Validation of a Red Blood Cell Transfusion Prediction Model in a Low Transfusion Rate Population.

Overview

The transfusion risk and clinical knowledge (TRACK) allogeneic blood transfusion prediction model was developed more than ten years ago and demonstrated good discriminative ability in patients with increased risk for allogeneic blood transfusion in an all Italian population. At the time of derivation, dual anti-platelet medication was suggested in the treatment of acute coronary syndrome, but not yet fully implemented. The aim of this study is to externally validate the TRACK blood transfusion prediction model in the cardiac surgery population of Medisch Spectrum Twente Thoraxcentrum Twente. Additionally, the impact of adding the preoperative use of dual anti-platelet medication, as additional predictive factor, to the TRACK blood transfusion prediction model will be investigated.

Cardiothoracic surgery is associated with increased perioperative blood loss and the need for allogeneic blood and blood product transfusions. This is due to three distinct factors, the invasive nature of the surgical procedures, the need for high dose anti-coagulation during extracorporeal circulation, and lastly the exposure of blood to the internal surface area of the heart-lung machine. Most cardiac patients use anti-platelet therapy before and after surgery, making hemostasis management in this specific patient population even more complex. European guidelines recommend administering anti-platelet therapy using Aspirin in combination with platelet receptor inhibitors such as Clopidogrel, Prasugrel or Ticagrelor, also known as dual anti-platelet therapy in patients with acute coronary syndrome. Making distinct decisions regarding individual patient hemostasis management remains challenging. Decision making supported by prediction models, such as, EuroSCORE is well established in the cardiac surgery population. A few models pertaining specifically to allogeneic blood transfusion have been created and externally validated. Most of these prediction models perform reasonably well, predicting red blood cell transfusions with seventy-eighty percent accuracy, depending on the model and number of prediction factors used. Some are even excellent for predicting the chance of severe post-operative bleeding. As the transfusion of even one unit of allogeneic blood transfusion impacts mortality, the choice for the best feasible prediction model for routine clinical practice that reflects daily practice, uses a limited number of predictive factors, has a predictive capacity of more than seventy percent, and discriminates between risk groups for allogeneic blood transfusion is desirable. Transfusion Risk and Clinical Knowledge (TRACK) model validation and optimization The TRACK model was developed more than ten years ago in an Italian adult cardiac surgery population. The decision to validate the TRACK model was based on its simplicity and relatively high predictive capacity, in comparison to other models with higher numbers of complex factors. This model has an allogeneic blood transfusion predictive capacity of seventy-two percent and uses a point system to divide patients into different risk groups, according to the total number of points allocated. During the derivation of this model, dual anti-platelet medication was included, but no significant association was found. In the twelve years since development, the popularity of dual anti-platelet medication used in acute coronary syndrome patients has significantly increased and its association with post-operative bleeding and allogeneic blood transfusion has been suggested. Recent studies suggest that platelet activity may play a significant role in the prediction of post-operative bleeding, and one research group found that adding platelet activity to the CRUSADE score showed a significant increase in predicting risk of major bleeding in acute coronary syndrome patients. A re-evaluation of the association between dual anti-platelet (DAPT) and allogeneic blood transfusion is necessary. This will be done by the addition of DAPT as an extra predictive factor to the TRACK model, during external validation. The negative association between mortality and transfusion products is well known. In addition, the related significant increase in hospital costs makes better perioperative hemostasis management crucial. Identifying cardiac surgery patients at risk for blood transfusion pre-operatively would aid clinicians in modifying the perioperative approach with goal the prevention of unnecessary allogeneic blood transfusion and the associated complications thereof. Validating this model might aid clinicians in reducing allogeneic blood transfusions, transfusion complications and associated costs. Ultimately this might aid for development of patient specific transfusion strategies and new blood management protocols. The aim of this study is to externally validate the TRACK blood transfusion prediction model in the cardiac surgery population of Medisch Spectrum Twente Thoraxcentrum Twente. Additionally, the impact of adding the preoperative use of dual anti-platelet medication will be studied, as additional predictive factor, to the TRACK blood transfusion prediction model.