Background: Gastro-oesophageal surgery is associated with appreciable postoperative morbidity and mortality. POSSUM (Physiological and Operative Severity Score for the enUmeration of Mortality and Morbidity) and its related models P-POSSUM and O-POSSUM have been developed to predict such events in general surgery. The aim was to undertake the first systematic review of the use of these models in gastro-oesophageal surgery patients.
Methods: An online database search was carried out from 1991 to December 2008.
Results: Twenty-two published studies in gastro-oesophageal cancer surgery were identified. Twelve studies were found not to address the above aim, leaving ten relevant publications for analysis. Pooled data from these studies showed the weighted observed-to-expected ratio (O/E) for postoperative mortality using POSSUM (n = 1189), P-POSSUM (n = 2314), and O-POSSUM (n = 1755) was 0.37, 0.83, and 0.51, respectively. The weighted O/E for morbidity using POSSUM (n = 1038) was 0.86.
Conclusion: POSSUM and O-POSSUM most significantly overestimated postoperative mortality in gastro-oesophageal cancer patients. In contrast, P-POSSUM had the least overestimation and may be the most useful predictor of likely postoperative mortality in these patients.