Exploration of three different nutritional scores in predicting postoperative complications after pancreaticoduodenectomy

Objectives: we used the Controlling Nutritional Status score (CONUT), Geriatric Nutritional Risk Index (GNRI), and Prognostic Nutritional Index (PNI) to explore three different nutritional scores in predicting postoperative complications after pancreaticoduodenectomy (PD). Methods: data were retrospectively reviewed from 113 patients who underwent PD to treat pancreatic cancer and periampullary neoplasms at the Second Affiliated Hospital of Soochow University between 2015 and 2020. Nutritional status was assessed by the CONUT, GNRI, and PNI scores, and patients were categorized as either at risk or not at risk for malnutrition by each score. Postoperative complications were defined according to the Clavien-Dindo classification. Data were analyzed using Fisher's exact probability method and multivariate logistic regression analysis. The relationships between the three nutritional scoring systems and postoperative complications were examined. Results: CONUT, GNRI and PNI scores were closely related to the occurrence of postoperative complications. CONUT (OR = 0.92, 95 % CI, 0.751.12, p = 0.043), GNRI (OR = 0.98, 95 % CI, 0.93-1.02, p = 0.036), PNI (OR = 0.96, 95 % CI, 0.89-1.03, p = 0.024), and operation periods (OR = 1.01, 95 % CI, 0.99-1.02, p = 0.034) were independent risk factors for complications in patients after PD. The predictive value of the three nutritional screening methods for overall complications in patients with PD had a sensitivity of 31.8 %, 56.06 % and 74.24 %, a specificity of 85.10 %, 68.08 % and 76.81 %, a Youden index of 0.17, 0.24 and 0.71, and a kappa value of 0.460, 0.389 and 0.472, respectively. The predictive value of the three nutritional screening methods in predicting the severity of complications in patients with PD had a sensitivity of 82.11 %, 58.95 % and 65.26 %, a specificity of 38.89 %, 55.56 % and 66.67 %, a Youden index of 0.21, 0.15 and 0.36, and a kappa value of 0.664, 0.416 and 0.645, respectively. Among the three nutrition scoring systems, PNI score had better diagnostic efficiency (0.660 area under the AUC curve), higher specificity (66.67 %), and was consistent with postoperative complications (KCONUT = 0.664, KGNRI = 0.416, KPNI = 0.645) when compared to the GNRI and CONUT scores. Conclusions: CONUT, GNRI and PNI scores, especially PNI score, have good predictive values for the occurrence and severity of postoperative complications in PD patients, and should be used as preoperative nutritional risk screening tools for PD patients.