Influence of vessel-depleted neck and risk factors on vascularized free flap failure: a retrospective cohort study and predictive model

Background The vessel-depleted neck represents a distinct clinical entity in free flap reconstruction, yet its impact on flap outcomes remains controversial. This study aimed to compare the rate of free flap failure between vessel-depleted neck cases and ordinary cases in head and neck reconstruction, and to identify risk factors for flap failure while developing a perioperative predictive model. Methods We analyzed all vascularized free flap cases performed at our institution between July 2021 and June 2024. Cases were divided into training and testing sets based on surgery time. The association between vessel-depleted neck and flap failure was assessed. Multivariate logistic regression analysis was performed to identify independent risk factors for flap failure in the training set. A predictive model was subsequently constructed and validated using the testing set. Results Among 1,965 free flap cases in analysis, 61 cases in the training set ( n = 1, 506) were vessel-depleted neck cases. The incidence of flap failure in this subgroup was 4.9% (3/61), which was not significantly different from that in ordinary cases ( P = 0.552). Multivariate logistic regression analysis revealed that liver disease ( P = 0.001), maxillary disease location ( P = 0.001), titanium plate use ( P = 0.033), and surgical site infection ( P < 0.001) were significant independent risk factors for flap failure. Details regarding recipient vessels were also recorded. A perioperative predictive model incorporating these four factors was developed and demonstrated an area under the curve (AUC) of 0.782 in the testing set. Conclusion There was no significant difference in flap failure rates between vessel-depleted neck cases and ordinary cases. Liver disease, titanium plate use, maxillary disease location, and surgical site infection were identified as significant risk factors for flap failure. The developed predictive model demonstrated clinically useful discriminatory ability.