Predictors of survival after Transjugular Intrahepatic Portosystemic Shunt (TIPS) in patients with refractory ascites in end-stage liver cirrhosis

Objectives: The aim of this study is to analyse predictors for overall survival (OS) after TIPS in patients with refractory ascites in end-stage liver cirrhosis. Methods: A total of 213 consecutive patients (129 male, 84 female, age 59.9±10.5 years) with TIPS for refractory ascites were included. Patient characteristics, clinical outcome and survival data were analysed from institutional databases and respective records from referring physicians and registration offi ces. Data were completed with a telephone questionnaire. Research Article Predictors of survival after Transjugular Intrahepatic Portosystemic Shunt (TIPS) in patients with refractory ascites in end-stage liver cirrhosis Pitton Michael Bernhard1*, Zimmermann Tim2, HoppeLotichius Maria3, Mildenberger Philipp4, Kloeckner Roman5, Mittler Jens6, Weinmann Arndt7, Otto Gerd8 and Düber Christoph9 1Department of Diagnostic and Interventional Radiology, University Medical Center, Langenbeckstr.1, 55131 Mainz, Germany 2Department of Internal Medicine I, University Medical Center, Langenbeckstr.1, 55131 Mainz, Germany 3Department of General and Visceral Surgery and Transplantation Surgery, University Medical Center, Langenbeckstr.1, 55131 Mainz, Germany 4Institute of Medical Biostatistics, Epidemiology and Informatics (IMBEI), University Medical Center, Langenbeckstr.1, 55131 Mainz, Germany 5Department of Diagnostic and Interventional Radiology, University Medical Center, Langenbeckstr.1, 55131 Mainz, Germany 6Department of General and Visceral Surgery and Transplantation Surgery, University Medical Center, Langenbeckstr.1, 55131 Mainz, Germany 7Department of Internal Medicine I University Medical Center, Langenbeckstr.1, 55131 Mainz, Germany 8Emeritus of the Division of Transplantation Surgery, University Medical Center, Langenbeckstr.1, 55131 Mainz, Germany 9Department of Diagnostic and Interventional Radiology, University Medical Center, 55131 Mainz, Germany Received: 04 May, 2021 Accepted: 15 June, 2021 Published: 16 June, 2021 *Corresponding author: Dr. Pitton Michael Bernhard, EBIR, MHBA, Professor, Department of Diagnostic and Interventional Radiology, University Medical Center, Johannes Gutenberg University Mainz, Langenbeckstr.1, 55131 Mainz, Germany, Tel: ++49 6131 172057; Fax: ++49 6131 176642; E-mail:


Introduction
Refractory ascites is a challenging complication of portal hypertension in end-stage liver cirrhosis. Transjugular Portosystemic Stent Shunt (TIPS) basically reduces portosystemic pressure gradient (PSG) and thereby reduces ascites and prevents bleeding from of varices [1][2][3]. Alternative techniques like alpha pump or abdominal drains might improve ascites [4], but do not reduce portal vein pressure. The introduction of Polytetrafl uoroethylene (PTFE)-covered stent-grafts has signifi cantly improved clinical outcome after TIPS [1,2,5]. TIPS is an effective option for patients who are no candidates for liver transplantation, for candidates on the waiting list, and in new-onset refractory ascites after liver transplantation [6].
Survival after TIPS signifi cantly depends on the clinical stage of the liver cirrhosis, however, the underlying liver disease and personal co-factors might also impact the clinical results [1].
The aim of this study is to analyse the clinical outcome after elective TIPS in refractory ascites. We particularly analyse the clinical pre-conditions and co-morbidities in order to identify predictive risk factors for Overall Survival (OS).

Methods
Between July 2002 and August 2016 all consecutive patients with TIPS for refractory ascites were included.
Patient characteristics and outcome were analysed from our institutional databases as well as from medical reports of subsequent treatments and interventions of referring hospitals. Survival data were obtained from our clinical records and/or registration offi ces. According to EASL guidelines, all patients suffered from grade 3 ascites [7]. We defi ned moderate grade 3 ascites as relevant amount of peritoneal fl uid in CT/ MRI fi ndings, no extensive abdominal distension but need for intermittent paracentesis under moderate doses of diuretics (e.g. Spironolactone <100mg and/or Furosemide <100mg/d and analogues). Extensive grade 3 ascites was defi ned as extensive abdominal distension, repetitive need for large volume paracentesis and high doses of diuretics (e.g. Spironolactone >100mg/d and/or Furosemide >100mg/d and analogous), or hepatorenal syndrome that did not allow for higher doses of diuretics [7]. The data were completed with a telephone questionnaire with the referring general practitioners, the patients or their relatives in case of patient's death. The study was carried out according to the Helsinki Declaration. The retrospective data analysis and the patient questionnaire were approved by the local ethics committee.
The basic technique of TIPS has been described in the 1990s [8]. After  The corresponding stratifi ed median survival and the hazard ratios between the resulting strata were calculated. Given the Results: Patients presented in clinical stage Child-Pugh A, B, and C in 4.2%, 72.3%, and 23.5%. The primary and secondary success rate for TIPS was 97.2% and 100%. 3d mortality was 0.9%. There were 10 major complications (4.7%) and 2 minor complications (1%). 1-, 2-, 3-, and 5-year OS was 64.1%, 54.5%, 51.6% and 44.6%, respectively. MELD>14, advanced age>70 years, cirrhosis caused by viral hepatitis, cardiac diseases and other co-morbidities, as well as early need for haemodialysis within 4 weeks after TIPS were predictive for inferior survival. Vice versa, early response of ascites after TIPS seems to be correlated with improved survival.
Conclusions: TIPS with PTFE-covered stent grafts is safe and effective for treatment of refractory ascites in patients with decompensated liver cirrhosis in this realworld setting. OS after TIPS is not only predicted by pre-interventional liver function but also by the aetiology of liver cirrhosis, personal factors like age, sex, and various co-morbidities. The demonstrated risk factors should be considered for decision making for TIPS. The status of early responder to TIPS (defi ned as ascites reduction during the fi rst three months) as well as the need for dialysis within the fi rst four weeks after TIPS use both future information. To prevent bias, analysis was conducted on the subset of patients who survived the fi rst 4 weeks or 3 months, respectively.  presented a history of former bleeding episodes which had been successfully treated. However, there was no evidence for repeated bleeding episodes during at least 3 months before TIPS, meaning that the indication for TIPS was dedicated to refractory ascites only.

Between
The primary technical success rate was 97.2%. In 6 cases the portal vein was not entered in the fi rst attempt but was completed in a second attempt, resulting in an overall technical success rate of 100%. The 3-day and 30-day mortality was 0.9% and 8.0%, respectively. There were 10 major complications (4.7%) and 2 minor complications (1%) ( Table   3). Major complications covered 6 bleeding complications and 4 cardiac events. Bleeding complications were caused by artifi cial puncture of segmental arteries along the puncture tracts. Four of six bleeings were treated with coil embolization of the respective segmental arteries, two of these patients died in multi-organ failure at one and three days after the TIPS procedure. Another two patients suffered from some hemobilia   Figure 1a,b.
The unadjusted univariate risk analysis identifi ed MELD >14, age>70 years, cirrhosis caused by viral hepatitis, cardiac diseases and other co-morbidities predictive for inferior survival. Child-Pugh score presented decreasing survival from Child A to C, however showed broad confi dence intervals and did not reach statistical signifi cance. With respect to laboratory fi ndings, increased levels of creatinine, bilirubin, and INR, and decreased levels of albumin and thrombocytes were signifi cant risk factors. After TIPS creation, an inadequate PSG of <5mmHg or >9mmHg or need for haemodialysis within 4 weeks after TIPS, each was predictive for an inferior outcome. Failed early response of ascites resolution within 3 months after TIPS seems predictive for inferior outcome, but this did not reach statistical signifi cance ( Table 5). The status of early ascites responder to TIPS as well as need for dialysis within the fi rst four weeks after TIPS use both future information. To prevent bias, a separate analysis was conducted on the subset of patients who survived the fi rst 4 weeks or 3 months, respectively. Results confi rmed both factors predictive for survival Figure 2a  and need for early haemodialysis after TIPS may not serve as prognostic factor before TIPS creation and were therefore not tested in the multivariate cox model (Table 5).

Discussion
The data presented herein demonstrate that overall survival depends on a complex synopsis pre-interventional liver function (MELD score), individual co-factors (sex, age), and co-morbidities.   fl ow obstruction, reduces porto-systemic pressure gradient, increases circulatory volume to the heart, and improves sodium excretion, renal function, and hepatic hydrothorax [1,[9][10][11]. In a randomized study, Bureau      one year were INR, bilirubin, platelet count, Child-Pugh score, MELD score, MELD-Na score, and patient age [14]. Our data show slightly reduced OS compared with these reports from the literature. This might be due to different patient cohorts including variant clinical stages of liver disease as well as personalized co-factors which are subject of this analysis. Our cohort, e.g., included nearly one quarter of patients with clinical stage Child-Pugh C.
Our data confi rmed an increased risk in patients with Child-Pugh stage C, increasing Child-Pugh score was also associated with increased mortality, however did not reach statistical signifi cance because of broad variances between the groups. MELD score was more predictive compared to Child-Pugh clinical stages to predict outcome. The best cut-off for impaired survival after TIPS was MELD >14 points. MELD score is known to reliably predict survival after TIPS in all causes of end-stage liver disease and is also used for allocation of liver transplantation. The score was suggested to be independent from the aetiology of the liver cirrhosis [15,16]. Our data, however, demonstrated a signifi cantly reduced survival after TIPS in cirrhosis from viral hepatitis compared to alcoholic liver cirrhosis and thus support the assessment from the initial Mayo-TIPS-model that scores TIPS in cirrhosis from viral hepatitis with reduced OS compared to alcoholic cirrhosis [17].
Cardiac diseases were also a risk factor after TIPS but were correlated with increasing age and therefore no independent predictor in the multivariate cox model. Acute right heart failure might occur shortly peri-interventionally after opening the TIPS-shunt and can be explained by signifi cant hemodynamic changes after TIPS. Fili, et al reported a signifi cant right atrial and pulmonary volume load and pressure increase, increase of capillary wedge pressure and cardiac index and a decrease in systemic vascular resistance accompanied with an increased rate of diastolic dysfunction [18]. TIPS placement should therefore be supported by an anaesthesiologist with experience in the management of acute cardiac volume load.
Personal factors like sex and advanced age has already been reported as risk factor. With respect to age, Parvinian, et al.
Moreover, advanced age might be accompanied by an increasing number of co-morbidities and combinations of them which typically result in reduced activities in daily living. The ADL score calculates these activities of daily living (ADL). The score includes and rates seven respective activities of daily living (to shop, do laundry, ambulate, clean, transfer, bathe, and dress); each score with 1 to 3 points, resulting in a total score between 7 and 21 points. Grunewald, et al. reported respective functional disability with an ADL score <21 as predictor of post-TIPS mortality and length of hospital stay [19]. In large patient cohorts, patient´s sex predicted clinical outcome to some extent with slightly inferior outcomes in males [1]. In the univariate analysis, we demonstrated a respective trend towards a slightly increased hazard ratio for males; however, the data demonstrated a crossover of Kaplan Meier survival curves of males and females and thereby did not meet the proportional hazard assumption of the cox model. The data were therefore stratifi ed by sex for further multivariate analysis.
The need for haemodialysis within 4 weeks after TIPS was correlated with reduced survival. Vice versa, early response to TIPS with early relief of ascites within 3 months was prognostically favourable in the univariate analysis. This allows for some further prognosis in a post-hoc perspective.
However, it is not suited for prognostic assumptions or decision making before TIPS creation. With this restriction, we thereby confi rm respective fi ndings from Tan, et al. who suggested complete ascites response and the use of covered stent grafts to be predictive for improved survival [10]. Similar data were reported by Taki, et al. who investigated the effect of early response of ascites after TIPS. The overall survival rate at 1, 2, and 3 years was 76%, 54%, and 45% with improved survival in responders at 4 weeks compared to non-responders. Serum creatinine <1.9mg/dl was identifi ed as a correlated predictor of early response [11].
Post-TIPS porto-systemic Pressure Gradients (PSG) are routinely used to confi rm pressure relief from the portal venous system. However, immediate PSG measurement might be signifi cantly infl uenced by right heart function. The amount of PSG reduction can either result from acute pressure relief in the portal vein or an accompanying increase of right atrial pressure due to the acute volume load from the abdominal space into the right atrium, or both. However, an inappropriate instantaneous post-interventional PSG <5mmHg or >9mmHg, each showed a trend to reduced survival, but did not reach statistical signifi cance. However, immediate post-TIPS measurement of PSG might be signifi cantly infl uenced by volume overload of the right heart and does not refl ect the hemodynamic results in steady state. Post-TIPS PSG measurements should therefore be performed in a new steady state situation two or three days after TIPS when a respective volume shift and right heart overload has been compensated.

Biases and limitations
This is a retrospective analysis over a long period of time. It

Conclusion
TIPS with PTFE-covered stent grafts is safe and effective for treatment of refractory ascites in patients with decompensated liver cirrhosis. Survival after TIPS is not only predicted by pre-interventional liver function but also by the etiology of liver cirrhosis, and individual co-factors like sex, age, and co-morbidities. The demonstrated risk factors should be considered for decision making for TIPS.