Sarcopenia is associated with blood transfusions in head and neck cancer free flap surgery.

OBJECTIVE: To determine if sarcopenia is a predictor of blood transfusion requirements in head and neck cancer free flap reconstruction (HNCFFR). METHODS: A single-institution, retrospective review was performed of HNCFFR patients with preoperative abdominal imaging from 2014 to 2019. Demographics, comorbidities (modified Charlson Comorbidity Index [mCCI]), skeletal muscle index (cm2/m2), oncologic history, intraoperative data, and 30-day postoperative complications (Clavien-Dindo score [CD]) were collected. Binary logistic regression was performed to determine predictors of transfusion. RESULTS: Eighty (33.5%), 66 (27.6%), and 110 (46.0%) of n = 239 total patients received an intraoperative, postoperative, or any perioperative blood transfusion, respectively. Sixty-two (25.9%) patients had sarcopenia. Patients receiving intraoperative transfusions had older age (P = .035), more frequent alcoholism (P = .028) and sarcopenia (P < .001), greater mCCI (P < .001), lower preoperative hemoglobin (P < .001), reconstruction with flaps other than forearm (P = .003), and greater operative times (P = .001), intravenous fluids (P < .001), and estimated blood loss (EBL, P < .001). Postoperative transfusions were associated with major complications (CD ≥ 3; P < .001). Multivariate regression determined sarcopenia (P = .023), mCCI (P = .013), preoperative hemoglobin (P = .002), operative time (P = .036), and EBL (P < .001) as independent predictors of intraoperative transfusion requirements. Postoperative transfusions were predicted by preoperative hemoglobin (P = .007), osseous flap (P = .036), and CD ≥ 3 (P < .001). A perioperative transfusion was predicted by sarcopenia (P = .021), preoperative hemoglobin (P < .001), operative time (P = .008), and CD ≥ 3 (P = .018). CONCLUSION: Sarcopenia is associated with increased blood transfusions in HNCFFR. Patients should be counseled preoperatively on the associated risks, and the increased blood product requirement should be accounted in resource-limited scenarios. LEVEL OF EVIDENCE: 4.