Prediction Models for Postoperative Reintubation in Patients With Acute Aortic Dissection (ADreintubate)

Huazhong University of Science and Technology

Status: Completed

Conditions: Prediction Model; Postoperative Reintubation; Type A Aortic Dissection; Nomogram

Treatments: Procedure: Type A aortic dissection surgery

Study type: Observational

Funder types: Other

Identifiers

NCT06415630

WHUH-AADreintu

Details and patient eligibility

About

Reintubation is an adverse postoperative complication in patients with Type A aortic dissection (AAD) that correlates to poor outcomes. This study aims to analyze the risk factors associated with reintubation and to create a fully automated score model to predict the incidence of reintubation. A total of 861 patients diagnosed with AAD and undergoing surgical procedures in a single institution between January 2018 and October 2023 were selected in wuhan Union Hospital. Preoperative and postoperative informmation was used for seeking risk factors and build prediction model for postoperative reintubation. Finally, 5 risk factors wasidentified and a nomogram was established for predicting postoperative reintubation in patients with AAD.

Enrollment: 861 patients

Sex: All

Ages: 19 to 90 years old

Volunteers: No Healthy Volunteers

Inclusion criteria

(1) patients diagnosed with AAD admitted for open surgery; (2) aged 18 years or older.

Exclusion criteria

(1) patients deceased during or within 24 hours after surgery; (2) patients with preoperative intubation.

Trial design

861 participants in 2 patient groups

training group

Description:

To determine potential predictive factors, the training group was subjected to LASSO regression analysis, which effectively eliminated several irrelevant or multicollinearity independent variables to reduce high-dimensional data.Seven models were initially constructed in the training group: multivariable logistics regression (MLR), decision-tree modeling, random forest, XGBoost, Support Vector Machines, k-nearest neighbors, and LightGBM.

Treatment:

Procedure: Type A aortic dissection surgery

testing group

Description:

Testing group was used to examine the performance of the seven prediction models, including discrimination and calibration performance. Finally, the model with best discrimination and calibration performance was used to construct nomogram for predicting reintubation.

Treatment:

Procedure: Type A aortic dissection surgery