Background & Aims: Budd-Chiari syndrome (BCS) is a rare and life-threatening disorder secondary to hepatic venous outflow obstruction. Small series of BCS patients indicate that transjugular intrahepatic portosystemic shunt (TIPS) may be useful. However, the influence of TIPS on patient survival and factors that predict the outcome of TIPS in BCS patients remain unknown. Methods: One hundred twenty-four consecutive BCS patients treated with TIPS in 6 European centers between July 1993 and March 2006 were followed until death, orthotopic liver transplantation (OLT), or last clinical evaluation. Results: Prior to treatment with TIPS, BCS patients had a high Model of End Stage Liver Disease and high Rotterdam BCS prognostic index (98% of patients at intermediate or high risk) indicating severity of liver dysfunction. However, 1- and 5-year OLT-free survival were 88% and 78%, respectively. In the high-risk patients, 5-year OLT-free survival was much better than that estimated by the Rotterdam BCS index (71% vs 42%, respectively). In the whole population, bilirubin, age, and international normalized ratio for prothrombin time independently predicted 1-year OLT-free survival. A prognostic score with a good discriminative capacity (area under the curve, 0.86) was developed from these variables. Seven out of 8 patients with a score >7 died or underwent transplantation vs 5 out of 114 patients with a score <7. Conclusions: Long-term outcome for patients with severe BCS treated with TIPS is excellent even in high-risk patients, suggesting that TIPS may improve survival. Furthermore, we identified a small subgroup of BCS patients with poor prognosis despite TIPS who might benefit from early OLT. Background & Aims: Budd-Chiari syndrome (BCS) is a rare and life-threatening disorder secondary to hepatic venous outflow obstruction. Small series of BCS patients indicate that transjugular intrahepatic portosystemic shunt (TIPS) may be useful. However, the influence of TIPS on patient survival and factors that predict the outcome of TIPS in BCS patients remain unknown. Methods: One hundred twenty-four consecutive BCS patients treated with TIPS in 6 European centers between July 1993 and March 2006 were followed until death, orthotopic liver transplantation (OLT), or last clinical evaluation. Results: Prior to treatment with TIPS, BCS patients had a high Model of End Stage Liver Disease and high Rotterdam BCS prognostic index (98% of patients at intermediate or high risk) indicating severity of liver dysfunction. However, 1- and 5-year OLT-free survival were 88% and 78%, respectively. In the high-risk patients, 5-year OLT-free survival was much better than that estimated by the Rotterdam BCS index (71% vs 42%, respectively). In the whole population, bilirubin, age, and international normalized ratio for prothrombin time independently predicted 1-year OLT-free survival. A prognostic score with a good discriminative capacity (area under the curve, 0.86) was developed from these variables. Seven out of 8 patients with a score >7 died or underwent transplantation vs 5 out of 114 patients with a score <7. Conclusions: Long-term outcome for patients with severe BCS treated with TIPS is excellent even in high-risk patients, suggesting that TIPS may improve survival. Furthermore, we identified a small subgroup of BCS patients with poor prognosis despite TIPS who might benefit from early OLT. Budd-Chiari Syndrome (BCS) is a rare disorder consisting of hepatic venous outflow obstruction at any level between the small hepatic veins and the right atrium.1Janssen H.L. Garcia-Pagan J.C. Elias E. et al.Budd-Chiari syndrome: a review by an expert panel.J Hepatol. 2003; 38: 364-371Abstract Full Text Full Text PDF PubMed Scopus (384) Google Scholar Clinical presentation ranges from asymptomatic to fulminant liver failure,2Hadengue A. Poliquin M. Vilgrain V. et al.The changing scene of hepatic vein thrombosis: recognition of asymptomatic cases.Gastroenterology. 1994; 106: 1042-1047PubMed Google Scholar most commonly as subacute liver disease, with manifestations of portal hypertension, such as ascites or upper gastrointestinal bleeding.3Menon K.V. Shah V. Kamath P.S. The Budd-Chiari syndrome.N Engl J Med. 2004; 350: 578-585Crossref PubMed Scopus (408) Google Scholar Conventional management of BCS patients is anticoagulation, medical treatment of the complications of portal hypertension, and treatment of any underlying (hematologic) disease.4Valla D.C. The diagnosis and management of the Budd-Chiari syndrome: consensus and controversies.Hepatology. 2003; 38: 793-803Crossref PubMed Scopus (0) Google Scholar, 5Zimmerman M.A. Cameron A.M. Ghobrial R.M. Budd-Chiari syndrome.Clin Liver Dis. 2006; 10: 259-273Abstract Full Text Full Text PDF PubMed Scopus (41) Google Scholar However, this approach is successful in only 18% of patients.6Plessier A. Sibert A. Consigny Y. et al.Aiming at minimal invasiveness as a therapeutic strategy for Budd-Chiari syndrome.Hepatology. 2006; 44: 1308-1316Crossref PubMed Scopus (204) Google Scholar Adding recanalization (thrombolysis, angioplasty, or stenting) of short-length hepatic vein stenoses, the success rate increases up to 32%.6Plessier A. Sibert A. Consigny Y. et al.Aiming at minimal invasiveness as a therapeutic strategy for Budd-Chiari syndrome.Hepatology. 2006; 44: 1308-1316Crossref PubMed Scopus (204) Google Scholar Thus, most patients need more aggressive treatments. Transjugular intrahepatic portosystemic shunt (TIPS) is a percutaneous interventional radiologic procedure that creates a portocaval shunt through the liver parenchyma.7Rosch J. Hanafee W. Snow H. et al.Transjugular intrahepatic portacaval shunt An experimental work.Am J Surg. 1971; 121: 588-592Abstract Full Text PDF PubMed Scopus (116) Google Scholar, 8Boyer T.D. Haskal Z.J. The role of transjugular intrahepatic portosystemic shunt in the management of portal hypertension.Hepatology. 2005; 41: 386-400Crossref PubMed Scopus (325) Google Scholar Although more challenging than in patients with cirrhosis, in experienced hands, TIPS is technically feasible in most patients with BCS.9Blum U. Rossle M. Haag K. et al.Budd-Chiari syndrome: technical, hemodynamic, and clinical results of treatment with transjugular intrahepatic portosystemic shunt.Radiology. 1995; 197: 805-811Crossref PubMed Scopus (141) Google Scholar, 10Eapen C.E. Velissaris D. Heydtmann M. et al.Favourable medium term outcome following hepatic vein recanalisation and/or transjugular intrahepatic portosystemic shunt for Budd-Chiari syndrome.Gut. 2006; 55: 878-884Crossref PubMed Scopus (134) Google Scholar, 11Perello A. Garcia-Pagan J.C. Gilabert R. et al.TIPS is a useful long-term derivative therapy for patients with Budd-Chiari syndrome uncontrolled by medical therapy.Hepatology. 2002; 35: 132-139Crossref PubMed Scopus (166) Google Scholar In patients without a remaining hepatic vein stump, a direct puncture from the intrahepatic inferior vena cava (IVC) into the liver parenchyma is an option. In small series of patients with BCS, TIPS has been shown to be an effective treatment.10Eapen C.E. Velissaris D. Heydtmann M. et al.Favourable medium term outcome following hepatic vein recanalisation and/or transjugular intrahepatic portosystemic shunt for Budd-Chiari syndrome.Gut. 2006; 55: 878-884Crossref PubMed Scopus (134) Google Scholar, 12Rossle M. Olschewski M. Siegerstetter V. et al.The Budd-Chiari syndrome: outcome after treatment with the transjugular intrahepatic portosystemic shunt.Surgery. 2004; 135: 394-403Abstract Full Text Full Text PDF PubMed Scopus (121) Google Scholar, 13Mancuso A. Fung K. Mela M. et al.TIPS for acute and chronic Budd-Chiari syndrome: a single-centre experience.J Hepatol. 2003; 38: 751-754Abstract Full Text Full Text PDF PubMed Google Scholar However, the impact of TIPS on survival is unknown. Some BCS patients do not improve after TIPS with clinical deterioration because of liver failure and early death or require liver transplantation.14Ryu R.K. Durham J.D. Krysl J. et al.Role of TIPS as a bridge to hepatic transplantation in Budd-Chiari syndrome.J Vasc Interv Radiol. 1999; 10: 799-805Abstract Full Text PDF PubMed Scopus (71) Google Scholar It is conceivable that the management of patients with BCS would be improved if therapeutic alternatives such as liver transplantation were offered to patients who do not benefit from TIPS. The aim of the present study was to evaluate the medium- and long-term outcome in the, at present, largest series of BCS patients treated with TIPS and to determine prognostic factors at 1 year. Between July 1993 and March 2006, 221 patients were diagnosed with BCS in the liver units of the 6 participating European tertiary referral centers. All centers have highly experienced clinicians performing TIPS. BCS was defined following the criteria established by the European network for vascular disorders of the liver (En-Vie)1Janssen H.L. Garcia-Pagan J.C. Elias E. et al.Budd-Chiari syndrome: a review by an expert panel.J Hepatol. 2003; 38: 364-371Abstract Full Text Full Text PDF PubMed Scopus (384) Google Scholar and ratified at the last Baveno consensus15de 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: 167-176Abstract Full Text Full Text PDF PubMed Scopus (935) Google Scholar as “hepatic venous outflow obstruction at any level between the small hepatic veins and the junction between the IVC and the right atrium.” Diagnosis of BCS was made by demonstrating an unequivocal obstruction by either Doppler ultrasonography, magnetic resonance imaging, or computerized tomography.1Janssen H.L. Garcia-Pagan J.C. Elias E. et al.Budd-Chiari syndrome: a review by an expert panel.J Hepatol. 2003; 38: 364-371Abstract Full Text Full Text PDF PubMed Scopus (384) Google Scholar Patients younger than 15 years or with tumoral BCS were excluded. All patients were initially treated with anticoagulation and received treatment for any other secondary disorders then percutaneous angioplasty, then TIPS, and, finally, liver transplantation depending on the response to the previous treatment option. One hundred forty-seven out of 221 BCS patients (66.5%) not responding to medical treatment or recanalization were eligible for TIPS. Treatment and outcome of BCS patients not receiving TIPS are reported in Figure 1. Clinical, biochemical, and radiologic information were extracted retrospectively from patient's notes into a predesigned case record form including clinical information, laboratory, and radiologic data collected at the time of BCS diagnosis, at the time of TIPS, and at the end of follow-up, which was either death, liver transplantation, or study closure (October 31, 2006). In addition, Child–Pugh score,16Pugh R.N. Murray-Lyon I.M. Dawson J.L. et al.Transection of the oesophagus for bleeding oesophageal varices.Br J Surg. 1973; 60: 646-649Crossref PubMed Scopus (6601) Google Scholar Model of End Stage Liver Disease (MELD) score,17Malinchoc M. Kamath P.S. Gordon F.D. et al.A model to predict poor survival in patients undergoing transjugular intrahepatic portosystemic shunts.Hepatology. 2000; 31: 864-871Crossref PubMed Scopus (2032) Google Scholar and specific prognostic indices developed for BCS, such as the original and revised Clichy scores18Zeitoun G. Escolano S. Hadengue A. et al.Outcome of Budd-Chiari syndrome: a multivariate analysis of factors related to survival including surgical portosystemic shunting.Hepatology. 1999; 30: 84-89Crossref PubMed Scopus (199) Google Scholar, 19Langlet P. Escolano S. Valla D. et al.Clinicopathological forms and prognostic index in Budd-Chiari syndrome.J Hepatol. 2003; 39: 496-501Abstract Full Text Full Text PDF PubMed Scopus (129) Google Scholar and Rotterdam score,20Murad S.D. Valla D.C. de Groen P.C. et al.Determinants of survival and the effect of portosystemic shunting in patients with Budd-Chiari syndrome.Hepatology. 2004; 39: 500-508Crossref PubMed Scopus (264) Google Scholar were calculated using the data obtained at the time of TIPS. Acute liver failure was defined according to the American Association for the Study of Liver Diseases 2005 guideline21Polson J. Lee W.M. AASLD position paper: the management of acute liver failure.Hepatology. 2005; 41: 1179-1197Crossref PubMed Scopus (727) Google Scholar: “evidence of coagulation abnormality (international normalized ratio for prothrombin time [INR] > 1.5) and hepatic encephalopathy (HE) in a patient with a disease that has only been recognized for <26 weeks.” Twelve out of 124 (9.7%) patients were on oral anticoagulation up until TIPS. In these patients, the INR was derived as done at the initial description of the MELD score17Malinchoc M. Kamath P.S. Gordon F.D. et al.A model to predict poor survival in patients undergoing transjugular intrahepatic portosystemic shunts.Hepatology. 2000; 31: 864-871Crossref PubMed Scopus (2032) Google Scholar from bilirubin and aspartate aminotransferase (AST) values by linear regression (r2 = 0.57). During hepatic venography using an internal jugular vein approach, a needle-catheter was placed into the hepatic vein remnant, and, under ultrasonographic and/or fluoroscopic guidance, the needle was directed through the liver parenchyma toward a branch of the portal vein. When a remnant was not present, the portal vein was punctured directly from the IVC.7Rosch J. Hanafee W. Snow H. et al.Transjugular intrahepatic portacaval shunt An experimental work.Am J Surg. 1971; 121: 588-592Abstract Full Text PDF PubMed Scopus (116) Google Scholar, 22Richter G.M. Palmaz J.C. Noldge G. et al.[The transjugular intrahepatic portosystemic stent-shunt. A new nonsurgical percutaneous method].Radiologe. 1989; 29: 406-411PubMed Google Scholar After the TIPS procedure, patients underwent anticoagulation, following the guidelines for BCS.15de 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: 167-176Abstract Full Text Full Text PDF PubMed Scopus (935) Google Scholar TIPS dysfunction during follow-up was defined as any of the following: increase in portal pressure gradient above 12 mm Hg, significant narrowing of the lumen on venography or (re-) appearance of ascites or other complication of portal hypertension. Patients were followed in the outpatient clinic, initially at 1 month after TIPS, then at 3 months, and every 6 months thereafter. An abdominal ultrasound was performed every 6 months. Hemodynamic evaluation was performed when clinical or ultrasound data suggested TIPS dysfunction.23Abraldes J.G. Gilabert R. Turnes J. et al.Utility of color Doppler ultrasonography predicting tips dysfunction.Am J Gastroenterol. 2005; 100: 2696-2701Crossref PubMed Scopus (39) Google Scholar Transplantation-free survival was calculated by the Kaplan–Meier method. The potential contribution of demographic, clinical, biochemical, radiologic, and procedure-related variables to the risk of death or liver transplantation (OLT) at 1 year was analyzed by Cox regression univariable analysis. Variables achieving a P value <.1 were introduced into a multivariate forward stepwise Cox regression analysis. A prognostic model was constructed with the regression coefficients of the variables that were independently related to 1-year OLT-free survival. The discriminative ability of the model was assessed by receiver operator curve analysis. A cut-off level with high specificity was chosen to determine which patients benefit from TIPS. The low number of events and the expected colinearity between variables, inevitable in a study of a rare disease, might result in an unstable model with low reproducibility. Therefore, to test the reliability of our predictors, a bootstrapping Cox regression analysis24Altman D.G. Andersen P.K. Bootstrap investigation of the stability of a Cox regression model.Stat Med. 1989; 8: 771-783Crossref PubMed Scopus (247) Google Scholar, 25Steyerberg E.W. Harrell Jr, F.E. Borsboom G.J. et al.Internal validation of predictive models: efficiency of some procedures for logistic regression analysis.J Clin Epidemiol. 2001; 54: 774-781Abstract Full Text Full Text PDF PubMed Scopus (1713) Google Scholar was conducted. Briefly, from our series of patients, 200 different samples of patients were generated by random patient selection with replacement with each sample of the same size as the original patient number. In each of the 200 samples, a Cox regression analysis was conducted. Variables that appeared in 50% or more of the models in the bootstrapping analysis were considered reliable predictors of outcome.26Blackstone E.H. Breaking down barriers: helpful breakthrough statistical methods you need to understand better.J Thorac Cardiovasc Surg. 2001; 122: 430-439Abstract Full Text PDF PubMed Scopus (192) Google Scholar Next, after evaluating which of the currently available prognostic scores (Child–Pugh, original and revised Clichy, Rotterdam, MELD, and the prognostic index derived from this study) best predicts 1-year OLT-free survival in patients treated with TIPS for BCS, we conducted a bootstrapping analysis introducing these 6 scores as covariates. All reported P values are 2-sided. The level of statistical significance was set at P < .05. Statistical analysis was performed using SPSS 12.0 package (SPSS, Inc, Chicago, IL). The bootstrapping was performed with the macro “regression bootstrap” (provided with the SPSS package) adapted by the investigators for Cox regression analysis. Table 1 summarizes the clinical and biochemical features of patients before TIPS. As shown, patients receiving TIPS had severe liver disease reflected by a high Child–Pugh, MELD, Clichy, and Rotterdam scores. The mean follow-up after the TIPS procedure was 36.7 months (range, 0.7–156 months).Table 1Clinical Characteristics of BCS Patients Prior to TIPSAge (y), mean (95% CI)38 (35−40)Radiologic findings, n/%Sex (male/female); n (%)46 (37)/78 (63) Associated IVC thrombosis18/15Etiology, n (%) Associated portal vein thrombosis12/10 Chronic myeloproliferative disorders64 (51.6)Blood tests mean (95% CI) Antiphospholipid syndrome15 (12.1) Bun (mg/100 mL)23.5 (20−27) Paroxysmal nocturnal hemoglobinuria13 (10.5) Creatinine (mg/100 mL)1.1 (1−1.3) Prothrombotic disorders12 (9.7) Sodium (mEq/L)133 (132−135) Idiopathic17 (13.7) AST (×ULN)7.2 (1.5−14) Others causesaBehcet's disease in 1 patient, myelodisplastic syndrome in 1 patient, and thrombotic thrombocytopenic purpura in 1 patient.3 (2.4) Alkaline phosphotase (×ULN)1.4 (1.2−1.5)Clinical manifestations, n (%) Bilirubin (mg/100 mL)3.7 (2.8−4.6) Ascites122 (98) Albumin (g/L)31 (29−32) Refractory ascites80 (64.5) Hemoglobin (g/L)12.5 (11.8−12.9) Gastrointestinal bleeding18 (14.5) Platelets (×106/mm3)2.3 (2.06−2.63) Hepatic encephalopathy27 (21.8) INR1.7 (1.5−1.9) Acute liver failure9 (7.3)Clinical P. I. mean (95% CI): Spontaneous bacterial peritonitis6 (4.8) Child–Pugh score9 (8.6−9.5) Hepatorenal syndrome5 (4) MELD score17 (15.7−18.7) Hepatopulmonary syndrome1 (0.8) Original Clichy score (high risk, >5.4)6.2 (6.1−6.5)Main TIPS indications, n (%) Revised Clichy score (high risk, >5.1)6.4 (6.2−7) Refractory ascites73 (59) Rotterdam score (high risk, ≥1.5)1.63 (1.4−1.7) Liver failure27 (22) Upper gastrointestinal bleeding12 (9.5)Other portal hypertension complicationsbLow-grade ascites in 11 patients and 1 patient with nonbleeding esophageal varices with transient hepatic encephalopathy episode.12 (9.5)ULN, upper limit of normal; P. I., prognostic index.a Behcet's disease in 1 patient, myelodisplastic syndrome in 1 patient, and thrombotic thrombocytopenic purpura in 1 patient.b Low-grade ascites in 11 patients and 1 patient with nonbleeding esophageal varices with transient hepatic encephalopathy episode. Open table in a new tab ULN, upper limit of normal; P. I., prognostic index. One hundred forty-seven BCS patients who did not respond to medical treatment/recanalization were considered eligible for TIPS. However, in 14 patients, a formal contraindication for TIPS was established; therefore, TIPS was finally indicated in 133 patients. It was successfully performed in 124 patients (success rates: 84% on intention to treat, 93% technical success). The portal vein was punctured via a hepatic vein stump in 47 patients (38%); from the IVC (transcaval approach) in 70 patients (56.5%); and, in 7 patients (5.5%), this information was not recorded. The number of stents placed was 1 in 46 patients (37%), 2 in 62 patients (50%), 3 in 13 patients (10.5%), and 4 in 3 patients (2.5%). The stents used were uncovered in 61 patients (49%), polytetrafluoroethylene covered in 48 patients (39%), and both types in 15 patients (12%). Twenty-two patients (17.7%) had complications associated with TIPS. Two resulted in death. One patient died after IVC injury, intraperitoneal bleeding, and hemothorax. The other death was related to partial occlusion of the IVC by the stent, with persisting of ascites, infection, and death. Both patients died within days after the procedure. The remaining 20 patients had medically controlled complications: subcapsular hematoma (n = 5), hemoperitoneum without hemodynamic instability (n = 3), hemobilia (n = 3), reversible heart failure (n = 3), procedure-associated infection (n = 2), supraventricular tachycardia (n = 1), jugular vein thrombosis (n = 1), IVC compression (n = 1), and mild hemolysis (n = 1). The mean baseline portal pressure and portoatrial pressure gradients were 33 mm Hg (range, 19–56 mm Hg) and 27 mm Hg (range, 14–50 mm Hg), respectively, and were reduced to 21 mm Hg (range, 8–33 mm Hg) and 9 mm Hg (range, 2–17 mm Hg), respectively, after TIPS. During follow-up, 16 patients died (13%), and 8 (6.5%) required OLT. The main causes of death were liver failure (n = 4), hematologic disorders (n = 4), sepsis (n = 3), procedure-related death (n = 2), stroke (n = 2), and upper gastrointestinal bleeding of unknown origin (n = 1). The main indications for OLT were liver failure (n = 5), recurrent HE (n = 2), and recurrent TIPS dysfunction (n = 1). OLT was performed at a median of 23 months after TIPS procedure (range, 20 days to 85 months). There was no perioperative death after OLT. In 1 transplant patient, the stent removal and graft anastomosis with the recipient IVC was technically difficult, causing operative bleeding and requiring a side-to-side cavo-caval anastomosis. No patient required retransplantation, and 7 out of 8 (87.5%) are alive at the study closure date. Of the 16 patients who died, 12 had an absolute contraindication to OLT, and the remaining 4 patients died with multiorgan failure before any possible evaluation for OLT. One-, 5-, and 10-year OLT-free survival rates were 88% (95% confidence interval [CI]: 81%–93%), 78% (95% CI: 68%–85%), and 69% (95% CI: 52%–80%), respectively (Figure 2). With 1 death among the transplantation patients, 1-, 5-, and 10-year survival rates were 90% (95% CI: 83%–94%), 84% (95% CI: 75%–90%), and 80% (95% CI: 66%–88%), respectively. Classification of our cohort of BCS patients according to the risk score for BCS described by the Rotterdam group20Murad S.D. Valla D.C. de Groen P.C. et al.Determinants of survival and the effect of portosystemic shunting in patients with Budd-Chiari syndrome.Hepatology. 2004; 39: 500-508Crossref PubMed Scopus (264) Google Scholar attributed 3 patients (2.4%) at low risk, 81 patients (65.3%) at intermediate risk, and 40 patients (32.2%) at high risk for death or OLT. In both the intermediate and high-risk groups, the survival of our patients cohort was better than predicted. This was especially evident in the high-risk group (Table 2).Table 2Five-Year OLT-Free Survival of BCS Patients According to Rotterdam Score Risk GroupsRotterdam cohort (n = 205)aIn the Rotterdam cohort, only 15% of patients received TIPS.20Present cohort (n = 124)Intermediate-risk group74% (95 % CI: 65%−83%)82% (95% CI: 69%−90%)High-risk group42% (95% CI: 28%−56%)71% (95% CI: 53%−84%)a In the Rotterdam cohort, only 15% of patients received TIPS.20Murad S.D. Valla D.C. de Groen P.C. et al.Determinants of survival and the effect of portosystemic shunting in patients with Budd-Chiari syndrome.Hepatology. 2004; 39: 500-508Crossref PubMed Scopus (264) Google Scholar Open table in a new tab Sixty-one of the 124 patients (41%) had TIPS dysfunction during follow-up. Thirty-two were detected during routine shunt assessment and 29 after developing recurrent ascites (n = 24), upper gastrointestinal bleeding (n = 3), abdominal pain (n = 1), and hydrothorax (n = 1). TIPS dysfunction required restenting in 35 patients, angioplasty in 20, and thrombolysis in 6. Actuarial probability of TIPS dysfunction was significantly lower in patients treated with PTFE-covered stents than with bare stents (P = .001) as shown in Figure 3. Those patients who were treated with bare and PTFE-covered stents (n = 15) were included in the covered group for the analysis. The actuarial probability of developing HE at 1-year post-TIPS was 21% (95% CI: 14%–29%). The median time between TIPS insertion and the first episode of HE was 1.2 months (range, 1 day to 32.5 months). In 22 patients, HE was transient and easily controlled with no reappearance after a median follow-up of 8 months. Four patients had recurrent HE, and all 4 were listed for OLT, with 1 still on the waiting list at the end of follow-up. The 2 patients who died within days because of direct complications of the TIPS procedure were excluded from the prognostic analysis. Variables associated with 1-year OLT-free survival after TIPS achieving a P value <.1 at univariate analysis (Table 3) were included into a multivariate Cox regression analysis that disclosed bilirubin, age, and INR as independent predictors of 1-year OLT-free survival after TIPS (Table 4).Table 3Univariate Analysis of Pre-TIPS Variables Predicting OLT-Free Survival at 1 YearVariableHR95% CIP valueVariableHR95% CIP valueMaterial centerPortal vein thrombosis0.040−184.16 Clinic10.50 Birmingham0.730.18−2.92 Clichy0.290.03−2.57 Rotterdam0.670.07−6 Milan1.530.28−8.4 Palermo0.00.0−∞Sex (male)1.80.55−5.3.31Number of hepatic veinsaThree hepatic veins vs 1 or 2 hepatic veins affected.0.350.1−1.88AgebVariables achieving a P value < .1 that were introduced into a multivariate Cox regression analysis.1.081.03−1.12< .001Stent typecPTFE-covered stent vs bare stent.1.940.6−6.9.26Etiology1.160.72−1.81.56CreatininebVariables achieving a P value < .1 that were introduced into a multivariate Cox regression analysis.2.71.41−5.17.00Time between diagnosis and TIPS10.99−1.88SodiumbVariables achieving a P value < .1 that were introduced into a multivariate Cox regression analysis.0.910.83−1.04Period of TIPS insertiondBefore vs after 2001.0.850.23−3.14.80AST0.970.85−1.1.40Ascites.75BilirubinbVariables achieving a P value < .1 that were introduced into a multivariate Cox regression analysis.1.221.13−1.3< .001Refractory ascites0.390.12−1.18.10Albumin0.970.88−1.07.52Gastrointestinal bleeding20.58−7.95.33Platelets11−1.16Hepatic encephalopathybVariables achieving a P value < .1 that were introduced into a multivariate Cox regression analysis.3.741.12−12.5.02INRbVariables achieving a P value < .1 that were introduced into a multivariate Cox regression analysis.1.611.14−2.3.00Acute liver failurebVariables achieving a P value < .1 that were introduced into a multivariate Cox regression analysis.10.73.17−36.15< .001Spontaneous bacterial peritonitisbVariables achieving a P value < .1 that were introduced into a multivariate Cox regression analysis.4.160.9−18.7.04Surgical shunting pre-TIPS0.050−8054.35Anticoagulation pre-TIPS0.440.14−1.4.16Clinicopathologic form (type III)eAcute on chronic clinical presentation.192.010.68−6.11.24HR, hazard ratio.a Three hepatic veins vs 1 or 2 hepatic veins affected.b Variables achieving a P value < .1 that were introduced into a multivariate Cox regression analysis.c PTFE-covered stent vs bare stent.d Before vs after 2001.e Acute on chronic clinical presentation.19Langlet P. Escolano S. Valla D. et al.Clinicopathological forms and prognostic index in Budd-Chiari syndrome.J Hepatol. 2003; 39: 496-501Abstract Full Text Full Text PDF PubMed Scopus (129) Google Scholar Open table in a new tab Table 4Results of Multivariate Cox Regression Analysis for the 122 BCS PatientsVariableHR95% CIP valueBootstrapping Cox regression analysis in 200 samplesBilirubin1.1711.08–1.27< .00192%Age1.0861.03–1.14.00264%INR1.8761.13–3.1.01450%NOTE. The BCS-TIPS prognostic index was derived from these variables and their natural logarithm of the Cox regression coefficients: BCS-TIPS PI = age (years) × 0.08 + bilirubin (mg/dL) × 0.16 + INR × 0.63. Open table in a new tab HR, hazard ratio. NOTE. The BCS-TIPS prognostic index was derived from these variables and their natural logarithm of the Cox regression coefficients: BCS-TIPS PI = age (years) × 0.08 + bilirubin (mg/dL) × 0.16 + INR × 0.63. In the bootstrapping Cox analysis, these 3 variables (bilirubin, age, and INR) were also identified as independent predictors of 1-year OLT-free survival after TIPS. Of the 200 bootstrap samples, bilirubin was selected in 92%, age in 64%, and INR in 50%. From these variables and their Cox regression coefficients, a BCS-TIPS prognostic index (BCS-TIPS PI) was derived: BCS-TIPS PI = age (years) × 0.08 + bilirubin (mg/dL) × 0.16 + INR × 0.63. This model shows a good prognostic ability with an area under the curve of 0.86 (95% CI: 0.72–0.99) (Figure 4). Bootstrapping C