Early identification of sepsis-induced coagulopathy in critical ill patients: an analysis from MIMIC IV database.

BACKGROUND

To develop a model for early identification of coagulopathy in septic patients.

METHODS

Patients with sepsis were identified from the Medical Information Mart for Intensive Care (MIMIC)-IV database. Patients who did not meet the sepsis-induced coagulopathy (SIC) scoring criteria upon admission but developed SIC within the subsequent 7 days were considered to be in a pre-SIC state at baseline. Baseline clinical features of the patients were screened by lasso regression. Subsequently, these features underwent multivariate logistic regression for model construction, followed by testing the stability of the model in the test set.

RESULTS

A total of 7,806 patients were included in the study from the MIMIC-IV database, comprising 7,080 without SIC and 726 with pre-SIC. Patients with pre-SIC had higher criticality scores compared to patients with Non-SIC. Pre-SIC was identified as an independent risk factor for hospitalization, 28-day, 90-day, and 1-year mortality in patients with sepsis. Patients with pre-SIC who received early heparin had lower 28-day mortality compared to those without treatment. The SIC scoring system demonstrated a sensitivity of 77.0% for identifying pre-SIC, a specificity of 53.9%, and an AUC of 0.694 (95% CI: 0.659–0.730). Based on SIC scoring system, additional clinical features were added to the pre-SIC model, ultimately yielding 70% sensitivity and 76.2% specificity with an AUC of 0.802 (95% CI: 0.773–0.830) in the validation set.

CONCLUSION

The development of SIC is associated with increased mortality rate in patients with sepsis, and precise identification of this group of patients and individualized treatment may be important for improving prognosis.