Choosing Temporary Mechanical Support for Postcardiotomy Shock
A 66-year-old man with severe aortic stenosis and an LVEF of 30% undergoes bioprosthetic aortic valve replacement and three-vessel CABG. Separation from bypass fails twice. He is now on epinephrine 0.15 mcg/kg/min, milrinone 0.5 mcg/kg/min, norepinephrine 0.1 mcg/kg/min, and an intra-aortic balloon pump. MAP is 60 mmHg, cardiac index 1.8 L/min/m², PCWP 26 mmHg, CVP 19 mmHg, SvO2 48%, and lactate 5.4 mmol/L and rising. TEE shows a globally hypokinetic LV with an EF of 20%, a moderately dilated right ventricle, TAPSE 11 mm, a well-seated bioprosthesis with no regurgitation, no residual MR, and no pericardial effusion. The rhythm is sinus at 105 beats/min. The sternum is open, the patient is systemically heparinized, and the ACT is therapeutic.
1. Select a temporary support strategy and justify it, comparing the plausible alternatives in this specific patient.
Read the worked answer
The failure is biventricular: a 20% LV and a TAPSE of 11 mm with a dilated RV mean that any device restricted to the left ventricle will be limited by right-sided output. The patient has actually received all of this in the initial period, so escalation to mechanical support is appropriate.
Peripheral VA-ECMO is the most practical first step. It can be deployed rapidly, provides biventricular support plus gas exchange, and is the standard salvage for refractory postcardiotomy shock. Its liabilities here are precisely the ones to plan for: a competent bioprosthesis means the LV has no egress route, so the afterload increase will cause LV distension, pulmonary edema, and eventually thrombus and arrhythmia unless a vent is placed. With the chest open, central cannulation with a direct left atrial or LV vent is straightforward, and femoral access is avoided.
Impella alone is inadequate. It adds roughly 3.5 L/min and unloads the LV, but the right ventricle remains the limiting chamber, and retrograde passage across a freshly implanted bioprosthesis risks leaflet injury and makes the device unstable.
Central ECMO with a vent, or ECMO plus an Impella (ECPELLA), both work. The choice should follow local expertise, and the vent decision should be made before distension occurs, not after.
2. Six hours later the patient is on peripheral VA-ECMO at 4 L/min. The echo shows an LV end-diastolic dimension increasing from 52 to 62 mm, the aortic bioprosthesis no longer opening, dense spontaneous echo contrast in the LV apex, worsening pulmonary edema, and rising airway pressures. What is happening, and what are your options?
Read the worked answer
This is LV distension from a competent aortic valve, not a device failure. Peripheral VA-ECMO raises LV afterload and, with a bioprosthesis that will not open, the LV has no egress: it fills from the Thebesian and bronchial circulation, dilates, and the consequences of distension are pulmonary edema, subendocardial ischemia, stasis and thrombus, and arrhythmias. Rising airway pressures and a non-opening valve are the confirmatory signs.
Options, by preference in this patient.
- Add an Impella CP (ECPELLA) to both vent and eject. This is the least invasive way to unload the LV and simultaneously raise systemic flow, and it is feasible because the bioprosthesis is competent; the risk is device-related hemolysis - track LDH and plasma-free hemoglobin.
- Place a surgical vent - left atrial or pulmonary artery - since the chest is open, or perform a balloon atrial septostomy if the chest is closed.
- Reduce ECMO flow only if native contractility is improving and perfusion is still adequate; do not reduce flow while the patient is not improving, because flow then falls below the threshold for adequate support.
- Increase inotropy modestly to encourage ejection. This is a supplement to unloading, not a substitute.
Whichever route you choose, monitor for differential hypoxemia - a right radial arterial line and cerebral oximetry will show whether the brain and heart are being perfused by native, poorly oxygenated blood - and re-examine the LV every few hours for recovery or thrombus.
3. On day 5 there is no meaningful recovery: the LVEF is 15% and the right ventricle remains dilated and dysfunctional, while the patient is awake and neurologically intact. How do you think about the exit strategy?
Read the worked answer
I would start by confirming that the assessment is valid, not just that the device has reached an endpoint. Perform a structured weaning trial: reduce ECMO flow and inotrope support with echo at each step, look at LVEF, aortic velocity-time integral, RV function, and mixed venous saturation, and confirm that the lactate is falling and end-organ function has recovered. Check the LV for thrombus and the aortic valve for opening before any trial.
Assuming there is genuinely no recovery, the patient has reached the decision point that should have been discussed with the family and the surgical team before support was started:
- Bridge to transplant if the patient is a candidate, and this is a young patient with a 15% LVEF - full transplant and durable LVAD workup would already be under way.
- Bridge to durable LVAD is the more realistic option in a patient of this age and size. However, an RV that is still dysfunctional after five days on full support is a serious concern. An RVAD or a biventricular device may be needed, and post-LVAD RV failure is a leading cause of death. Plan for biventricular support rather than assuming LVAD alone will be enough.
- Withdrawal of support is a legitimate option if the patient is not a candidate or if irreversible organ injury is present. This should be a decision made by the team, the patient and the family, not a decision made in isolation.
What matters most is setting that decision point early and explicitly, instead of drifting from day 5 to day 15 on ECMO.