Transjugular Intrahepatic Portosystemic Shunt

Transcription

1 카J Korean Radiol Soc 1997; 36 : Transjugular Intrahepatic Portosystemic Shunt Results and Prognostic Factors in Patients with Post-necrotic Liver Cirrhosis 1 Jae Hyung Park, M.D., Yong Joo Kim, M.D. 2, Jin Wool Chung, M.D. Joon Koo Han, M.D., Man Chung Han, M.D. Purpose : To evaluate the effectiveness of transjugular intrahepatic portosystemic shunt(tips) in the management of gastroesophageal variceal bleeding and predictive factors for long-term survival in patients with post-necrotic liver cirrhosis. Materials and Methods : A total of 49 patients with post-necrotic liver cirrhosis underwent TIPS over a recent three-year period. Forty-five had a history ofhepatitis B viral infection, and four, ofhepatitis c viral infection. In all patients, the indication for the procedure was variceal bleeding. Child-Pugh class was A in seven patients, B in 16 and c in 26 patients at the time ofthe last bleeding. The effectiveness ofportal decompression and bleeding control was evaluated. Long-term survival was calculated by the Kaplan-Meier method and predictive factors were analyzed using the Wilcoxon test. Results : The procedure was technically successful in all cases. The portosystemic pressure gradient decreased significantly from 21.4::!: 6.4mmHg to 12.0::!: 5.1 mmhg (N=45). Active variceal bleeding was controlled in 34 of the 37 emergency patients. The totallength of follow-up was from one day to three and a half years(mean: 383:!: 357 days). Rebleeding developed in 17 patients (35 %). Hepatic encephalopathy, either newly developed or aggravate,d, occurred in 16(32.7 %). The thirty-day mortality rate was 20.4%, and the one-year survival rate was 63.8 %. The significant predictive factors for poor prognosis were Child-Pugh class c and post-tips hepatic encephalopathy. Conclusion : TIPS is effective in portal decompression in the patients with variceal bleeding due to post-necrotic liver cirrhosis. The Child-Pugh classification and hepatic encephalopathy after TIPS are considered to be significant predictive factors for long-term survival. Index Words : Hypertension, portal Liver, cirrhosis Liver, interventional procedure Shunts, portosystemic Introduction Transjugular intrahepatic portosystemic shunt 'Department of Radiology. Seoul National University Coll ege of Medicine ' Department of Ra diology. Ky ungbook National University Co llege of Medi cine. Taegu, Korea Received August ; Accepted October Ad dress reprint requests to: Jae Hyung Park, M.D., Department of Radiology Seoul National Un iversity College of Medicine, # 28 Yongon-Dong Chongno-Ku, Seoul Korea (TIPS) has been reported as an alternative to sclerotherapy or surgical portosystemic shunt in the treatment of acute variceal bleeding in liver cirrhosis patients (1-3). In many Asian countries, the majority of liver cirrhosis is due to hepatitis B or C viral infection, while in western countries it is due to alcoholic abuse (4). To our knowledge, there have been few reports of portal decompression using the TIPS procedure in patients with post-necrotic liver cirrhosis. Because of differing etiologies of liver cirrhosis, there

2 Jae Hyung Park, et al : Transiugular Intrahepatic Portosystemic Shunt might be differences in procedural techniques, the effectiveness of portal decompression, frequency of rebleeding, and shunt related complications In order to evaluate the effectiveness of TIPS in the management of gastroesophageal variceal bleeding and long-term survival in patients with post-necrotic liver cirrhosis, we retrospectively reviewed clinical and radiological data ofthose patients. Materials and Methods Patients Of 70 patients who underwent TIPS procedure during the three years from June 1991 to April 1994, 49 were suffering from post-necrotic liver cirrhosis. Thirty-four were men and 15 were women ; their mean age was 49 years (range, 30-69). All patients had a history of between one and nine variceal bleeding episodes (mean 2.9 times). There was endoscopic evidence of esophageal or esophagogastric varices(n = 20 ; N = 26, respectively). In three patients, only gastric varices were detected. Congestive gastropathy was associated with gastric varices in three, and with esophageal varices in two. The serologic criteria ofprevious viral infection were positive surface antigen and core antibody for hepatitis B infection, and positive antibody for hepatitis C viral infection. There was evidence of previous hepatitis B viral infection in 45 patients and of hepatitis C viral infection in four. The liver function reservoir at the time of the last bleeding was estimated by the Child-Pugh modified scoring system. Child-Pugh class was A in seven patients, B in 16 and C in 26 patients, and 30 were positive for ascites. Imaging diagnoses including sonography (N=49), CT (N=25) and arterial portography (N=16), were performed to assess liver size, portal vein patency and the presence of associated hepatomas. In all cases, endoscopically confirmed bleeding from esophageal or gastric varices was the indication oftips procedure. In these cases, endoscopic sclerotherapy had failed to control the gastric or esophageal varices. The pro cedure was performed on an emergency basis for active bleeding in 37 patients and was elective in 12. In seven Child-Pugh class A patients, the procedure was performed electively in three patients and on an emergency basis in four, because of active bleeding from esophageal or gastric varices. Procedure The right internal jugular vein was. punctured at the middle of the right neck. After introducing a 9-F arterial sheath (Cook Bloomington, Ind) into the jugular vein, an angiographic catheter was inserted into either the right or middle large hepatic vein. We punctured blindly with Colapinto needle in the initial stage (N=31), but in later stages punctures were guided by a fine guide-wire inserted in the right and main portal vein via the percutaneous transhepatic route (N = 12) To demonstrate the portal system, however, we have recently performed superior mesenteric arterial portography (N=6) with or without transfemoral fine guide-wire in the right hepatic via the celiac artery. In all 49 cases, the stent shunt was constructed successfully. Shunts were from the right hepatic (N=35) to the right portal vein (N=27), to the portal bifurcation (N=5), or to the left portal vein (N=3). The shunt was from the middle hepatic (N= 12), to the right portal (N=lO), or to the left portal vein (N=2) (Fig. 1), or from left hepatic to the left portal vein (N=2) Pressure was measured at the main portal vein and hepatic venous pressure was measured through the sheath after pull-back. Portography was performed with a catheter in the splenic and superior mesenteric veins. The balloon was positioned across the parenchymal tract and dilated. In 47 patients we used Wallstent (Schneider, Switzerland). A 7F Wallstent r 빼 A / J 빼 B m 1 률 ; 철닐 Fig. 1. Forty-year-old male patient 1 멸 11. with post-necrotic liver cirrhosis... and variceal bleeding. After punc., ture of left portal vein from middle ' hepatic vein, portal venography was performed. Multiple gastric varicose veins were revealed (A) After creation of shunt between middle hepatic vein and right portal vein with Wallstent, lomm in di J ameter, effective portal decom.. pression was obtained. The portosy,... stemic gradient was 9mmHg (B).

3 J Korean Radiol Soc 1997; 36 : was used. Ifthere was still significant hepatofugal flow into the varices, metallic coil embolization with or without Gelfoam pellets was selectively applied to the varices especially in cases with incomplete portal decompression. To evaluate shunt patency, follow-up duplex sonography was performed every three month. Data anajysis Technical success was defined as the construction of a stent shunt between the hepatic and portal veins and effective portal decompression with a portosystemic pressure gradient less than 15mmHg after the procedure. Successful bleeding control was defined as cessation of bleeding within 24 hours after TIPS. Rebleeding was defined as a new episode of gastrointestinal hemorrhage occurring more than 24 hours of TIPS. The severity of encephalopathy was graded as o-n, based on level of consciousness, intellectual function, and neurologic abnormalities (5). The effectiveness of TIPS for portal decompression and bleeding control was evaluated. Thirty-day mortality and long-term survival were calculated by the Kaplan-Meier method and prognostic factors were analyzed with a pc-sas 6.04 and an IBM compatible PC. Various potential prognostic factors such as the Child-Pugh classification (A & B versus C), emergency versus elective basis ofthe procedure, stent size (8mm n 1 versus 10 아mm미 1 s잉 i펴 dua떠 1 po야 rta 려 vein pressure > 20 아mmHg versus < 20 mmhg), rebleeding (positive versus negative) and post-tips encephalopathy (positive versus negative) were evaluated in terms of the influence of prognostic factors on survival rate. The Wilcoxon test was used for statistical evaluation ofthe groups. Results PortaJ decompression Hemodynamic data was available for 45 patients Portal vein pressure before shunt creation was 29.2 :t 7.0 mmhg and the hepatic vein pressure was 7.8 :t 4.7 mmhg after TIPS procedure ; these changed to 20.3 :t 6.2mmHg and 8.3 :t 5.2mmHg, respectively. The portosystemic gradient improved from 21 :t 6.4mmHg before the procedure to 12.0 :t 5.1 mmhg after ; even in the seven of nine patients with an 8mm stent, the pressure gradient significantly improved from 21.3 :t 4.3mmHg to 13.0 :t 5.7mmHg. Coil embolizations for varices that persisted after TIPS were performed in 12 cases, with additional Gelfoam embolization in two. Successful TIPS creation resulted in control of acute variceal bleeding in 34 cases (92 %) among the 37 emergency patients. Effective portal decompression with a portosystemic pressure gradient of less than 15 mmhg was obtained in 35 ofthe 49 patients (78 %) (Fig. 1). RebJeeding Rebleeding occurred in 17 patients(35 %) during the follow-up period of three and a half years. Rebleeding started within six months in 15 patients, 8 months later in one case, and 2.2years later in one case. Follow-up evaluation with transjugular portal venography revealed stent occlusion in both patients ; redilation of the shunt tract and shunt surgery were successfully performed in both cases. In six of the 15 patients who experienced rebleeding within six months, this occurred within one month of TIPS. The causes of rebleeding in six cases were pre-existing portal vein thrombosis (N=2), acute thrombotic occlusion (N=I) and unknown etiology (N=3). Of the six, only two were successfully controlled by shunt surgery and endoscopic sclerotherapy. The remaining nine cases (18 %) underwent shunt stenosis on transjugular portal venography at the time of rebleeding. In two cases, redilatation of the stent shunt with a balloon catheter was successful in controlling the second bleeding. A second Wallstent was inserted into the first stent shunt in two cases, and in two patients a new second stent shunt was created after making a new tract. In three cases surgery was performed to make a splenorenal shunt. In one case, the shunt was occluded due to portal vein thrombosis, 없 ld in the remaining two cases, rebleeding developed due to stent occlusion. Two patients underwent sclerotherapy and conservative treatment ; in three cases, death ensued due to rebleeding. Hepatic encephajopathy Before the procedure, thirteen patients had hepatic encephalopathy, and in 16 cases (32.7 %) it was aggravated (N=7) or newly developed (N=9) after creation of the intrahepatic shunt. Various degrees of hepatic encephalopathy were present in 20 patients after the procedure. In ten ofthese, conservative treatment with lactulose was effective in controlling the symptoms. In six, however, hepatic encephalopathy progressed and death d ue to hepatic coma ensued within one moqth. Other compj ications In patients with post-necrotic liver cirrhosis, there were several complications other than hepatic encephalopathy and rebleeding after creation of TIPS. These were acute renal failure (N= 2), hemoperitoneum

4 Jae Hyung Park, et al : Transiugular Intrahepatic Portosystemic Shunt catheter was positioned across the tract and the stent was deployed so that it covered the tract and extended 1-3 cm into the hepatic and portal veins. In nine patients, the diameter ofthe stent was 8mm and in 42 it was lomm. An 8-mm balloon was inserted and inf1ated in the stent. Portal vein pressure was measured and if the portal decompression was not satisfactory, a 10mm balloon (N=l), pulmonary edema (N= l), sepsis (N=l), and progressive hepatic failure (N=4). In the patient with hemoperitoneum, f1uoroscopy during the procedure showed extravasation to the peritoneum due to a hepatic capsular puncture ; he died three days later. Long-term surv;val At the end of May 1994, 24 patients had died, 22 were alive and three were no longer included in survival calculation. None of these patients had undergone a liver transplantation. The three who dropped out had undergone a surgical shunt operation after the failure of effective portal decompression with TIPS. The follow-up period of the 49 patients was from one day to three and a half years (mean 383 ::!::: 357 days). The overall one year survival rate, calculated by the Kaplan-Meier method, was 63.8 % and the thirty-day mortality rate was 20.4 %. The causes of death were progressive hepatic failure or hepatic coma (N=9), rebleeding (N=9), renal failure (N=2), hemoperitoneum (N= 1), sepsis (N= 1) and unknown causes (N=2). Prognost;c factors Among many prognostic factors, Child-Pugh class and hepatic encephalopathy were significant. The survival rate of Child-Pugh class A and B patients (mean 633 days) was significantly better than that of Child-Pugh class C patients (mean 279 days), as determined by Wilcoxon analysis (p = ). The one year survival rate of Child-Pugh class A or B patients was 90.0 %, significantly better than the rate of 40.8 % for those who were Child-Pugh class C (Fig. 2). The thirty-day mortality rate ofchild-pugh class A or B patients, 4.4 %, was significantly lower than the rate of 34.6 %, noted for those who were Child-Pugh class C. The survival rate of elective cases was significantly better than that of emergency cases. The presence of hepatic encephalopathy after TIPS, either aggrevated status of preexisting one or newlydeveloped one, was a significant predictive factor for poor prognosis. The mean survival period of patients without hepatic encephalopathy was 450 days, significantly better than the 289 days of those with this condition (p=0.0093) (Fig. 3). All other potential prognostic factors were not significant(table 1). Discussion TIPS has been accepted as a portal decompression procedure in the management ofvariceal bleeding and the efficacy of the procedure has been proven in various types of liver cirrhosis (1-3). The postnecrotic liver, especially after hepatitis B viral infection, is typically shrunken, distorted in shape, and composed of nodules of liver cells separated by dense and broad bands of fi brosis (6). The right hepatic encephalopathy (-) Child C o DIl\ Fig. 2. Long-term survival curve of patients with postnecrotic cirrhosis after TIPS. The survival rate of patients with Child-Pugh class A and B was significantly higher than that of patients with Child-pugh class C o Day Fig. 3. Long-term survival curve of patients with postnecrotic cirrhosis after TIPS. The survival rate of patients with presence of hepatic encephalopathy after TIPS was significantly lower than that of patients without hepatic encephalopathy after TIPS

5 J Korean Radiol Soc '997; 36: Table 1. Variables and Groups in Consideration of Prognostic Factors of TIPS in 49 Patients with Post-necrotic Liver Cirrhosis Variables N Grups I n m P value+ Child-Pugh classification 49 Class A & B (47 %) Class C (53 %) * Endoscopic grade of esophaged varix 49 Grade 0-1 (14 %) Grade II (29 %) Grade ]]J (57 %) Stent diameter 49 8mm (18 %) 10mm (82 %) Post-TIPS portal vein pressure 45 늘 20mmHg (56 %) 20mmHg (44 %) Post-TIPS portosystemic gradient 45 > 12mmHg (42 %) < 12mmHg (58 %) Rebleeding 49 Positive (29 %) Negative (71 %) Post-TIPS hepatic encephalopathy 49 Positive (43 %) Negative (57 %) * + P value by Wilcoxon method, * statistically significant < 0.05 vein sometimes becomes sma11 and narrow due to the shrunken right lobe of the liver. Even with such anatomic distortion, TIPS could be created in a11 patients with post-necrotic liver cirrhosis. In many Asian countries, liver transplantation has not been widely accepted, so TIPS procedure is used only for the purpose of permanent portal decompression, not for the bridging procedure before a subsequent liver transplant. Long-term survival after TIPS is, therefore, important in these patients. There have been many reports of successful portal decompression after TIPS that resulted in a significantly lower residual pressure gradient between the portal and hepatic veins (1-3); after successful TIPS, the mean residual portosystemic gradient was usua11y lower than 12mmHg. In recent TIPS reports, most cases were Laennecs cirrhosis (1-3). In this series with post-necrotic cirrhosis, a high technical success rate was also obtained. The technique for creating a stent shunt between the hepatic and portal veins was diffi cult in those cases with severe atrophy ofthe right lobe due to post-necrotic liver cirrhosis, and the resulting anatomical distortion may have resulted in relatively poor maintenance ofthe shunt. Emergency TIPS successfu11y contro11ed acute variceal bleeding in a11 except three patients. The incidence of rebleeding was reported to be between 6.7 and 36 % (7); it developed in 17 patients(35 %) in our series, mostly within one month of TIPS. The main cause of rebleeding was considered to be restenosis of stent shunt, which was successfu11y managed in six cases. The relatively high incidence of early rebleeding in our series seemed to be due to coagulopathy in those patients who were Child-Pugh class C. In addition to hepatoce11ular carcinoma and bleeding varices, hepatic encephalopathy is one of the serious sequelae of the progressive post-necrotic liver cirrhosis. It is also one of the serious complications of TIPS and its frequency has been reported as being between 5 and 35 % (7). In a series of 100 patients, LaBerge et a1. reported 17 with new or aggravated hep atic encephalopathy out of a total of 96 patients who had successfu11y undergone TIPS (2). In our series, the incidence of hepatic encephalopathy was 32.7 %, relatively high compared to previous reports. In many of our patients it was most likely due to poor liver function, Child-Pugh class C. Ofthe 14 patients with new or aggravated hepatic encephalopathy, eight died because of progressive hepatic failure and hepatic coma. Excluding hepatic encephalopathy and delayed stenosis or occlusion of the stent, the overa11 complication rate in the report of LaBerge et a1. was less than 10 %. Complications were related to techniques, stents or contrast media (7). In our series, complications other than hepatic encephalopathy and rebleeding were hemoperitoneum, renal failure, pulmonary edema and sepsls. We assumed many potential prognostic factors, such as Child-Pugh classification, emergency or elective basis for the procedure, stent size, post-procedural hemodynamic data, rebleeding, and post-tips encephalopathy. Of these, Child-Pugh class and hepatic post-procedural encephalopathy were significant prognostic factors in multivariate analysis. In any treatment modality the prognosis of Child-Pugh class C patients is very poor. The reported operative mortality rate associated with surgical portocaval shunts in cir rhotic patients who are Child-Pugh class C is as high as 50 % (8). LaBerge et a1. reported that the 30-day mor tality rate after TIPS was 24 % in Child-Pugh class C patients (2). In our series, the survival rate of such patients was very poor with a mean survival period of 279 days ; due to hepatic coma, failed bleeding control and other complications, the 30-day mortality rate of Child-Pugh C patients was as high as 34.6 %. A11 except one of such patients underwent the procedure on an 41

6 Jae Hyung Park, et al : Transiugular Intrahepatic Portosystemic Shunt emergency basis. In summary, TIPS is considered to be effective in portal decompression in patients with variceal bleeding due to post-necrotic liver cirrhosis. The overall one year survival rate after TIPS is 63.8 % in patients with post-necrotic liver cirrhosis. During long-term followup, it should be considered that the significant prognostic factors are Child-Pugh classification and post TIPS hepatic encephalopathy. References 1. Richter GM, Noeldge G, Palmaz JC, et aj. Transjugular intrahepatic stent shunt: preliminary clinical results. Radiology 1990; 174: LaBerge JM, Ring EJ, Gordon RL, et aj. Creation of transjugular intrahepatic portosystemic shunts with the Wallstent endoprosthesis : results in 100 patients. Radiology 1993; Hausegger KA, Sternthal HM, Klein GE, Karaic R, Stauber R, Zenker G. Transjugular intrahepatic portosystemic shunt: angiographic follow-up and secondary interventions. Radiology 1994; 191: Podolsky DK, Isselbacher KJ. Cirrhosis. In. Braunwald E, Isselbacher KJ, Petersdorf RG, Wilson JD, Martin JB, Fauci AS ed. Harrison s Principles of Internal Medicine. 11 th ed. New York, McGraw-Hill Book Company, 1987; Schafer DF, Jones EA. Hepatic encephalopathy. In: Zakim D and Boyer TD, ed. Hepatology. 2nd ed. Philadelphia: W. B Saunders Company, 1990 ; Edmondson HA, Peters RL. Liver. In Kissane JM, ed Anderson' s Pathology. 8th ed. St. Louis, Toronto, Princeton. The c. V. Mosby Company, 1985; Freedman AM, Sanyal AJ, Tisnado J, et aj. Complications of transjugular intrahepatic portosystemic shunt: a comprehensive review. RadioGraphics 1993; 13: Cello JP, Grendel JN, Crass RA, et al. Endoscopic sclerotherapy versus portacaval shunt in patients with severe cirrhosis and variceal hemorrhage. N Engl Med 1984; 311 : Warren WD, Henderson JM, Millikan WJ, et aj. Distal splenorenal shunt versus endoscopic sclerotherapy for long-term management of variceal bleeding. Ann Surg 1986; 203 : 대한방사선의학호 IXI 1997; 36: 경경정맥간내문맥간정맥단락술 : 괴사후간경변환자에서의성적과예후인자 1 l 서울대학교의과대학방사선과학교실 2 경북대학교의과대학방사선과학교실 박재형 검용주 2 정진욱 한준구 한만청 목적 : 목정맥을통한간내문맥간정맥단락술 (TIPS) 을이용한괴사후간경변환자의위 식도정맥류출혈치료에있어서그성적과예후인자를알아보고자한다. 대상및방법 : 최근 3년간 TIPS시술을받은괴사후간경변환자 49 명을대상으로하였다. 이들중 B 형감염표식자양성환자는 45 명이었고 C 형간염표식자양성은 4 명이었다. 전례에서시술의적응증은위 식도정맥류출혈이었다. 시술전최종출혈시 Child-Pugh 분류는 A가 7 명, B가 16 명, C 가 26 명이었다. 시술에의한문맥감압과지혈효과를분석하였으며 Kaplan-Meier 방법으로 TIPS 시술후의장기생존율을구하였으며 Wilcoxon 테스트로서예후인자를검정하였다. 결과시술은전례에서기술상성공하였다. 문맥체정맥압력차가시술전 21.4 :t 6.4 mmhg에서수술후 12.0 :t 5.1mmHg (N = 45) 로감소하였다.37 명의응급환자중 34 명에서정맥류의활동성출혈이지혈되었다. 추적기간은 3년 6 개월 ( 평균 383 :t 357 일 ) 이였으며이기간중재출혈은 17 명 (35%) 에서있었고간성뇌병증은 16 명 (32.7%) 에서새로나타나거나악화되었다.30 일사망율은 20.4% 이며일년생존율은 63.8% 이었다. 예후가불량할것으로예상할수있는예후인자는 Child-Pugh 분류 C 와시술후의간성뇌병증이었다. 결론 : TIPS 는위 식도정맥류출혈이있는괴사후간경변환자에있어서효과적인문맥감압방법이다. 의미있는예후인자는 Child-Pugh 분류 C 와시술후의간성뇌병증이었다