Relation between nutritional status on clinical outcomes of critically ill patients: emphasizing nutritional screening tools in a prospective cohort investigation

Background Malnutrition is a significant concern reported in adult critically ill patients, yet there is no gold standard to assess nutritional status in this population. This study examines the association between nutritional status and clinical outcomes in intensive care unit (ICU) patients using nutritional risk assessment tools and aims to look for the best tool.Method In a single-center prospective cohort study among 165 patients, the predictive performance of high or low malnutrition risk assessed by Nutritional Risk Screening (NRS), Modified Nutrition Risk in Critically Ill (m-NUTRIC), Mini-Nutritional-Assessment Short-Form (MNA-SF), Controlling Nutritional status (CONUT), and Prognostic Nutritional Index (PNI) were evaluated and compared for mortality, organ failure, length of hospitalization, and mechanical ventilation (MV).Results Different assessment tools showed various nutritional statuses. m-NUTRIC and NRS-2002 were found to be associated more strongly relative to other tools with mortality (RR = 1.72; 95% CI, 1.42-2.08) and (RR = 1.37; 95% CI, 1.08-1.72), organ failure (RR = 1.69; 95% CI, 1.44-1.96) and (RR = 1.22; 95% CI, 0.99-1.48), MV (RR = 1.46; 95% CI, 1.27-1.65) and (RR = 1.21; 95% CI, 1.04-1.39) respectively. There was no correlation between malnutrition levels assessed by mentioned tools except for NRS-2002 and length of hospitalization. In predicting mortality or illness severity, the cut points were different for some tools like NUTRIC-score and all assessed outcomes (3.5), MNA-SF and mortality (6.5), CONUT with mortality, and MV (6.5).Conclusions A considerable proportion of patients admitted to the ICU are at high risk for malnutrition. Compared to other tools, m-NUTRIC and NRS-2002 proved superior in predicting clinical outcomes in critically ill patients. Other tools overestimated the risk of malnutrition in the ICU so couldn't predict clinical outcomes correctly.