Nomogram-Derived Prediction of Postoperative Ileus after Colectomy: An Assessment from Nationwide Procedure-Targeted Cohort

Abstract

Postoperative ileus (POI) is a clinical burden to health-care system. This study aims to evaluate the incidence and predictors of POI in patients undergoing colectomy and create a nomogram by using recently released procedure-targeted nationwide database. Patients who underwent elective colectomy in 2012 and 2013 were identified from American College of Surgeons National Surgical Quality Improvement Program using the new procedure-targeted database. Demographics, comorbidities, and 30-day postoperative outcomes were evaluated. Variables in the final step-wise multiple logistic regression model for each outcome were selected in a stepwise fashion using Akaike's information criterion. A nomogram was created to aid in the calculation of POI risk for individual patients. A total of 29,201 patients met the inclusion criteria; 3834 (13.1%) developed POI with a male predominance (55.9%). Patients who developed ileus had longer length of hospital stay (11 vs 5 days; P < 0.001) and operative time (200 vs 174 minutes; P < 0.001). In the stepwise logistic regression model, the following variables were found to be independent risk factors for POI: older age (P < 0.001), male gender (P < 0.001), American Society of Anesthesiologists class III/IV (P < 0.001), open approach (P < 0.001), preoperative septic conditions (P < 0.001), omission of oral antibiotic before surgery (P < 0.001), right colectomy or total colectomy vs other procedures (P < 0.001), smoking (P = 0.001), decreased preoperative serum albumin level (P < 0.001), and prolonged operating time (P < 0.001). All postoperative complications were more frequently occurred in patients with POI. The nomogram accurately predicted POI with a concordant index for this model of 0.69. The use of minimal invasive techniques, control of preoperative septic conditions, oral antibiotic bowel preparation and shorter operative time are associated with a decreased rate of POI. External validation is essential for the confirmation and further evaluation of our logistic regression model and nomogram.