Comparison of the predictive value of NUTRIC and modified NUTRIC scores for ICU mortality in patients with sepsis: a single-center prospective cohort study

Both the Nutrition Risk in the Critically Ill (NUTRIC) score and modified NUTRIC (mNUTRIC) scores are commonly used tools for evaluating nutritional risk and prognosis in critically ill patients. However, direct comparative studies on their predictive value for ICU mortality in patients with sepsis remain scarce, and no consensus has been reached regarding differences in their predictive efficacy. This study aimed to compare the predictive value of the NUTRIC and mNUTRIC scores for ICU mortality in patients with sepsis, providing a reliable reference for clinical prognosis assessment of these patients. This was a single-center prospective cohort study that enrolled patients with sepsis admitted to the Department of Critical Care Medicine, the First Affiliated Hospital of Guangxi Medical University, between November 2024 and June 2025. The primary outcome was ICU mortality among sepsis patients during their ICU stay. Cox regression models were employed to analyze the associations of the NUTRIC score and mNUTRIC score with ICU mortality. Restricted cubic spline (RCS) analysis was performed to investigate the dose-response relationship between the two scores and the risk of ICU mortality. Receiver operating characteristic (ROC) curves were constructed, and the Delong test was used to compare the area under the curve (AUC) for assessing the discriminative ability of the scores. Decision curve analysis (DCA) was conducted to quantify the clinical net benefit of the two scores at various risk thresholds. A total of 259 patients with sepsis were enrolled, of whom 47 died during ICU stay, resulting in an ICU mortality rate of 18.1%. Multivariate Cox regression analysis showed that both the NUTRIC and mNUTRIC scores were significantly associated with ICU mortality risk. After full adjustment for confounding factors, each 1-point increase in the NUTRIC score was associated with a 1.83-fold higher risk of ICU mortality (HR = 1.83, 95%CI: 1.48–2.26, P < 0.001), and each 1-point increase in the mNUTRIC score was associated with an 72% higher risk of ICU mortality (HR = 1.72, 95%CI: 1.37–2.17, P < 0.001). RCS analysis demonstrated a significant linear association between both scores and ICU mortality risk (NUTRIC score: Pnon−linearity=0.313; mNUTRIC score: Pnon−linearity=0.343). ROC curve analysis showed that the AUC values of the NUTRIC and mNUTRIC scores for predicting ICU mortality were 0.784 (95%CI: 0.719–0.848) and 0.761 (95%CI: 0.693–0.830), respectively. The Delong test indicated no statistically significant difference in discriminative efficacy between the two scores (P = 0.069). The optimal cut-off values for predicting ICU mortality were 5.5 points for both the NUTRIC and mNUTRIC scores. DCA revealed that when the threshold probability ranged from 1% to 80%, the clinical net benefit of using the NUTRIC score to predict ICU mortality was superior to the “all predict” and “no predict” strategies; when the threshold probability ranged from 2% to 70%, the clinical net benefit of using the mNUTRIC score was superior to the “all predict” and “no predict” strategies. Both the NUTRIC and mNUTRIC scores are independent influencing factors for ICU mortality risk in patients with sepsis, and both scores are linearly and positively correlated with ICU mortality. There is no statistically significant difference in their predictive value for ICU mortality in patients with sepsis. In clinical practice, either score can be selected according to actual needs for early prognostic screening of patients with sepsis, thereby providing a basis for improving patient outcomes.