Polytetrafluoroethylene-Covered Stent Graft Versus Bare Stent in Transjugular Intrahepatic Portosystemic Shunt: Systematic Review and Meta-Analysis

Abstract

Background & Aims:

Transjugular Intrahepatic Portosystemic Shunt (TIPS) is used to control refractory variceal bleeding secondary to portal hypertension. This meta-analysis was conducted to systematically review polytetrafluoroethylene-covered stent grafts (CS) versus bare stents (BS) in TIPS procedure.

Methods:

Systematic search of literature databases was done from January-1990 till April-2017, using predecided keywords. Outcome measures studied were (1) primary-patency (PP) at 1 year (defined as absence of shunt insufficiency at 1 year), (2) rebleeding (RE) (3) new-onset hepatic encephalopathy ([HE] new-onset or worsening encephalopathy following the procedure), and (4) survival at 1 year (SU). Odds ratio (OR) was calculated for each outcome variable. Between-study heterogeneity was assessed by the I² statistics and χ² Q-test.

Results:

Fourteen studies (4 RCTs, 2 prospective nonrandomized, and 8 retrospective) were included with 2519 patients (1548 patients in BS group and 971 patients in CS group). Three-quarter outcome measures showed significantly better results with CS. PP was pooled from 13 studies and showed an OR = 4.75 (95% confidence interval [CI] = 3.32–6.79; P < .00001; I² = 44%) in favor of CS. RE was pooled from six studies with odds ratio (OR) = 0.37(95% CI = 0.24–0.56; P < .00001; I² = 0%) in favor of CS. SU was pooled from 11 studies with OR = 1.85 (95% CI = 1.44–2.38; P < .00001; I² = 0%) in favor of CS. On subset analysis for RCTs, three outcome variables favored CS with minimal heterogeneity [PP: OR = 4.18 (95% CI = 2.66–6.55; P < .00001; I² = 0%). RE: OR = 0.43 (95% CI = 0.25–0.72; P < .001; I² = 0%). SU: OR = 1.85 (95% CI = 1.44–2.38, P < .00001; I² = 0%)]. HE showed no difference between two stents on the overall and subset pooled analysis. Overall: OR = 0.86 (95% CI = 0.68–1.08; P = .19; I² = 4%). Only RCTs: OR = 0.91 (95% CI = 0.63–1.32, P = .63; I² = 0%).

Conclusions:

CS is associated with better primary patency and survival and lesser rate of rebleeding than BS in patients undergoing TIPS procedure. There is no difference in new-onset hepatic encephalopathy.

Introduction

Acute variceal hemorrhage is one of the most common causes of upper gastrointestinal hemorrhage and associated mortality in patients with cirrhosis.¹ Endoscopy is both diagnostic and therapeutic. Success rate following endoscopic intervention approaches 90% and with acceptable rate of complications.^2,3 In a subset of patients with recurrent bleeding following two attempts at endoscopic interventions, transjugular intrahepatic portosystemic shunt (TIPS) is considered the procedure of choice.

TIPS is a percutaneously placed nonselective intrahepatic hepatoportal shunt to decompress high-pressure portal venous flow into the low-pressure hepatic vein and thus into inferior vena cava. The patency of this low-resistance shunt is maintained by using expandable and nonexpandable metallic stents. Ever since its conception, the TIPS procedure has gained wide acceptance and has reduced morbidity and mortality in patients with portal hypertension.⁴

TIPS has been well established as an emergency procedure for acute variceal hemorrhage as well as for refractory ascites, Budd–Chiari syndrome, and refractory hepatic hydrothorax. Major or minor complications are reported in up to 4% of the patients, common complications being hepatic encephalopathy (HE), stent dysfunction, variceal hemorrhage, and a relatively higher rate of mortality than selective shunts.⁵ Covered stent grafts were introduced in an effort to curtail the common side effects of shunt dysfunction and improve survival. Polytetrafluoroethylene (PTFE)-covered stent grafts are the most popular and widely used covered grafts. Several randomized and nonrandomized studies have compared clinical outcomes, safety, and efficacy between covered and bare stents, although producing inadequately powered studies to substantiate a clear supremacy of one type of stent over the other. The significant difference of cost between the two types of stents is an important consideration, especially if two stents were to have comparable results. The aim of this meta-analysis was to establish weighted-pooled analysis of the available literature, and to determine the superiority of one stent over another over four domains as the surrogate markers of efficacy of the procedure.

Materials and Methods

PRISMA criteria

This systematic review and meta-analysis were planned and undertaken in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analysis) guidelines⁶ and Cochrane handbook for systematic reviews and interventions.⁷ The checklist items from PRISMA as relevant to this review, such as those related to search and writing approaches, and conduct of meta-analysis were included. It is to be noted that no ethics approval was necessary for this type of study.

Study search

Three of the authors (S.S., P.A., and E.G.) independently conducted systematic literature search through the major indexed literature databases, including Embase, Medline, Pubmed, Cochrane, and Google Scholar, from January 1990 to March 2017, to identify all the publications using the following keywords: “TIPS”, “transjugular intrahepatic portosystemic shunt”, “PTFE stent”, “polytetrafluoroethylene”, “stent graft,” and “bare stent” as search items. The keywords were used in all possible combinations to avoid missing any article.

Study selection

Inclusion criteria were as follows: studies in the English language, comparing bare stents (uncovered stents) and PTFE-covered stent grafts for TIPS procedure done for any indications related to portal hypertension, full-length articles, and reporting at least one of the four outcome variables included in the meta-analysis. In view of the paucity of an adequate number of prospective randomized controlled trials (RCTs), all types of studies (including nonrandomized prospective comparisons as well as retrospective case–control studies) were included. The searches were limited to studies conducted in humans. All experimental studies on animal models, case reports, case series with a total sample size of 20 or less, technical reports, abstracts, editorials, review articles, and previous articles with duplicate patient populations were excluded. The references of the included studies were further reviewed to identify any additional studies on the topic. The reporting quality of the study was critically assessed according to the PRISMA checklist.⁶

Data abstraction

The formal full-length publications of all included studies were first reviewed independently by three authors (S.S., P.A., and E.G.) according to a predecided standardized approach. The extracted data were charted in an Excel spreadsheet (Microsoft Corp., Redmond, WA) and the discrepancies were mutually resolved with consensus after discussion with the other authors. The following parameters for each included study were noted: first author, year of publication, journal name, country of origin (in case of multiple countries, country of the corresponding author), type of study design, sample size in each limb, preprocedure clinical status of included subjects, type (and brand, if available) and diameter of stents used, and outcome reports. Attempts were made to contact the corresponding authors of the respective studies for any controversy, missing data points, or data for information only available as graphs in the original articles. If the data from authors could not be retrieved even after multiple attempts at contact, the data available in the published articles and graphs/charts were used. We referenced our approach to this meta-analysis from the meta-analysis published by Yang et al.⁸

Quality review of included studies

The recommendations in the Quality of Reporting of Meta-Analyses (QUOROM) statement⁹ were used for quality assessment and associated publication bias of the included studies. The studies with two patient cohorts from the same time period, similar baseline parameters, prospective studies, randomization, detailed and high level of data reporting, and less than 10% patients lost to follow-up at 1 year were considered high quality. In addition, the methodological index for nonrandomized studies (MINORS) tool¹⁰ was used to assess individual nonrandomized studies for the quality of evidence. A summary of the quality evaluation and risk of bias for studies is shown in Table 1.

Table 1.

Quality Assessment of Nonrandomized Studies Using MINORS Criteria

	Included nonrandomized studies
	Angermayr B, 2003	Angeloni S, 2004	Hernández- Guerra M, 2004	Tripathi D, 2014	Jung HS, 2009	Wu X, 2010	Clark W, 2011	Sommer, 2012	Tan HK, 2015	Lauermann J, 2016
Methodological items
1. A clearly stated aim	2	2	2	2	2	2	2	2	2	2
2. Inclusion of consecutive patients	1	2	1	1	1	2	1	1	2	2
3. Prospective collection of data	1	2	2	2	1	1	2	1	1	1
4. Endpoints appropriate to the aim of the study	2	2	2	2	2	2	1	2	2	2
5. Unbiased assessment of the study endpoint	1	2	1	2	2	1	2	1	2	2
6. Follow-up period appropriate to the study aim	2	2	2	2	2	1	2	2	1	0
7. Loss to follow-up less than 5%	0	2	2	0	0	0	0	0	0	1
8. Prospective calculation of the study size	0	0	0	0	0	0	0	0	0	0
Additional criteria
9. An adequate control group	2	2	1	2	2	2	2	2	2	2
10. Contemporary groups	2	1	0	1	2	2	2	2	1	2
11. Baseline equivalence of groups	1	2	2	2	2	1	1	2	2	0
12. Adequate statistical analyses	2	2	2	2	2	2	1	2	2	2
Total score	16	21	17	18	18	16	16	17	17	16

MINORS, methodological index for nonrandomized studies.

Outcome measures

The primary summary outcome measures were primary shunt patency at 1 year and survival at 1 year. Secondary outcome measures were rebleeding after the TIPS procedure and new-onset HE.

The outcome measures were defined as follows:

(1) Primary patency at 1 year: absence of shunt insufficiency at 1-year follow-up following the procedure.

(2) Rebleeding: hemorrhage following the procedure that was associated with the disease process or attributable to the failure of the procedure.

(3) New-onset HE: appearance of HE in patients who did not have HE before the procedure or reported increase in severity of encephalopathy following TIPS.

(4) Survival at 1 year: subset of patients alive at 1-year follow-up following the procedure.

All recorded data were in the form of dichotomous variables with specific point data and no time-to-event data.

Statistical analyses

All analyses were performed using Review Manager (RevMan) v5.3 (The Cochrane Collaboration, Oxford, United Kingdom).¹¹ Effect sizes for variables were calculated and expressed as odds ratio (OR) with 95% confidence interval (CI). Between-study heterogeneity was assessed by the I² statistics and χ² Q-test. A heterogeneity of more than 40% or P < .10 was considered substantial and important to the interpretation of the outcome. The random-effect model was used in presence of heterogeneity (DerSimonian and Laird),¹² while a fixed-effect model was used in the absence of significant heterogeneity (Mantel–Haenszel method).¹³ In the presence of significant heterogeneity across studies, each study was sequentially omitted during the sensitivity analysis to look at the influence of individual studies on pooled analysis. In the scenario where data were not clearly published or were not available from authors on request, the published Kaplan–Meier curve was used to collect data, as described and used by Yang et al.⁸ The overlapping data from sequential publications by the same group were confirmed with authors and thus rectified. The presence of publication bias was assessed and quantified by the Egger's test.

Results

Search results

The systematic database search and stepwise exclusion of studies is shown in Figure 1. A search through the database and review of references of previously published articles returned 126 results related to the study topic. The abstracts of all studies were systematically screened. After removing 38 duplicate results, 88 studies were screened for inclusion according to the eligibility criteria; 17 studies were identified as eligible for review. Of these, three studies were excluded due to insufficient published data, and further, the data could not be retrieved from the investigators on multiple attempts. Fourteen studies were included in the cohort of this review.

FIG. 1.

The flowchart reflecting the stepwise exclusion and selection of studies for the meta-analysis.

The 14 included studies were published between 2003 and 2016, on different cohorts of patients from various countries.^14–27 These included eight retrospective studies, two prospective nonrandomized comparative studies, and four RCTs. Total of 2519 patients were pooled, including 1548 patients in bare stent group and 971 patients in PTFE stent graft group. The study designs and patient characteristics are summarized in Tables 2 and 3.

Table 2.

Characteristics of Included Studies

Country of origin First Author, Year,	Type of study	Study population	Inclusion and exclusion criteria	Type of stent used (Bare/PTFE)	Diameter of the shunt, mm (Bare/PTFE)	Primary/Secondary Outcome measures reported ^a	TIPS patency surveillance	Mean follow-up (months)
Randomized studies
Bureau et al. 2007,¹⁹ France, Spain, Canada	Randomized multicenter	Cirrhotics	Inclusions: Uncontrolled or recurrent variceal bleeding after failure of medical or endoscopic treatment, refractory ascites, age: 18–75 years	Memotherm, Wallstent, Luminex, or Sinus/Viatorr	11.7/10.5^b	1/2–4	Clinical examination, USG, angiography	14.3 ± 12.2/19.5 ± 14.6^b
Bureau et al. 2007,¹⁹ France, Spain, Canada	Randomized multicenter	Cirrhotics	Exclusions: chronic or recurrent encephalopathy, portal vein thrombosis, HCC, cardiac failure, hepatic polycystosis, intrahepatic bile duct dilatation, patients unable to give consent form, pregnant or breastfeeding women	Memotherm, Wallstent, Luminex, or Sinus/Viatorr	11.7/10.5^b	1/2–4	Clinical examination, USG, angiography	14.3 ± 12.2/19.5 ± 14.6^b
Huang et al. 2010,²⁷ China	Randomized single center	Cirrhotics	Inclusions: Biopsy-proven cirrhosis uncontrolled variceal bleeding; refractory ascites, recurrent refractory variceal bleeding; age: 18–75 years.	Wallstent/Fluency	10/8	1–4	Shunt venography	8.3 ± 4.4/6.2 ± 3.9
Huang et al. 2010,²⁷ China	Randomized single center	Cirrhotics	Exclusions: congestive heart failure, multiple hepatic cysts, uncontrolled systemic infection, unrelieved biliary obstruction, severe pulmonary hypertension, hepatoma, portal vein thrombosis, and severe coagulopathy.	Wallstent/Fluency	10/8	1–4	Shunt venography	8.3 ± 4.4/6.2 ± 3.9
Perarnau et al., 2014,²⁵ France	Randomized	Cirrhotics	Inclusion: Baveno criteria, 18–75 years old, cirrhosis by historical/clinical signs, Child-Pugh score < C12.	Luminexx, Palmaz, Genesis, Smart, Wallstent, Zilver, Fluency alone Fluency+bare stent, Viatorr	NA	1/2–4	Clinical follow-up, Doppler ultrasound assessment at 1 month and then every 3 months over a total follow-up period of 2 years, angiography every 6 months	21.8/23.6^b
Perarnau et al., 2014,²⁵ France	Randomized	Cirrhotics	Exclusion: acute uncontrolled bleeding, portal vein thrombosis. Hepatocellular carcinoma, heart failure, pulmonary hypertension, hepatic polycystosis, dilatation of intrahepatic bile ducts, history of recurrent spontaneous hepatic encephalopathy (HE), hepatic arterial insufficiency, pregnancy, breastfeeding, or childbearing potential without contraception.		NA	1/2–4		21.8/23.6^b
Wang et al., 2016,²⁶ China	Randomized single center	Cirrhotics with portal hypertension	Inclusion: portal hypertension with defined indications for TIPS, scheduled for elective TIPS, 18–70 years old	EV3, protégé; Cordis, Smart/Bard, Fluency	NA	1–4	Clinial follow-up at 3 and 6 months, liver function, coagulation, blood ammonia, blood routine, color ultrasonography, esophagography, CT, and gastroscopy	NA
Wang et al., 2016,²⁶ China	Randomized single center	Cirrhotics with portal hypertension	Exclusion: combined with hepatic encephalopathy before treatment/portal vein thrombosis/malignant liver tumor or other malignancy/hemorrhage/gastrointestinal ulcer	EV3, protégé; Cordis, Smart/Bard, Fluency	NA	1–4		NA
Nonrandomized studies
Angermayr et al. 2003,¹⁸ Austria	Retrospective analysis	Cirrhotics	Inclusions: h/o variceal bleeding, refractory ascites	—/Viatorr endoprosthesis	72.1/80	4	NA	32.9/13.8
Angermayr et al. 2003,¹⁸ Austria	Retrospective analysis	Cirrhotics	Exclusions: emergency TIPS, severe infections, malignant tumors, organic renal disease, CP comorbidity, no cirrhosis, <18 years	—/Viatorr endoprosthesis	72.1/80	4	NA	32.9/13.8
Angeloni et al. 2004,¹⁴ Italy	Retrospective BS control, Prospective PTFE	Cirrhotics	Cirrhotics	—/Viatorr endoprosthesis	73.6/53.1	1/2–4	Clinical examination, laboratory tests, ultrasonography, angiography	37.5 ± 29.6 vs. 11.2 ± 7.4
Hernández-Guerra et al. 2004,¹⁵ Spain/UK	Prospective nonrandomized study	Budd-Chiari Syndrome	Budd-Chiari Syndrome with refractory ascites and esophageal variceal bleeding	Wallstent (Boston Scientific)/Viatorr (Gore)	28.6% (overall)	¼	USG, Portography	21.2/19.1
Tripathi D. 2014, UK	Retrospective BS control, prospective PTFE	Cirrhotics	Exclusion criteria: technical failure of TIPS, use of PTFE to restent bare metal stent, clinical trial of only 1-year follow-up, follow-up at a different center, h/o OLT, non-Viatorr stent	—/Viatorr (Gore)	63/71.3	1–4	Doppler, portography, clinical, hematological, biochemical assessment	22.8 ± 25.4 vs. 13.1 ± 12.5
Jung et al. 2009,²¹ USA	Retrospective	Liver disease^c	De novo TIPS for variceal bleeding, refractory ascites, or hydrothorax	Wallstents, Smart Stents/Viatorr	10/10^b	1/2–4	Doppler USG/portography, clinical, hematologic, biochemical assessment	12 months
Wu et al. 2010,²⁴ China	Retrospective	Portal Hypertension	Inclusion: portal HTN	NiTi/Fluency	8/8	1–4	Doppler, portography, clinical, hematological, biochemical assessment	8.34 ± 4.42 vs. 6.16 ± 3.89
Clark et al. 2011,²⁰ USA	Retrospective review of prospective data	Cirrhotics	Inclusion: portal HTN	None mentioned	NA	¼	Angiography, clinical assessment, recurrent/refractory ascites, liver failure	48 ± 44.7 vs. 24 ± 21.1
Clark et al. 2011,²⁰ USA	Retrospective review of prospective data	Cirrhotics	Exclusion: previous liver transplant	None mentioned	NA	¼		48 ± 44.7 vs. 24 ± 21.1
Sommer et al. 2012,²³ Germany	Retrospective analysis	Cirrhotics	Inclusion criteria: elective, in-house follow-up	Cordis/Viatorr	NA	1/3.4	Clinical evaluation, angiography, portography (not all patients)	11.2^d + 15.3^e/14.5^d + 21.7^e
Tan et al. 2015,¹⁷ Canada	Retrospective chart review	Cirrhotics	Inclusion: cirrhotic patients undergoing elective TIPS for refractory ascites	None mentioned	10/10	1/3.4	USG, Portography	25.1 vs. 14.1
Tan et al. 2015,¹⁷ Canada	Retrospective chart review	Cirrhotics	Exclusion: spontaneous hepatic encephalopathy, organic renal disease, cardiopulmonary disease (EF <55%, diastolic dysfunction, abnormal right-sided cardiac pressures)	None mentioned	10/10	1/3.4	USG, Portography	25.1 vs. 14.1
Lauermann et al. 2016,²² Germany, Ireland	Retrospective single center study	Portal Hypertension	No criteria mentioned (all patients)	Wallstent/Viatorr	12.0/9.7	1/3	Clinical assessment, USG (not in all patients)	12 months

1: Primary patency at 1 year, 2: Rebleeding, 3: Hepatic encephalopathy, 4: Survival at 1 year.

Median.

Liver diseases included were cirrhosis, Budd-Chiari syndrome, Cryptogenic, Hepatitis, Primary biliary cirrhosis, Primary sclerosing cholangitis, Hemochromatosis.

Angiographic follow-up, not available for all patients.

Clinical follow-up, not available for all patients.

CP, cardiopulmonary; EF, ejection fraction; HCC, hepatocelluar carcinoma; HTN, hypertension; NA, not available; NiTi, Nitinol; OLT, orthotopic liver transplant; PTFE, Polytetrafluoroethylene; TIPS, Transjugular Intrahepatic Portosystemic Shunt; USG, ultrasonography.

Table 3.

Characteristics of Patient Population in the Included Studies

First Author, Year	Sample size (n)	Mean age (years) (Bare/PTFE)	(Bare/PTFE) Male patients, %	Alcohol-induced cirrhosis, % (Bare/PTFE)	(Bare/PTFE) Child-Pugh C%	(Bare/PTFE) Mean Child-Pugh score	(Bare/PTFE) mean MELD score	(Bare/PTFE) indication for TIPS, %		(Bare/PTFE) Emergency TIPS, %	(Bare/PTFE) Serum bilirubin, mg/dL	(Bare/PTFE) Serum albumin, g/dL	(Bare/PTFE) Serum creatinine, mg/dL	(Bare/PTFE) Prothrombin index, %	(Bare/PTFE) PPG before TIPS, mmHg	(Bare/PTFE) PPG after TIPS, mmHg
First Author, Year	Sample size (n)	Mean age (years) (Bare/PTFE)	(Bare/PTFE) Male patients, %	Alcohol-induced cirrhosis, % (Bare/PTFE)	(Bare/PTFE) Child-Pugh C%	(Bare/PTFE) Mean Child-Pugh score	(Bare/PTFE) mean MELD score	Bleeding	Ascites	(Bare/PTFE) Emergency TIPS, %	(Bare/PTFE) Serum bilirubin, mg/dL	(Bare/PTFE) Serum albumin, g/dL	(Bare/PTFE) Serum creatinine, mg/dL	(Bare/PTFE) Prothrombin index, %	(Bare/PTFE) PPG before TIPS, mmHg	(Bare/PTFE) PPG after TIPS, mmHg
Randomized studies
Bureau et al. 2007,¹⁹ France, Spain, Canada	41/39	55 ± 12/55 ± 10	73.1/69.2	54/56	NA	9 ± 2/9 ± 2	NA	34.1/23^a36.5/25.6^b	29.2/51.2	NA	3.6 ± 4.2/2.6 ± 2.6	2.7 ± 0.6/3.0 ± 0.5	1.0 ± 0.4/1.1 ± 0.4	61 ± 18/63 ± 17	20 ± 5/20 ± 7	7 ± 4/7 ± 4
Huang et al. 2010,²⁷ China	30/30	50 ± 10/47 ± 11	76,6/73,3	3,3/6,6	23,3/26,6	8,4 ± 2,0/8,6 ± 2,1	NA	86,6/83,3	13,3/16,6^d	NA	NA	NA	NA	NA	26.5 ± 6 5.1/25.7 ± 66.6	11.0 6 ± 1.5/10.3 ± 4.4 (short term)9.5 ± 2.9/13.2 ± 1.5
Perarnau et al., 2014,²⁵ France	71/66	59/57 median	83,1/66,1	85,9/80,0	12,9/24,6	NA	11,6/11,2 median	28,1/33,8^a	76,1/70,7	NA	NA	NA	NA	NA	15/16	5,5/6
Wang et al., 2016, China	127/131	46,7 ± 5,0/45,4 ± 7,0	59,8/67,1	NA	29,2/26,7	7,1 ± 1,8/6,98 ± 1,4	NA	96/93,8	17,3/15,2	NA	NA	NA	NA	58 ± 29,5/55 ± 21,7 (platelet count)	3,77 ± 0,59/3,64 ± 0,83	2,51 ± 0,26/2,34 ± 0,38
Nonrandomized studies
Angermayr et al. 2003,¹⁸ Austria	419/89	54.9/56.1 (median)	72/79.7	68.2/71.9	23.6/29.2	NA	NA	68.2/49.4^a	18.8/50.5	0/0	NA	NA	NA	NA	NA/20.5 (median)	NA/7
Angeloni 2004,¹⁴ Italy	87/32	58.8 ± 10.4/57.2 ± 11.5	72/79.7	17.2/28.1	16.06/9.3	NA	9.4/10.1	83/66	17/28	23/19	1.5 ± 1.0/1.9 ± 2.3	3.3 ± 0.6/3.2 ± 0.5	1.2 ± 1.1/1.2 ± 0.7	64 ± 16/66 ± 17	22.2 ± 4.1/19.8 ± 4.7	9.3 ± 3.8/7.2 ± 2.4
Hernández-Guerra et al.¹⁵ 2004, Spain/UK	16/9	NA	NA	NA	NA	9/9	NA	6/22	81/56	NA	2.4 ± 1.0/3.6 ± 3.2	3.2 ± 0.3/3.2 ± 0.9	1.1 ± 0.3/0.9 ± 0.3	56 ± 16/52 ± 21	NA	NA
Tripathi D. 2014, UK	316/157	54.0 ± 12.7/53.1 ± 12.2	62.9/71.3	61.0/68.7	NA	9.8 ± 2.4/9.6 ± 2.3	NA	68.9/73.2	12.6/11.4	NA	NA	NA	NA	NA	22.2 ± 7.5/21.4 ± 6.8	6.8 ± 4.4/6.3 ± 4.0
Jung et al. 2009,²¹ USA	41/40	54.8 ± 11.9/56.3 ± 10.7	73.1/52.5	29.2/40	46.3/40	9.7 ± 2.2/9.3 ± 1.5	19.3 ± 9.9/17.1 ± 7.2	53.6/40	46.3/47.5^c	NA	NA	NA	NA	NA	18.7 ± 5.7/17.8 ± 5.2	6.7 ± 4.2/5.5 ± 3.3
Wu et al. 2010,²⁴ China	30/30	47 ± 11/50 ± 10	73.3/76.6	NA	26.6/23.3	NA	8.4 ± 1.5/8.6 ± 1.9	86.6/83.3	13.3/16.6	16.6/20	1.5 ± 0.7/1.4 ± 0.4	3.5 ± 0.5/3.5 ± 0.4	0.9 ± 0.1/0.9 ± 0.2	17 ± 4/16 ± 4^e	36 ± 7/35 ± 9	17 ± 2/13 ± 4
Clark et al. 2011,²⁰ USA	70/176	55/56 (median)	64.2/60.2	23/27	41/13	9/8	12.6/11.1	NA	NA	NA	NA	NA	NA	NA	19/17	6/5
Sommer et al. 2012,²³ Germany	116/58	56.3 ± 11.1/53.4 ± 16.1	63.7/60.3	69/51.7	10.3/19	7.6 ± 1.6/8.0 ± 1.8	11.2 ± 3.9/11.9 ± 5.0	21.6/27.6^d	51.7/36.2^d	0/0	NA	NA	NA	NA	20.6 ± 6.8/22.3 ± 8.9	7.6 ± 3.5/7.8 ± 3.6
Tan et al. 2015,¹⁷ Canada	104/32	54 ± 1/55 ± 1	78.8/75	67.3/65.6	NA	8.7 ± 0.2/8.2 ± 0.2	12.9 ± 0.5/10.4 ± 0.8	NA	NA	NA	1.9 ± 0.1/1.6 ± 0.2	3.3 ± 0.1/3.3 ± 0.1	1.1 ± 0.1/0.9 ± 0.05	1.46 ± 0.03/1.33 ± 0.4^f	18.8 ± 0.6/16.1 ± 0.9	6.9 ± 0.3/6.3 ± 0.5
Lauermann et al. 2016,²² Germany, Ireland	80/83	56.3 ± 12.3/56.8 ± 10.7	72.5/61.4	54.7	24.2	NA	NA	12.4	66.6	NA	NA	NA	NA	NA	15.7/16.3	4.6/5.5

Prevention.

Active bleed.

Hydrothorax

Refractory ascites and recurrent variceal bleeding.

Prothrombin Time (sec).

INR.

NA, not available.

Primary shunt patency at 1 year

Thirteen studies reported the rate of primary shunt patency at 1-year follow-up following the procedure (Fig. 2A); these included all four RCTs. The total number of included patients was 1938 (1097 in bare stent group and 841 in PTFE stent group) with primary patency of the shunt reported in 84% of patients with PTFE-covered stent graft and 52% of patients with bare stent at 1-year follow-up. The pooled analysis significantly favored PTFE stent grafts (odds ratio [OR] = 4.75, 95% confidence interval [CI] = 3.32–6.79, P < .00001), with a high between-study heterogeneity (χ² = 21.62, df = 12, I² = 44%, P = .04). The random-effect model was thus used.

FIG. 2.

Weighted pooled analysis of all included studies for the four outcome variables; right axis favors PTFE stent group. (A) Primary Patency at 1-year. (B) Rebleeding. (C) New-onset hepatic encephalopathy. (D) Survival at 1-year.

Interestingly, on subset analysis for the four RCTs only, there was minimum heterogeneity seen between the included studies (Fig. 3A). Eighty-six percent of patients with PTFE stents and 61% of patients with bare stents were reported to have primarily patent stents at 1 year. Using the Mantel–Haenszel method with fixed-effect model, the PTFE stent group had significantly higher odds of primary patency at 1 year (OR = 4.18, 95% CI = 2.66–6.55, P < .00001) with low heterogeneity (χ² = 0.77, df = 3, I² = 0%, P = .86).

FIG. 3.

Weighted pooled analysis of only randomized controlled trials for the four outcome variables; right axis favors PTFE stent group. (A) Primary Patency at 1-year. (B) Rebleeding. (C) New-onset hepatic encephalopathy. (D) Survival at 1-year.

Rebleeding

Rebleeding following the TIPS procedure was the least reported parameter, being reported by only six studies, including three of the four RCTs (Fig. 2B). The total number of included patients was 1050 (631 in bare stent group and 419 in PTFE stent group). Rebleeding occurred in 9% patients following PTFE stent placement compared with 18% patients with bare stent. The pooled analysis significantly favored PTFE stent grafts (OR = 0.37, 95% CI = 0.24–0.56, P < .00001), with low between-study heterogeneity (χ² = 3.18, df = 5, I² = 0%, P = .67). The three RCTs included 398 subjects (PTFE stent, 200, and bare stent, 198), and showed a higher rate of rebleeding for both PTFE stent (14%) and bare stent (27%) compared with the total pooled rates for all six studies (Fig. 3B). The weighted pooled analysis continued to favor PTFE stents over bare stents (OR = 0.43, 95% CI = 0.25–0.72, P < .001) with a low heterogeneity (χ² = 1.58, df = 2, I² = 0%, P = .45).

Hepatic encephalopathy

The onset of HE in patients following a TIPS procedure was reported by 12 studies, including all four RCTs (Fig. 2C). There were a total of 1638 pooled patients (980 in bare stent group and 658 in PTFE stent group). The rate of HE was not significantly different (26% in PTFE stent vs. 29% in bare stent) and the pooled analysis failed to favor any one stent type over another (OR = 0.86, 95% CI = 0.68–1.08, P = .19). There was low between-study heterogeneity (χ² = 11.49, df = 11, I² = 4%, P = .40), and there was no publication bias seen on the funnel plot. The subset pooled analysis of only RCTs continued to show no difference in the rate of developing HE following the use of either stent (PTFE stent group, 33%, and bare stent group, 35%; OR = 0.91, 95% CI = 0.63–1.32, P = .63; χ² = 2.21, df = 3, I² = 0%, P = .53) (Fig. 3C).

Survival at 1 year

The data for survival at 1 year was available from 11 studies and included 2152 subjects (1413 in bare stent group and 739 in PTFE stent group) (Figs. 2D and 4). The pooled rate of 1-year survival was higher for PTFE stent grafts (85%) compared with bare stents (73%); the weighted analysis significantly favored PTFE stent grafts (OR = 1.85, 95% CI = 1.44–2.38, P < .00001) with low between-study heterogeneity (χ² = 9.82, df = 10, I² = 0%; P = .46).

FIG. 4.

Funnel plot with 95% CI lines for survival at 1 year.

All four randomized trials reported 1-year survival following the placement of a TIPS stent (Fig. 3D). A total of 519 subjects were available with 258 subjects having PTFE stent grafts and 261 subjects with bare stents. The rate of 1-year survival in these patients was higher than the overall pooled rates from all 11 included studies and continued to be in favor of PTFE stent grafts (PTFE stent group, 91%, vs. bare stent group, 83%). The pooled analysis significantly favored PTFE stent group (OR = 2.09, 95% CI = 1.18–3.69, P < .001), with low between-study heterogeneity (χ² = 1.27, df = 3, I² = 0%; P = .74).

Discussion

TIPS is a nonselective portosystemic shunt, which has gained wide acceptance and has helped reduce morbidity and mortality in patients with portal hypertension. The procedure was first described by Dr Josef Rösch in 1969 in a canine model, in what he described as a combination of science, curiosity, and serendipity.^28–30 Following pioneering works by Colapinto et al.,³¹ Palmaz et al. reported the first successful clinical application of TIPS in human subjects.^32–34

The emergency selective surgical shunts are precluded in patients with decompensated liver function (Child-Turcotte-Pugh class C) and poor operative risk. Although TIPS was initially developed as an emergency procedure for refractory acute variceal bleed, its indications are no more limited to the acute hemorrhage. It has gained popularity as an elective procedure in patients with refractory portal hypertension or ascites, who are poor candidates for surgical shunts or liver transplantation.^4,35 TIPS is also considered in selected patients awaiting liver transplantation; however, the benefits are debatable, and a prior TIPS procedure may increase technical difficulties during liver transplant.^35,36

The high rates of severe adverse effects that were reported in the initial years of use of bare stents led to the birth of a new generation of stents, PTFE-covered stent grafts. The 2009 update of the American Association for the Study of Liver Diseases (AASLD) recommended the use of PTFE stents over bare stents.³⁷ TIPS dysfunction was the most common complication with bare stents; however, there was unclear evidence supporting either stent over the other in terms of survival and risk of encephalopathy.³⁷

In the present meta-analysis, we aimed to provide current evidence on clinical outcome following the use of bare stents and PTFE stent grafts for the TIPS procedure. We have reported the primary patency at one year, risk of rebleeding, new-onset HE following the procedure, and 1-year survival rates. Yang et al. reported improved primary patency and survival and lower risk of encephalopathy with PTFE stents.⁸ However, due to the lack of adequate number of well-powered studies, their meta-analysis did not reach significance in all studied domains.

Qi et al. published the pooled analysis of two RCTs comparing the two types of stents.³⁸ Their results showed an improved patency and the trend toward a better overall survival for covered stents over bare stents. Their updated analysis consisted of four RCTs from both East and West and presented subgroup analyses on different types of covered stents.³⁹ Overall survival and shunt patency were better with covered stents. Interestingly, the improved results of shunt patency with covered stents did not affect the risk of encephalopathy. Nonetheless, these meta-analyses presented several limitations, such as small and heterogenous populations, inclusion of various indications for TIPS, and missing data points. By far, there is an observable lack of systematic approach to all these prior published meta-analyses.

Occlusion or stenosis of the stent defines TIPS dysfunction. This is, in fact, the most common complication of the TIPS procedure.^40,41 Occlusion can either be due to an acute thrombosis, which causes the early stent dysfunction, or pseudointimal hyperplasia of the hepatic vein outflow tract, which is more likely to occur at 3 months to 1 year following the procedure.^42–44 The evaluation of shunt patency is routinely done by direct portography and less often by doppler ultrasound. Thirteen out of 14 studies (Fig. 2A) reported data on stent sufficiency at 1 year, and all of them favored PTFE stents with regard to better primary patency and lower rates of occlusion. Interestingly, although Angeloni et al. concluded that PTFE stents presented with lower rate of stenosis of the stent tract, they were followed by a higher incidence of hepatic or portal vein stenosis.¹⁴ Our pooled analysis showed a high heterogeneity among the included studies, compelling the use of random-effect model. However, the heterogeneity was minimal when only four RCTs were used for pooled analysis (Fig. 3A), thus showing a clear benefit of PTFE stent grafts over bare stents. Covered stents have an antithrombotic effect and tend to diminish the incidence of pseudointimal hyperplasia by eliminating the inflammatory process. Overall, the advantages of stent sufficiency that covered stents provide seem to be undebatable.

The association of recurrent hemorrhage with increased mortality requires early diagnosis and control with repeat endoscopy. The clear superiority of PTFE stent grafts over bare stents with respect to lesser chances of recurrent hemorrhage was clearly exemplified by subgroup analysis of three available RCTs (Figs. 2B and 3B).

By far, the most interesting results in our study were seen with regard to the new-onset encephalopathy following TIPS procedure. Data were available from 12 studies (Fig. 2C). Authors have reported a high incidence of new diagnoses of HE after PTFE stent placement.⁴⁵ The parameters found to correlate with new-onset HE following TIPS procedure were male sex, older age, higher creatinine level, and lower serum sodium and albumin levels.^16,19,45 Increasing diameter of the stent and portal pressure gradient <5 mm Hg relate with a lower incidence of encephalopathy.²⁴ We believe that the occurrence of new-onset encephalopathy (newly diagnosed or worsening HE postprocedure), rather than the absolute presence of encephalopathy, is a more direct indicator of functional outcome following TIPS procedure. In our meta-analysis, two-thirds of the studies favored covered stents, while the others favored bare stents with respect to lower rates of new-onset HE. There was a temporal trend in outcomes with earlier studies favoring bare stent, while the more recent studies favoring PTFE stent grafts. The technical expertise, quality of stents, surveillance protocols, and ability to restent a blocked stent have improved over time. The pooled analysis finds no clear superiority of one type of stent over another with regard to prevention of postprocedure encephalopathy, which highlights the role of shunting of blood bypassing liver parenchyma being the underlying mechanism common among both types of stents. Almost 28% of all included patients developed encephalopathy postprocedure, which highlights the inherent shortcomings of TIPS to be ever used as a definitive procedure.

Most importantly, our analysis indicated a trend toward better 1-year survival following the new generation of stents. Overall, 1-year survival of the covered stents was estimated between 79.1% and 90.4%, while the rates for the bare stents were 73%–87.5%. Age, Child-Pugh score, and Model for End-Stage Liver Disease (MELD) score are important predictors of survival among patients with end-stage liver disease.^4,18,46,47 The use of covered stents seems more reasonable as was reported by all included studies, although the mean MELD and Child-Pugh scores did not differ significantly between the two groups of patients.

Caution should be taken to explain the pooled results from varied study designs. We performed a subanalysis that only included the RCTs to refine the quality of evidence. It is notable that in seven of the total included articles, the sample size of patients in the PTFE stent group was much lower than that in the bare stent group. Furthermore, most of the nonrandomized studies had adapted bare stents earlier than the introduction of PTFE stent grafts. The new-generation covered stents have clear superiority in three of the four studied variables over bare stents, which may, atleast in part, be attributed to the technological advances in the field of interventional gastroenterology over the past decade.

This meta-analysis lays down a clear advantage of PTFE-covered stents over bare stents with improved primary patency and short to mid-term survival, while risk of recurrent hemorrhage tends to disseminate with the former. The use of bare stents seems limited in these therapeutic protocols. The management of these complications should be reassessed in more randomized controlled studies with larger sample sizes to precisely and independently evaluate the possible advantages of the new-generation stents in view of clinical outcomes.

Footnotes

Authors' Contributions

Study concept and design: Saurabh Singhal, Piyush Aggarwal, and Ekansh Gupta

Literature search and acquisition of data: Piyush Aggarwal, Ekansh Gupta, and Saurabh Singhal

Statistical analysis and interpretation of data: Piyush Aggarwal and Saurabh Singhal

Drafting of the article: Tania Triantafyllou, Piyush Aggarwal, Wendy Jo Svetanoff, Ekansh Gupta, Deepak Prakash Bhirud, and Saurabh Singhal

Critical revision of the article: Tania Triantafyllou, Wendy Jo Svetanoff, Deepak Prakash Bhirud, and Saurabh Singhal

Administrative, technical, and material support: Saurabh Singhal

Study supervision: Saurabh Singhal

Disclosure Statement

No competing financial interests exist.

References

Graham

, Smith

. The course of patients after variceal hemorrhage. Gastroenterology, 1981; 80:800–8009.

Ali

, Wu

, Xu

, Liu

, Hao

, Qin

. A prospective study of endoscopic injection sclerotherapy and endoscopic variceal ligation in the treatment of esophageal varices. J Laparoendosc Adv Surg Tech A, 2017; 27:333–341.

Van

, Cambre

, Sun

. A new endoscopic elastic band ligating device. Gastrointest Endosc, 1986; 32:230–233.

Pateria

, Jeffrey

, Garas

, Tibballs

, Ferguson

, Delriviere

, et al. Transjugular intrahepatic portosystemic shunt: Indications, complications, survival and its use as a bridging therapy to liver transplant in Western Australia. J Med Imaging Radiat Oncol, 2017; 61:441–447.

Shah

, Sze

. Complications during transjugular intrahepatic portosystemic shunt creation. Tech Vasc Interv Radiol, 2016; 19:61–73.

PRISMA Checklist at: http://prisma-statement.org/PRISMAStatement/Checklist.aspx (accessed April 10, 2017).

Cochrane handbook for systematic reviews and interventions v5.1.0 at: http://handbook.cochrane.org (accessed April 10, 2017).

Yang

, Han

, Wu

, Ye

, Jin

, Yin

, et al. Patency and clinical outcomes of transjugular intrahepatic portosystemic shunt with polytetrafluoroethylene-covered stents versus bare stents: A meta-analysis. J Gastroenterol Hepatol, 2010; 25:1718–1725.

Moher

, Cook

, Eastwood

, Olkin

, Rennie

, Stroup

. Improving the quality of reports of meta-analyses of randomised controlled trials: The QUOROM statement. Quality of Reporting of Meta-analyses. Lancet, 1999; 354:1896–1900.

10.

Slim

, Nini

, Forestier

, Kwiatkowski

, Panis

, Chipponi

. Methodological index for non-randomized studies (minors): Development and validation of a new instrument. ANZ J Surg, 2003; 73:712–716.

11.

Review Manager (RevMan) [Computer program]. Version 5.3. Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014.

12.

DerSimonian

, Laird

. Meta-analysis in clinical trials. Controll Clin Trials, 1986; 7:177–188.

13.

Mantel

, Haenszel

. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst, 1959; 22:719–748.

14.

Angeloni

, Merli

, Salvatori

, De Santis

, Fanelli

, Pepino

, et al. Polytetrafluoroethylene-covered stent grafts for TIPS procedure: 1-year patency and clinical results. Am J Gastroenterol, 2004; 99:280–285.

15.

Hernández-Guerra

, Turnes

, Rubinstein

, Olliff

, Elias

, Bosch

, et al. PTFE-covered stents improve TIPS patency in Budd-Chiari syndrome. Hepatology, 2004; 40:1197–1202.

16.

Tripathi

, Macnicholas

, Kothari

, Sunderraj

, Al-Hilou

, Rangarajan

, et al. Good clinical outcomes following transjugular intrahepatic portosystemic stent-shunts in Budd-Chiari syndrome. Aliment Pharmacol Ther, 2014; 39:864–872.

17.

Tan

, James

, Sniderman

, Wong

. Long-term clinical outcome of patients with cirrhosis and refractory ascites treated with transjugular intrahepatic portosystemic shunt insertion. J Gastroenterol Hepatol, 2015; 30:389–395.

18.

Angermayr

, Cejna

, Koenig

, Karnel

, Hackl

, Gangl

, et al. Survival in patients undergoing transjugular intrahepatic portosystemic shunt: ePTFE-covered stentgrafts versus bare stents. Hepatology, 2003; 38:1043–1050.

19.

Bureau

, Pagan

, Layrargues

, Metivier

, Bellot

, Perreault

, 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; 27:742–747.

20.

Clark

, Golkar

, Luberice

, Toomey

, Paul

, Marcadis

, 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; 202:561–564.

21.

Jung

, Kalva

, Greenfield

, Waltman

, Walker

, Athanasoulis

, et al. TIPS: Comparison of shunt patency and clinical outcomes between bare stents and expanded polytetrafluoroethylene stent-grafts. J Vasc Interv Radiol, 2009; 20:180–185.

22.

Lauermann

, Potthoff

, Mc Cavert

, Marquardt

, Vaske

, Rosenthal

, et al. Comparison of Technical and Clinical Outcome of Transjugular Portosystemic Shunt Placement Between a Bare Metal Stent and a PTFE-Stentgraft Device. Cardiovasc Intervent Radiol, 2016; 39:547–556.

23.

Sommer

, Gockner

, Stampfl

, Bellemann

, Sauer

, Ganten

, et al. Technical and clinical outcome of transjugular intrahepatic portosystemic stent shunt: Bare metal stents (BMS) versus viatorr stent-grafts (VSG). Eur J Radiol, 2012; 81:2273–2280.

24.

, Ding

, Cao

, Han

, Huang

, Li

, et al. Favorable clinical outcome using a covered stent following transjugular intrahepatic portosystemic shunt in patients with portal hypertension. J Hepatobiliary Pancreat Sci, 2010; 17:701–708.

25.

Perarnau

, Le Gouge

, Nicolas

, d'Alteroche

, Borentain

, Saliba

, et al. Covered vs. uncovered stents for transjugular intrahepatic portosystemic shunt: A randomized controlled trial. J Hepatol, 2014; 60:962–968.

26.

Wang

, Xiao

, Yue

, Zhao

, Fan

, Zhao

, et al. Efficacy of covered and bare stent in TIPS for cirrhotic portal hypertension: A single-center randomized trial. Sci Rep, 2016; 6:21011.

27.

Huang

, Wu

, Fan

, Cao

, Han

, Xu

, Li

. Comparison study of Doppler ultrasound surveillance of expanded polytetrafluoroethylene-covered stent versus bare stent in transjugular intrahepatic portosystemic shunt. J Clin Ultrasound, 2010; 38:353–360.

28.

Rösch

, Hanafee

, Snow

. Transjugular portal venography and radiologic portacaval shunt: An experimental study. Radiology, 1969; 92:1112–1114.

29.

Rösch

, Hanafee

, Snow

, Barenfus

, Gray

. Transjugular intrahepatic portacaval shunt—an experimental work. Am J Surg, 1971; 121:588–592.

30.

Rösch

. Development of transjugular intrahepatic portosystemic shunt. J Vasc Interv Radiol, 2015; 26:220–222.

31.

Colapinto

, Stronell

, Gildiner

, et al. Formation of intrahepatic portosystemic shunts using a balloon dilation catheter: Preliminary clinical experience. AJR, 1983; 140:709–714.

32.

Palmaz

, Garcia

, Sibbitt

, et al. Expandable intrahepatic portacaval shunt stents in dogs with chronic portal hypertension. AJR, 1986; 147:1251–1254.

33.

Rössle

, Richter

, Nöldge

, Palmaz

, Wenz

, Gerok

. (1989). New non-operative treatment for variceal haemorrhage. Lancet, 1989; 2:153.

34.

Richter

, Palmaz

, Nöldge

, et al. The transjugular intrahepatic portosystemic stent-shunt (TIPSS): A new non-operative, transjugular percutaneous procedure. Radiologe, 1989; 29:406–411.

35.

Millis

, Martin

, Gomes

, Shaked

, Colquhoun

, Jurim

, et al. Transjugular intrahepatic portosystemic shunts: Impact on liver transplantation. Liver Transpl Surg, 1995; 1:229–233.

36.

Fagiuoli

, Bruno

, Debernardi Venon

, Schepis

, Vizzutti

, Toniutto

, et al. Consensus conference on TIPS management: Techniques, indications, contraindications. Dig Liver Dis, 2017; 49:121–137.

37.

Boyer

, Haskal

. American Association for the Study of Liver Diseases. The Role of Transjugular Intrahepatic Portosystemic Shunt (TIPS) in the management of portal hypertension: Update 2009. Hepatology, 2010; 51:306.

38.

, Guo

, Fan

. A trend toward the improvement of survival after TIPS by the use of covered stents: A meta-analysis of two randomized controlled trials. Cardiovasc Intervent Radiol, 2015; 38:1363–1364.

39.

, Tian

, Zhang

, Yang

, Guo

. Covered versus bare stents for transjugular intrahepatic portosystemic shunt: An updated meta-analysis of randomized controlled trials. Therap Adv Gastroenterol, 2017; 10:32–41.

40.

Boyer

, Vargus

. Transjugular intrahepatic portosystemic shunt. In: Schiff

, Sorrell

, Maddrey

(eds), Schiff's Diseases of the Liver, 9th ed. Philadelphia: Lippincott, Williams & Wilkins Co., 2003, pp. 369–382.

41.

Rössle

, Siegerstetter

, Huber

, Ochs

. The first decade of the transjugular intrahepatic portosystemic shunt (TIPS): State of the art. Liver, 1998; 18:73–89.

42.

Zhuang

, Bettmann

, Teng

, Jeffery

, Langdon

. Long-term results in patients with TIPS: An outcome study. J Vasc Interv Radiol, 2001; 12 (Suppl):S9–S10.

43.

Haskal

, Pentecost

, Soulen

, Shlansky-Goldberg

, Baum

, Cope

. Transjugular intrahepatic portosystemic shunt stenosis and revision: Early and midterm results. AJR Am J Roentgenol, 1994; 163:439–444.

44.

Saxon

, Ross

, Mendel-Hartvig

, Barton

, Benner

, Flora

, et al. Transjugular intrahepatic portosystemic shunt patency and the importance of stenosis location in the development of recurrent symptoms. Radiology, 1998; 207:683–693.

45.

Riggio

, Angeloni

, Salvatori

, De Santis

, Cerini

, Farcomeni

, et al. Incidence, natural history, and risk factors of hepatic encephalopathy after transjugular intrahepatic portosystemic shunt with polytetrafluoroethylene-covered stent grafts. Am J Gastroenterol, 2008; 103:2738–2746.

46.

Dunn

, Jamil

, Brown

, Wiesner

, Kim

, Menon

, et al. MELD accurately predicts mortality in patients with alcoholic hepatitis. Hepatology, 2005; 41:353–358.

47.

Ferral

, Gamboa

, Postoak

, Albernaz

, Young

, Speeg

, McMahan

. Survival after elective transjugular intrahepatic portosystemic shunt creation: Prediction with model for end-stage liver disease score. Radiology, 2004; 231:231–236.