Abstract

Objective: To document factors that significantly effect operative mortality and survival following oesophageal resection for carcinoma.

Methods: Retrospective analysis of 591 patients (408 males, 183 females, mean age 66) notes under the care of one surgeon who underwent an oesophageal resection for carcinoma between 1974 and 1996.

Results: In-hospital mortality was 52 (8.7%). This has decreased to less than 5% for the past decade. Age (P = 0.01), preoperative serum alkaline phosphatase level (P = 0.04) and site of tumour (P = 0.02) were significantly associated with in-hospital mortality (logistic regression). Survival (± SEM) at 1,2 and 5 years was 53.9% (± 2.1), 33.3% ± 2.0), and 15.4% (± 1.6) respectively. There was a significant difference in survival between patients grouped according to TNM stage (P < 0.0001), tumour differentiation (P < 0.0001), completeness of resection (P < 0.0001) and performance of splenectomy (P < 0.0001). Using a Cox proportional hazards model, which makes allowance for each variable, age (P = 0.01), TNM stage (P < 0.0001), completeness of resection (P = 0.02) and splenectomy (P = 0.008) were found to significantly affect survival.

Conclusion: Early diagnosis is important if reasonable survival is to be attained. Irrelevant of the stage of the tumour splenectomy adversly affects survival. Care must therefore be taken to preserve the spleen.