A predictive model for the identification of the risk of sepsis in patients with Gram-positive bacteria in the intensive care unit.

BACKGROUND

Gram-positive bacterial infections are very common in the intensive care unit (ICU) and may lead to sepsis. However, there are no models to predict the risk of sepsis in persons with Gram-positive bacterial infections. Therefore, the purpose of this study was to create and validate a nomogram for predicting the risk of sepsis in patients with common gram-positive bacterial infections.

METHODS

Patients infected with three common Gram-positive bacteria who were admitted to the Multiparameter Intelligent Monitoring in Intensive Care IV (MIMIC IV) database were included in this retrospective cohort study. A Cox regression model was used to develop a nomogram for predicting 3-day, 1-week, 2-week, and 1-month sepsis probability. The performance of the nomogram was analyzed using receiver operating characteristic (ROC) curves, calibration curves, and decision curves.

RESULTS

In total, 19,961 eligible patients were enrolled from MIMIC IV datasets. All participants were allocated to training and validation cohorts at random in a 7:3 ratio. The use of more than 3 types of antibiotics, dementia, ethnicity, aspartate aminotransferase (AST), neutrophils, the use of antifungal drug, ventilation and need for vasopressors were all discovered to be highly correlated with enhanced probability of sepsis in patients with Gram-positive bacteria. A prediction nomogram was constructed using these 8 predictors. The area under the curve (AUC) for predicting 3-day, 1-week, 2-week, and 1-month sepsis risk in the training cohort was 0.857, 0.774, 0.740, and 0.728, respectively, and that in the validation cohort was 0.855, 0.781, 0.742, and 0.742, respectively. The predictive power of our model is better than the SOFA score. The model had good predictive performance in all three classes of Gram-positive bacteria. Based on the calibration and clinical decision curves, the nomogram correctly predicted sepsis in patients with Gram-positive bacteria.

CONCLUSIONS

We were able to build a nomogram to predict the probability of sepsis in patients with Gram-positive bacteria, particularly those infected with spp. and spp. This model performs effectively, and it might be used clinically to manage patients with Gram-positive bacteria.