A model for predicting transfusion requirements in head and neck surgery

Area of Science:

• Oncology
• Surgical Oncology
• Transfusion Medicine

Background:

• Allogeneic blood transfusions, while enabling complex cancer resections, carry risks like infection and immunosuppression.
• Patient reluctance towards transfusions increases demand for preoperative autologous blood donation (PABD), often resulting in significant blood wastage.

Purpose of the Study:

• To develop a Transfusion Prediction and Risk Assessment (TPRA) model for perioperative blood transfusion needs in head and neck oncologic surgery.
• To enable informed decisions regarding PABD by providing patients and physicians with transfusion probability.

Main Methods:

• A prospective study of 436 patients undergoing major head and neck oncologic surgery over 4 years.
• Logistic regression analysis of 12 pre-operative variables to identify predictors of transfusion need.
• Development of a TPRA model and algorithm based on significant variables: flap reconstruction, low preoperative hemoglobin, and T3/T4 tumor stage.

Main Results:

• 11.7% of patients required transfusions, with a median of 2 units.
• Flap reconstruction, low preoperative hemoglobin, and advanced tumor stage (T3/T4) were significant predictors of transfusion need.
• The TPRA model identified lowest risk (0.02) for patients with normal hemoglobin, no flap reconstruction, and non-advanced tumors, and highest risk (0.65) for those with low hemoglobin, flap reconstruction, and advanced tumors.

Conclusions:

• The TPRA model can potentially reduce transfusion costs by 50% without increasing allogeneic transfusion risks.
• Clinical utility lies in aiding surgeons with cost-effective, evidence-based preoperative transfusion planning.
• Further validation studies are recommended to confirm the TPRA model's accuracy and clinical applicability.