The choice between urgent surgical revascularization (CABG) or percutaneous coronary intervention (PCI) in patients undergoing ECPR after acute coronary syndrome (ACS) remains controversial. 

Following the SHOCK trial and similar studies, PCI became a common strategy for patients with refractory cardiac arrest associated with ACS, aiming to reduce the time from symptom onset to revascularization—a delay inevitably associated with surgery. However, survival rates in that study were similar between both revascularization methods. Combined with the widespread use of ECMO and advancements in surgical techniques in recent years, this controversy persists, particularly in patients with complex coronary anatomies, such as triple vessel disease. The absence of robust comparative studies in this population further underscores this ongoing debate. 

This study analyzed retrospective data from two tertiary hospitals in Taiwan, including patients treated with ECPR between 2010 and 2022 for both in-hospital and out-of-hospital cardiac arrest. Of 327 patients initially selected, 215 were included (40 CABG and 175 PCI). Propensity score matching (1:1) was used to balance baseline characteristics (demographics and resuscitation parameters) among the 40 CABG patients and an equal number of PCI patients, with 95% presenting triple vessel disease. Outcomes such as in-hospital and midterm survival and successful ECMO weaning rates were evaluated. 

The analysis revealed higher success rates for ECMO weaning (71.1% vs. 48.7%; p = .05) and in-hospital survival (56.4% vs. 32.4%; p = .04) for CABG compared to PCI. However, no significant differences were observed in midterm survival among hospital survivors, though CABG demonstrated a trend toward fewer reinterventions (p = .07). 

COMMENTARY: 

While this study provides valuable insights into revascularization strategies in the complex context of ECPR, its methodological limitations should be considered. 

The primary limitation is the retrospective design, which inherently limits causal inferences for various reasons. Patient allocation to treatment was determined by the attending medical team’s decisions, suggesting that treatment modalities were likely influenced by multiple factors beyond those included in the propensity score analysis (e.g., ventricular function was not assessed). Another key limitation of the design is the selection of 40 CABG-treated patients and their comparison to a matched subset of 40 PCI-treated patients, excluding 72% of initially included PCI-treated patients. Moreover, the relatively small sample size reduces the statistical power to detect differences in both outcomes and baseline characteristics, despite matching. For instance, the CABG group had twice the coronary disease burden and four times the proportion of patients with NYHA functional class III or IV, with p = .05 and .08, respectively. 

Another significant consideration is the longer ECMO-to-revascularization time in the CABG group. Cases defined as “emergent” included all revascularizations performed within 48 hours, an arbitrary criterion that does not align with the common definition of an emergency. This discrepancy could be a critical factor requiring further investigation, as delays in revascularization are typically associated with worse outcomes. 

Despite these limitations, the study’s findings suggest that CABG may offer initial advantages in terms of in-hospital survival and reduced need for reinterventions in selected patients. Future studies should aim for prospective designs and randomized comparisons between CABG and PCI. Additionally, incorporating post-hospitalization functional status variables would provide a more comprehensive view of each strategy’s impact on patients with such critical conditions. 

REFERENCE:

Fu HY, Chen YS, Yu HY, Chi NH, Wei LY, Chen KP-H, et al. Emergent coronary revascularization with percutaneous coronary intervention and coronary artery bypass grafting in patients receiving extracorporeal cardiopulmonary resuscitation. Eur J Cardiothorac Surg. 2024; doi:10.1093/ejcts/ezae290.