Nutritional assessment of critically ill patients: validation of the modified NUTRIC score.

Background/objectives: In order to identify critically ill patients with high nutritional risk the modified NUTrition Risk in the Critically ill (mNUTRIC)-score was developed. This score aims to identify patients that will benefit from nutritional interventions. Few data are available on its validity. In The Netherlands, the MUST-score, a nutritional assessment tool for non-ICU patients, is commonly used in the ICU. To validate the mNUTRIC-score in Dutch ICU patients and compare its prognostic performance with the MUST-score.
Subjects/methods: A single-center retrospective cohort study among 475 mechanically ventilated patients. Prognostic performance of the mNUTRIC and MUST-scores were assessed and compared for discriminative abilities for 28-day mortality and prolonged mechanical ventilation (>2 days).
Results: The discriminative ability of the mNUTRIC-score for 28-day mortality is (ROC-AUC) 0.768 (95% CI 0.722-0.814) with an associated LR+ of 1.73 (95% CI 1.53-1.95) and LR- of 0.24 (95% CI 0.14-0.39) when comparing low with high (>4) scores. Comparing low with high MUST-scores (>1) a ROC-AUC of 0.513 (95% CI 0.445-0.587) and LR+ of 1.05 (95%CI 0.77-1.45) and LR- of 0.97 (95% CI 0.71-1.17) for mortality were found. The discriminative ability for prolonged ventilation was 0.666 (95% CI 0.616-0.716) and 0.532 (95% CI 0.469-0.594) for the mNUTRIC and MUST-scores, respectively.
Conclusions: The prognostic performance of the mNUTRIC-score for 28-day mortality is fair and comparable to other validation studies. The association with prolonged ventilation was not confirmed by our results. The mNUTRIC-score has better performance than the commonly used MUST-score. Therefore, we suggest abandoning use of the MUST-score and to recommend introduction of the mNUTRIC-score for the nutritional risk assessment of critically ill patients.