The J-shaped association between the ratio of neutrophil counts to prognostic nutritional index and mortality in ICU patients with sepsis: a retrospective study based on the MIMIC database
Jiaqi Lou, Hong Kong, Ziyi Xiang, Xiaoyu Zhu, Shengyong Cui, Ji‐Liang Li, Guoying Jin, Neng Huang, Le Xin, Youfen Fan, Sida Xu
Abstract
Background: The ICU faces persistent challenges with sepsis, marked by systemic inflammation and metabolic disruptions, often leading to poor outcomes. Despite advances, reliable biomarkers for predicting sepsis outcomes are needed. This study introduces a novel indicator combining neutrophil count and prognostic nutritional index (PNI) to improve predictive accuracy by addressing both inflammatory and nutritional-immune aspects. Methods: We conducted a retrospective cohort study analyzing data from the MIMIC-IV database, focusing on adults diagnosed with sepsis per Sepsis 3.0 criteria. We excluded those younger than 18, with ICU stays under 48 hours, multiple ICU admissions, or incomplete data. Participants' neutrophil counts/PNI ratios were calculated and correlated with 30, 60, and 90-day hospital and ICU mortality, utilising Kaplan-Meier survival analysis, Cox proportional hazards models, restricted cubic spline (RCS) models and subgroup analysis. Results: We included 2,116 patients from 22,517 eligible cases. Survival analysis demonstrated lower survival probabilities for higher neutrophil counts/PNI ratios across all observed time windows. Cox regression models revealed a significant association between higher neutrophil counts/PNI ratios and increased short- to medium-term mortality. The restricted cubic spline regression models illustrated a J-shaped relationship between neutrophil counts/PNI and mortality. Conclusion: The neutrophil counts/PNI ratio is a promising prognostic biomarker for sepsis-related outcomes in ICU settings, offering improved risk stratification and potentially guiding therapeutic interventions. Further research is warranted to validate these findings across diverse populations.