In Reply:-The letters from Oppenheim and Pizov and from Findlay et al. indicate an important potential drawback of expiratory washout: auto-PEEP resulting in an increase in inspiratory plateau airway pressure, which may reintroduce a risk of lung barotrauma.
Different solutions exist to overcome EWO-induced auto-PEEP. As suggested by Findlay et al., one is to deliver EWO during pressure-controlled ventilation. In such circumstances, because EWO-induced increase in airway pressure automatically results in a decrease in the tidal volume, the lack of EWO-induced auto-PEEP is associated with a reduction in the efficiency of EWO on CO2elimination. In Findlay's study, a 6 l/min EWO flow delivered during pressure-controlled ventilation was associated with a modest reduction in PaCO2(-14%). As a consequence, this technique for counteracting EWO-induced auto-PEEP appears to have serious limitations. It is more simple and efficient to reduce the PEEP level. In a recent study, we could demonstrate in a series of patients with ARDS that inspiratory plateau airway pressure, mean airway pressure, and consequently arterial oxygenation and pulmonary shunt could be maintained constant when the PEEP level was reduced of an amount equal to EWO-induced auto-PEEP. In this study wherein a 15 l/min EWO flow was used, PaCO2could be reduced by 26% if compared with control values without EWO. 
As emphasized by Findlay et al., it is technically possible to measure EWO-induced auto-PEEP by synchronizing suppression of EWO and occlusion of the expiratory valve. This technical option, already present in prototypes, should be available on any medical device providing EWO and should be considered as a critical element of safety. Findlay et al. have outlined an important limitation of our study, although the indication for EWO was unquestionnable-severe ARDS with lung hyperdensities involving 50% of the lung parenchyma, low respiratory compliance, and profound hypercarbia with acid-base compromise resulting in increased pulmonary artery pressure -the technical conditions of EWO administration were not optimized. EWO-induced auto-PEEP should have been counterbalanced by a concomittant reduction in PEEP, a technique that does not compromise CO2elimination and arterial oxygenation. We agree with Findlay et al. that our study should not bias intensivists against the use of EWO in patients with severe ARDS.
Another solution proposed by Oppenheim and Pizov to limit the auto-PEEP resulting from tracheal gas insufflation is to replace oxygen by helium. In an interesting study, these authors have demonstrated that helium is more effective than oxygen for improving CO sub 2 elimination and limiting auto-PEEP. However, helium does not totally avoid the increase in peak inspiratory airway pressure and therefore appears less effective than decreasing PEEP. In addition, its high cost and the risk of decreasing FIO2and worsening arterial oxygenation in patients with ARDS render problematic its routine use in clinical practice.
Although we agree that EWO is a peculiar mode of tracheal gas insufflation, we think that it should be clearly individualized from a semantic point of view. The administration of an intratracheal gas flow during inspiration-inspiratory bypass-increases tidal volume and therefore cannot be considered as a new mode of CO2elimination. In contrast, EWO that removes the carbon dioxide-laden gas occupying the instrumental dead space without inducing major changes in tidal volume appears as an original technique for enhancing CO2clearance in patients treated by permissive hypercapnia. In our opinion, EWO rather than inspiratory bypass or continuous tracheal gas insufflation could be clinically used in the near future if technical problems related to the measurement of auto-PEEP are solved by manufacturers of ICU ventilators.
Pierre Kalfon, M.D.
Chef du Service de Reanimation; Polyvalente Hopital des Diaconesses; Paris, France
Jean-Jacques Rouby, M.D., Ph.D.
Professor of Anesthesia and Critical Care; Chef de l'Unite de Reanimation Chirurgicale; Hopital de la Pitie-Salpetriere; Universite Paris VI; Paris, France
(Accepted for publication November 18, 1997.)