Comparison of Anticoagulants in the Prevention of Arterial Thromboembolism in Patients with Cancer and Venous Thromboembolism: A Population-Based Cohort Study

Background:

The clinical course of patients with cancer is frequently complicated by venous thromboembolism (VTE), and emerging evidence suggests that the incidence of arterial thromboembolism (ATE) may also be elevated in this patient population. However, it is not established whether the development of VTE later predicts the development of ATE or whether one class of anticoagulant is more effective in preventing subsequent arterial thrombotic complications.

Aim:

We investigated the incidence of ATE including ischemic stroke, transient ischemic attack (TIA), and systemic thromboembolism among patients with cancer following diagnosis of VTE (deep vein thrombosis [DVT] or pulmonary embolism [PE]) and initiation of therapeutic anticoagulation. We sought to determine if the risk of subsequent ATE differed based on receipt of direct oral anticoagulant (DOAC), low-molecular-weight heparin (LWMH), or warfarin for treatment of VTE.

Methods:

This study was a retrospective, population-based cohort study of patients diagnosed with primary brain, colorectal, gastric, pancreatic, lung, or ovarian cancer from January 1, 2007 to December 31, 2015 using Surveillance, Epidemiology, and End Results (SEER) data linked with Medicare claims. Patients were eligible for study inclusion if they had a qualifying VTE diagnosis within one month before or six months following cancer diagnosis, were at least 66 years of age at time of VTE diagnosis, survived at least 14 days following the VTE, and had a prescription claim for a DOAC, LWMH, or warfarin within 30 days of VTE diagnosis. To account for patients who were bridged from LMWH to a second anticoagulant, the LMWH group was defined as patients whose first anticoagulant prescription was a LMWH product and who did not subsequently receive a prescription for warfarin or DOAC within 14 days. The date of first anticoagulant prescription was used as the index date for each study participant. Patients were propensity-score matched 1:1:1 into DOAC, LMWH, and warfarin treatment groups based on a scoring algorithm using age, sex, race, ethnicity, marital status, comorbidities (atrial fibrillation, hypertension, congestive heart failure, peripheral artery disease, coronary artery disease, cerebrovascular disease, atherosclerosis, diabetes, tobacco use, prior ATE), cancer site and stage, type of VTE (DVT or PE), and receipt of systemic cancer treatment within three months of study index. The primary outcome of the study was to determine the odds ratio (OR) for development of ischemic stroke or TIA within 12 months with each anticoagulant treatment group. Secondary outcomes included the risk of all ATE, overall survival (OS), and the identification of risk factors associated with ATE.

Results:

The study cohort comprised of 12,952 total eligible patients – after propensity-score matching, 4,641 were included for final analysis with 1,547 in each of DOAC, LMWH, and warfarin treatment groups. At 12 months following diagnosis of VTE and initiation of anticoagulation, the OR for ischemic stroke and TIA for LMWH was 0.92 (95% confidence interval [CI] 0.79 – 1.07) and 0.91 (95% CI 0.78 – 1.06) for warfarin with DOAC as the reference group. The median survival for the cohort was 10.4 months, 10.7 months for the DOAC group, 9.6 months for the LMWH group, and 11.4 months for the warfarin group (p = 0.007). On univariable modeling, age > 75 years (OR 1.38, 95% CI 1.22-1.57), hypertension (OR 8.22, 95% CI 5.92 – 11.79), atrial fibrillation (OR 2.51, 95% CI 2.05 – 3.08), congestive heart failure (OR 3.70, 95% CI 3.24 – 4.21), peripheral artery disease (OR 3.98, 95% CI 3.10 – 5.12), diabetes (OR 2.27, 95% CI 2.00 – 2.58), tobacco use (OR 2.53, 95% CI 1.53 – 4.18), and receipt of systemic cancer therapy (OR 1.14, 95% CI 1.00 – 1.29) were associated with ATE.

Conclusions:

Ischemic stroke and ATE are common among patients with cancer treated with anticoagulation for VTE. In this SEER-Medicare based study, the type of anticoagulant did not impact the risk of ATE. Traditional cardiovascular risk factors and systemic anticancer treatment were predictive of ATE among this patient population. Further investigation is required to delineate optimal preventative approaches for patients with cancer, VTE, and risk factors for subsequent ATE.