Reperfusion strategies while on extracorporeal membrane oxygenation for catastrophic pulmonary embolism: keep fishing or cut bait?

Pulmonary embolism (PE) results in an acute increase in right ­ventricular afterload that can lead to a downward spiral of progressive right ventricular (RV) failure and cardiogenic shock. These high-risk patients are characterized by hemodynamic instability, ongoing evidence of RV dysfunction, elevated troponins, and ­elevated PE severity scores.1 This high-risk group is sometimes subdivided into catastrophic PE—following cardiac arrest or progressive organ dysfunction in the face of high or escalating vasopressor requirements. In the US National Pulmonary Embolism Response Team (PERT) Registry, the unadjusted in-hospital mortality was significantly higher in patients with catastrophic compared with patients with high-risk PE without hemodynamic collapse (42.1% vs 17.2%; P < .001).2 Length of stay was also significantly longer in patients with catastrophic PE. The rationale for acute thrombolysis relates to the need to quickly reduce the afterload of the failing RV. The use of systemic thrombolysis in high-risk PE is primarily based on physiologic principles, expert opinion, 1 small randomized trial of 8 patients, and retrospective studies.1,2 However, lytic therapy alone in ­patients with high-risk or catastrophic PE may be insufficient to improve outcomes.3 Extracorporeal support is increasingly used in patients with catastrophic PE or persistent refractory shock. Early identification, triage, and treatment of high-risk patients are tenets of PERT programs. Unfortunately, even in high-risk patients, advanced therapies remain underutilized. In the US National PERT consortium registry, although high-risk patients were more likely to receive mechanical circulatory support with extracorporeal membrane oxygenation (ECMO) (6.3% vs 0.4%), only 39% of patients had advanced interventional and lytic therapies.2 Utilization of recent shock criteria put forth by the Shock Academic Consortium may better delineate these high-risk ­patients and inform clinical decision-making by incorporating better estimates of changes in clinical status and markers of ­tissue perfusion.4 For patients with catastrophic PE who are initiated on ECMO, the role of treatments to reduce clot burden vs allowing time for endogenous clot clearance in patients is unclear.5-8 The study by Levy and colleagues9 in this issue of the Journal attempts to address this question. They conducted a large international ­retrospective cohort study of patients who were supported on ECMO for ­catastrophic pulmonary embolism to determine if surgical (ST) or catheter-directed therapy (CDT) improved the outcomes of patients supported with ECMO for catastrophic PE. The mode of CDT was at the discretion of the center and likely varied over the 10-year period, with the benefit of a less invasive approach using CDT (2014 to 2024). Patients who died early (less than 12 hours on ECMO) or had CDT or surgical lysis 48 hours after initiation of ECMO were excluded. Additionally, patients who received CDT or surgical thrombectomy prior to ECMO were also excluded. The authors are to be congratulated on this logistical effort. They included a total of 492 patients across 39 centers (15 countries). This was a very sick cohort of patients, with 70% having had a cardiac arrest prior to ECMO, and 39% were placed on ECMO during cardiopulmonary resuscitation. While subject to the limitations of a retrospective study, their dataset was not missing many variables across their cohort, and they were able to ascertain the main outcomes of their entire cohort. Adding strength to the current proposal, the authors strove to mitigate the influence of confounding as well as competing outcomes through propensity score matching, inverse probability of treatment weighting, and a genetic matching procedure—to better balance covariates. Immortal time bias was addressed through not only inverse probability of treatment weighting but also excluding patients who died within 12 hours of initiating ECMO (although the 12-hour cutoff was admittedly somewhat arbitrary). This effort was crucial as their cohort was slightly imbalanced, with those who received ECMO alone being sicker having higher severity of illness scores, vasopressor requirements, liver dysfunction, and lactate levels at baseline. Indeed, the benefit of ST/CDT in 90-day mortality and decannulation was not seen after propensity matching. In the subgroup of 180 patients who did not receive systemic lysis prior to ECMO, the use of ECMO plus either ST or CDT was associated with improved survival and a shorter duration of ECMO. Bleeding rates and limb ischemia were similar across the different subgroups. No differences in adverse events were reported between ST and CDT. While there was no direct comparison between ST and CDT, the hazard ratio for ST for ECMO weaning at 90 days was significant, whereas for CDT it was not (Table S7). Whether improvements in CDT methods and newer catheter embolectomy technology were responsible for the lower risk for mortality in their later (2020-2024) cohort cannot be determined (Table S8b). Current guidelines support the role of early thrombolysis in massive PE, as it likely improves outcomes. The study affirms that systemic lysis prior to ECMO should be considered unless cannulation is imminent, and furthermore, that early lytic therapy may mitigate the need for additional interventions during ECMO. For the subgroup that did not receive medical thrombolysis prior to ECMO, this study suggests that the use of either surgical or CDT may improve survival and help reduce the duration of ECMO support. Unfortunately, longer-term outcomes (disability, return to work, cognitive function) were not captured. The definition of brain injury was very broad, defined as either intracerebral bleed or stroke, and did not consider more subtle or longer-term cognitive performance measures or patient disposition. These outcomes are ­important in critically ill patients and in those after cardiac arrest. Although the relevance of this study may only relate to centers with ECMO capabilities, it builds on existing evidence that we need to do a better job at identifying high-risk patients earlier and, where available, provide extracorporeal support for those who are developing progressive shock in the face of failures of, or contraindications to, medical therapies. Earlier interventions, including systemic lysis, should not be delayed based on concerns that they may require cannulation. Most centers are moving away from surgical thrombectomy, given advances in catheter-directed thrombectomy. The potential benefits in the different subgroups in the Levy et al. study may serve to inform future research or at the very least reinforce the need to tailor treatments in a high-risk patient with PE. The role of CDT is evolving and may have a role in selected cases in the absence of prior thrombolysis. However, further evidence will be needed to define the specific role of CDT in the context of catastrophic PE. Supplementary material is available at American Journal of ­Respiratory and Critical Care Medicine online. Please see the ICMJE disclosure forms, which have been provided as supplementary material. No artificial intelligence tools were used in writing this manuscript.