The problem with treating ARDS using mechanical ventilation is that treatment starts too late in the disease process. We have recently published two papers in a high fidelity clinically applicable animal model showing that preemptive application of the appropriate ventilation strategy actually PREVENTS ARDS. In our first paper (Roy S et al J Trauma 2012) we tested early application of airway pressure release ventilation (APRV) using very precise settings immediately post injury (peritoneal sepsis plus gut ischemia/reperfusion) and showed that we could prevent the development of ARDS. In the second paper (Roy S et al Shock 2013) we compared preemptive APRV with the standard of care ARDSnet low tidlal volume (Vt) ventilation, which was applied late after injury developed. Again APRV prevented ARDS while the low Vt did not stop disease progression and all animals in the ARDSnet group developed established-ARDS. Our efforts should not be at treating established ARDS because it is refractory to treatment. In the recent ALIEN paper by Villar he showed even with low Vt ventilation mortality is still greater than 40%, clearly showing that low Vt is not very effective. Indeed, nothing is very effective at treating established ARDS therefore we must attempt to prevent the disease from occurring. In the paper by Roy et J Trauma we discuss ARDS being a staged disease similar to cancer and the key to reducing ARDS mortality is application of treatment in an early ARDS stage, similar to reducing mortality for cancer.

I agree we have to discuss these results, and in particular those of us who use HFOV. We have used this technique as a rescue technique in severe ARDS for 15 years now and have had good results in our unit (long-term survival around 50%). Our indication to use it is however different from the two studies that used a PaO2/FiO2 ratio of 26,7 kPa (200 mmHg), a value that after the new Berlin ARDS definition is the start of moderate severity ARDS. What these two studies demonstrates is that using HFOV in moderate severity ARDS probably have no effect, and may even worsen outcome. The studies does not tell us the effect of HFOV in a group of severe ARDS only (ratio

Thanks for your comments.

Hans, I have been asking myself why was the higher mortality seen in the HFOV group for oscillate? It sounds like you have more HFOV experience than I do and would be interested in your thoughts. Based on my understanding of the physiology and respiratory mechanics of ARDS, in moderate disease you might not see an improvement with HFOV, but should have at least similar outcomes. I wonder if the more aggressive HFOV strategy used in Oscillate created more hemodynamic disturbance, resulting in more fluids as well as the reported higher inotrope usage. A more positive fluid balance in the HFOV group would be detrimental.

I agree absolutely Gary, prevent ARDS in the first place. I don't use APRV, but had similar observations with BiLevel. I would be interested in those animal papers, do you have the full reference to save a small amount of searching time? I have started treating all my patients at risk of ARDS (sepsis, pneumonia, massive transfusion etc) as through they have ARDS, reaching for BiLevel quickly if standard LPV is not well tolerated. I haven't had a patient progress to ARDS recently, but I take care of a low risk population now (cardiac/thoracic surgery pts). As a result, these days I primarily get involved in ARDS when someone else has pushed the ECMO button. I have been pleasantly surprised how well our ARDS patients do on vv-ecmo, provided we provide total lung rest. This fits with your observation, recognise the at risk patient and prevent ARDS and once established perhaps no vent based strategy can provide a satisfactory mortality outcome, although there are clearly bad vent strategies. I have been impressed with the recent ultra-low TV techniques supported by ECCOR in various studies. Mostly observational but there was a small study in 2009 (Terragni et al) that was particularly compelling.

Hi Matthew. Good question, I am also a bit surprised with the results from the study of Ferguson with increased mortality in HFOV group, which differs from the other (and larger) UK based study. At least we can conclude that we can put a question-mark with the increased mortality finding. In the Ferguson study some clue may be in the supplementary appendix. The HFOV group was more heavliy sedated, they had a more positive fluid balance and reveived more inotropes, indicating that haemodynamic stability indees was more problematic in that group. I am also surprised by the very high frequency of neuromuscular blocking agent use in both groups, this is very rarely necessary in my unit (HFOV or IPPV). Some spontaneous respiratory activity is possible and maybe good in patients with HFOV (you just have to adjust the lower pressure alarm limit since it drops during an inspiratory effort). In fact we have occationally weaned patients directly using HFOV and gradually lowering Mean Airway Pressure and allow for more sponatenous ventilation. Interesting is also their post-hoc subgraoup analysis where groups of more severe ARDS (low O2 ratio/low compliance) had more equal results regarding mortality.

I read these studies primarily as giving evidence that the supposed lung protective effects of HFOV does not translate into clinically relevant benefits. The question of HFOV as rescue therapy remains unanswered. The treatment of severe ARDS is complicated despite large improvements in evidence the last decade. I think it is important to aknowledge the recent recommandations in the Berlin defintions. There is now overwhelming evidence that protocolized prone position markedly reduces mortality in severe ARDS. Very impressive numbers were presented by Guerin at ESICM-Lisbon. At in least in early ARDS, prone position should be more widely implemented. In my experience high dose cistratracurium infusions are also useful. There is good evidence from both animal and human studies that support ventilation in contrast to controlled ventilation is detrimental in severe ARDS, an effect clearly not seen in moderate ARDS. Addressing the alternative modes of ventilation (HFOV, ECCO2, APRV), there is still a profound lack of evidence. With the current evidence, ECMO seems as the modality of choice. Hopefully technical advances will reduce the present barriers to providing ECMO in the future.

One thing to keep in mind is that the HFO ventilators used in these 2 studies do not adequately allow for spontaneous breathing and this is probably the explanation for the greater use of sedatives and neuromuscular blockers in the HFO group. The design of the Sensormedics HFO vent has not changed substantially since the first neonatal trials in the 1980s. It may be that the results of these studies reflect the limitations of current technology rather than high frequency ventilation per se.

Hans, thanks for adding the appendix. I wasn't as surprised to note the frequency of neuromuscular blocker use in the HFOV. In the different places I have worked I have noticed a lot of providers reach for this at the same time as starting HFOV. I have not adopted this practice, and try to only use NMB in accordance with the french study, if at all. I am interested to note your experience of allowing some spontaneous breathing on HFOV and directly weaning from HFOV. Your centre certainly has a lot of experience with HFOV. Based on your experience what do you think about OSCAR, where they rolled out HFOV in centres with no previous HFOV experience? Do think that could impact HFOV outcomes?

Matthew, another interesting point not easy to retrieve from the study. A majority of the ICUs were as you said new to this treatment. At least in our unit it took some time to get used to this very different form of ventilation, I will say a couple of years since we did not use it as frequently when we started compared with today. This fact may of influence the results, which the researchers have tried to answer in a post-hoc analysis of results according to the number of included patients to the study. That analysis did not reveal any significant diifferences, but we do not know if the rate of recruitement to the study had anything to do with prior experience or not. Such data is not revealed, but would be interesting to see.

Anoher comment regarding treatment in general. We, as dr. Strand above, also have good experience with using prone position as our first action to treat moderate ARDS. We practice at least 16 hours prone before turning supine and repeat 1-2 times if necessary (responders). Those with little/no sustained response are then candidate for HFOV when PaO2/FiO2 ratio approaches the limit for severe ARDS (13,3 kPa/100 mmHg). We might also add inhaled NO if we have evidence of pulmonary hypertension, but this is infrequently necessary. Then of course, if patient deteriorates we turn to ECMO whith is a ratio < 8 kPa on two occations with 4-6 hours space. In some patients with severe respiratory acidocis we use Novalung (extracorporel CO2 removal) to keep pCO2 normal. Since we started to use this algoritm our need to use ECMO have been very low, less than once a year (ECMO for pulmonary indications only).

I think it is important to be clear on what the oscillate trial involved. These were patients in which HFO was instituted much earlier than is the practice in most places. It stands to reason that some of the patients in this group may never have received HFO using a more conventional/rescue approach. I don't believe that it would be justified to use these results to move away from the use of HFO when conventional ventilator methods are failing. I believe the message here is that in patients with early or developing ARDS the negative effects of HFO (Hemodynamic effects, increased need for sedation or paralysis) outweigh the potential gains from there may be from lung protection. In more advanced cases in which the conventional methods are failing this may not be the case.