ELSO-score: A practical pre-ECMO risk model for predicting in-hospital mortality in patients receiving VA-ECMO.
Despite advancements in management strategies, patients with refractory cardiogenic shock (CS) have high mortality. Veno-arterial extracorporeal membrane oxygenation (VA-ECMO) is increasingly utilized as temporary mechanical support in these patients. Current predictive models for mortality suffer from practical limitations, including complexity and the extensive variables required.
We aimed to develop a simplified, practical predictive model, the Extracorporeal Life Support Outcome score (ELSO-Score), using readily available pre-ECMO variables to predict in-hospital mortality among VA-ECMO patients.
This retrospective study utilized data from 8495 VA-ECMO patients collected by the Extracorporeal Life Support Organization (ELSO) registry between January 2017 and December 2022. We developed a simple neural predictive model, validated on training, validation, and test cohorts.
The training cohort comprised 6029 analyzed cases, with an overall in-hospital mortality rate of 55.5%. Significant predictors of mortality were elevated lactate levels, age, bilirubinemia, acute kidney injury, and the requirement for renal replacement therapy at the time of ECMO cannulation. The predictive model demonstrated moderate discriminatory performance, achieving area under the ROC curve values of 0.70, 0.69, and 0.68, 95% CI [0.63-0.73], in the training, validation, and test cohorts, respectively.
Our study demonstrates that the ELSO-Score may be a practical and effective predictive tool facilitating informed clinical decision-making and resource allocation for patients considered for VA-ECMO therapy.
We aimed to develop a simplified, practical predictive model, the Extracorporeal Life Support Outcome score (ELSO-Score), using readily available pre-ECMO variables to predict in-hospital mortality among VA-ECMO patients.
This retrospective study utilized data from 8495 VA-ECMO patients collected by the Extracorporeal Life Support Organization (ELSO) registry between January 2017 and December 2022. We developed a simple neural predictive model, validated on training, validation, and test cohorts.
The training cohort comprised 6029 analyzed cases, with an overall in-hospital mortality rate of 55.5%. Significant predictors of mortality were elevated lactate levels, age, bilirubinemia, acute kidney injury, and the requirement for renal replacement therapy at the time of ECMO cannulation. The predictive model demonstrated moderate discriminatory performance, achieving area under the ROC curve values of 0.70, 0.69, and 0.68, 95% CI [0.63-0.73], in the training, validation, and test cohorts, respectively.
Our study demonstrates that the ELSO-Score may be a practical and effective predictive tool facilitating informed clinical decision-making and resource allocation for patients considered for VA-ECMO therapy.
Authors
Alarfaj Alarfaj, Alahmad Alahmad, Wallisch Wallisch, Flynn Flynn, Rali Rali, Tonna Tonna, Shah Shah
View on Pubmed