Purpose: Mortality after veno-arterial extracorporeal membrane oxygenation (VA-ECMO) implantation remains a major problem in patients with cardiogenic shock. Our objective was to assess the utility of the SOFA score in combination with markers of right ventricular (RV) dysfunction in predicting mortality in the ICU. Materials and methods: Data were retrospectively obtained from all adult patients (n=103) who were treated with VA-ECMO between November 2004 and January 2016. The primary outcome of this study was ICU mortality after VA-ECMO implantation. Using the clinical, demographic and echocardiographic data, we developed a novel mortality risk score, the SOFA-RV score, which combine RV-function to the SOFA score at the time of VA-ECMO implantation. Results: Out of 103 patients, 37 (36%) died in the ICU. The median duration of VA-ECMO support was 7 days [IQR 4-11], mean age 49 ± 16 years, and 54% were male. SOFA-RV score has an AUC of 0.70, and was significantly better than SOFA alone (AUC of 0.57) in predicting ICU mortality. In addition, SAVE and MELD scores were not able to predict ICU mortality. Conclusion: Adding RV-function to the existing SOFA score improves significantly the prediction of ICU mortality in patients on VA-ECMO. Dedicated evaluation of RV function in patients with VA-ECMO is therefore recommended.

Additional Metadata
Keywords Cardiogenic shock, Heart failure, ICU mortality, Right ventricular function, SOFA score, VA-ECMO
Persistent URL dx.doi.org/10.1016/j.jcrc.2019.09.017, hdl.handle.net/1765/120813
Journal Journal of Critical Care: improving patient care by integrating critical care systems knowledge into practice behavior
Citation
Akin, S, Caliskan, K, Soliman, O.I.I, Muslem, R, Guven, G. (Goksel), van Thiel, R.J, … dos Reis Miranda, D. (2020). A novel mortality risk score predicting intensive care mortality in cardiogenic shock patients treated with veno-arterial extracorporeal membrane oxygenation. Journal of Critical Care: improving patient care by integrating critical care systems knowledge into practice behavior, 55, 35–41. doi:10.1016/j.jcrc.2019.09.017