Enhanced recovery after surgery in colon and rectal surgery: identification of predictive variables of failure in a monocentric series including 733 patients

To identify factors associated with early deviation and delayed discharge within an Enhanced Recovery after Surgery (ERAS) pathway. This is a retrospective review of prospectively collected data of consecutive patients who underwent laparoscopic or open colorectal surgery and managed with a standardized ERAS pathway between April 2015 and October 2018. ERAS items were assessed within 48 h after surgery. Patients with early complications were excluded. The influence of factors on length of stay was calculated by univariate and multivariate analysis. A binary logistic regression was used to model a predicting score. Seven hundred and thirty-three patients met the inclusion criteria. Multivariate analysis showed that age >= 75 years (P = 0.02), ASA score >= 3 (P = 0.03), open surgery or conversion to open (P = 0.001), non-compliance with the intra-operative balanced fluid therapy (P = 0.049), failure to early removal of the urinary catheter (P = 0.001), to discontinue IV fluid (P = 0.02) and to early mobilization (P = 0.001) were independently associated with ERAS failure. The generated score had a specificity of 84% and a positive predictive value of 72%. Patients who would have a length of stay longer than the median for each surgical procedure were properly identified (Area under ROC Curve = 0.753,P < 0.001). The delayed discharge could be predicted at 48 h from the intervention. The ability of the model to weight the specific role of each statistically significant variable might be a useful tool to identify the most frail patients.