The progressive increase in experience with short-term circulatory support devices, such as VA-ECMO, has provided a survival opportunity for patients with heart failure in its various presentations. Although less common due to increased experience among surgical teams and new perfusion and cardioplegia techniques, postcardiotomy shock remains one of the etiologies associated with the poorest outcomes in this field. Beyond intraoperative complications, the growing complexity of patients, as well as graft failures in the context of heart transplantation, are some reasons why certain patients continue to require this type of support.

There are still significant gaps today regarding the criteria and management of these patients. In fact, if any consensus exists, it pertains to increasingly specific indications for candidates for this type of support. However, some situations, such as postcardiotomy VA-ECMO support, are generally considered a bridge to decision in extremely severe conditions, potentially resulting in the desirable recovery, transition to other support, or even heart transplantation in rare cases. Thus, once the support is initiated, there is substantial variability, changes in plans, and management difficulties concerning the initially proposed bridge goal. Here, the skill and expertise of the team remain essential in the care of these critically ill patients. No consensus exists either on what may be considered a successful weaning from this support, nor do we have predictors to assess factors that might forecast successful weaning or protocols based on solid scientific evidence to guide this process. We continue to rely on individual protocols from reference centers and reports of various experiences.

The authors describe their experience from 2015 to 2020 with patients who received VA-ECMO support at their institution (Columbia University Irving Medical Center). These patients were managed by a multidisciplinary team of surgeons and cardiologists, including both interventional and heart failure specialists. The center has developed a highly protocolized strategy, specifically in managing postcardiotomy shock as well as other forms of cardiogenic shock, with distinct protocols detailed in this publication. Some notable aspects of their protocol include:

They describe different short-term support device indications, including intra-aortic balloon pump (IABP), ImpellaCP® or 5.5®, or VA-ECMO, depending on etiology, univentricular or biventricular involvement, and shock severity (as per the SCAI classification previously discussed on the blog), along with technical aspects related to cannulation approach and the necessity of associated respiratory support. All this information is detailed in a practical decision-making protocol worth reviewing as an algorithm.

The decision for left ventricular unloading was systematically applied when the pulmonary capillary wedge pressure exceeded 25 mmHg, there was no pulsatility in invasive arterial monitoring, and/or the aortic valve did not open on echocardiographic evaluation (a topic also discussed in previous blog posts). Distal limb perfusion was monitored using near-infrared spectroscopy (NIRS), with a distal perfusion catheter inserted only in cases of poor perfusion/tissue saturation results.

Decisions on weaning were made by a multidisciplinary “shock team,” where one of the objectives before conducting support flow reduction tests was to minimize vasopressor doses. To initiate weaning, several criteria had to be met:

• Patient phenotype compatibility (although few details were provided on this, it likely involves aspects such as clinical viability, absence of irreversible neurological damage, and no severe or uncontrolled infections).
• Recovery of organ failures, or at least non-substitutable ones like respiratory failure, requiring a PaO2/FiO2 ratio greater than 100. However, they also seem to require the absence of failure in other systems, such as renal function, which is replaceable.
• Pharmacological support, including both inotropic and vasopressor agents, was required to be at “reasonably” low levels.

To initiate weaning, once all criteria were met, a systematic three-step approach was followed, aligned with ELSO recommendations:

1. Protocolized Weaning Assessment: This involved reducing the support flow rate by increments of 0.5 L/min, with at least a 1-minute wait to observe hemodynamic responses after each change. If the patient responded well, they would proceed to the next step. If failure was demonstrated, flow would be increased to the minimum level where stability was maintained, holding it for 8-24 hours before the next assessment. The target was to reach 2 L/min, at which point attention should be given to the anticoagulation doses used.
2. Support Discontinuation Assessment: Following 8-24 hours of stability at the 2 L/min level achieved in the previous phase, support flow was reduced further in 0.5 L/min increments, waiting at least 1 minute for observation between changes until reaching 1.5 L/min, ideally reducing to 1 L/min. If the patient failed to maintain stability, support would be increased back to 2 L/min for an 8-24 hour stabilization period.
3. Decannulation: If the patient demonstrated adequate hemodynamic response, they were considered candidates for support discontinuation and decannulation, a procedure that should also be protocolized according to patient characteristics, clinical circumstances, and technical aspects of the previous cannulation (operating room vs. bedside), aiming to prevent complications during this procedure in such critically ill patients.

Of the 538 patients reported to require VA-ECMO support, 510 were deemed eligible for the study. The etiologies for which this support was required were diverse, with postcardiotomy shock (36.2%, although the number of surgeries performed at the institution in this period, which would allow calculation of this rate, is unknown), primary graft dysfunction post-transplantation (21.5%), acute myocardial infarction, and extracorporeal cardiopulmonary resuscitation (13% each) predominating. A previous blog entry provides an interesting analysis of extracorporeal CPR.

A total of 249 patients (48.8%) were successfully decannulated. However, with no clear criterion for successful weaning, of these 249 patients, 120 survived at least 30 days post-decannulation, and 129 were discharged within 30 days following circulatory support weaning. The remaining patients included 227 (44.5%) who died, of whom 42 died while on circulatory support and 185 following weaning within 30 days. Thus, a total of 476 patients (93.3%) were decannulated, highlighting the outstanding outcomes of this team. The remaining 34 patients (6.7%) received destination or long-term therapy, which consisted of heart transplantation in 3 cases and an intracorporeal ventricular assist device in 31 cases.

Based on these data and the clinical characteristics of the cohort, the authors conducted various analyses using multivariate models focused on successful weaning and short-term survival, with limited validity due to the heterogeneity of the population. The multivariate analysis identified younger age, etiologies like acute myocardial infarction, heart failure decompensation, or extracorporeal CPR, absence of renal failure, adequate albumin and bilirubin levels, and normal mean hemodynamic parameters as independent predictors of successful decannulation. Postcardiotomy shock (HR 2.6) and primary graft dysfunction (HR 7.5) were adverse predictors.

Among decannulated patients, they identified poor prognosis predictors for in-hospital mortality before discharge or within 30 days post-decannulation, which would formally question the classification as successful weaning: concomitant IABP (likely due to inherent left ventricular dysfunction, HR 1.3), acute myocardial infarction (HR 4.7), renal failure history (HR 3.4), and/or need for renal replacement therapy (HR 3.1), and cerebrovascular accident (HR 1.9).

One-year outcomes indicated survival in 33 out of 34 patients with long-term ventricular assist or heart transplant, and in 218 out of 249 (87.5%) patients who met successful weaning criteria with in-hospital survival beyond 30 days.

COMMENTARY:

The outcomes demonstrated by this team are exceptionally challenging to achieve, particularly in terms of the decannulation rate, underscoring their expertise and experience. Their multidisciplinary, protocolized approach stands out, following the available consensus evidence while adapting to their internal workflow. They serve as a model to emulate and may be paving the way toward optimizing results by following objective criteria and consensual protocols, rather than the variability that still prevails in settings like ours, across hospitals and even within individual centers. In this way, far from the approach of “this is how we do things here,” a homogeneous response to common scenarios is fostered, corrective measures are applied when results fall short, and generational knowledge transfer is promoted since the protocols belong to the institution rather than to individual professionals. No matter how expert or interdisciplinary a team may be, they cannot afford different standards of care depending on individual judgment, especially when managing such high-cost resources and critically ill patients.

Despite these benefits, the relevance of predictors for successful weaning and survival is less significant. Although extensive experiences like this are limited in the literature, it still represents a highly heterogeneous sample of VA-ECMO indications for cardiogenic shock across various etiologies. These etiologies carry with them unique patient clinical profiles and morbidities associated with hemodynamic failure and the critical context, which do not necessarily translate to the experience of other centers. Reported VA-ECMO weaning rates range from 30-60%. According to recent ELSO registry data, successful weaning was achieved in 59% of cases, but only 44% survived to hospital discharge. The authors do not improve weaning outcomes but do achieve better survival, extending across follow-up. This suggests an excellent patient selection protocol for successful weaning, which, in my view, represents one of the study’s most valuable lessons.

However, it may be a highly restrictive protocol that guarantees results but may exclude other patients who might benefit from weaning, thus prolonging support and its associated morbidity. Indeed, the absence of significance as a poor prognostic factor post-successful weaning in postcardiotomy shock likely reflects selective progression of candidates to support discontinuation. As a result, we are again dependent on institutional practices, which lead to varying outcomes. Conversely, we might critique the increasingly early use of VA-ECMO for postcardiotomy shock, a trend that persists today, sometimes even without maximum inotropic doses or progression from IABP. This may yield better outcomes in one of the worst support indications, potentially leading to overuse.

The responsible, protocolized use of mechanical circulatory support remains a necessity. For now, we find ourselves at a bridge-to-decision, awaiting further studies like this one to continue teaching effective practices that result in favorable outcomes, distancing from heroic and personalized care of these patients. We must also remember that none of these mechanical support systems are actual “therapies.” Therefore, I believe we should abandon this term and stop perceiving them as solutions, as they cure nothing but merely maintain sufficient circulatory status for recovery or long-term cardiac function replacement.

REFERENCE:

Zhang A, Kurlansky P, Ning Y, Wang A, Kaku Y, Fried J, et al. Outcomes following successful decannulation from extracorporeal life support for cardiogenic shock. J Thorac Cardiovasc Surg. 2024 Mar;167(3):1033-1046.e8. doi: 10.1016/j.jtcvs.2022.08.007.