[Perioperative risk factors in esophageal cancer: a retrospective study of independent variables]

The postoperative outcome of 127 patients with an esophageal carcinoma was investigated retrospectively, to identify independent factors for the perioperative mortality (hospital mortality). Chi-square single factor analysis was employed to test 35 independent variables, in a second analysis stepwise logistic regression was used to determine the factors correlating with the morality. In 41 patients the esophagus was resected by a thoraco-abdominal approach, in 86 by blunt dissection. The hospital mortality was 17.3%, the 30-day mortality 12.6%. Single factor analysis revealed a significant correlation with the variable smoking (p < 0.003), post-operative morbidity (p < 0.008), CEA (p < 0.02), time of operation (p < 0.02) and surgical procedures (p < 0.02). The influence of the surgeon's experience was significant. In the stepwise logistic regression the duration of operation (p < 0.0039), the surgical procedures (p < 0.016), and the units of blood (p < 0.03) were correlated with mortality. The logistic regression showed a significant increase of mortality for the thoracoabdominal approach with increasing duration of the operation. The estimation of survival time according to Kaplan and Meier revealed a significant correlation with the stage of the tumor, but not with the surgical procedure.