Survival in patients with cancer‐associated thrombosis in relation to anticoagulants: Re‐vitalization of warfarin?

Abstract

Cancerassociated thrombosis is a wellknown complication. Patients with cancer have a 12fold increased risk of venous thromboembolism (VTE) compared to noncancer population, and the risks further increase to 23fold in those receiving chemotherapy or targeted therapy.1 Thromboembolism is the second leading cause of death in patients with cancer.2 The CLOT trial established lowmolecularweight heparin (LMWH) as the recommended anticoagulant for cancerassociated thrombosis, based on the findings that LMWH (dalteparin) was associated with a significantly reduced risk of recurrent VTE without an increased risk of bleeding events when compared with warfarin.3 More recently, in the similarly designed CATCH trial using Tinzaparin as the LMWH of choice, a trend towards a reduction of VTE recurrence was also found in patients receiving LMWH, with a significant reduction in clinically relevant nonmajor bleeding (CRNMB) (but not major bleeding) events.4 Although LMWH is viewed as the standard of care for cancerassociated thrombosis for close to two decades, it is associated with high cost, burdensome daily injections, and decreasing adherence over time.5 Despite a reduction of recurrent VTE, mortality benefit has mostly not been shown with LMWH in randomized controlled trials (RCTs) (Table 1), as most deaths (85% to >90%) were attributed to cancer progression. However, available RCTs were limited in sample size and followup duration (6– 12 months) to assess a potential survival benefit. With these premises in mind, in this issue of the Journal, Chiasakul and colleagues assessed the overall survival in 9706 patients with cancerassociated thrombosis prescribed LMWH compared to warfarin within 30 days of VTE, using data from the United States SEERMedicare database from 2007 to 2016.6 Six cancer types (gastric, colorectal, pancreatic, lung, ovarian, or brain) with known high rates of VTE were included. As expected from a nonrandomized observational study, baseline characteristics in patients prescribed warfarin were different from those on LMWH. To account for confounders, patients in either group were exact matched on cancer stage and propensity score matched on age, cancer site, year of VTE diagnosis, and time from cancer diagnosis to index VTE. The primary outcome of the study was overall survival from the diagnosis of VTE to death from any cause or to December 31, 2016. Although the investigators hypothesized that LMWH would be associated with a superior survival compared with warfarin, surprisingly, they found the opposite, that patients prescribed warfarin had an improved overall survival compared to LMWH, after a median follow up of 61 months (range 0.5– 119). Patients prescribed warfarin had a median survival of 9.8 months (95% confidence interval [CI] 9.1– 10.4), compared with 7.2 months (95% CI 6.8– 7.8) in patients prescribed LMWH.6 The hazard ratio (HR) for death (warfarin vs. LMWH) was 0.86 (95% CI 0.83– 0.90, p < .001). The survival benefit of warfarin over LMWH was consistent across multiple subgroup analyses, including cancer stage, cancer type, age, gender, type of VTE, and after excluding pancreatic cancer. Interestingly, patients with earlier stage malignancy (stage 1– 2) derived more survival benefit from warfarin over LMWH (10.5 months longer) compared to stage IV disease (1 month longer).6 As expected, patients prescribed warfarin had poorer socioeconomic status such as lower household income, shorter years of education, etc. Despite that, the survival benefits were seen with the warfarin group.6 This work is certainly interesting and results provocative, but has the usual limitations related to a retrospective database analysis. First is the potential of residual confounding from unmeasured factors, despite the investigators extensive efforts to adjust for possible confounders. LMWH is likely preferred and prescribed more in patients considered as “higher risk”, evident by more pancreatic cancer (22% vs. 14%), more systemic anticancer therapy within 3 months