Gastrointestinal Bleed Post HeartMate-2 LVAD Implant

Abstract

Purpose: Left ventricular assist device (LVAD) augments systemic circulation and improves quality of life in patients suffering from end stage heart failure. Until recently, most LVAD implants have been with pulsatile devices. Newer continuous-flow (CF) pumps have resulted in superior outcomes with reduction in size, complications and improved durability over first-generation pulsatile pumps. Despite these improvements gastrointestinal bleeding (GIB) continues to be problematic. Reports have implicated angioectasia as one of the culprit lesions leading to such high rates of GIB. The purpose of this single center retrospective study is to report the etiologies and incidence of GIB in patients following the implantation of HeartMate-2 (HM2) CF-LVAD. Methods: Medical records of 84 patients, ages 28-77, implanted with HM-2 LVAD at Penn State Hershey Medical Center (PSHMC) from July 2006 through November 2011, were reviewed. GIB was defined as one guiac-positive stool with a subsequent transfusion of at least two units of PRBC's. Results: A total of 20 (24%) of 84 patients developed GIB after HM2 LVAD implant. Of those patients,2 (10%) had a prior history of GIB. Median interquartile range [IQR] time to onset of bleed was 51.5 [25.5-347] days. In the multivariable logistic regression, age, sex, and extracorporeal membrane oxygenation (ECMO) prior to LVAD placement were not associated with increased incidence of GI bleed. (See Table). Conclusion: This study demonstrates that the incidence of GIB post HM-2 LVAD implant at PSHMC is 24%, 20 of 84 patients. Our result implicates the majority of GIB were due to angioectasia from the small intestine (33%) and gastric ulcers (20%). Active bleeding sites were identified in 12 (60%) of 20 patients. EGD identified most lesion 11 of 20 GIB patients (55%). 7 (35%) patients were admitted on ≥3 separate occasions for GI bleed. Of those 7 patients, 4 (57%) had different sites of GIB identified on subsequent admissions. 6 of 20 patients (30%) underwent capsule studies when endoscopic studies were non-revealing but no active sites of bleeding were identified on the capsule studies. However, 3 (50%) of those 6 patients had angioectasia in various regions of the GI tract. Video capsule endoscopy (VCE) studies were completed safely in all patients.