Application of four nutritional risk indexes in perioperative management for esophageal cancer patients

PURPOSE: The Prognostic Nutritional Index (PNI), Nutritional Risk Index (NRI), Geriatric Nutritional Risk Index (GNRI), and Controlling Nutritional Status (CONUT) score were devised for quantifying nutritional risk. This study evaluated their properties in detecting compromised nutrition and guiding perioperative management of esophageal cancer patients. METHODS: A prospective institutional database of esophageal cancer patients was reviewed and analyzed. Compromised nutritional status was defined as PNI < 50, NRI < 97.5, GNRI < 92, or CONUT score ≥ 4, respectively. The malnutrition diagnosis consensus established by the European Society of Clinical Nutrition and Metabolism (ESPEN 2015) was selected as reference. Multivariable logistic regression and receiver operating characteristic curve analysis were used. External validation was conducted. RESULTS: After reviewing the 212-patient database, 192 patients were finally included. Among the four nutritional indexes, the GNRI < 92 showed highest sensitivity (72.0%), specificity (78.9%), and consistency (AUC 0.754, 95% CI 0.672–0.836) with malnutrition diagnosed by ESPEN 2015. The GNRI < 92 showed comparable performance with ESPEN 2015 in recognizing decreased fat mass, fat-free mass, and skeletal muscle mass (all P < 0.01). Both the GNRI < 92 and ESPEN 2015 showed good property in predicting major complications, infectious complications, overall complications and delayed hospital discharge (all P < 0.01), better than PNI < 50, NRI < 97.5, and CONUT score ≥ 4. Regarding the external validation, a retrospective analysis of 155 esophageal cancer patients confirmed the better performance of GNRI < 92 in predicting perioperative morbidities than other 3 nutritional indexes. CONCLUSION: The GNRI was optimal in perioperative management of esophageal cancer patients among the four nutritional indexes and was an appropriate alternative to ESPEN 2015 for simplifying nutritional assessment. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1007/s00432-021-03585-8.