Treat the VTE and worry less about the hemorrhage: A sigh of relief or is the jury still out for therapeutic anticoagulation in high-grade glioma?

Management of venous thromboembolic events (VTE) is a complicated, but critical, aspect of the care of glioma patients. Though the incidence is highest in glioblastoma, IDH wildtype, cited at about 30%, there is also an increased risk in lower-grade gliomas that harbor the IDH mutation and extend outside the postoperative period.¹ Several patient factors—age, hemiparesis, extent of resection, tumor size—and biologic factors—such as the tumor hypoxia, upregulation of vascular endothelial growth factor receptor (VEGFR), and increase in tumor-specific tissue—also contribute to the lifelong incidence increase risk of vascular events.²

However, treatment of a VTE with therapeutic anticoagulation is always balanced against the risk of hemorrhage. Though intracranial hemorrhage (ICH) is perhaps the most feared complication from anticoagulation, gastrointestinal bleeding—the risk of which may be confounded in patients with chronic steroids at risk for ulcer formation—hematoma from a fall, or worsening of any of these complications with chemotherapy-induced thrombocytopenia or anti-VEGF therapy, can make the decision to anticoagulate or not anticoagulate, a less than straightforward one.

In this edition of Neuro-Oncology, Jo et al³ take on this challenging question in a single institution, retrospective, matched cohort study of therapeutic anticoagulation as risk factor for ICH. They matched 110 glioma patients with VTE, 40% treated with low-molecular-weight heparin (LMWH) and 10% without anticoagulation, and with 110 glioma patients without VTE, matched on age (±5 years), histology, extent of resection, and sex. They found cumulative risk at 1 year for development of ICH in VTE group treated with LMWH of 17% (95% CI 0.10-0.26) vs 9% (95% CI 0.01-0.26) in VTE patients not treated with anticoagulation vs 13% (95% CI 0.07-0.20) of patients without VTE, and no difference in median survival. Importantly, there was an independent, blinded imaging review by neuroradiologist who graded ICH by lesional volume, most of which were <30 cm³, and petechial hemorrhage seen only on susceptibility imaging were appropriately excluded given they were unlikely to be clinically significant. However, the incidence should also be interpreted with caution, given that imaging frequency may have varied between groups. The median time from diagnosis to development of VTE was 3.3-5 months, as expected, given the higher risk in the postoperative setting, and median time from start of anticoagulation to ICH of 8.8 weeks in VTE patients with treated LMWH vs 3.7 weeks’ non-AC VTE group.

As a neuro-oncologist who has been paged on more than one occasion in the middle of the night by an ED physician checking to see if it is safe to anticoagulate a glioblastoma patient who presented with a deep vein thrombosis (DVT) or pulmonary embolism (PE), does this mean I’ll get more hours of sleep on call? Well, maybe. Given the unlikelihood of a prospective, randomized study to address this question, these results may probably as good as we’re going to get in addressing cumulative risk of developing ICH on LMWH. However, there are still clinical questions left unanswered.

These data should be reviewed in the context of other studies, which did find increased risk of hemorrhage in anticoagulating for VTE, including a recent meta-analysis that did identify increased risk.⁴ Particularly, data from Mantia et al,⁵ who found the PANWARDS (platelets, albumin, no congestive heart failure, warfarin, age, race, diastolic blood pressure, stroke) risk score to predict major ICHs on enoxaparin occurred with score ≥25 with a 100% sensitivity (95% CI 63% to 100%) and 40% specificity (95% CI 25% to 56%). A recently published analysis in brain metastasis also found higher risk of ICH (HR 1.80, 95% CI 1.01-3.22, P = .05) with anticoagulation for VTE (including warfarin, LMWH, and direct oral anticoagulants [DOACs]), particularly in patients with prior history (HR 2.20 vs 0.68, respectively, P_interaction < .001) or in melanoma (HR 6.46 vs 1.36, respectively, P_interaction = .02).⁶ These studies, and others referenced in the manuscript highlight the variability and severity of ICH in brain tumor populations, and lack of algorithms for treatment.

The National Comprehensive Cancer Network (NCCN) guidelines for Cancer-Associated Venous Thromboembolic Disease acknowledge brain tumor patients as high risk for developing VTE,⁷ but site only brain metastases as relative contraindication to anticoagulation.⁷ The guidelines recommend treatment for at least 3 months or “as long as active cancer or cancer therapy.” Of course, for high-grade gliomas (HGG) the cancer risk does not go away and factors such as immobility or treatment with anti-angiogenic therapy may increase the risk of recurrent VTE. However, this must be balanced with tumor progression leading to fall risk, chemotherapy-induced thrombocytopenia, anti-angiogenic agents, need for recurrent surgery, or long-term post-radiation vascular changes, which may put patients on long-term anticoagulation at higher risk for ICH. Jo et al,³ provided a 1-year cumulative risk, but acknowledge they did not address risk past 1 year, or comment on incidence of hemorrhage risk other than ICH.

An important elephant in the room that remains is the decision process on what anticoagulant to use. The NCCN guidelines support either DOACs or LMWH above warfarin.⁷ The population presented from Jo et al³ spans from 2005 to 2016 and DOACs received FDA approval in 2010 VTE, likely limiting their ability to comment on risk of ICH with DOAC, though this will be an important question going forward. While patients prefer oral agents over injections, the half-life of these medications is much longer and, therefore, poses a higher risk in managing ICH or other hemorrhage.

Many questions in the management of VTE for HGG patients remain unknown. The authors, in fact, start their secondary introductory paragraph with: “There are no established guidelines on the treatment of VTE in HGG patients” for the field.³ Should hemodynamically stable patients be just started on DOACs? Or LMWH and transitioned to DOACs? What about hemodynamically unstable patients? The role for heparin drip? Should brain imaging be done at some interval once starting anticoagulation? Is there a role for repeat ultrasounds or chest imaging after a certain time to discuss stopping anticoagulation? As we work to provide practice standards to our community—that is, seizure management in glioma patients⁸ and response assessments for MRIs and neurologic exams—let us also consider this important topic that affects the lives of so many of our patients, another call to action.