Outcomes of Patients with Left Ventricular Assist Devices Undergoing Mesenteric Angiography for Gastrointestinal Bleeding.

ABSTRACT

PURPOSE: To compare measures of clinical success, such as the need for subsequent intervention and mortality, in patients with left ventricular assist devices (LVADs) undergoing mesenteric angiography for gastrointestinal (GI) bleeding with respect to a control group. MATERIALS AND METHODS: A retrospective study was conducted on 48 consecutive patients undergoing anticoagulation whose GI bleeding was assessed with angiography between August 2007 and June 2014 24 patients with LVADs and 24 control patients without LVADs. The χ2 and t tests were used for statistical analysis. RESULTS: Mean ages were 62.1 years ± 9.6 and 74.5 years ± 11.3 in the LVAD and control groups, respectively. No significant difference was observed in hemodynamic instability, presenting hemoglobin level and International Normalized Ratio, or hemoglobin nadir. Two patients with LVADs (8.3%) and 8 control patients (33.3%) had bleeding detected on angiograms (P = .032). Six embolizations were performed in patients with LVADs and 8 were performed in control patients. Clinical success was achieved in 2 of 6 patients with LVADs (33.3%) and 7 of 8 control patients (87.5%; P = .036). Seven patients with LVADs (29.2%) and 1 control patient (4.5%) underwent repeat angiography within 14 days (P = .020). Seven patients with LVADs (29.2%) and 4 control patients (18.2%) required postprocedural endoscopic or operative intervention as definitive therapy (P = .302). All-cause in-hospital mortality rates were 16.7% in the LVAD group and 4.2% in the control group (P = .032), and the respective all-cause 1-year mortality rates were 33.3% and 9.1% (P = .080). CONCLUSIONS: A higher rate of clinical failure is observed in patients with LVADs presenting with GI bleeding compared with those without LVADs, with a more frequent need for subsequent endoscopic or surgical intervention.