A novel simplified scoring system for predicting mortality in emergency colorectal surgery

prediction model development

Abstract

BACKGROUND: Despite advances in surgical approaches, emergency colorectal surgery has high mor-tality and morbidity. OBJECTIVE: We aimed to create a simple and distinctive scoring system, for predicting mortality among patients undergoing emergency colorectal surgery. DESIGN AND SETTING: Prediction model development study based on retrospective data-gathering. METHODS: Patients who underwent emergency colorectal surgery between March 2014 and December 2016 at a single tertiary-level referral center were included in our study. Patient demographics, comorbidities, type of surgery, etiology and laboratory and radiological findings were collected retrospectively and ana-lyzed. A new clinical score (named the Numune emergency colorectal resection score) was constructed from the last logistic regression model, in which one point was assigned for the presence of each predictive factor. RESULTS: 138 patients underwent emergency colorectal surgery. These comprised 64 males (46.4%) and 74 females (53.6%), with a mean age of 64 years. Multivariate analysis revealed that blood urea nitrogen level > 65 mg/dl (odds ratio, OR: 8.03; 95% confidence interval, CI: 2.16-15.77), albumin level < 0.7 mg/­dl­ (OR: 4.43; 95% CI: 1.96-14.39) and American Society of Anesthesiologists score ≥ 3 (OR: 3.47; 95% CI: 0.81‑9.18) were associated with postoperative complications. The Numune score was graded from I to III. The risk of mortality was found to be 63.2% in the group with grade III, which accounted for 35.2% of the subjects. There were 37 postoperative deaths. CONCLUSIONS: Surgeons need scoring systems, especially to predict postoperative mortality. We pro-pose the Numune emergency colorectal resection score for emergency surgical procedures as a practical, usable and effective system for predicting postoperative morbidity.