Toward catecholamine responsiveness in cardiogenic shock: insights from the CRASH score.

Purpose: To model a prognostic score in patients with cardiogenic shock that incorporates catecholamine responsiveness and hemodynamics.

Methods: Assessment of the best accuracy of the score (CRASH score: Catecholamine Refractoriness and Assistance guide based on cardiogenic Shock Hemodynamics) to predict in-hospital mortality and/or extracorporeal life support, based on a retrospective study performed in a medical-surgical intensive care unit. We included 66 patients with cardiogenic shock from various etiologies. We modeled the CRASH score to be a simple score with the best accuracy to predict in-hospital mortality and/or extracorporeal life support. The CRASH score was defined as the cardiac power index (CPI) divided by the square root of 1 + inotropic score (IS, sum of different catecholamines pondered with their variable efficiency): CRASH score = CPI/square root (IS + 1).

Results: A CRASH score <0.0375 predicted death and/or extracorporeal life support with a sensitivity of 68% and a specificity of 92%. The area under the ROC curve was 0.851, achieving an overall accuracy of 0.833. The CRASH score had improved accuracy when compared with Simplified Acute Physiology, Sleeper, and CardSHOCK scores (p<0.05 for all). With a threshold of 0.300 we found even higher specificity (95%) at the price of decreased sensibility (52%) and accuracy.

Conclusions: In patients with cardiogenic shock, the CRASH score (= CPI/square root (IS + 1)) <0.0375 allowed patients at high risk of adverse outcome to be identified.