Effect of Lung Recruitment and Titrated PositiveEnd-Expiratory Pressure (PEEP) vs Low PEEP on Mortalityin Patients With Acute Respiratory Distress Syndrome. A Randomized Clinical Trial, by The ART Investigators. 
Well, before we really start, and this is going to be a bit unusual, I gotta tell you how proud I am to write my thoughts about this specific trial. We know how bloody hard is to run a major trial like this in a third world country like Brazil. And see it published in a major journal fills my heart with joy. Amazing! Good job guys! BUUUUTTTT, we are what we are. And boy, when comes to discussing high end critical care research, we can be a pain in the ass.
We already discussed lung recruitment here few months ago (link), and how Carol Hodgson’s systematic review didn’t show any real benefit of recruitment maneuvers (RM) in both 28-day mortality and in-hospital mortality. Also, RM “appeared” to be safe (didn’t increased the risk of barotrauma). I mean, “appeared”, because if you take a blood gas during a RM you’ll never recruit a patient again. Anyhow, the point is, there is a good physiological background in “open the lung and keep it open” strategy (minimizing atelectrauma and possibly dynamic
overdistention) and the ART investigators knew it. They also kne a good trial evaluating the specific role of lung recruitment with PEEP titration based on pulmonary compliance was still missing. The stage for the ART trial was set.
The ART trial was a multicenter (9 countries and 120 ICUs), randomized controlled trial. Patients receiving mechanical ventilation (MV) with ARDS (moderate to severe) for less than 72h were evaluated for inclusion. The inclusion process had 2 steps: screening phase and confirmatory phase. For the screening phase all potential patients received at least 3h of MV based on the ARDSNet protocol , after that, the FiO2 was set at 100% and the PEEP at 10cmH2O or more for 30 minutes. Eligibility was confirmed if the P/F ratio was 200 or lower after this phase. The exclusion criteria were: <18y/o, MAP<65mmHg or increasing use of vasopressors over the past 2h, contraindications to hypercapnia, pneumothorax, subcutaneous emphysema or pneumomediastinum.
Patients were randomized either to RM or low-PEEP strategy. They used block randomization stratified by site, age ( ≤55 or >55y) and P/F ratio (≤100 or >100).
The low-PEEP strategy group (control) continued receiving MV under the ARDSNet protocol -> Vt 4-6ml/kg of predicted body weight to maintain plateau pressure ≤30cmH2O, with PEEP and FiO2 adjusted according the ARDSNet table(below).
The recruitment strategy in the intervention group resambles the strategy in the OLA trial . All patients were sedated and received neuromuscular blocking agents before the RM.
-Maximum alveolar recruitment maneuver*: PCV, RR 15/min, FiO2 100%, I:E ratio 1:1 -> start with PEEP=25 and delta pressure=15 for 1 min -> increase PEEP to 35 keeping delta pressure=15 for 1 min -> increase PEEP to 45 keeping delta pressure=15 for 2 min. Note the driving pressure was always 15.
-PEEP titration*: immediately after maximum alveolar recruitment, the following ventilation settings should be done: PEEP 23, Vt=5ml/kg of PBW, RR=20, mode:VCV, flow 30L/min (square wave) and FiO2 100%. All the parameters should remain constant, except for the PEEP. They start the titration with PEEP of 23, and after 4 min the static compliance was calculated (C= Vt/(plateau pressure-PEEP), the PEEP was decreased in steps of 3 until 11, always calculating the C after 4 min. The optimal PEEP was the PEEP value at maximum compliance + 2cmH2O.
-New recruitment after PEEP titration*: settings were again changed (PCV, RR 15/min, FiO2 100%, I:E ratio 1:1) and a new recruitment with PEEP of 45 for 1 min was made. Than the ventilation was set to VCV, with the optimal PEEP, Vt 6ml/kg of PBW.
*In June 2015, with 555 patients randomized, “after 3 cases of resuscitated cardiac arrest possibly associated with the experimental group treatment were observed”, the steering committee decided to modify the RM protocol. Now, PEEP in maximum alveolar recruitment should be 25->30->35, in steps of 1 min (always keeping the driving pressure <15); the decremental PEEP should be in steps of 3 min and the new recruitment with PEEP of 35. The possible causes for the arrests were: 2 respiratory acidosis and 1 hemodynamic collapse, all of them possible related to study interventions . Any protocol changes during a trial is a drawback, but in this case it had to be done. Safety comes first.
If P/F ratio was stable or increasing for 24h after RM, PEEP was weaned by 2cmH2O every 8h.
The primary outcome was 28-day mortality. It was an event driven study, meaning, the study was designed to continue until 520 events were reached. When it comes to survival analysis, the study power relies on the number of events by the trial’s end. Therefore, this number of events, assuming a hazard-ratio of 0.75 and type I error of 5%, would provide 90% power. Two interim analysis were performed (after 1/3 and 2/3 of patients were recruited). The trial would have been stopped if there was evidence of harm with 1-sided p <0.01. Early stop due to intervention benefits was not planned. An interesting point is that although the authors published in the statistical protocol that “we will not adjust p values for multiple comparisons”, they stated in the article: “the significance level for the primary outcome final analysis was .042, to maintain overall alpha at 0.05”. The itch in my ass is because why this adjusted p value of 0.042 wasn’t specified in the analysis plan [4,5] and on a more holistic bio-psycho-social-statistical way of thinking, although the alpha spending is a reality in interim analysis, it might not be entirely true in safety-only interim analysis, in these cases, some recommend that the p-value should not be adjusted for the final analysis. For me, it was clear in both statistical analysis plans [4,5] the final p value would be <0.05. In the end it doesn’t matter, but can you imagine if the final p (SPOILER ALERT), instead of 0.041 was 0.043? Since we’re living “the p dictatorship”, it would have been crazy!
From November 2011 to April 2017 1013 patients were randomized (501 intervention vs 512 control). Table below shows the baseline characteristics of the patients:
We can see that near 2/3 of patients had septic shock, the major causes for ARDS were: pneumonia, sepsis, gastric aspiration and major surgery (other than cardiac). The table below shows the baseline respiratory measures before recruitment:
I gotta one thing to say: WHY IN THE HELL only 10% of prone position in both groups??? Really???? Well, maybe that’s because prone positioning was a resource for refractory hypoxemia, as described in the protocol . But why? Why deprive patients of such amazing resource in ARDS treatment?
The Real Deal
Almost every patient in the intervention group (95.8%) was recruited. In almost 16% recruitment had to be stopped, mostly due to hypotension or decrease in SpO2. The mean PEEP after titration was 16.8cmH2O. The recruitment protocol demanded a new recruitment maneuver after PEEP titration, however, only 78.4% of patients were re-recruited. Patients in the intervention group had higher PEEP, higher plateau pressure, P/F ratio and lower driving pressure than controls from day 1 to 7 (below).
Although more patients received neuromuscular blocking agents (NMBA), as required by protocol (96,8%) in the intervention group, we still had lot of patients using NMBA in the control group (73.3%). Yeah, its easier to pump some NMBA than prone a patient.
The primary outcome favored control group (49.3% vs 55.3%, HR 1.2, 95% CI, 1.01-1.42; P = 0.041), with the effect persisting after adjusting for covariates (HR 1.22, 95% CI, 1.03-1.45; P = 0.02). The Kaplan-Meier below shows the curves separating between day 3 and 4. The exploratory secondary outcomes showed no difference in refractory hipoxemia or severe acidosis. The intervention group had more pneumothorax and barotrauma. None of the subgroup analysis showed benefit favoring the intervention.
I must tell you, there’s an unmeasurable beauty in trials showing opposite results than expected, like this one, when everybody thought lung recruitment would save lives but it ended up killing people. Because that’s what it did. And that’s not a bad thing, I mean… A question has been answered. “But the physiology, recruitment maneuvers….” C’mon! We used to give prophylactic antiarrhythmics to MI and you know how it ended up! There are more things in ARDS than meet the eyes, or in this case, the way we think a disease process work. “If I improve the P/F ratio I’ll save lives… blah blah blah…”. Recruitment for ARDS is no more and prone positioning rules! Go PRONE!
1. Cavalcanti AB, Suzumura A, Laranjeira LN et al. Effect of Lung Recruitment and Titrated Positive End-Expiratory Pressure (PEEP) vs Low PEEP on Mortality in Patients With Acute Respiratory Distress Syndrome JAMA. 2017; 318(14):1335-.
2. Ventilation with Lower Tidal Volumes as Compared with Traditional Tidal Volumes for Acute Lung Injury and the Acute Respiratory Distress Syndrome N Engl J Med. 2000; 342(18):1301-1308.
3. Kacmarek RM, Villar J, Sulemanji D et al. Open Lung Approach for the Acute Respiratory Distress Syndrome Critical Care Medicine. 2016; 44(1):32-42.
4. Damiani LP, Berwanger O, Paisani D et al. Statistical analysis plan for the Alveolar Recruitment for Acute Respiratory Distress Syndrome Trial (ART). A randomized controlled trial . 2017; 29(2).
5. The ART Investigators. Rationale, study design, and analysis plan of the Alveolar Recruitment for ARDS Trial (ART): Study protocol for a randomized controlled trial Trials. 2012; 13(1).