CARLO MERKEL, MASSIMO BOLOGNESI, DAVID SACERDOTI, GIANCARLO BOMBONATO, BARBARA BELLINI, RAFFAELLA BIGHIN, AND ANGELO GATTA

The Hemodynamic Response to Medical Treatment of Portal Hypertension as a Predictor of Clinical Effectiveness in the Primary Prophylaxis of Variceal Bleeding in Cirrhosis CARLO MERKEL, MASSIMO BOLOGNESI, DAVID SACERDOTI, GIANCARLO BOMBONATO, BARBARA BELLINI, RAFFAELLA BIGHIN, AND ANGELO GATTA SEE EDITORIAL ON PAGE 1175 In the prevention of variceal rebleeding, it is already established that hemodynamic response to drug treatment (decrease in hepatic venous pressure gradient [HVPG] to 12 mm Hg or by >20%) is predictive of clinical effectiveness. In primary prophylaxis very few clinical data are available. We assessed the role of the hemodynamic response to beta-blockers or beta-blockers plus nitrates in predicting clinical efficacy of prophylaxis. A total of 49 cirrhotic patients with varices at risk of bleeding, without prior variceal bleeding, were investigated by hepatic vein catheterization before and after 1 to 3 months of chronic treatment with nadolol or nadolol plus isosorbide mononitrate, and were followed during treatment for up to 5 years. A total of 30 patients (61%) were good hemodynamic responders, and among them in 12 (24%) HVPG was <12 mm Hg during treatment. During treatment 9 patients had variceal bleeding: 7 were poor responders and 2 were good responders. The probability of bleeding at 3 years of follow-up was significantly higher in poor responders (41%) than in good responders (7%; P.0008). No patient reaching an HVPG of 12 mm Hg or less during treatment had variceal bleeding during follow-up. Cox s regression analysis showed that poor hemodynamic response was the main factor predicting bleeding ( 1.91; SE( ) 0.80; P.01). During follow-up 11 patients died of hepatic causes. Survival was related to Child-Pugh class and to initial value of HVPG, according to Cox s analysis. In conclusion, the assessment of hemodynamic response to drugs in terms of HVPG is the best predictor of efficacy of prophylaxis of variceal bleeding in patients treated with beta-blockers or beta-blockers plus nitrates. (HEPATOLOGY 2000;32:930-934.) It is clearly established that prophylaxis with beta-blockers decreases the risk of first variceal bleeding in cirrhotic patients Abbreviations: HVPG, hepatic venous pressure gradient. From the Department of Clinical and Experimental Medicine, University of Padua, Padua, Italy. Received June 20, 2000; accepted August 17, 2000. Supported in part by a grant from the Italian Ministry of University and Scientific Research, Rome. Address reprint requests to: Carlo Merkel, M.D., Dipartimento di Medicina Clinica e Sperimentale, Università di Padova, Via Giustiniani 2, I-35100 Padova, Italy. E-mail: cmerkel@ux1.unipd.it; fax: (39) 049-975-4179. Copyright 2000 by the American Association for the Study of Liver Diseases. 0270-9139/00/3205-0008$3.00/0 doi:10.1053/jhep.2000.19322 930 with portal hypertension and esophageal varices at risk of bleeding. 1 Recently, it was also shown that the efficacy of beta-blockers may be enhanced by the addition of long-acting nitrates. 2 These treatments, however, do not abolish the risk of bleeding but decrease it by half or by three quarters, respectively. For these reasons, it may be useful to know factors predicting clinical response or nonresponse to treatment to allow treatment of nonresponders with alternative methodologies. In patients treated with beta-blockers 3-5 or beta-blockers plus nitrates 6 for prevention of rebleeding, it was observed that a decrease in portal pressure under chronic treatment, expressed as a decrease in hepatic venous pressure gradient (HVPG), is a strong predictor of clinical effectiveness. Indeed, decreases of HVPG to 12 mm Hg or less, or decreases by at least 20% from baseline values were associated with a negligible risk of rebleeding, whereas risk of rebleeding was confined to patients not meeting these hemodynamic end points. A single report was unable to confirm the clinical value of these hemodynamic end points. 7 In the clinical setting of prophylaxis of first bleeding, available evidence is very small. In a single study it was documented 8 that decrease in HVPG to 12 mm Hg or less is an effective predictor of success of prophylaxis, but percent variations did not allow definition of any useful cutoff value. Further data might be derived from the series by Escorsell et al., 5 which included patients treated in prevention of rebleeding and in primary prophylaxis, but no separate analysis is available. In the present study we tried to assess the role of the hemodynamic response to chronic treatment with beta-blockers or beta-blockers plus nitrates as a prognostic indicator of clinical efficacy of the primary prophylaxis of variceal bleeding in cirrhotic patients with esophageal varices at risk of bleeding. PATIENTS AND METHODS A total of 49 patients with cirrhosis, portal hypertension, and esophageal varices (F2 or F3 varices, or F1 associated with red weal marks), without previous variceal bleeding, with no previous treatment for portal hypertension were consecutively observed at the Hepatic Hemodynamic Laboratory of the Department of Clinical and Experimental Medicine of the University of Padua, and were included in the present study. The study conformed to the Declaration of Helsinki, and was approved by the competent Ethics Authorities. Written informed consent was obtained from each subject. Main clinical and biochemical data of these patients are reported in Table 1. Before starting a pharmacologic prophylaxis of variceal bleeding, in all patients HVPG was measured by hepatic vein catheterization according to the methodology used in our laboratory. 9 In brief, an F8

HEPATOLOGY Vol. 32, No. 5, 2000 MERKEL ET AL. 931 TABLE 1. Main Clinical and Biochemical Characteristics in the 49 Patients With Cirrhosis Age (years) 60.0 9.9 Gender (m/f) 34/15 Etiology Alcoholic 18 Virus-related 18 Mixed 8 Other 5 Ascites 22 s-bilirubin ( mol/l) 26.7 18.5 s-albumin (g/l) 37.4 9.7 Prothrombin activity (%) 67.4 15.3 Child-Pugh class (A/B/C) 21/24/4 Esophageal varices (F1/F2/F3) 2/34/14 Red weal marks 20 Portal hypertensive gastropathy 31 Baseline HVPG (mm Hg) 18.4 4.5 NOTE. Data are expressed as frequencies or mean SD. venous introducer was inserted in the right femoral vein under local anesthesia, and a preshaped F5 Cobra or Simmons catheter (Radifocus, Terumo Corp., Tokyo, Japan) was positioned under fluoroscopic control in a hepatic vein. The catheter was exchanged, with the aid of a Teflon-coated guide-wire, with an F7 balloon catheter (Meditech; Boston Scientific, Watertown, MA), which was used for all measurements. Hepatic venous pressure was measured in the occluded position, filling the balloon with 2 ml of diluted contrast medium, and then in the free position, after deflation, and after checking that the catheter tip was floating in the middle of the hepatic vein. Adequacy of occlusion was always checked by distal injection of a small amount of radiographic contrast medium. HVPG, expression of portal pressure in sinusoidal and postsinusoidal outflow block, as it occurs in cirrhosis, 10-13 was calculated as the difference between occluded and free HVPG. At least 3 measurements were always performed, electronic means were obtained during one minute recordings, and results were given as arithmetic means of these 3 measurements. Permanent tracings were always obtained. During the whole procedure heart rate and arterial pressure were continuously monitored. The following day, 29 patients started a chronic nadolol treatment at a dose reducing resting heart rate by approximately 25% (mean dose, 72 mg/d; range, 40-80) and underwent a second hemodynamic investigation after 1 to 3 months of continuous treatment (median, 42 days). A total of 20 patients, 6 of whom participated in a clinical trial of prophylaxis of variceal bleeding with beta-blockers plus nitrates 2 and 9 with insufficient response to beta-blockers, included in a pathophysiologic study on the hemodynamic effects of the association in this subgroup of patients, 14 were treated with nadolol (40-80 mg/d) and isosorbide-5-mononitrate (20-40 mg/d) and underwent a second hemodynamic investigation after 1 to 3 months of continuous treatment (median, 50 days). Procedures were done in the morning approximately at the same time, and with the same methodology of the first investigation. During the initial period of treatment, before the second hemodynamic investigation, 6 of 29 patients treated with nadolol had side effects requesting withdrawal of treatment, and 3 out of 20 patients treated with the association had side effects requesting withdrawal of the pertinent drug (1 with side effects of nadolol, 1 of nitrates, and 1 of both). Therefore, 7 patients had a second hemodynamic investigation without any treatment. Of these patients, two had a decrease in HVPG, one an increase, and the remainders were unchanged. During treatment patients were seen as outpatients at fortnight intervals in the first 3 months, then every 6 months. Compliance to treatment was assessed according to anamnesis and to heart rate measurements. A further patient withdrew from treatment after 10 months because of marked asthenia. No patient was lost to follow-up. Statistical Analysis. To assess the prognostic significance of the hemodynamic response to treatment, patients were classified, according to the literature 4-6 as good responders, if during chronic treatment HVPG was 12 mm Hg or if it was decreased by more than 20% from baseline values, or as poor responders, if these hemodynamic end points were not met. Cumulative probabilities of not bleeding from varices and of survival in good and poor responders were described by Kaplan-Meier curves 15 and were compared by log-rank test. 16 The possible role of confounding variables was investigated by Cox s multiple regression analysis. 17 Considering the limited number of observed end points, to avoid possible bias due to overfitting, size of varices and presence of red weal marks were only considered as possible confounding factors for bleeding, and Child-Pugh class was considered as a confounding factor for survival. In a further analysis, NIEC index 18 was tested as a possible confounding factor for bleeding. RESULTS A total of 30 patients (61%) were considered good responders according to the predefined hemodynamic criteria, and 19 (39%) were considered poor responders. In 12 patients the definition of good response was related to a decrease in HVPG to 12 mm Hg or less. The probability of a good response was similar in patients who started a treatment with nadolol alone (16 of 29; 55%) and in those who started a treatment with the association (12 of 20; 60%). Poor responders did not appear to be more severe in any variable related to liver dysfunction or portal hypertension (Table 2). During follow-up 9 patients experienced a variceal bleeding: 7 were poor responders and 2 were good responders. Of the 2 patients classified as good responders, in 1, treatment had to be withdrawn because of marked asthenia after the second hemodynamic investigation and 2 months before the occurrence of bleeding. In the other, a hepatocellular carcinoma complicated with portal vein thrombosis was diagnosed at ultrasonography 1 month before bleeding. Of the poor responders who bled during follow-up, none had to be withdrawn from treatment after the second hemodynamic investigation, and 1 developed hepatocellular carcinoma without portal vein thrombosis. Cumulative probability of variceal bleeding was significantly higher in poor responders (41% at 3 years) than in good responders (7% at 3 years; P.008; log-rank test) (Fig. 1). Probability of variceal bleeding was also higher in patients who had an HVPG during treatment higher than 12 mm Hg (9 of 37; 3-year risk of 28%), than in those who reached the threshold value of 12 mm Hg (0 of 12; log-rank test: P.06). The probability of variceal bleeding was marginally higher in patients with F3 varices (P.10) or with red weal marks (P.09); no difference was observed in TABLE 2. Clinical, Endoscopic, and Hemodynamic Characteristics of Good and Poor Responders to Medical Treatment of Portal Hypertension Variable Good Responders Poor Responders Age (years) 60.3 10.7 59.6 8.6 t 0.23; P 0.62 Child-Pugh score 6.8 1.7 7.4 1.6 t 1.19; P 0.24 HVPG (mm Hg) 18.7 5.2 18.9 3.6 t 0.15; P 0.88 Gender (m/f) 20/10 14/5 2 0.27; P.60 Ascites (yes/no) 10/18 12/7 2 3.42; P.06 Varices (F1/F2/F3) 1/21/8 1/13/6 2 1.75; P.91 Red weal marks (yes/no) 12/18 8/11 2 0.21; P.88

932 MERKEL ET AL. HEPATOLOGY November 2000 FIG. 1. Cumulative probability of being free of a first variceal bleeding in good responders (solid line) and poor responders (dashed line) according to hemodynamic criteria. FIG. 3. Cumulative probability of survival in patients with baseline HVPG below (solid line) or above (dashed line) 16mmHg. relation to the initial value of HVPG (P.50) (Fig. 2). Cox s regression analysis showed that belonging to the group of poor responders was the main factor predicting variceal bleeding ( 1.91; SE( ) 0.80; P.01), and the presence of red weal marks had a marginally significant additive value ( 1.21; SE( ) 0.70; P.07). This implies that the relative risk of bleeding in poor responders compared with good responders was 6.75 (95% CI, 1.47-30.8), with other confounding factors being equal. When NIEC index 18 was tested as a possible confounding factor instead of size of varices and red weal marks, -coefficient for poor response was nearly identical ( 1.93; SE( ) 0.80; P.01), and NIEC index was near to the significance level ( 0.078; SE( ) 0.042); P.07). During follow-up 11 patients died for causes related to liver disease: 4 of gastrointestinal bleeding (2 after the first bleeding), 4 of liver failure, 1 of hepato-renal syndrome, and 2 of hepatocellular carcinoma. Mortality was related to Child- Pugh class, to the initial value of HVPG (Fig. 3), and to an HVPG during treatment 12 mm Hg. Cox s regression analysis showed that HVPG ( 0.16; SE( ) 0.046; P.002) and Child-Pugh class ( 0.31; SE( ) 0.15; P.05) were independent predictors of outcome. FIG. 2. Cumulative probability of being free of a first variceal bleeding in patients with baseline HVPG below (solid line) or above (dashed line) 16mm Hg. DISCUSSION Prophylaxis of first variceal bleeding with beta-blockers, alone or combined with nitrates, plays a central role in the management of patients with cirrhosis and esophageal varices at risk of bleeding. 19 It is clearly established that risk of bleeding is decreased by this treatment, but not abolished. Therefore, it would be extremely helpful to be able to discriminate patients who are going to benefit from treatment from those in which no benefit will result and who may be proposed for alternative prophylactic treatments, such as band ligation. 20 Such treatments were shown to be effective in prophylaxis but cannot be considered first line treatments. In fact, in patients who are likely not to benefit from first line treatments, band ligation might prove to be a rational strategy if in this subgroup of patients it may prove to be as effective as in the general population of cirrhotic patients with large varices. The hemodynamic response to treatment was found to be the main predictor of clinical effectiveness of medical treatment in patients undergoing prevention of rebleeding with beta-blockers. 4 Indeed, a decrease in HVPG to 12 mm Hg or less or a percent decrease in HVPG of 20% or more during chronic therapy was found to be associated with a negligible risk of rebleeding. These results were confirmed by Villanueva et al. 6 in a clinical trial with beta-blockers plus nitrates, which showed that a decrease in HVPG by 20% or more is a strong predictor of effectiveness. A small clinical experience from our group was also in keeping with these statements, but the best cut-off value between good and poor responders was fixed at a 12% decrease in HVPG. 3 Recently a contrasting observation was published 7 : in that series a decrease in HVPG to 12 mm Hg or less was a predictor of efficacy only in the subgroup of patients with alcoholic cirrhosis who were compliant to treatment, but not in the general population. Some methodologic problems, such as a long interval between start of treatment and assessment, and the occurrence of many cases who bled before the second hemodynamic evaluation, may contribute to this difference. So far, a single study evaluated the value of HVPG in the prediction of first bleeding in patients undergoing medical treatment for portal hypertension. 8 In that study, a decrease in HVPG during chronic treatment with propranolol to 12 mm Hg or below identified patients with a good response to treatment. However, the clinical value of this prediction was limited, because this hemodynamic end point was reached in a

HEPATOLOGY Vol. 32, No. 5, 2000 MERKEL ET AL. 933 minority of patients (less than a quarter), whereas patients who were not going to bleed were the majority (more than two thirds). In other words, this prediction rule was sensitive but not specific in predicting failure of treatment. In the present study we could confirm the observation by Groszmann et al. 8 that patients who decrease their HVPG to 12 mm Hg or below are protected from bleeding (indeed, no patients with this feature bled during follow-up), but we were able to add a second criterion. Indeed, similar to the condition of prevention of rebleeding, a decrease in HVPG by 20% or more was also a strong predictor of success of prophylaxis. Sensitivity in predicting bleeding was 78% (7 of 9) and specificity was 70% (28 of 40), i.e. much higher than in the previous report. The clinical value of HVPG changes remained stable, also when other confounding factors, such as size of varices and presence of red weal marks or NIEC index 18 were taken into consideration by Cox s regression analysis. This method allowed us to estimate that the relative risk of bleeding in poor responders compared with that of good responders, with size of varices and red weal marks being equal, approached the value of 7. The regression analysis also concluded that in the prediction of effectiveness of treatment, the hemodynamic response is much more useful than the initial size of varices and the presence of red weal marks. According to our data, it appears that good responders are at negligible risk of bleeding (2 of 30; 6.6%), whereas poor responders have a rather high risk (7 of 29; 25%), which is similar to that expected in untreated patients. This difference may appear more relevant, if one considers that the 2 good responders who bled had precipitating factors that probably contributed to bleeding, such as the need to withdraw treatment for side effects, or the occurrence of a portal vein thrombosis. Finally, from a clinical point of view it should be considered that, despite an elevated predictive value for success of treatment of being a good hemodynamic responder (28 of 30; 93%), out of the 19 poor responders, only 7 bled during follow-up (positive predictive value for bleeding, 36%); therefore it may be questioned if the condition of poor responder may be an indication for an alternative treatment by itself considering that good alternative treatments are incompletely assessed at the present time. Only clinical trials specifically addressing this question may clarify this point. The short interval (less than 2 months as an average), shorter than that of the previous reports, that we allowed between the 2 hemodynamic studies, permitted us to assess all patients, and to avoid the problem of how to consider possible patients who bleed before the second hemodynamic investigation. From a pathophysiologic point of view, such cases should be deleted from the analysis, but from a clinical point of view, one should consider them as patients who are under investigation and for whom a diagnosis of lack of response has not been done yet; consequently they cannot be proposed for alternative treatments yet and should be considered together with good responders. For these reasons, we consider of particular relevance the issue of shortening the interval between the two hemodynamic examinations as much as possible, keeping it the shortest, which is associated to chronic hemodynamic effects. Considering that 2 poor responders bled during the second month of treatment, and 1 during the fourth month, and that we have already shown that no relevant difference in the hemodynamic effect occurs between 1 and 3 months of continuous nadolol administration, 21 we feel that 1 month may be the optimal time frame. The present research confirms the clinical value of measuring HVPG as a prognostic indicator for survival, which was consistently reported in many series. 9,22-25 In the present group of patients, the prognostic value of HVPG was additive to that of Child-Pugh classification, as assessed by Cox s regression analysis. A more severe pattern of portal hypertension, therefore, appears to be peculiar of the advanced stages of cirrhosis and is an indicator of a more advanced disease with a shorter survival. The present research could not assess the cost effectiveness of measuring HVPG response to medical treatment in patients undergoing prophylaxis. Indeed, costs can easily be evaluated, but effectiveness is related to the efficacy of alternative treatments in patients classified as poor responders. This information is currently unavailable, and may be the object of specific investigations. In conclusion, the evaluation of the hemodynamic response to medical therapy of portal hypertension, as expressed by a decrease in HVPG to 12 mm Hg or less, or by a decrease exceeding 20% of the initial value, is the most important predictor of effectiveness of prophylaxis of bleeding from esophageal varices in cirrhotic patients treated with beta-blockers or the association of beta-blockers plus nitrates. REFERENCES 1. Poynard T, Calés P, Pasta L, Ideo G, Pascal JP, Pagliaro L, Lebrec D, 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 from 589 patients from four randomized clinical trials. N Engl J Med 1991;324:1532-1538. 2. Merkel C, Marin R, Enzo E, Donada C, Cavallarin G, Torboli P, Amodio P, et al. Randomised trial of nadolol alone or with isosorbide mononitrate for primary prophylaxis of variceal bleeding in cirrhosis. Lancet 1996; 348:1677-1681. 3. Sacerdoti D, Merkel C, Gatta A. Importance of the 1-month-effect of nadolol on portal pressure in predicting failure of prevention of rebleeding in cirrhosis. J Hepatol 1991;12:124-125. 4. Feu F, Garcia-Pagan JC, Bosch J, Luca A, Terés J, Escorsell A, Rodes J. Relation between portal pressure response to pharmacotherapy and risk of recurrent variceal haemorrhage in patients with cirrhosis. Lancet 1995;346:1056-1059. 5. Escorsell A, Bordas JM, Castaneda B, Llach J, Garcìa-Pagàn JC, Rodés J, Bosch J. Predictive value of variceal pressure response to continued pharmacological therapy in patients with cirrhosis and portal hypertension. HEPATOLOGY 2000;31:1061-1067. 6. Villanueva C, Balanzò J, Novella MT, Soriano G, Sàinz S, Torras X, Cusso X, et al. Nadolol plus isosorbide mononitrate compared with sclerotherapy for the prevention of variceal rebleeding. N Engl J Med 1996;334: 1624-1629. 7. McCormick PA, Patch D, Greenslade L, Chin J, McIntyre N, Burroughs AK. Clinical vs. hemodynamic response to drugs in portal hypertension. J Hepatol 1998;28:1015-1019. 8. Groszmann RJ, Bosch J, Grace N, Conn HO, Garcia-Tsao G, Navasa M, Alberts J, et al. Hemodynamic events in a prospective randomized trial of propranolol vs. placebo in the prevention of the first variceal hemorrhage. HEPATOLOGY 1990;99:1401-1407. 9. Merkel C, Bolognesi M, Bellon S, Zuin R, Noventa F, Finucci G, Sacerdoti D, et al. Prognostic usefulness of hepatic vein catheterization in patients with cirrhosis and esophageal varices. Gastroenterology 1992;102:973-979. 10. Reynolds TB, Ito S, Iwatsuki S. Measurement of portal pressure and its clinical application. Am J Med 1970;49:649-657. 11. Lin HC, Tsai YT, Lee FY, Chang TT, Wang SS, Lay CS, Lee SD, et al. Comparison between portal vein pressure and wedged hepatic vein pressure in hepatitis B-related cirrhosis. J Hepatol 1989;9:326-330. 12. Lebrec D. Methods to evaluate portal hypertension. Gastroenterol Clin North Am 1992;21:41-59.

934 MERKEL ET AL. HEPATOLOGY November 2000 13. Perello A, Escorsell A, Bru C, Gilabert R, Moitinho E, Garcia-Pagan JC, Bosch J. Wedged hepatic venous pressure adequately reflects portal pressure in hepatitis C virus-related cirrhosis. HEPATOLOGY 1999;30:1393-1397. 14. Merkel C, Sacerdoti D, Bolognesi M, Enzo E, Marin R, Bombonato G, Angeli P, et al. Hemodynamic evaluation of the addition of isosorbide-5- mononitrate to nadolol in cirrhotic patients with insufficient response to the -blocker alone. HEPATOLOGY 1997;26:34-39. 15. Kaplan EL, Meier P. Non-parametric estimation from incomplete observations. J Am Stat Assoc 1958;53:458-481. 16. Breslow N. A generalized Kruskal-Wallis test for comparing k samples subject to unequal pattern of censorship. Biometrika 1970;57:579-594. 17. Cox DR. Regression model and life tables. J Roy Statist Soc 1972;34:187-220. 18. 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; 319:983-989. 19. Committee for Portal Hypertension of the Italian Association for the Study of the Liver. AISF practice guidelines for portal hypertension. Ital J Gastroenterol Hepatol 1999;31:224-234. 20. Sarin SK, Lamba GS, Kumar M, Mirsa A, Murthy NS. Comparison of endoscopic ligation and propranolol for the primary prevention of variceal bleeding. N Engl J Med 1999;340:988-993. 21. Gatta A, Bolognesi M, Merkel C, Finucci GF, Angeli P, Sacerdoti D, Ruol A. Long-term effects of beta-adrenergic blockade with nadolol on hepatic and renal hemodynamic and function in cirrhotic patients. Clin Physiol 1987;7:377-387. 22. Vinel JP, Cassigneul J, Louis A, Levade M, Pascal JP. Clinical and prognostic significance of portohepatic gradient in patients with cirrhosis. Surg Gynecol Obst 1982;155:347-352. 23. Gluud C, Henriksen JH, Nielsen G, Copenhagen Study Group for Liver Diseases. Prognostic indicators in alcoholic cirrhotic men. HEPATOLOGY 1988;8:222-227. 24. Vorobioff J, Groszmann RJ, Picabea E, Gamen M, Villavicencio R, Bordato J, Morel I, et al. Prognostic value of hepatic venous pressure gradient measurement in alcoholic cirrhosis. A 10-year prospective study. Gastroenterology 1996;111:701-709. 25. Armonis A, Patch D, Burroughs AK. Hepatic venous pressure measurement: an old test as a new prognostic marker in cirrhosis? HEPATOLOGY 1997;25:245-248.