Prognostic models for mortality risk in patients requiring extracorporeal membrane oxygenation (ECMO)
Authors' conclusion
A large number of mortality prediction models have been developed for ECMO patients, yet only a minority has been externally validated. Furthermore, we observed only moderate predictive performance, large heterogeneity between-study populations and model performance, and poor methodological quality overall.
Study context
- Population
- Adults requiring or receiving veno-arterial (V-A) or V–V ECMO for refractory cardiocirculatory or respiratory failure96 studies
- Comparison
- ECMO-specific mortality prediction models vs general ICU models
- Outcome
- Performance of mortality prediction models (O:E ratio and c-statistic) for all-cause mortality
Methodological detailsrandom effects models, meta-regression, PROBAST risk of bias assessment, All predicted outcomes conditional on ECMO already initiated, reducing applicability for patient selection
- Random effects models
- Meta-regression
- PROBAST risk of bias assessment
- All predicted outcomes conditional on ECMO already initiated, reducing applicability for patient selection
Pooled result
As reported in the abstract
Discrimination (pooled c-statistic)
SAVE and RESP score
between 0.66 and 0.70
pooled c-statistic
SAPS II, APACHE II, SOFA
between 0.66 and 0.69
pooled c-statistic
Further pooled results
Models and external validations
- Models described
- 58
- External validations
- 225
- ECMO-specific models ever externally validated
- 14 (24%)
Methodological quality (PROBAST)
- Models meeting PROBAST criteria of good quality
- 1 of the 58
Heterogeneity of external validations
- Wide variability in reported performance, large between-study heterogeneity
Original publication
Prognostic models for mortality risk in patients requiring ECMO. Intensive Care Med. Published 4 January 2023. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →