554
HEPATOLOGY Elsewhere
given late in the course of the hepatic necrosis. The similarity in beneficial effects of prostacyclin with N-acetylcysteine may be the result of the fact that prostacyclin is also known to inhibit platelet aggregation, reduce production and release of proinflammatory cytokines and maintain vascular patency. Unfortunately, the authors have not provided any data to either support or refute this contention, and therefore this conclusion is not warranted at this time. Nevertheless, the findings of this study are potentially interesting and may provide insights into the mechanism for the reported beneficial effect of late administration of acetylcysteine in acetaminophen-induced FHF. However, controlled studies are required before one can indeed conclude that late infusion of acetylcysteine is of benefit in the treatment of patients with acetaminophen overdose.
PATRICIA SHEINER Liver Transplant Program WILFREDDE MAJO Department of Anesthesia GARYA. LEVY Liver Transplant Program University of Toronto The Toronto Hospital (Toronto General Division) Toronto, Ontario, Canada M5G 2C4 REFERENCES 1. Kaplowitz .4W, Simon FR, Stolz A. Drug induced hepatotoxicity. Ann Int Med 1986;104:826-839. 2. Prescott L. Treatment of severe acetaminophen poisoning with intravenous acetylcysteine. Arch Intern Med 1981;41:386-389. 3. O’Grady JG, Alexander GJ, Hayllar KM, Williams R. Early indications of prognosis in fulminant hepatic failure. Gastroenterology 1989;97:439-435. 4. Harrison PM, Keays R, Bray GP, Alexander GJM, Williams R. Improved outcome of paracetamol-induced fulminant hepatic failure by late administration of acetylcysteine. Lancet 1990;335: 1572-1573. 5. Sinclair S13, Greig PD, Blendis LM, Abecassis M, Roberts EA, Phillips MJ, Cameron R, et al. Biochemical and clinical response to fulminant viral hepatitis to administration of prostaglandin E. J Clin Invest 1989;84:1063-1069. 6. Could SA, Rosen AL, Sehgal LR, Rice CL, Levine H, Sehgal HL, Dulski R, et al. Is cardiac output the principal determinant of PvO,? Crit. Care Med 1981;9:273. 7 . Archie JP Jr. Mathematic coupling of data: a common source of error. Ann Surg 1981;193:296-303. 8. Stratton HH, Feustel PJ, Newel1 JC. Regression of calculated variables in the presence of shared measurement error. J Appl Physiol 1987;62:2083-2093. 9. Bihari DJ, Smithies M, Pozniak A, Gimson A. A comparison of direct and indirect measurements of oxygen delivery and consumption: the effects of prostacyclin in two human volunteers. Scan J Clin Lab Invest Supplement 1987;188:37-45. 10. Ronco JJ, Phang T, Walley KR, Wiggs B, Fenwick JC, Russell JA. Oxygen consuinption is independent of changes in oxygen delivery in severe adult respiratory distress syndrome. Am Rev Respir Dis 1991;143:1267-1273. 1I . Sinclair SB, Levy GA. Eicosanoids and the liver. Ital J Gastroenterol 1990;22:205-213. 12. Rubanyi GY, Parker Botelho LH. Endothelins. FASEB J 1991;5: 2713-2720.
HEPATOI.OGY
PROPRANOLOL FOR PORTAL HYPERTENSIVE GASTROPATW: ANOTHER VIRTUE OF P-BLOCKADE?
Pe‘rez-AyusoRM, Pique‘ JM, Bosch J , Pane‘s J, Gonzalez A, Pe‘rez R, Rigau J , et al. Propranolol in prevention of recurrent bleeding from severe portal hypertensive gastropathy in cirrhosis. Lancet 1991;337:1431-1434. ABSTRACT The two main causes of gastrointestinal bleeding in cirrhosis are oesophageal varices and portal hypertensive gastropathy (PHG). Rebleeding from varices can be prevented by beta-blockers, but it is not clear whether these drugs effectively reduce rebleeding from PHG. 54 cirrhotic patients with acute or chronic bleeding from severe PHG took part in a randomised, controlled trial to investigate the efficacy of propranolol in prevention of rebleeding from PHG. 26 patients were randomised to receive propranolol daily at a dose that reduced the resting heart rate by 25%or to 55 bpm (20-160 mg twice daily), throughout mean follow-up of 21 (SD 11) months. 28 untreated controls were followed-up, with the same examinations, for 18 (13) months. The actuarial percentages of patients free of rebleeding from PHG were significantly higher in the propranolol-treated patients than in the untreated controls at 12 months (65%vs 38%; p < 0.05) and at 30 months of follow-up (52% vs 7%; p < 0.05). Propranololtreated patients had fewer episodes of acute bleeding than controls ( 0 010 [O * 0041 vs 0 * 120 [ O * 0401 per patient per month). Multivariate analysis showed that absence of propranolol treatment was the only predictive variable for rebleeding. Actuarial survival was slightly higher in the propranolol group than in the controls, but the difference was not significant. Thus, long-term propranolol treatment significantly reduces the frequency of rebleeding from severe PHG, and may improve the prognosis of cirrhotic patients with this disorder.
COMMENTS A variety of macroscopic abnormalities are often found in the gastric mucosa of cirrhotic patients. Descriptive terms such as erythematous gastropathy of cirrhosis, scarlatina rash of the stomach, congestive gastropathy, hemorrhagic gastritis of portal hypertension, mosaic gastropathy, snake-skin stomach and others are currently in use as substitutes for our lack of understanding of the pathogenesis of these lesions (1,2). The denomination of portal hypertensive gastropathy (PHG) has become popular because of the impression that portal hypertension is a necessary requirement for the development of PHG. However, this may be a simplistic view. The lack of correlation with the hepatic vein wedge pressure suggests that portal hypertension may be only a permissive factor for PHG (2). Furthermore, with the normalization of the hepatic function after orthotopic liver transplantation, resolution of PHG is the rule, indicating a possible relation to the hepatocellular functional status. The precise nature of this “hepatogastric” connection remains undefined and would undoubtedly be the subject of
Vol. 15, No. 3, 1992
HEPATOLOGY Elsewhere
555
interesting research. Recent observations indicate that therapy. An important figure in P6rez-Ayuso et al.’s depressed prostaglandin E, levels in the gastric mucosa study is the significantly greater number of patients free of patients with PHG may be an important factor in of PHG bleedingat 30 mo in the propranolol group (52%) preventing further injury to the microcirculation of the vs. controls (7%). The results at 12 mo, although stomach that had already been altered by increased significant, are coincident with an abrupt drop in the portal pressure (4,5 ) . Whatever the initiating events, number of propranolol patients, making their interprethe gastric lesions seen in PHG are responsible for many tation difficult. Patients were enrolled in the treatment nonvariceal hemorrhages in cirrhotic patients. Occa- or control groups within a month of diagnosis of PHG. sionally, patients have recurrent and severe upper The time interval between the bleeding episode and gastrointestinal hemorrhage associated only with the entry of patients into the study is an important factor in presence of prominent PHG changes. More frequently, relation to the outcome of cirrhotic patients with however, a simultaneous presence of esophagogastric hemorrhage from esophagogastric varices (7, 8). Calcuvarices with signs of recent bleeding or imminent lation of survival beginning several weeks after the rupture is seen, in addition to the gastric mucosal bleeding episode results in the exclusion of patients changes of PHG. In these instances, determination of with a high early mortality rate. Although no specific the site of bleeding is difficult and often made on an information exists regarding PHG in this respect, there educated but subjective guess. Management for bleeding is no apparent reason to believe that it behaves differfrom portal hypertensive gastropathy consists of non- ently to variceal bleeding. A l-mo interval between specific elements including intravenous fluid resusci- diagnosis and enrollment (zero-time survival) could tation, antacids, H, receptor antagonists and removal of have introduced considerable selection of the study additional noxious factors (i.e., nonsteroidal antiinflam- groups. Finally, because half of the patients had alcomatory drugs, aspirin and ethanol). Unfortunately, the holic cirrhosis, information on the number of patients response to this antigastritis cocktail is generally rather who continued t o actively drink during the study period unsatisfactory. That propranolol may have a possible in each group would have been desirable. Ethanolbeneficial effect on bleeding from PHG was suspected induced injury to the gastric mucosa may have caused from the reduction in upper gastrointestinal bleeding episodes of acute or chronic gastrointestinal blood loss rates observed in a recent large multicenter trial of unrelated to PHG. Despite these limitations, many of which are inherent p-blockers in cirrhosis (6). Therefore the study by Perez-Ayuso et al. is a welcome effort to improve our to clinical trials with cirrhotic patients, Perez-Ayuso et current management of PHG. The investigators used al.’s study indicates that it would seem appropriate to propranolol as prophylaxis for bleeding from PHG in the consider a trial with propranolol for patients who have only randomized controlled trial performed to date. The small or medium-size gastroesophageal varices and who study is well designed and conducted, following the have had at least one episode of acute upper gastrointestradition of the Barcelona liver unit. Patients ran- tinal bleeding without discernable cause other than domized to propranolol ended with a lower recurrence of severe PHG. In my opinion, the indication for prophybleeding and also had a lower overall mortality rate. The lactic propranolol is less clear in cirrhotic patients with conclusions are straightforward, and certain aspects chronic anemia presumably related to PHG. Because deserve additional discussion. It appears reasonable to many patients have both severe PHG and large esophstudy separately the patients with acute upper gas- agogastric varices with “high risk” endoscopic signs, trointestinal hemorrhage caused by PHG, which may be propranolol may also be useful for the prevention of of a life-threatening nature, and those with chronic rebleeding from PHG when it is administered as part of anemia caused by intermittent andfor occult bleeding prophylactic therapy for variceal bleeding. Several clinical questions in PHG remain unanfrom PHG. However, it is unknown whether clinically important differences exist in the natural history of swered, including an assessment of the efficacy of these two forms of PHG. A pretrial estimate of sample selective p-blockade, the therapeutic potential of parensize indicated that 50 was adequate to detect differences teral administration of P-adrenergic receptor antagoin survival and rebleeding. However, the subsequent nists in the management of acute bleeding from PHG subdivision of the treatment and control groups in and the effects other splanchnic vasodilatory drugs have patients with acute hemorrhage and those with chronic on the gastric lesions of PHG. Furthermore, because anemia may have reduced the ability to assess treatment sclerotherapy of varices interferes with the blood flow of effect on these subgroups. As the subgroups become the venous collaterals of the esophagogastric region, one smaller, the chance of observing a significant difference, wonders about the consequences of acute and long-term which may not be real (type I error), increases substan- variceal sclerotherapy on the gastric lesions of PHG. tially. In a related aspect, the definition of a rebleeding Previous studies have suggested that PHG is found more event in patients with chronic anemia was judged from frequently after programs of sclerotherapy (9, 10). blood transfusion requirements or need for continuous Additional knowledge of the natural history of PHG iron therapy for longer than 50%of follow-up time. In an should indicate whether clinically relevant subtypes of unblinded study, the latter criteria becomes a subjective PHG can be identified. Finally, advances in the underend point and raises the possibility of investigator biases standing of pathogenetic factors at the cellular, biohecause thev had effective control over iron replacement chemical and hemodvnamic level will enable us to
556
HEPATOLOGY Elsewhere
implement, it is hoped, a more rational and effective management, of patients with severe PHG.
SANTIAGO J. MuNoz, M.D. Division of Gastroenterology and Hepatology Jefferson Medical College Philadelphia, Pennsylvania 19107 REFERENCES 1. Vigneri S, Termini R, Piraino A, Scialabba A, Pisciotta G, Fontana N. The stomach in liver cirrhosis: endoscopic, morphological and clinical correlations. Gastroenterology 1991;101:472-478. 2. Lin W, Lee F, Lin H, Tsai Y, Lee S, Lai K, Hsia H, et al. Snake skin pattern gastropathy in cirrhotic patients. J Gastroenterol Hepatol 1991;6:145-149. 3. Babb RR, Mitchell RL. Persistent hemorrhagic gastritis in a patient with portal hypertension and esophagogastric varices: the role of port.al decompressive surgery. Am J Gastroenterol 1988; 83:777-779. 4. Weiler H, Gerok W. Gastric mucosal PGE2 levels are decreased in liver cirrhosis with but not without portal hypertension [Abstract I. Gastroenterology 1990;98:645. 5. Tarnawski A, Sheffield MF, Sarfeh IJ, Douglas TG, Stachura J. Microvascular endothelium of portal hypertensive gastric mucosa is a major t.arget for alcohol injury and prostaglandin protection [Abstract]. Gastroenterology 1990;98:638. 6. Poynard T, Cales P, Pasta L, Ideo G, Pascal JP, Pagliaro L, Lebrec D, et al. Beta-adrenergic-antagonistdrugs in the prevention of gastrointestinal bleeding in patients with cirrhosis and esophageal varices. N 13ngl J Med 1991;324:1532-1538. 7. Burroughs AK, Mezzanotte G, Phillips A, McCormick A, McIntyre N. Cirrhotics with variceal hemorrhage: the importance of the time interval between admission and the start of analysis for survival and rebleedine 1989:9:801-807. u rates. HEPATOLOGY 8. Graham DY, Smith JL. The course of patients after variceal hemorrhage. Gastroenterology 1981;80:800-809. 9. Sarin SK, Misra SP, Singal A, Thorat V, Broor SL. Evaluation of the incidence and significance of the “mosaic pattern” in patients with cirrhosis, non-cirrhotic portal fibrosis, and extrahepatic obstruction. Am J Gastroenterol 1988;83:1235-1239. 10. McCormack TP, Sims I, Eyre-Brook I, Kennedy H, Goepel J, Johnson AG, l’riger DR. Gastric lesions in portal hypertension: inflammatory gastritis or congestive gastropathy? Gut 1985;26: 1226-1232. ~I
PORTOSYSTEMIC SHUNT WITHOUT SURGERY
Zernel G, Katzen BT, Becker GJ, Benenati JF, Sallee S. Percutaneous transjugular portosystemic shunt. JAMA 1991;266:390-393. ABSTRACT Objective.-To determine the effectiveness of the Palmaz balloon expandable stent for the creation of a transjugular intrahepatic portosystemic shunt. The device is designed to achieve portal decompression in patients with variceal hemorrhage secondary to portal hypertension. Design. -Transjugular intrahepatic portosystemic shunting was performed in eight patients during a 9-month period. Mean follow-up was 5 months. Patients. -All patients had cirrhosis with portal hypertension and varices. Bleeding occurred in seven patients from esophageal varices and in one patient from hemorrhoids.
HEPATOLOGY
Main Outcome Measures.-Shunt patency and recurrent variceal hemorrhage. Results. -Shunts created from a transjugular approach between a hepatic and a portal vein (diameters of 8 to 12 mm) lowered the average portosystemic pressure gradient from 36 to 11mm Hg. Mean postoperative hospital stay was 7.7 days. Complete variceal decompression after transjugular intrahepatic portosystemic shunt placement was identified endoscopically in all eight patients. The patient treated for hemorrhoids rebled and was treated successfully by transfemoral balloon expansion of the shunt diameter from 8 to 12 mm. All shunts were patent at 1to 9 months (mean, 5 months) of follow-up. Conclusion. -Initial results suggest that transjugular intrahepatic portosystemic shunt is a safe and effective method of portal decompression for the treatment of variceal hemorrhage. COMMENTS
The wide variety of treatments available for the management of variceal bleeding attests to the complexity of this problem. Treatments are commonly classified according to whether they arrest active bleeding, prevent rebleeding or prophylax against potential bleeding. Not all modalities have comparable efficacy for achieving these objectives. This study summarizes an early experience with a new and innovative technique that has potential broad applications for managing variceal hemorrhage. Briefly, the technique consists of nonsurgically creating a portosystemic shunt. Further studies will be needed to establish whether the safety, efficacy and cost of a transjugular intrahepatic portosystemic shunt (TIPS) compare favorably with currently available therapy. The first issue will be safety. This study reported no acute complications of TIPS although we are not provided with data regarding the transfusion requirements associated with the procedure. A prior study of nine patients similarly treated had one procedurerelated fatality (11%)(1). The study used a recent technical advance that allows TIPS placement without concomitant transhepatic portal venous puncture; this appears to simplify the procedure and may reduce morbidity. The explanation for an average postprocedure hospital stay of 7.7 days in the absence of complications is not stated in the manuscript. Previous case reports have described TIPS in patients with poor hepatic function, in patients who had failed sclerotherapy and were deemed inoperable (2) and in a patient with intractable ascites caused by malignant biliary obstruction (3). Unlike the individuals treated in the reviewed study (mostly Child’s class A or B), most patients who have variceal hemorrhage are class C cirrhotic patients (4).For TIPS to become an accepted alternative for reasonable operative candidates, the morbidity and mortality rates associated with the procedure must be no higher than portosystemic shunting in patients with similar severity of liver disease. The second issue will be efficacy in the prevention of
HEPATOLOGY Elsewhere
given late in the course of the hepatic necrosis. The similarity in beneficial effects of prostacyclin with N-acetylcysteine may be the result of the fact that prostacyclin is also known to inhibit platelet aggregation, reduce production and release of proinflammatory cytokines and maintain vascular patency. Unfortunately, the authors have not provided any data to either support or refute this contention, and therefore this conclusion is not warranted at this time. Nevertheless, the findings of this study are potentially interesting and may provide insights into the mechanism for the reported beneficial effect of late administration of acetylcysteine in acetaminophen-induced FHF. However, controlled studies are required before one can indeed conclude that late infusion of acetylcysteine is of benefit in the treatment of patients with acetaminophen overdose.
PATRICIA SHEINER Liver Transplant Program WILFREDDE MAJO Department of Anesthesia GARYA. LEVY Liver Transplant Program University of Toronto The Toronto Hospital (Toronto General Division) Toronto, Ontario, Canada M5G 2C4 REFERENCES 1. Kaplowitz .4W, Simon FR, Stolz A. Drug induced hepatotoxicity. Ann Int Med 1986;104:826-839. 2. Prescott L. Treatment of severe acetaminophen poisoning with intravenous acetylcysteine. Arch Intern Med 1981;41:386-389. 3. O’Grady JG, Alexander GJ, Hayllar KM, Williams R. Early indications of prognosis in fulminant hepatic failure. Gastroenterology 1989;97:439-435. 4. Harrison PM, Keays R, Bray GP, Alexander GJM, Williams R. Improved outcome of paracetamol-induced fulminant hepatic failure by late administration of acetylcysteine. Lancet 1990;335: 1572-1573. 5. Sinclair S13, Greig PD, Blendis LM, Abecassis M, Roberts EA, Phillips MJ, Cameron R, et al. Biochemical and clinical response to fulminant viral hepatitis to administration of prostaglandin E. J Clin Invest 1989;84:1063-1069. 6. Could SA, Rosen AL, Sehgal LR, Rice CL, Levine H, Sehgal HL, Dulski R, et al. Is cardiac output the principal determinant of PvO,? Crit. Care Med 1981;9:273. 7 . Archie JP Jr. Mathematic coupling of data: a common source of error. Ann Surg 1981;193:296-303. 8. Stratton HH, Feustel PJ, Newel1 JC. Regression of calculated variables in the presence of shared measurement error. J Appl Physiol 1987;62:2083-2093. 9. Bihari DJ, Smithies M, Pozniak A, Gimson A. A comparison of direct and indirect measurements of oxygen delivery and consumption: the effects of prostacyclin in two human volunteers. Scan J Clin Lab Invest Supplement 1987;188:37-45. 10. Ronco JJ, Phang T, Walley KR, Wiggs B, Fenwick JC, Russell JA. Oxygen consuinption is independent of changes in oxygen delivery in severe adult respiratory distress syndrome. Am Rev Respir Dis 1991;143:1267-1273. 1I . Sinclair SB, Levy GA. Eicosanoids and the liver. Ital J Gastroenterol 1990;22:205-213. 12. Rubanyi GY, Parker Botelho LH. Endothelins. FASEB J 1991;5: 2713-2720.
HEPATOI.OGY
PROPRANOLOL FOR PORTAL HYPERTENSIVE GASTROPATW: ANOTHER VIRTUE OF P-BLOCKADE?
Pe‘rez-AyusoRM, Pique‘ JM, Bosch J , Pane‘s J, Gonzalez A, Pe‘rez R, Rigau J , et al. Propranolol in prevention of recurrent bleeding from severe portal hypertensive gastropathy in cirrhosis. Lancet 1991;337:1431-1434. ABSTRACT The two main causes of gastrointestinal bleeding in cirrhosis are oesophageal varices and portal hypertensive gastropathy (PHG). Rebleeding from varices can be prevented by beta-blockers, but it is not clear whether these drugs effectively reduce rebleeding from PHG. 54 cirrhotic patients with acute or chronic bleeding from severe PHG took part in a randomised, controlled trial to investigate the efficacy of propranolol in prevention of rebleeding from PHG. 26 patients were randomised to receive propranolol daily at a dose that reduced the resting heart rate by 25%or to 55 bpm (20-160 mg twice daily), throughout mean follow-up of 21 (SD 11) months. 28 untreated controls were followed-up, with the same examinations, for 18 (13) months. The actuarial percentages of patients free of rebleeding from PHG were significantly higher in the propranolol-treated patients than in the untreated controls at 12 months (65%vs 38%; p < 0.05) and at 30 months of follow-up (52% vs 7%; p < 0.05). Propranololtreated patients had fewer episodes of acute bleeding than controls ( 0 010 [O * 0041 vs 0 * 120 [ O * 0401 per patient per month). Multivariate analysis showed that absence of propranolol treatment was the only predictive variable for rebleeding. Actuarial survival was slightly higher in the propranolol group than in the controls, but the difference was not significant. Thus, long-term propranolol treatment significantly reduces the frequency of rebleeding from severe PHG, and may improve the prognosis of cirrhotic patients with this disorder.
COMMENTS A variety of macroscopic abnormalities are often found in the gastric mucosa of cirrhotic patients. Descriptive terms such as erythematous gastropathy of cirrhosis, scarlatina rash of the stomach, congestive gastropathy, hemorrhagic gastritis of portal hypertension, mosaic gastropathy, snake-skin stomach and others are currently in use as substitutes for our lack of understanding of the pathogenesis of these lesions (1,2). The denomination of portal hypertensive gastropathy (PHG) has become popular because of the impression that portal hypertension is a necessary requirement for the development of PHG. However, this may be a simplistic view. The lack of correlation with the hepatic vein wedge pressure suggests that portal hypertension may be only a permissive factor for PHG (2). Furthermore, with the normalization of the hepatic function after orthotopic liver transplantation, resolution of PHG is the rule, indicating a possible relation to the hepatocellular functional status. The precise nature of this “hepatogastric” connection remains undefined and would undoubtedly be the subject of
Vol. 15, No. 3, 1992
HEPATOLOGY Elsewhere
555
interesting research. Recent observations indicate that therapy. An important figure in P6rez-Ayuso et al.’s depressed prostaglandin E, levels in the gastric mucosa study is the significantly greater number of patients free of patients with PHG may be an important factor in of PHG bleedingat 30 mo in the propranolol group (52%) preventing further injury to the microcirculation of the vs. controls (7%). The results at 12 mo, although stomach that had already been altered by increased significant, are coincident with an abrupt drop in the portal pressure (4,5 ) . Whatever the initiating events, number of propranolol patients, making their interprethe gastric lesions seen in PHG are responsible for many tation difficult. Patients were enrolled in the treatment nonvariceal hemorrhages in cirrhotic patients. Occa- or control groups within a month of diagnosis of PHG. sionally, patients have recurrent and severe upper The time interval between the bleeding episode and gastrointestinal hemorrhage associated only with the entry of patients into the study is an important factor in presence of prominent PHG changes. More frequently, relation to the outcome of cirrhotic patients with however, a simultaneous presence of esophagogastric hemorrhage from esophagogastric varices (7, 8). Calcuvarices with signs of recent bleeding or imminent lation of survival beginning several weeks after the rupture is seen, in addition to the gastric mucosal bleeding episode results in the exclusion of patients changes of PHG. In these instances, determination of with a high early mortality rate. Although no specific the site of bleeding is difficult and often made on an information exists regarding PHG in this respect, there educated but subjective guess. Management for bleeding is no apparent reason to believe that it behaves differfrom portal hypertensive gastropathy consists of non- ently to variceal bleeding. A l-mo interval between specific elements including intravenous fluid resusci- diagnosis and enrollment (zero-time survival) could tation, antacids, H, receptor antagonists and removal of have introduced considerable selection of the study additional noxious factors (i.e., nonsteroidal antiinflam- groups. Finally, because half of the patients had alcomatory drugs, aspirin and ethanol). Unfortunately, the holic cirrhosis, information on the number of patients response to this antigastritis cocktail is generally rather who continued t o actively drink during the study period unsatisfactory. That propranolol may have a possible in each group would have been desirable. Ethanolbeneficial effect on bleeding from PHG was suspected induced injury to the gastric mucosa may have caused from the reduction in upper gastrointestinal bleeding episodes of acute or chronic gastrointestinal blood loss rates observed in a recent large multicenter trial of unrelated to PHG. Despite these limitations, many of which are inherent p-blockers in cirrhosis (6). Therefore the study by Perez-Ayuso et al. is a welcome effort to improve our to clinical trials with cirrhotic patients, Perez-Ayuso et current management of PHG. The investigators used al.’s study indicates that it would seem appropriate to propranolol as prophylaxis for bleeding from PHG in the consider a trial with propranolol for patients who have only randomized controlled trial performed to date. The small or medium-size gastroesophageal varices and who study is well designed and conducted, following the have had at least one episode of acute upper gastrointestradition of the Barcelona liver unit. Patients ran- tinal bleeding without discernable cause other than domized to propranolol ended with a lower recurrence of severe PHG. In my opinion, the indication for prophybleeding and also had a lower overall mortality rate. The lactic propranolol is less clear in cirrhotic patients with conclusions are straightforward, and certain aspects chronic anemia presumably related to PHG. Because deserve additional discussion. It appears reasonable to many patients have both severe PHG and large esophstudy separately the patients with acute upper gas- agogastric varices with “high risk” endoscopic signs, trointestinal hemorrhage caused by PHG, which may be propranolol may also be useful for the prevention of of a life-threatening nature, and those with chronic rebleeding from PHG when it is administered as part of anemia caused by intermittent andfor occult bleeding prophylactic therapy for variceal bleeding. Several clinical questions in PHG remain unanfrom PHG. However, it is unknown whether clinically important differences exist in the natural history of swered, including an assessment of the efficacy of these two forms of PHG. A pretrial estimate of sample selective p-blockade, the therapeutic potential of parensize indicated that 50 was adequate to detect differences teral administration of P-adrenergic receptor antagoin survival and rebleeding. However, the subsequent nists in the management of acute bleeding from PHG subdivision of the treatment and control groups in and the effects other splanchnic vasodilatory drugs have patients with acute hemorrhage and those with chronic on the gastric lesions of PHG. Furthermore, because anemia may have reduced the ability to assess treatment sclerotherapy of varices interferes with the blood flow of effect on these subgroups. As the subgroups become the venous collaterals of the esophagogastric region, one smaller, the chance of observing a significant difference, wonders about the consequences of acute and long-term which may not be real (type I error), increases substan- variceal sclerotherapy on the gastric lesions of PHG. tially. In a related aspect, the definition of a rebleeding Previous studies have suggested that PHG is found more event in patients with chronic anemia was judged from frequently after programs of sclerotherapy (9, 10). blood transfusion requirements or need for continuous Additional knowledge of the natural history of PHG iron therapy for longer than 50%of follow-up time. In an should indicate whether clinically relevant subtypes of unblinded study, the latter criteria becomes a subjective PHG can be identified. Finally, advances in the underend point and raises the possibility of investigator biases standing of pathogenetic factors at the cellular, biohecause thev had effective control over iron replacement chemical and hemodvnamic level will enable us to
556
HEPATOLOGY Elsewhere
implement, it is hoped, a more rational and effective management, of patients with severe PHG.
SANTIAGO J. MuNoz, M.D. Division of Gastroenterology and Hepatology Jefferson Medical College Philadelphia, Pennsylvania 19107 REFERENCES 1. Vigneri S, Termini R, Piraino A, Scialabba A, Pisciotta G, Fontana N. The stomach in liver cirrhosis: endoscopic, morphological and clinical correlations. Gastroenterology 1991;101:472-478. 2. Lin W, Lee F, Lin H, Tsai Y, Lee S, Lai K, Hsia H, et al. Snake skin pattern gastropathy in cirrhotic patients. J Gastroenterol Hepatol 1991;6:145-149. 3. Babb RR, Mitchell RL. Persistent hemorrhagic gastritis in a patient with portal hypertension and esophagogastric varices: the role of port.al decompressive surgery. Am J Gastroenterol 1988; 83:777-779. 4. Weiler H, Gerok W. Gastric mucosal PGE2 levels are decreased in liver cirrhosis with but not without portal hypertension [Abstract I. Gastroenterology 1990;98:645. 5. Tarnawski A, Sheffield MF, Sarfeh IJ, Douglas TG, Stachura J. Microvascular endothelium of portal hypertensive gastric mucosa is a major t.arget for alcohol injury and prostaglandin protection [Abstract]. Gastroenterology 1990;98:638. 6. Poynard T, Cales P, Pasta L, Ideo G, Pascal JP, Pagliaro L, Lebrec D, et al. Beta-adrenergic-antagonistdrugs in the prevention of gastrointestinal bleeding in patients with cirrhosis and esophageal varices. N 13ngl J Med 1991;324:1532-1538. 7. Burroughs AK, Mezzanotte G, Phillips A, McCormick A, McIntyre N. Cirrhotics with variceal hemorrhage: the importance of the time interval between admission and the start of analysis for survival and rebleedine 1989:9:801-807. u rates. HEPATOLOGY 8. Graham DY, Smith JL. The course of patients after variceal hemorrhage. Gastroenterology 1981;80:800-809. 9. Sarin SK, Misra SP, Singal A, Thorat V, Broor SL. Evaluation of the incidence and significance of the “mosaic pattern” in patients with cirrhosis, non-cirrhotic portal fibrosis, and extrahepatic obstruction. Am J Gastroenterol 1988;83:1235-1239. 10. McCormack TP, Sims I, Eyre-Brook I, Kennedy H, Goepel J, Johnson AG, l’riger DR. Gastric lesions in portal hypertension: inflammatory gastritis or congestive gastropathy? Gut 1985;26: 1226-1232. ~I
PORTOSYSTEMIC SHUNT WITHOUT SURGERY
Zernel G, Katzen BT, Becker GJ, Benenati JF, Sallee S. Percutaneous transjugular portosystemic shunt. JAMA 1991;266:390-393. ABSTRACT Objective.-To determine the effectiveness of the Palmaz balloon expandable stent for the creation of a transjugular intrahepatic portosystemic shunt. The device is designed to achieve portal decompression in patients with variceal hemorrhage secondary to portal hypertension. Design. -Transjugular intrahepatic portosystemic shunting was performed in eight patients during a 9-month period. Mean follow-up was 5 months. Patients. -All patients had cirrhosis with portal hypertension and varices. Bleeding occurred in seven patients from esophageal varices and in one patient from hemorrhoids.
HEPATOLOGY
Main Outcome Measures.-Shunt patency and recurrent variceal hemorrhage. Results. -Shunts created from a transjugular approach between a hepatic and a portal vein (diameters of 8 to 12 mm) lowered the average portosystemic pressure gradient from 36 to 11mm Hg. Mean postoperative hospital stay was 7.7 days. Complete variceal decompression after transjugular intrahepatic portosystemic shunt placement was identified endoscopically in all eight patients. The patient treated for hemorrhoids rebled and was treated successfully by transfemoral balloon expansion of the shunt diameter from 8 to 12 mm. All shunts were patent at 1to 9 months (mean, 5 months) of follow-up. Conclusion. -Initial results suggest that transjugular intrahepatic portosystemic shunt is a safe and effective method of portal decompression for the treatment of variceal hemorrhage. COMMENTS
The wide variety of treatments available for the management of variceal bleeding attests to the complexity of this problem. Treatments are commonly classified according to whether they arrest active bleeding, prevent rebleeding or prophylax against potential bleeding. Not all modalities have comparable efficacy for achieving these objectives. This study summarizes an early experience with a new and innovative technique that has potential broad applications for managing variceal hemorrhage. Briefly, the technique consists of nonsurgically creating a portosystemic shunt. Further studies will be needed to establish whether the safety, efficacy and cost of a transjugular intrahepatic portosystemic shunt (TIPS) compare favorably with currently available therapy. The first issue will be safety. This study reported no acute complications of TIPS although we are not provided with data regarding the transfusion requirements associated with the procedure. A prior study of nine patients similarly treated had one procedurerelated fatality (11%)(1). The study used a recent technical advance that allows TIPS placement without concomitant transhepatic portal venous puncture; this appears to simplify the procedure and may reduce morbidity. The explanation for an average postprocedure hospital stay of 7.7 days in the absence of complications is not stated in the manuscript. Previous case reports have described TIPS in patients with poor hepatic function, in patients who had failed sclerotherapy and were deemed inoperable (2) and in a patient with intractable ascites caused by malignant biliary obstruction (3). Unlike the individuals treated in the reviewed study (mostly Child’s class A or B), most patients who have variceal hemorrhage are class C cirrhotic patients (4).For TIPS to become an accepted alternative for reasonable operative candidates, the morbidity and mortality rates associated with the procedure must be no higher than portosystemic shunting in patients with similar severity of liver disease. The second issue will be efficacy in the prevention of
