Prognostic factors for peritoneovenous shunt placement for refractory ascites in liver cirrhosis

Abstract

BACKGROUND

Refractory ascites severely impairs quality of life in patients with liver cirrhosis (LC) and cancer-related peritonitis. For individuals who are intolerant to medical therapy and require frequent large-volume paracentesis, a peritoneovenous shunt (PVS) offers a potential treatment option. However, PVS placement is associated with high complication rates, perioperative mortality, and lacks well-defined indications.

AIM

To identify prognostic factors for PVS placement and develop a novel postoperative survival scoring model for LC with refractory ascites.

METHODS

A total of 100 patients who underwent PVS placement for refractory ascites due to LC in our department between 1998 and 2024 were analyzed. Patients were stratified into two groups: Those who survived more than 180 days after PVS placement (L-group) and those who survived for less than 180 days (S-group). Prognostic factors were compared between groups, and four variables (sex, age, Child-Pugh score, and liver volume) were selected for the creation of a new scoring system.

RESULTS

Significant differences between the S- and L-groups were observed for age, sex, Child-Pugh score and preoperative liver volume. Based on these variables, we developed a scoring system as follows: 1 point each for age ≥ 60 years, Child-Pugh score ≥ 10, female sex, and preoperative liver volume < 1057 mL. Patients scoring 0-2 points were classified as PVS grade A, and those scoring 3-4 points as PVS grade B. Survival analysis showed that overall survival was significantly higher in PVS grade A compared with PVS grade B. Multivariate analysis confirmed PVS grade as an independent prognostic factor.

CONCLUSION

The proposed PVS scoring system may be a useful tool for predicting postoperative prognosis following PVS placement in patients with LC and refractory ascites.

Key Words: Peritoneovenous shunt; Liver cirrhosis; Refractory ascites; Prognostic factor; Scoring system

Core Tip: This study culminated in the proposal of a novel scoring system to predict prognosis after peritoneovenous shunt placement for refractory ascites in patients with liver cirrhosis. In an analysis of 100 cases, four prognostic factors were identified: Age ≥ 60 years; Child-Pugh score ≥ 10; female sex; and liver volume < 1057 mL. Patients were stratified into peritoneovenous shunt grade A (0-2 points) and grade B (3-4 points), with significantly better survival observed in grade A. This scoring system may support clinical decision-making and improve patient outcomes.

INTRODUCTION

Refractory ascites is a serious complication of liver cirrhosis (LC), associated with impaired quality of life and poor prognosis[1]. Patients experience daily activity limitations and significant distress from abdominal distension. Ascites develops through multiple interrelated mechanisms involving visceral and systemic hemodynamics, as well as dysfunction of the liver, kidneys, and heart[2]. Vasodilation leads to reduced effective circulating volume, decreased renal blood flow, and activation of the renin-angiotensin system, further promoting sodium and water retention and ascites formation[3].

Current treatment options include large-volume paracentesis, transjugular intrahepatic portosystemic shunt (commonly known as TIPS), automated low-flow ascites pump (Alfapump^®; Sequana Medical, Zürich, Switzerland) and peritoneovenous shunt (PVS)[4]. The PVS (LeVeen or Denver), introduced in the 1970s, diverts ascitic fluid into the central venous system based on the peritoneal-venous pressure gradient[3,5]. While PVSs can provide permanent symptomatic relief and preserve patient lifestyle, they carry substantial risks, including disseminated intravascular coagulation, sepsis, and heart failure[6-8].

Although PVS is primarily a palliative intervention, careful patient selection is essential given its high complication rate. Notably, PVS placement has been reported to mitigate sarcopenia in patients with refractory ascites[9], and early intervention may improve outcomes and quality of life. Thus, patient selection remains the most critical step in management. However, no standardized criteria nor guidelines currently exist. The aim of this study was to identify prognostic indicators from perioperative data and to establish a preoperative model for predicting prognosis after PVS placement.

MATERIALS AND METHODS

Patients

Between 1998 and 2024, 100 patients underwent PVS placement for refractory ascites due to LC at our institution. Patients who died within 180 days after PVS placement were classified into the short-term survival group (S-group), while those who survived longer were classified into the long-term survival group (L-group). Prognostic factors, including liver volume, were compared between the groups. Based on prior literature[4], clinical relevance, and dataset availability, the four variables of sex, age, Child-Pugh score, and liver volume were selected as prognostic factors. Cutoff values for each variable were determined by receiver operating characteristic analysis. Using these factors, we developed a scoring system and stratified patients into two categories: PVS grade A (low score) and PVS grade B (high score). Long-term survival and multivariate analyses were performed to assess prognostic validity. Concomitant hepatocellular carcinoma was present in 43 patients (43%). No cases were excluded from PVS placement on the basis of malignant tumors. Data collection and analysis adhered to institutional guidelines and the ethical standards of the Declaration of Helsinki. Ethical approval was obtained from the Nippon Medical School Hospital Ethics Committee[10].

Imaging and volume measurements

Liver and spleen volumes before and after PVS placement were measured using the SyNAPSE VINCENT analyzer (Fujifilm Medical Co., Tokyo, Japan)[11,12]. Plain or delayed-phase computed tomography images reconstructed at 5-mm intervals were analyzed. Images were imported into the analyzer, the liver region was delineated automatically or manually, and volumes were calculated.

Indication for PVS placement

At our institution, the indication for PVS placement is ascites refractory to medical therapy, including diuretics and albumin, requiring repeated abdominal paracentesis. The primary contraindication is bacterial peritonitis, defined by the presence of endotoxin or bacteria in ascitic fluid, or an ascitic white blood cell count ≥ 500/μL. Patients with concurrent heart failure, pulmonary edema, markedly impaired liver function (Child-Pugh score ≥ 14), severe renal dysfunction, or poor performance status (≥ 2) were considered high risk and required caution (Table 1). Within this cohort, 1 patient had a Child-Pugh score of 14 and 3 had severe renal dysfunction requiring dialysis. No cases of heart failure, pulmonary edema, nor poor performance status were observed.

Table 1 Indications for peritoneovenous shunt placement.

Indication for PVS placement
Ascites refractory to medical treatment
Multiple CART or multiple abdominal paracenteses performed
Contraindication
Presence of bacterial peritonitis
Positive findings of endotoxin and bacteria in ascitic fluid
White blood cell counts in ascitic fluid of 500/μL or higher
Patients requiring caution
Patients with concurrent heart failure or pulmonary edema
Patients with significantly impaired liver function (Child-Pugh score of 14 or higher)
Patients with severe renal dysfunction
PS 2 or higher

PVS: Peritoneovenous shunt; CART: Concentrated ascites reinfusion therapy; PS: Performance status.

PVS placement

All procedures were performed under local anesthesia using 1% lidocaine, using equipment supplied in the PVS placement kit (BD, Franklin Lakes, NJ, United States). Abdominal ultrasonography was first performed to evaluate ascites and confirm a safe paracentesis site. Under real-time ultrasound guidance, the right internal jugular vein was accessed with a 21-gauge needle and micropuncture set, and a 0.035-inch guidewire was advanced into the superior vena cava. Separately, an 18-gauge needle was inserted into the peritoneal cavity under ultrasound guidance, and a 0.035-inch guidewire was advanced. The peritoneal limb of the PVS was then introduced via a 16-F peel-away sheath into the peritoneal cavity and connected to wall suction, allowing drainage of 1-2 L of ascites.

A pump pocket was created over the lower ribs by blunt dissection, and the pump was placed within the pump pocket. The peritoneal portion of the shunt was tunneled subcutaneously to the peritoneal access site, while the jugular portion was tunneled to the venous access site. The pump chamber was manually pumped until fluid flowed freely through the venous limb. Finally, a 12-F or 16-F peel-away sheath was introduced into the jugular vein over the guidewire, and the jugular limb of the shunt was advanced, with its tip positioned in the mid-right atrium.

Statistical analysis

Quantitative data are expressed as mean ± SD. Comparisons between two groups were performed using Student’s t-test, while multiple group comparisons were analyzed with one-way ANOVA. Correlations between categorical variables were assessed using the χ² test or Fisher’s exact test, as appropriate. Cumulative recurrence-free survival rate was estimated by the Kaplan-Meier method, and differences in overall survival were evaluated using the generalized log-rank test. Multivariate analysis was conducted with a Cox proportional hazards regression model to assess the combined effects of prognostic factors. A P value < 0.05 was considered statistically significant. Statistical analysis was performed in SPSS software (IBM Corp., Armonk, NY, United States). For comparisons between the S-group and L-group, a Student’s t-test was used for continuous variables, including age, Child-Pugh score, albumin-bilirubin (commonly known as ALBI) score, platelet count, operation time, postoperative hospital stay, liver volume, and spleen volume. For categorical variables such as sex, Child-Pugh classification, and morbidity, Fisher’s exact test was applied.

RESULTS

Patient groups

The clinical characteristics of the study participants are summarized in Table 2. The mean age was 65.5 years, and the median survival time was 227 days. The 30-day mortality rate was 9%, and overall morbidity was high. Postoperative complications included 4 cases of liver failure and 2 cases of hemorrhage. Morbidity was defined as Clavien-Dindo grade III or higher. For analysis, the disease severity index (commonly known as DSI) and hepatic reserve were substituted with the albumin-bilirubin score and Child-Pugh score, respectively. Portal-systemic shunting was replaced by spleen volume and platelet count. Shunt dysfunction occurred in 6 patients (6%), including 4 cases of occlusion (3 requiring reinsertion) and 2 cases of infection necessitating shunt removal. Patients who died within 180 days of PVS placement (S-group) numbered 43, and those who survived longer (L-group) numbered 57.

Table 2 Patient summary, mean ± SD.

Characteristic	Finding
Age	65.5 ± 10.1
Female/male	69/31
Etiology of HBV/HCV/alcohol/MASLD/others	41/7/32/9/11
Child-Pugh classification of A/B/C	0/59/41
Child-Pugh score	9.41 ± 1.46
ALBI score	-1.19 ± 0.53
Platelet as 103/μL	109 ± 52.7
Operation time in minutes	83.5 ± 23.7
Morbidity, +/-	15/85
Post operative hospital stay in days	24.6 ± 28.4
Liver volume in mL	1024 ± 324
Spleen volume in mL	344 ± 182

HBV: Hepatitis B virus; HCV: Hepatitis C virus; MASLD: Metabolic dysfunction-associated steatotic liver disease; ALBI: Albumin-bilirubin.

Scoring system

A comparison of the S- and L- groups is presented in Table 3. Patients in the S-group were more often female, with higher mean age and Child-Pugh scores than those in the L-group (P = 0.001). Preoperative liver volume was significantly greater in the L-group than in the S-group (P = 0.004). Based on these findings, we developed a new scoring system. Cutoff values for age, Child-Pugh score, and liver volume were determined by receiver operating characteristic analysis: Age, 60 years (sensitivity 0.837, specificity 0.451); Child-Pugh score, 9.0 (sensitivity 0.551, specificity 0.725); and liver volume, 1057 mL (sensitivity 0.621, specificity 0.771) (Figure 1). The scoring system was defined as 1 point each for age ≥ 60 years, Child-Pugh score ≥ 10, female sex, and preoperative liver volume < 1057 mL. Total scores of 0-2 points were classified as PVS grade A, and scores of 3-4 points as PVS grade B (Table 4).

Figure 1 Receiver operating characteristic analysis. A: Age [cutoff value: 60, sensitivity: 0.837, specificity: 0.451, area under the curve (AUC): 0.676]; B: Child-Pugh score (cutoff value: 9, sensitivity: 0.551, specificity: 0.725, AUC: 0.694); C: Liver volume (cutoff value: 1057, sensitivity: 0.621, specificity: 0.771, AUC: 0.692).

Table 3 Comparison of the S-group and L-group, mean ± SD.

Variable	S-group1, n = 43	L-group2, n = 57	P value
Age	69.3 ± 9.11	62.6 ± 10.1	0.001
Female/male	26/17	43/14	0.083
Child-Pugh classification A/B/C	0/18/25	0/41/16	0.004
Child-Pugh score	10.1 ± 1.65	8.93 ± 1.10	0.001
ALBI score	-1.30 ± 0.45	-1.58 ± 0.55	0.009
Platelet as 103/μL	117 ± 63.2	133 ± 64.3	0.231
Operation time in minute	86.1 ± 32.2	81.8 ± 16.2	0.472
Morbidity	9	6	0.243
Post operative hospital stay in days	28.1 ± 28.4	24.3 ± 29.0	0.575
Liver volume in mL	897 ± 211	1129 ± 368	0.004
Spleen volume in mL	396 ± 207	378 ± 262	0.773

¹L-group, patients who survived more than 180 days after peritoneovenous shunt placement.

²S-group, patients who survived for less than 180 days after peritoneovenous shunt placement.

ALBI: Albumin-bilirubin.

Table 4 Peritoneovenous shunt grade.

Variable	Point assigned
Parameter	0	1
Age	≤ 60	> 60
Sex	Male	Female
Child-Pugh score	5-9	10-15
Liver volume in mL	> 1057	≤ 1057
Grade
A	0-2
B	3-4

Overall survival rate, multivariate analysis

Evaluation of the scoring system in this cohort demonstrated that overall survival was significantly higher in grade A patients compared with grade B patients (P = 0.001) (Figure 2). In multivariate analysis, PVS grade was identified as an independent prognostic factor (Table 5).

Figure 2 Overall survival. The overall survival rate was significantly higher in grade A patients than in grade B patients.

Table 5 Multivariate analysis.

Variable	Hazard ratio	95%CI	P value
Age	1.047	1.005-1.090	0.027
Sex	0.374	0.165-0.866	0.022
Child-Pugh score	1.286	1.016-1.626	0.036
Liver volume	1.001	0.999-1.002	0.264
PVS grade	0.292	0.103-0.828	0.021

CI: Confidence interval; PVS: Peritoneovenous shunt.

DISCUSSION

We evaluated prognostic factors for PVS placement and developed a new scoring system to aid in patient selection. Although PVS can provide effective symptomatic relief for refractory ascites, it is associated with high mortality and morbidity. A meta-analysis reported 12- and 48-month mortality rates of 44% and 55%, respectively[4], whereas in our cohort the corresponding rates were 52% and 54%, respectively. Reported complications included disseminated intravascular coagulation in approximately 6% of patients and infections in up to 15%[4].

In our study, patients who survived more than 180 days after PVS placement were defined as the long-term survival group (L-group). This cutoff was chosen because the median survival time in our cohort was 227 days. Given the high risks associated with PVS, we believe the procedure should be offered primarily to patients with an anticipated long-term survival. Our scoring system helps identify such patients, as those classified as PVS grade A were more likely to survive beyond 180 days following PVS placement. This model may therefore assist clinicians in selecting appropriate candidates and optimizing outcomes. We compared several parameters between the S- and L-groups and found that age, sex, Child-Pugh score, and liver volume differed significantly. Using these factors, we developed a new prognostic scoring system. In our study, advanced age and female sex emerged as risk factors for poor perioperative recovery after PVS placement.

Aging is known to exacerbate chronic liver disease[13]. The prevalence and severity of many liver disorders, including cirrhosis and hepatocellular carcinoma, increase with age, and the concept of “age-related liver disease” has been introduced to encompass a spectrum of disorders, such as metabolic dysfunction-associated steatotic liver disease (MASLD), alcoholic liver disease, viral hepatitis, fibrosis, and cirrhosis[14].

Sex-related differences in liver health and disease are also well-established. For example, MASLD is influenced by obesity, insulin resistance, hyperlipidemia, and sex-related factors[15]. Recent studies have highlighted sex disparities in MASLD development, underscoring the need to better understand sex-specific responses to liver injury[16,17]. Notably, MASLD in women has been associated with increased overall mortality, as well as higher risk of death from cancer, cardiovascular disease, and liver disease[18]. Similarly, female sex is a recognized factor for the progression of alcohol-related disease[19]. Women are more vulnerable to alcohol-related liver injury due to differences in alcohol metabolism, including lower gastric alcohol dehydrogenase activity and altered enzymatic pathways which lead to higher concentrations of acetaldehyde and other hepatotoxic metabolites[20]. In addition, smaller body size, lower lean body mass, and higher fat content contribute to higher systemic alcohol concentrations in women compared with men. These biological differences may partly explain why female sex was associated with poorer outcomes following PVS placement in our study.

Liver volume was also identified as a prognostic factor in our study. Previous reports have shown that computed tomography can help detect features associated with cirrhosis-related complications[21]. Several studies have demonstrated that platelet count, together with total liver volume, right liver volume and spleen volume, correlates with Child-Pugh classification and the presence of esophageal varices[22-26]. Lee et al[27] further reported that a greater rate of liver volume reduction was significantly associated with an increased risk of decompensation events in patients with LC. Similarly, Hagan et al[28] identified low liver volume as a prognostic predictor for both transplantation and mortality. Spleen size has also been linked to disease etiology, with splenomegaly observed more frequently in patients with post-hepatitis cirrhosis than in those with alcoholic cirrhosis[26]. In our cohort, we examined splenic volume in patients with LC but found no statistically significant relationship between spleen size and prognosis after PVS placement.

We developed a new scoring system using these prognostic parameters and stratified patients into two groups: Grade A and grade B. We recommend PVS placement for grade A patients, who are more likely to survive beyond 180 days after the procedure. However, we do not regard grade B status as a strict contraindication. Given the palliative role of PVS, patients in the grade B group may still derive meaningful symptomatic benefit, such as relief of abdominal distension and reduced need for repeated paracentesis, even though their life expectancy is typically shorter than 180 days.

Despite the significant findings of this study, several limitations should be acknowledged. First, this was a single-center analysis with a relatively small sample size, which may limit the generalizability of the results. Second, the extended enrollment period (1998-2024) introduces potential heterogeneity, as advances in patient management, surgical techniques, and supportive care over time may have influenced outcomes. To enhance external validity, future studies should include larger, more diverse populations and be conducted prospectively across multiple centers.

CONCLUSION

This study identified key prognostic factors for PVS placement and introduced a novel classification system based on age, sex, Child-Pugh classification, and liver volume. Patients were stratified into grade A and grade B, with this grading system serving as a useful indicator for predicting prognosis after PVS placement.