1479P - Predictive factors for use of parenteral versus oral anticoagulants in the treatment of venous thromboembolism in patients with active cancer

Date 09 October 2016
Event ESMO 2016 Congress
Session Poster display
Topics Supportive Measures
Presenter Anja Katholing
Citation Annals of Oncology (2016) 27 (6): 497-521. 10.1093/annonc/mdw390
Authors A. Katholing1, K. Folkerts2, M. Bach3, C. Martinez4
  • 1Epidemiology, Institute for Epidemiology, Statistics and Informatics GmbH, 60388 Frankfurt - Frankfurt am Main/DE
  • 2Global Market Access Gheor Gm, Bayer Pharma AG, 42096 - Wuppertal/DE
  • 3Global Medical Affairs, Bayer Pharma AG, 13353 - Berlin/DE
  • 4Epidemiology, Institute for Epidemiology, Statistics and Informatics GmbH, 60388 - Frankfurt am Main/DE



In patients with active cancer, the risk of venous thromboembolism (VTE) is up to five times higher. Anticoagulants (ACs) reduce the risk of VTE. We explored predictors for parenteral versus oral ACs in patients with cancer-associated VTE (Ca-VTE).


Patient data were retrieved from general practices in England contributing to the Clinical Practice Research Datalink (CPRD) with additional hospital discharge and cause of death data. The study cohort consisted of patients with a specified cancer recorded within the 90 days before or after the first VTE between 2008 and 2014, and with a subsequent record of either LMWH or fondaparinux defined as parenteral ACs (PACs), of vitamin K antagonists (VKA) or non-VKA oral ACs (NOACs). To allow for transition from PACs to VKAs, PAC only treatment was defined as initial PAC treatment without any VKA within the first 15 days. In a case-control analysis, cases were VKA users and controls PAC only users, matched 1:1 on the date of the Ca-VTE. Adjusted incidence rate ratios were estimated from the odds ratios (ORs) using conditional logistic regression for matched case–control data. Adjustment included age, gender, tumour type, cancer therapy, lifestyle factors and comorbidities.


The study cohort consisted of 5419 patients with a Ca-VTE. Of those, 34.5% were given PACs (mean age 66.5), 22.7% VKAs (mean age 70.2), 0.6% NOACs (mean age 67.6), and 42.1% received no ACs. A total of 1228 PAC users were matched to 1228 VKA users. With prostate cancer as the reference group, all other types of cancer were associated with preferential use of PACs. Pancreatic cancer (OR 7.69, 3.85-14.3), stomach cancer (OR 7.14; 3.70-14.3) and ovarian cancer (OR 6.67, 3.57–12.5) were the strongest predictors for initiation of PAC only therapy. History of diabetes (OR 1.37, 1.03-1.81) also predicted preferential PAC use. Predictors for preferential use of VKAs were age ≥80 compared with age 60-69 (OR 1.48, 1.08 –2.03) and radiation therapy compared with chemotherapy (OR 1.63, 1.02–2.60).


Treatment with only PACs depends on the primary tumour and is more likely in patients with a history of diabetes. However, advanced age, and radiation therapy are predictors of VKA use.

Clinical trial identification

Legal entity responsible for the study

Institute for Epidemiology, Stastistics and Informatics GmbH


Bayer Pharma AG


A. Katholing: Grants from Bayer Pharma AG and grants from BMS/Pfizer outside the submitted work. K. Folkerts, M. Bach: Employee of the sponsor of this study. C. Martinez: Personal fees from Boehringer Ingelheim, grants from CSL Behring, grants from Bayer Pharma AG, grants from BMS-Pfizer, outside the submitted work.