Causes and Rates of 30-Day Readmissions After Transjugular Intrahepatic Portosystemic Shunts

OBJECTIVE. The purpose of this study was to investigate the causes and rates of 30-day readmission after transjugular intrahepatic portosystemic shunt (TIPS) at a single liver transplant center. MATERIALS AND METHODS. We reviewed 165 TIPS procedures performed between 2003 and 2013. After excluding patients who died during the index admission (n = 16), any readmission within 30 days of discharge was identified, and cause of readmission was determined. Causes were categorized as planned or unplanned and interventional radiology (IR)-related or IR-unrelated. Unplanned readmissions were independently categorized as preventable or unpreventable by two interventional radiologists. Discrepant opinions were resolved by consensus. Factors predictive of 30-day readmission were identified by univariate and multivariate analysis. RESULTS. The reviewed TIPS procedures were performed in 165 patients (mean age ± SD, 56 ± 11 years; 69% male, 31% female). TIPS were placed for ascites or hydrothorax in 82 patients (50%) and variceal bleeding in 83 patients (50%). The 30-day readmission rate was 21% (31/149). The most common causes for readmissions were ascites or hydrothorax (23%, 7/31) and hepatic encephalopathy (23%, 7/31). All 30-day readmissions were unplanned; 17 (55%) of them were potentially preventable. Of the 17 potentially preventable readmissions, five (29%) were IR-related and 12 (71%) were IR-unrelated. In IR-related readmissions, all patients presented with a recurrence of symptoms (rebleeding or ascites) and were found to have TIPS stenosis or occlusion. Mortality rates were similar between patients who were and were not readmitted (p = 0.23). On multivariate analysis, spontaneous bacterial peritonitis during the index admission was the only variable associated with 30-day readmission (odds ratio = 4.81, p = 0.02). CONCLUSION. Over half of 30-day readmissions after TIPS could have been prevented by early outpatient follow-up and intraprocedural technique to optimize stent landing zones.