Prognostic value of glycaemic variability for mortality in critically ill atrial fibrillation patients and mortality prediction model using machine learning
Authors' conclusion
GV is a significant predictor of ICU short-term, mid-term, and long-term all-cause mortality in patients with AF (the potential risk stratification threshold is 20.0%). ML models incorporating GV demonstrated high efficiency in predicting short-term mortality and GV was ranked anterior in importance. These findings underscore the potential of GV as a valuable biomarker in guiding clinical decisions and improving patient outcomes in this high-risk population.
Study context
- Population
- Critically ill atrial fibrillation (AF) patients in the intensive care unit (ICU), MIMIC-IV 3.0 database; primary cohort admissions 2008–2019 (age 76.5 [67.7–84.3] years; 57.8% male), external validation cohort admissions 2020–2022837 in external validation cohort
- Participants
- 8,989 patients
- Exposure
- Glycaemic variability (GV) quartiles
- Outcome
- All-cause mortality after ICU admission
- Follow-up
- 30-day, 90-day and 360-day
Methodological detailsMIMIC-IV 3.0 database, external validation cohort, machine learning prediction models
- MIMIC-IV 3.0 database
- External validation cohort
- Machine learning prediction models
Main finding
As reported in the abstract
Observational data: associations, not effects.
Further findings
Machine learning (ML) prediction of 30-day all-cause mortality: light gradient boosting machine (best model)
- Internal validation AUC
- 0.780
- Internal validation G-mean
- 0.551
- Internal validation F1-score
- 0.533
- External validation AUC
- 0.788
- External validation G-mean
- 0.578
- External validation F1-score
- 0.568
Original publication
Composed by AI from the published abstract and checked against it. How Briefs are made →