Controlling nutritional status score as a novel prognostic tool for hospital mortality risk stratification in sepsis: a large-scale retrospective cohort analysis

Malnutrition is associated with poor outcomes in critically ill patients. However, the prognostic value of the Controlling Nutritional Status (CONUT) score in sepsis remains unclear. This study aimed to investigate the association between CONUT score and hospital death in sepsis patients. In this retrospective cohort study, 2,143 sepsis patients from the eICU Collaborative Research Database were categorized into Mild (2–4 points), Moderate (5–8 points), and Severe (9–12 points) malnutrition groups based on the CONUT score. Multivariable logistic regression analysis was used to evaluate the relationship between CONUT score groups and hospital death. Subgroup analyses were performed to assess the stability of the association across different populations. Restricted cubic spline (RCS) analysis was conducted to examine the potential non-linear relationship between CONUT score and hospital death. Hospital death rates increased progressively across malnutrition groups: 8.13% in Mild, 16.14% in Moderate, and 23.27% in Severe groups ( P < 0.001). After fully adjusting for confounding factors, compared with the Mild group, the risk of hospital mortality was significantly higher in the Moderate group (adjusted OR: 1.60, 95% CI: 1.07–2.38, P = 0.021) and Severe group (adjusted OR: 2.66, 95% CI: 1.73–4.07, P < 0.001). Subgroup analyses demonstrated consistent associations across different demographic and clinical characteristics (all P for interaction > 0.05). RCS analysis revealed a significant linear dose–response relationship between CONUT score and risk of hospital mortality (P for overall < 0.001, P for nonlinearity = 0.685). Each one-point increase in CONUT score was associated with a 15% higher adjusted odds of hospital mortality (adjusted OR 1.15, 95% CI 1.09–1.21). The fully adjusted model showed good discrimination (AUC 0.773, 95% CI 0.747–0.799) and excellent calibration (Hosmer-Lemeshow P = 0.390). Higher CONUT scores were independently associated with increased risk of hospital mortality in sepsis patients. The CONUT score might serve as a valuable prognostic indicator for risk stratification in sepsis.