Development and validation of clinical prediction models for acute kidney injury recovery at hospital discharge in critically ill adults
Authors' conclusion
Models to predict AKI recovery upon hospital discharge in critically ill patients with AKI-3 showed poor performance in the general ICU population, similar to the biomarker NGAL. In cardiac surgery patients, discrimination was acceptable, and better than NGAL.
Study context
- Population
- Critically ill adults with ICU-acquired acute kidney injury (AKI) stage 3 (AKI-3), multicentre EPaNIC databasedevelopment 229, validation 244
- Comparison
- Prediction models vs neutrophil gelatinase-associated lipocalin (NGAL)_AKI3 vs Reference model
- Outcome
- AKI recovery (complete; complete or partial) — model discrimination (AUROC)
- Follow-up
- Hospital discharge
Methodological detailsLeast absolute shrinkage and selection operator (Lasso) machine-learning algorithm, matched validation cohort, multicentre database
- Lasso machine-learning algorithm
- Matched validation cohort
- Multicentre database
Performance
As reported in the abstract
Discrimination for AKI recovery, validation cohort
| Test or model | AUROC, complete recovery | AUROC, complete or partial recovery |
|---|---|---|
| Prediction models | 0.53 | 0.61 |
| NGAL_AKI3 | 0.55 | 0.53 |
AUROC, complete recovery
Prediction models0.53
NGAL_AKI30.55
AUROC, complete or partial recovery
Prediction models0.61
NGAL_AKI30.53
Further findings
| Outcome | Prediction models | NGAL_AKI3 |
|---|---|---|
| Discrimination in cardiac patients (subgroup) · hospital discharge | ||
| AUROC, complete recovery | 0.60 | 0.52 |
| AUROC, complete or partial recovery | 0.71 | 0.54 |
Recovery rates, validation cohort · hospital discharge
- Complete recovery
- 33.20%
- Complete or partial recovery
- 51.23%
Discrimination in cardiac patients (subgroup) · hospital discharge
- Developed models better than the reference model (age only) for cardiac surgery patients
Original publication
Development and validation of clinical prediction models for acute kidney injury recovery at hospital discharge in critically ill adults. J Clin Monit Comput. Published 9 May 2022. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →