Complication rates, economic consequences, and resource limitations associated with the
use of venovenous ECMO (VV-ECMO) are widely recognized. Decannulation at the earliest and
safest possible time would be expected to improve clinical outcomes, reduce cost, and
optimize resource allocation. Yet, there are no data comparing weaning strategies for
decannulation from VV-ECMO, and there is significant variation between centers in
approaches to weaning VV-ECMO.
Current approaches to weaning VV-ECMO generally rely on clinicians to identify signs of
lung recovery and initiate incremental reductions in blood flow rate, fraction of
delivered oxygen (FdO2), and sweep gas flow rate4-6. This approach has been previously
outlined in guidelines distributed by the Extracorporeal Life Support Organization,
expert opinion, and in small descriptive studies, though little data exist to support
this strategy. Further, these approaches run counter to the large body of literature for
assessing readiness for "liberation" from sedation and mechanical ventilation in which
incremental reductions (weaning) have repeatedly been shown to be inferior to
protocolized daily assessments (spontaneous awakening trials and spontaneous breathing
trials7-11).
Prior data suggest that clinicians underestimate readiness for liberation from organ
support and suggest that protocols to identify readiness for liberation are superior to
clinician judgement9,11. Compared to incremental weaning, spontaneous awakening trials
and spontaneous breathing trials have been shown to dramatically shorten the duration of
support, reduce intensive care costs, and improve outcomes7-13. Until recently, this
approach to liberating patients from a therapy had not been applied to ECMO. Our groups
recently conducted a 26-patient, prospective, single-arm, safety and feasibility study to
develop and refine a protocol for daily assessment of readiness to liberate from VV-ECMO
at a single center14. The results of this study, published in CHEST, suggested that a
protocolized daily assessment of readiness for liberation from VV-ECMO is feasible and
safe. Further, the median time from first passed trial to decannulation was 2 days,
suggesting that a daily protocolized assessment might identify candidates for
decannulation earlier than occurs in usual care. However, as a single-arm feasibility
study, the prior study was insufficient to determine whether dedicating resources to a
protocolized daily assessment of readiness to liberate from VV-ECMO affects patient
outcomes. Further, the manner in which ECMO is provided, weaned, and discontinued varies
significantly between centers, raising persistent concerns regarding the feasibility of
widespread adoption of protocolized daily assessment of readiness for liberation from
VV-ECMO.
Additional data from a large, multi-center randomized controlled trial are needed to
compare the effects of a protocolized daily assessment of readiness for liberation from
VV-ECMO versus usual care on duration of ECMO support, measures of unsafe liberation, and
other clinical outcomes.
Primary aim: Compare the effects of a protocolized daily assessment of readiness for
liberation from VV-ECMO (ECMO-free protocol) versus liberation strategy directed by the
clinical team (usual care) on time to successful decannulation via a large multi-site
randomized controlled trial.
Secondary aim: To compare the effect of a once daily protocolized assessment of readiness
to liberate from ECMO (ECMO-free protocol) versus a liberation strategy directed by the
clinical team (usual care) on the number of days alive and free of ECMO by day 60
(ECMO-free days).
To address these aims, we propose a multi-center, open-label, parallel-group, randomized
controlled trial comparing a protocolized daily assessment of readiness for liberation
from VV-ECMO (ECMO-free protocol) to usual care. All patients who receive VV-ECMO in a
participating unit of an adult hospital and meet all inclusion criteria and no exclusion
criteria will be eligible for participation. Eligible participants or surrogate decision
makers will be approached for consent. Following documentation of written informed
consent, patients will be enrolled and randomly assigned to receive the ECMO-free
protocol or usual care. The study will control VV-ECMO weaning strategy until the first
of decannulation or death. All other decisions regarding critical care support,
interventional therapies, and medical treatment will remain at the discretion of the
treating physician and consulting teams.