Hospital Utilization for Patients With Cirrhosis and Severe Ascites in a Model of Outpatient Paracentesis by Interventional Radiology.

backgroundParacentesis is currently performed by interventional radiologists (IR) rather than gastroenterologists/hepatologists or internists. In this model of care, there is usually no evaluation of patients' renal function or adjustment of their medications at the time of paracentesis. The objectives of this study were to analyze hospital utilization and cirrhosis complications within six months of index outpatient paracentesis by IR and to identify potential areas of improvement in care.

methodsThis is a retrospective study of patients with cirrhosis and ascites who underwent outpatient paracentesis by IR between October 15, 2015, and October 15, 2018, at a tertiary academic medical center. We collected demographics, data on cirrhosis etiology/complications, laboratory tests, provider notes, outpatient paracentesis dates, emergency department (ED) visits, hospitalizations, and ICU admissions within the following six months post index paracentesis. Associations between categorical predictors and clinical outcomes were analyzed using the chi-square test. Associations between quantitative predictors and clinical outcomes were analyzed using the Wilcoxon rank sum test.

resultsOur study included 69 unique patients who had at least one outpatient encounter for paracentesis by IR in the study period. Most patients were men (71%), had alcohol-related cirrhosis as primary etiology (53.6%), an average age of 60 years, and an average Model for End-Stage Liver Disease-sodium (MELDNa) score at baseline of 16. Within six months from index paracentesis, 44 patients (64.7%) underwent repeat IR outpatient paracentesis (total 187 paracenteses, 4.25 paracenteses/patient), 43 patients (62.3%) had ER visits (total 118 ER visits, 2.8/patient), 41 patients (59.4%) had hospital admissions (total 88 admissions, 2.2/patient), and 11 patients required ICU admission. Complications of cirrhosis noted during follow-up included hepatic encephalopathy (40.5%), acute kidney injury (38.2%), upper gastrointestinal (UGI) bleeding (16%), and spontaneous bacterial peritonitis (SBP) in 15%. The mortality rate at six months was 20%. On multivariate analysis, the predictive factors for mortality were older age (p = 0.03) and MELDNa score (p = 0.02). Baseline MELDNa was predictive of acute kidney injury (p = 0.02), UGI bleed (p < 0.01), and ICU admission (p < 0.01), but not of SBP, encephalopathy, ED visit, or hospital admissions. Among patients with more than one paracentesis (64%),six patients underwent transjugular portosystemic shunt (TIPS), but there was no documentation of TIPS consideration in 31 patients (70.4%). A total of 20 patients (29%) were waitlisted for liver transplantation.

conclusionIn this contemporary cohort of patients with cirrhosis undergoing outpatient IR paracentesis, we found a high rate of short-term cirrhosis complications and hospital utilization, while TIPS consideration was very low. Further data are needed to identify specific gaps in care, but IR paracentesis should be integrated within a multidisciplinary management model, with emphasis on early TIPS in eligible patients, as recommended by the current practice guidelines.