The PHARLAP Trial: Recruitment Maneuvers Buy Stairway to Heaven?
Jon-Emile S. Kenny MD [@heart_lung]
“Dying is a wild night and a new road.”
In 2006, a recruitment maneuver to achieve an ‘open lung’ was circulated; it has since been dubbed ‘stairway recruitment’ given its stepwise upthrusts in airway pressure [see figure 1]:
Figure 1: The maneuver begins with a PEEP of 25 and is carried out in pressure control with a driving pressure of 15 cm H2O. The first phase is 4 minutes long oscillating between 25 and 40 cm H2O. The entire maneuver is 20 minutes long. The y-axis is pressure in cm H2O.
Plodding up this ladder of airway pressure was meant to completely open collapsed alveoli in the ARDS lung. As described previously, surface tension, ambient interstitial pressure and the chest wall all conspire against an open alveolus and must be overcome; this may demand 45 to 60 cm H2O of pressure.
Thus, the ‘stairway’ recruitment approach – with modifications – provided the foundation for three investigations discussed below: The Open Lung Approach [OLA], the Alveolar Recruitment for Acute Respiratory Distress Syndrome Trial [ART] and the most current Permissive Hypercapnia, Alveolar Recruitment, and Low Airway Pressure Trial [PHARLAP].
What They Did
PHARLAP recruited patients with moderate-to-severe ARDS [i.e. P/F ratio of less than 200] for less than 72 hours duration and randomized them to either a modified ‘stairway’ recruitment, or conventional care as delineated by the original ARDSnet ‘low PEEP’ titration table – an experimental protocol very similar to the ART trial.
Pertinent exclusion criteria were as follows: ventilated for more than 10 days, younger than 16 years of age, evidence of barotrauma, active bronchospasm or significant obstructive or restrictive pulmonary disease, suspicion of raised intracranial pressure, unstable cardiovascular status, pregnancy, or were receiving extracorporeal membrane oxygenation or high-frequency oscillatory ventilation, imminent death, or if there was a ‘lack of treating physician equipoise.’
The stairway recruitment employed did away with resetting PEEP to 25 cm H2O between each jump in PEEP as described in the 2006 protocol. In PHARLAP, PEEP was increased by 10 cm H2O from 20 cm to 40 cm H2O for 2 minutes each – all with a driving pressure of 15 cm H2O in pressure control. This step was followed by decremental PEEP titration from 25 cm H2O in 2.5 cm H2O steps for 3 minutes until the SpO2 decreased by 2% or more or to a minimum of 15 cm H2O PEEP if desaturation did not occur. Last, a brief recruitment maneuver returned the PEEP to a level 2.5 cm H2O above the level of desaturation.
The primary outcome was ventilator-free days at Day 28. Secondary outcomes included mortality, barotrauma, new use of hypoxemic adjuvant therapies, and ICU and hospital stay.
What They Found
Enrolment halted October 2, 2017, after publication of ART; 115 patients were randomized, 58 to the intervention group and 57 to the control group. Notably, 175 additional were eligible but not enrolled – 88 of which were omitted because of ‘clinician preference.’
At Day 28, there was no difference in ventilator free days in the PHARLAP intervention and control groups. Further, there were no differences in mortality, rate of barotrauma, rate of pneumothorax requiring a chest drain, or length of stay.
In roughly 13% of patients in the intervention group [13/102] there was clinically significant hypotension during the stairway recruitment that required further management – typically an increased dose of vasopressors.
Since the publication of ART, we have been left wondering why that trial demonstrated a clinically and statistically significant increase in mortality in those randomized to ‘stairway recruitment’ and higher PEEP – that is, an open lung strategy.
Does PHARLAP help dissect out the mechanism of harm in ART? Not definitely. Yet I agree with the thoughtful commentary included with the publication of PHARLAP. There is an undeniable suggestion of harm with the open lung strategy in addition to its cumbersome application at the bedside. Indeed, the authors of PHARLAP conducted their own mini meta-analysis including ART, OLA, PHARLAP and a smaller proof-of-concept study that they had previously published; each of these trials employed a ‘stairway’ type recruitment maneuver with higher PEEP. In totality, these 4 trials showed no difference in mortality, but an increased risk of barotrauma in the intervention group. Yet one may still rightfully ask: is it the stairway recruitment or higher PEEP that increased mortality in ART?
To better discern the cause, I have created a comparison table below – limited to patients with ARDS and a P/F ratio of less than 200 [i.e. moderate-to-severe ARDS]. Briel et al. is an earlier meta-analysis including the ALVEOLI, LOVS and EXPRESS trials [note that the PEEP level presented for Briel et al. includes all patients in that analysis – most of which had moderate-to-severe ARDS]. The three trials that employed ‘stairway’ recruitment [only in the high PEEP/intervention arms] are highlighted in pink. Additionally, I have included the control arms of four important studies – OSCILLATE, OSCAR, EPVENT2 and ROSE. The former two studied the effect of high frequency ventilation, while the latter two investigated the effects of PEEP titration by esophageal manometry and neuromuscular blockade, respectively. The control groups for OSCILLATE, EPVENT2 and ROSE employed the ‘higher PEEP’ table initially used in ALVEOLI and LOVS, while OSCAR used the ‘lower PEEP’ table used in the landmark ARDSnet trial.
Table 1: Briel et al. represents a meta-analysis including ALVEOLI, LOVS and EXPRESS - the PEEP levels displayed [+] represent all patients in the meta-analysis while the 'n=' and mortality displayed [*] are only for those with moderate-to-severe ARDS and represent 'in house' mortality. The control group of OSCAR [**] reported 30-day mortality. The three trials highlighted in pink [OLA, ART and PHARLAP] employed a modified 'staircase' recruitment maneuver only in the 'high PEEP' arms. Physiological data represents 24 hours of intervention. Note driving pressure is mediated by respiratory system compliance and applied tidal volume - making comparisons challenging.A cursory comparison of weighted-mortality shows that within the ‘high PEEP’ trials, mortality was greater for those three that employed stairway recruitment compared to those that used ‘high PEEP’ titration tables alone [48% versus 34%, respectively – note these are my rough calculations]. Within the lower PEEP groups, the control arms from OLA, PHARLAP and ART [these arms did not receive stairway recruitment] had a mortality rate similar to the other lower PEEP trials 44% versus 40%, respectively.
Yet the weighted-barotrauma incidences did not appear to mirror this mortality trend – the rate of barotrauma was similar amongst the higher PEEP trials [7% versus 8.3% for ‘stairway’ versus ‘non-stairway’ trials, respectively] compared to the lower PEEP trials [3% versus 6.8%, respectively – again, no stairway recruitment was used in any of the lower PEEP arms and these are my rough, non-systematic calculations].
If it is not barotrauma driving the mortality difference, what might it be? I think the answer lies in hemodynamics. As above, 13% of the intervention arm in PHARLAP required increased vasopressors during recruitment. In OLA, 17 of 99 in the stairway arm either developed hypotension during recruitment or were considered ‘too unstable’ to undergo recruitment. Similarly, in ART, 35% receiving stairway recruitment had either hypotension or need for vasopressors within 1 hour of protocol initiation compared to 28% in its control arm. As a comparison, in the control arm of OSCILLATE [which was higher PEEP only] just 24% required an increase in vasopressors upon protocol initiation. Additionally, recall that in ART, the stairway recruitment protocol was lightened after 50% of patients had been randomized because of three cases of resuscitated cardiac arrest! In a previous entry, I presented a thought experiment illustrating that airway pressures cycling between 45 and 60 cm H2O [see figure 1] could quite easily place the entire lung in West Zone II – making acute cor pulmonale a real risk.
If we consider moderate-to-severe ARDS to have a physiology akin to acute pulmonary embolus – we can understand why hypotension is hemodynamically hazardous. Acute cor pulmonale physiology may also explain the seemingly higher PaCO2 observed in the arms receiving ‘stairway’ recruitment; that is, high dead space reflecting iatrogenic right ventricular afterload [see table].
Finally, we may wonder why the results of ART were disparate from OLA and PHARLAP? While it may be due to different populations, disease severity and healthcare resources – as suggested by the authors of PHARLAP – I think there is another important consideration. There was a much greater fraction of patients in PHARLAP who were eligible but not enrolled because of ‘clinician preference;’ that is, of 115 total patients randomized, 88 additional patients were eligible but were withheld at the discretion of the treating physician. In ART, there were over 1000 patients randomized but only 17 additional patients were withheld for ‘clinician preference.’ One wonders if this preference was over hemodynamic instability?
Given the lack of benefit, complexity and potential harm, it seems that ‘stairway recruitment’ should pass over to the ICU intervention afterlife. The authors of PHARLAP may argue that those who have evidence of pre-maneuver lung recruitment may still be eligible, but I wonder if pushing a struggling right ventricle to the precipice of hypotension makes it worthwhile. I also wonder how prone position mediates these hemodynamic connections.
Dr. Kenny is the cofounder and Chief Medical Officer of Flosonics Medical; he also the creator and author of a free hemodynamic curriculum at heart-lung.org