Background: This study assessed the APACHE II (Acute Physiology and Chronic Health Evaluation II), SAPS II (Simplified Acute Physiology Score-II), POSSUM (Physiologic and Operative Severity Score for Enumeration of Morbidity and Mortality), and P-POSSUM (Portsmouth-POSSUM) in patients with colorectal cancer undergoing curative or palliative resection.
Methods: Predicted mortality rates and the observed/expected mortality ratio were computed by means of each scoring system. The results were compared between survivors and nonsurvivors and between elective and emergency operations. Each model was assessed for its accuracy to predict the risk of death using receiver operator characteristic (ROC) curve analysis, and risk stratification was generated as well.
Results: Some 224 patients were enrolled in the study. The overall 30-day mortality rate was 3.6% (n = 8). Predicted mortality rates generated by APACHE II, SAPS II, POSSUM, and P-POSSUM were 9.1%, 3.7%, 13.4%, and 5.2%, respectively. All the scoring systems assigned higher scores to those patients who died than to those who survived. Areas under the curve calculated by ROC curve analysis for APACHE II, SAPS II, POSSUM, and P-POSSUM were 0.786, 0.854, 0.793, and 0.831, respectively. Best stratification was achieved by the SAPS II score.
Conclusions: SAPS II and P-POSSUM were determined to be better predictors for patients with colorectal cancer undergoing resection. SAPS II also was found to have a higher degree of discriminatory power in colorectal resection for carcinoma. The predictive value of this useful severity score in several surgical subgroups must be examined to evaluate its routine use in risk-adjusted audit.