Back to Program | 2010 Program and Abstracts Overview | 2010 Posters

Background: The aim of the study was to evaluate the use of the Physiological and Operative Severity Score for the enUmeration of Mortality and morbidity (POSSUM), Portsmouth (P) POSSUM and upper gastrointestinal (O) POSSUM models in patients undergoing elective resection for gastro-oesophageal cancer. Patients and methods: Elective gastro-oesophageal cancer resections in Glasgow Royal infirmary from January 2005 to May 2009 were scored by POSSUM, P-POSSUM and O-POSSUM methods. Observed mortality rates were compared to predicted mortality rates in three risk groups for each model using the Hosmer-Lemeshow goodness-of-fit test. The power to discriminate between patients who died and those who survived was assessed using the area under the receiver-operator characteristic (ROC) curve.Results: The observed mortality rate was 4.1 per cent, whereas rates predicted by POSSUM, P-POSSUM and O-POSSUM were 16.5, 5.8 and 9.9 per cent respectively. POSSUM over predicted mortality in all risk groups, whereas O-POSSUM and P-POSSUM over predicted mortality in low risk groups. The area under ROC curve for P-POSSUM (0.808) showed good discriminatory power than POSSUM (0.759), and O-POSSUM (0.715). The observed morbidity was 48.8 per cent, whereas POSSUM predicted post-operative complication in 59.5 percent. POSSUM morbidity equation showed poor discriminatory power (AUC 0.639).Conclusion: All the POSSUM models over-predicted post-operative mortality accurately. P-POSSUM least over predicted the mortality with better discriminatory power in patients undergoing gastro-oesophageal cancer resection.