Introduction: Previous observational studies in patients with cancer, subarachnoid hemorrhage, or undergoing surgery suggested an association between red blood cell (RBC) transfusion and an increased risk for arterial and venous thrombosis, findings warranting further evaluation.

Methods: The risk of thrombosis in all hospitalized patients was assessed in patients admitted to 12 hospitals located across the United States in New Haven, Pittsburgh, Milwaukee, and San Francisco using the Recipient Database from the NHLBI Recipient Epidemiology and Donor Evaluation Study-III. All patients admitted between January 2013 and September 2015 were included in analyses. A thrombosis event was defined as an admission with an ICD-9 code of arterial or venous thrombosis as a primary or secondary diagnosis and administration of a therapeutic dose of an anticoagulant or antiplatelet agent during that admission. The timing of the thrombosis was based upon medication administration time. Timing of transfusion was based upon release of RBC from the blood bank. Patients using anticoagulant medications at home or with a history of thrombosis diagnostic code were excluded. Encounters with a thrombotic event within the first 24 hours after admission were also excluded as thrombosis was likely the etiology of the admission. A proportional hazards regression model was used to evaluate the effect of transfusion on the risk of thrombosis with transfusion treated as a time-varying predictor. Discharge without thrombosis, in-hospital thrombosis and death were considered competing risks. Estimates were adjusted for age, sex, hospital, medical comorbidities, and surgical procedures.

Results: Of 657,412 inpatient admissions, 67,176 (10.2%) received at least one RBC transfusion; median age was 58 and 53.6% were women. Two percent (12,927) of patients experienced a thrombosis. Of these, 2,587 developed thrombosis after a RBC transfusion. In a univariable analysis, many surgical procedures were associated with a risk of thrombosis including cardiac surgery [Hazard Ratio (HR)=3.25 (3.05-3.47)] and vascular surgery, HR=1.90 (1.82-1.99). Comorbidities were also associated with thrombosis including heart arrhythmias HR=1.60 (1.54-1.66), hypertension HR=1.31 (1.26-1.37), anemia HR=1.32 (1.17-1.49), and metastatic cancer HR=1.36 (1.28-1.45). In unadjusted analyses, RBC transfusion was associated with an increased risk of thrombosis [HR=1.3 (95% CI 1.25-1.36)] (Table 1). The risk of thrombosis increased with the number of RBC units transfused. However, after adjustment for surgical procedures, age, sex, hospital, and comorbidities, no effect of RBC transfusion on risk of thrombosis was found [HR 0.98 (95% CI: 0.94-1.03)] (Table 1).

Conclusions: In univariate analyses there are multiple risks for venous and arterial thrombosis, including many surgical procedures and medical comorbidities, as well as RBC transfusion. However, after adjustment for these multiple risk factors, RBC transfusion was not associated with a risk for arterial or venous thrombosis. Thus, this study of all patients admitted to 12 hospitals in the US did not identify RBC transfusion as an independent risk factor for thrombosis. The REDS-III database contains detailed information about comorbidities, medications, and timing of transfusion which allowed us to carefully define a thrombotic episode and its temporal relationship to RBC transfusion, and adjust for confounders that may not have been available in other datasets. Our findings do not rule out the possibility of a patient subset where RBC transfusion increases risk for thrombosis, such as patients with an underlying genetic hypercoagulable risk, but they do indicate that RBC transfusion does not appear to be an important risk factor for thrombosis in most patients. As such, receiving a blood transfusion should not influence decisions about venous thrombosis prophylaxis in hospitalized patients.

Disclosures

Mast:Novo Nordisk: Research Funding.

Author notes

*

Asterisk with author names denotes non-ASH members.

This icon denotes a clinically relevant abstract

Sign in via your Institution