Abstract 503 Long- and short-term mortality of Dresden Impella Registry – LV unloading improved outcome compared to predicted mortality

Background: Cardiogenic shock (CS) is a state of end-organ hypoperfusion due to cardiac output failure and is characterized by high mortality. Percutaneous left ventricular assist devices (pLVAD), such as the Impella^® system, support the left ventricular function and provide a sufficient supply of oxygen to all tissues. It is still unknown whether its use improves mortality of CS. In the current study, we compared the outcome of patients receiving an Impella CP^® in CS with the predicted mortality estimated by Apache II score. We also investigated long- and short-term mortality of different subgroups.

Methods: The Dresden Impella Registry is a large, ongoing registry including more than 350 patients since 2014. Until January 2018, a total of 183 patients received an Impella CP^® in refractory CS. The Cox proportional-hazards model was used for analysis of mortality at 30 days and at 1 year. Predicted mortality was estimated by the Apache II score. Analysis was performed in the total registry population and in different subgroups.

Results: The analyzed patients of the Dresden Impella Registry (n=183) were aged 66.1 ± 0.1 years, with males predominating (69.9%). The cardiovascular risk factor showed a distribution typical for industrial countries. CS was mainly caused by acute myocardial infarction (64.3%). A cardiopulmonary resuscitation (CPR) before pLVAD was performed in 48.1%, with a mean duration of 23 ± 2 minutes. At admission, the mean serum lactate measured 6.5 ± 0.4 mM.

Mortality at 30 days and 1 year were high with 58.9% and 70.8%, respectively. However, the predicted mortality estimated by the Apache II score was still above. The score calculated an intrahospital mortality of 96.0% ± 0.3% for the total registry population. At the mean duration of hospital stay (22.7 d, survivors only), mortality was markedly lower, at 56.6%. CPR before implantation of pLVAD significantly increased mortality (30 d: no CPR 50.6% ± 5.5% vs. CPR 73.1% ± 5.1%; p<0.001) without any difference between in-hospital cardiac arrest (IHCA) or out-of-hospital cardiac arrest (OHCA) (30 d: IHCA 69.2% ± 6.5% vs. 79.8% ± 8.2%; p=0.321). This result might be influenced by a higher disease burden of the IHCA cohort with significantly increased serum lactate, norepinephrine dosage at the time of pLVAD implantation, and a more severe coronary artery disease (CAD).

Furthermore, an elevated serum lactate during pLVAD implantation was associated with increased mortality (30 d: <5mM 47.2% ± 6.1%, 5-10mM 79.3% ± 6.1%, >10mM 68.2% ± 6.9%; p<0.001). Except the cohort <50a, higher age was concomitant with higher mortality. Patients below 50a distinctly benefited from LV unloading (mean survival <50a: 207d ± 8 vs. >50a: 114d ± 1d; p=0.064). Sex category, type of CAD, or culprit lesion showed no significant effect on outcome. However, there was a trend to lower mortality of patient without CAD (30 d: no CAD 54.5% ± 15.8% vs. CAD 80.5 ± 4.2%; P=0.074). All subgroups showed an improved outcome compared to predicted mortality. All results were consistent at 30 days and 1 year. Furthermore, a Cox regression showed serum lactate, CPR before pLVAD implantation, and older age as independent risk factors for increased mortality.

Conclusion: LV unloading improved outcome of refractory CS compared to predicted mortality estimated by Apache II score in Dresden Impella Registry.


Long- and short-term mortality of Dresden Impella Registry – LV unloading improved outcome compared to predicted mortality
J. Mierke¹, T. Löhn², G. Ende¹, F. Kluge¹, S. Jellinghaus¹, U. Speiser¹, K. Ibrahim³, A. Linke¹, C. Pflücke¹
¹Technische Universität Dresden, Herzzentrum Dresden Universitätsklinik, Dresden; ²Kreiskrankenhaus Freiberg, Klinik für Innere Medizin II, Freiberg; ³Klinik für Innere Medizin I, Klinikum Chemnitz gGmbH, Chemnitz;
Background: Cardiogenic shock (CS) is a state of end-organ hypoperfusion due to cardiac output failure and is characterized by high mortality. Percutaneous left ventricular assist devices (pLVAD), such as the Impella^® system, support the left ventricular function and provide a sufficient supply of oxygen to all tissues. It is still unknown whether its use improves mortality of CS. In the current study, we compared the outcome of patients receiving an Impella CP^® in CS with the predicted mortality estimated by Apache II score. We also investigated long- and short-term mortality of different subgroups. Methods: The Dresden Impella Registry is a large, ongoing registry including more than 350 patients since 2014. Until January 2018, a total of 183 patients received an Impella CP^® in refractory CS. The Cox proportional-hazards model was used for analysis of mortality at 30 days and at 1 year. Predicted mortality was estimated by the Apache II score. Analysis was performed in the total registry population and in different subgroups. Results: The analyzed patients of the Dresden Impella Registry (n=183) were aged 66.1 ± 0.1 years, with males predominating (69.9%). The cardiovascular risk factor showed a distribution typical for industrial countries. CS was mainly caused by acute myocardial infarction (64.3%). A cardiopulmonary resuscitation (CPR) before pLVAD was performed in 48.1%, with a mean duration of 23 ± 2 minutes. At admission, the mean serum lactate measured 6.5 ± 0.4 mM. Mortality at 30 days and 1 year were high with 58.9% and 70.8%, respectively. However, the predicted mortality estimated by the Apache II score was still above. The score calculated an intrahospital mortality of 96.0% ± 0.3% for the total registry population. At the mean duration of hospital stay (22.7 d, survivors only), mortality was markedly lower, at 56.6%. CPR before implantation of pLVAD significantly increased mortality (30 d: no CPR 50.6% ± 5.5% vs. CPR 73.1% ± 5.1%; p<0.001) without any difference between in-hospital cardiac arrest (IHCA) or out-of-hospital cardiac arrest (OHCA) (30 d: IHCA 69.2% ± 6.5% vs. 79.8% ± 8.2%; p=0.321). This result might be influenced by a higher disease burden of the IHCA cohort with significantly increased serum lactate, norepinephrine dosage at the time of pLVAD implantation, and a more severe coronary artery disease (CAD). Furthermore, an elevated serum lactate during pLVAD implantation was associated with increased mortality (30 d: <5mM 47.2% ± 6.1%, 5-10mM 79.3% ± 6.1%, >10mM 68.2% ± 6.9%; p<0.001). Except the cohort <50a, higher age was concomitant with higher mortality. Patients below 50a distinctly benefited from LV unloading (mean survival <50a: 207d ± 8 vs. >50a: 114d ± 1d; p=0.064). Sex category, type of CAD, or culprit lesion showed no significant effect on outcome. However, there was a trend to lower mortality of patient without CAD (30 d: no CAD 54.5% ± 15.8% vs. CAD 80.5 ± 4.2%; P=0.074). All subgroups showed an improved outcome compared to predicted mortality. All results were consistent at 30 days and 1 year. Furthermore, a Cox regression showed serum lactate, CPR before pLVAD implantation, and older age as independent risk factors for increased mortality. Conclusion: LV unloading improved outcome of refractory CS compared to predicted mortality estimated by Apache II score in Dresden Impella Registry.
/V405.htm