Author + information
- Thomas Fux, MD∗ (, )
- Manne Holm, MD,
- Matthias Corbascio, MD, PhD,
- Jan van der Linden, MD, PhD and
- Lars H. Lund, MD, PhD
- ↵∗Department of Medicine, Solna, Karolinska Institutet, Karolinska University Hospital, SE-171 76 Stockholm, Sweden
Cardiogenic shock complicates acute myocardial infarction, cardiac surgery, and decompensated heart failure of other etiologies, and has an in-hospital mortality rate ranging widely from 40% to 90%. Venoarterial extracorporeal membrane oxygenation (VA-ECMO) offers acute temporary life support to patients with refractory cardiogenic shock (RCS) who are unresponsive to conventional medical therapy, although the in-hospital mortality rate remains around 50%. Optimal selection and timing of VA-ECMO implantation are vital because of the high demand of resources. Few studies have focused specifically on pre–VA-ECMO factors (those that can be used for selection) for risk prediction in an undifferentiated RCS population (1,2). For this purpose, identifying independent pre–VA-ECMO predictors of mortality is essential, and consequently we refrained from combining pre–VA-ECMO and on–VA-ECMO variables. Instead, the latest available clinical and laboratory data before VA-ECMO initiation were analyzed. The primary aim of this study was to identify pre–VA-ECMO predictors of 90-day mortality in patients with RCS who were supported by VA-ECMO.
We analyzed 181 consecutive patients with RCS (median age 58 years; interquartile range [IQR]: 47 to 66 years; 75% men) who underwent VA-ECMO between September 2006 and April 2015 as salvage therapy for persistent tissue hypoperfusion despite adequate intravascular volume loading and support with inotropes and vasopressors (milrinone, levosimendan, epinephrine, norepinephrine, dobutamine, dopamine, and vasopressin). The 3 main causes were post-cardiotomy failure (58%), acute myocardial infarction (22%), and other medical causes (20%). Cardiopulmonary resuscitation was performed in 40% of patients before initiation of VA-ECMO. The median duration of VA-ECMO support was 7 days (IQR: 3 to 13 days). The 90-day mortality rate was 54%; 46% of these patients died while undergoing VA-ECMO, 45% were successfully weaned, 6% were bridged to heart transplantation, and 4% were bridged to left ventricular assist devices. The in-hospital mortality rate was 54%, and 46% of the patients were discharged home. Thirty-eight pre–VA-ECMO variables were analyzed. Age, ischemic heart disease (IHD), multiorgan failure, left ventricular ejection fraction, mean arterial pressure, arterial pH, arterial lactate, international normalized ratio, and number of inotropes and vasopressors were significant in univariable regression analysis. There was a strong negative correlation between pH and lactate; hence pH was excluded in the multivariable analysis. Multivariable logistic regression analysis identified arterial lactate (odds ratio [OR] per unit: 1.14; 95% confidence interval [CI]: 1.06 to 1.23; p <0.001), number of inotropes and vasopressors (OR per agent: 1.58; 95% CI: 1.13 to 2.21; p = 0.008) (Figure 1), and IHD (OR: 2.90; 95% CI: 1.31 to 6.39; p = 0.008) as independent predictors of 90-day mortality. Multiorgan failure was the main cause of death on (45%) and after explantation of VA-ECMO (44%). The main complications were renal failure necessitating hemodialysis (60%) and pneumonia (51%).
In RCS, arterial lactate indicates the degree of tissue hypoperfusion secondary to low cardiac output. The pharmacological approach to counteract RCS is administration of inotropes and vasopressors, which was also found to be a strong independent predictor of 90-day mortality, as a marker of a more severe clinical condition and perhaps also as the cause of worse outcomes. For clinical utility, we graded the amount of intravenous inotropic and vasopressor support in a clinically relevant “number” of ongoing agents at initiation of VA-ECMO, instead of using more complex vasoactive agent scoring systems. To our knowledge, our study is the first to use this approach. IHD was the only organ-specific predictor, which was expected because patients with IHD generally have worse predictable outcomes compared with patients who do not have IHD.
Our finding of arterial lactate as an independent pre–VA-ECMO predictor of survival is supported by other single-center studies with undifferentiated RCS populations (2,3). Additionally, we identified that specific intervals of arterial lactate levels progressively indicated worse outcome. None of the 3 etiologic categories used in this study significantly predicted 90-day mortality. The hemodynamic state during RCS, as indicated by lactate levels, the level of pharmacological support, and the presence of IHD, appeared to be more important than the specific cardiac conditions causing RCS, similar to the situation in end-stage chronic heart failure. In contrast to several other studies, age did not reach significance in multivariable analysis. However, a type II error cannot be excluded. With the screening of a large number of potential risk markers in a relatively modest sample size, the findings should be considered hypothesis generating and should be validated in a separate sample.
In conclusion, this observational study of heterogeneous patients reflects the clinical reality when providing VA-ECMO support to patients with RCS and identified arterial lactate, number of inotropes and vasopressors, and presence of IHD as independent pre–VA-ECMO predictors of 90-day mortality after VA-ECMO initiation. These predictors are easily available for pre–VA-ECMO risk prediction and may help clinicians to use VA-ECMO more effectively for adult patients with RCS.
Please note: The authors thank Anders Franco-Cereceda, Professor of Cardiothoracic Surgery, Head of Department of Molecular Medicine and Surgery, Karolinska Institutet for supporting this project. This work was supported by grants from the Stockholm County Council and Karolinska Institutet. Dr. Lund has received grants from AstraZeneca, Novartis, and Boston Scientific; and is a consultant for AstraZeneca, Novartis, Bayer, Vifor, and Relypsa. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- 2017 American College of Cardiology Foundation