Nutritional indices at admission are associated with mortality rates of patients in the intensive care unit.

Abstract

Malnutrition is a common occurrence in critically ill patients, and has been related to poor prognosis in various diseases. Here, we assess the prognostic value of malnutrition using nutritional indices in intensive care units (ICU) patients. We retrieved information on 2060 patients from the Medical Information Mart for Intensive Care III, and randomized the patients into training and validation cohorts, at a ratio of 7:3. We estimated their nutritional indices using prognostic nutritional index (PNI), geriatric nutritional risk index (GNRI), and controlling nutritional status (CONUT) score. Both multivariate regression analysis and the Kaplan-Meier (KM) survival curve were used to examine the prognostic role of nutritional indices in ICU mortality. Then we evaluated the additional predictive significance of each nutritional index beyond the baseline model including conventional risk factors. Multivariate regression analysis revealed that PNI, GNRI, and CONUT were independent predictors of in-hospital and 1-year mortality (all P < 0.001). KM curves showed higher 1-year mortality rates in having nutritional risk patients (PNI ≤ 38 or GNRI ≤ 98 or CONUT ≥ 2). Moreover, subgroup analyses revealed a significant association between each nutritional index and 1-year mortality in patients with different comorbidities. We also observed a pronounced additional impact on the predictive value of 1-year mortality when PNI, GNRI, and CONUT were separately added to the baseline model. The additional role of each nutritional index was further verified in the validation cohort. Our results revealed that the nutritional indices at admission are significantly correlated with increased mortality rates in ICU adult patients.