Association analysis of sepsis progression to sepsis-induced coagulopathy: a study based on the MIMIC-IV database

Abstract

Background: Sepsis-induced coagulopathy (SIC) is a severe complication of sepsis, characterized by poor prognosis and high mortality. However, the predictive factors for the development of SIC in sepsis patients remain to be determined. The aim of this study was to develop an easy-to-use and efficient nomogram for predicting the risk of sepsis patients developing SIC in the intensive care unit (ICU), based on common indicators and complications observed at admission.

Methods: A total of 12, 455 sepsis patients from the MIMIC database were screened and randomly divided into training and validation cohorts. In the training cohort, LASSO regression was used for variable selection and regularization. The selected variables were then incorporated into a multivariable logistic regression model to construct the nomogram for predicting the risk of sepsis patients developing sepsis-induced coagulopathy (SIC). The model's predictive performance was evaluated using the area under the receiver operating characteristic curve (AUC), and its calibration was assessed through a calibration curve. Additionally, decision curve analysis (DCA) was performed to evaluate the clinical applicability of the model. External validation was conducted using data from the ICU database of Xingtai People's Hospital.

Results: Among the 12, 455 sepsis patients, 5, 145 (41. 3%) developed SIC. The occurrence of SIC was significantly associated with the SOFA score, red blood cell count, red cell distribution width (RDW), white blood cell count, platelet count, INR, and lactate levels. Additionally, hypertension was identified as a potential protective factor. A nomogram was developed to predict the risk of SIC, which showed an AUC of 0. 81 (95% CI: 0. 79-0. 83) in the training set, 0. 83 (95% CI: 0. 82-0. 84) in the validation set, and 0. 79 (95% CI: 0. 74-0. 84) in the external validation. The calibration curve of the nomogram showed good consistency between the observed and predicted probabilities of SIC.

Conclusions: The novel nomogram demonstrates excellent predictive performance for the incidence of SIC in ICU patients with sepsis and holds promise for assisting clinicians in early identification and intervention of SIC.

Clinical trial: Not applicable.

Keywords: Disseminated intravascular coagulation; Prediction; Regression analysis; Sepsis.

Fig. 1

The schematic diagram shows the data inclusion, exclusion, and grouping process for the study

Fig. 2

(a) The cross-validation result of LASSO-Logistic regression. (b) LASSO coefficient profiles of the variables. (a) The figure illustrates the cross-validation results of the LASSO-Logistic regression. The x-axis represents the logarithmic values of the regularization parameter λ, while the y-axis represents the mean error from the cross-validation. The solid line shows the trend of the mean error, and the two dashed lines indicate the upper and lower limits of the error, reflecting the stability of the model’s performance

Fig. 3

Development of a predictive risk model for SIC

Fig. 4

(a) The ROC curve of the training group; (b) The ROC curve of the validation group; (c) The external validation curve is shown in the figure

Fig. 5

(a) The calibration model of the training group; (b) The calibration model of the validation group; (c) The calibration model of the validation group

Fig. 6

(a) The decision curve analysis of the training group (b) The decision curve analysis of the validation group (c) The decision curve analysis of the external validation group