To establish a model to predict the risk of acute respiratory distress syndrome (ARDS) after cardiac surgery.
Data were collected on 132 ARDS patients, who received valvular or coronary artery bypass grafting surgery from January 2009 to December 2019. We developed the prediction model by multivariable logistic regression. Then, we used the coefficients for developing a nomogram that predicts ARDS occurrence. Internal validation was performed using resampling techniques to evaluate and optimize the model.
All variables fit into the model, including albumin level before surgery (odds ratio [OR]: 0.96; 95% confidence interval [CI]: 0.92, 0.99; P = .01), cardiopulmonary bypass time (OR: 1.01; 95% CI: 1.00, 1.02; P = .02), APACHE II after surgery (OR: 1.21; 95% CI: 1.13, 1.29; P < .001), and history of diabetes (OR: 2.31; 95% CI: 1.88, 3.87; P < .001); these were considered to build the nomogram. The score distinguished ARDS patients from non-ARDS patients with an AUC of 0.785 (95% CI: 0.740, 0.830) and was well calibrated (Hosmer–Lemeshow P = .53).
Our developed model predicted ARDS in patients undergoing cardiac surgery and may serve as a tool for identifying patients at high risk for ARDS after cardiac surgery.