The TIMELESS Project: A Mixed-Methods Analysis of Modifiable Risk Factors and Transfusion Practices in 24,000 Cardiac...

Approximately 10% to 15% of all red blood cell (RBC) transfusions in the United States and United Kingdom occur during cardiac surgery. Despite efforts to standardize patient blood management, transfusion practices remain highly variable across providers and institutions. We launched the Transfusion Improvement to Minimizing Excess in Cardiac Surgery (TIMELESS) initiative to identify modifiable drivers of transfusion at our high-volume cardiac surgery center.

Our mixed-methods approach combined a retrospective data review and a survey to assess perioperative RBC and blood component (platelets, plasma, cryoprecipitate) transfusion through three key domains: (1) preoperative assessment/optimization (evaluation of modifiable risk factors for transfusion: preoperative anemia, malnutrition, anticoagulant/platelet inhibitor use); (2) perioperative blood conservation (modifiable risk factors: intraoperative antifibrinolytic use, hyperglycemia, perioperative hypothermia); (3) transfusion behaviors (laboratory guidance and transfusion triggers). These domains were investigated through (i) a retrospective database analysis that examined the association of risk factors with total intraoperative and early postoperative blood product transfusion in adult cardiac surgical patients between January 2018 and June 2024; and (ii) a structured front-line provider survey that examined opinions and behaviors that drive blood transfusion (including transfusion thresholds, triggers, and barriers to laboratory-guided practice) in anesthesia, surgical, and intensive care unit caregivers.

Of 24,005 cardiac surgical patients in the database review, 13,420 (55.9%) received transfusions during surgery and/or early postoperatively. Of 258 front-line providers surveyed, 112 (43%) responded. Preoperative assessment/optimization (domain #1) and perioperative blood conservation (#2) found that five of six modifiable factors were associated with greater risk-adjusted transfusion: preoperative anemia (incidence rate ratio [IRR] 1.74, 95% confidence interval [CI], 1.67-1.83), malnutrition (IRR 1.44, 95% CI, 1.36-1.54), perioperative hypothermia (IRR 1.66, 95% CI, 1.59-1.73), and hyperglycemia (IRR 1.02 per 10 mg/dL, 95% CI, 1.02-1.02), whereas antifibrinolytic therapy was associated with reduced transfusion (IRR 0.90, 95% CI, 0.86-0.93). Transfusion triggers/laboratory guidance (#3) found that many transfusions were given empirically without laboratory-guidance (RBC 32.1%, platelets 44.9%, cryoprecipitate 55.0%, plasma 82.4%). Survey found that most providers (91/112, 81%) endorsed restrictive hemoglobin thresholds (<7.5 g/dL) for stable, non-bleeding patients; platelet and cryoprecipitate transfusion thresholds varied by specialty; laboratory-guidance was omitted for urgent bleeding/coagulopathy (99/112, 88%), prolonged surgical time (49/112, 44%), and long laboratory turnaround time (49/112, 44%).

This large mixed-methods quality initiative identified modifiable risk factors, provider opinions, and transfusion behaviors to understand key drivers of transfusion in our cardiac surgery center. These results lay the groundwork to modify risk factors and transfusion behaviors and ultimately optimize perioperative blood transfusion therapy.