Nutritional screening for risk prediction in patients scheduled for extra-abdominal surgery.

The Nutritional Risk Screening-2002 (NRS-2000) is currently recommended by the European Society of Parenteral and Enteral Nutrition as a screening tool in hospitalized patients. However, for preoperative risk prediction, the usefulness of this tool is uncertain and may depend on the type of surgical disease. The present study investigated the relative prognostic importance of the NRS-2002 and of established medical and surgical predictors for postoperative complications in patients scheduled for non-abdominal procedures.In this prospective observational study, we enrolled 581 patients scheduled for elective non-abdominal surgery. Data were collected on nutritional variables (body mass index, weight loss, and food intake), age, gender, type of surgery, extent of surgery, underlying disease, American Society of Anesthesiologists class, and comorbidity. We also evaluated a modification of the NRS-2002 (ordinal graduation according to <2 or ≥2 points) and the importance of individual parameter values. Relative complication rates were calculated with generalized linear models and cumulative proportional odds models.Forty-four patients (7.6%) sustained at least one postoperative complication. The frequency of this event increased significantly with a higher NRS-2002 score. However, the model that performed the best (sensitivity 81.8%, specificity 78.6%) included the modified NRS-2002 graduation (<2 or ≥2 points) and other factors such as American Society of Anesthesiologists class, the duration of the procedure, and the need for red blood cell transfusion.In surgical patients with non-abdominal diseases, a modified NRS-2002 classification may be required to preoperatively identify patients at a high nutritional risk. The NRS-2002 alone is insufficient to precisely predict complications.