Perioperative mortality and morbidity prediction using POSSUM, P-POSSUM and APACHE II in Chinese gastric cancer patients: surgical method is a key independent factor affecting prognosis

Abstract

Gastric cancer is the fourth most common cancer worldwide. Predicting morbidity and mortality is important in deciding timing of surgery and type of surgery offered. APACHE II, POSSUM, and P-POSSUM are the most reliable scoring methods in use today. This is the first paper to evaluate the utility of all three scoring systems in China. We collected data on 851 patients (583 male and 268 female) who underwent surgery between 1991 and 2011. Physiological and pathological data was entered in spreadsheet format and analyzed using STATA version 11.0 to generate ROC curves for each scoring system. In predicting mortality, P-POSSUM and POSSUM were most effective and APACHE II was ineffective. POSSUM predicted a higher morbidity risk than was actually encountered. Age and type of operation were found to be independent risk factors for mortality. The utility of the APACHE II score in gastric cancer patients is limited. APACHE II is suitable for considering group versus individual effect. The POSSUM score is useful in general surgery, but needs improvement. We found the P-POSSUM score to be superior for morbidity and mortality prediction. P-POSSUM is useful for both the general population and for a specific cohort. The type of surgery is a key decision point for surgeons, and independently affects prognosis. Based upon these findings and clinical scoring systems, clinicians can develop individualized treatment algorithms.