Anticoagulation for the Prevention of Arterial Thromboembolism in Cancer Patients by Primary Tumor Site: A Systematic Review and Meta-Analysis of Randomized Trials.

AIMS Incidence of arterial thromboembolism (ATE) among ambulatory cancer patients varies by primary tumor site. However, it is unclear whether this alters the benefit-to-harm profile of prophylactic anticoagulation for ATE prevention. Therefore, we systematically evaluated the efficacy and safety of anticoagulants for ATE prevention among ambulatory cancer patients according to the primary tumor site. METHODS AND RESULTS We conducted a systematic review using Medline, Embase, SCOPUS, and CENTRAL, and included randomized trials comparing prophylactic anticoagulation to no anticoagulation among ambulatory cancer patients who initiated tumor-directed systemic therapy. Incidence of symptomatic ATE (acute ischemic stroke, acute myocardial infarction or peripheral artery occlusion) and major bleeding, as well as risk differences (RDs) attributable to anticoagulation were meta-analyzed by primary tumor site using random-effects modeling. We included 10 randomized controlled trials with 9,875 patients with follow-up ranging from 3.3 to 68 (median 6.6) months. While prophylactic anticoagulation did not reduce ATE risks overall (RD -0.49%; 95% CI -0.49% to 0.01%; I2=0%), it conferred a protective effect among pancreatic cancer patients (RD -3.2%; 95%CI -5.7% to -0.8%; I2=0%) without a detectable increase in major bleeding (RD -1.4%; 95% CI -4.6% to 1.8%; I2=0%). Prophylactic anticoagulation was not associated with ATE risk reduction in other tumor sites. CONCLUSION Based on available evidence, prophylactic anticoagulation did not reduce ATE risk among ambulatory cancer patients overall. However, we observed lower incidence of ATE among pancreatic cancer patients randomized to receive anticoagulation. Prophylactic anticoagulant use to reduce ATEs in pancreatic cancer should be evaluated in future research.