Combined mechanical circulatory support (Impella + ECMO) in cardiogenic shock caused by fulminant myocarditis

Acute myocarditis affects around 4–14 people per 100 000 each year globally. Approximately 2% to 9% of patients have haemodynamic instability and require inotropic agents or mechanical circulatory support (MCS) devices to facilitate functional recovery. These patients have an approximately 28% rate of mortality or heart transplant at 60 days.1 This case study underscores the key decision points, considerations, and advantages of combining both Impella and veno-arterial extracorporeal membrane oxygenation (VA ECMO) in the treatment of patients suffering from advanced stages of cardiogenic shock (CS) due to fulminant myocarditis. A 22-year-old male was admitted to the intensive care unit due to sustained ventricular tachycardia (VT). He experienced fatigue for the past 3 days, and on the night of admission, he reported palpitations and dyspnoea. The patient had a fever on the first day of hospitalization, which he suffered from for 3 days, with a high of 39.1°C. He also vomited for the first two nights. The initial ECG obtained from the ambulance revealed sustained VT (Figure 1A). Upon admission, the subsequent ECG displayed sinus rhythm, into which he spontaneously converted during transport; pathological ST elevations were seen in leads I, aVL, and V1–V5, and deep Q waves were noted in leads V1-V3 (Figure 1B). Despite these manifestations, vital signs upon admission were a blood pressure of 93/53 mmHg, a pulse rate of 101 beats per minute, and a lactate level of 1.9 mmol/L. He was free of symptoms during the initial assessment. The patient was diagnosed with Addison's disease, autoimmune hypothyroidism, and heart failure with a mildly reduced left ventricle ejection fraction (LVEF) of 45% in 2020. The treatment included prednisone, levothyroxine, ramipril, and bisoprolol. Cardiac magnetic resonance (CMR) 3 months after the treatment initiation showed normal left ventricular (LV) function without the presence of late gadolinium enhancement (LGE), and the temporary LV dysfunction was attributed to an endocrine disorder. Ramipril and bisoprolol were discontinued after the improvement of LVEF, and the patient continued to use levothyroxine and prednisone. Since then, the heart function has been assessed as normal with the use of echocardiography during regular check-ups with his cardiologist. The echocardiography performed on admission revealed significant reduction of LVEF (28%) due to severe global hypokinesis (Video S1). The right ventricle (RV) was of normal size with decreased systolic function [tricuspid annular plane systolic excursion (TAPSE) 14 mm and 9 cm s′ tissue Doppler imaging (TDI) of tricuspid annulus]. The initial value of highly sensitive troponin I was 18 752 ng/L with a peak value of 27 157 ng/L on the next day. The NT-proBNP level was 10 125 ng/L, and the peak lactate was 3.9 mmol/L. Selective coronary angiography was normal (Video S2A and Video S2B). On the second day, the patient developed signs of low cardiac output syndrome. Urgent CMR was performed, which revealed several non-contiguous midmyocardial LGE within the interventricular septum and LV free walls suggesting non-ischaemic damage of the LV myocardium (Video S3, Figure 2A). All laboratory tests on possible viral agents (including hepatitis and HIV) came back negative. For the suspected Addisonian crisis, we administered corticosteroids from the first day, starting with a bolus of 100 mg of hydrocortisone followed by 200 mg administered daily in four doses.2 On the second morning, the patient received inotropic support due to low cardiac output syndrome (ScV02 36.9%, VTI [velocity time integral] LVOT 8 cm) using levosimendan. In the evening, the patient experienced recurrent episodes of sustained VT with four defibrillations; intravenous amiodarone was administered to prevent VT recurrence. After this episode, his echocardiography showed progressively worsening LV dysfunction with an LVEF of 10%, and the patient experienced severe pulmonary oedema, which required treatment with non-invasive ventilation support (FiO2 100%). A rapid shock-team call was made, and the decision was to initiate Impella CP Smart support, inserted through the right femoral artery with an initial flow of 3.2 L/min (P8), which led to rapid regression of pulmonary oedema, lactate normalization, and restoration of urine output. Since the Impella implantation, the patient has been free from any arrhythmias. However, on the morning of the next day, the persistence of severely reduced LV systolic function, together with worsened RV function (TAPSE 12 mm) and high haemolysis (serum free haemoglobin 3058.5 mg/L) on Impella support, led to the decision to start awake veno-arterial extracorporeal membrane oxygenation (Maquet Cardiohelp, 19F arterial cannula, 23F vein cannula) with a flow of 4.2 L/min. The Impella flow was then lowered to 1.2 L (P2) as an active LV unloading strategy. This approach resolved the problem of haemolysis (the serum free haemoglobin was 79.8 mg/L the next day). Immediately after the ECMO implantation, an EMB was performed with preliminary results suggestive of acute necrotizing lymphocytic myocarditis (Figure 3). Based on the EMB results, treatment with high pulses of corticosteroids was started on the same day (a pulse of 1 g of methylprednisolone on the first day, followed by a dose of 100 mg of methylprednisolone on the second day, then reducing the dose to 80 mg a day for the next 7 days) along with high-dose intravenous immunoglobulins (1 mg/kg in total, divided into 3 doses given over 3 days). After starting with higher doses of corticosteroids for 7 days, the patient was switched to hydrocortisone at 200 mg per day due to ongoing stressful circumstances requiring substitution. After another 7 days, the dosage was reduced to prednisone at 20 mg per day. Fludrocortisone at a dosage of 0.1 mg per day was added later, and the dosage was gradually reduced over the following months with the endocrinologist's guidance. The patient was on ECMO and Impella for 5 days with minimal LV pulsatility (Figure 4). On day 6, he began to show the primary signs of significant RV recovery, and over the next few days, we observed further gradual improvement in both LV and RV function. This enabled us to slowly wean off ECMO, with successful fully percutaneous explantation of both the Impella and ECMO on the 8th and 9th days post-implantation, respectively (Figure 5).Details of the laboratory results are presented in Figure 6. This case study has several important implications. First, it describes the key steps of the diagnostic and decision-making process in fulminant myocarditis. Second, it underscores the advantages of combining both Impella and ECMO, often referred to as ECPELLA or ECMELLA, in the treatment of patients suffering from advanced stages of CS due to fulminant myocarditis. Given the limited availability of data concerning the management of fulminant myocarditis with CS, we posit that this case report has the potential to illuminate the importance of early initiation of left ventricular (LV) unloading and combined support in cases with deteriorating CS. Both the European Society of Cardiology (ESC)3 and the American Heart Association (AHA)4 currently advise the use of MCS in cases of acute myocarditis complicated by refractory heart failure or CS.5 Fulminant myocarditis often proves to be reversible, making the temporary utilization of short-term MCS devices an appealing therapeutic strategy. An analysis of myocarditis management trends in the United States from 2005 to 2014 revealed an increasing rate of temporary MCS utilization, growing from 4.5% to 8.6%.6 V-A ECMO remains the most widely employed MCS in fulminant myocarditis complicated by refractory CS.7 V-A ECMO provides robust circulatory support at the expense of elevated LV afterload and considerable risks of bleeding, vascular, and ischaemic complications. That's why we opted for the Impella CP Smart device implantation due to its advantages over ECMO and other MCS. Notably, Impella actively unloads the LV, favouring myocardial recovery and improvement in pulmonary oedema compared with VA-ECMO. Direct LV unloading, reduced mechanical workload, lowered myocardial oxygen demand, decreased wall stress, and improved subendocardial coronary blood flow may explain the immediate effect of device insertion on the disappearance of VT in our patient. However, all MCS devices have inherent limitations and potential complications. Impella CP provides limited LV support, and high pump flows may lead to significant haemolysis. Given the high haemolysis, worsened RV function, and expected recovery timeline spanning days to weeks, we selected peripheral percutaneous V-A ECMO. This choice enabled us to decrease Impella flow and ensure complete biventricular circulatory and respiratory support until the appearance of cardiac recovery. There are also other possible MCS combinations; one of the most deployed is the use of an intra-aortic balloon pump (IABP) with VA ECMO. However, the IABP provides only passive and limited LV support, which may not be sufficient in severe CS cases. Additionally, a recently published retrospective registry study from Japan revealed that a substantial proportion of patients with myocarditis complicated by CS can be managed by Impella alone without VA ECMO, and the survival rate for the Impella standalone group in this study was high (83.2%).8 Despite the lack of prospective and randomized data, there are experimental and clinical studies showing better efficacy of ECPELLA compared with the IABP combination with VA ECMO, although at the expense of higher complication rates with Impella compared with IABP.9 The survival advantage of ECPELLA over those solely treated with VA ECMO has been highlighted in various observational studies.10 However, contrasting these encouraging findings, a case series from a high-volume centre in Hannover, Germany, involving seven patients with influenza-associated myocarditis supported by ECPELLA, indicated a zero survival rate.11 Considering the variability in myocarditis presentation and severity, gathering robust evidence remains challenging. To achieve optimal outcomes, we recommend a comprehensive case assessment, shock-team deliberation, and decision-making involving strategies for haemodynamic deterioration. Tailoring the indication and timing of MCS devices to individual patients is essential, dependent on numerous factors outlined in Figure 7. Emphasis should be placed on LV unloading, rapid diagnosis, and treatment. Furthermore, it is important to note that the combination of ECMO and Impella is associated with elevated complication rates, primarily bleeding and vascular complications.12 To mitigate these issues, comprehensive strategies, including ultrasound and X-ray guided procedures and MCS insertions, fully percutaneous closure techniques, and intensive monitoring of bleeding, coagulation, and haemolysis, should be employed. Additionally, the patient was kept awake throughout the course of hospitalization, and the respiratory failure in this case was effectively managed by non-invasive ventilation and early Impella CP. The awake MCS strategy is used for most CS patients in our hospital. In a retrospective observational study from Paris, the awake ECMO group had significantly lower rates of pneumonia, tracheostomy, renal replacement therapy, less antibiotic and sedative consumption, and even reduced short-term and long-term mortality compared with ventilated patients.13 This study confirms previous reports suggesting that an awake approach to patients treated with MCS is feasible and effective for a significant proportion of patients with CS. The patient was discharged with an LVEF of 58%. After an 8-week follow-up, the patient was free of symptoms, and his LV function was 57% (Video S4). The CMR performed 8 weeks after discharge from the hospital showed no signs of LGE (Figure 2B). At the 6-month follow-up, the patient was still free of symptoms with normal LVEF. In conclusion, the combined use of Impella and ECMO holds potential for reversing the lethal course of refractory CS due to fulminant myocarditis. Success hinges on appropriate patient selection, timing of implantation, active LV unloading, and mitigation of potential complications associated with MCS. In this case, early Impella and VA ECMO implantation proved pivotal in reversing cardiogenic shock and facilitating successful bridge-to-cardiac recovery. Decisions regarding device selection and timing should be tailored to individual patients and coordinated within an experienced shock team. This study was supported by MH CZ-DRO-VFN64165 VFN: General University Hospital in Prague and the Charles University Research Program Cooperation – Intensive Care Medicine. Daniel Rob received consulting honoraria from Abiomed. Jan Belohlavek received consulting honoraria from Abiomed, Getinge, Resuscitec and Xenios. Michaela Zemkova, Milan Dusik, Jan Pudil, Tomas Palecek and Ivana Vitkova declare that they have no conflict of interest. Video S1. The initial echocardiography – apical four chamber projection. The initial echocardiography revealed significant reduction of ejection fraction (28%) of left ventricle with severe global hypokinesis with slightly better kinesis of the basal parts of left ventricle. Video S2A. Selective coronary angiography. Selective coronary angiography showed coronary arteries without any stenosis. A) right coronary artery. Video S2B. Left coronary artery, left anterior descending artery and left circumflex artery. Video S3. CMR. This video from cardiac magnetic resonance shows reduced ejection fraction of the left ventricle (around 30%), akinesis of the apex and apical halves of anterior and lateral wall of LV and anterior part of interventricular septum. Video S4. Transthoracic echocardiography during follow-up after eight weeks – apical four chamber projection. The echocardiography shows no kinetics disorders of the LV, and its ejection fraction is 57% according to Simpson. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.