Systematic review with meta-analysis: automated low-flow ascites pump therapy for refractory ascites

Background Few effective treatments are available for patients with cirrhosis and refractory ascites. New treatment modalities are needed for these patients. Aim To synthesise the available evidence on the efficacy and safety of automated low-flow ascites pump therapy in patients with cirrhosis and refractory ascites. Methods Electronic databases were searched for trials evaluating automated low-flow ascites pump therapy in patients with refractory ascites. Results Nine studies were included. Eight were case series, one was a randomised controlled trial. Pooled estimate rates were 0.62 (95% CI = 0.49-0.74) for the absence of requirement of large volume paracentesis (LVP) after pump insertion, 0.30 (95% CI = 0.17-0.47) for acute kidney injury, 0.27 (95% CI = 0.13-0.49) for bacterial peritonitis and 0.20 (95% CI = 0.09-0.37) for urinary tract infection. There was high heterogeneity between studies which was often reduced or eliminated in sensitivity analyses by excluding studies of patients with a mean or median model for end-stage liver disease (MELD) score > 15. Results of sensitivity analyses were similar to those of overall analyses. Mean increase in serum creatinine level after pump insertion was 23 mu mol/L (95% CI = 10-35) with no heterogeneity between studies. The pooled estimate rate for pump-related side effects was 0.77 (95% CI = 0.64-0.87) with low heterogeneity between studies. Conclusion This meta-analysis demonstrates that most patients treated with automated low-flow ascites pump therapy do not require LVP after pump insertion. Acute kidney injury occurs in 30% of patients and creatinine levels increase by a mean of 23 mu mol/L after pump insertion. Bacterial peritonitis and urinary tract infection occur in 27% and 20% of patients respectively.