A superior tool for predicting sepsis in SAH patients: The nomogram outperforms SOFA score

Area of Science:

• Critical Care Medicine
• Medical Informatics
• Biostatistics

Background:

• Sepsis is a life-threatening complication in patients with non-traumatic subarachnoid hemorrhage (SAH).
• Accurate prediction of sepsis risk is crucial for timely intervention and improved patient outcomes.
• Existing scoring systems may not fully capture the complexity of sepsis development in SAH patients.

Purpose of the Study:

• To develop and validate a predictive nomogram for sepsis risk in non-traumatic SAH patients.
• To compare the performance of the developed nomogram against the established SOFA score.
• To assess the clinical utility and discriminative ability of the nomogram.

Main Methods:

• Utilized data from the MIMIC-IV database for a cohort of 803 SAH patients.
• Employed forward stepwise logistic regression and LASSO regression to identify independent prognostic factors.
• Validated the nomogram using receiver operating characteristic curves (AUC), C-index, NRI, and IDI, comparing it with the SOFA score.

Main Results:

• Identified five key predictors: mechanical ventilation, hyperlipidemia, temperature, white blood cell count, and red blood cell count.
• The nomogram demonstrated high discriminative ability with AUCs of 0.854 (training) and 0.824 (validation), exceeding the SOFA score.
• The nomogram showed superior performance and better calibration compared to the SOFA score, with significant net reclassification improvement.

Conclusions:

• The developed nomogram is a robust tool for predicting sepsis risk in SAH patients.
• The nomogram offers superior predictive accuracy and clinical applicability compared to the SOFA score.
• This predictive model holds significant potential for guiding clinical management and improving outcomes in SAH patients.