Machine learning derived serum creatinine trajectories in acute kidney injury in critically ill patients with sepsis
Main finding · Development of AKDResult — go to the resultOdds ratio 5.16 95% CI 2.87–9.24Authors' conclusion
These 8 classes of AKI in critically ill patients with sepsis, stratified by early creatinine trajectories, were good predictors for key outcomes in patients with AKI in critically ill patients with sepsis independent of their AKI staging.
Study context
- Population
- Critically ill patients with Sepsis-3 who developed acute kidney injury (AKI) within 48 h of ICU admission (MIMIC-IV database)8 trajectory classes
- Participants
- 4,197 patients
- Exposure
- Early creatinine trajectory-based AKI classes vs transient AKI class
- Outcome
- Development of acute kidney disease (AKD)
Methodological detailslatent class mixed modelling, multivariable regression, external validation (eICU database)
- Latent class mixed modelling
- Multivariable regression
- External validation (eICU database)
Main finding
Development of AKD
Odds ratio
5.16
95% CI 2.87–9.24
Severe AKI with mild improvement but persistence vs Transient AKI
OR 5.16 (95% CI 2.87–9.24)
As reported in the abstract
Observational data: associations, not effects.
Secondary outcomes
| Outcome | HR (95% CI) | |
|---|---|---|
| Composite of AKD or all-cause in-hospital mortality by day 7Severe AKI with mild improvement but persistence vs Transient AKI · adjusted; highest risk among classes | 4.51 (2.69–7.56) | |
| Composite of AKD or all-cause in-hospital mortality by hospital dischargeLate mild AKI with persistence and worsening vs Transient AKI · highest risk among classes | 2.04 (1.41–2.94) | |
| 0.10.31310Hazard ratio (log scale) |
External validation (eICU)
- Associations similar on external validation
Original publication
Machine learning derived serum creatinine trajectories in acute kidney injury in critically ill patients with sepsis. Crit Care. Published 10 May 2024. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →