- Email: [email protected]
Recurrent venous thromboembolism in glioblastoma
Thrombosis Research 137 (2016) 184–188
Contents lists available at ScienceDirect
Thrombosis Research journal homepage: www.elsevier.com/locate/thromres
Full Length Article
Recurrent venous thromboembolism in glioblastoma Natasha Catherine Edwin a,1, Michael N. Khoury b,2, Davendra Sohal c, Keith R. McCrae c, Manmeet S. Ahluwalia b, Alok A. Khorana c,⁎ a b c
Division of Hematology and Medical Oncology, Washington University in St Louis School of Medicine, 660 S Euclid Avenue, Campus Box 8056, St Louis, MO 63110, USA Burkhardt Brain Tumor and Neuro Oncology Center, Neurological Institute Cleveland Clinic, 9500 Euclid Avenue, S73 Cleveland, OH 44195, USA Department of Hematology and Medical Oncology, Taussig Cancer Institute Cleveland Clinic, 9500 Euclid Avenue, R35 Cleveland, OH 44195, USA
a r t i c l e
i n f o
Article history: Received 14 July 2015 Received in revised form 17 November 2015 Accepted 19 November 2015 Available online 22 November 2015 Keywords: Deep vein thrombosis Cancer associated thrombosis Glioblastoma multiforme Pulmonary embolism Venous thromboembolism
a b s t r a c t Background: Patients with glioblastoma (GBM) are at increased risk of initial and recurrent venous thromboembolism (VTE) but rates of recurrence and real-world treatment choices are incompletely understood. Objectives: We aim to describe the treatment of incident VTE, report incidence and risk factors for recurrence. Patients/methods: We conducted a retrospective cohort study of consecutive Cleveland Clinic patients with GBM presenting with objectively diagnosed deep vein thrombosis (DVT) or pulmonary embolism (PE) from 2007 to 2013 with at least 6-month follow-up. We collected information on patient demographics, VTE incidence, treatment and recurrence. Data were analyzed using multivariate logistic regression analysis. Results: Of 450 patients with GBM, 145 (32.2%) developed VTE and comprised the study population. Of these, 11 (7.6%) experienced PE, 117 (80.7%) had DVT and 16 (11%) had DVT as well as PE. Fifty five (37.9%) VTE events occurred in the first 30 post-operative days and 56 (38.6%) during chemotherapy. Thirty one (21.4%) patients were untreated. Treatments included enoxaparin (N = 36, 24.8%), warfarin (15, 10.3%) or vena cava filters either alone (N = 39, 26.9%) or in combination with anticoagulation (N = 21, 14.5%). Recurrent VTE occurred in 39 patients (26.9%).In multivariate analysis, lack of long term anticoagulation (HR 11.2, CI 1.5–86.3, p b 0.05) and the presence of second primary malignancy (HR 3.69, CI 1.2–11.1, p b 0.05) were significantly associated with recurrent VTE. Conclusion: VTE and recurrent VTE are highly prevalent throughout the disease course among patients with GBM. Long term anticoagulation is associated with reduced risk of recurrent VTE but is often not utilized. © 2015 Elsevier Ltd. All rights reserved.
1. Introduction Venous thromboembolism (VTE) is a common complication of glioblastoma (GBM) [1]. The incidence of symptomatic VTE among patients with GBM ranges between 3 and 60% depending on the methods of detection and modalities of thromboprophylaxis employed [2,3]. A prospective randomized control trial of thromboprophylaxis in the setting of high grade glioma reported 17% incidence of VTE in the placebo arm within the first 6 months of diagnosis [4]. While the risk of thrombosis is highest post operatively, the risk of VTE remains elevated throughout the course of the disease with 1.5–2% risk of events per month of survival [5]. While the incidence of recurrent thrombosis ranges between 11 and 17%, risk factors for recurrence are incompletely understood [6,7]. ⁎ Corresponding author at: 9500 Euclid Avenue, Cleveland, OH 44195, USA. E-mail address: [email protected] (A.A. Khorana). Department of Hematology and Medical Oncology, Washington University in St. Louis School of Medicine, 660 S Euclid Avenue, St. Louis, Missouri 63108, USA. 2 Department of Neuro Oncology, H Lee Moffitt Cancer Center & Research Institute – Tampa FL, Department of Oncological Sciences at University of South Florida, 12902 Magnolia Drive, Tampa, Florida 33612, USA. 1
http://dx.doi.org/10.1016/j.thromres.2015.11.027 0049-3848/© 2015 Elsevier Ltd. All rights reserved.
There is a wide range of therapeutic options for patients with VTE in the setting of GBM. The American Society of Clinical Oncology (ASCO) guidelines recommend administration of anticoagulation with careful monitoring given the known risk of bleeding, if the risk of recurrent or progressive thrombosis is greater than that of hemorrhage [8]. The glioma prophylaxis study cited above reported a 5.1% incidence of intracranial bleeding in the arm receiving low molecular weight heparin (LMWH) at prophylactic doses [4]. With a relative scarcity of good evidence, clinicians are uncertain about how best to balance risks and benefits of anticoagulation in this setting. In this study, we describe our experience with recurrent VTE in GBM in the contemporary era of antiangiogenic therapy. In this, the largest such series to date in the literature, we describe the incidence, profile the characteristics of patients with VTE and also compare modalities employed for prophylaxis and treatment. We aim to characterize the incidence and risk factors for recurrent VTE in the setting of GBM. 2. Materials and methods The Cleveland Clinic is an academic medical center providing patient care in a nonprofit group practice setting. It is among the five largest
N.C. Edwin et al. / Thrombosis Research 137 (2016) 184–188
185
of patients in our study were not managed by the Thrombosis service and we did not use any database maintained or operated by this service. Regardless of primary team, all thrombotic events would be documented on the electronic medical record. It is possible but unlikely that patients receiving care at the Cleveland Clinic would be managed for thrombotic events at another center outside of the Cleveland Clinic network. In order to ensure capture of all VTE events, we excluded patients who did not maintain a minimum of 6 months follow-up at Cleveland Clinic. Patients with spinal glioblastoma were not included. Patient characteristics were summarized using frequencies for categorical variables and by using medians and ranges for continuous variables. Categorical variables were compared using the chi square test. Continuous variables were compared using Student's t test. Variables that were found to be significantly associated with VTE were fitted into a multivariate logistic regression model undertaken to determine the contribution of individual variables to the overall risk. Analyses were performed using MedCalc Statistical Software version 14.8.1 (MedCalc Software, Ostend, Belgium; http://www.medcalc.org; 2014). 3. Results
Fig. 1. Flow sheet of patients with VTE in the setting of GBM.
group practices in the US. The Rose Ella Burkhardt Brain Tumor and Neuro-Oncology center records over 8000 patient visits and 900 surgeries each year. Cleveland Clinic pathology records were queried to obtain medical record numbers of consecutive adult patients presenting between 2007 and 2013 with a histologic diagnosis of GBM. We reviewed individual patient electronic medical records for each identified patient specifically reviewing radiology reports for evidence of documented VTE. VTE events included deep vein thrombosis (DVT) confirmed by extremity ultrasonography and pulmonary embolism (PE) confirmed by computerized tomography (CT) and cerebral sinus thrombosis as detected by magnetic resonance venography (MRV). The study cohort comprised all patients with confirmed VTE (Fig. 1). Variables collected included demographic information and disease specific information such as type of surgery, IVC filter placement etc. Treatment history including history of anticoagulant use and thromboprophylaxis was obtained through review of the prescription history and medication administration record. We also collected data regarding objectively diagnosed recurrent VTE. We defined recurrent thrombosis as new deep vein thrombosis or pulmonary embolism at a site distinct from the index VTE that developed after the index thrombosis. Data was extracted by NE, a resident physician, through direct review of the patients' electronic medical record for inpatient and ambulatory visits. While the Cleveland Clinic has a Thrombosis service, clinical care of patients with GBM is provided by the Oncology service. The majority
A total of 483 patients had been reviewed at our center. We excluded 33 of these based on the fact that they presented only for a single visit and obtained ongoing care elsewhere. Of 450 GBM patients, 145 (32.2%) developed VTE and formed the study population. Baseline characteristics of the patients are described in Table 1. The age of the patients ranged from 39 to 94 years with a median of 67 years. Eightyfive (58.6%) patients were male. Second primary cancers were seen in 16 patients (11%). All of these were diagnosed prior to diagnosis of GBM (range 5–20 years, median 7 years) Five patients had a history of prostate cancer, one patient had a history of prostate cancer as well as sarcoma and one patient had a history of dermatofibrosarcoma. Two patients had previously diagnosed breast cancer and one had previously treated breast cancer along with anaplastic T cell lymphoma. Three patients had history of renal cell cancer. One patient had a history of colon cancer and two others had head and neck malignancies. 3.1. Characteristics of VTE events Of 145 patients with confirmed VTE, 11 (7.6%) experienced PE and 117 (80.7%) developed DVT (Table 2). Thirteen patients (8.9%) were known to have VTE prior to the diagnosis of GBM. Sixty six (45.5%) of Table 1 Baseline characteristics of patients with VTE in the setting of GBM. VTE — venous thromboembolism, GBM — glioblastoma. Characteristics
Median (range)
N (%)
Age (years) Body mass index Gender Male Female Comorbid illnesses Chronic kidney disease Chronic obstructive lung disease Congestive heart failure Smoking Second primary malignancy Tumor size (largest dimension in cm) Karnofsky performance status Type of surgery Biopsy Near total resection Subtotal resection Gross total resection
67 (39–94) 29 (15.6–45.1) –
–
85 (58.6) 60 (41.4) –
– 4 (0.9–8.5) 80 (10–100) –
1 (0.7) 7 (4.8) 2 (1.4) 2 (1.4) 16 (11) –
56 (38.6) 12 (8.3) 43 (29.7) 34 (23.4)
186
N.C. Edwin et al. / Thrombosis Research 137 (2016) 184–188 Table 2 Characteristics of VTE. VTE — venous thromboembolism, GBM — glioblastoma, VEGF — vascular endothelial growth factor. Characteristics Setting: VTE prior to diagnosis of GBM Perioperative VTE Inpatient VTE Chemotherapy at time of VTE VEGF inhibitor at time of VTE Location of VTE: DVT Proximal Distal Proximal and distal Upper extremity Upper and lower extremity Pulmonary embolism DVT and Pulmonary embolism Dural sinus thrombosis
Table 4 Comparison between patients with and without recurrent VTE in the setting of GBM. VTE — venous thromboembolism, GBM — glioblastoma, VEGF — vascular endothelial growth factor, IVC — inferior vena cava.
N (%) 13 (8.9) 55 (37.9) 62 (42.8) 56 (38.6) 14 (9.6) 117 (80.7) 19 66 24 6 2 11 (7.6) 16 (11) 1 (0.7)
117 patients with DVT had only distal lower extremity DVT whereas the rest developed either proximal lower extremity clots or upper extremity VTE. Sixteen (11%) patients developed both DVT and PE. One (0.7%) patient had a dural sinus thrombosis. Fifty-five (37.9%) patients with VTE were noted to develop events in the 30 day perioperative period while 56 (38.6%) developed VTE while receiving chemotherapy. Sixty-two (42.8%) patients were diagnosed to have a VTE while hospitalized. Of these, 11 (17.7%) received pharmacologic thromboprophylaxis. 3.2. Treatment of VTE VTE treatment approaches included IVC filters, anticoagulation or a combination of both modalities. (Table 3) Thirty nine (26.9%) patients received IVC filters alone. Twenty one (14.5%) patients received IVC filters with anticoagulation. Thirty-one patients (21.4%) were not treated. Most of these (N = 20, 64.5%) had only distal lower extremity DVT and were followed up with serial ultrasounds. Four patients had concomitant intracranial bleeds and were therefore not anticoagulated. Fifty four (37.2%) patients were only administered anticoagulation. Of the 75 patients receiving anticoagulation (alone or in the presence of IVC filters), 54 (72%) were treated with heparin or LMWH, 20 (26.7%) patients were treated with warfarin and one was treated with dabigatran (1.4%).
Recurrent VTE No recurrent VTE p-Value N Body mass index (mean) Age (mean) Gender Female Blood group (N = 140) A and AB B and O VTE prior to diagnosis of GBM Use of VEGF inhibitors Antecedent cancer IVC filter placement Use of anticoagulation for index VTE Lifelong anticoagulation
39 28.6 63.2
106 29.7 64.1
– 0.226 0.783
12 (30.8)
48 (45.3)
0.166 0.06
20 19 5 (12.8) 9 (23.1) 9 (23.1) 20 (51.3) 16 (41) 1 (2.6)
61 40 8 (7.5) 26 (24.5) 7 (6.6) 40 (37.7) 59 (55.7) 26 (24.5)
0.52 0.77 0.0012 0.20 0.17 0.005
one of the three had a concomitant PE. One patient developed bilateral extensive DVT, two patients had distal lower limb clots and one patient had a PE. Three others developed mechanical complications of filters necessitating their removal/replacement. The incidence of recurrent VTE among patients treated with anticoagulation was 21.3% (16/75). Specifically, the incidence of recurrent VTE among patients treated with anticoagulation alone was 18.5% (10/54) and 28.6% among those treated with IVC filters with anticoagulation (6/21). At the time of recurrent VTE, most patients were either not anticoagulated (N = 36), had sub therapeutic INR on warfarin (N = 2) or were receiving low dose subcutaneous heparin (N = 1). Twenty seven of the patients treated with anticoagulation received it lifelong (or till last follow up) and only one (3.7%) of them had a recurrent VTE. There was nosignificant association identified between recurrent VTE and age, BMI, gender, blood group, VEGF inhibitor therapy, VTE (prior to diagnosis of GBM) and modality of treatment for index VTE. Multivariate analysis demonstrated a significant association between patients with second primary malignancies and recurrent VTE (HR 3.69, CI 1.2–11.1, p b 0.05). Risk of recurrent VTE was lower in patients receiving long term anticoagulation (HR 11.2, CI 1.5–86.3, p b 0.05). (Table 5) The median survival of patients with recurrent VTE was 173 days (95% CI 147–226 days) compared to 253 days (95% CI 123– 331 days). This difference was however, not statistically significant.
4. Discussion 3.3. Recurrent VTE Thirty-nine patients (26.9%) developed recurrent VTE (Table 4). Of the sixty patients treated with IVC filters, twenty developed recurrent VTE (30%).Three of these developed thrombus within the filter and Table 3 Treatment of VTE. VTE — venous thromboembolism, IVC — inferior vena cava. Treatment modality
N (%)
IVC filter alone IVC filter with AC Anticoagulation alone Low molecular weight heparin Warfarin Heparin Target specific oral anticoagulant None Method of anticoagulation — total Low molecular weight heparin/heparin Warfarin Target specific oral anticoagulant
39 (26.9) 21 (14.5) 54 (37.2) 36 (24.8) 15 (10.3) 2 (1.4) 1 (0.7) 31 (21.4) 75 54 (72) 20 (26.7) 1 (1.3)
In this, the largest single institution series of patients with recurrent VTE in GBM published to date, we report a prevalence of primary VTE of 32.2% and recurrent VTE at 26.9%. This is consistent with contemporaneous cohorts which describe incidence of symptomatic of VTE ranging from 17 to 34% [9]. The majority of primary VTE occurred in the ambulatory setting. While the majority of patients were treated with anticoagulation (75, 51.7%), 21 (14.5%) of these received an IVC filter as well. In addition, thirty-nine patients had IVC filters placed in the absence of systemic anticoagulation. We found that patients treated with long term anticoagulation had a significant lower risk of recurrent VTE. We also
Table 5 Significant predictive factors for VTE in logistic regression analysis. Adjusted for gender, p b 0.05 p-values are 2-sided and derived by t-test for continuous and chi-square for categorical variables.
Lack of long term anticoagulation History of second primary
HR
95% confidence intervals
11.2 3.69
1.45–86.27 1.23–11.07
N.C. Edwin et al. / Thrombosis Research 137 (2016) 184–188
identified a significant increased risk of recurrent VTE among patients with second primary malignancies. The prevalence of VTE in our study population is consistent with the published literature, which describes a risk comparable to that of patients with pancreatic cancer [6]. Notably, the risk of VTE was present throughout the clinical course and not merely in the perioperative period. Several factors contribute to the hypercoagulable state in GBM. Most patients with GBM tend to have reduced mobility or limb paresis [5]. Poor functional status is an established risk factor for VTE in this population [10]. Treatment options for GBM include steroids and osmotically active materials which are thrombogenic [6]. Use of steroids has been established to be a risk factor for VTE in GBM [10]. Chronic activation of the coagulation system — both locally and systemically — is a known characteristic of GBM. Histologically, GBM is characterized by intratumoral thrombosis. This results in tissue hypoxia leading to active angiogenesis and tissue factor expression. Circulating tissue factor and tissue factor bearing micro particles (MP) have been implicated in hypercoagulability associated with malignancies [11]. Prospective studies in GBM have described elevated levels of circulating microparticles preoperatively and a week after surgery. A modest but statistically significant correlation exists between the increase in MP levels above the upper limit of normal and the incidence of VTE in this population [12]. GBM itself is known to alter the nature of the tumor–vascular interface thereby increasing the contact between procoagulant tumor and the circulation [11]. Anticoagulation remains an underutilized therapeutic modality in patients with GBM, likely due to concerns regarding risk of potentially catastrophic intracranial bleeding [5]. In a retrospective cohort study of patients with VTE in the setting of malignant glioma from 1977 to 1986, 61.1% patients received warfarin with no documented ICH [13]. Nearly 30 years later, the proportion of patients receiving anticoagulation remained 60.9% [7]. In our study, just over half of the population received systemic anticoagulation. Concerns over medication compliance and drug interactions — particularly between warfarin and antiepileptic agents may further limit the prescription of oral anticoagulants in this population. Use of vascular endothelial growth factor (VEGF) inhibitors adds an additional layer of complexity as these agents are known to promote both thrombosis and hemorrhage. In our study, we did not observe any difference between the incidence of recurrent VTE within the groups of patients treated with bevacizumab and those that were not, although we were underpowered to detect small differences in risk. A large number of patients with VTE receive IVC filters as the sole modality of treatment. Given the high failure rate of IVC filters without decreased incidence of ICH in this population, ASCO guidelines do not support routine usage [8]. A contemporaneous retrospective study from MD Anderson report that 32.8% patients with VTE in the setting of GBM had IVC filters placed [10]. This is consistent with our report, where 41.4% patients received IVC filters. The incidence of recurrent VTE in patients with IVC filters in our population was 11.7% which is lower than that described in older CNS neoplasm studies which report incidences as high as 40–50% [14,15]. More modern cohorts of patients with malignancies and IVC filters report a 13.5% recurrence rate at 30 days [16]. This trend towards reduction in the incidence of filter associated thrombosis with time might be explained by the improvements in techniques and type of filters. Notably, the risk of recurrence remains over 10% reinforcing the fact that filters do not, by themselves, alter the procoagulant milieu. Secondary prophylaxis against VTE requires a careful assessment of risk factors for recurrent VTE. The only risk assessment model developed for the assessment of recurrent thrombotic risk in the general oncology population, the Ottawa score identified female gender and previous VTE as significant predictors [17]. In our study, we did not identify any of these as significant risk factors for recurrent thrombosis. We found a significantly increased risk of recurrent VTE among patients with non-cutaneous second primary malignancies prior to the
187
diagnosis of GBM. Our findings might be partially explained by the fact that hypercoagulability associated with intra-abdominal malignancies persists in the long term (N9 months) after complete surgical resection [18]. We found that lifelong anticoagulation (defined as anticoagulation continued till the time of last follow up or death) was associated with a reduced risk of recurrent VTE. This reinforces the need for development of a system of risk stratification for judicious prescription of anticoagulation. Our study is limited by its retrospective nature and that it involves patients seen at a single institution. However, our incidence rates are similar to those previously reported in the literature suggesting generalizability of the data. Screening for VTE was not undertaken. This could potentially lead to an underestimation of recurrent events. However, this provides a more real-world description of the incidence and prevalence of initial and recurrent events. While anticoagulation in general and lifelong anticoagulation in particular, is protective against recurrent VTE; we lack assessment tools to identify patients at highest risk of VTE in the setting of GBM. Despite advances in therapeutic anticoagulation, IVC filters continue to remain an important part of the therapeutic armamentarium. Additional studies would be needed to clarify their role in management of VTE. While direct oral anticoagulants show promise in the treatment of VTE, their place in the world of cancer associated thrombosis is still under investigation. Funding Dr. Khorana is supported by the Sondra and Stephen Hardis Chair in Oncology Research and the Scott Hamilton CARES Initiative. References [1] J.R. Perry, Anticoagulation of malignant glioma patients in the era of novel antiangiogenic agents, Curr. Opin. Neurol. 23 (2010) 592–596. [2] L.C. Marras, W.H. Geerts, J.R. Perry, The risk of venous thromboembolism is increased throughout the course of malignant glioma, Cancer 89 (2000) 640–646. [3] D.E. Gerber, S.A. Grossman, M.B. Streiff, Management of venous thromboembolism in patients with primary and metastatic brain tumors, J. Clin. Oncol. 24 (2006) 1310–1318. [4] J.R. Perry, J.A. Julian, N.J. Lapierriere, W. Geerts, G. Agnelli, L.R. Rogers, M.G. Malkin, R. Sawaya, R. Baker, A. Falanga, S. Parpia, T. Finch, M.N. Levine, PRODIGE: a randomized placebo-controlled trial of dalteparin low-molecular-weight heparin thromboprophyloaxis in patients with newly diagnosed malignant glioma, J. Thromb. Haemost. 8 (2010) 1959–1965. [5] J.R. Perry, Thromboembolic disease in patients with high-grade glioma, Neuro-Oncology 14 (2012) iv73–iv80. [6] T.J. Semrad, R. O'Donnell, T. Wun, H. Chew, D. Harvey, H. Zhou, R.H. White, Epidemiology of venous thromboembolism in 9489 patients with malignant glioma, J. Neurosurg. 106 (2007) 601–608. [7] E. Pan, J.S. Tsai, S.B. Mitchell, Retrospective study of venous thromboembolic and intraceerebral hemorrhagic events in glioblastoma patients, Anticancer Res. 29 (2009) 4309–4314. [8] G.H. Lyman, A.A. Khorana, N.M. Kuderer, A.Y. Lee, J.I. Arcelus, E.P. Balaban, J.M. Clarke, C.R. Flowers, C.W. Francis, L.E. Gates, A.K. Kakkar, N.S. Key, M.N. Levine, H.A. Liebman, M.A. Tempero, S.L. Wong, M.R. Somerfield, A. Falanga, Venous thromboembolism prophylaxis and treatment in patients with cancer: American society of clinical oncology clinical practice guideline update, J. Clin. Oncol. 31 (2013) 2189–2204. [9] M.B. Streiff, X. Ye, T.S. Kickler, S. Desideri, J. Jani, J. Fisher, S.A. Grossman, A prospective multicenter study of venous thromboembolism in patients with newly diagnosed high grade glioma: hazard rate and risk factors, J. Neuro-Oncol. (2015)http://dx.doi.org/10.1007/s11060-015-1840-z (Advance online publication). [10] S. Yust-Katz, J.J. Mandel, J. Wu, Y. Yuan, C.C. Webre, T.A. Pawar, H.S. Lhadha, M.R. Gilbert, T.S. Armstrong, Venous thromboemboliam and glioblastoma, J. NeuroOncol. 124 (2015) 87–94. [11] N. Magnus, E. D'Asti, D. Garnier, B. Meehan, J. Rak, Brain neoplasms and coagulation, Semin. Thromb. Hemost. 39 (2013) 881–895. [12] M.T. Sartori, A.D. Puppa, A. Ballin, G. Saggiorato, D. Bernardi, A. Padoan, R. Scienza, D. d'Avella, G. Cella, Prothrombotic state in glioblastoma multiforme: an evaluation of the procoagulant activity of circulating microparticles, J. Neuro-Oncol. 104 (2011) 225–231. [13] A.K. Choucair, P. Silver, V.A. Levin, Risk of intracranial hemorrhage in glioma patients receiving anticoagulant therapy for venous thromboembolism, J. Neurosurg. 66 (1987) 357–358. [14] D. Schiff, L.M. De Angelis, Therapy of venous thromboembolism in patients with brain metastases, Cancer 73 (1994) 493–498.
188
N.C. Edwin et al. / Thrombosis Research 137 (2016) 184–188
[15] J.M. Levin, D. Schiff, J.S. Loeffler, H.A. Fine, P.M. Black, P.Y. Wen, Complications of therapy for venous thromboembolic disease in patients with brain tumors, Neurology 43 (1993) 1111–1114. [16] A. Narayan, K. Hong, M. Streiff, R. Shinohara, C. Frangakis, J. Coresh, H.S. Kim, The impact of cancer on the clinical outcome of patients after inferior vena cava filter placement: a retrospective cohort study, Am. J. Clin. Oncol. (2014) (epub ahead of print). [17] M.L. Louzada, M. Carrier, A. Lazo-Langner, V. Dao, M.J. Kovacs, T.O. Ramsy, M.A. Rodger, J. Zhang, A.Y. Lee, G. Meyer, P.S. Wells, Development of a clinical risk
prediction rule for risk stratification of recurrent venous thromboembolism in patients with cancer associated venous thromboembolism, Circulation 126 (2012) 448–454. [18] R.M. Van Haren, E.J. Valle, C.M. Thorson, J.A. Guarch, J.M. Jouria, D.M. Andrews, D. Sleeman, J.U. Levi, A.S. Livingstone, K.G. Proctor, Long-term coagulation changes after resection of thoracoabdominal malignancies, J. Am. Coll. Surg. 21 (2014) 846–855.
Contents lists available at ScienceDirect
Thrombosis Research journal homepage: www.elsevier.com/locate/thromres
Full Length Article
Recurrent venous thromboembolism in glioblastoma Natasha Catherine Edwin a,1, Michael N. Khoury b,2, Davendra Sohal c, Keith R. McCrae c, Manmeet S. Ahluwalia b, Alok A. Khorana c,⁎ a b c
Division of Hematology and Medical Oncology, Washington University in St Louis School of Medicine, 660 S Euclid Avenue, Campus Box 8056, St Louis, MO 63110, USA Burkhardt Brain Tumor and Neuro Oncology Center, Neurological Institute Cleveland Clinic, 9500 Euclid Avenue, S73 Cleveland, OH 44195, USA Department of Hematology and Medical Oncology, Taussig Cancer Institute Cleveland Clinic, 9500 Euclid Avenue, R35 Cleveland, OH 44195, USA
a r t i c l e
i n f o
Article history: Received 14 July 2015 Received in revised form 17 November 2015 Accepted 19 November 2015 Available online 22 November 2015 Keywords: Deep vein thrombosis Cancer associated thrombosis Glioblastoma multiforme Pulmonary embolism Venous thromboembolism
a b s t r a c t Background: Patients with glioblastoma (GBM) are at increased risk of initial and recurrent venous thromboembolism (VTE) but rates of recurrence and real-world treatment choices are incompletely understood. Objectives: We aim to describe the treatment of incident VTE, report incidence and risk factors for recurrence. Patients/methods: We conducted a retrospective cohort study of consecutive Cleveland Clinic patients with GBM presenting with objectively diagnosed deep vein thrombosis (DVT) or pulmonary embolism (PE) from 2007 to 2013 with at least 6-month follow-up. We collected information on patient demographics, VTE incidence, treatment and recurrence. Data were analyzed using multivariate logistic regression analysis. Results: Of 450 patients with GBM, 145 (32.2%) developed VTE and comprised the study population. Of these, 11 (7.6%) experienced PE, 117 (80.7%) had DVT and 16 (11%) had DVT as well as PE. Fifty five (37.9%) VTE events occurred in the first 30 post-operative days and 56 (38.6%) during chemotherapy. Thirty one (21.4%) patients were untreated. Treatments included enoxaparin (N = 36, 24.8%), warfarin (15, 10.3%) or vena cava filters either alone (N = 39, 26.9%) or in combination with anticoagulation (N = 21, 14.5%). Recurrent VTE occurred in 39 patients (26.9%).In multivariate analysis, lack of long term anticoagulation (HR 11.2, CI 1.5–86.3, p b 0.05) and the presence of second primary malignancy (HR 3.69, CI 1.2–11.1, p b 0.05) were significantly associated with recurrent VTE. Conclusion: VTE and recurrent VTE are highly prevalent throughout the disease course among patients with GBM. Long term anticoagulation is associated with reduced risk of recurrent VTE but is often not utilized. © 2015 Elsevier Ltd. All rights reserved.
1. Introduction Venous thromboembolism (VTE) is a common complication of glioblastoma (GBM) [1]. The incidence of symptomatic VTE among patients with GBM ranges between 3 and 60% depending on the methods of detection and modalities of thromboprophylaxis employed [2,3]. A prospective randomized control trial of thromboprophylaxis in the setting of high grade glioma reported 17% incidence of VTE in the placebo arm within the first 6 months of diagnosis [4]. While the risk of thrombosis is highest post operatively, the risk of VTE remains elevated throughout the course of the disease with 1.5–2% risk of events per month of survival [5]. While the incidence of recurrent thrombosis ranges between 11 and 17%, risk factors for recurrence are incompletely understood [6,7]. ⁎ Corresponding author at: 9500 Euclid Avenue, Cleveland, OH 44195, USA. E-mail address: [email protected] (A.A. Khorana). Department of Hematology and Medical Oncology, Washington University in St. Louis School of Medicine, 660 S Euclid Avenue, St. Louis, Missouri 63108, USA. 2 Department of Neuro Oncology, H Lee Moffitt Cancer Center & Research Institute – Tampa FL, Department of Oncological Sciences at University of South Florida, 12902 Magnolia Drive, Tampa, Florida 33612, USA. 1
http://dx.doi.org/10.1016/j.thromres.2015.11.027 0049-3848/© 2015 Elsevier Ltd. All rights reserved.
There is a wide range of therapeutic options for patients with VTE in the setting of GBM. The American Society of Clinical Oncology (ASCO) guidelines recommend administration of anticoagulation with careful monitoring given the known risk of bleeding, if the risk of recurrent or progressive thrombosis is greater than that of hemorrhage [8]. The glioma prophylaxis study cited above reported a 5.1% incidence of intracranial bleeding in the arm receiving low molecular weight heparin (LMWH) at prophylactic doses [4]. With a relative scarcity of good evidence, clinicians are uncertain about how best to balance risks and benefits of anticoagulation in this setting. In this study, we describe our experience with recurrent VTE in GBM in the contemporary era of antiangiogenic therapy. In this, the largest such series to date in the literature, we describe the incidence, profile the characteristics of patients with VTE and also compare modalities employed for prophylaxis and treatment. We aim to characterize the incidence and risk factors for recurrent VTE in the setting of GBM. 2. Materials and methods The Cleveland Clinic is an academic medical center providing patient care in a nonprofit group practice setting. It is among the five largest
N.C. Edwin et al. / Thrombosis Research 137 (2016) 184–188
185
of patients in our study were not managed by the Thrombosis service and we did not use any database maintained or operated by this service. Regardless of primary team, all thrombotic events would be documented on the electronic medical record. It is possible but unlikely that patients receiving care at the Cleveland Clinic would be managed for thrombotic events at another center outside of the Cleveland Clinic network. In order to ensure capture of all VTE events, we excluded patients who did not maintain a minimum of 6 months follow-up at Cleveland Clinic. Patients with spinal glioblastoma were not included. Patient characteristics were summarized using frequencies for categorical variables and by using medians and ranges for continuous variables. Categorical variables were compared using the chi square test. Continuous variables were compared using Student's t test. Variables that were found to be significantly associated with VTE were fitted into a multivariate logistic regression model undertaken to determine the contribution of individual variables to the overall risk. Analyses were performed using MedCalc Statistical Software version 14.8.1 (MedCalc Software, Ostend, Belgium; http://www.medcalc.org; 2014). 3. Results
Fig. 1. Flow sheet of patients with VTE in the setting of GBM.
group practices in the US. The Rose Ella Burkhardt Brain Tumor and Neuro-Oncology center records over 8000 patient visits and 900 surgeries each year. Cleveland Clinic pathology records were queried to obtain medical record numbers of consecutive adult patients presenting between 2007 and 2013 with a histologic diagnosis of GBM. We reviewed individual patient electronic medical records for each identified patient specifically reviewing radiology reports for evidence of documented VTE. VTE events included deep vein thrombosis (DVT) confirmed by extremity ultrasonography and pulmonary embolism (PE) confirmed by computerized tomography (CT) and cerebral sinus thrombosis as detected by magnetic resonance venography (MRV). The study cohort comprised all patients with confirmed VTE (Fig. 1). Variables collected included demographic information and disease specific information such as type of surgery, IVC filter placement etc. Treatment history including history of anticoagulant use and thromboprophylaxis was obtained through review of the prescription history and medication administration record. We also collected data regarding objectively diagnosed recurrent VTE. We defined recurrent thrombosis as new deep vein thrombosis or pulmonary embolism at a site distinct from the index VTE that developed after the index thrombosis. Data was extracted by NE, a resident physician, through direct review of the patients' electronic medical record for inpatient and ambulatory visits. While the Cleveland Clinic has a Thrombosis service, clinical care of patients with GBM is provided by the Oncology service. The majority
A total of 483 patients had been reviewed at our center. We excluded 33 of these based on the fact that they presented only for a single visit and obtained ongoing care elsewhere. Of 450 GBM patients, 145 (32.2%) developed VTE and formed the study population. Baseline characteristics of the patients are described in Table 1. The age of the patients ranged from 39 to 94 years with a median of 67 years. Eightyfive (58.6%) patients were male. Second primary cancers were seen in 16 patients (11%). All of these were diagnosed prior to diagnosis of GBM (range 5–20 years, median 7 years) Five patients had a history of prostate cancer, one patient had a history of prostate cancer as well as sarcoma and one patient had a history of dermatofibrosarcoma. Two patients had previously diagnosed breast cancer and one had previously treated breast cancer along with anaplastic T cell lymphoma. Three patients had history of renal cell cancer. One patient had a history of colon cancer and two others had head and neck malignancies. 3.1. Characteristics of VTE events Of 145 patients with confirmed VTE, 11 (7.6%) experienced PE and 117 (80.7%) developed DVT (Table 2). Thirteen patients (8.9%) were known to have VTE prior to the diagnosis of GBM. Sixty six (45.5%) of Table 1 Baseline characteristics of patients with VTE in the setting of GBM. VTE — venous thromboembolism, GBM — glioblastoma. Characteristics
Median (range)
N (%)
Age (years) Body mass index Gender Male Female Comorbid illnesses Chronic kidney disease Chronic obstructive lung disease Congestive heart failure Smoking Second primary malignancy Tumor size (largest dimension in cm) Karnofsky performance status Type of surgery Biopsy Near total resection Subtotal resection Gross total resection
67 (39–94) 29 (15.6–45.1) –
–
85 (58.6) 60 (41.4) –
– 4 (0.9–8.5) 80 (10–100) –
1 (0.7) 7 (4.8) 2 (1.4) 2 (1.4) 16 (11) –
56 (38.6) 12 (8.3) 43 (29.7) 34 (23.4)
186
N.C. Edwin et al. / Thrombosis Research 137 (2016) 184–188 Table 2 Characteristics of VTE. VTE — venous thromboembolism, GBM — glioblastoma, VEGF — vascular endothelial growth factor. Characteristics Setting: VTE prior to diagnosis of GBM Perioperative VTE Inpatient VTE Chemotherapy at time of VTE VEGF inhibitor at time of VTE Location of VTE: DVT Proximal Distal Proximal and distal Upper extremity Upper and lower extremity Pulmonary embolism DVT and Pulmonary embolism Dural sinus thrombosis
Table 4 Comparison between patients with and without recurrent VTE in the setting of GBM. VTE — venous thromboembolism, GBM — glioblastoma, VEGF — vascular endothelial growth factor, IVC — inferior vena cava.
N (%) 13 (8.9) 55 (37.9) 62 (42.8) 56 (38.6) 14 (9.6) 117 (80.7) 19 66 24 6 2 11 (7.6) 16 (11) 1 (0.7)
117 patients with DVT had only distal lower extremity DVT whereas the rest developed either proximal lower extremity clots or upper extremity VTE. Sixteen (11%) patients developed both DVT and PE. One (0.7%) patient had a dural sinus thrombosis. Fifty-five (37.9%) patients with VTE were noted to develop events in the 30 day perioperative period while 56 (38.6%) developed VTE while receiving chemotherapy. Sixty-two (42.8%) patients were diagnosed to have a VTE while hospitalized. Of these, 11 (17.7%) received pharmacologic thromboprophylaxis. 3.2. Treatment of VTE VTE treatment approaches included IVC filters, anticoagulation or a combination of both modalities. (Table 3) Thirty nine (26.9%) patients received IVC filters alone. Twenty one (14.5%) patients received IVC filters with anticoagulation. Thirty-one patients (21.4%) were not treated. Most of these (N = 20, 64.5%) had only distal lower extremity DVT and were followed up with serial ultrasounds. Four patients had concomitant intracranial bleeds and were therefore not anticoagulated. Fifty four (37.2%) patients were only administered anticoagulation. Of the 75 patients receiving anticoagulation (alone or in the presence of IVC filters), 54 (72%) were treated with heparin or LMWH, 20 (26.7%) patients were treated with warfarin and one was treated with dabigatran (1.4%).
Recurrent VTE No recurrent VTE p-Value N Body mass index (mean) Age (mean) Gender Female Blood group (N = 140) A and AB B and O VTE prior to diagnosis of GBM Use of VEGF inhibitors Antecedent cancer IVC filter placement Use of anticoagulation for index VTE Lifelong anticoagulation
39 28.6 63.2
106 29.7 64.1
– 0.226 0.783
12 (30.8)
48 (45.3)
0.166 0.06
20 19 5 (12.8) 9 (23.1) 9 (23.1) 20 (51.3) 16 (41) 1 (2.6)
61 40 8 (7.5) 26 (24.5) 7 (6.6) 40 (37.7) 59 (55.7) 26 (24.5)
0.52 0.77 0.0012 0.20 0.17 0.005
one of the three had a concomitant PE. One patient developed bilateral extensive DVT, two patients had distal lower limb clots and one patient had a PE. Three others developed mechanical complications of filters necessitating their removal/replacement. The incidence of recurrent VTE among patients treated with anticoagulation was 21.3% (16/75). Specifically, the incidence of recurrent VTE among patients treated with anticoagulation alone was 18.5% (10/54) and 28.6% among those treated with IVC filters with anticoagulation (6/21). At the time of recurrent VTE, most patients were either not anticoagulated (N = 36), had sub therapeutic INR on warfarin (N = 2) or were receiving low dose subcutaneous heparin (N = 1). Twenty seven of the patients treated with anticoagulation received it lifelong (or till last follow up) and only one (3.7%) of them had a recurrent VTE. There was nosignificant association identified between recurrent VTE and age, BMI, gender, blood group, VEGF inhibitor therapy, VTE (prior to diagnosis of GBM) and modality of treatment for index VTE. Multivariate analysis demonstrated a significant association between patients with second primary malignancies and recurrent VTE (HR 3.69, CI 1.2–11.1, p b 0.05). Risk of recurrent VTE was lower in patients receiving long term anticoagulation (HR 11.2, CI 1.5–86.3, p b 0.05). (Table 5) The median survival of patients with recurrent VTE was 173 days (95% CI 147–226 days) compared to 253 days (95% CI 123– 331 days). This difference was however, not statistically significant.
4. Discussion 3.3. Recurrent VTE Thirty-nine patients (26.9%) developed recurrent VTE (Table 4). Of the sixty patients treated with IVC filters, twenty developed recurrent VTE (30%).Three of these developed thrombus within the filter and Table 3 Treatment of VTE. VTE — venous thromboembolism, IVC — inferior vena cava. Treatment modality
N (%)
IVC filter alone IVC filter with AC Anticoagulation alone Low molecular weight heparin Warfarin Heparin Target specific oral anticoagulant None Method of anticoagulation — total Low molecular weight heparin/heparin Warfarin Target specific oral anticoagulant
39 (26.9) 21 (14.5) 54 (37.2) 36 (24.8) 15 (10.3) 2 (1.4) 1 (0.7) 31 (21.4) 75 54 (72) 20 (26.7) 1 (1.3)
In this, the largest single institution series of patients with recurrent VTE in GBM published to date, we report a prevalence of primary VTE of 32.2% and recurrent VTE at 26.9%. This is consistent with contemporaneous cohorts which describe incidence of symptomatic of VTE ranging from 17 to 34% [9]. The majority of primary VTE occurred in the ambulatory setting. While the majority of patients were treated with anticoagulation (75, 51.7%), 21 (14.5%) of these received an IVC filter as well. In addition, thirty-nine patients had IVC filters placed in the absence of systemic anticoagulation. We found that patients treated with long term anticoagulation had a significant lower risk of recurrent VTE. We also
Table 5 Significant predictive factors for VTE in logistic regression analysis. Adjusted for gender, p b 0.05 p-values are 2-sided and derived by t-test for continuous and chi-square for categorical variables.
Lack of long term anticoagulation History of second primary
HR
95% confidence intervals
11.2 3.69
1.45–86.27 1.23–11.07
N.C. Edwin et al. / Thrombosis Research 137 (2016) 184–188
identified a significant increased risk of recurrent VTE among patients with second primary malignancies. The prevalence of VTE in our study population is consistent with the published literature, which describes a risk comparable to that of patients with pancreatic cancer [6]. Notably, the risk of VTE was present throughout the clinical course and not merely in the perioperative period. Several factors contribute to the hypercoagulable state in GBM. Most patients with GBM tend to have reduced mobility or limb paresis [5]. Poor functional status is an established risk factor for VTE in this population [10]. Treatment options for GBM include steroids and osmotically active materials which are thrombogenic [6]. Use of steroids has been established to be a risk factor for VTE in GBM [10]. Chronic activation of the coagulation system — both locally and systemically — is a known characteristic of GBM. Histologically, GBM is characterized by intratumoral thrombosis. This results in tissue hypoxia leading to active angiogenesis and tissue factor expression. Circulating tissue factor and tissue factor bearing micro particles (MP) have been implicated in hypercoagulability associated with malignancies [11]. Prospective studies in GBM have described elevated levels of circulating microparticles preoperatively and a week after surgery. A modest but statistically significant correlation exists between the increase in MP levels above the upper limit of normal and the incidence of VTE in this population [12]. GBM itself is known to alter the nature of the tumor–vascular interface thereby increasing the contact between procoagulant tumor and the circulation [11]. Anticoagulation remains an underutilized therapeutic modality in patients with GBM, likely due to concerns regarding risk of potentially catastrophic intracranial bleeding [5]. In a retrospective cohort study of patients with VTE in the setting of malignant glioma from 1977 to 1986, 61.1% patients received warfarin with no documented ICH [13]. Nearly 30 years later, the proportion of patients receiving anticoagulation remained 60.9% [7]. In our study, just over half of the population received systemic anticoagulation. Concerns over medication compliance and drug interactions — particularly between warfarin and antiepileptic agents may further limit the prescription of oral anticoagulants in this population. Use of vascular endothelial growth factor (VEGF) inhibitors adds an additional layer of complexity as these agents are known to promote both thrombosis and hemorrhage. In our study, we did not observe any difference between the incidence of recurrent VTE within the groups of patients treated with bevacizumab and those that were not, although we were underpowered to detect small differences in risk. A large number of patients with VTE receive IVC filters as the sole modality of treatment. Given the high failure rate of IVC filters without decreased incidence of ICH in this population, ASCO guidelines do not support routine usage [8]. A contemporaneous retrospective study from MD Anderson report that 32.8% patients with VTE in the setting of GBM had IVC filters placed [10]. This is consistent with our report, where 41.4% patients received IVC filters. The incidence of recurrent VTE in patients with IVC filters in our population was 11.7% which is lower than that described in older CNS neoplasm studies which report incidences as high as 40–50% [14,15]. More modern cohorts of patients with malignancies and IVC filters report a 13.5% recurrence rate at 30 days [16]. This trend towards reduction in the incidence of filter associated thrombosis with time might be explained by the improvements in techniques and type of filters. Notably, the risk of recurrence remains over 10% reinforcing the fact that filters do not, by themselves, alter the procoagulant milieu. Secondary prophylaxis against VTE requires a careful assessment of risk factors for recurrent VTE. The only risk assessment model developed for the assessment of recurrent thrombotic risk in the general oncology population, the Ottawa score identified female gender and previous VTE as significant predictors [17]. In our study, we did not identify any of these as significant risk factors for recurrent thrombosis. We found a significantly increased risk of recurrent VTE among patients with non-cutaneous second primary malignancies prior to the
187
diagnosis of GBM. Our findings might be partially explained by the fact that hypercoagulability associated with intra-abdominal malignancies persists in the long term (N9 months) after complete surgical resection [18]. We found that lifelong anticoagulation (defined as anticoagulation continued till the time of last follow up or death) was associated with a reduced risk of recurrent VTE. This reinforces the need for development of a system of risk stratification for judicious prescription of anticoagulation. Our study is limited by its retrospective nature and that it involves patients seen at a single institution. However, our incidence rates are similar to those previously reported in the literature suggesting generalizability of the data. Screening for VTE was not undertaken. This could potentially lead to an underestimation of recurrent events. However, this provides a more real-world description of the incidence and prevalence of initial and recurrent events. While anticoagulation in general and lifelong anticoagulation in particular, is protective against recurrent VTE; we lack assessment tools to identify patients at highest risk of VTE in the setting of GBM. Despite advances in therapeutic anticoagulation, IVC filters continue to remain an important part of the therapeutic armamentarium. Additional studies would be needed to clarify their role in management of VTE. While direct oral anticoagulants show promise in the treatment of VTE, their place in the world of cancer associated thrombosis is still under investigation. Funding Dr. Khorana is supported by the Sondra and Stephen Hardis Chair in Oncology Research and the Scott Hamilton CARES Initiative. References [1] J.R. Perry, Anticoagulation of malignant glioma patients in the era of novel antiangiogenic agents, Curr. Opin. Neurol. 23 (2010) 592–596. [2] L.C. Marras, W.H. Geerts, J.R. Perry, The risk of venous thromboembolism is increased throughout the course of malignant glioma, Cancer 89 (2000) 640–646. [3] D.E. Gerber, S.A. Grossman, M.B. Streiff, Management of venous thromboembolism in patients with primary and metastatic brain tumors, J. Clin. Oncol. 24 (2006) 1310–1318. [4] J.R. Perry, J.A. Julian, N.J. Lapierriere, W. Geerts, G. Agnelli, L.R. Rogers, M.G. Malkin, R. Sawaya, R. Baker, A. Falanga, S. Parpia, T. Finch, M.N. Levine, PRODIGE: a randomized placebo-controlled trial of dalteparin low-molecular-weight heparin thromboprophyloaxis in patients with newly diagnosed malignant glioma, J. Thromb. Haemost. 8 (2010) 1959–1965. [5] J.R. Perry, Thromboembolic disease in patients with high-grade glioma, Neuro-Oncology 14 (2012) iv73–iv80. [6] T.J. Semrad, R. O'Donnell, T. Wun, H. Chew, D. Harvey, H. Zhou, R.H. White, Epidemiology of venous thromboembolism in 9489 patients with malignant glioma, J. Neurosurg. 106 (2007) 601–608. [7] E. Pan, J.S. Tsai, S.B. Mitchell, Retrospective study of venous thromboembolic and intraceerebral hemorrhagic events in glioblastoma patients, Anticancer Res. 29 (2009) 4309–4314. [8] G.H. Lyman, A.A. Khorana, N.M. Kuderer, A.Y. Lee, J.I. Arcelus, E.P. Balaban, J.M. Clarke, C.R. Flowers, C.W. Francis, L.E. Gates, A.K. Kakkar, N.S. Key, M.N. Levine, H.A. Liebman, M.A. Tempero, S.L. Wong, M.R. Somerfield, A. Falanga, Venous thromboembolism prophylaxis and treatment in patients with cancer: American society of clinical oncology clinical practice guideline update, J. Clin. Oncol. 31 (2013) 2189–2204. [9] M.B. Streiff, X. Ye, T.S. Kickler, S. Desideri, J. Jani, J. Fisher, S.A. Grossman, A prospective multicenter study of venous thromboembolism in patients with newly diagnosed high grade glioma: hazard rate and risk factors, J. Neuro-Oncol. (2015)http://dx.doi.org/10.1007/s11060-015-1840-z (Advance online publication). [10] S. Yust-Katz, J.J. Mandel, J. Wu, Y. Yuan, C.C. Webre, T.A. Pawar, H.S. Lhadha, M.R. Gilbert, T.S. Armstrong, Venous thromboemboliam and glioblastoma, J. NeuroOncol. 124 (2015) 87–94. [11] N. Magnus, E. D'Asti, D. Garnier, B. Meehan, J. Rak, Brain neoplasms and coagulation, Semin. Thromb. Hemost. 39 (2013) 881–895. [12] M.T. Sartori, A.D. Puppa, A. Ballin, G. Saggiorato, D. Bernardi, A. Padoan, R. Scienza, D. d'Avella, G. Cella, Prothrombotic state in glioblastoma multiforme: an evaluation of the procoagulant activity of circulating microparticles, J. Neuro-Oncol. 104 (2011) 225–231. [13] A.K. Choucair, P. Silver, V.A. Levin, Risk of intracranial hemorrhage in glioma patients receiving anticoagulant therapy for venous thromboembolism, J. Neurosurg. 66 (1987) 357–358. [14] D. Schiff, L.M. De Angelis, Therapy of venous thromboembolism in patients with brain metastases, Cancer 73 (1994) 493–498.
188
N.C. Edwin et al. / Thrombosis Research 137 (2016) 184–188
[15] J.M. Levin, D. Schiff, J.S. Loeffler, H.A. Fine, P.M. Black, P.Y. Wen, Complications of therapy for venous thromboembolic disease in patients with brain tumors, Neurology 43 (1993) 1111–1114. [16] A. Narayan, K. Hong, M. Streiff, R. Shinohara, C. Frangakis, J. Coresh, H.S. Kim, The impact of cancer on the clinical outcome of patients after inferior vena cava filter placement: a retrospective cohort study, Am. J. Clin. Oncol. (2014) (epub ahead of print). [17] M.L. Louzada, M. Carrier, A. Lazo-Langner, V. Dao, M.J. Kovacs, T.O. Ramsy, M.A. Rodger, J. Zhang, A.Y. Lee, G. Meyer, P.S. Wells, Development of a clinical risk
prediction rule for risk stratification of recurrent venous thromboembolism in patients with cancer associated venous thromboembolism, Circulation 126 (2012) 448–454. [18] R.M. Van Haren, E.J. Valle, C.M. Thorson, J.A. Guarch, J.M. Jouria, D.M. Andrews, D. Sleeman, J.U. Levi, A.S. Livingstone, K.G. Proctor, Long-term coagulation changes after resection of thoracoabdominal malignancies, J. Am. Coll. Surg. 21 (2014) 846–855.