Neuromuscular blockade for ARDS was no help, in supine patients
Continuous neuromuscular blockade for severe ARDS became common practice after the ACURASYS trial (2010) showed it reduced mortality by an absolute 9%. A larger trial, ROSE, now finds no benefit of the therapy over usual care -- but leaves lingering questions due to major design differences.
The ROSE trial was a multicenter randomized trial in which 1,066 patients with ARDS received either early 48-hour infusions of cisatracurium and deep sedation, or usual care (it was not blinded; the control patients did not receive placebo). Patients in the usual care arm could get as-needed doses of cisatracurium (and 17% did), but were kept to the lightest sedation possible in accordance with current standard practice. Results were published in the New England Journal of Medicine and reported at a society conference.
The trial was stopped early for futility after it was obvious there was no benefit from the continuous neuromuscular blockade infusions. At 90 days, all-cause mortality was 42.5% in the neuromuscular blockade group vs 42.8% in the control group. There were no benefits detected in secondary endpoints such as 28-day mortality, ventilator-free days, or early discharge from the hospital.
There were also very few differences observed between groups in adverse events. There were no differences in the rates of pneumothorax. There were an excess of 10 cardiovascular events in the neuromuscular blockade group, largely conduction system failures (bradycardia, atrial fibrillation, heart block), with an excess of four cardiac arrests.
Neuromuscular Blockade for ARDS: Huge Benefit, Or None?
Why the discordant findings between this trial and ACURASYS, which was smaller (n=340) but also multicenter, randomized, and placebo-controlled, but found a 9% absolute mortality reduction? Differences in design include:
The control patients in ROSE got lighter sedation, while in ACURASYS both arms were deeply sedated.
Higher PEEP was used in ROSE than in ACURASYS.
ACURASYS was conducted in France, ROSE in the U.S.
ROSE enrolled patients within 8 hours, ACURASYS within 16 hours.
Authors suggest that improvements in critical care, such as reduced sedation, resulted in the negative result in ROSE relative to ACURASYS. But assuming ROSE is "true" and there was proper randomization in ACURASYS, one would have to also conclude either 1) that the control arm in ACURASYS got particularly worse care, i.e., inadequate blinding led to care differences between groups; or 2) paralysis is lifesaving for ARDS to precisely the same degree that deep sedation is lethal.
That is, if ACURASYS is “true” and paralytics save lives in ARDS, then lighter sedation in ROSE’s control arm would have to have saved an implausibly high number of lives (about 45), relative to deep sedation, to erase paralytics’ survival benefits in the deeply sedated arm.
But one doesn't need to resort to such convoluted reasoning, because there was another important difference between the two trials: almost half of patients in ACURASYS were placed in prone position, while only 16% of patients in ROSE were proned.
Prone positioning for ARDS was found to be lifesaving in another French trial and in a meta-analysis, and is recommended by U.S.-based professional societies for severe ARDS, but U.S. centers have been slow to adopt the practice. In ROSE, it was left up to the individual treating physician’s discretion, and in 5 of 6 patients, they did not use the option (likely because their centers weren’t experienced with proning).
Authors acknowledged this in their discussion,
Whether early continuous neuromuscular blockade is more effective with prone positioning is unknown, but it is a possible explanation for the different results of our trial and the ACURASYS trial."
Observing the similar mortality in ACURASYS’s control arm (40.7%) with ROSE’s (42.8%) with similar severity of ARDS, the huge variance in the use of prone positioning, and the absence of other good reasons, that is a possibility that must be considered.
Also plausible is that continuous neuromuscular blockade doesn’t help at all, and ACURASYS’s positive findings were due to unmeasurable bias (like care teams noticing which patients were paralyzed). This would be a disappointing but familiar repetition of other ‘practice-changing’ trials that were contradicted by larger subsequent studies.
Thanks to ROSE, it does seem clear that continuous neuromuscular blockade is not superior to as-needed use in supine ventilated patients with ARDS, and it may cause cardiovascular harm, which is very helpful to know. (Of course, any cardiovascular harm was insufficient to change all-cause mortality, meaning the practice can be considered appropriate on an individualized basis.)
Now, it would be good to see U.S.-based investigators finally turn their attention to prone positioning for ARDS, whose benefits have remained just unproven enough to allow hospitals to defer the hassle of deployment. Which in turn (because the lack of widespread use stops proning from being considered standard care) gives academicians ethical cover to conduct ARDS studies (on ECMO, paralytics, and other more-fun stuff) without the logistical burdens of including prone positioning in the design.
PROSEVA was published in 2013; ROSE started enrollment in 2016, according to clinicaltrials.gov. U.S. societies endorsed prone positioning in 2017. Yet there are no large U.S.-based trials on ARDS and prone positioning in process or planned on clinicaltrials.gov. That choice by researchers and policymakers is to practicing physicians’ dismay, and may be to patients' detriment.
Until that changes, the awkward question trailing ROSE, EPVent-2 and any upcoming study of ECMO for ARDS will be: “Sure, but what if they had proned the patients?” (And why didn’t they?)