Predicting ICU Mortality in Patients with Acute-on-Chronic Liver Failure
Author Information
Author(s): Pollmanns Maike R., Kister Bastian, Abu Jhaisha Samira, Adams Jule K., Kabak Elena, Brozat Jonathan F., Schneider Carolin V., Hohlstein Philipp, Bruns Tony, Küpfer Lars, Trautwein Christian, Koch Alexander, Wirtz Theresa H.
Primary Institution: RWTH Aachen University Hospital
Hypothesis
Can a machine learning model accurately predict ICU mortality in patients with acute-on-chronic liver failure?
Conclusion
The Aachen ACLF ICU score demonstrates high predictive accuracy for ICU mortality in critically ill patients with acute-on-chronic liver failure.
Supporting Evidence
- ICU mortality was 60% among the studied patients.
- The ACICU score achieved an AUROC of 0.96, indicating excellent predictive accuracy.
- Patients with an ACICU score above 0.94 had a 100% mortality rate in the ICU.
- Significant differences in laboratory parameters were observed between ICU survivors and non-survivors.
- The study included a diverse cohort of patients with varying degrees of liver failure.
Takeaway
Doctors can use a new score to better predict which liver failure patients might not survive in the ICU, helping them make better treatment decisions.
Methodology
Data from 206 patients was analyzed using logistic regression and machine learning to develop a mortality prediction model.
Potential Biases
Potential overfitting of the model due to the high AUROC.
Limitations
The study's retrospective and single-center design may limit the generalizability of the findings.
Participant Demographics
Median age of patients was 56 years, with 61% male and 28% having diabetes.
Statistical Information
P-Value
0.004
Confidence Interval
95% CI: 0.87; 1.0
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website