Related Experiment Video
Updated: Jul 2, 2026
06:02
A Contemporary Warming/Restraining Device for Efficient Tail Vein Injections in a Murine Fungal Sepsis Model
Published on: November 6, 2020
Dynamic Temperature Modeling Predicts Mortality in Murine Sepsis
Zengli Xiao1,2, Luhao Wang3, Sandra L Carpenter1
1Department of Surgery and Emory Critical Care Center, Emory University School of Medicine, Atlanta, Georgia, USA.
Shock (Augusta, Ga.)
|July 1, 2026
Summary
Serial body temperature monitoring in mice with sepsis can predict mortality. Dynamic Bayesian modeling of temperature trajectories offers a more accurate method than single measurements for predicting survival outcomes in pre-clinical sepsis studies.
Area of Science:
- Veterinary Medicine
- Pre-clinical Research
- Sepsis Modeling
Background:
- Single body temperature measurements in mouse sepsis lack predictive accuracy for mortality.
- Temperature trajectories, however, show promise in predicting outcomes in human sepsis patients.
- Accurate prediction of mortality is crucial for establishing humane endpoints in pre-clinical sepsis studies.
Purpose of the Study:
- To determine if temperature trajectories can predict subsequent mortality in septic mice.
- To develop a quantitative model for predicting mortality risk using longitudinal temperature data.
Main Methods:
- A large, heterogeneous cohort of 511 C57Bl/6 mice across three laboratories was used.
- Serial body temperature was measured every 12 hours for up to 7 days post-sepsis onset.
- A Bayesian joint model incorporated longitudinal temperature data (baseline, 12, 24 hours) to predict mortality risk (12-72 hours post-sepsis).
Main Results:
- The Bayesian model demonstrated strong predictive accuracy, with AUCs increasing from 0.864 to 0.932 for longer prediction horizons.
- Model performance was not significantly improved by incorporating sepsis model type, age, or sex.
- Specific temperature thresholds (e.g., 27.0°C at 24 hours) identified mice with 100% certainty of subsequent death, establishing irreversible terminal states.
Conclusions:
- Temperature trajectories, reflecting longitudinal hypothermia, are powerful predictors of mortality in heterogeneous murine sepsis models.
- Dynamic Bayesian modeling enables individualized mortality risk estimation, improving upon single temperature measurements.
- This approach can aid in establishing more precise and humane endpoints for pre-clinical sepsis research.
