Survival in patients undergoing transjugular ... - Wiley Online Library

Survival in Patients Undergoing Transjugular lntrahepatic Portosystemic Shunt: ePTFE-Covered Stentgrafts Versus Bare Stents Bernhard Angermay,' Manfred Cejna,2Franz K~enig,~ Franz Karnel,*Franz H a ~ k lAlfred , ~ Gangl,' and Markus Peck-Radosavljevic' for the Vienna TIPS Study Group In patients with liver cirrhosis, implantation of a transjugular intrahepatic shunt (TIPS) leads to reduction of portal pressure, but not of mortality compared with other therapies. The high stenosis rates of conventional bare stents causes high reintervention rates and costs and may be correlated with poor survival. ePTFE-covered stentgrafts provide much improved patency rates, but their impact on survival is unclear. All suitable patients receiving either bare TIPS (419/466)or undergoing implantation of ePTFE endoprostheses (89/100) in several centers in Austria up to 2002 were included in this retrospective analysis. Both patient groups were compared regarding survival with Kaplan-Meier and Cox regression analysis. Unmatched and 1:1-matchedsurvival analyses were performed. Patients undergoing ePTFE stentgraft implantation had significantly higher survival rates in all analyses. The 3-month, 1-year, and 2-year survival rates were 93%, 88%, and 76% for the ePTFE-group and 83%, 73%, and 62% for conventional TIPS patients, respectively. The matched survival analyses validated these findings. The model of the stent, patient age, and Child-Pugh Class (CPC) were independent predictors of survival. In conclusion, patients undergoing ePTFEendoprosthesis implantation had higher survival rates within 2 years after TIPS-implantation. This may be the result of improved patency rates after correct placement (up to the inferior caval vein [Icv]) of the ePTFE stentgraft. These data should be validated in a prospective series. (HEPATOLOGY 2003;38:1043-1050.)

I

mplantation of a transjugular intrahepatic portosystemic shunt (TIPS) is a minimally invasive radiologic procedure resulting in a side-to-side portocaval shunt to lower portal pressure.l Currently, TIPS-implantation is performed to prevent recurrent variceal bleeding, for control of ascites refractory to sodium restriction and

Abbreviations: TIPS, transjugular intrahepaticportosystemic shunt; ePTFE, expanded polytPtajuoroethykne; ICK inferior caval vein; CPC, Child-Pugh class; HVPG, hepatic venous pressure gradient. From the 'Departments of Gartroenterology and Hepatology, 21nterventional Radioloo, and 3Medical Statistics, University of Vienna Medical School; the 4Department ofRadiology, &her Franz-JosefSpital, Vienna;and the 5Departmentof Gastroenterology and Hepatology, Krankenhaw der Elisabetbinen, Linz, Austria. Received April 4, 2002; acceptedJuly 22, 2003. The other members of the Vienna TIPS S d y Group are Harald Brunner, Peter Ferenci, ArnurFerlitsch, Michael Gschwantler, Anna Kreil, Johanna Lammer, Emanuel Lipinski, Helmuih Mendel, Ludwig Pichler, Monika Schmid, Wol&ang Sieghafl, and Martina Wichh. Address reprint requests to: Markus Peck-Raabsavljevic M.D., University of Vienna Medical School, Department of Medicine 4, Division of Gastroenterology and Hepatoloo, Waehringer Guertel 18-20, A-1 090 Vienna, Austria. E-mail: [email protected];fa: (43) I 40400 4735. Copyright 0 2003 by the American Associationfor the Study of Liver Diseases. 0270-9139/03/3804-0030$30.00/0

d0i:lO.1053ljhep.2003.50423

medical therapy, and for therapy of variceal bleeding unresponsive to endoscopic and medical thera~y.~-5 In nearly all published series, implantation of a TIPS led to reduction of rebleeding rates627 and improvement or resolution of ascites.8 Despite this therapeutic benefit, TIPS could not improve survival in patients undergoing TIPS for prevention of rebleeding in most studies7 and there are contradictory results of the effect of TIPS on survival in patients undergoing TIPS for refractory ascites.'-'' TIPS implantation is complicated by high reintervention rates, mainly caused by stenosis of the stent or its outflow tract.' These findings in regard to survival and complications are based on studies using bare stents for creation of a TIPS. Recently, an expanded polytetrafluoroethylene (ePTFE)covered endoprosthesis has become available for TIPS implantation. The ePTFE-covered stentgraft minimizes transmural permeation of bile and mucin (which are common causes of patency loss) and reduces tissue ingrowth into the graft, resulting in a much improved patency rate after correct placement.12-'5 The centers participating in our study were among the first centers to use ePTFE stentgrafts starting in 1999.12,13,16 Until now, no studies in1043

1044

ANGERMAYR ET AL.

HEPATOLOGY, October 2003

cluding more patients and providing longer followup data tation to have a reasonable patient number in the ePTFE for patients undergoing ePTFE endoprostheses have been group for comparison. Patients participating in a prospective, multicenter published. Furthermore, no definitive data comparing survival of patients undergoing bare and ePTFE-TIPS are trial of ePTFE-TIPS endoprosthesisls (n = 16) provided currently available. The aim of our study was to evaluate written informed consent to participate in the study after the survival of patients undergoing bare TIPS and the protocol had been reviewed by an institutional ethics committee and had been found to comply with the DecePTFE-covered endoprostheses. laration of Helsinki.21 All other patients (bare TIPS or ePTFE endoprosthesis) provided written informed conPatients and Methods sent to undergo the procedure as clinically indicated. We followed patients who underwent elective ePTFE StatisticaZAnalysis. The starting time for all survival stentgraft (June, 1999 to November, 2002) and bare analyses was the date of the TIPS procedure. ‘The primary TIPS placement (May, 1991 to November, 2002) in sev- endpoint was death within 2 years, so for all statistical eral centers in Austria. Elective was defined as TIPS being tests the observation period was resticted to 2 years to have performed for the prevention of variceal rebleeding or for a reasonable patient number in the ePTFE group. Theretreatment of refractory ascites. Patients in whom TIPS fore, all patients surviving longer than 2 years were cenwas performed for the control of active variceal bleeding sored 2 years after TIPS placement. Patients undergoing after failure of 2 sessions of endoscopic therapy within a orthotopic liver transplantation were censored at the date 24-hour period were designated emergency TIPS recipi- of transplantation. The main statistical tests for survival ents and excluded from the current analysis. Refractory analyses were Kaplan-Meier analysis, the log-rank test, ascites was defined as ascites that could not be controlled and Cox proportional-hazards regression analyses. by paracentesis every 2 weeks despite a sodium-restricted Survival Analysis of All Patients. Both patient diet and intensive diuretic therapy, consistent with the groups were compared regarding survival and regarding consensus statement.” The diagnosis of cirrhosis either baseline characteristics using test and t test for depenwas confirmed by biopsy or was based on clinical, radio- dent and independent variables. logic, and laboratory findings. The TIPS procedures were SurvivaZ Analysis of All Patients Since 1998. The performed with standard techniques.8.18~19In all patients study period for the bare TIPS patients was, at the most, undergoing ePTFE-covered stentgrafts, a Viatorr endo- 10 years, and for the ePTFE group 4 years. To exclude prosthesis (W.L. Gore, Flagstaff, AZ) was implanted. changes in the therapeutic management of patients with Each individual hospital record was reviewed to verify the advanced liver disease and portal hypertension effecting diagnosis and indication for TIPS and receive all relevant survival over time, an additional survival analysis includclinical and laboratory data. The patients were followed ing only patients since 1998 was performed. from their date of TIPS procedure until death, liver transOne-to-one Matched Survival AnaZyses. The plantation, or study closure (October 31,2002). Survival matching criteria were selected according to risk factors of data of all patients included in the analyses were obtained mortality after TIPS placement published eIse~here8,~~-2* from the national death index (Statistik Austria), which and according to the independent predictors of survival keeps accurate and up-to-date records of all demographic yielded by the Cox analysis performed. Age and Childdata in Austria. Pugh class (CPC) primarily were chosen as matching criFour hundred sixty-six patients underwent bare TIPS teria. In addition, we included the patients’ sex. In this placement during the study period and 100 patients had analysis, patients undergoing ePTFE stentgraft were an implantation of an ePTFE stentgraft. Patients with matched 1:1 according to CPC, age ( 21 year), and sex severe infection, malignant tumors, organic renal disease, with patients undergoing bare TIPS. or significant cardiopulmonary comorbidity at time of Although ascites is included into the CPC classificaTIPS placement were not included into the analysis, and tion as a variable and in Cox analysis indication for TIPS also excluded were patients without cirrhosis16,20and pa- was no independent predictor of survival in our patients, tients younger than 18 years of age or patients undergoing a further 1:l matched analysis including indication for emergency TIPS placement (n = 47). Patients receiving TIPS was performed to fully rule out refractory ascites as an ePTFE endoprosthesis for revision of a bare TIPS also an additional risk factor. Therefore, patients undergoing were excluded from the analysis (n = 15). Thus, a total of ePTFE endoprostheses were matched 1:1 according to 419 bare-TIPS patients and 89 patients who underwent age (51 year), CPC, and indication for TIPS. In this ePTFE stentgrafts were included into the analysis. We analysis, matching according to sex in a 1:1 manner was provide here the 2-year followup data after TIPS implan- not possible.

x2

HEPATOLOGY, Vol. 38, No. 4, 2003

ANGERMAYR ET AL.

1045

Table 1. Patient Characteristics of Several Subgroups of Patients

Number of patients Age (median, range) Gender (M/F) Etiology of cirrhosis (n/%) Alcoholic Viral CIyptogenic PBC Autoimmune Unspecified indication for TIPS (n/%) Prevention bleeding Refractoly ascites Other indications Unspecified Child-Pugh class (n/%): A B C Unspecified

ePlFE

1:l Bare TIPS (Age, Gender, CPS)

1:l Bare TIPS (Age, Indlcatlon, CPS)

All Bare TIPS

89 56.0 (35.3-76.6) 71/18

89 56.2 (35.1-76.8) 71/18

89 56.1 (35.1-75.6) 24/65

419 54.9 (20-81.2) 302/117

64/72 18/20 4/5 3/3 O/O

74/83 9/10 5/ 6 O/ 0 1/ 1

o/ 0

o/o

74/83 11/12 3/3 O/O 1/ 1 O/ 0

286/68 50/12 22/5 3/ 1 2/0.5 56/13

44/49 45/51

60/67 29/33

44/49 45/51

o/o o/ 0

o/o

o/o o/ 0

286/68 79/19 3/ 1 51/12

15/17 48/54 26/29

15/17 48/54 26/29

o/o

o/o

15/17 48/54 26/29 o/ 0

83/20 113/30 99/23 124/30

Comparison of the Dzrment Data Sets. To make the results of the different data sets comparable, Cox proportional-hazards regression analyses were performed. For the matched data sets, univariate Cox regression analyses were performed with TIPS model as the independent variable. For the full sample, a multivariate analysis was performed using age, sex, CPC, TIPS model (bare ePTFE), and indication for TIPS as independent variables. The hazard ratio and the correspondig 95% CI of the TIPS model are reported for all Cox models. We used the SAS statistical software system (Version 8.2; SAS Institute Inc., Cary, NC) to carry out the calculations. P-values < -05 were considered to indicate statistical significance.

O/O

The 3-month survival rate was loo%, 93%, and 87% for class A, B, and C; the 1 year survival rate was 1000~6, 87%, and 8 1%; and the 2 year survival rate was 1 OO%, 73%, and 54%, respectively. Liver ischemia after ePTFE stentgrafi placement, as previously reported,25 was not observed in our patients.

Patient Characteristics in the Bare TIPS Group (n = 419). Of all patients receiving a bare TIPS (I 17

female and 302 male), 286 underwent TIPS for prevention of variceal rebleeding, 79 for therapy of refractory ascites, and 3 for other indications. The underlying liver disease was caused by alcohol in 286 cases, by virus infection in 50 cases, and otherwise in 27 patients (Table 1). Eighty-three, 113, and 99 patients had CPC A, B, and C liver disease, respectively. The median age was 54.9 years Results (range, 20 to 8 1.2); the median follow up was 1,OO1 days Patient Characteristics in the ePTFE Group (n = (range, 1 to 3,527). The 3-month survival rate was 92%, 89). Of 89 patients (18 female and 71 male), 44 patients 89%, and 71% for class A, B, and C; the 1-year survival underwent TIPS for prevention of variceal bleeding and rate was 84%, 73%, and 64%; and the 2-year survival rate 45 patients for therapy of refractory ascites. The etiology was 76%, 60%, and 50% for classes A, B, and C, respecof the liver disease was alcohol-related in 64 patients and tively. Bare TIPS Group, 1:I Matched According to Age, virus-related in 18 patients; 7 patients presented with Sex, and CPC. The indication for TIPS was prevention other etiologies (Table 1). At the time of TIPS implantaof variceal reebleeding in 60 patients and therapy of retion, 15 patients had CPC A, 48 class B, and 26 class C liver disease. The median age was 56.1 years (range, 35.3 fractory ascites in 29 patients. The etiology of the cirrhosis to 76.6), median followup was 421 days (range, 5 to of the liver was alcohol-related in 74 patients and virus1,238 days). In this group, the hepatic venous pressure related in 9 patients; in 6 patients, cirrhosis was caused gradient (HWG) could be lowered from a median of otherwise (Table 1). The median age was 56.2 years (range, 35.1 to 76.8); 20.5 mm Hg (range, 8-40) to a median of 7 mm Hg (range, 2 to 17). the median followup was 2,155 days (range, 2 to 2,936).

1046


ANGERMAYR ET AL.

The HVPG could be lowered from a median of 26 mm Hg (range, 15 to 35) to a median of 12 mm Hg (range, 5 to 25) in this patient group. The 3-month survival rate was 84%) 81%, and 62%; the 1-year survival rate was 769'0, 70%) and 58%; and the 2-year survival rate was 76%, 58%) and 42% for classes A, B, and C, respectively. Bare TIPS Group, 1:1 Matched According to Age, Indication for TIPS, and CPC. The etiology of liver disease was alcohol-related in 74 patients and virus-related in 11 patients; in 4 patients the liver disease was caused otherwise (Table 1). Twenty-four female and 65 male patients underwent bare TIPS placement in this group. The median age was 56.1 years (range, 35.1 to 75.3) in this patient group; the median followup was 1,741 days (range, 2 to 2,936). The 3-month survival rate was 86%, 79%) and 65%; the 1-year survival rate was 80%, 62%, and 58%; and the 2-year survival rate was 80%) 509'0, and 42% for classes A, B, and C, respectively.

Survival Analysis (ePTFE Versus All Bare TIPS -89 Vmsm 419). Comparing both patient groups according to baseline characteristics (age, sex, etiology of liver disease, CPC, and indication for TIPS) yielded significant differences in CPS and indication for TIPS: In the ePTFE group, patients had more advanced liver disease (reflected by CPC), and more patients underwent TIPS for therapy of refractory ascites. The survival rates in the bare TIPS group were 83%, 73%, and 62% for 3 months, 1 year, and 2 years, and survival rates in the ePTFE group were 93%, 88%, and 76940, respectively (Table 2, Fig. 1). Comparing both groups in regard to survival yielded significantly higher survival rates in the ePTFE group (P = .01). Survival Analysis of All Patients Since 1998. Reviewing all single-patient records yielded no difference in the periprocedural management of portal hypertension in both patient groups. Nevertheless, to fully rule out possible changes in the therapeutic management of patients with advanced liver disease and portal hypertension over Table 2. Cumulative Sutvival Rates of Different Groups of Patients

ePTFE group (n = 89) 1:l matched bare-TIPS group (age, gender, CPS) 1:l matched bare-TIPS group (age, CPS, indication) all bare-TIPS patients (n = 419)

3-Month Sunlval

1-Year Sunlval

2-Year Sunlval

93%

88%

76%

77%

68%

60%

76%

64%

53%

83%

73%

62%

NOTE. Cumulative 3-month, 1-year, and 2-year survival rates of patients undergoing ePTFE-TIPS and corresponding sutvival rates of groups of patients undergoing bare TIPS.

0.1

1

00 0

100

200

3W

--

400

500

600

700

800

time (in days)

Fig. 1. Kaplan-Meier analysis of the whole patient sample. The continuous line shows the survival of patients undergoing ePTFE-TIPS, the dotted line of patients undergoing bare TIPS. The x-axis shows the days of follow-up. Patients undergoing ePTFE-TIPS had a significantly higher survival rate (P = .01).

time on survival, we performed a survival analysis including only patients since 1998. Since 1998, 82 patients underwent bare TIPS and 89 patients underwent ePTFETIPS. In regard to age, CPS, sex, and etiology of liver disease, there was no significant difference between these groups. The groups differed in indication for TIPS, with more patients with refractory ascites in the ePTFE group. Again, patients undergoing ePTFE-TIPS had a significantly higher survival rate than patients undergoing bare TIPS ( P = .004). One-to-one Matched Suryival AnaLysis (Age 2 I Year, Sex, CPC). Patients undergoing ePTFE-covered stentgrafts had a significantly higher survival rate than patients undergoing bare TIPS in this matched analysis ( P = .009). The cumulative survival rates for 3 months, 1 year, and 2 years were 77%, 68%, and 60% for the bare TIPS group and 93%, 88%, and 76% for the ePTFE group, respectively (Table 2, Fig. 2). Patients undergoing ePTFE stentgrafts had significantly (I'< .001) lower levels of HVPG after TIPS placement than patients undergoing bare TIPS. The mean reduction of HVPG was 66.7% in patients undergoing ePTFE endoprostheses and 52.2% in patients undergoing bare TIPS. In all patients (ePTFE and bare), the H W G could be lowered by at least 20% of baseline values. One-to-one Matched Survival Analysis (Age 2 1 Year, Indication for TIPS, CPC). Including the indication for TIPS into the matching criteria yielded the same results as the other analyses with a significantly (P= .OOl) higher survival rate for the ePTFE group. The survival rates for the bare group included in this analysis (n =

ANGERMAYR ET AL.


Fig. 2. (A) Kaplan-Meier analysis of the 1:l-matched analysis according to age, sex, and CPC. (B) Kaplan-Meier analysis of the 1:lmatched analysis according to age, indication, and CPC.The continuous lines indicate the sufvival of patients undergoing ePTFE-stentgrafts, the dotted lines of patients undergoing bare TIPS. The x-axis shows the days of follow-up. The ePTFEgroup had a significantly higher surviva1 rate than the bare group in both analyses.

1047

A 1: 8 'I'

3 05

3

04

u3 o1

no o

irn

ZUI

im

4ca

son

6w

'M

I

rrm (m days)

89) were 76%, 64%, and 53% for 3 months, 1 year, and 2 years, respectively (Table 2, Figure 2). Cox Proportional-Hazarcls Regression Analyses. In multivariate analysis, model of the stent, age, sex, CPC, and indication for TIPS were included as variables. Of these determinants, model of the stent, age, and CPC were independent predictors of survival (Table 3). The patients' sex as well as the indication for TIPS did not correlate with survival. To make the results of the different data sets comparable, further Cox regression analyses were performed. In all models, the confidence interval for the hazard ratio of the TIPS model covers the same area (Table 3), so there is no substantial difference in the results of the 3 models. This substantiates the robustness of the effect of the TIPS model on survival.

Discussion TIPS, by preventing complications of portal hypertension such as variceal bleeding or refractory ascites and hepatorenal syndrome, theoretically could have a beneficial effect on survival of patients with cirrhosis and portal hypertension. Interestingly, TIPS was not able to improve survival compared with other therapies in most studies to date.'J,8 This might have been attributable to insufficient patient sample sizes or observation periods or to the study design allowing cross-over between the randomized

groups from conventional or endoscopic therapy to TIPS in case of treatment failure in the non-TIPS group with concomitant intention-to-treat analysis.26 But it might be attributable also to the fact that TIPS has its own detrimental effects, which result in comparable patient survival despite a reduction in rebleeding episodes and in some studies also in bleeding-related deaths in the TIPS groups.3J7-30 These studies also showed that the high rate of TIPS stenosis is the worst factor of the TIPS procedure. TIPS dyshnction may result from various causes: acute thrombosis, often caused by technical failure; pseudointimal hyperplasia, resulting from biliary leaks of transsected bile ducts into the shunt lumen; and intimal hyperplasia of the hepatic vein outflow tract occurring between 3 months and 1 year after TIPS p1acement.l4,3'-33These mechanisms lead to recurrence of portal hypertension, usually revealed by variceal rebleeding or recurrence of ascites. The reported patency rates of bare TIPS are between 25% and 85% at 12-month followup, depending on the definition of patency and the method of evaluation, with an average of 50% in larger ~tudies.3~4-3~ Rebleeding of patients with TIPS is associated with reobstruction of the shunt in 90% of the cases. The high restenosis rate necessitates frequent followup visits to control shunt function and revisions when indicated and determines many of the relevant outcome parameters such as rebleeding, recurrence of ascites, hepatorenal syn-

Table 3. Univariate and Multivariate Cox Proportional Hazards Regression Analysis Whole Sample

Variable

Bare vs. ePTFE Age Child-Pugh score Gender Indication

Pr

>

2

,0118 ,0011 ,0038 ,7336 .0696

Hazard Ratio (Ci)

2.240 [1.2-4.21 1.038 1.218 0.916 1.597

1:l According to Age, Gender, and CPS

Pr > ,$

Hazard Ratlo (Cl)

.0111

2.300 [1.2-4.41

*

*

*

* *

* *

*

1:1 According to Age, Indicatlon, and CPS

Pr

> ,$

Hazard Ratio (CI)

,0011

2.834 [1.5-5.31

* *

* *

*

*

*

*

NOTE. Of the candidate variables shown in the left column, "gender" and "indication for TIPS did not correlate with suIvivaI. In all models, the confidence interval for the hazard ratio of the TIPS-model covers the same area, so there is no substantial difference in the results of the 3 models. *This variable was not used in the model.

1048

ANGERMAYR ET AL.

drome, spontaneous bacterial peritonitis, hospitalization, and cost. These parameters also are closely related to survival.37-39 Taking these considerations into account, it seems to be reasonable to speculate that prevention of TIPS stenosis and recurrence of portal hypertension and its complications may lead to a survival benefit. Implantation of an ePTFE-covered TIPS endoprosthesis has been shown to improve TIPS patency rates dramatically,'2-15 but definitive data on patient survival have not been presented so far. Because of the lack of prospective data with comparable patient numbers and followup times, we report our retrospective data, which are important for their clear implications. In our study, patients undergoing ePTFE-covered TIPS had higher survival rates than patients undergoing bare TIPS in any survival analyses performed. These data seem to be valid and robust, because we found uniform results in all the analyses and included a large number of patients from several different institutions. A Cox analysis was performed to evaluate independent predictors of survival and to evaluate possible differences in the results of the survival analyses. This analysis showed that the model of the TIPS used for implantation was an independent predictor of survival and that there are no substantial differences in the results of any of the survival analyses used here. This model also showed that the variables used as matching criteria were valid. In Cox analysis, baseline patient characteristics were used as candidate variables. Since several (more than 2) etiologies of liver disease could be observed in our patients, etiology of liver disease could not be included in the Cox model as variable. However, previous studies (also including our own patients) showed that in patients undergoing TIPS, etiology of liver disease does not determine ~urvival.*O,~~ When looking at baseline patient characteristics of all patient groups, no significant difference could be observed, except in 2 analyses: in analysis of the whole sample, patients of the ePTFE group had higher Child-Pugh scores and significantly more patients of this group underwent TIPS for therapy of refractory ascites. In 1:l matched analysis (according to age, sex, and CPC), significantly more patients of the ePTFE group underwent TIPS for therapy of refractory ascites. Because occurrence of refractory ascites is known to be a poor prognostic sign in patients with cirrhosis,42-44 in particular in patients undergoing TIPS placement when compared to bleeding patients,*O and CPC is known to correlate with survival, one would expect a higher mortality rate for the ePTFE group. Despite this negative selection bias at baseline, these patients had higher survival rates, which substantiates the beneficial effects of ePTFE-covered TIPS endoprotheses on survival.


The median HVPG after TIPS placement was significantly lower in the 1: 1-matched (age, sex, CPC) ePTFE group. Nevertheless, the HVPG could be lowered by 20% from baseline values in all patients of both groups, which is known to be effective in the therapy of complications of portal h~pertension.~5-~7 We cannot fully rule out at the moment that the differences in the final median HVPG may play a role for the difference in survival. If so, it would seem that a sufficient pressure reduction can be achieved more easily by using the ePTFE-covered stentgrafts with the same diameters as the bare stents. Although indication for TIPS was no independent predictor of death in our patients, a further I:l-matched analysis including this variable was performed to completely match the patients at baseline. This analysis yielded a survival benefit for patients undergoing ePTFE stentgrafts even more clearly. Bare TIPS and ePTFE endoprostheses mainly differ in patency rates and probably also in encephalopathy rates. Because selection criteria and followup regimen did not change within the study period at the participating centers, it seems to be most probable that the improved survival in patients undergoing TIPS placement with the ePTFE endoprothesis is caused by the much improved patency rates, which does not seem to be offset by the higher encephalopathy rates to be expected with a stent that does not reduce its flow over time. Our study was performed outside a prospective study setting, and complete followup data according to shunt dysfunction and recurrence of complications of portal hypertension are not available in all patients, although higher patency rates of ePTFE stentgrafts, including some patients of this study, have been reported elsewhere.13 This is a limitation in this analysis, but, on the other hand, may be a stronghold too, because it reflects real life and clinical practice for the following reason: Shunt dysfunction can be associated with rebleeding, recurrence of ascites and its concomitant complications, or both. In patients under close surveillance, such as in patients included into a prospective series, these complications may be managed by early reinterventions. Shunt dysfunction and its impact on survival may play a minor role if detected and managed in due time before a clinically relevant complication such as variceal rebleeding or intractable ascites occurs. Although all patients in the participating hospitals in our study are intended for followup examinations every 6 months, a similar surveillance rate such as in prospective series cannot be achieved in routine medical practice without study protocols and aggressive patient recall procedures. Especially in patients not undergoing close surveillance, the primary patency rate of the stent implanted may determine survival.


ANGERMAYR ET AL.

1049

13. Cejna M, Peck-Radosavljevic M, Thurnher SA, Hittmair K, Schoder M, Lammer J. Creation of transjugular intraheparic portosystemicshunts with stent-grafts:initial experiences with a polytetrafluoroethylene-coverednitino1 endoprosthesis. Radiology 2001;221:437-446. 14. HaskaI ZJ. Improved patency of transjugular intrahepatic portosystemic shunts in humans: creation and revision with PTFE stent-grafts.Radiology 1999;213:759-766. 15. Otal P, Smayra T, Bureau C, Peron JM, Chabbert V, Chemla P, Joffre F, et al. Preliminary results of a new expanded-polytetrahoroethylene-covered stent-graft for transjugular intrahepatic portosystemic shunt procedures. AJRAm J Roentgen01 2002;178:141-147. 16. Cejna M, Peck-RadosavljevicM, Schoder M, Thurnher S, Ba-Ssalamah A, Angermayr B, Kaserer K, et al. Repeat interventions for maintenance of transjugular intrahepatic portosystemic shunt function in patients with Budd-Chiari syndrome. J Vasc Interv Radiol 2002;13:193-199. 17. Arroyo V, Gines P, Gerbes AL, Dudley FJ, Gentilini P, L& G, Reynolds TB, et al. Definition and diagnostic criteria of refractory ascites and hepatorenal syndrome in cirrhosis. International Ascites Club. Hepatology 1996;23:164-176. 18. Cejna M, Thurnher S, Pidlich J, Kaserer K, Schoder M, Lammer J. PriAcknowledgment: The authors thank Professor Peter mary implantation of polyester-covered stent-grafts for transjugular intraBauer for his expert statistical assistance and Barbara hepatic portosystemic stent shunts (TIPSS): A pilot study. Cardiovasc Intervent Radiol 1993;22:305-310. Tichy for her help in ascertaining patient data. 19. LaBerge JM, Ring EJ, Gordon RL, Lake JR, Doherty MM, Somberg KA, Roberts JP, et al. Creation of transjugular intrahepatic portosystemic shunts with the wallstent endoprosthesis:results in 100patients. Radiology 1993;187:413-420. 1. Bilbao JI, Quiroga J, Herrero JI, Benito A.Transjugular intrahepatic portosystemic shunt (TIPS): current status and future possibilities. Cardiovasc 20. Angermayr B, Cejna M, Schoder M, Wrba F, Valent P, Gang1 A, PeckRadosavljevic M. Transjugular intrahepatic portosystemic shunt for treatIntervent Radiol 2002;25:251-269. ment of portal hypertension due to extramedullary hematopoiesis in 2. Ochs A, Rossle M, Haag K, Hauenstein KH, Deibert P, SiegerstetterV, idiopathic myelofibrosis. Blood 2002;99:4246-4247. Huonker M, et al. The transjugular intrahepatic portosystemicstent-shunt procedure for refractory ascites. N Engl J Med 1995;3321192-1197. 21. World Medical Organization. Declaration of Helsinki. Br Med J 1996; 313:1448-1449. 3. R o d e M, Deibert P, Haag K, Ochs A, Olschewski M, Siegerstetter V, Hauenstein K-H, et al. Randomised trial of transjugular-intrahepatic-por- 22. Infante-Rivard C, Esnaola S, Villeneuve JP. Clinical and statisticalvalidity tosystemic shunt versus endoscopy plus propranolol for prevention of of conventional prognostic factors in predicting short-term survival among variceal rebleeding. Lancet 1997;349:1043-1049. cirrhotics. HEPATOLOGY 1987;7:660-664. 4. R o d e M, Haag K, Ochs A, Sellinger M, Noldge G, Perarnau JM, Berger 23. Malinchoc M, Kamath PS, Gordon FD, Peine CJ, Rank J, ter Borg PC. A E, et al. The transjugular intrahepatic portosystemicstent-shunt procedure model to predict poor survival in patients undergoing transjugular intrafor variceal bleeding. N Engl J Med 1994;330:165-171. hepatic portosystemic shunts. HEPATOLOGY 2000;3 1:864-871. 5. Sanyal AJ.The use and misuse of transjugular intrahepatic portasystemic 24. Sanyal AJ, Freedman AM, Luketic VA, Purdum PP 3rd, Shiffman ML, shunts. Curr Gastroenterol Rep 2000;2:61-71. DeMeo J, Cole PE, et al. The natural history of portal hypertension after 6. Luca A, DAmico G, La Galla R, Midiri M, Morabito A, Pagliaro L.TIPS transjugular intrahepatic portosystemic shunts. Gastroenterology 1997; for prevention of recurrent bleeding in patients with cirrhosis: meta-anal112:889-898. ysis of randomized clinical trials. Radiology 1999;212:411-421. 25. Bureau C, Otal P, Chabbert V, Peron JM, Rousseau H, Vine1JP. Segmen7. Papatheodoridis GV, Goulis J, Leandro G, Patch D, Burroughs AK.Transtal liver ischemia after TIPS procedure using a new PTFE-covered stent. jugular intrahepatic portosystemic shunt compared with endoscopic treatHEPATOLOGY 2002;36: 1554. ment for prevention of variceal rebleeding: A meta-analysis. HEPATOLOGY 26. Rossle M. Prevention of rebleeding from oesophageal-gastric varices. Eur J 1999;30:612-622. Gastroenterol Hepatol2001; 13:343-348. 8. Rossle M, Siegerstetter V, Huber M, Ochs A. The first decade of the 27. Cabrera J, Maynar M, Granados R, Gorriz E, Reyes R, Pulido-Duque JM, transjugular intrahepatic portosystemicshunt (TIPS): state of the art. Liver Rodriguez SanRoman JL, et al. Transjugular intrahepatic portosystemic 1998;18:73-89. shunt versus sclerotherapyin the elective treatment of variceal hemorrhage. 9. Gines P, Uriz J, Calahorra B, Garcia-Tsao G, Kamath PS, Del Arbol LR, Gastroenterology 1996;110:832-839. Planas R, et al. Transjugular intrahepatic portosystemic shunting versus paracentesis plus albumin for refractory ascites in cirrhosis. Gastroenterol- 28. EscorsellA, Banares R, Garcia-PaganJC, Gilabert R, Moitinho E, Piqueras B, Bru C, et al. TIPS versus drug therapy in preventing variceal rebleeding ogy 2002;123: 1839-1847. in advanced cirrhosis: a randomized controlled trial. HEPATOLOGY 2002; 10 Rossle M, Ochs A, Gulberg V, Siegerstetter V, Holl J, Deibert P, 0135:385-392. schewski M, et al. A comparison of paracentesis and transjugular intrahepatic portosystemicshunting in patients with ascites. N Engl J Med 2000; 29. Garcia-Villarreal L, Martinez-Lagares F, Sierra A, Guevara C, Marrero JM, Jimenez E, Monescillo A, et al. Transjugular intrahepatic portosystemic 342: 1701-1707. shunt versus endoscopic sclerotherapy for the prevention of variceal re11. Sanyal AJ, Genning C, Reddy KR, Wong F, Kowdley KV, Benner K, bleeding after recent variceal hemorrhage. HEPATOLOGY 1999;29:27-32. McCashland T. The North American Study for the Treatment of Refrac30. Merli M, Salerno F, Riggio 0, de Franchis R, Fiaccadori F, Meddi P, tory Ascites. Gastroenterology 2003; 124:634-641. Primignani M, et al. Transjugular intrahepatic portosystemic shunt versus 12. Cejna M, Peck-Radosavljevic M, Thurnher S, Schoder M, Rand T, Anendoscopic sclerotherapyfor the prevention of variceal bleeding in cirrhogermayr B, Lammer J. ePTFE-covered stent-grafts for revision of obsis: a randomized multicenter trial. Gruppo Italian0 Studio TIPS structed transjugular intrahepatic portosystemic shunt. Cardiovasc (G.I.S.T.). HEPATOLOGY 1998;27:48-53. Intervent Radiol 2002;25:365-372.

Prospective randomized trials comparing ePTFE-covered and bare endoprostheses are needed to clarify whether the covered endoprosthesis provides a survival benefit and to evaluate the reasons for that. These trials are underway but will require substantial more followup time for long-term survival analysis. In the meantime, we think that our data suggest strongly that ePTFE-covered endoprotheses should be used in cases in which TIPS implantation is considered. If our data, showing a survival benefit for patients with ePTFE-covered TIPS endoprostheses, are confirmed in randomized trials, the role of TIPS versus interventional or medical therapies would need serious reassessment in randomized trials comparing these therapies against the much-improved covered TIPS endoprosthesis.

References

1050

ANGERMAYR ET AL.

31. Haskal ZJ, Davis A, McAllisterA, Furth EE. PTFE-encapsulated endovascular stem-graft for transjugular intrahepatic portosystemic shunts: experimental evaluation. Radiology 1997;205:682-688. 32. Saxon RR, Timmermans HA, Uchida BT, Petersen BD, Benner KG, Rabkin J, Keller FS. Stent-grafts for revision of TIPS stenoses and occlusions: a clinical pilot study. J Vasc Interv Radiol 1997;8:539-548. 33. Sze DY, Vestring T, Liddell RP, Kato N, Semba CP, Razavi MK, Kee ST, et al. Recurrent TIPS failure associated with biliary fistulae: treatment with PTFE-covered stents. Cardiovasc Intervent Radiol 1999;22:298-304. 34. Haskal ZJ, Pentecost MJ, Soulen MC, Shlansky-Goldberg RD, Baum RA, Cope C. Transjugular intrahepatic portosystemic shunt stenosis and revision: early and midterm results. AJRAm J Roentgen01 1994;163:439-444. 35. Latimer J, Bawa SM, Rees CJ, Hudson M, Rose JD. Patency and reintervention rates during routine TIPSS surveillance. Cardiovasc Intervent Radiol 1998;21:234-239. 36. Stanley AJ, Jalan R, Forrest EH, Redhead DN, Hayes PC. Long-term follow up of transjugular intrahepatic portosystemic s e n t shunt (TIPSS) for the treatment ofpottal hypertension: results in 130 patients. Gut 1996; 39:479-485. 37. de Franchis R, Primignani M. Natural history of portal hypertension in patients with cirrhosis. Clin Liver Dis 2001;5:645-663. 38. Garcia-Tsao G. Current management of the complications of cirrhosis and portal hypertension: variceal hemorrhage, ascites, and spontaneous bacterial peritonitis. Gastroenterology 200 1;120:726-748. 39. Kramer L, Horl WH. Hepatorenal syndrome. Semin Nephrol 2002;22: 290-30 1.


40. Angermayr B, Cejna M, Karnel F, Gschwantler M, Koenig F, Pidlich J, Mendel H, et al. Child-Pugh versus MELD score in predicting survival in patients undergoing transjugular intraheparic portosystemic shunt. Gut 2003;52:879-885. 41. Kamath PS, Wiesner RH, Malinchoc M, Kremers W, Therneau TM, Kosberg CL, D’Amico G, et al. A model to predict survival in patients with end-stage liver disease. HEPATOLOGY 2001 ;33:464-470. 42. D’Amico G , Morabito A, Pagliaro L, Marubini E. Survival and prognostic indicators in compensated and decompensated cirrhosis. Dig Dis Sci 1986; 3 1:468-475. 43. Pagliaro L, D’Amico G, Pasta L. Portal hypertension in cirrhosis: natural history. In: Bosch J, Groszmann RJ, eds. Portal Hypertension: I’athophysiology and Treatment. Oxford; Blackwell Scientific Publications: 1994;7292. 44. Salerno F, Borroni G, Moser P, Badalamenti S, Cassara L., Maggi A, Fusini M, et al. Survival and prognostic factors of cirrhotic patients with ascites: a study of 134 outpatients. Am J Gastroenterol 1993;88:514-519. 45. Abraldes JG, Tarantino 1, Turnes J, Garcia-Pagan JC, Rodes J, Bosch J. Hemodynamic response to pharmacological treatment of portal hypertension and long-term prognosis of cirrhosis. HEPATOLOGY 2003;37:902908. 46. Bosch J. Medical treatment of portal hypertension. Digestion 1998;59: 547-5 5 5. 47. Feu F, Garcia-Pagan JC, Bosch J, Luca A, Teres J, Escorsell A, Rodes J. Relation between portal pressure response to pharmacotherapy and risk of recurrent variceal haemorrhage in patients with cirrhosis. Lancet 1995; 346:1056-1059.