Nonlinear correlation between prognostic nutritional indices (PNI) and patients with sepsis: a retrospective study based on the MIMIC database.

Background: The objective of this study was to explore the association between PNI and mortality among sepsis patients. Methods: Data in the present study were obtained from MIMIC-IV. PNI was calculated as follows: serum albumin concentration (g/L) + 0.005 × lymphocyte count. The primary outcome of this study was in-hospital mortality. COX proportional hazard regression analysis was conducted to examine the association between PNI and in-hospital mortality. A linear trend was evaluated by including the median PNI of each group as a continuous variable in the model. Restricted cubic spline (RCS) analysis was employed to explore the linear relationship between PNI and the risk of in-hospital mortality and to investigate the interaction between PNI and different factors. Results: A total of 2794 patients were included in this study and divided into four groups (Q1-Q4) according to PNI quartile values. In the fully adjusted model, in-hospital mortality of patients in the highest quartile group of PNI values was 49.4% ( HR = 0.506, 95% CI : 0.342-0.747, P = 0.001) lower than those in the lowest quartile group, respectively, with a statistically significant trend toward increased risk, P _trend < 0.001. RCS analysis showed that an L-shaped association between PNI and in-hospital mortality. Subgroup analyses showed a association between PNI and in-hospital mortality in different strata of patients, with a negative correlation between PNI and in-hospital mortality in all groups ( HR <1 in each group). Conclusions: There is a strong correlation between low PNI and an increased risk of death during hospitalization in patients with sepsis. An L-shaped association was observed between PNI and in-hospital mortality in patients with sepsis, with an inflection point at 33.99.