The aim for this systematic review was to summarize evidence regarding the prevalence of NR and the predictive validity of different tools applied for NR screening of critically ill patients.
The PubMed, Embase, and Scopus databases were searched up to December 2019 using the subject headings related to critically ill patients and NR screening. The current systematic review is registered with PROSPERO (identifier: CRD42019129668).
Data on NR prevalence, predictive validity of nutritional screening tools, and interaction between caloric-protein balance and NR in outcome prediction were collected.
Results were summarized qualitatively in text and tables, considering the outcomes of interest.
From 15 669 articles initially identified, 36 fulfilled the inclusion criteria, providing data from 8 nutritional screening tools: modified Nutrition Risk in the Critically Ill (mNUTRIC; n = 26 studies) and Nutritional Risk Screening-2002 (NRS-2002; n = 7 studies) were the most frequent; the NR prevalence was 55.9% (range, 16.0% to 99.5%). Nutritional risk was a predictor of 28-day and ICU mortality in 8 studies. Interactions between caloric-protein balance and NR on outcome prediction presented were scarcely tested and presented heterogeneous results (n = 8).
Prevalence of NR in patients in the ICU varies widely; a satisfactory predictive validity was observed, especially when mNUTRIC or NRS-2002 were applied.