Red blood cell transfusion may not increase risk for thrombosis

Issue: February 25, 2019

SAN DIEGO — Red blood cell transfusion did not increase risk for thrombotic events in adjusted analyses, despite reports of increased risk in previous observational studies, according to findings presented at ASH Annual Meeting and Exposition.

There is biologic plausibility for the association between red blood cell transfusion and thrombotic events, according to Lisa M. Baumann Kreuziger, MD, MS, associate medical director of BloodCenter of Wisconsin in Milwaukee and assistant professor in the division of hematology and oncology of Medical College of Wisconsin. However, she noted that there have been limitations to previous epidemiologic studies exploring the associations, such as “the inability to associate the time of transfusion with the time of thrombosis,” she said.

“There was some limitation in the ability to adjust for comorbidities,” she added. “All of these studies used logistic regression models, so that gives you either a ‘yes’ or ‘no’ event. Either they were transfused or not, or they had thrombosis or not, and it doesn’t give you the association of the timing of those events.”

Baumann Kreuziger and colleagues assessed risk for thrombosis among 755,000 patients admitted to any of 12 academic or VA hospitals in four states across the U.S. between January 2013 and September 2015.

Researchers used a database from the NHLBI Recipient Epidemiology and Donor Evaluation Study-III to obtain information about the time of transfusion, medication dosing and comorbidities, along with ICD-9 diagnostic codes.

“This allowed us to really carefully define a thrombotic episode, and also establish a temporal relationship between the red cell transfusion and the thrombotic event,” Baumann Kreuziger said. “We were able to adjust for confounders that were not available in other data sets.”

The final analysis included 579,896 admitted patients, 64,589 of whom underwent red blood cell transfusion and 8,779 of whom had a thrombotic event.

Baumann Kreuziger reported that 1,561 patients were admitted, had a thrombotic event and then a transfusion, whereas 2,587 were admitted, had a transfusion and then a thrombotic event.

Surgical procedures such as cardiac surgery (HR = 2.58; 95% CI, 2.4-2.77) and vascular surgery (HR =2.15; 95% CI, 2.02-2.29) were largely associated with risk for thrombosis.

“The only one that showed a decreased risk was orthopedic surgery,” Baumann Kreuziger said.

A number of comorbidities also appeared associated with thrombosis, from arrhythmia (HR = 1.62; 95% CI, 1.56-1.68) and hypertension (HR = 1.36; 95% CI, 1.31-1.41), to anemia (HR = 1.33; 95% CI, 1.18-1.51), metastatic cancer (HR = 1.31; 95% CI, 1.23-1.39), heart failure (HR = 1.18; 95% CI, 1.13-1.23) and diabetes (HR = 1.13; 95% CI, 1.08-1.17).

In the unadjusted model, red blood cell transfusion was associated with increased risk for thrombosis (HR = 1.3; 95% CI, 1.23-1.36). However, when researchers adjusted that model for surgical procedures, the association decreased (HR = 1.18; 95% CI, 1.12-1.23). It decreased further and lost significance when adjusted for sex, age, hospital, surgical procedures and comorbidities (HR = 0.98; 95% CI, 0.94-1.03).

“Surgical procedures, medical comorbidities and red cell transfusion are associated with a risk for thrombosis,” Baumann Kreuziger said. “After appropriate adjustment, red cell transfusion was not associated with a risk for arterial and venous thrombosis. Really, the risk for thrombosis is driven by the underlying comorbidities and the surgical procedures other than the red cell transfusion.” – by Rob Volansky

Reference:

Baumann Kreuziger L, et al. Abstract 415. Presented at: ASH Annual Meeting and Exposition; Dec. 1-4, 2018; San Diego.

Disclosures : Baumann Kreuziger reports no relevant financial disclosures. One researcher reports research funding from Novo Nordisk.