Discussion
The dramatic increase in ECMO utilization over the last decade has provided a wealth of data and experience that has enhanced physician comfort with treating these complicated patients. However, it has also brought unique challenges associated with increasingly complex physiology. Some patients supported with VA ECMO may develop differential hypoxemia resulting in poor cerebral, myocardial, and upper body oxygenation and many whose initial presentation is combined cardiopulmonary failure cannot be adequately supported with standard percutaneous configurations. While central cannulation is another possible strategy, it is usually not a preferred approach unless the patient is already undergoing cardiac surgery or has severely diseased peripheral arteries precluding placement of adequately sized cannulae for optimal support. Another strategy used in such circumstances is axillary artery cannulation, which can be a good alternative to femoral artery reinfusion. Its advantages include less differential hypoxemia, less secondary left ventricular distension and greater mobility on ECMO. However, it is usually performed utilizing a side graft technique, is not suitable for urgent or emergent situations, and can potentially lead to upper extremity hyperperfusion and compartment syndrome. In a patient supported with percutaneous VA ECMO utilizing a femoral arterial reinfusion cannula, an additional venous reinfusion cannula ensures oxygenated cardiac preload and hemodynamic support that satisfies these physiologic requirements.
As with previous studies, our cohort is small and heterogeneous although largely older and obese with frequent cardiac and pulmonary comorbidities. The median SOFA-0 score of 15 and SAVE score of -12 (95.2% predicted ICU mortality and 18% in-hospital survival respectively) highlight the extreme severity of illness prior to ECMO cannulation.(15) Half of our patients were placed directly on VA-V due to comorbid cardiopulmonary failure at presentation, and the remaining patients were initiated on VA with differential hypoxemia as the most common indication for VA to VA-V conversion. Although it was initially hypothesized that VA-V conversion may reduce neurologic complications, the 14% incidence of anoxic brain injury is consistent with other reports across both standard and VA-V configurations.11 However, the survival rate of 64% compares favorably with 2016 ELSO Report adult statistics (43-65% in respiratory failure and 42-51% in shock and cardiomyopathy) as well as previous VA-V case series (39-42.9%). Notably, during the same period, three patients received an axillary graft for reinfusion as a substitute for traditional femoral arterial cannulation, but their data is not included in our analysis.
ECMO provides respiratory and circulatory support to facilitate native organ recovery (or less commonly bridge to transplant). As it is utilized in increasingly complex patients, it is not unexpected that the incidence of comorbid cardiopulmonary failure and sequelae such as differential hypoxemia will continue to grow. This is supported by the use of VA-V in 7.7% (14 out of 181) of our ECMO patients from 2016 to 2019, compared to less than 1.5% in the largest previous case series.11 It is important to note that the cannulation strategy may not be fixed for the duration of ECMO support. Patient physiology, or clinical conditions and requirements, may change over time (e.g. the heart may recover faster than the lungs) and modifications in ECMO configurations may occasionally be necessary. The conversion from the initial ECMO strategy to a different modality should always be strongly considered if the patient’s perfusion is inadequate, gas exchange is suboptimal, or complications result from the initial cannulation strategy. However, one should always be cautious with additional cannula placement due to the ongoing anticoagulation for ECMO, higher risk of bleeding (particularly in arterial vasculature), and risk of another port for infection or thrombosis.(9) Since the small number of patients with presentations prompting consideration of VA-V cannulation precludes larger randomized trials, clinical decision making is heavily influenced by smaller case series and expert opinion. Despite lack of reduction in neurologic complications, we believe that combined cardiopulmonary failure and differential hypoxemia necessitate additional venous reinfusion to supply oxygenated cardiac preload and subsequent upper body delivery. At minimum, our data indicates maintained survival with VA-V cannulation, which is notable as these are the ECMO patients with the highest severity and complexity of illness. As such, we advocate for early consideration and proactive VA-V cannulation in these situations.