LAUSR

BACKGROUND There are multiple definitions for malnutrition, without evidence of superiority of any one definition to assess preoperative risk. Therefore, to aid in identification of patients that might warrant prehabilitation we aimed to determine the optimal definition of malnutrition before major oncologic resection for six cancer types. METHODS The ACS-NSQIP database was queried for patients undergoing elective major oncologic surgery from 2005-2017. Nutritional status was evaluated using the European Society for Parenteral and Enteral Nutrition (ESPEN) definitions, NSQIP's variable for >10% weight loss over the prior 6 months, and the World Health Organization body mass index (BMI) classification system. Multivariable logistic regression was performed to evaluate the adjusted effect of nutritional status on mortality and major morbidity. RESULTS 205,840 operations were identified (74% colorectal, 10% pancreatic, 9% lung, 3% gastric, 3% esophageal, and 2% liver). A minority (16%) of patients met criteria for malnutrition (0.6% severe malnutrition, 1% ESPEN 1, 2% ESPEN 2, 6% NSQIP, 6% mild malnutrition), 31% were obese, and the remaining 54% had a normal nutrition status. Mortality and major morbidity both varied significantly between the nutrition groups (both p<0.0001). An interaction between nutritional status and cancer type was observed in the models for mortality and major morbidity (interaction term p<0.0001 for both) indicating the optimal definition of malnutrition varied by cancer type. CONCLUSION The definition of malnutrition used to assess postoperative risk is specific to the type of cancer being treated. These findings can be used to enhance nutritional preparedness in the preoperative setting.