A meta-analysis of transjugular intrahepatic portosystemic shunt versus paracentesis for refractory ascites

Share Embed

Descrição do Produto

Journal of Hepatology 43 (2005) 990–996 www.elsevier.com/locate/jhep

A meta-analysis of transjugular intrahepatic portosystemic shunt versus paracentesis for refractory ascites Agustı´n Albillos1,†,*, Rafael Ban˜ares2,†, Mo´nica Gonza´lez1, Marı´a-Vega Catalina2, Luis-Miguel Molinero3 2

1 Servicio de Gastroenterologı´a, Hospital Universitario Ramo´n y Cajal, Departamento de Medicina, Universidad de Alcala´, Madrid, Spain Servicio de Gastroenterologı´a, Hospital Universitario General Gregorio Maran˜o´n, Departamento de Medicina, Universidad Complutense, Madrid, Spain 3 Alce Ingenierı´a, Madrid, Spain

See Editorial, pages 924–925 Background/Aims: Meta-analysis designed to provide evidence-based guidance on the effect of TIPS and paracentesis on mortality and encephalopathy in cirrhotic patients with refractory ascites. Methods: Five randomized trials published between 1989 and 2005 were identified. Results: The five trials involved 330 patients, and none included patients O76 years, with bilirubin O5–10 mg/dl or creatinine O3 mg/dl. Ascites recurrence was lower in the TIPS arm (RR 0.56; 95% CI 0.47–0.66). TIPS was associated with a greater risk of encephalopathy (RR 1.36; 95% CI 1.1–1.68) and severe encephalopathy (RR 1.72; 95% CI 1.14– 2.58). TIPS did not affect mortality, as estimated by the RR (0.93; 95% CI 0.67–1.28, random effect model) and pooled hazard ratio (RR 1.09; 95% CI 0.84–1.88). Analysis of this outcome measure was limited by significant heterogeneity among trials. Liver-related mortality was homogenous and similar in both arms. Results were unaffected by excluding trials of lower quality or with a greater number of alcoholics. Meta-analysis of trials including patients with recidivant ascites revealed a lower mortality in the TIPS arm (RR 0.68; 95% CI 0.49–0.93). Conclusions: In patients with refractory ascites, a better control of ascites by TIPS does not translate into improved survival and worsens encephalopathy. q 2005 European Association for the Study of the Liver. Published by Elsevier B.V. All rights reserved. Keywords: Cirrhosis; Encephalopathy; Pooled hazard ratio; Inconsistency index; Systematic review 1. Introduction Refractory ascites complicates advanced cirrhosis in 5–10% of patients, and is associated with a high rate of hepatorenal syndrome and poor survival [1–3]. Largevolume paracentesis, regarded as the first-line treatment for refractory ascites [4], does not prevent ascites recurrence.

Received 20 February 2005; received in revised form 9 May 2005; accepted 7 June 2005; available online 5 July 2005 * Corresponding author. Tel.: C34 9 188 54870; fax: C34 9 188 545 26. E-mail address: [email protected] (A. Albillos). Abbreviations: TIPS, transjugular intrahepatic portosystemic shunt; RR, pooled relative risk; CI, confidence interval; NNT, number needed to be treated. † Both the authors share the first authorship of this article.

Transjugular intrahepatic portosystemic shunt (TIPS) has been proposed as an alternative to paracentesis [5–7]. The TIPS reduces portal pressure and ascites recurrence rate [8], but its use is related to the potential development of encephalopathy, and worsening of liver function with diminished survival [9]. Five randomized controlled trials have compared TIPS and paracentesis in the management of refractory ascites [10–14]. These trials indicate that TIPS lowers the ascites recurrence rate yet they differ in the impact of TIPS on survival and encephalopathy risk. In the first of these trials, performed in 1996 on a small population of patients, TIPS increased encephalopathy and mortality [10]. Of the other trials, one revealed no difference in encephalopathy and a trend towards improved survival

0168-8278/$30.00 q 2005 European Association for the Study of the Liver. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.jhep.2005.06.005

A. Albillos et al. / Journal of Hepatology 43 (2005) 990–996

using TIPS [11], whereas two indicated increased encephalopathy with TIPS and no difference in survival [12,13]. The fifth study attributed TIPS a greater risk of severe encephalopathy and improved survival [14]. Herein, we report a meta-analysis of data derived from the published literature on the use of TIPS and largevolume paracentesis in the treatment of cirrhotic patients with refractory ascites. We aimed at providing evidencebased guidance regarding the effects of these therapies on survival and encephalopathy risk, and to find plausible explanations for the variable results among trials.

2. Materials and methods Our meta-analysis was conducted according to Quorum statements [15].

2.1. Searching and selection of studies English-language studies were identified by a comprehensive search of MEDLINE and EMBASE (1989–2005), and the controlled trial Cochrane Registry. A keyword search for ascites adjacent to refractory or resistant was done. This search was supplemented by combinations of truncated keywords, such as random, random allocation, clinical trial, and multicenter study and was limited to English-language studies in humans. Abstracts from the meetings of the American Gastroenterological Association, the American Association for the Study of Liver Disease, the European Association for the Study of the Liver and the British Society of Gastroenterology were revised. The following selection criteria were applied: (1) study design: randomized controlled trial; (2) study population: cirrhotic patients with refractory ascites; (3) intervention: TIPS; (4) comparison intervention: repeated large-volume paracentesis; and (5) one or more of the following outcome measures: recurrence of ascites, encephalopathy, and mortality.

2.2. Data abstraction Trials considered for inclusion were analyzed independently by two observers (AA, RB), and disagreement was resolved by consensus among the team members.

2.3. Qualitative assessment The qualitative assessment of trials is controversial [16–18]. We decided a priori on five criteria to assess the methodological quality of each trial: (1) allocation sequence generated by computer random numbers, (2) allocation concealment, (3) baseline similarity of treatment and control groups, (4) rigorous analysis by the intention-to-treat principle, and (5) adequate potency of the treatments to detect a treatment-group difference in mortality. Baseline equivalency was judged to exist if none of the baseline variables differed by O25% among treatment groups. Each criterion was rated as present (1 point) or absent/not described (0 points) except for criterion 2, which was worth 3 points if randomization was central and 1 point if conducted using sealed, opaque envelopes. Any trial that satisfied at least four of the five criteria was judged as high quality. Double-blinding was not included among the criteria since the intervention tested is difficult to blind.

2.4. Quantitative data synthesis Three primary outcomes were evaluated: (1) recurrence of ascites, defined as ascites requiring paracentesis, (2) mortality, defined as any death occurring during follow-up or the need for liver transplantation, and (3) hepatic encephalopathy, defined as any episode of clinically overt encephalopathy.


The effect of treatment on the defined outcome measures was calculated from the raw data using the fixed and random effect models (Der Simonian and Laird) [19,20]. Differences observed between the treatment groups were expressed as the pooled relative risk (RR) with 95% confidence interval (CI). For survival analysis, an indirect log hazard ratio and variance estimation was calculated using the log-rank test result and the total number of events observed in the two groups, and expressed as the pooled hazard ratio [21]. Statistical heterogeneity between trials was evaluated by the Cochran x2 test and was considered to exist when P!0.1. In the absence of statistically significant heterogeneity, only the RR by the fixed model is given in the results. We calculated the I2 index to measure the degree of inconsistency in the studies’ results [22]. The number needed to be treated (NNT) or to be harmed, respectively, was calculated as the inverse of the absolute risk difference between the treatment groups. We identified potential differences in treatments across the studies through sensitivity tests. The influence of the quality of the trials, and the etiology of liver disease were analyzed by excluding those trials with a quality score !4 or with a large proportion of alcoholics (O40%), respectively. The influence of albumin infusion after paracentesis was analyzed by excluding trials in which albumin was not routinely infused. Patients with recidivant ascites do not have such a poor prognosis as those with refractory ascites, and the sensitivity analysis evaluated the trials by inclusion of patients with recidivant ascites.

3. Results 3.1. Trial flow Our bibliographic searches yielded 23 references. Of these, five randomized, controlled trials published from October 1989 to January 2005 met the inclusion criteria. The reasons for excluding articles (followed by the number excluded) were: reference not related to the study objective (3), reference not a randomized trial (6) or reference did not include paracentesis or TIPS as the interventions (9). 3.2. Description of the selected trials Table 1 summarizes the details of the five included trials. Four out of the five trials were multicenter studies. The baseline condition in all the trials was ascites of difficult control (refractory or recidivant). In the five trials, refractory ascites was defined according to the International Ascites Club criteria [1]. Recidivant ascites was defined as tense ascites that recurred on at least three occasions within 12 months despite standard treatment in the studies by Ro¨ssle and Salerno. Patients with recidivant ascites accounted for 45 and 32% of the total number of patients included in the Ro¨ssle and Salerno trials, respectively. Patients O70–75 years were excluded in three trials [10,12,14], but no trial included any patient O76 years. All studies excluded patients with encephalopathy Ograde 2, portal vein thrombosis, hepatocellular carcinoma or parenchymal renal disease. Except the trial by Lebrec, the other four studies excluded patients with terminal liver disease (established by a serum bilirubin O5–10 mg/dl and/ or Child-Pugh O11). Trials had two arms: TIPS or paracentesis. Patients allocated to the TIPS arm received a non-covered stent. In all studies, follow-up Doppler ultrasonography was used


A. Albillos et al. / Journal of Hepatology 43 (2005) 990–996

Table 1 Design characteristics of the selected trials Study, year [reference]

Type of publication


Study population

Exclusion criteria

TIPS intervention

Paracentesis intervention

End Points (primary/ secondary)

Lebrec, 1996 [10]

Peerreviewed article

Single-center, randomized controlled

Cirrhosis with refractory ascites

O70 years, HERgrade II, portal vein thrombosis, HCC, active bacterial infection, alcoholic hepatitis, biliary obstruction, creatinine O1.7 mg/dl, severe non-hepatic disease, pulmonary hypertension

Balloon expandable Palmaz stent

Recurrence of ascites Mortality

Ro¨ssle, 2000 [11]

Peerreviewed article

Multi-center, randomized controlled

Cirrhosis with refractory or recidivant (45% of the patients) ascites

HERgrade II, bilirubin O5 mg/dl, portal-vein thrombosis, advanced HCC, failed paracentesis, creatinine O3 mg/dl

Balloon-Palmaz (21 patients) or self-expandable nitinol (8 patients) stent

Gine´s, 2002 [12]

Peerreviewed article

Multi-center, randomized controlled

Cirrhosis with refractory ascites

Balloon expandable Wallstent

Sanyal, 2003 [13]

Peerreviewed article

Multi-center, randomized controlled

Cirrhosis with refractory ascites

!18 or O75 years, HERgrade II, portal-vein thrombosis, bilirubin O10 mg/dl, HCC, prothrombin time !40%, platelets !40000/mm3, active bacterial infection, creatinine O3 mg/dl, parenchymal renal disease, cardiac failure HERgrade II, portal-vein thrombosis, bilirubin O5 mg/dl, HCC, active bacterial infection, alcoholic hepatitis, GI bleeding within 6 weeks, parenchymal renal disease, cardiac or respiratory failure, pulmonary hypertension

Largevolume paracentesis Albumin infusion if creatinine clearance !60 ml/min Largevolume paracentesis Albumin infusion ‘when clinically indicated’ Largevolume paracentesis followed by albumin infusion

Salerno, 2004 [14]

Peerreviewed article

Multi-center, randomized controlled

Cirrhosis with refractory or recidivant (32% of the patients) ascites

O72 years, HE Rgrade II, portal vein thrombosis, bilirubin O6 mg/dl, ChildPugh O11, advanced HCC, active bacterial infection, creatinine O3 mg/dl, cardiac or pulmonary failure, GI bleeding within 2 weeks

Not stated

Largevolume paracentesis followed of albumin infusion

Balloon expandable Wallstent or Memotherm stent

Largevolume paracentesis followed of albumin infusion

Mortality/ recurrence of ascites Other complications of cirrhosis

Mortality/ recurrence of ascites Other complications of cirrhosis Cost Recurrence of ascites mortality/ other complications of cirrhosis quality of life Mortality/ recurrence of ascites, other complications of cirrhosis, need for rehospitalization

HE, hepatic encephalopathy; HCC, hepatocellular carcinoma. Other complications of cirrhosis include encephalopathy, variceal bleeding, hepatorenal syndrome.

to assess shunt patency. Paracentesis was systematically followed by albumin infusion in the studies by Gine´s, Sanyal and Salerno. Albumin was infused in the study by Lebrec if creatinine clearance was below 60 ml/min, and in the study by Ro¨ssle ‘when clinically indicated’. Two studies allowed treatment failures to switch to the other treatment [11,14]. The number of subjects ranged from 25 to 109, totaling 330 (Table 2). The proportion of Child C patients was 34% (range 22–43%) in three trials [10–12], and of 76% in the trial of Salerno. The proportion of patients with alcoholic cirrhosis was above 70% in two studies [10,11]. Recurrence of ascites and mortality were assessed in the five trials. The criteria used to define ascites recurrence were fairly similar across the studies: tense, symptomatic ascites

requiring paracentesis. Hepatic and non-hepatic causes of death were reported in four studies [10,11,13,14]. The incidence of any degree of encephalopathy was reported in the five studies, and that of severe encephalopathy (grade III or IV) in four [10,12–14]. The number of patients with variceal hemorrhage was reported in three trials [12–14]. The incidence of hepatorenal syndrome was only reported in the Gine´s and Salerno trials. The way in which worsening hepatic function during follow-up was reported (change in mean serum bilirubin or in Child-Pugh score) precluded the pooling of data. The five trials reviewed concluded that TIPS was more effective than paracentesis for ascites control. Mortality was higher in the TIPS arm in the Lebrec trial, and in the paracentesis arm in the Salerno trial, whereas it was

A. Albillos et al. / Journal of Hepatology 43 (2005) 990–996


Table 2 Characteristics of the patients in the selected trials Study, year [reference]

Number of patients

Mean age (year)

Patients with alcoholic cirrhosis (%)

Patients with Child C (%)

Mean Child-Pugh score (points)

Lebrec, 1996 [10] Ro¨ssle, 2000 [11] Gine´s, 2002 [12] Sanyal, 2003 [13] Salerno, 2004 [14]

25 (13/12) 60 (29/31) 70 (35/35) 109 (52/57) 66 (33/33)

50/52 58/61 59/56 56/52 58/60

77/83 83/75 51/60 62/58 45/39

31/33 38/22 37/43 – 79/73

9.3/9.2 9.1/8.7 9.3/9.2 9.2/9.3 9.4/9.4

Values indicate TIPS/paracentesis group. Table 3 Quality score of the selected trials Study, year [reference]

Generation of allocation sequence

Allocation concealment

Baseline equivalency of groups

Analysis by intention-to-treat

Potency of intervention

Quality score

Lebrec, 1996 [10] Ro¨ssle, 2000 [11] Gine´s, 2002 [12] Sanyal, 2003 [13] Salerno, 2004 [14]

0 0 0 0 0

1 1 3 3 3

1 1 1 1 1

1 1 1 1 1

0 1 1 1 1

3 4 6 6 6

identical in both arms in the Sanyal and Gine´s trials. In the trial of Ro¨ssle no significant differences in survival were observed between treatments through univariate analysis, but assignment to TIPS was a predictor of survival in their multivariate analysis. The incidence of encephalopathy was greater in the TIPS group in three of the studies [10,12,13], and similar in both therapeutic arms in the remaining two [11,14]. 3.3. Quality assessment of the selected trials Our quality scores for the five trials ranged from 3 to 6, out of a highest score of 7 (Table 3). None of the trials was investigator-blinded. Randomization was centralized in the trials of Gine´s, Sanyal and Salerno. The study by Lebrec was a small trial in which the sample size was not calculated to estimate a change in survival, neither potential dropouts. The therapeutic arms within each trial were comparable at baseline, and analyzed by intention-to-treat.

3.4. Quantitative assessment of the selected trials TIPS insertion was unsuccessful in three patients in each of the studies by Lebrec, Sanyal, and Salerno, and in one patient in the Gine´s trial. Different protocols were used to detect stenosis, but stent-assisted patency at 1 year was O90% in the surviving patients of the five reports. Sixtyseven percent and 58% of the patients in the Ro¨ssle and Salerno trials, respectively, failing to respond to paracentesis received a TIPS as rescue treatment. Ascites recurred in a lower proportion of patients treated with TIPS (42 vs. 80%) (Table 4). Values for this variable showed no statistically significant heterogeneity. The RR of achieving ascites control by TIPS insertion was 0.56 (95% CI 0.47–0.66), and the NNT was 3 (95% CI 2.3–4.2) (Fig. 1). Mortality rates during follow-up were 46 and 50% for patients treated with TIPS and paracentesis, respectively. The P value for the heterogeneity test performed on this outcome was significant (PZ0.09). The percentage total

Table 4 Meta-analysis of TIPS vs. paracentesis in the treatment of patients with refractory ascites Outcome measure

No. of trials

No. of patients



Ascites recurrence Mortality Liver-related mortality Encephalopathy Severe encephalopathy




5 4

330 260

5 4

330 270

a b


P-value for homogeneity

Inconsistency I2 (CI)


0.56 (0.47–0.66)


0 (0–54)

0.46 0.31

0.50 0.51

0.90 (0.72–1.12) 0.75 (0.53–1.04)

0.09b 0.22

48 (0–81) 32 (0–75)

0.54 0.38

0.36 0.21

1.36 (1.1–1.68) 1.72 (1.14–2.58)

0.48 0.79

0 (0–76) 0 (0–55)

Proportion of the outcome measure. RR (95% CI), 0.93 (0.67–1.28) using the random effect model.


A. Albillos et al. / Journal of Hepatology 43 (2005) 990–996

Recurrence of ascites



Lebrec, 1996 Rössle, 2000 Ginés, 2002 Sanyal, 2003 Salerno, 2004 Pooled RR fixed effects Pooled RR random effects Lower with TIPS





Lower with Lower with paracentesis TIPS


Lower with paracentesis

Lower with TIPS


20 40



Lower with paracentesis






Fig. 1. Recurrence rates for ascites, encephalopathy and mortality in trials comparing TIPS and large-volume paracentesis in cirrhotic patients with refractory ascites. Relative risks and 95% confidence intervals are represented.

variation across studies due to heterogeneity (I2) was 48% (95% CI 0–81). The RR of dying related to TIPS was 0.93 (95% CI 0.67–1.28, using the random-effect model), and the pooled hazard ratio was 1.09 (95% CI 0.84–1.88) (Fig. 1). Hepatic causes of death were of 31 and 51% in the TIPS and paracentesis arms, respectively (Tables 4 and 5). The RR of mortality attributed to liver disease in TIPS patients was 0.75 (95% CI 0.53–1.04, using the fixed-effect model; heterogeneity test, PZ0.22). The risk of developing any degree of encephalopathy was greater in TIPS patients (54 vs. 36%, RR 1.36, 95% CI 1.1–1.68, test for heterogeneity, PZ0.48); the number needed to be harmed being 6 (95% CI 1.8–19.2) (Fig. 1). Severe encephalopathy, assessed in four of the studies (270 patients), was also more frequent in the TIPS arm (38 vs. 21%, RR 1.72, 95% CI 1.14–2.58; heterogeneity test, PZ 0.79) (10,12–14). The number of patients with variceal

bleeding was similar in both arms (7 vs. 10%; RR 0.8, 95% CI 0.41–1.56), as reported in three of the trials including 245 patients [12–14]. As reported in two of the trials including 136 patients, the number of patients that developed hepatorenal syndrome was significantly lower in the TIPS arm (9 vs. 24%; RR 0.38, 95% CI 0.16–0.94) [12,14]. Subgroup analyses were performed in relation to the quality of the trials, the proportion of alcoholic patients, the albumin infusion policy, and the presence of patients with recidivant ascites. All tests for homogeneity were non-significant. The exclusion of the trials by Lebrec and Ro¨ssle, which had quality scores of 3 and 4, respectively, included O70% of alcoholics and did not routinely infuse albumin after paracentesis, had no effect on the mortality rates (RR 0.93, 95% CI 0.70–1.22), encephalopathy (RR 1.27, 95% CI 1–1.61) or ascites recurrence (RR 0.52, 95% CI 0.41–0.65). Meta-analyses of the trials

Table 5 Causes of death Causes of death

Hepatic Hepatic failure Massive variceal bleeding Hepatorenal syndrome type I Hepatocellular carcinoma Non-hepatic Sepsis Cerebrovascular accident Cardiac dysfunction Other diseases/ unknown

Ro¨ssle [11]

Lebrec [10]

Sanyal [13]

Salerno [14]

TIPS (nZ13)

Paracentesis (nZ12)

TIPS (nZ29)

Paracentesis (nZ31)

TIPS (nZ52)

Paracentesis (nZ57)

TIPS (nZ33)

Paracentesis (nZ33)

6 4

4 1

10 9

20 16

20 16

13 9









3 2

4 1




3 3












The study by Gine´s et al. does not specify the causes of death.

5 4 1

A. Albillos et al. / Journal of Hepatology 43 (2005) 990–996

by Rossle and Salerno that included patients with recidivant ascites showed lower mortality in the TIPS arm (45 vs. 67%; RR 0.68, 95% CI 0.49–0.93). Mortality was similar in both therapeutic arms in the three other trials (47 vs. 39%; RR 1.17, 95% CI 0.86–1.6).

4. Discussion This meta-analysis of five randomized controlled trials comparing TIPS and paracentesis in the treatment of refractory ascites indicates that the improved ability of TIPS to control ascites does not modify survival, yet occurs at the expense of a greater incidence of encephalopathy. The outcome measures selected were homogeneously defined and could be examined in the five studies. TIPS reduced the overall ascites recurrence and encephalopathy rates by approximately 50 and 40%, respectively. One would need to treat three patients with TIPS instead of paracentesis to prevent one episode of ascites, but encephalopathy develops in every six patients undergoing TIPS. The impact of TIPS on the survival of patients with refractory ascites has been controversial [10–14,23,24]. We could expect both a detrimental effect, given that shunting of portal blood can lead to liver failure, and a beneficial effect, since TIPS can theoretically prevent other complications of portal hypertension. The similar RR of death in both arms determined using the fixed and random effect models and the pooled hazard ratio of death—the summary estimate in our meta-analysis—do not support the hypothesis that TIPS negatively affects the survival of cirrhotic patients with refractory ascites. Noteworthy, the only trial indicating a beneficial effect of TIPS on survival included the largest proportion of Child C patients [14]. Neither was TIPS associated with increased liver-related mortality, an outcome assessed in four of the five trials [10,11,13,14]. The lower incidence of hepatorenal syndrome in the TIPS arm, an outcome that could be analyzed in two trials, might have contributed to the lower mortality in TIPS patients. The combined results of the mortality estimate were heterogeneous, and roughly half the variability in effect estimates could be attributed to genuine variation across studies, as assessed by the I2 index [22]. The heterogeneity of the trials limits drawing firm conclusions regarding this outcome measure. There are several possible sources of heterogeneity among the studies: (i) differences in the study populations. First, the proportion of Child C patients was about 30% in three trials [10,11,12], and 70% in the most recently published study [14]. Second, two of the five trials included patients with recidivant ascites. This group of patients with ascites of difficult control does not show such a dismal prognosis than patients with refractory ascites [1]. The distinctive nature of patients with recidivant ascites is supported by the results of the sensitivity analysis, which revealed a favorable effect of TIPS on survival in the two trials including patients with this condition. Future studies


should address the impact of TIPS in this population. Third, patients with refractory ascites do not form a homogeneous population despite most having a poor prognosis; in some, sodium retention may even spontaneously improve leading to their survival beyond 6–12 months with no intervention. Variables with an impact on the outcome of patients with refractory ascites should be considered in future trials on this topic [25]. (ii) Differences in the design. Two of the trials allowed the use of TIPS as rescue treatment during follow-up in patients not responding to paracentesis [11,14]. This could affect analysis of the effects of treatment on mortality. (iii) Differences in interventions. The practice of albumin infusion after paracentesis varied among reports. In two of the studies [10,11], albumin was not routinely infused; this policy being a potential cause of postparacentesis dysfunction and increased mortality [26]. (iv) Differences in population size. The much higher mortality reported by Lebrec might be the result of the small number of patients enrolled in their study (only four were Child C), and a high rate of failure of TIPS insertion (3 out of 13 patients) [10]. The impact of these differences in the metaanalysis results could not be demonstrated by our sensitivity analysis, because of the low number of patients included in each trial and the relatively low weight of each individual trial. Thus, results were unchanged despite excluding discrepant trials. The methodological limitations of the trials warrant some discussion. First, no study was double-blinded. Although it is true that the tested intervention is difficult to blind, it is unlikely that the lack of blinding could affect the outcomes assessed (i.e. mortality) [18]. Secondly, none of the trials described the method used to generate the allocation sequence, and the trial by Ro¨ssle failed to describe the concealment of the allocation sequence. Despite these potential sources of bias, randomization was adequate in the five trials as shown by the baseline equivalency of treatment groups. The subgroup analysis comparing the trials according to their quality scores did not modify the outcome estimates. A third limitation was that only one provided raw data by Child class [12], which precluded the subgroup analysis of this variable according to liver function. Combining individual patient-level data could be useful for subgroup analysis on the basis of the degree of liver failure and establishing time-to-event outcomes. Fourthly, since a given intervention (i.e. TIPS) on the one hand might alleviate symptoms (i.e. ascites), yet on the other, exacerbate other symptoms (i.e. encephalopathy) or lead to an increase in costs, future trials should assess health-related quality of life and care costs. These variables could not be addressed in the present metaanalysis, since quality of life and cost were each evaluated in only one trial [12,13]. According to this meta-analysis, the improved ability of TIPS to control ascites recurrence is associated with a greater risk of encephalopathy. Shunting of portal blood by TIPS does not increase overall or liver-related mortality.


A. Albillos et al. / Journal of Hepatology 43 (2005) 990–996

The sources of heterogeneity among studies were differences in the patients included, the design and size of the trials, and the policy of albumin administration in the paracentesis arm.



Acknowledgements We are indebted to Ana Burton for editorial assistance. This study was supported by grants from the Ministerio de Ciencia y Tecnologı´a (BFI 2003-03858), and from the Ministerio de Sanidad y Consumo, Instituto de Salud Carlos III (C03/02).

References [1] Arroyo V, Gine´s P, Gerbes AL, Dudley FJ, Gentilini P, Laffi G, et al. Definition and diagnostic criteria of refractory ascites and hepatorenal syndrome in cirrhosis. International Ascites Club. Hepatology 1996; 23:164–176. [2] Gatta A, Angeli P, Caregaro L, Menon F, Sacerdoti D, Merkel C. A pathophysiological interpretation of unresponsiveness to spironolactone in a stepped-care approach to the diuretic treatment of ascites in nonazotemic cirrhotic patients. Hepatology 1991;14:231–236. [3] Arroyo V, Colmenero J. Ascites and hepatorenal syndrome in cirrhosis: pathophysiological basis of therapy and current management. J Hepatol 2003;38:S69–S89. [4] Garcia-Tsao G. Current management of the complications of cirrhosis and portal hypertension: variceal hemorrhage, ascites, and spontaneous bacterial peritonitis. Gastroenterology 2001;120:726–748. [5] Ochs A, Rossle M, Haag K, Hauenstein KH, Deibert P, Siegerstetter V, et al. The transjugular intrahepatic portosystemic stent-shunt procedure for refractory ascites. N Engl J Med 1995;332: 1192–1197. [6] Quiroga J, Sangro B, Nun˜ez M, Bilbao I, Longo J, Garcia-Villarreal L, et al. Transjugular intrahepatic portal-systemic shunt in the treatment of refractory ascites: effect on clinical, renal, humoral, and hemodynamic parameters. Hepatology 1995;21:986–994. [7] Forrest EH, Stanley AJ, Redhead DN, McGilchrist AJ, Hayes PC. Clinical response after transjugular intrahepatic portosystemic stent shunt insertion for refractory ascites in cirrhosis. Aliment Pharmacol Ther 1996;10:801–806. [8] Wong F, Sniderman K, Liu P, Allidina Y, Sherman M, Blendis L. Transjugular intrahepatic portosystemic stent shunt: effects on hemodynamics and sodium hemostasis in cirrhosis and refractory ascites. Ann Intern Med 1995;122:816–822. [9] Boyer TD. Transjugular intrahepatic portosystemic shunt: current status. Gastroenterology 2003;124:1700–1710. [10] Lebrec D, Giuily N, Hadengue A, Vilgrain V, Moreau R, Poynard T, et al. Transjugular intrahepatic portosystemic shunts: comparison with






[18] [19]

[20] [21]

[22] [23]




paracentesis in patients with cirrhosis and refractory ascites: a randomized trial. J Hepatol 1996;25:135–144. Rossle M, Ochs A, Gulberg V, Siegerstetter V, Holl J, Deibert P, et al. A comparison of paracentesis and transjugular intrahepatic portosystemic shunting in patients with ascites. N Engl J Med 2000;342: 1701–1707. Gines P, Uriz J, Calahorra B, Garcia-Tsao G, Kamath PS, Del Arbol L R, et al. Transjugular intrahepatic portosystemic shunting versus paracentesis plus albumin for refractory ascites in cirrhosis. Gastroenterology 2002;123:1839–1847. Sanyal AJ, Genning C, Reddy KR, Wong F, Kowdley KV, Benner K, et al. The North American study for the Treatment of Refractory Ascites. Gastroenterology 2003;124:634–641. Salerno F, Merli M, Riggio O, Cazzaniga M, Valeriano V, Pozzi M, et al. Randomized controlled study of TIPS versus paracentesis plus albumin in cirrhosis with severe ascites. Hepatology 2004;40: 629–635. Moher D, Cook DJ, Eastwood S, Olkin I, Rennie D, Stroup DF. 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. Moher D, Pham B, Jones A, Cook DJ, Jadad AR, Moher M, et al. Does quality of reports of randomised trials affect estimates of intervention efficacy reported in meta-analyses? Lancet 1998;352:609–613. Kjaergard LL, Villumsen J, Gluud C. Reported methodologic quality and discrepancies between large and small randomized trials in metaanalyses. Ann Intern Med 2001;135:982–989. Juni P, Altman DG, Egger M. Assessing the quality of controlled clinical trials. Br Med J 2001;323:42–46. Robbins J, Brenslow N, Greenland S. Estimators of the Mantel– Haenszel variance consistent in both sparse data and large strata limiting models. Biometrics 1986;42:311–323. Der Simonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986;7:177–188. Parmar MK, Torri V, Stewart L. Extracting summary statistics to perform meta-analyses of the published literature for survival endpoints. Stat Med 1998;17:2815–2834. Higgins J, Thompson SG. Quantifying heterogeneity in a metaanalysis. Stat Med 2002;21:1539–1558. Russo MW, Sood A, Jacobson IM, Brown RS. Transjugular intrahepatic portosystemic shunt for refractory ascites: an analysis of the literature on efficacy, morbidity, and mortality. Am J Gastroenterol 2003;98:2521–2527. Saab S, Nieto JM, Ly D, Runyon BA. TIPS versus paracentesis for cirrhotic patients with refractory ascites. Cochrane Database Syst Rev 2004;CD004889. Moreau R, Delegue P, Pessione F, Hillaire S, Durand F, Lebrec D, et al. Clinical characteristics and outcome of patients with cirrhosis and refractory ascites. Liver Int 2004;24:457–464. Gines A, Fernandez-Esparrach G, Monescillo A, Vila C, Dome´nech E, Abecasis R, et al. Randomized trial comparing albumin, dextran 70, and polygeline in cirrhotic patients with ascites treated by paracentesis. Gastroenterology 1996;111:1002–1010.

Lihat lebih banyak...


Copyright © 2017 DADOSPDF Inc.