The correlation of nutrition risk index, nutrition risk score, and bioimpedance analysis with postoperative complications in patients undergoing gastrointestinal surgery

Abstract

Malnutrition in gastrointestinal (GI) surgery is associated with increased morbidity. Therefore, careful screening remains crucial to identify patients at risk for malnutrition and consequently postoperative complications. The aim of this study was to evaluate the ability of 3 established score systems to identify patients at risk of developing postoperative complications in GI surgery and to assess the correlation among the score systems. We evaluated prospectively 200 patients admitted for elective GI surgery using (1) nutrition risk index, (2) nutrition risk score, and (3) bioelectrical impedance analysis. Complications were assessed using a standardized complication classification. The findings of the score systems were correlated with the incidence and severity of complications. Parametric and nonparametric correlation analysis was performed among the different score systems. All 3 score systems correlated significantly with the incidence and severity of postoperative complications and the duration of hospital stay. Using multiple regression analysis, only nutrition risk score and malignancy remained prognostic factors for the development of complications with odds ratios of 4.2 (P = .024) and 5.6 (P < .001), respectively. The correlation between nutrition risk score and nutrition risk index was only moderate (Pearson coefficient = 0.54). Bioelectrical impedance analysis displayed only weak to trivial correlation to the nutrition risk index (0.32) and nutrition risk score (0.19), respectively. The nutrition risk score, nutrition risk index, and bioimpedance analysis correlate with the incidence and severity of perioperative complications in GI surgery. The nutrition risk score was the best score in predicting patients who will develop complications in this study population. The correlation between the individual scores was only moderate, and therefore, they do not necessarily identify the same patients.