Risk evaluation of postoperative complication in patients undergoing rectal cancer surgery.

Abstract

756 Background: Colorectal cancer patients with postoperative complications have poor long-term outcomes. Rectal cancer is associated with a higher incidence of complications than colon cancer. Reliable predictors of complications would facilitate determination of surgical procedures such as stoma creation. We examined the relations between five kinds of risk score and postoperative complications in patients(pts) with rectal cancer. Methods: The subjects were 261 pts with cStage II/III rectal cancer from 2003 to 2013. We categorized complications into all complications, leakage, infectious complications and intestinal obstruction. Estimation of physiologic ability and surgical stress comprehensive risk scores (E-PASS CRS), surgical Apgar scores (SAS), prognostic nutritional index (PNI), colorectal physiological and operative severity scores for the enumeration of mortality and morbidity (CR-POSSUM), and neutrophil-lymphocyte ratios (NLR) were assessed. Clavien-Dindo (CD) Grade 3a or higher requiring surgically invasive treatment were considered complications. For leakage, CD Grade 3b or higher were considered complications. Results: Complications occurred in 56 pts (21%), leakage (L) in 12 (14%), infectious complications (IC) in 19 (7%), and intestinal obstruction (IO) in 16 (6%). E-PASS CRS significantly correlated with all complications (OR 3.45; p < 0.001), IC (OR 0.26; p = 0.008), L (OR 4.94; p = 0.027), and IO (OR 3.92; p = 0.007). PNI correlated with all complications (OR 0.38; p = 0.002) and IO (OR 3.08; p = 0.024). NLR correlated with all complications (OR 0.40; p = 0.003), IC (OR 0.25; p = 0.004), L (OR 8.66; p < 0.001), and IO (OR 3.86; p = 0.007). SAS correlated with IC (OR 0.19; p = 0.004). CR-POSSUM correlated with all complications, (OR 2.26; p = 0.009), IC (OR 5.02; p < 0.001), and IO (OR 0.28; p = 0.014). Multivariate analysis revealed E-PASS CRS (OR 6.17 p = 0.020) and NLR (OR 7.11 p = 0.011) were independent risk factors for L. Conclusions: Five kinds of risk score were all useful for assessing the risk of complications in rectal cancer. NLR was the only risk factor for leakage that could be used before surgery. Our results suggest NLR might be an auxiliary means of evaluating the need for creation of diverting stoma.