Bleeding in Patients Receiving Low-Molecular-Weight Heparin for Cancer-Associated Thrombosis

Background: Bleeding is a major concern in patients who are treated with anticoagulants. To date, there are no studies of predictors of bleeding in cancer-associated thrombosis patients who receive extended duration low-molecular-weight heparin (LMWH). This study aims to determine the incidence and predictors of bleedings in these patients in a real-world setting.

Methods: A retrospective cohort (chart review) study was conducted from January 2011 to January 2014 at Juravinski Cancer Center, Hamilton, Canada. Consecutive objectively proven venous thromboembolism (VTE) patients were included if they had active cancer and were planned to receive extended duration LMWH (longer than 4 weeks). The primary outcome measure was the incidence of clinically relevant bleeding, which was defined as bleeding that required investigation, an invasive procedure, hospital admission or withholding LMWH for greater than or equal to 3 days. Secondary outcome measures included incidence of major bleeding (defined by bleeding that was associated with drop of hemoglobin at least 20 g/L or required at least 2 units of red blood cell transfusion, bleeding in critical site or fatal bleeding) and incidence of objectively proven recurrent VTE.

Results: Data were available for 1,144 patients with a median follow-up of 8.5 months. The average age (standard deviation, [SD]) of the patients was 63.6 (12.2) years, and 53.6% were female. Concomitant antineoplastic treatment consisted of chemotherapy (45.1%), radiotherapy (5.7%), targeted therapy (1.9%) and combination therapy (17.5%). No antineoplastic treatment was given during the study period to 29.7% of the patients. The cumulative incidence of clinically relevant bleeding was 4.6% at 3 months, 7.3% at 6 months and 10.3% at 12 months, Figure 1. Sites of bleeding were gastrointestinal tract (49.6%), genitourinary tract (16.2%), intracranial (9.0%), muscle and retroperitoneal (8.2%), and others (17.1%). Of the gastrointestinal bleeds, 52.3% occurred in patients who were not documented to have the GI tract as the primary site of malignancy. The cumulative incidence of major bleeding was 5.5% (1.6% of the study cohort had a fatal bleed). At the time of the bleeding event, the mean (SD) hemoglobin was 87.2 (23.1) g/L, mean platelet count (SD) was 251.8 (158.4) x10⁹/L and 61% of patients received a red blood cell transfusion (median 2 unit [range, 1-7]). The independent predictors of bleeding in a multivariable model were hypertension, metastatic disease, prostate cancer, soft tissue sarcoma and recurrent VTE (Table 1.). The presence of brain tumour (primary or secondary) was not statistically significantly associated with an increased risk of clinically relevant bleeding. The cumulative incidence of recurrent VTE was 12.0%. Seventy-six percent of recurrent VTE occurred while patients were receiving anticoagulant therapy.

Conclusions: This study suggests that predictors for clinically relevant bleeding in patients who receive extended duration LMWH for treatment of cancer-associated thrombosis include hypertension, metastatic disease, and recurrent VTE, in addition to tumour-site specific characteristics. Investigation into measures to reduce the frequency of gastrointestinal bleeding, the most common form of anticoagulant-related bleeding in both gastrointestinal cancer and non-gastrointestinal cancer, would be beneficial.

Figure 1.

View largeDownload slide

Cumulative risk of clinically relevant bleeding in cancer-associated thrombosis patients who received low-molecular weight heparin

Figure 1.

View largeDownload slide

Cumulative risk of clinically relevant bleeding in cancer-associated thrombosis patients who received low-molecular weight heparin

Close modal