Venous Thromboembolism and Intracranial Hemorrhage in Patients with High-grade Glioma

AbstractBackgroundPatients with high-grade glioma (HGG) are at increased risk of venous thromboembolism (VTE), throughout the course of disease. Prophylactic anticoagulation is not established among patients with HGG, outside of perioperative context, due to potential for intracranial hemorrhage (ICH) and the limited data available for predictive VTE scores in primary brain tumors. The aim of this study was to characterize VTE prevalence in adult patients with HGG and assess ICH risk during therapeutic anticoagulation.MethodsRetrospective analysis of patients diagnosed with HGG at our institution, between 2009 and 2018. All adult patients proposed to systemic treatment were included into this study. Exclusion criteria was anticoagulation previous to diagnosis. VTE was defined as any radiographic-confirmed thrombus in the venous system. Risk factors of interest for VTE and bleeding risk scores were analyzed by chi-squared test and multivariate logistic regression. Survival analysis was performed using Kaplan-Meier method.ResultsA total of 410 patients were included of whom 31 (7,8%) developed a VTE, including 22 deep vein, 6 pulmonary and 3 central venous thrombosis. Twenty-nine patients with VTE had a WHO grade 4 glioma and 2 patients had grade 3 (anaplastic astrocytoma and oligodendroglioma).In 22 cases, the VTE occurred during systemic treatment, more frequently during Temozolomide (n=15), followed by Irinotecan+Bevacizumab (n=6), Lomustine+Bevacizumab (n=1) and PCV (n=1). The median time between diagnosis and VTE was 10,11 months (95CI 6,46-14,79). Khorana score, age, ECOG performance status, smoking and obesity did not significantly differ in the VTE population. All VTE were initially treated with low molecular weight heparin (LMWH), of which 64.5% maintained LMWH, and the remainder switched to warfarin (19.4%) or to direct oral anticoagulant (16.1%). Six patients (19,4%) had spontaneous ICH under anticoagulation. Patients with grade 3 glioma (p=0,032) had significantly higher rates of ICH than grade 4. Patients with higher ECOG had significantly higher risk of ICH (OR 3,23 (95CI 1,18-8,81), p=0,022). HAS-BLED and ACCP bleeding scores were not associated with ICH. There was no significant difference in overall survival for TVE or ICH.ConclusionAccording to our data, ICH occurred in nearly 20% anticoagulated patients with HGG, as described in literature, and did not correlate with poorer prognosis. High ECOG performance status was an independent risk factor for ICH. Further effort towards better prediction models for VTE and ICH in HGG is warranted.