Comparative analysis of four nutritional scores in predicting hospital stay duration for EICU Patients with acute pancreatitis

Background Nutritional status significantly impacts outcomes in critically ill patients. Nutritional assessment tools such as the prognostic nutritional index (PNI), controlling nutritional status (CONUT) scores, the Nutritional Risk Screening (NRS-2002), and the Triglyceride and Body Weight Index (TCBI) scores might have prognostic significance in acute pancreatitis (AP). This study aims to investigate the association between these admission scores and the length of hospital stay, as well as their predictive value, in AP patients admitted to the Emergency Intensive Care Unit (EICU). Methods Clinical data were retrospectively collected from AP patients admitted to the EICU. Four nutritional scores (including the NRS-2002, PNI, TCBI, and CONUT scores) were calculated within 24 h of admission. Patients were divided into quartiles based on total hospital stay. Linear regression analyses were used to evaluate the associations, and ROC analyses were performed to assess predictive performance. Subgroup analyses were also conducted. Results A total of 147 patients were included and divided into quartiles (Q1-Q4) based on hospital stay duration. The NRS-2002 score, the PNI, and the CONUT score showed significant differences across the quartiles. After multivariable adjustment, only NRS-2002 remained independently associated with prolonged hospitalization (MD = 2.98, 95% CI: 1.69–4.27, p < 0.001). The ROC analysis demonstrated good predictive ability for NRS-2002 (AUC = 0.813). Subgroup analyses confirmed consistent performance across most patient strata. Conclusions The NRS-2002 is the most robust predictor of hospital stay duration among the four nutritional scores evaluated in EICU patients with acute pancreatitis. Routine NRS-2002 screening at admission may help identify high-risk patients for early nutritional intervention.