Pohlmann, Joshua R.*; Brant, David O.*; Daul, Morgan A.*; Reoma, Junewai L.*; Kim, Anne C.*; Osterholzer, Kathryn R.†; Johnson, Kent J.‡; Bartlett, Robert H.*; Cook, Keith E. PhD*; Hirschl, Ronald B.*
Despite improvements in management, the mortality of acute respiratory distress syndrome (ARDS) is 30%–40%.1–3 Clinical strategies that hasten the resolution of lung injury may improve both survival and long-term outcomes for patients with ARDS.4 Ventilation with liquid perfluorocarbon (PFC) has been investigated as a potential treatment for many respiratory diseases. Perfluorocarbon has high respiratory gas solubilities, low surface tension, and cytoprotective properties that allow for improved gas exchange at lower ventilation pressures while protecting the lung from inflammation. This may promote the resolution of lung injury and improve long-term outcomes. Partial liquid ventilation (partial filling followed by gas ventilation) supported this hypothesis in short-term (<4 hours) laboratory experiments and clinical trials. Despite positive results, studies indicate that total liquid ventilation (TLV) may be a better approach.
In TLV, the lungs are both filled and ventilated with PFC using a ventilator designed specifically for delivery of liquid tidal volumes. Early laboratory studies of TLV identified safe and effective ventilation strategies and described the basic physiology in animal models of neonatal pulmonary support.5 More recently, TLV has been evaluated in animals up to 25 kg with ARDS for as long as 4 hours.6–9 However, TLV is not yet sufficiently developed to be evaluated in clinical ARDS. Additional laboratory studies are needed to evaluate whether TLV can be successful in larger subjects with lung injury over a full course of treatment before progressing to clinical trials. We hypothesized that TLV could provide gas exchange superior to conventional mechanical ventilation (CMV) and improve survival in a longer-term model of severe respiratory failure. Therefore, the purpose of this study was to compare the performance of TLV and CMV in large animals with severe ARDS for 24 hours.
Materials and Methods
Animal Preparation and ARDS Model
The experimental protocol and procedures were approved by the committee for the use and care of animals at the University of Michigan in accordance with the National Institutes of Health guidelines for ethical animal research. Ten sheep [weight 53 ± 4 (SD) kg] were anesthetized with sodium thiopental (15–20 mg/kg IV, Hospira Inc., Lake Forest, IL), placed supine, intubated, and gas ventilated with 100% oxygen. Anesthesia was maintained with propofol (6–12 mg/kg/h, American Pharmaceuticals Partners, Schaumburg, IL) and ketamine hydrochloride (2–10 mg/kg/h, Hospira, Inc.) for the entire duration of ventilator support. An arterial line and pulmonary artery catheter (CCombo V777HF8, Edwards Lifesciences, Irvine, CA) were placed using aseptic technique. Paralysis was induced with 0.1 mg/kg IV pancuronium bromide (Hospira Inc.) and the following baseline data were taken: arterial blood gases, mean arterial pressure (MAP), central venous pressure (CVP), cardiac output (CO), mean pulmonary artery pressure (PAP), and venous saturation of oxygen (SvO2).
After collecting baseline data, lung injury was induced by injecting oleic acid (0.07 ml/kg/dose; C18H34O2; Fisher Scientific, Pittsburgh, PA) into the right atrium until PaO2/FiO2 ≤60 mm Hg using previously published methods.9 Continuous trending of SvO2 was used to indicate the onset and stabilization of lung injury followed by an arterial blood gas to quantify the severity. If PaO2 was >60 mm Hg, additional doses were given. Once the criterion was met, animals were transitioned to a protective CMV protocol (n = 5) or TLV (n = 5) and supported for 24 hours.
Conventional Mechanical Ventilation
In the CMV group, the ventilator (Nellcor 7200, Puritan Bennett, Boulder, CO) was set to deliver a tidal volume (VT) of 6 ml/kg, a frequency of 12 bpm, and FiO2 of 1.0. Positive end-expiratory pressure (PEEP) was set between 5 and 13 to optimize PaO2. Frequency and VT were subsequently adjusted to maintain PaCO2 between 35 and 50 mm Hg, while limiting frequency and peak inspiratory pressure (PIP) to 40 bpm and 35 cm H2O to minimize ventilator-induced lung injury. Ventilator settings over the course of the experiment are shown in Table 1.
Total Liquid Ventilation
To initiate TLV, the gas ventilator was disconnected, the lungs were rapidly filled with warm preoxygenated perfluorocarbon (FC77, Fluorinert, 3M Corporation, St. Paul, MN) to a volume of 30 ml/kg. Liquid ventilation was initiated using a frequency of 5 bpm and VT of 15 ml/kg, and then VT was adjusted between 15 and 20 ml/kg to maintain PaCO2 between 35 and 50 mm Hg. Gas exchange was optimized during TLV by maintaining an end-expiratory lung volume (EELV) that was low to maximize liquid exchange in the lungs, yet great enough to prevent airway collapse during expiration. Supplemental PFC volume (8–10 ml/min) was continuously infused into the ventilator circuit to overcompensate for evaporative loss of PFC. This caused EELV to slowly increase with time. Lung volumes were then manually adjusted every 1–2 hours by withdrawing volume from the system until the airway pressure curve indicated early signs of airway collapse. Airway collapse is characterized by a sudden decrease in endotracheal tube pressures with a decrease or no change in flow rate.10 Once identified, PFC volume was then returned until the airway pressure waveform indicated resolution of the airway collapse. Ventilator settings over the course of the experiment are shown in Table 1.
The study applied a modular liquid ventilator configuration constructed by MC3, Inc. (Ann Arbor, MI) using a combination of custom-built and off-the-shelf components. It was designed to be modular with two parallel circuits for providing tidal PFC flow to the subject and continuous PFC flow to the oxygenators to supply O2 to and remove CO2 from the PFC. One module is illustrated in Figure 1. The tidal flow portion of the circuit consists of a custom-built piston with a commercial linear actuator (Exlar SR21, Chanhassen, MN) for controlling piston movement. The linear actuator is controlled by a custom LabVIEW program (National Instruments, Austin, TX). The continuous flow PFC refreshing circuit includes a spiral-wound, silicone membrane oxygenator with integral heat exchanger (I-4500-2A, Medtronic, Inc., Minneapolis, MN), an adult bubble trap (Quest Medical, Allen, TX), and a roller pump (Cobe Perfusion System, Cobe BCT Inc., Lakewood, CO). In this study, four modules were used. PFC outflow from each piston converged into a single connection near the endotracheal tube (ETT) to provide adequate tidal volumes to support 60-kg subjects. Circuit temperature was maintained at 37°C.
Data Collection and Animal Management
The following parameters were recorded hourly: arterial blood gases, CO, MAP, CVP, PAP, and ventilator parameters RR, VT, FiO2, PEEP, and PIP. A maintenance infusion of 5% dextrose solution was given intravenously at 2.5–4.0 ml/kg/h. Gentamicin sulfate (120 mg IM, Abraxis, Schaumburg, IL) and nafcillin sodium (500 mg IV, Sandoz, Inc., Broomfield, CO) were given at 8- and 6-hour intervals, respectively. Phenylephrine (0–20 mg/kg/h IV, American Regent, Inc., Shirley, NY) was given as needed to treat hypotension unresponsive to volume administration, and supplemental bicarbonate (Hospira, Inc.) was given as needed.
After 24 hours, animals were euthanized with pentobarbital sodium (85 mg/kg IV, Fatal Plus, Vortech Pharmaceuticals, Dearborn, MI), and the heart and lungs were removed en bloc. Consistent regional samples were obtained from the dependent and nondependent regions of the apex, middle, and base of the right lung. Hematoxylin and eosin-stained slides were prepared and evaluated.
All data are presented as the mean and standard error of the mean (SEM) unless indicated otherwise. A t test was used to compare baseline and post-oleic acid injury data between the CMV and TLV groups. To determine differences between treatment modes (CMV or TLV), all longitudinal data was compared using a mixed-model analysis with repeated measures within SPSS (Chicago, IL). Finally, survival between CMV and TLV groups was compared using the Kaplan-Meier technique within SPSS.
Survival at 24 hours in the TLV and CMV groups were 100% and 40%, respectively. Three of five animals in the CMV group died (asystole) while on maximum ventilator support. These deaths all featured marked hypoxia and hypercapnia and occurred during the 9th hour (PaO2 = 38 mm Hg, PaCO2 = 116 mm Hg), the 12th hour (PaO2 = 41 mm Hg, PaCO2 = 216 mm Hg), and the 24th hour (PaO2 = 64 mm Hg, PaCO2 = 89 mm Hg). Survival was statistically longer in the TLV group than the CMV group (p < 0.05).
The average ventilation settings for the CMV and TLV groups are shown in Table 1. Gas exchange data for PaO2 and PaCO2 are shown in Figure 2. There was no significant difference between groups in PaO2 or PaCO2 at baseline (p = 0.12 and 0.29, respectively) or after oleic acid injury (p = 0.92 and 0.84, respectively). Over 24 hours of support, however, PaO2 was significantly higher (p < 10−9) and PaCO2 was significantly lower in the TLV group (p < 10−3).
The systemic hemodynamic variables of MAP and CO are shown in Figure 3. The average MAP decreased on lung injury and throughout the course of the study, whereas cardiac output increased slightly throughout the course of the study. There was no significant difference in MAP or CO between groups at baseline (p = 0.37 and 0.78, respectively). On injection of oleic acid, no significant difference in cardiac output was observed between groups (p = 0.45), but MAP was significantly lower in the TLV group (p < 0.05). During 24 hours of ventilator support, there was no significant difference in MAP or CO between groups (p = 0.94 and 0.99, respectively). Average phenylephrine doses are shown in Figure 4. There was no significant differences between the groups (p = 0.80).
Mean PAP and CVP data are shown in Figure 5. There was no significant difference between groups with respect to PAP at baseline (p = 0.68), on stabilization of lung injury (p = 0.15), and throughout the 24 hours of ventilatory support (p = 0.78). There was no significant difference in CVP between groups at baseline (p = 0.57). On injection of oleic acid, CVP approached a significantly lower value in the TLV group when compared with the CMV group (p = 0.056) and was significantly lower in the TLV group for the remainder of the study (p < 0.05).
Gross Pathology and Histology
The gross appearance of the lungs from the CMV and TLV groups are shown in Figure 6. The lungs from TLV-treated animals showed less evidence of injury than CMV-treated animals. In the CMV group, consolidation and edema were evident in the dependent lung regions as manifested by the dark red coloration in Figure 6. On cut section, these regions were relatively airless and filled with fluid. By comparison, the lungs of the TLV-treated animals showed much less consolidation and edema.
The histologic changes in the lungs of these animals are shown in Figure 7. As expected based on the gross appearance of the lungs, animals treated with TLV showed much less injury than the animals treated with CMV. The dependent lobes are shown as they represent the regions with the greatest degree of lung injury. The histologic appearance of the CMV-treated animals is shown in panel A. There is a prominent degree of leukocyte infiltration with edema, hemorrhage, and focal infarction (arrow). The inset demonstrates the focal hemorrhage (arrow). By comparison, the lung injury seen in the TLV-treated animals is shown in panel B. There was some inflammation present with neutrophils, but inflammation, edema, and hemorrhage were much less severe in the TLV group. The inset shows preservation of the lung architecture with some edema and congestion.
This study demonstrates superior gas exchange and survival using TLV in this animal model of severe ARDS. In the TLV group, arterial PO2 is significantly greater and PCO2 is significantly lower. This difference seems to be the main cause of the mortality difference between the groups. There are several explanations for improved gas exchange using TLV in the setting of respiratory failure. During the onset of ARDS, hemorrhagic edema fluid floods the airspace, presenting a barrier to oxygen delivery and creating regional lung consolidation, especially in the dependent regions of the lung where perfusion is greatest. Mechanical ventilation of these lungs causes the cyclic opening and closing of unstable alveoli and overdistension of functional alveoli, which can further induce lung injury.11,12 In this study, approximately 100–300 ml of hemorrhagic, edema fluid was collected in bubble traps and removed from the circuit during the initial hours of TLV. In contrast to CMV, which ventilates gas against the edema fluid, ventilation with immiscible PFC provides pulmonary lavage, replacing edema fluid with a dense, low-surface-tension liquid that can recruit or stabilize injured lung while supporting gas exchange. In the PFC-breathing lung, the air-liquid interface is completely removed, thus ventilation occurs at lower peak inspiratory pressures, and the density of PFC causes a redistribution of pulmonary blood flow to less injured, nondependent lung regions. The cumulative result is better V:Q matching compared with gas ventilation.
All sheep demonstrated decreasing arterial pressure over time that is unrelated to ventilation technique. After an initial drop in MAP associated with the oleic acid lung injury, MAP continued to decline despite increasing CO (see Figure 3). The cause of the decrease in MAP, therefore, is marked systemic vasodilation in both groups. This is likely an effect caused by the use of propofol anesthesia. Propofol has been shown to cause MAP decreases of 13–17 mm Hg when used over 10- to 30-minute periods of sedation or anesthesia.13,14 The vasodilation was likely worse in this experiment due to the longer duration of the protocol. For optimization of long-term TLV and improvements in future comparative studies, a better method of sedation must be used. In this study, a surgical plane of anesthesia was used for the entire experiment. It may be possible to use a lower dose of propofol and achieve more stable arterial pressures, but other means of sedation such as a ketamine and propofol with diazepam should be explored. To compensate for the arterial hypotension, phenylephrine and IV fluids were given. The IV fluids lead to a progressive increase in CVP over the course of the experiment, but neither were sufficient to maintain MAP.
The average mean PAP rose significantly in all 10 animals during lung injury with no effect on cardiac output. Mean PAP was slightly higher after injury in the TLV group, perhaps related to the higher average dose of oleic acid needed to meet the injury criteria in this group. However, the transition to either ventilation strategy did not further affect the PAP in this study.
Limitations of TLV and CMV protocols
The CMV protocol used in this study was unable to provide adequate respiratory support in this model of severe lung injury model. Based on ARDSnet recommendations, a high frequency, low tidal volume, peak pressure, and rate-limited strategy was used to minimize ventilator-induced lung injury. Adjustments in PEEP were allowed, but increases in PEEP were not associated with increased oxygenation, perhaps because no recruitment maneuvers preceded the increase. Furthermore, because a pressure-limited ventilation mode was used in the face of severely impaired lung compliance, any increase to PEEP caused a reduction in the tidal volume achieved with each breath. Therefore, because all animals required maximum respiratory rate, increases in PEEP caused decreased minute ventilation and thus accelerated hypercapnia. This problem may be attributed to severely impaired lung compliance associated with a severe lung injury.
Likewise, the performance of TLV was not optimal in this study. The PFC used (FC77) is a more economical alternative to perflubron with a much higher vapor pressure, which requires more careful monitoring of lung volumes. Second, TLV was performed with a respiratory rate of only 5/min, and expiratory flows were simple square waves. Unlike CMV, expiration is not passive during TLV. The ventilator drives the rate of expiration and must withdraw liquid slow enough to not exceed a critical negative airway pressure (approximately −15 cm H2O)10,15,16 at which large airways begin to close and expiration becomes restricted. Despite being noninjurious,17 more complex, “sculpted” expiratory flow patterns have been developed to optimize expiration and increase minute ventilation. These advances were not applied or needed in this study. Finally, lung volume measurements are traditionally derived indirectly by measuring body weight as it varies cyclically during TLV or by maintaining a desired end-expiratory pause pressure. Techniques exist to monitor lung volumes and the rate PFC is lost from the ventilator during TLV, which may be practical in future clinical applications.18,19 No such system was used in this study, and lung volumes were adequately maintained as described. Despite these limitations, TLV was effective in providing gas exchange in the setting of a severe lung injury.
Total liquid ventilation demonstrated superior gas exchange and reduced mortality when compared with CMV in a sheep model of severe ARDS. Future studies will prolong the duration of TLV to provide support through the reparative phase of ARDS to allow survival and long-term outcomes studies.
Supported by the NIH grants R01HL64373 and 2R42HL064987 and The University of Michigan Biomedical Research Council.
1. Rubenfeld GD, Caldwell E, Peabody E, et al
: Incidence and outcomes of acute lung injury. N Engl J Med
353: 1685–1693, 2005.
2. Milberg JA, Davis DR, Steinberg KP, et al
: Improved survival of patients with acute respiratory distress syndrome (ARDS): 1983–1993. JAMA
273: 306–309, 1995.
3. The Acute Respiratory Distress Syndrome Network. Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. N Engl J Med
4. Heyland DK, Groll D, Caeser M: Survivors of acute respiratory distress syndrome: Relationship between pulmonary dysfunction and long-term health-related quality of life. Crit Care Med
33: 1549–1556, 2005.
5. Wolfson MR, Shaffer TH: Liquid ventilation during early development: Theory, physiologic processes and application. J Dev Physiol
13: 1–12, 1990.
6. Hirschl RB, Parent A, Tooley R, et al
: Lung management with perfluorocarbon liquid ventilation improves pulmonary function and gas exchange during extracorporeal membrane oxygenation (ECMO). Artif Cells Blood Substit Immobil Biotechnol
22: 1389–1396, 1994.
7. Hirschl RB, Parent A, Tooley R, et al
: Liquid ventilation improves pulmonary function, gas exchange, and lung injury in a model of respiratory failure. Ann Surg
221: 79–88, 1995.
8. Hirschl RB, Tooley R, Parent A, et al
: Evaluation of gas exchange, pulmonary compliance, and lung injury during total and partial liquid ventilation in the acute respiratory distress syndrome. Crit Care Med
24: 1001–1008, 1996.
9. Wolfson MR, Hirschl RB, Jackson JC, et al
: Multicenter comparative study of conventional mechanical gas ventilation to tidal liquid ventilation in oleic acid injured sheep. ASAIO J
54: 256–269, 2008.
10. Foley DS, Brah R, Bull JL, et al
: Total liquid ventilation: Dynamic airway pressure and the development of expiratory flow limitation. ASAIO J
50: 485–490, 2004.
11. Ware LB, Matthay MA: The acute respiratory distress syndrome. N Engl J Med
342: 1334–1349, 2000.
12. Slutsky AS, Tremblay LN: Multiple system organ failure: Is mechanical ventilation a contributing factor? Am J Respir Crit Care Med
157: 1721–1725, 1998.
13. Upton RN, Martinez AM, Grant C: Comparison of the sedative properties of CNS 7056, midazolam, and propofol in sheep. Br J Anaesth
103: 848–857, 2009.
14. Zheng D, Upton RN, Martinez AM: The contribution of the coronary concentrations of propofol to its cardiovascular effects in anesthetized sheep. Anesth Analg
96: 1589–1597, 2003.
15. Komori E, Tredici S, Bull JL, et al
: Expiratory flow limitation during gravitational drainage of perfluorocarbons from liquid-filled lungs. ASAIO J
51: 795–801, 2005.
16. Bull JL, Foley DS, Bagnoli P, et al
: Location of flow limitation in liquid-filled rabbit lungs. ASAIO J
51: 781–788, 2005.
17. Bagnoli P, Tredici S, Seetharamaiah R, et al
: Effect of repeated induced airway collapse during total liquid ventilation. ASAIO J
53: 549–555, 2007.
18. Libros R, Philips CM, Wolfson MR, Shaffer TH: A perfluorocarbon loss/restoration (L/R) system for tidal liquid ventilation. Biomed Instrum Technol
34: 351–360, 2000.
19. Heckman JL, Hoffman J, Shaffer TH, et al
: Software for real-time control of a tidal liquid ventilator. Biomed Instrum Technol
33: 260–267, 1999.