Intravenous Bevacizumab for Refractory Hereditary Hemorrhagic Telangiectasia–Related Epistaxis and Gastrointestinal Bleeding

Abstract

ObjectiveTo present a multiyear clinical experience with intravenous bevacizumab for the management of severe gastrointestinal bleeding and/or epistaxis in patients with hereditary hemorrhagic telangiectasia (HHT). Patients and MethodsAll patients treated with intravenous bevacizumab for severe hereditary hemorrhagic telangiectasia–related bleeding from June 1, 2013, through January 31, 2017, were included in this report. Severity of epistaxis (determined using the Epistaxis Severity Score questionnaire); hemoglobin, iron, and ferritin levels; and quality of life data were collected serially in all patients. ResultsIntravenous bevacizumab was administered to 34 patients using a standardized treatment protocol. Anemia was primarily related to severe epistaxis (n=15, 44%), severe gastrointestinal bleeding (n=4, 12%), or both (n=15, 44%), with a median baseline hemoglobin level of 9.1 g/dL (range, 8.3-10.5 gm/dL; to convert to mmol/L, multiply by 0.62). Red blood cell (RBC) transfusions had been administered to 28 patients (82%). Of these, 16 patients (47%) were RBC transfusion dependent and had received a median of 75 RBC transfusions (range, 4->500 RBC units) before bevacizumab initiation. The median length of follow-up was 17.6 months from the beginning of bevacizumab treatment (range, 3-42.5 months). There was a significant reduction in epistaxis severity scores (P<.001) and RBC transfusion requirements (P=.007) after completion of the initial bevacizumab treatment cycle. New-onset or worsened hypertension was noted in 4 patients, with 1 patient experiencing hypertensive urgency with a temporary decline in renal function. ConclusionIntravenous bevacizumab is an effective treatment option for patients with severe anemia related to epistaxis and/or gastrointestinal bleeding. Further studies are needed to establish a dose-response relationship as well as clinical, genetic, and biomarker predictors of response.