Prediction and prognosis of reintubation after surgery for Stanford type A aortic dissection

Abstract

BackgroundReintubation is a serious adverse respiratory event after Stanford type A aortic dissection surgery (AADS), however, published studies focused on reintubation after AADS are very limited worldwide. The objectives of the current study were to establish an early risk prediction model for reintubation after AADS and to clarify its relationship with short-term and long-term prognosis.MethodsPatients undergoing AADS between 2016–2019 in a single institution were identified and divided into two groups based on whether reintubation was performed. Independent predictors were identified by univariable and multivariable analysis and a clinical prediction model was then established. Internal validation was performed using bootstrap method with 1,000 replications. The relationship between reintubation and clinical outcomes was determined by univariable and propensity score matching analysis.ResultsReintubation were performed in 72 of the 492 included patients (14.6%). Three preoperative and one intraoperative predictors for reintubation were identified by multivariable analysis, including older age, smoking history, renal insufficiency and transfusion of intraoperative red blood cells. The model established using the above four predictors showed moderate discrimination (AUC = 0.753, 95% CI, [0.695–0.811]), good calibration (Hosmer-Lemeshow χ2 value = 3.282, P = 0.915) and clinical utility. Risk stratification was performed and three risk intervals were identified. Reintubation was closely associated with poorer in-hospital outcomes, however, no statistically significant association between reintubation and long-term outcomes has been observed in patients who were discharged successfully after surgery.ConclusionsThe requirement of reintubation after AADS is prevalent, closely related to adverse in-hospital outcomes, but there is no statistically significant association between reintubation and long-term outcomes. Predictors were identified and a risk model predicting reintubation was established, which may have clinical utility in early individualized risk assessment and targeted intervention.