Author + information
- Nikolay Vdovin, MD,
- Sabina P.W. Günther, MD,
- Suzanne de Waha, MD,
- Peter Seizer, MD,
- Stefan Brunner, MD,
- Christian Schlensak, MD,
- Holger Thiele, MD,
- Christian Hagl, MD,
- Steffen Massberg, MD and
- Axel Bauer, MD∗ ()
- ↵∗Munich University Clinic, German Center for Cardiovascular Research (DZHK), Medizinische Klinik und Poliklinik I, Ziemssenstrasse 1, Munich 80336, Germany
Despite advances of medical and interventional treatment mortality in patients with acute myocardial infarction (MI) complicated by cardiogenic shock (CS) remains high. Mechanical hemodynamic support in addition to immediate revascularization and optimal medical therapy may improve prognosis in these patients, although not supported by randomized trials (1). Extracorporeal life support by means of venoarterial extracorporeal membrane oxygenation (vaECMO) provides the maximum hemodynamic support but also demands the highest logistic requirements (2). For these reasons, vaECMO is often restricted to patients in advanced stages of CS as a last resort. It is presently unknown which patients benefit from this most invasive approach. Early risk stratification based on markers available at time of decision making might help to identify patients who could qualify for this treatment strategy.
We report on 104 consecutive patients with acute MI complicated by CS who were treated between January 2013 and February 2017 with primary percutaneous coronary intervention (PCI) and vaECMO at 4 university centers (Online Table 1). In total, 85 (81.7%) underwent cardiopulmonary resuscitation (CPR) before vaECMO implantation. In 45 patients (43.3%) vaECMO was implanted during resuscitation. All patients underwent immediate primary PCI. Final Thrombolysis In Myocardial Infarction (TIMI) flow grade 3 was achieved in 83 (79.8%) patients (p < 0.001 for difference from baseline). In 29 patients (27.9%) vaECMO was implanted before PCI; in 75 patients (72.1%) vaECMO was implanted during PCI or directly thereafter. Primary endpoint was 30-day mortality.
We constructed 2 risk prediction models by Cox regression analysis. Model 1 considered only factors that were available at time of decision making: patient’s age, sex, body mass index, cardiovascular risk factors, known cardiovascular diseases (coronary artery disease, previous MI, history of stroke, peripheral artery disease), CPR, CPR duration, ongoing CPR, arterial pH, and arterial lactate (mM). Model 2 additionally considered success of reperfusion (final TIMI flow grade 3).
Of the factors considered, only CPR duration (hazard ratio: 1.010 [95% confidence intervals (CI): 1.005 to 1.016] per minute increase; p < 0.001), arterial lactate (1.080 [95% CI: 1.030 to 1.132] per mM increase; p = 0.001), and presence of diabetes mellitus (1.882 [95% CI: 1.078 to 3.285]; p = 0.026) were significant predictors of Model 1, yielding a C statistic of 79.8% (p < 0.001).
Using the beta-coefficients derived from the Cox model the predicted probability P of dying within 30 days was calculated by: where Lac is arterial lactate in mM and CPRd is the CPR duration in minutes (Figure 1A). Achievement of final TIMI flow grade 3 was associated with favorable outcome (Figure 1B) and an independent predictor of Model 2, providing a hazard ratio of 0.532 (95% CI: 0.298 to 0.951; p = 0.033).
The limitations of our study need to be recognized. Patient selection and treatment represents clinical practice at the study centers. Although we report on 1 of the largest cohorts of patients treated with vaECMO in the setting of acute MI, the sample size of our study is limited. Furthermore, our findings should be independently validated. Finally, the findings of our study are purely observational and therefore do not allow conclusions about vaECMO efficacy.
In summary, our study indicates the adverse prognosis of patients in advanced stages of CS cannot be reversed even by maximum treatment. It should be considered, however, that compared with previous studies (3) our patients were in an advanced stage of CS.
In lack of randomized trials vaECMO implantation remains an individual decision. The findings of our study can help identify those patients who are unlikely to benefit. The proposed risk model can be easily calculated and is based on 3 markers available at time of decision making. Although arterial lactate and CPR duration are physiologically linked, they reflect different aspects. Whereas arterial lactate directly measures cellular hypoxia, CPR duration additionally reflects refractoriness to treatment. In our cohort, no patient survived with a CPR duration ≥80 min or an arterial lactate level of ≥23 mM. Achievement of final TIMI flow grade 3 was independently associated with favorable outcome, underlining the importance of immediate reperfusion as a main principle of MI treatment.
Please note: All authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- 2017 American College of Cardiology Foundation