Independent predictors for early and midterm mortality after thoracic surgery

Background: The purpose of the present study was to determine independent predictors for early and midterm mortality for the whole context of thoracic surgery. Methods: We studied 1453 consecutive patients who underwent thoracic surgery between 2002 and 2005. Operations included lung resections (n = 504), mediastinal (n = 468), pleural and pericardial (n = 226), esophageal (n = 83), chest wall (n = 85), tracheal (n = 50) and other procedures (n = 37). Midterm survival data (mean follow-up 2.0 ± 1.1 years) were obtained from the National Death Index. Multivariate logistic regression was used to assess in-hospital mortality. Independent predictors for midterm mortality were determined by multivariate Cox regression analysis. Results: There were 47 (3.2%) in-hospital and 312 (21.5%) late deaths. Independent predictors for in-hospital mortality included Zubrod score (OR 2.72, p < 0.001), ASA score (OR 3.42, p < 0.001), pneumonectomy (OR 20.71, p = 0.001) and no history of cerebrovascular events (OR 0.27, p = 0.011). Independent predictors for midterm mortality included age (HR 1.03, p < 0.001), weight loss (HR 1.57, p = 0.005), Zubrod score (HR 1.47, p < 0.001), primary lung cancer (HR 1.98 p < 0.001), intrathoracic extrapulmonary metastases (HR 2.78, p < 0.001), primary chest wall tumor (HR 0.14, p = 0.008), diabetes requiring insulin (HR 1.71, p = 0.017), no preoperative renal failure (HR 0.57, p = 0.004), no comorbidities (HR 0.54, p = 0.009), ASA score (HR 1.69, p < 0.001), postoperative radiation treatment (HR 1.90, p = 0.016), pneumonectomy (HR 2.18, p = 0.040), reoperation for bleeding and/or postoperative transfusion (HR 3.10, p = 0.027) and postoperative pulmonary complications (HR 1.89, p = 0.013). Conclusions: We determined independent predictors for in-hospital and midterm mortality for the whole context of thoracic surgery. Zubrod and ASA scores affect both early and midterm mortality.