Stefano Tredici

A prototype of a liquid ventilator using a novel hollow-fiber oxygenator in a rabbit model

Objective:A functional total liquid ventilator should be simple in design to minimize operating errors and have a low priming volume to minimize the amount of perfluorocarbon needed. Closed system circuits using a membrane oxygenator have partially met these requirements but have high resistance to perfluorocarbon flow and high priming volume. To further this goal, a single piston prototype ventilator with a low priming volume and a new high-efficiency hollow-fiber oxygenator in a circuit with a check valve flow control system was developed. Design:Prospective, controlled animal laboratory study. Setting:Research facility at a university medical center. Subjects:Seven anesthetized, paralyzed, normal New Zealand rabbits Interventions:The prototype oxygenator, consisting of cross-wound silicone hollow fibers with a surface area of 1.5 m2 with a priming volume of 190 mL, was tested in a bench-top model followed by an in vivo rabbit model. Total liquid ventilation was performed for 3 hrs with 20 mL·kg−1 initial fill volume, 17.5–20 mL·kg−1 tidal volume, respiratory rate of 5 breaths/min, inspiratory/expiratory ratio 1:2, and countercurrent sweep gas of 100% oxygen. Measurements and Main Results:Bench top experiments demonstrated 66–81% elimination of Co2 and 0.64–0.76 mL·min−1 loss of perfluorocarbon across the fibers. No significant changes in Paco2 and Pao2 were observed. Dynamic airway pressures were in a safe range in which ventilator lung injury or airway closure was unlikely (3.6 ± 0.5 and −7.8 ± 0.3 cm H2O, respectively, for mean peak inspiratory pressure and mean end expiratory pressure). No leakage of perfluorocarbon was noted in the new silicone fiber gas exchange device. Estimated in vivo perfluorocarbon loss from the device was 1.2 mL·min−1. Conclusions:These data demonstrate the ability of this novel single-piston, nonporous hollow silicone fiber oxygenator to adequately support gas exchange, allowing successful performance of total liquid ventilation.

Clinical design functions: round table discussions on the bioengineering of liquid ventilators.

During the 6th International Symposium on Perfluorocarbon Application and Liquid Ventilation, a round table discussion on bioengineering was held in which different experts shared their opinions and experiences about the use of a total liquid ventilator design for clinical applications. To structure the discussion, all experts were invited to contribute their knowledge within the context of three matrixes related to the liquid ventilators: 1) function and technology, 2) ventilation modes, and 3) risk analyses. The outcome of this international conference recommends continued development of a total liquid ventilator toward clinical applications.

Effect of repeated induced airway collapse during total liquid ventilation.

Negative pressure generated during the expiratory phase of total liquid ventilation (TLV) may induce airway collapse. Evaluation of the effect of repeated airway collapse is crucial to optimize this technique. A total of 24 New Zealand White rabbits were randomly divided into four groups. Ventilation was performed for 6 hours with different strategies: conventional gas ventilation, TLV without airway collapse, and TLV with collapse induced in either 75 or 150 sequential breaths. In the treated groups, airway collapse was induced by increasing the perfluorocarbon drainage velocity while maintaining the minute ventilation constant. Airway pressure, gas exchange, and blood pressure were monitored at 30-minute intervals. At the end of the experiment, airway and lung parenchyma specimens were processed for light microscopy. No evidence of fluorothorax was noticed in any of the four groups at autopsy examination. Minimal signs of inflammation were noticed in all airway and lung parenchyma specimens, but no evident structural alteration was visible. Adequate gas exchange and systemic blood pressure were maintained during all the studies. Repeated airway collapse is not associated with structural changes in the respiratory system and does not alter the gas exchange ability of the lungs.

Location of flow limitation in liquid-filled rabbit lungs

The effects of end-inspiratory lung volume (EILV) and expiratory flow rate (Q) on the location of flow limitation in liquid-filled lungs were investigated by measuring pressure along the airways and by radiographic imaging. The lungs of New Zealand white rabbits were filled with perfluorocarbon to the randomly selected EILV of 20, 30, or 40 ml/kg, and the volume was actively drained at one of three Q: 2.5, 5.0, or 7.5 ml/s. The minimum pressures recorded by a movable catheter at locations along the airways show that flow limitation occurred in the main bronchi and trachea, and was independent of EILV and Q. The minimum pressure at the trachea was –80 mm Hg compared with values that were more positive than –10 mm Hg at a location 3 cm distal to the carina for all EILV and Q combinations. This location was confirmed by the lung images. The airway diameters gradually decreased with time, until flow limitation occurred. In airways distal to the collapse, there was not a significant decrease in diameter. Based on these data, we conclude that flow limitation in liquid-filled lungs occurs in the trachea and main bronchi and its location is independent of EILV or Q.

Effect of Viscosity on Instilled Perfluorocarbon Distribution in Rabbit Lungs

The effect of viscosity on the distribution of perfluorocarbon instilled into the lungs for liquid ventilation was investigated. Perfluorocarbon (either perfluorodecalin or FC-3283) was instilled into the trachea during ventilation at a constant infusion rate of 40 ml/min and radiographic images were obtained at 30 frames/s. Image analysis was performed and the homogeneity index of the distribution was computed for images at the end of inspiration of each breath to evaluate the evolution of perfluorocarbon distribution during filling. The higher viscosity perfluorocarbon (perfluorodecalin) resulted in a more homogeneous distribution. This was attributed to perfluorodecalins higher propensity to form liquid plugs in large airways and to those plugs leaving behind a thicker liquid layer as they propagated through the lungs.

Effects of respiratory rate and tidal volume on gas exchange in total liquid ventilation.

Using a rabbit model of total liquid ventilation (TLV), and in a corresponding theoretical model, we compared nine tidal volume-respiratory rate combinations to identify a ventilator strategy to maximize gas exchange, while avoiding choked flow, during TLV. Nine different ventilation strategies were tested in each animal (n = 12): low [LR = 2.5 breath/min (bpm)], medium (MR = 5 bpm), or high (HR = 7.5 bpm) respiratory rates were combined with a low (LV = 10 ml/kg), medium (MV = 15 ml/kg), or high (HV = 20 ml/kg) tidal volumes. Blood gases and partial pressures, perfluorocarbon gas content, and airway pressures were measured for each combination. Choked flow occurred in all high respiratory rate-high volume animals, 71% of high respiratory rate-medium volume (HRMV) animals, and 50% of medium respiratory rate-high volume (MRHV) animals but in no other combinations. Medium respiratory rate-medium volume (MRMV) resulted in the highest gas exchange of the combinations that did not induce choke. The HRMV and MRHV animals that did not choke had similar or higher gas exchange than MRMV. The theory predicted this behavior, along with spatial and temporal variations in alveolar gas partial pressures. Of the combinations that did not induce choked flow, MRMV provided the highest gas exchange. Alveolar gas transport is diffusion dominated and rapid during gas ventilation but is convection dominated and slow during TLV. Consequently, the usual alveolar gas equation is not applicable for TLV.

Expiratory flow limitation during gravitational drainage of perfluorocarbons from liquid-filled lungs.

Flow limitation during pressure-driven expiration in liquid-filled lungs was examined in intact, euthanized New Zealand white rabbits. The aim of this study was to further characterize expiratory flow limitation during gravitational drainage of perfluorocarbon liquids from the lungs, and to study the effect of perfluorocarbon type and negative mouth pressure on this phenomenon. Four different perfluorocarbons (PP4, perfluorodecalin, perfluoro-octyl-bromide, and FC-77) were used to examine the effects of density and kinematic viscosity on volume recovered and maximum expiratory flow. It was demonstrated that flow limitation occurs during gravitational drainage when the airway pressure is ≤ −15 cm H2O, and that this critical value of pressure did not depend on mouth pressure or perfluorocarbon type. The perfluorocarbon properties affect the volume recovered, maximum expiratory flow, and the time to drain, with the most viscous perfluorocarbon (perfluorodecalin) taking the longest time to drain and resulting in lowest maximum expiratory flow. Perfluoro-octyl-bromide resulted in the highest recovered volume. The findings of this study are relevant to the selection of perfluorocarbons to reduce the occurrence of flow limitation and provide adequate minute ventilation during total liquid ventilation.

THEORETICAL STUDY OF GAS EXCHANGE IN TOTAL LIQUID VENTILATION

Adult respiratory distress syndrome and neonatal respiratory distress syndrome are characterized by less compliant lungs, a result of insufficient surfactant production or effectiveness. Total liquid ventilation (TLV) is an artificial ventilation system which uses perfluorocarbon (PFC) liquid to eliminate the air-liquid interface. PFC has high solubilities of oxygen and carbon-dioxide. In TLV, the convection is important for gas transport in the alveolar region because the diffusivities of gases in PFC are four orders of magnitude lower than in air. In this study, a computational model for gas exchange in the lung was developed: a conducting airways branching network was modelled as a trumpet-shaped tube; the terminal alveolar sac was modelled as an oscillating spherical shell; a tissue and capillary blood were modelled as well-mixed compartments; and, mixed-venous gas partial pressures were calculated assuming a constant oxygen consumption and carbon dioxide production. Since the convection dominates the transport in the sac as well as in the conducting airways, steep partial pressure gradients in the sac exist from the middle of inspiration to the beginning of expiration. Our computational results for the arterial gas partial pressures agreed well with the experimental results of TLV for rabbits. This work is supported by NIH grant HL64373. VAPORIZED PERFLUOROHEXANE VS. PARTIAL LIQUID VENTILATION IN EXPERIMENTAL LUNG INJURY