Risk prediction model for respiratory complications after lung resection: An observational multicentre study

Area of Science:

• Thoracic Surgery
• Pulmonary Medicine
• Clinical Prediction Models

Background:

• Postoperative pulmonary complications (PPCs) are a risk for patients undergoing lung surgery.
• Accurate identification of high-risk patients is crucial for optimizing perioperative management.
• The Clinical Prediction Rule for Pulmonary Complications (CPRPC) was developed for thoracic surgery but required external validation.

Purpose of the Study:

• To externally validate the CPRPC's performance in predicting PPCs after lung resection for primary tumors.
• To derive a new predictive index for PPCs if the CPRPC demonstrated poor discriminatory power.

Main Methods:

• Retrospective, observational, multicenter study involving 559 adult patients undergoing pulmonary resection across 13 Spanish hospitals.
• Data collected on PPCs, defined as atelectasis, pneumonia, pulmonary embolism, respiratory failure, or need for oxygen at discharge.
• Performance of the CPRPC assessed for discrimination and calibration using receiver operating characteristic (ROC) curve analysis.

Main Results:

• The CPRPC showed insufficient discriminatory power, with an area under the ROC curve of 0.47.
• Stepwise multivariate logistic regression identified age, smoking status, and predicted postoperative forced expiratory volume in 1 second as key predictors.
• A new risk score incorporating these factors achieved an area under the ROC curve of 0.74, indicating improved predictive accuracy.

Conclusions:

• The CPRPC demonstrated poor performance in this external validation cohort and is not a useful tool for predicting PPCs.
• A novel, more accurate predictive score based on age, smoking status, and predicted postoperative forced expiratory volume in 1 second was developed.
• External validation of the newly derived predictive formula is recommended.