Predicting death in necrotizing soft tissue infections: a clinical score

Area of Science:

• Infectious Diseases
• Critical Care Medicine
• Clinical Prediction Modeling

Background:

• Necrotizing soft tissue infections (NSTIs) carry a high mortality rate.
• Current methods lack a standardized approach for early disease severity assessment.
• Effective early risk stratification is crucial for improving patient outcomes.

Purpose of the Study:

• To develop and validate a clinical score for predicting mortality in patients with NSTI.
• To identify key predictors of death based on data available at initial patient assessment.
• To provide a tool for stratifying NSTI patients by risk.

Main Methods:

• Retrospective analysis of 350 NSTI patients over nine years.
• Data split into prediction and validation sets using random sampling.
• Multivariable stepwise regression identified independent predictors of mortality.
• Clinical score development and validation using TRISS methodology.

Main Results:

• Six independent predictors of death identified: age, heart rate, temperature, WBC count, creatinine, and hematocrit.
• The developed model demonstrated 86.8% accuracy and an AUC of 0.81.
• The score showed excellent performance on the validation dataset (z-statistic 0.23 to -0.83).

Conclusions:

• A novel clinical score effectively stratifies NSTI patients by mortality risk.
• The score utilizes simple, readily available admission parameters.
• This tool can guide aggressive therapeutic strategies and patient selection for clinical trials.