Predicting postoperative mortality in patients undergoing colorectal surgery.

Well-known and suitable instruments for surgical audit are the POSSUM and P-POSSUM scoring systems. But these scores have not been well validated across the countries. The objective of the present study was to assess the predictive value of scores for colorectal surgery in France. Patients operated on for colorectal malignant or diverticular diseases, whether electively or on emergency basis, within a 4-month period were included in a prospective multicenter study conducted by the French Association for Surgery (Association Française de Chirurgie, AFC). The main outcome measure was postoperative in-hospital mortality. Independent factors leading to death were assessed by multivariate logistic regression analysis (AFC-index). The ratio of expected versus observed deaths was calculated, and the predictive value of the POSSUM and P-POSSUM scores were analyzed by the receiver operating characteristic (ROC) curve. A total of 1426 patients were included. The in-hospital death rate was 3.4%. Four independent preoperative factors (AFC-index) have been found: emergency surgery, loss of more than 10% of weight, neurological disease history, and age > 70 years. POSSUM had a poor predictive value; it overestimated postoperative death in all cases. P-POSSUM had a good predictive value, except for elective surgery, where it overestimated postoperative death twofold. The predictive value of the AFC-index was also good. It had the same sensitivity and specificity as the P-POSSUM. POSSUM has not been validated in France in the field of colorectal surgery. P-POSSUM was as predictive as the AFC-index which is a simpler instrument based on four clinical parameters (without any mathematical formulas).