Incidence, Morbidity, and Mortality of Pulmonary Complications in Free Flap Reconstruction: Limitations of Predictive Models

Abstract Objective Herein, we evaluate the incidence, risk factors, and prognostic implications of postoperative pulmonary complications (PPCs) in head and neck microvascular free flap (MVFF) reconstruction. Current prediction models were assessed, and a head and neck MVFF‐specific model is proposed. Study Design Retrospective review of 638 head and neck MVFF cases from August 2019 to May 2024. Setting Tertiary academic center. Methods Data were collected via chart review focusing on preoperative, intraoperative, and postoperative risk factors for PPCs within 30 days of surgery. Results Grades 2 to 5 PPCs occurred in 27% of patients. Predictors of PPCs in univariate analysis include prolonged surgery (mean: 10.06 ± 2.67 hours, P = .006), estimated blood loss ≥ 200 mL (n = 111, P = .006), advanced tumor stage (III/IV: n = 123, P = .013), hematoma (n = 25, P < .001), and postoperative transfusion (n = 20, P = .037). Tumor stage (odds ratio [OR] 1.29, 95% CI 1.06‐1.57, P = .012), surgery duration (OR 1.08, 95% CI 1.01‐1.17, P = .031), and hematoma (OR 2.98, 95% CI 1.50‐5.94, P = .002) remained significant predictors of grades 2 to 5 PPCs on multivariable analysis. In‐hospital mortality was 1.4% (n = 9), and all experienced grade 5 PPCs. The 1‐year mortality rate was 13.48 per 100 patients, with significantly lower survival in patients with grades 2 to 5 PPCs (75.6% vs 89.4%). PPC was independently associated with mortality (hazard ratio [HR] 3.94, 95% CI 1.69‐9.22, P = .002). Our model (area under the curve [AUC] 0.65) outperformed the ARISCAT (AUC = 0.51) and Gupta scores (AUC = 0.45) in predicting PPCs. Conclusion PPCs are common after MVFF, contributing to significant morbidity and mortality. Current models are inadequate, highlighting the need for a tailored model specific to oncologic head and neck surgery.