Are all low-NUTRIC-score patients the same? Analysis of a multi-center observational study to determine the relationship between nutrition intake and outcome

Abstract

The NUTrition Risk in the Critically Ill (NUTRIC) scoring system is a tool useful, discriminating critically-ill patients benefiting from optimal nutrition intake (>80% of prescription). Recent recommendations advocate for withholding artificial nutrition among low-NUTRIC patients, however, we hypothesized that some low-NUTRIC patients would show an association between nutrition intake and outcome. Patients were selected from the 2013-2014 International Nutrition Surveys when ICU length of stay (LICU) ≥72h, baseline mNUTRIC score ≤4 and had at least three evaluable nutrition days (N=2781). Proportion of prescription received during evaluable days was associated to 60-day hospital mortality by a logistic regression modelling. A priori, we expected that the association between proportion of prescription received and mortality might differ according to: LICU, BMI and prior unintentional weight loss or reduced oral intake. A total of 2781 patients fulfilled the inclusion criteria and participated in the study. Ten percent of the sample had a BMI <20kg/m2 and 20% experienced either unintentional weight loss during the last 3 months, or reduced food intake over the last week. Sixty-day hospital mortality was 15% and median LICU reached 11.3 [6.3-21.7] days. Mean total prescription received by any means of nutritional support during the first 12 evaluable days was 57.4±28.1% for energy and 53.7±29.2% for protein. In the pooled, subgroup and sensitivity analyses, no significant associations were identified. Low-NUTRIC (≤4) patients demonstrate a prolonged LICU, while experiencing significant mortality and a high prevalence of malnutrition risk factors. Although improvements in mortality were not achieved with increased nutritional intake, this should not be construed as a rationale for withholding artificial nutrition among this patient group.