TIPS for Refractory Ascites: A 6-Year Single-Center Experience With Expanded Polytetrafluoroethylene Covered Stent-Grafts

Transcription

1 Vascular and Interventional Radiology Original Research Bercu et al. TIPS for Refractory Ascites Vascular and Interventional Radiology Original Research Zachary L. Bercu 1 Aaron M. Fischman 1 Edward Kim 1 F. Scott Nowakowski 1 Rahul S. Patel 1 Thomas D. Schiano 2 Charissa Y. Chang 2 Robert A. Lookstein 1 Bercu ZL, Fischman AM, Kim E, et al. Keywords: cirrhosis, expanded polytetrafluoroethylene (eptfe) covered stent, portal hypertension, refractory ascites, transjugular intrahepatic portosystemic shunt (TIPS) DOI: /AJR Received March 14, 2014; accepted after revision October 4, Z. L. Bercu received a grant from 4D Technology. A. M. Fischman is a consultant to Terumo Interventional Systems, Surefire Medical, and CeloNova BioSciences and is a lecturer for Philips Healthcare. E. Kim is a consultant to BTG International and Philips Healthcare; is on the scientific advisory board of Onyx Pharmaceuticals; and is a lecturer for BTG International and Philips Healthcare. R. S. Patel is a consultant to Sirtex and Arstasis. T. D. Schiano is a consultant to Biotest International, Bristol-Myers Squibb/Sanofi-Aventis Partnership, Novartis, Pfizer, and Salix Pharmaceuticals; is on the scientific advisory board of Gilead Sciences, Idenix Pharmaceuticals, Janssen Pharmaceuticals, and Merck; is a lecturer for Novartis; and is involved in research support for Genentech. C. Y. Chang is a consultant to Gilead Sciences. R. A. Lookstein is a consultant to Cordis, Boston Scientific, Bayer Healthcare, and The Medicines Company; has received education grants from Philips Healthcare and Terumo Medical; and has received research support from Surefire Medical. This article is available for credit. AJR 2015; 204: X/15/ American Roentgen Ray Society TIPS for Refractory Ascites: A 6-Year Single-Center Experience With Expanded Polytetrafluoroethylene Covered Stent-Grafts OBJECTIVE. This single-center study evaluated the use of expanded polytetrafluoroethylene (eptfe) covered stent-grafts for transjugular intrahepatic portosystemic shunt (TIPS) placement to manage portal hypertension related refractory ascites. MATERIALS AND METHODS. One hundred patients at a single tertiary care center in a major metropolitan hospital underwent TIPS placement with an eptfe-covered stent-graft (Viatorr TIPS Endoprosthesis). Patients with portal hypertension related ascites and preexisting hepatocellular carcinoma or liver transplant were excluded from the analysis. Records were reviewed for demographic characteristics, technical success of the TIPS procedures, and stent follow-up findings. Clinical results were assessed at 90- and 180-day intervals. RESULTS. Immediate technical success of the TIPS procedure was 100%. Of the 61 patients with documented follow-up, 55 (90.2%) had a partial or complete ascites response to TIPS creation. Of these 55 patients, nine experienced severe encephalopathy. Six of 61 patients (9.8%) did not experience a significant ascites response. Overall survival was 78.7% at 365-day follow-up. The 365-day survival was 84.2% for patients with a model for end-stage liver disease (MELD) score of less than 15, 67.0% for those with a score of 15 18, and 53.8% for those with a score of greater than 18 (p = 0.01). For patients with a MELD score of less than 18, the 365-day survival was 88.0% for those with an albumin value of 3 mg/dl or greater and 72.8% for those with an albumin value of less than 3 mg/dl (p = 0.04). CONCLUSION. TIPS placement using an eptfe-covered stent-graft is an efficacious therapy for refractory ascites. Patients with preserved liver function characterized by a MELD score of less than 15 or a MELD score of less than 18 and an albumin value of 3 mg/ dl or greater experience the greatest survival benefit. I n 2009, cirrhosis was the cause of 31,522 deaths and was the 12th leading cause of death in the United States [1]. Portal hypertension is caused by intrahepatic obstruction of portal blood flow and transmitted pressure from the inferior vena cava (IVC) [2 4]. Complications include ascites, hepatorenal syndrome [5 9], variceal bleeding, and spontaneous bacterial peritonitis [5, 6]. Refractory ascites is defined by an inability to mobilize ascites despite maximal doses of diuretics [10] and has a significant morbidity [7]. Therapeutic options for refractory ascites include frequent paracentesis, transjugular intrahepatic portosystemic shunt (TIPS) place- ment, and liver transplant [8, 9, 11]. TIPS improves serum sodium and creatinine levels [12, 13] and is more effective in controlling ascites than large-volume paracentesis [14]. Early investigations have shown lower rates of ascites recurrence and lower hepatorenal syndrome risk in patients with TIPS as compared with patients who undergo repeated paracentesis [15 20]. Several metaanalyses have shown that TIPS is associated with reduced ascites recurrence and improved transplant-free survival in comparison with large-volume paracentesis but that TIPS is also associated with an increased rate of hepatic encephalopathy [13, 21 23]. Improved technique focuses on portosystemic gradient 1 Department of Radiology, Division of Interventional Radiology, Icahn School of Medicine at Mount Sinai, 1 Gustave L. Levy Pl, Box 1234, New York, NY Address correspondence to Z. L. Bercu (zachary.bercu@mountsinai.org). 2 Division of Liver Diseases, Recanati Miller Transplantation Institute, Icahn School of Medicine at Mount Sinai, New York, NY. 654 AJR:204, March 2015

2 TIPS for Refractory Ascites (PSG) optimization through submaximal dilatation with balloon angioplasty [24]. Improved outcomes after TIPS placement relies on the selection of appropriate candidates for which post-tips mortality is expected to be low. Factors associated with increased mortality include Child-Pugh class C cirrhosis, model for end-stage liver disease (MELD) score of greater than 25, PSG of less than 8 mm Hg [25], international normalized ratio (INR) of greater than 2 [26], total serum bilirubin value of greater than 3 mg/dl, and platelet count of less than /L [27]. Factors associated with improved survival are a MELD score of less than 15 [28], serum creatinine value of less than 2 mg/dl, and serum bilirubin value of less than 2 mg/dl [29]. Serum sodium, bilirubin, and creatinine values may predict the likelihood of encephalopathy [30]. The use of a covered stent-graft for TIPS placement has led to studies revisiting predictors of post-tips outcomes. Polytetrafluoroethylene is believed to improve shunt patency by reducing ingrowth of liver parenchyma and transmural deposition of bile and mucin [31]. The use of TIPS with an expanded polytetrafluoroethylene (eptfe) covered stent-graft for refractory ascites is widespread despite the lack of large controlled studies [28, 32, 33]. Nevertheless, the 2009 American Association for the Study of Liver Diseases guidelines update recommends the use of eptfe-covered TIPS to decrease the incidence of shunt dysfunction and of portal hypertension complication recurrence [34]. In a randomized study, patients with eptfe-covered TIPS had significantly lower rates of shunt dysfunction and clinical relapse than patients with uncovered TIPS; however, only 32 of 80 patients underwent TIPS for refractory ascites [35]. The effect of eptfe-covered TIPS on survival is controversial [36]. Cost-effectiveness has yet to be determined, although preliminary data suggest a reduction in costs with eptfe-covered TIPS [37]. The purpose of the current study was to review a single-center series experience with eptfe-covered TIPS for the management of refractory ascites in the setting of portal hypertension. Materials and Methods Study Population Institutional review board approval was obtained for a retrospective study that met the criteria for patient privacy and HIPAA maintenance. Over a 6-year period, 100 consecutive patients at a single quaternary liver center in a major metropolitan hospital underwent eptfe-covered TIPS placement for refractory ascites in the setting of portal hypertension. Patients with a prior liver transplant and those with hepatocellular carcinoma were excluded. All patients underwent echocardiography before TIPS to confirm the absence of significant right or left cardiac dysfunction given the high rates of hepatopulmonary syndrome in this population [38]. Significant cardiac dysfunction may suggest that the patient has increased right atrial pressures, precluding an effective outcome from TIPS placement. No TIPS procedure was performed in a patient with clinically significant cardiac dysfunction, which was defined as diastolic dysfunction. Dysfunction was diagnosed by the patient s internist or cardiologist or by a radiologist based on echocardiography measures suggestive of diastolic dysfunction, such as early maximal ventricular filling velocity to late filling velocity ratio of 1 or less. TIPS Placement Technique One hundred patients underwent eptfe-covered TIPS placement using a 10-mm-diameter Viatorr TIPS Endoprosthesis (W. L. Gore & Associates) for the management of refractory ascites between January 24, 2006, and June 21, The TIPS procedures were performed as elective outpatient procedures while the patient was under general anesthesia, and all patients were admitted to the hospital after TIPS placement for close observation. All operators were interventional radiologists experienced in performing TIPS. Stent-graft lengths ranged from 60 to 100 mm. Paracentesis was performed in every patient referred for TIPS for refractory ascites in the angiography suite just before TIPS placement to reduce the risk of uncontrolled bleeding in the event of capsule puncture and to reduce the effect of ascites on hepatic vessel anatomy distortion. A standard technique was performed using a 16-gauge needle (Ross Modified Colapinto Needle, which is included in the Haskal Transjugular Liver Access Set, Cook Medical) to cross from either the right or middle hepatic vein to the right portal vein. The target portal vein was visualized in advance of the puncture via CO 2 portography. All patients underwent a single procedure at the time of the initial TIPS procedure. No adjunct technique, such as transhepatic or transsplenic access, was used. For six of the 100 patients, TIPS was performed from the left hepatic vein to the left portal vein. For two patients, TIPS was performed from the middle hepatic vein to the left portal vein. For two patients, TIPS was performed from the right hepatic vein to the left portal vein. For one patient, TIPS was performed from the IVC directly to the left portal vein. For the remaining 89 patients, TIPS was performed from the right hepatic vein to the right portal vein. For three patients, the stentgraft was extended into the IVC due to a small hepatic vein landing zone from the site of puncture to the IVC or due to the presence of hepatic vein thrombosis. Variations in TIPS tract creation were likely secondary to evaluation of pre-tips crosssectional imaging leading to tract planning. Additional stents were required in nine patients at the time of initial TIPS placement. Secondary stents included bare metal stents in six patients and covered stent-grafts in two patients. One patient had both a bare metal stent placed along the proximal portal edge and a second covered stent-graft added to the distal hepatic venous end. Pressure gradients were measured before and after stent placement. An intraabdominal reference point of right atrial pressure was used. The PSG was measured as the difference between the hepatic vein wedge or portal vein pressure and the right atrial pressure, representing the standard measurements obtained at our institution. All stents were expanded to 6 mm using a balloon. Increasing balloon expansion was then performed until the optimal pressure gradient was achieved (target PSG, 7 12 mm Hg). For 75 patients, balloon expansion was less than the diameter of the stent-graft. For five patients, target PSG could not be obtained despite maximum balloon dilatation of 10 mm. Routine TIPS screening was performed at approximate 6-month intervals with Doppler sonography. The reasons for nonroutine TIPS ultrasound evaluation included recurrent ascites, increasing abdominal girth, and new or recurrent gastrointestinal bleeding. The reasons for TIPS revision included TIPS stenosis or occlusion and severe hepatic encephalopathy. TIPS revisions for stenosis or occlusion involved venoplasty with or without new stent placement targeting a reduction in PSG and rheolytic thrombolysis in the setting of occluded TIPS. The placement and choice of a stent were based on operator assessment and preference after initial venoplasty. The TIPS revision for hepatic encephalopathy involved an attempted increase in PSG through placement of a crimped stent. The standard technique for crimped stent placement at our institution is deployment and 6-mm balloon angioplasty of the portal one third of a stent while the remaining stent continues to be sheathed, followed by unsheathing, and 6-mm balloon angioplasty of the hepatic venous one third of the stent. This creates an hourglass configuration of the bare metal stent. Outcome Measures Records were reviewed for patient demographics, immediate technical success, and clinical results. Immediate technical success was defined as AJR:204, March

3 Bercu et al. successful shunt deployment with confirmed portosystemic shunt patency on completion direct portography. At approximately 90- and 180-day intervals, the following were assessed: the presence and severity of ascites; presence and severity of encephalopathy; and serum sodium, creatinine, total bilirubin, albumin, and INR. Ascites severity was defined on a 3-point scale as absent, mild to moderate (present but controlled medically or with infrequent paracentesis), or severe (controlled with frequent paracentesis or requiring hospitalization). Encephalopathy severity was defined on a 3-point scale as absent, mild to moderate (present but controlled with two or fewer medications), or severe (requiring hospitalization or controlled with more than two medications). Complete response was defined as new absence of ascites after TIPS. Partial response was defined as a decrease in the 3-point severity score of ascites from pre-tips assessment. Development of significant encephalopathy was defined as a severity score increase from the pre-tips assessment. The development of severe encephalopathy was defined as a 2-point increase in the severity score. The MELD score was calculated at pre-tips consultation. The serum creatinine value was recorded at 30-day follow-up to assess for acute renal failure, which was defined as an increase of 0.5 mg/dl or greater. Survival was assessed using hospital records and the Social Security Death Index (SSDI) based on an intention-to-treat model. Separate analyses were performed to assess transplant-free survival and to provide data for comparison with prior studies. Revision-free survival was also assessed. Statistical Analysis Comparison of clinical values before and after TIPS placement, risk factors for severe encephalopathy, and changes in clinical values for patients with independent postprocedure severe encephalopathy were assessed using the paired Student t test. Missing data were excluded using a standard delimiter. Survival, transplant-free survival, revisionfree survival, and both transplant- and revision-free survival were calculated using Kaplan-Meier analysis. A multivariate analysis was performed using a Cox proportional hazards regression model with Breslow s method for ties to assess statistically significant pre-tips covariates. The probability of the chi-square test on the log ratio was used to determine the statistical significance of the multivariate analysis. Hazard ratios (HRs) were determined for all clinically significant covariates. The overall survival distribution function was then stratified by MELD scores of less than 15, 15 18, and greater than 18. For patients with a MELD score of less than 18, the overall survival distribution function TABLE 1: Clinical Characteristics of the Study Population Characteristic was stratified for those with an albumin value of 3 mg/dl or greater and those with an albumin value of less than 3 mg/dl. Statistics software (R, version 1.63, R Project for Statistical Computing) was used for statistical analysis. Results Eight patients were excluded because of preexisting hepatocellular carcinoma or liver transplant before TIPS; thus, 92 patients remained in the analysis. Immediate technical success at TIPS creation was 100% (92 patients). The average age of the 92 patients was 55.8 years (range, years), and 38% were female and 62% were male. The causes of portal hypertension in the study group are shown in Table 1. The mean Child-Pugh score was 9.9 (range, 7 14). Disease was defined as Child-Pugh B in 41 patients (44.6%) and as Child-Pugh C in 48 patients (52.2%); the remaining three patients (3.3%) did not have sufficient data for calculation of a pre-tips Child-Pugh score. The mean MELD score before TIPS placement was 13.8 (range, 7 33). Value Total no. of eligible patients who underwent eptfe-covered TIPS creation for 92 refractory ascites Age (y) Mean ± SD 55.8 ± 9.5 Sex, no. (%) of patients (n = 100 patients) Female 63 (63) Male 37 (37) Cause of portal hypertension, no. (%) of patients (n = 100 patients) Hepatitis C cirrhosis 31 (31.0) Alcoholic cirrhosis 31 (31.0) Cryptogenic cirrhosis 10 (10.0) Budd-Chiari syndrome or other venoocclusive disease 8 (8.0) Nonalcoholic steatohepatitis cirrhosis 8 (8.0) Hepatitis B cirrhosis 6 (6.0) Primary biliary cirrhosis 3 (3.0) Autoimmune cirrhosis 3 (3.0) Child-Pugh score, mean ± SD 9.9 ± 1.5 MELD score before TIPS creation, mean ± SD 13.8 ± 4.8 Pressure gradient (mm Hg), mean ± SD Before TIPS creation 19.2 ± 6.1 After TIPS creation 7.8 ± 3.5 Clinical follow-up time (d), mean ± SD 473 ± 579 Note eptfe = expanded polytetrafluoroethylene, TIPS = transjugular intrahepatic portosystemic shunt, MELD = model for end-stage liver disease. The mean pre-tips and post-tips pressure gradients were 19.2 and 7.8 mm Hg, respectively. Independent pre-tips right atrial, free hepatic vein, and wedged hepatic vein pressures were reported in 58, seven, and 64 patients, respectively. Independent post-tips right atrial and portal vein pressures were reported in 69 patients. Average pre-tips right atrial, free hepatic vein, and wedged hepatic vein pressures were 11.9 mm Hg (range, 0 23 mm Hg), 14.3 mm Hg (range, 3 23 mm Hg), and 29.8 mm Hg (range, mm Hg). The average post-tips right atrial and portal vein pressures were 16.3 mm Hg (range, 7 28 mm Hg) and 23.3 mm Hg (range, mm Hg). The mean clinical follow-up time was 473 days (range, days). Eight of 92 patients (8.7%) underwent postprocedural liver transplant during the study period. Ten of 92 patients (10.9%) were lost to follow-up immediately after discharge after TIPS. For the five patients who did not reach targeted PSG after maximal balloon dilatation of 10 mm, three experienced partial or complete reduction in ascites and 656 AJR:204, March 2015

4 TIPS for Refractory Ascites two experienced worsening encephalopathy. None of these patients developed severe encephalopathy. The average pre- and post- TIPS PSG for these five patients were 21.3 and 17.6 mm Hg, respectively. Twelve of 92 patients (13.0%) required TIPS revision over the study period. Only one of the four patients for whom the target PSG was not obtained required TIPS revision. One patient (8.3% of revisions) underwent unsuccessful placement of a crimped 10-mm-diameter bare metal stent (stent length not documented) to target an increased PSG. After the original TIPS placement, PSG was 6 mm Hg; PSG before TIPS had been 25 mm Hg. The specific PSG value after TIPS revision was not documented. Given persistent severe encephalopathy in this patient, the TIPS eventually was deliberately occluded using a 16-mm plug (Amplatzer Vascular Plug II, St. Jude Medical). The patient subsequently died 32 days after complete TIPS occlusion and 507 days after the initial TIPS placement. Of the four patients (33.3% of revisions) with recurrent ascites or increasing abdominal girth requiring TIPS revision, two patients were found to have occlusion of the TIPS on sonography, one patient had an increase in velocities on sonography from the proximal to middle portion of the TIPS, and one patient had normal findings on ultrasound. One of two patients who had TIPS occlusion underwent successful rheolytic thrombolysis using an AngioJet Thrombectomy System (Boston Scientific) and placement of a new proximal 12 8 mm bare metal stent to the inferior mesenteric vein and a new distal 10 8 mm eptfe-covered Viatorr TIPS to the right atrium given a small landing zone free of thrombus in the IVC. The patient experienced a reduction in the PSG from 26 to 24 mm Hg but was noted to have angiographic improvement. The other patient did not have multifocal stenoses on venography and underwent 8-mm balloon venoplasty and placement of a new mm eptfe-covered Viatorr TABLE 2: Acute Complications Within 30 Days of Transjugular Intrahepatic Portosystemic Shunt (TIPS) Placement Complication No. (%) of Patients (n = 100) TIPS occlusion or malfunction 2 (2.0) Acute renal failure 7 (7.0) Death 5 (5.0) Cumulative Survival Time After TIPS Placement (d) Fig. 1 Cumulative overall survival results in number of days after transjugular intrahepatic portosystemic shunt (TIPS) placement for refractory ascites are shown. Overall survival was 85.1% at 180-day follow-up after TIPS and 78.7% at 365-day follow-up after TIPS. Solid line = survival curve, dashed line = 95% confidence limits. TIPS; the patient experienced a reduction in the PSG from 14 to 9 mm Hg. The other two patients with recurrent ascites or increasing abdominal girth underwent 8- and 10-mm balloon venoplasty; at follow-up, their symptoms had improved and the mean PSG value was reduced from 10.5 to 4.0 mm Hg. For five patients (41.7% of revisions) with abnormal findings on routine sonography requiring TIPS revision, one patient had hepatic venous stenosis confirmed on both sonography and venography. The patient underwent distal shunt placement with a mm bare metal stent, and PSG decreased from 13 mm Hg before placement to 7 mm Hg after placement. One patient had interval diminished velocities on sonography that correlated with multifocal stenoses on venography. The patient underwent successful venoplasty and placement of a new mm eptfe-covered Viatorr TIPS; a PSG reduction from 24 to 9 mm Hg was noted. One patient had nonocclusive thrombosis on sonography and underwent rheolytic thrombolysis with placement of a new mm eptfe-covered Viatorr TIPS. No PSG measurements were reported for this patient. One patient had hepatic venous stenosis and underwent balloon venoplasty. No PSG measurements were reported for this patient. For one patient, the results of ultrasound performed at an outside facility were not available. On venography, the patient was found to have multifocal stenoses requiring 10-mm balloon venoplasty followed by placement of a mm bare metal Viatorr TIPS and a mm eptfe-covered Viatorr TIPS distally; a PSG reduction from 16 to 9 mm Hg was noted. The revision ultimately failed, and the patient underwent placement of new parallel TIPS stent-grafts. The last patient was found to have a proximal TIPS stenosis and underwent successful venoplasty; a PSG reduction from 12 to 6 mm Hg was noted. One patient (8.3% of revisions) with new gastrointestinal bleeding and persistent recurrent ascites requiring TIPS revision was found to have interval elevated distal velocities on sonography correlating with hemodynamically significant hepatic venous stenosis on venography. The TIPS was extended using a new mm eptfe-covered Viatorr TIPS to the right atrium given a small distal landing zone beyond the hepatic venous stenosis. PSG was reduced from 13 mm Hg before the procedure to 7 mm Hg after the procedure. Seven of 67 patients with follow-up laboratory data at 30 days after TIPS placement (10.4%) experienced acute renal failure (serum creatinine increase, 0.5 mg/dl). Three of these patients experienced reversal of acute renal failure at 60 days, and the remaining four were lost to follow-up. Table 2 shows the number of patients with acute complications detected within 30 days of TIPS. Thirty-one patients did not have ascites documentation at follow-up; 55 of 61 patients (90.2%) experienced a partial or complete response at follow-up with respect to ascites. For the six patients who did not experience a partial or complete response with respect to ascites, four (66.7%) had a postintervention PSG of greater than 8 mm Hg. All six patients were managed with frequent largevolume paracentesis. AJR:204, March

5 Bercu et al. Thirty-one of the 92 patients included in the study did not have encephalopathy documentation at follow-up. Twenty-five of 61 patients (41.0%) had no encephalopathy at follow-up. Thirty-six of 61 patients (59.0%) had developed encephalopathy at follow-up, 12 (19.7%) of whom had severe encephalopathy. Of the 55 patients with partial or complete ascites response at follow-up, 32 (58.2%) experienced postprocedure encephalopathy. Nine of the 55 patients (16.4%) with partial or complete ascites response experienced severe postprocedure encephalopathy. Three patients with severe encephalopathy had no ascites response at follow-up. The risk factors for severe encephalopathy in univariate analysis were age, Child-Pugh score, MELD score, total bilirubin value, creatinine value, sodium value, and albumin value (p < 0.01 for all factors). Of the patients with a postintervention PSG of 8 mm Hg or less, 47.8% experienced worsening of encephalopathy in comparison with only 8.7% of patients with a higher PSG. Table 3 shows the mean clinical values (MELD scores and laboratory values) before TIPS and at follow-up. The overall survival was 85.1% at 180-day follow-up and 78.7% at 365-day follow-up. Figure 1 shows overall survival. Transplant-free survival was 85.1% and 74.5% at 180- and 365-day follow-up, respectively. Revision-free survival was 80.9% and 75.5% at 180- and 365-day follow-up, respectively. Transplant- and revision-free survival was 79.8% and 70.2% at 180- and 365- day follow-up, respectively. Figure 2 shows results for transplant-free, revision-free, and transplant- and revision-free survival. The multivariate analysis for overall survival shows that the clinically significant pre- TIPS covariates are MELD score (p = 0.02) TABLE 3: Mean Clinical Values in 92 Patients Before and After Transjugular Intrahepatic Portosystemic Shunt (TIPS) Placement Clinical Variable Before TIPS Procedure 90-Day Follow-Up 180-Day Follow-Up MELD score 13.7 ± ± 7.1 a 16.9 ± 9.0 b INR 1.4 ± ± 0.5 a 1.7 ± 0.8 Total bilirubin (mg/dl) 1.9 ± ± 8.0 b 3.4 ± 4.1 b Creatinine (mg/dl) 1.2 ± ± ± 1.3 Sodium (mmol/l) 135 ± ± ± 5.2 b Albumin (mg/dl) 3.0 ± ± 0.7 b 3.1 ± 0.8 Note MELD = model for end-stage liver disease, INR = international normalized ratio. a p < b p < and albumin value (p = 0.01). Table 4 shows the HRs for all measured covariates of overall survival before TIPS. The probability for the chi-square test on the log ratio for overall survival was 0.02, representing a good fit. TABLE 4: Cox Proportional Hazards Model for Covariates of Overall Survival Characteristic Cumulative Survival Probability Greater Than χ 2 (p value equivalent) Hazard Ratio Transplant-free survival Revision-free survival Transplant- and revision-free survival Time After TIPS Placement (d) Fig. 2 Cumulative transplant-free survival, revision-free survival, and transplant- and revision-free survival results in number of days after transjugular intrahepatic portosystemic shunt (TIPS) placement for refractory ascites are shown. At 180- and 365-day follow-up after TIPS, transplant-free survival was 85.1% and 74.5%, respectively; revision-free survival was 80.9% and 75.5%; and transplant- and revision-free survival was 79.8% and 70.2%. The overall survival distribution function was stratified by MELD score (Fig. 3). The 365-day survival was found to be 84.2%, 67.0%, and 53.8% for MELD scores less than 15, 15 18, and greater than 18, respec- Hazard Ratio (95% Confidence Limits) Lower Limit Upper Limit Age Presence of varices on endoscopy, CT, or MRI Prior gastrointestinal bleed Child-Pugh score MELD score Total serum bilirubin Creatinine Sodium Albumin Note MELD = model for end-stage liver disease. 658 AJR:204, March 2015

6 TIPS for Refractory Ascites tively (p = 0.01). For patients with MELD scores of less than 18, the overall survival distribution function was stratified by serum albumin value (Fig. 4): The 365-day survival was 88.0% for those with an albumin value of 3 mg/dl or greater and 72.8% for those with an albumin value of less than 3 mg/dl (p = 0.04). Discussion Prior studies of TIPS placement for refractory ascites have focused on bare metal stents and small populations with refractory ascites [15, 20 23, 35, 39]. This study confirms the immediate technical success of eptfe-covered TIPS at shunt creation and 1-year survival was very high. Most patients experienced partial or complete resolution of refractory ascites. Many experienced some form of encephalopathy. A low but significant number experienced severe encephalopathy. The rate of nonsevere encephalopathy emphasizes the narrow therapeutic window between ascites reduction and encephalopathy in this patient population. The relatively high rate of encephalopathy in this study among patients with partial or complete ascites response (58.2%) appears to be higher than rates reported in the general literature dealing with TIPS using covered stents for both refractory ascites and variceal bleeding (21 27%); however, this finding may reflect better detection of nonsevere encephalopathy and specific encephalopathy risk in the population with recurrent ascites [39, 40]. A target PSG of greater than 8 mm Hg confers a lower risk of encephalopathy and should especially be considered in patients with a history of encephalopathy. Given the low rates of persistent ascites in the study, the presence of a TIPS and relative PSG reduction likely are more critical to ascites reduction than targeting a specific PSG number for all patients. The optimal PSG may be highly individualized. For the six patients who did not experience a partial or complete response with regard to ascites, four had postintervention PSG of greater than 8 mm Hg. It is likely that the target PSG that is needed to achieve ascites reduction in these patients was not reached. For the other two patients, it is likely that the advanced nature of the recurrent ascites could not be addressed through optimal PSG reduction. At our institution, most operators will leave postintervention PSG on the high end of the targeted range given the ease with which one can subsequently lower PSG in most cases with balloon dilatation. Cumulative Survival Time After TIPS Placement (d) < > 18 Fig. 3 Cumulative overall survival results in number of days after transjugular intrahepatic portosystemic shunt (TIPS) placement for refractory ascites stratified by model for end-stage liver disease (MELD) score are shown. Survival at 365 days after TIPS placement was 84.2%, 67.0%, and 53.8% for patients with MELD score of less than 15, 15 18, and greater than 18, respectively. Differences between curves normalize at approximately 1100 days ( 3 years). Cumulative Survival MELD score < 18 and albumin value 3 mg/dl MELD score < 18 and albumin value < 3 mg/dl Time After TIPS Placement (d) Fig. 4 Cumulative overall survival results in number of days after transjugular intrahepatic portosystemic shunt (TIPS) placement for refractory ascites stratified by albumin value are shown for patients with model for end-stage liver disease (MELD) score of less than 18. Survival at 365 days after TIPS placement was 88.0% for patients with albumin value of 3 mg/dl or greater and 72.8% for patients with albumin value of less than 3 mg/ dl. Difference between curves normalizes at approximately 1000 days, which is slightly less than 3 years. Increasing PSG typically is more complicated and requires advanced techniques such as placement of a crimped stent or deliberate occlusion of a TIPS. Although seven of 67 patients had acute renal failure after TIPS placement, renal failure did not persist after 30 days for the three patients with follow-up. The cause of the acute renal failure is uncertain, however, in these patients. Given these results, TIPS placement in the setting of refractory ascites is not associated with a chronic decline in renal function. Slight worsening of MELD scores and INR may be ascribable to TIPS placement in the short term but likely reflects the natural progression of liver disease over time. Worsening of total bilirubin values is slightly more than previously reported; however, this finding may reflect a sicker population as shown by higher Child-Pugh scores [41]. Improved sodium levels at follow-up may reflect improved oncotic or hydrostatic pressure. The slight improvements in albumin values appear transient, which may reflect the natural progression of liver disease. Although the average MELD score before TIPS was low, Child-Pugh scores were relatively high. Of note, patients with lower MELD scores may still qualify as Child-Pugh class C with a low albumin value. Given the severity of disease in this patient population, earlier TIPS placement may be warranted. AJR:204, March

7 Bercu et al. In the current study, results for immediate overall, transplant-free, revision-free, and transplant- and revision-free survival remained high. An overall survival of 78.7% at 1 year was found in the current population. In comparison, Membreno et al. [40] reported in 2005 a 58.0% overall survival in patients with ascites and a Child-Pugh score of less than 10 at 1 year after TIPS with bare metal stents. Survival measure improvements with eptfecovered stents over bare metal stents may be related to increased patency rates as previously described [31]. In 2004, Bureau et al. [39] reported 70.9% transplant-free survival at 1 year with eptfe-covered stents in patients with ascites or variceal bleeding [39]. Transplant-free survival at 1 year was 74.5% in the current population of patients with refractory ascites. Improvements in survival measures seen in our study compared with the results of prior studies may reflect improved technical success and pressure gradient optimization. The use of eptfe-covered TIPS also likely contributes to the higher 1-year survival seen in our study, especially given the low Child- Pugh scores reported by Membreno and colleagues for their study group. Increasing MELD scores and decreasing albumin levels are related to overall survival. In particular, MELD scores of less than 15 confer the greatest likelihood of survival at 1 year after TIPS placement in the setting of refractory ascites. MELD scores of confer a higher likelihood of survival at 1 year compared with higher MELD scores. Similar findings have been suggested previously by extrapolating meta-analysis data for hypothetic patients with specific MELD scores [21]. Given the absence of albumin values in MELD score calculations, interventionalists should consider measures of albumin independently when assessing risk in this setting. In patients with a MELD score of less than 18, albumin levels of 3 mg/dl or greater confer the greatest likelihood of survival at 1 year. At approximately 3 years, however, the survival curves begin to overlap and no significant survival benefit is apparent for patients with lower MELD scores and higher albumin levels before intervention. Patients with preserved liver function characterized by a MELD score of less than 15 or a MELD score of less than 18 with an albumin value of 3 mg/ dl or greater experience the highest levels of survival and are ideally suited to undergo TIPS for refractory ascites. A postintervention PSG of 8 mm Hg or less was associated with worsening encephalopathy; this result reaffirms the need for pressure gradient optimization, particularly with regard to gradual balloon dilatation. Age, Child-Pugh score, total bilirubin value, creatinine value, and sodium value do not appear to be related to overall survival but, along with MELD score and albumin value, may contribute to the risk of severe encephalopathy. The limitations of the study include the use of retrospective analysis, single-center evaluation, and clinical reporting of ascites and encephalopathy by providers. The reasons for variations in TIPS placement and stent extension were not apparent given the retrospective nature of this study. Reliance on the SSDI may have missed patients who died outside the United States or those who did not have legal status in the United States, although this number is suspected to be low and of minimal clinical significance. Patients who had portal hypertension due to Budd-Chiari syndrome or other venoocclusive disease may have a different need for liver transplantation and likelihood of portosystemic encephalopathy in comparison with patients with cirrhosis. Prior studies suggest an increased risk of shunt dysfunction after uncovered TIPS placement with a lower rate of hepatic encephalopathy [42]. Nevertheless, we think that it is important to include this population given the use of TIPS for refractory ascites in this population and because of the presumed greater technical complexity of TIPS placement in this setting. Another limitation is the lack of a comparison with uncovered TIPS because of the near ubiquity in the use of eptfe-covered stents for TIPS. Most patients with recurrent ascites improvement after TIPS likely experience a reduced need for ascites-related medications such as spironolactone. Despite medication dose requirement reductions, most patients probably remain on a smaller dose of these medications. Assessment of medication follow-up was beyond the scope of the current study. Future prospective studies should address quantitative measurements of ascites and encephalopathy in addition to long-term detailed follow-up. Refractory ascites in the setting of cirrhosis portends a poor outcome for patients, particularly with regard to morbidity and mortality. The use of an eptfe-covered stent-graft for TIPS provides an efficacious opportunity and a critical therapy, albeit with high rates of encephalopathy, to intervene in this process. References 1. Yoon Y, Yi H. Liver cirrhosis mortality in the United States, Arlington, VA: National Institute on Alcohol Abuse and Alcoholism, Division of Epidemiology and Prevention Research, Alcohol Epidemiologic Data System, 2012: Surveillance report Cichoz-Lach H, Celinski K, Slomka M, Kasztelan-Szczerbinska B. Pathophysiology of portal hypertension. J Physiol Pharmacol 2008; 59(suppl 2): de Franchis R; Baveno V Faculty. Revising consensus in portal hypertension: report of the Baveno V consensus workshop on methodology of diagnosis and therapy in portal hypertension. J Hepatol 2010; 53: 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: Dib N, Oberti F, Calès P. Current management of the complications of portal hypertension: variceal bleeding and ascites. Can Med Assoc J 2006; 174: Garcia-Tsao G. Current management of the complications of cirrhosis and portal hypertension: variceal hemorrhage, ascites, and spontaneous bacterial peritonitis. Gastroenterology 2001; 120: Arroyo V, Rodes J. A rational approach to the treatment of ascites. Postgrad Med J 1975; 51: Khungar V, Saab S. Cirrhosis with refractory ascites: serial large volume paracentesis, TIPS, or transplantation? Clin Gastroenterol Hepatol 2011; 9: Owen AR, Stanley AJ, Vijayananthan A, Moss JG. The transjugular intrahepatic portosystemic shunt (TIPS). Clin Radiol 2009; 64: Arroyo V, Ginès A, Saló J. A European survey on the treatment of ascites in cirrhosis. J Hepatol 1994; 21: Senousy B, Draganov PV. Evaluation and management of patients with refractory ascites. World J Gastroenterol 2009; 15: Rössle M, Gerbes AL. TIPS for the treatment of refractory ascites, hepatorenal syndrome and hepatic hydrothorax: a critical update. Gut 2010; 59: Russo MW, Sood A, Jacobson IM, Brown RS Jr. Transjugular intrahepatic portosystemic shunt for refractory ascites: an analysis of the literature on efficacy, morbidity, and mortality. Am J Gastroenterol 2003; 98: Saab S, Nieto JM, Lewis SK, Runyon BA. TIPS versus paracentesis for cirrhotic patients with refractory ascites. Cochrane Database Syst Rev 2006; 18:CD AJR:204, March 2015

8 TIPS for Refractory Ascites 15. Ginès P, Uriz J, Calahorra B, et al. Transjugular intrahepatic portosystemic shunting versus paracentesis plus albumin for refractory ascites in cirrhosis. Gastroenterology 2002; 123: Boyer TD. Transjugular intrahepatic portosystemic shunt in the management of complications of portal hypertension. Curr Gastroenterol Rep 2008; 10: Busk TM, Bendtsen F, Møller S. Cardiac and renal effects of a transjugular intrahepatic portosystemic shunt in cirrhosis. Eur J Gastroenterol Hepatol 2013; 25: Garcia-Tsao G. Transjugular intrahepatic portosystemic shunt in the management of refractory ascites. Semin Intervent Radiol 2005; 22: LaBerge JM. Transjugular intrahepatic portosystemic shunt: role in treating intractable variceal bleeding, ascites, and hepatic hydrothorax. Clin Liver Dis 2006; 10: Narahara Y, Kanazawa H, Fukuda T, et al. Transjugular intrahepatic portosystemic shunt versus paracentesis plus albumin in patients with refractory ascites who have good hepatic and renal function: a prospective randomized trial. J Gastroenterol 2011; 46: Salerno F, Cammà C, Enea M, Rössle M, Wong F. Transjugular intrahepatic portosystemic shunt for refractory ascites: a meta-analysis of individual patient data. Gastroenterology 2007; 133: Albillos A, Bañares R, González M, Catalina MV, Molinero LM. A meta-analysis of transjugular intrahepatic portosystemic shunt versus paracentesis for refractory ascites. J Hepatol 2005; 43: D Amico G, Luca A, Morabito A, Miraglia R, D Amico M. Uncovered transjugular intrahepatic portosystemic shunt for refractory ascites: a metaanalysis. Gastroenterology 2005; 129: Thalheimer U, Leandro G, Samonakis DN, et al. TIPS for refractory ascites: a single-centre experience. J Gastroenterol 2009; 44: Harrod-Kim P, Saad W, Waldman D. Predictors of early mortality after transjugular intrahepatic portosystemic shunt creation for the treatment of refractory ascites. J Vasc Interv Radiol 2006; 17: Sanyal AJ. Pros and cons of TIPS for refractory ascites. J Hepatol 2005; 43: Bureau C, Metivier S, D Amico M, et al. Serum bilirubin and platelet count: a simple predictive model for survival in patients with refractory ascites treated by TIPS. J Hepatol 2011; 54: Wu X, Ding W, Cao J, Fan X, Li J. Clinical outcome using the Fluency stent graft for transjugular intrahepatic portosystemic shunt in patients with portal hypertension. Am Surg 2013; 79: Dhanasekaran R, Gonzales P, West JK, et al. Abstract no. 82: transjugular intrahepatic portosystemic shunt (TIPS) for refractory ascites. J Vasc Interv Radiol 2010; 21:S Guevara M, Baccaro ME, Ríos J, et al. Risk factors for hepatic encephalopathy in patients with cirrhosis and refractory ascites: relevance of serum sodium concentration. Liver Int 2010; 30: Cejna M. Should stent-grafts replace bare stents for primary transjugular intrahepatic portosystemic shunts? Semin Intervent Radiol 2005; 22: Tripathi D, Redhead D. Transjugular intrahepatic portosystemic stent-shunt: technical factors and new developments. Eur J Gastroenterol Hepatol 2006; 18: Bhogal HK, Sanyal AJ. Using transjugular intrahepatic portosystemic shunts for complications of cirrhosis. Clin Gastroenterol Hepatol 2011; 9: Boyer TD, Haskal Z; American Association for the Study of Liver Diseases.. The role of transjugular intrahepatic portosystemic shunt (TIPS) in the management of portal hypertension: update Hepatology 2010; 51: Bureau C, Pagan JCG, Layrargues GP, et al. Patency of stents covered with polytetrafluoroethylene in patients treated by transjugular intrahepatic portosystemic shunts: long-term results of a randomized multicentre study. Liver Int 2007; 27: Clark W, Golkar F, Luberice K, et al. Uncovering the truth about covered stents: is there a difference between covered versus uncovered stents with transjugular intrahepatic portosystemic shunts? Am J Surg 2011; 202: Ockenga J, Ockenga T, Kroencke T, et al. 222 cost analysis of E-PTFE covered tips stents compared to bare stents. J Hepatol 2004; 40(suppl 1): Hoeper MM, Krowka MJ, Strassburg CP. Portopulmonary hypertension and hepatopulmonary syndrome. Lancet 2004; 363: Bureau C, Garcia-Pagan JC, Otal P, et al. Improved clinical outcome using polytetrafluoroethylene-coated stents for TIPS: results of a randomized study. Gastroenterology 2004; 126: Membreno F, Baez AL, Pandula R, Walser E, Lau DTY. Differences in long-term survival after transjugular intrahepatic portosystemic shunt for refractory ascites and variceal bleed. J Gastroenterol Hepatol 2005; 20: Nazarian GK, Ferral H, Bjarnason H, et al. Effect of transjugular intrahepatic portosystemic shunt on quality of life. AJR 1996; 167: Qi X, Yang M, Fan D, Han G. Transjugular intrahepatic portosystemic shunt in the treatment of Budd-Chiari syndrome: a critical review of literatures. Scand J Gastroenterol 2013; 48: FOR YOUR INFORMATION This article is available for CME and Self-Assessment (SA-CME) credit that satisfies Part II requirements for maintenance of certification (MOC). To access the examination for this article, follow the prompts associated with the online version of the article. AJR:204, March