In reply: Non-ventilated lung airway occlusion during one-lung ventilation: a need for further research?

Abstract

To the Editor, Dr. Pfitzner suggests that when occluding the nonventilated limb of the double-lumen endotracheal tube (DL-ETT) prior to pleural opening, an event similar to tension pneumothorax (TP) could happen, where hyperinflation of the non-ventilated lung (instead of the pleural space) would cause a cardiovascular collapse. We wish to reassure him that this scenario could not happen since it would violate several laws of physics. First, developing a TP or TP-like event requires a defect acting as a one-way valve (OWV), allowing gases to enter the pleural cavity (or non-ventilated lung) during inspiration while preventing egress of gas during expiration. The OWV is created by the collapse or flapping of soft tissues at the leak location during expiration. In the scenario described by Dr. Pfitzner, the leak location around the bronchial cuff of the DL-ETT sits inside a fairly rigid main bronchus where no soft or floppy tissues could act as an OWV. Another OWV mechanism is a ball-valve mechanism, which requires a blood clot or very thick secretions to act as the closing part of the valve. The blood clot or secretions would need to form in a matter of minutes. While such phenomena have been described with a regular ETT, they are rarities. For such a rare phenomenon to appear with a DL-ETT within a few minutes would be an extreme rarity. Second, even when assuming that such an improbable OWV mechanism will happen, some important factors are to be considered. In the cited swine model, air injected in the pleural cavity is virtually not absorbable and hemodynamic stability was observed for TP up to 57% of the total lung capacity (representing 3,400 mL for humans). Considering a starting point at residual capacity (Dr. Piftzner showed 96.6% gas venting completion within 60 sec of initiation of one-lung ventilation) and 500 mL oxygen absorption within ten minutes, hemodynamic stability is expected for leaked volume of at least 3,900 mL. Since we are ventilating the patients with highly absorbable 100% oxygen rather than air, as the nonventilated lung reinflates, additional oxygen would be absorbed, and it could easily be envisioned that a leak of 5,000 mL would be required for hemodynamic instability. Third, the flow rate of the leak is not constant but a function, in part, of the pressure gradient and compliance of the non-ventilated lung. As the non-ventilated lung reinflates, compliance will decrease and intra-bronchus pressure will increase (reducing the pressure gradient). A rapid exponential decrease of the leak flow rate will ensue. An average ‘‘reduced 25 mL per breath’’ (375 mL per min) over ten minutes could require the initial leak to be three or four times higher and be so large that it would be nearly impossible to miss within the first few breaths (either by J. Somma, MD, BEng J. S. Bussières, MD (&) Department of Anesthesiology, Institut Universitaire de Cardiologie et de Pneumologie de Québec, Quebec City, QC, Canada e-mail: [email protected]