Predictors of operative death after oesophagectomy for carcinoma

Background: Oesophagectomy for carcinoma provides a chance of cure but carries significant risk. This study defined risk factors for death after oesophageal resection for malignant disease. Methods: Between 1990 and 2003, 773 oesophagectomies for oesophageal cancer were performed. Continuous variables were categorized into quartiles for analysis. Predictors of operative mortality were identified by univariate and multiple logistic regression analysis. Results: The operative mortality rate was 4.8 per cent (37 of 773). In univariate analysis, advanced age, reduced forced expiratory volume in 1 s (FEV1), reduced forced vital capacity, presence of diabetes and tumour located in the upper third of the oesophagus were associated with a higher mortality rate. Multivariate analysis identified age (highest relative to lowest quartile, odds ratio (OR) 4.87 (95 per cent confidence interval (c.i.) 1.35 to 17.55); P = 0.009), tumour position (upper third relative to other locations, OR 4.23 (95 per cent c.i. 1.06 to 16.86); P = 0.041) and FEV1 (lowest relative to highest quartile, OR 4.72 (95 per cent c.i. 1.01 to 21.99); P = 0.018) as independent predictors of death. Conclusion: Advanced age, impaired preoperative respiratory function and a tumour high in the oesophagus are associated with a significantly increased risk of death after oesophagectomy for carcinoma.