Early mobilization of patients receiving extracorporeal membrane oxygenation: a retrospective cohort study

Neuromuscular weakness and the accompanying impairment in physical functioning are common sequelae of immobilization during critical illness [1‐5]. Active participation in physical and occupational therapy is increasingly recognized as not only safe and feasible, but also the preferred approach to minimize debilitation in critically ill patients, including those who require invasive mechanical ventilation [6‐8]. Patients who receive early rehabilitation have shown improved rates of returning to independent functioning, decreased rates of delirium, and shorter durations of mechanical ventilation, intensive care unit length of stay, and hospital length of stay [9‐11].

Extracorporeal membrane oxygenation (ECMO) is increasingly being used in patients with respiratory failure [12, 13]. Patients receiving venovenous or venoarterial ECMO have traditionally been considered too unstable for active physical therapy, frequently are heavily sedated, and occasionally are administered neuromuscular blocking agents. However, the ability to ambulate while receiving ECMO support has been facilitated by advances in extracorporeal technology and cannulation techniques [14‐17]. Additionally, mobilization may be facilitated when ECMO allows for weaning from invasive mechanical ventilation [18]. Although patients receiving ECMO as bridge to transplantation (BTT) are obvious targets for early rehabilitation to maintain their transplant candidacy, those requiring ECMO as a bridge to recovery (BTR) from acute respiratory failure should theoretically benefit similarly from early mobilization, though there are few published reports of success in such populations [18‐21]. We describe our center’s experience with a multidisciplinary approach to early physical therapy, including ambulation, in patients requiring ECMO as either BTR from acute respiratory failure or BTT in cases of end-stage lung disease.

There is increasing evidence demonstrating improved outcomes from early rehabilitation, including ambulation, in critically ill patients [1, 7‐10]. With the increasing use of ECMO for patients with either acute, potentially reversible, respiratory failure or end-stage lung disease awaiting transplantation, there is a growing need to evaluate the feasibility, safety, and functional outcomes of performing physical rehabilitation and early mobilization within this patient population.

Previously published case series describing attempts at active physical therapy and mobilization in patients receiving ECMO have reported small cohorts with varying degrees of success, and these were mostly pre-transplant patients [18‐21, 23‐27]. In order to maximize the success of physical therapy and minimize the risk of complications, we have developed a multidisciplinary team approach to physical therapy for patients receiving ECMO.

Older generations of extracorporeal technology were cumbersome and difficult to mobilize. Additionally, traditional ECMO configurations involved femoral cannulation, making mobilization, especially ambulation, a high-risk intervention because ECMO catheters are larger and less flexible than other catheters typically used in medical intensive care [28] and dislodgment has more severe consequences. Despite these concerns, we were successful in achieving standing or ambulation in two patients with femoral venous ECMO cannulae, demonstrating feasibility. The risks associated with mobilization while receiving ECMO have been lessened by the concurrent development of more compact circuitry and dual-lumen catheters that can provide venovenous extracorporeal support via an upper-body approach [14, 15, 29]; however, these dual-lumen cannulae require particular expertise for proper placement and are associated with complications unique to this approach [30]. Additionally, the combination of internal jugular venous and subclavian arterial cannulation permits venoarterial ECMO support via an upper-body configuration [17]. Our experience suggests that while femoral cannulation is not an absolute contraindication to ambulation, it is certainly not optimal, and an upper-body configuration is our recommended approach to facilitate ambulation and avoid cannula complications during mobilization.

With increases in extracorporeal gas exchange efficiency, select patients are eligible for endotracheal extubation while receiving ECMO. Extubation eliminates the accumulation of complications associated with invasive mechanical ventilation (ventilator-associated pneumonia, ventilator-associated lung injury, dynamic hyperinflation, impaired delivery of aerosolized medications), reduces the need for sedatives commonly employed to maximize patient-ventilator synchrony, improves patient comfort, and further facilitates ambulation by having one less apparatus to mobilize, all of which may contribute to improved outcomes [31‐35]. We were successful in liberating 66% of the ECMO patients receiving active physical therapy from invasive mechanical ventilation, with similar rates in both BTT and BTR groups. Liberation from invasive mechanical ventilation occurred in parallel with the physical therapy program.

Our ability to safely and repeatedly mobilize patients being bridged to either transplant or recovery was demonstrated despite a high severity of illness, as evidenced by very severe pre-ECMO gas exchange abnormalities and high APACHE II scores. These scores may underestimate the true severity of illness in patients transferred from outside institutions due to missing data representing the most severe physiologic derangements prior to the initiation of ECMO. Patients were able to engage in active physical therapy despite the need for low-dose vasopressors, and no significant changes in ECMO settings were needed during activity with the exception of one patient who engaged in cycling. Physical therapy sessions and ambulation were more frequent in the BTT group, which may be explained, in part, by the longer duration of ECMO (18.7 versus 9.1 days), lower average APACHE II score (20.4 versus 28.8), lower percentage of patients with hypoxemic respiratory failure (58% versus 75%) and, to a much lesser extent, less severe hypoxemia (PaO₂:FIO₂ of 62 versus 55) in the BTT group than the BTR group. As expected, BTT patients had lower severity of illness scores than BTR patients, in part because those who are too critically ill may be unsuitable for transplantation and, therefore, not considered candidates for ECMO as BTT.

Despite the advances in extracorporeal technology, not all ECMO patients may be appropriate candidates for early rehabilitation. Overall, approximately one third of our total ECMO cohort was able to participate actively in physical therapy, although our rate of participation has increased by more than 240% from 2009 to the present, with more than 40% of recent ECMO patients participating in active physical therapy. The increased rate is due, in part, to increased experience and comfort with mobilizing critically ill patients in general, which was aided by the initiation of a global early mobilization program in our MICU.

Over the course of the study period, we relaxed our criteria for participation in rehabilitation, in parallel with increasing comfort by our staff in treating patients with marked physiologic derangements. Patients on low-dose vasopressors or requiring high ventilatory support who, early in our experience, were deemed too unstable for mobilization are now more routinely engaged in physical therapy. However, there are clinical parameters that may limit physical therapy. Patients who are hemodynamically unstable requiring high-dose vasopressor support are not engaged in physical therapy. Additionally, those who are deeply sedated, receiving neuromuscular blockade, or severely hypoxemic despite extracorporeal support remain a population that is unsuitable for our mobilization protocol, and accounted for the vast majority of patients who did not participate in active physical therapy in our cohort. However, compared with the earlier study period, our approach to critically ill patients in general, and ECMO patients specifically, has evolved to incorporate less sedation and neuromuscular blockade, allowing more patients to remain awake during their MICU stay, accounting for the significant difference in patients participating in active physical therapy. Because the more critically ill patients, with a higher expected mortality, were the ones unable to participate in mobilization, comparison of outcomes between those who did and did not receive active physical therapy would not reflect just the impact of our protocol on those outcomes. Larger, prospective studies are needed to better define the barriers to physical therapy and mobilization in ECMO patients and to understand the effect of these interventions on functional outcomes. Future studies may also help establish protocols regarding adjustments to ECMO settings as necessary to compensate for increases in activity that can be standardized and applied across different ECMO patient populations. Cost-benefit analyses are also warranted given the cost associated with increased staffing needs for early mobilization protocols. Although occupational therapy statistics are not represented in this study, it would be beneficial for future studies to examine the impact of occupational therapy on the ECMO population. Occupational therapists are now participating in the early mobilization of these patients with a focus on activities of daily living, upper extremity strengthening and coordination, cognition and communication.

Active physical therapy, including early mobilization, may be performed in patients receiving ECMO. Advances in technology and a multidisciplinary team approach facilitate the safe and reliable implementation of such an intervention, even among patients with a very high severity of illness. Additional research is needed to characterize the long-term functional, neurocognitive and psychiatric impact of physical and occupational therapy on patients receiving ECMO.

Co-senor authors are Matthew Bacchetta and Daniel Brodie.