If esophageal varices were found to be the cause of bleed, these patients underwent band ligation and also were given vasoactive drugs (terlipressin or somatostatin) for 5 days. Those patients who had a failure of primary hemostasis during acute bleed were excluded from the study. On day 6 of acute bleed, patients were screened and if found to satisfy the inclusion and exclusion criteria, were enrolled and randomly assigned to receive either EVL or propranolol for secondary prophylaxis. Thus, day 6 of acute bleed formed day 1 of randomization in patients presenting with acute bleed. Those patients with NCPH who presented with a history of recent bleed from esophageal varices and satisfying the inclusion and exclusion criteria were enrolled and randomized on day 1 to receive either EVL or propranolol. Patients who previously had received more than one session of ligation/sclerotherapy were excluded. Randomization This was an open-label, randomized, controlled trial. Patients were randomized using computer-generated random numbers to receive either of the 2 therapies (EVL or propranolol). The randomization sequence re-

3 1240 SARIN ET AL GASTROENTEROLOGY Vol. 139, No. 4 obliteration was defined as nonvisualization of varices, or grade I varices. 22 Once varices were obliterated, repeat endoscopy was performed at 3-month intervals to check for recurrent varices. Recurrence of varices was defined as appearance or an increase in the grade of varices after achieving successful eradication or reduction in size of varices. Propranolol Therapy Patients randomized to drug therapy were started on propranolol on day 1 of randomization after a baseline electrocardiogram and cardiac evaluation. Treatment was started with propranolol at a dose of 20 mg twice a day. The heart rate and blood pressure were checked after 12 and 24 hours. The dose of propranolol was increased at increments of 20 to 40 mg on alternate days until the patient achieved a heart rate of 55 bpm, or a maximum dose of 320 mg/day was achieved or side effects of therapy were observed. The dosage was reduced to the previous maximum tolerated dose if any side effects occurred. Patients were monitored daily until -blockade was adequate, then monthly for the first 3 months, and every 3 months subsequently. Drug compliance was ascertained by interviewing the patient and by measuring the resting heart rate. Endoscopy was repeated every 6 months for 1 year and then annually thereafter. End Points Primary end points of the study were rebleed and death. Secondary end points were adverse effects of EVL or drug therapy, variceal eradication on EVL, variceal recurrence after eradication on EVL, and decrease in the variceal grade in the propranolol limb. Figure 1. Study design. mained with the statistician, and the sequence remained concealed from the investigators until the intervention was assigned. EVL Patients assigned to the EVL group underwent ligation on day 1 of randomization (if not previously banded). Variceal ligation was performed by using a multiband ligator. EVL was repeated at intervals of 3 4 weeks until the varices were obliterated or reduced to grade I size and deemed unbandable by the endoscopist. In each session, as many bands as possible were placed on the varices in the lower 5 7 cm of the esophagus. The adverse effects including chest pain, dysphagia, fever, and gastrointestinal bleeding were carefully recorded. Variceal Follow-up Evaluation The study was continued until 6 months after enrollment of the last patient. The follow-up period was defined from the date of randomization until the date of rebleed or death or date of last follow-up evaluation. For analyses involving time to rebleeding, follow-up evaluation in those who did not rebleed was censored on the date when the patient was last seen or at death. For analyses involving time to death, follow-up evaluation in those who remained alive were censored on the date when the patient was last seen. Statistical Analysis Quantitative data were expressed as the mean ( standard deviation) or median and analyzed using the Student t test or the Mann Whitney test. Qualitative data were analyzed by the 2 or the Fisher exact test. The actuarial probabilities of rebleeding from varices and of death from bleeding were calculated for all patients by the Kaplan Meier method, and comparisons were made using the log-rank test. Rebleed was analyzed for efficacy of secondary prophylaxis as per-protocol analysis and

4 October 2010 EVL VERSUS BETA-BLOCKERS FOR NCPH 1241 intention-to-treat analysis. Statistical analysis was performed using the SPSS 15.0 software package (SPSS Inc, Chicago, IL). Sample Size Calculation The required sample size was calculated assuming a 5% rebleed rate with variceal banding and a 30% rebleed rate with propranolol on the basis of previous studies. 6,20 With an value of 5% and power of 80%, 47 patients were required in each limb, assuming 10% loss to follow-up evaluation. Results Patients A total of 2038 new patients with a diagnosis of portal hypertension presented to our liver follow-up clinic between January 2005 and April 2009 (Figure 2). Of these, 246 were of noncirrhotic etiology and were screened. A total of 107 patients never had upper gastrointestinal bleed and were excluded from the study. A total of 139 patients presented with acute variceal bleed (29 patients) or had a history of variceal bleed in the recent past (110 patients). Thirty-eight patients who did not fulfill the inclusion criteria also were excluded. The remaining 101 patients were included in this prospective trial. These patients were randomized to either the EVL arm (51 patients) or the -blocker arm (50 patients) and were analyzed on intention-to-treat analysis. Of the 51 patients in the EVL arm, 4 were lost to follow-up evaluation before the second EVL session, and in the -blocker arm, 3 patients were lost to follow-up evaluation before target heart rate or the maximum -blocker dose was achieved. These patients along with those who rebled were considered as treatment failures on intention-totreat analysis and were excluded on per-protocol analysis. Thus, 94 patients were available for evaluation on perprotocol analysis: 47 in the EVL arm and an equal number in the -blocker arm. Figure 2. Trial profile. Table 1. Baseline Characteristics of the Patients Parameter EVL Propranolol P value Patients (n 101) (n 51) (n 50) Mean age ( SD), y 18.8 ( 13.9) 22.5 ( 17.4).251 Sex, n (%).268 Male 38 (74) 32 (64) Female 13 (26) 18 (36) Etiology of portal.228 hypertension, n (%) EHPVO NCPF 8 13 Presentation as acute bleed 16 (31%) 9 (18%).167 Number of bleeds before 1.7 ( 1.1) 1.7 ( 0.9).978 randomization ( SD) Number of blood 2.2 ( 2.8) 1.7 ( 1.3).342 transfusions in last bleed ( SD) Median grade of esophageal varices Grade of varices.441 II 12 (24%) 17 (34%) III 20 (39%) 19 (38%) IV 19 (37%) 14 (28%) Patients with gastric varices, 24 (47) 21 (42).690 n(%) Patients with PHG, n (%) 22 (43) 19 (38).687 Mean ( SD) platelet count, 132 ( 42) 140 ( 45) /mm 3 Child s score, median (range) 5 (5 9) 5 (5 7).429 PHG, portal hypertensive gastropathy; SD, standard deviation. Baseline Characteristics The baseline characteristics of the patients are listed in Table 1. The 2 groups were well matched at the time of entry into the trial with no significant differences in the demographic, clinical, or endoscopic features. The etiology of NCPH was similar in both groups. The grade of esophageal varices and the proportion of patients with gastric varices or portal hypertensive gastropathy also was comparable in the 2 groups. Twenty-five patients presented with acute variceal bleed (16 in the EVL and 9 in the -blocker group; P.167). The severity of acute bleed as assessed by the number of blood transfusions in the last bleed also was comparable in both groups (EVL, 2.2; -blocker, 1.7; P.342). The median follow-up period was 23 months (range, 7 days to 56 months). This was comparable in both groups: the median follow-up period in the EVL group was 23 months (range, 7 days to 48 months); the median follow-up period in the -blocker group was 23 months (range, 1 56 mo; P.629). EVL was repeated at a median interval of 3 weeks in the EVL arm. The median number of sessions required to eradicate varices was 3 (range, 1 7) including the index bleed banding. The median dose of propranolol used in the study was 160 mg/day (range, mg). For children younger than 5 years of age, the dose required was 80 mg (range, mg). The time for adequate -blockade was 13 days. At the time of

5 1242 SARIN ET AL GASTROENTEROLOGY Vol. 139, No. 4 Figure 3. Changes in heart rate (HR) and mean arterial pressure (MAP) on follow-up evaluation after administration of propranolol. -blockade there was a decrease in heart rate from 92 to 61 bpm, which then was maintained throughout the course of the study. We did not observe any significant decrease in mean arterial pressure on administration of propranolol (85 mm Hg at baseline to 81 mm Hg at the time of -blockade) (Figure 3). Primary End Points Rebleed. The rebleed rate as per intention-totreat analysis was 18% (9 of 50) on -blocker therapy, compared with 23.5% (12 of 51) on EVL therapy (P.625) (Table 2). Kaplan Meier plots revealed that the actuarial probability of remaining free of rebleed at 2 and 3 years in the intention-to-treat analysis was 82% and 78%, respectively, on propranolol. This was comparable with 76% at 2 and 3 years on EVL (P.530, log-rank test) (Figure 4). As per-protocol analysis (Supplementary Figure 1), 8 (17%) patients in the EVL arm rebled during the study period compared with 6 patients (12.7%) in the -blocker arm (P.775). Four patients (50% of all rebleeds) bled from esophageal varices before variceal eradication could be achieved in the EVL arm compared with 5 patients (83% of all rebleeds) in the -blocker arm (P.333). Two patients in the EVL group and 1 patient in the propranolol group bled from gastric varices (P Figure 4. Kaplan Meier: probability of rebleed (intention-to-treat analysis)..707). In the EVL group, 1 patient bled from EVL-induced ulcer, and 1 patient bled from esophageal varices that had recurred after eradication. In those who rebled from the esophageal varices, bleed occurred with a median duration of 1.5 months from the day of randomization in both the EVL (range, mo) and the -blocker groups (range, 1 25 mo; P.556). Of the 21 patients in NCPF, 8 were randomized to the EVL group and 13 were randomized to propranolol. The mean follow-up duration was 23 months in the EVL group and 18 months in the propranolol group. One patient in the propranolol group and none in the EVL group bled on follow-up evaluation (P.619 by the Fisher exact test). Death. There were no deaths in either group during the study period. Secondary End Points Adverse effects of therapy. Six patients (12%) in the EVL arm experienced chest discomfort after EVL that Table 2. Comparison of the EVL and Propranolol Groups Parameters EVL Propranolol P value Median follow-up period, mo (range) 23 (7 days to 48 mo) 23 (7 days to 56 mo).629 Lost to follow-up evaluation, n (%) 4 (7.8) 3 (6) Rebleed rate Per-protocol, n (%) 8 (17) 6 (12.7).775 Intention-to-treat, n (%) 12 (23.5) 9 (18).625 Spectrum of rebleed, n (% of all rebleeds) Esophageal varices 4 (50) 5 (83).333 Recurrent esophageal varices after eradication 1 (13) EVL ulcer 1 (13).707 Gastric varices 2 (25) 1 (17) PHG 0 0 Median time from randomization to rebleed, mo Patients died, n (%) 0 0 Side-effects of therapy, n (%) 6 (12) 9 (18).635 PHG, portal hypertensive gastropathy.

6 October 2010 EVL VERSUS BETA-BLOCKERS FOR NCPH 1243 was transient, lasting fewer than 24 hours (Table 2). One patient developed bleed from EVL-induced ulcers, which was managed successfully with conservative therapy. The median dose of -blockers administered in our study was 160 mg; the decrease in the heart rate after the maximum tolerated dose was 29% (from 92 to 65 bpm). Nine (18%) patients in the -blocker arm developed side effects: 2 had hypotension, 2 had dizziness, and 5 had extreme lethargy. All of these patients improved with dose reduction and did not require withdrawal of therapy. Variceal eradication. On EVL therapy, esophageal varices could be eradicated in 43 of the 51 patients (84%). Of these 8 patients, 4 patients were lost to follow-up evaluation. In 4 patients, bleeding from esophageal varices occurred before varices could be eradicated. Two patients bled from gastric varices (after eradication of esophageal varices) and one bled from an esophageal ulcer. The patient with ulcer bleeding had small nonbandable varices. One patient had recurrence of varices on follow-up evaluation, which caused variceal bleed. Recurrence of varices after eradication. Of the patients who had eradication of varices on EVL, 10 patients (23%) had recurrence of varices, 1 of whom rebled at 6 months of randomization. This patient had not received band ligation after 3 months, when recurrence of varices first was noticed. Decrease in size of esophageal varices on propranolol. Forty-five of the 50 (90%) patients on -blocker therapy had repeat endoscopy performed; 47% of these patients had a decrease in the grade of varices. The remaining patients either had no decrease or an increase in the size of varices. Six of these patients had esophageal variceal bleed on follow-up evaluation. None of the patients in whom a reduction in the size of varices was documented experienced bleeding from varices during the study period. Predictors of rebleed. Different variables such as age, sex, treatment group, platelet count, grade of esophageal varices, presence of gastric varices, presentation as acute bleed, and severity of last bleed as assessed by number of blood transfusions were analyzed to assess the predictors of rebleed. None of the variables was found to be predictive of rebleeding on univariate analysis (Table 3). Table 3. Factors Associated With Rebleed (Univariate Analysis) Parameters Relative hazard 95% Confidence interval P value Age, y Sex Treatment group Presentation as acute bleed Number of blood transfusions Platelet count Grade of esophageal varices Presence of gastric varices The etiology of portal hypertension (EHPVO or NCPF) also was not a predictor of rebleed on univariate analysis. Discussion This novel and large prospective randomized controlled trial clearly shows that propranolol is as effective as endoscopic variceal band ligation in preventing variceal rebleed in patients with noncirrhotic portal hypertension. Side effects resulting from either therapy are minor and comparable with each other. There were no major side effects, suggesting that propranolol and EVL are well tolerated in NCPH patients. The 2 most commonly used treatment modalities for prevention of esophageal variceal rebleed in patients with portal hypertension are reducing portal pressure by nonselective -blockade with the help of -blockers, propranolol, or nadolol, or directly obliterating the varices by local therapy such as endoscopic variceal band ligation. However, a majority of studies comparing these 2 therapies have been in cirrhotic patients. Patch et al 13 found a rebleed rate of 37% with -blockers comparable with 51% with EVL at a median follow-up period of 307 days. Villanueva et al 14 found a rebleed rate of 33% at a 21- month median follow-up evaluation. The probability of recurrent bleeding is lower for patients with a hemodynamic response to therapy, defined as a decrease in the hepatic venous pressure gradient of more than 20% from the baseline value or to less than 12 mm Hg. This is one major limitation with the use of -blockers in noncirrhotic patients because HVPG cannot be used to define responders. Studies on prevention of rebleeding in NCPH patients are either noncomparative descriptive studies 5,7 9,22 or results of a subgroup analysis of all patients with portal hypertension in which NCPH patients were a minority. 20 We found that after a median follow-up period of 23 months, the variceal rebleed rate was 17% in the EVL group and 12.7% in the propranolol group (P.775). According to the intention-to-treat analysis, the rate of treatment failures also was comparable between the 2 groups; 23% on EVL and 18% on -blockers (P.625). The rebleeding rate in the endoscopic band ligation group (17%) was similar to that reported in other studies with sclerotherapy in NCPH patients. In another study, 8 a rebleed rate of 26% was reported after 29 months of follow-up evaluation in 101 patients with EHPVO who were entered into the sclerotherapy program after presenting with variceal bleeding. In another study 9 in EHPVO patients treated with sclerotherapy, the rate of variceal rebleeding at 5 years owing to all causes and owing to esophagogastric varices was 35% and 28%, respectively. 9 In a recent retrospective observational study in patients with EHPVO, the overall variceal rebleeding rate on long-term sclerotherapy and/or band ligation protocol was 23% at 1 year and 37% at 5 years. 22 In another prospective study, 23 the rebleed rate on EVL was

7 1244 SARIN ET AL GASTROENTEROLOGY Vol. 139, No % after a follow-up period of 12 months and on sclerotherapy was 19% at 14 months. In this study, the initial EVL was performed at an interval of 7 10 days and every 2 weeks thereafter. In previous studies in noncirrhotic patients, the use of -blockers to prevent rebleed has been rather scarce. There are anecdotal case reports of response to -blockers in NCPH in which propranolol was used in a small proportion of patients with good results. 24,25 There have been 2 placebo-controlled trials on propranolol in preventing rebleed in NCPH patients. Kiire 6 reported a 1-year rebleed rate of 20% on propranolol compared with 80% on placebo, similar to our study (12.7% rebleed rate at 20 mo). However, the patient population in their study was quite different from our study because they included patients with schistosomiasis of the liver, which besides presinusoidal, has a strong perisinusoidal component. 26 In the other controlled trial on schistosomiasis there was evidence of less rebleeding in the propranolol-treated patients compared with placebo. 19 More patients on propranolol survived than those on placebo. On the other hand, the majority of our patients had extrahepatic portal vein obstruction. This disease is quite distinct from schistosomiasis and is quite prevalent in children and young adults. The results of our study clearly show that -blockers are as effective as band ligation in preventing rebleed from esophageal varices in noncirrhotic patients. The results are slightly different from a subgroup analysis on NCPH patients performed by our group previously, which showed EVL to be superior to propranolol. 20 However, in that study, the proportion of NCPH patients was quite small and also was limited by a small sample size. This difference in the patient enrollment could explain the differences between the beneficial effects of propranolol in the present study compared with the previous study. Propranolol works predominantly by decreasing the portal inflow and portal pressure and more importantly by decreasing the portosystemic shunting. This can be inferred from the finding in our study because continued use of propranolol led to a decrease in the variceal size (by at least one grade on Conn s 21 classification) in 47% of patients and because variceal rebleed did not occur in any of the patients who had a decrease in variceal size. The cause of rebleed also was comparable in both the groups. In the -blocker group, of the 6 patients who rebled, 5 bled from esophageal varices and 1 bled from gastric varices. Of the 8 rebleeds in the EVL group, 5 occurred from esophageal varices and 2 from gastric varices. One patient rebled from EVL ulcers 1 week after the banding session. The median time from randomization to rebleed was 1 month in the EVL group (range, 7 days to 6 months). This was slightly lower compared with 6 months in the -blocker group (range, 1 25 mo) (P.142). This suggests that variceal bleed occurs earlier in patients on EVL, before eradication of varices, compared with -blockers, in which bleed occurs later. Variceal eradication could be achieved in 84% of patients in the EVL group. Four (16%) patients bled before varices could be eradicated. A median of 3 sessions were required for eradication of varices (range, 1 7). After variceal obliteration was achieved, rebleed was rare. Recurrence was noticed in 23% of patients, of whom 1 patient rebled. These data suggest that EVL is efficacious in eradication of esophageal varices. However, a proportion of patients can have recurrence of varices. It is well accepted that a repeat endoscopy should be performed every 3 6 months to screen for emergence of varices. 27 A timely surveillance endoscopy and band ligation could help prevent rebleed from recurred varix. The median dose of propranolol administered was 160 mg/day. This led to a decrease in heart rate by 33%. This compares well with the recommended reduction in heart rate in patients with cirrhosis to achieve a reduction in variceal bleeding. Prolonged administration of propranolol was successful in achieving a decrease in the variceal grade in approximately half of the patients. None of these patients rebled, confirming the data in cirrhosis that patients who respond to -blocker therapy rarely bleed. Because HVPG is not a good measure of assessment of response to -blocker therapy in NCPH patients, the surrogate index of reduction in variceal size could be used to prognosticate low likelihood of rebleed. None of the patients in either group died during the course of the study, suggesting that the bleeds are well tolerated in patients with NCPH. The adverse effects of therapy also were comparable in both groups (12% in EVL and 18% in -blocker; P.635). In the EVL group, there was 1 rebleed from the EVL-induced ulcerations. The patient did not decompensate from the bleed, suggesting that the EVL ulcer bleeds are well tolerated in noncirrhotic patients. Other complications were transient; chest discomfort lasted fewer than 24 hours, which resolved without any specific therapy. None of the patients developed esophageal strictures. Complications in the -blocker arm included dizziness, lethargy, and hypotension. All these were well tolerated and when dose reduction to the previous tolerated dose was performed, none of the patients required withdrawal of therapy, suggesting that -blockers are well tolerated in noncirrhotic patients. In conclusion, the results of this prospective study clearly show that -blockers are as effective as endoscopic variceal band ligation in the secondary prevention of variceal bleeding in patients with NCPH. Both propranolol and EVL are safe and well tolerated. Further, there is no survival advantage of one treatment modality over the other. Whether a combination of -blockers and EVL is

8 October 2010 EVL VERSUS BETA-BLOCKERS FOR NCPH 1245 superior to either of the therapies given alone needs to be studied. Supplementary Material Note: To access the supplementary material accompanying this article, visit the online version of Gastroenterology, atwww.gastrojournal.org, and at doi: /j.gastro References 1. Condat B, Pessione F, Hillaire S, et al. Current outcome of portal vein thrombosis in adults: risk and benefit of anticoagulant therapy. Gastroenterology 2001;120: Bhargava DK, Dasarathy S, Atmakuri SP, et al. Comparative efficacy of emergency endoscopic sclerotherapy for active variceal bleeding due to cirrhosis of the liver, non-cirrhotic portal fibrosis and extrahepatic portal venous obstruction. J Gastroenterol Hepatol 1990;5: de Franchis R. Evolving consensus in portal hypertension. Report of the Baveno IV consensus workshop on methodology of diagnosis and therapy in portal hypertension. J Hepatol 2005;43: Webb LJ, Sherlock S. The aetiology, presentation and natural history of extra-hepatic portal venous obstruction. QJM 1979;48: Vleggaar FP, van Buuren HR, Schalm SW. Endoscopic sclerotherapy for bleeding oesophagogastric varices secondary to extrahepatic portal vein obstruction in an adult Caucasian population. Eur J Gastroenterol Hepatol 1998;10: Kiire CF. Controlled trial of propranolol to prevent recurrent variceal bleeding in patients with noncirrhotic portal fibrosis. BMJ 1989;298: Chawla YK, Dilawari JB, Dhiman RK, et al. Sclerotherapy in noncirrhotic portal fibrosis. Dig Dis Sci 1997;42: Chawla YK, Dilawari JB, Ramesh GN, et al. Sclerotherapy in extrahepatic portal venous obstruction. Gut 1990;31: Vleggaar FP, van Buuren HR, Schalm SW. Endoscopic sclerotherapy for bleeding oesophagogastric varices secondary to extrahepatic portal vein obstruction in an adult Caucasian population. Eur J Gastroenterol Hepatol 1998;10: Laine L, Cook D. Endoscopic ligation compared with sclerotherapy for treatment of esophageal variceal bleeding. A meta-analysis. Ann Intern Med 1995;123: Fakrly S, Omar M, Gannam AI. EST versus EVL in management of bleeding esophageal varices: a prospective randomized study in schiolosomal hepatic fibrosis. Endoscopic Arch Ed 2000;1: Sarin SK, Kumar A, Ckawla YK, et al. Noncirrhotic portal fibrosis/ idiopathic portal hypertension: APASL recommendations for diagnosis and treatment. Hepatol Int 2007;1: Patch D, Sabin CA, Goulis J, et al. A randomized, controlled trial of medical therapy versus endoscopic ligation for the prevention of variceal rebleeding in patients with cirrhosis. Gastroenterology 2002;123: Villanueva C, Balanzo J, Novella MT, et al. Nadolol plus isosorbide mononitrate compared with sclerotherapy for the prevention of variceal rebleeding. N Engl J Med 1996;334: Braillon A, Moreau R, Hadengue A, et al. Hyperkinetic circulatory syndrome in patients with presinusoidal portal hypertension. Effect of propranolol. J Hepatol 1989;9: Sarin SK, Groszmann RJ, Mosca PG, et al. Propranolol ameliorates the development of portal-systemic shunting in a chronic murine schistosomiasis model of portal hypertension. Clin Invest 1991;87: Mies S, Neto OB, Beer A Jr, et al. Systemic and hepatic hemodynamics in hepatosplenic Manson s schistosomiasis with and without propranolol. Dig Dis Sci 1997;42: Farias AQ, Kassab F, da Rocha EC, et al. Propranolol reduces variceal pressure and wall tension in schistosomiasis presinusoidal portal hypertension. J Gastroenterol Hepatol 2009;24: el Tourabi H, el Amin AA, Shaheen M, et al. Propranolol reduces mortality in patients with portal hypertension secondary to schistosomiasis. Ann Trop Med Parasitol 1994;88: Sarin SK, Wadhawan M, Gupta R, et al. Evaluation of endoscopic variceal ligation (EVL) versus propanolol plus isosorbide mononitrate/nadolol (ISMN) in the prevention of variceal rebleeding: comparison of cirrhotic and noncirrhotic patients. Dig Dis Sci 2005;50: Conn HO. Ammonia tolerance in the diagnosis of esophageal varices: a comparison of the endoscopic, radiologic and biochemical techniques. J Lab Clin Med 1967;70: Spaander CW, Murad SD, van Buuren HR, et al. Endoscopic treatment of esophagogastric variceal bleeding in patients with noncirrhotic extrahepatic portal vein thrombosis: a long-term follow-up study. Gastrointest Endosc 2008;67: Zargar SA, Javid G, Khan BA, et al. Endoscopic ligation vs. sclerotherapy in adults with extrahepatic portal venous obstruction: a prospective randomized study. Gastrointest Endosc 2005; 61: Ozsoylu S, Koçak N, Yüce A. Propranolol therapy for portal hypertension in children. J Pediatr 1985;106: Toba N, Honjou I, Kadohara M, et al. Effect of propranolol on oesophageal variceal pressure in a patient with extrahepatic portal obstruction. J Gastroenterol Hepatol 1989;4(Suppl 1): Canto AL, Sesso A, De Brito T. Human chronic Mansonian schistosomiasis-cell proliferation and fibre formation in the hepatic sinusoidal wall: a morphometric, light and electron-microscopy study. J Pathol 1977;123: Sarin SK, Lamba GS, Kumar M, et al. Comparison of endoscopic ligation and propranolol for the primary prevention of variceal bleeding. N Engl J Med 1999;340: Received January 4, Accepted June 8, Reprint requests Address requests for reprints to: Shiv Kumar Sarin, MD, DM, Project Director, Institute of Liver and Biliary Sciences, D-1, Vasant Kunj, New Delhi, , India. shivsarin@gmail.com; fax: (91) Acknowledgments The trial is registered with ClinicalTrial.gov vide NCT Conflicts of interest The authors disclose no conflicts.

9 1245.e1 SARIN ET AL GASTROENTEROLOGY Vol. 139, No. 4 Supplementary Figure 1. Kaplan Meier analysis: probability of rebleed (per protocol analysis).