Comparison of different definitions of feeding intolerance: A retrospective observational study

A. Reintam Blaser*, L. Starkopf, A.M. Deane, M. Poeze, J. Starkopf

*Corresponding author for this work

Research output: Contribution to journalArticleAcademicpeer-review

Abstract

BACKGROUND & AIMS: While feeding intolerance (FI) is clinically important in the critically ill it is inconsistently defined. By evaluating definitions of FI based on relationships between symptoms and signs of gastrointestinal (GI) dysfunction and mortality the objective was to define FI using the definition that was most strongly associated with subsequent mortality. METHODS: Data from all adult patients admitted to a single ICU between 2004 and 2011, and who were receiving enteral nutrition (EN), were analysed. The amount of EN administered, presence of absent bowel sounds (BS), vomiting and/or regurgitation, diarrhoea, bowel distension, and large gastric residual volumes (GRVs) were documented daily. A GRV >/=500 ml/day was considered as large and the sum of gastrointestinal (GI) symptoms including large GRV was calculated daily. Various definitions of FI were modelled. Definitions using only GRV, or GRV with other GI symptoms, or GRV and failure to reach preset EN targets were evaluated. The predictive power of FI on mortality was tested by adding the presence of FI (different definitions were tested one-by-one) into multiple regression analyses together with admission day demographic and severity of illness variables. RESULTS: Of the 1712 patients included, 221 (12.9%) died in ICU and 495 (28.9%) had died within 90 days after ICU admission. The definition of FI based on the presence of at least three out of five GI symptoms was most strongly related to ICU-mortality (6.3% prevalence in survivors vs. 23.5% in non-survivors, p < 0.001, odds ratio (95%CI) 3.39 (2.23-5.14)), whereas EN <23% of caloric target was the strongest predictor for mortality 90 days after admission (50.7% prevalence among survivors vs 75.2% in non-survivors, p < 0.001, odds ratio (95% CI) 2.34 (1.80-3.04)). CONCLUSIONS: FI is associated with increased mortality but the strength of this relationship depends on the definition used. The 'best' definition of FI for prediction of ICU-mortality is based on a complex assessment of GI symptoms (including large GRV), whereas enteral underfeeding is the definition of FI that is the strongest predictor of death within 90 days of admission. Our 'best' definitions are not immediately generalizable, but should help building up future studies.
Original languageEnglish
Pages (from-to)956-961
JournalClinical Nutrition
Volume34
Issue number5
DOIs
Publication statusPublished - 1 Jan 2015

Cite this