Perioperative risk scores and predictors for postoperative pulmonary complications in thoracic surgery with one-lung ventilation: A prospective observational study

Background Postoperative pulmonary complications (PPCs) are common after thoracic surgery performed with one-lung ventilation (OLV). Although risk scores such as ARISCAT, ASA physical status, and LAS VEGAS are used to predict PPC in general surgery, their performance in thoracic surgery remains unclear. This study aimed to evaluate the ability of these scores to predict PPCs. Methods This prospective observational study included 115 adult patients undergoing noncardiac thoracic surgery requiring OLV. The primary outcome was the incidence of in-hospital PPCs within 30 days. We assessed the discriminatory ability of risk scores using receiver operating characteristic curve analysis (area under the curve (AUC)) and identified independent predictors with multivariable logistic regression. Results PPCs occurred in 43 patients (37.4%). The LAS VEGAS score demonstrated the best discriminatory performance (AUC = 0.648, p = .005), followed by ARISCAT (AUC = 0.598, p = .067) and ASA (AUC = 0.562, p = .123). Independent predictors of PPCs were preoperative hemoglobin (OR 0.77, 95% CI 0.62–0.95) and intraoperative noradrenaline use (OR 5.93, 95% CI 2.36–15.95). A combined model including these two variables achieved an AUC of 0.772, which was significantly superior to the individual scores. Conclusions The ARISCAT, ASA, and LAS VEGAS scores showed only weak-to-moderate predictive performance for PPCs after thoracic surgery with OLV. These findings do not support their routine use in this setting and highlight the need for thoracic surgery-specific risk prediction tools.