Prevention of ventilator-induced lung injury with partial liquid ventilation

Prevention of Ventilator-Induced Lung Injury With Partial Liquid Ventilation By Dorothy A. Lewis, Danny Colton, Kent Johnson, and Ronald B. Hirschl Ann Arbor, Michigan

Background/Purpose: Pulmonary injury from mechanical ventilation has been attributed to application of excess alveolar pressure (barotrauma) or volume (volutrauma). The authors questioned whether partial liquid ventilation (gas ventilation of the perfluorocarbon filled lung, PLV) would reduce ventilator-induced lung injury. Methods: A tracheostomy tube and carotid artery catheter were placed in anesthetized Sprague-Dawley rats (500 ⫾ 50 g). Bovine serum albumin (BSA) labeled with Iodine (I) 125 was administered intraarterially. Ventilation with tidal volume (TV) of 5 mL/kg was initiated. The rats were then selected randomly to a 30-minute experimental period of one of the following ventilation protocols: continued atraumatic gas ventilation (GV, TV, 5 mL/kg; n ⫽ 10); atraumatic gas ventilation combined with intratracheal administration of 10 mL/kg perfluorocarbon (GV-PLV, TV, 5 mL/kg, n ⫽ 10); barotrauma (BT, peak inspiratory pressure [PIP], 45 cm H2O; n ⫽ 10); barotrauma with PLV (BT-PLV, PIP, 45 cm H2O; n ⫽ 8); volutrauma (VT, TV, 30 mL/kg; n ⫽ 8); or volutrauma with

E

ARLY STUDIES evaluating the safety and efficacy of total liquid ventilation (TLV, perfluorocarbon ventilation of the perfluorocarbon-filled lungs utilizing a “liquid ventilator”) and partial liquid ventilation (PLV, gas ventilation of the perfluorocarbon-filled lung with a conventional mechanical ventilator) noted that a pulmonary protective effect was present when PLV or TLV were applied in lung injury models.1-3 This phenomenon has been reproduced in a variety of in vivo and in vitro models, including in the setting of cobra venom factor lung injury in the rat.4,5 During these studies, a reduction in lung neutrophil accumulation was noted during partial liquid ventilation when compared with gas ventilation in uninjured, control animals. We questioned whether this protective effect observed in normal lungs might be related to prevention of ventilator-induced lung injury by the presence of perflubron (LiquiVent, Alliance Pharmaceutical Corp, CA) in the lungs during PLV. To investigate this potential protective effect we applied a rat model of high-pressure ventilation developed by Dreyfuss et al.6 We explored in this study situations in which animals were gas and liquid ventilated in a volumelimited or pressure-limited mode such that injury was induced by either volutrauma or barotrauma, respectively. This study, then, addresses the question of whether PLV with perflubron provides a protective lung

PLV (VT-PLV, TV, 30 mL/kg; n ⫽ 10). Animals were killed and the amount of radiolabeled BSA in both lungs was measured and normalized to the counts in 1 mL of blood from that animal (injury index). Data were analyzed by analysis of variance (ANOVA) with post-hoc t test comparison between groups.

Results: There was a significant difference in the 125I-BSA injury index when all groups were compared (P ⬍ .001 by ANOVA). Post-hoc analysis showed a significant decrease in the injury index when comparing BT versus BT-PLV (P ⫽ .024) and VT versus VT-PLV (P ⫽ .014). Conclusion: 125I-BSA leak produced during high-pressure or high-volume mechanical ventilation is reduced by partial liquid ventilation. J Pediatr Surg 36:1333-1336. Copyright © 2001 by W.B. Saunders Company. INDEX WORDS: Ventilator-induced lung injury, barotrauma, partial liquid ventilation.

effect in the setting of both barotrauma and volutrauma in the uninjured rat lung. MATERIALS AND METHODS

Experimental Methods Sprague-Dawley pathogen-free male rats (Harlan Industries Inc, Indianapolis, IN) 500 ⫾ 50 g in weight were injected with an intramuscular mixture of xylazine (15 mg/kg; Lloyd Laboratories, Shenandoah, IA) and ketamine (100 mg/kg; Fort Dodge Laboratories, Fort Dodge, IA) for anesthesia. Neck dissection was performed followed by insertion of a 14-gauge tracheostomy tube. Volume-controlled ventilation with a standard gas mechanical ventilator (model 683, Harvard Rodent Ventilator, Harvard Apparatus, Inc, South Natick, MA) was initiated at the following settings: tidal volume, 5 mL; respiratory rate, 60 breaths per minute; FIO2, 1.0; positive end-expiratory pressure PEEP), 0 cm H2O. Animals were paralyzed with a single intramuscular injection of succinylcholine chloride (8 mg/kg; Abbott Laboratories, North Chicago, IL). Carotid artery access was performed using poly-

From the Department of Surgery and Pathology at the University of Michigan Medical Center, Ann Arbor, MI. Presented at the American College of Surgeons Surgical Forum, October 7, 1996. Address reprint requests to Ronald B. Hirschl, MD, Pediatric Surgery, University of Michigan, F3970 Mott, Box 0245, 1500 E Medical Center Dr, Ann Arbor, MI 48109-0245. Copyright © 2001 by W.B. Saunders Company 0022-3468/01/3609-0003$35.00/0 doi:10.1053/jpsu.2001.26361

Journal of Pediatric Surgery, Vol 36, No 9 (September), 2001: pp 1333-1336

1333

1334

LEWIS ET AL

ethylene tubing (PE-50, Clay Adams, Parsippany, NJ). An intravenous bolus of saline, 10 mL/kg, was administered after venous cannulation. The animals then were assigned randomly to 1 of 5 groups: (1) control (n ⫽ 10)—animals were gas ventilated at ventilator settings as noted above, which included a relatively low tidal volume of 5 mL7; (2) barotrauma (BT, n ⫽ 10)—animals were gas ventilated for 45 minutes at peak inspiratory pressure (PIP) of 45 cm H2O; (3) barotrauma and partial liquid ventilation (BT-PLV, n ⫽ 8)—partial liquid ventilation was maintained at a PIP of 45 cm H2O for 45 minutes; (4) volutrauma (VT, n ⫽ 8)—animals were ventilated at a tidal volume of 30 mL/kg for 45 minutes; (5) volutrauma and partial liquid ventilation (VT-PLV, n ⫽ 10)—animals underwent partial liquid ventilation at a tidal volume of 30 mL/kg for 45 minutes. All animals received an intravenous injection of iodine 125—labeled bovine serum albumin (125I-BSA) at the time of initiation of the 45-minute experimental period. After the 45-minute experimental period, the animals were killed by exsanguination of the abdominal inferior vena cava.

Procedure for Determining Albumin Leak Bovine serum albumin tagged with 125I (Reproductive Sciences Core, University of Michigan, Ann Arbor, MI) was diluted in 1 mL of normal saline, aliquot to a count of 800 K, and stored at room temperature. This aliquot of 125I-BSA was injected intravenously followed at the end of the experiment by assessment of counts in a sample of blood and both right and left lungs by gamma scintillation. An albumin leak index was determined by the following formula: Total CPM of 125I ⫺ (right ⫹ left lung) Total CPM of 125I in 1 mL blood

Partial Liquid Ventilation For the animals that underwent liquid ventilation the ventilator was temporarily disconnected and 5 mL (approximately 10 mL/kg) perflubron was instilled via the tracheostomy. Gas ventilation then was reinstituted.

Physiologic Assessment Airway pressures were determined at baseline and at the end of the 45-minute experimental period by a Sechrist airway pressure monitor (Model 400, Sechrist Industries, Inc, Anaheim, CA) in the ventilating circuit at a point just proximal to the endotracheal tube. Tidal volume was measured as the stroke volume of the Harvard ventilator; no correction for tubing compliance was applied. However, the same ventilator tubing and system was used in all studies.

Statistical Analysis ANOVA followed by post hoc analysis was applied using Dunnett’s t test to compare to the control group with significance at P less than .05. An independent t test was used to compare the BT or VT groups to the BT-PLV or VT-PLV groups, respectively. A Bonferroni-corrected P value of .025 was used to identify statistical significance for these latter comparisons because the number of t tests was 2. All values are reported as the mean ⫾ SEM. This study was approved by the University of Michigan Animal Care and Use Committee. National Institutes of Health guidelines for animal care and use were followed for all experiments.

RESULTS

Figure 1 illustrates the injury indices by group. There was a significant difference in the 125I-BSA injury index when all groups were compared (P ⬍ .001 by analysis of

Fig 1. 125I-BSA injury index in the control, BT, BT-PLV, VT, and VT-PLV animal groups. (*P < .05 by Dunnett’s t test when compared with GV; **P < .025 by independent t test when BT or VT are compared with BT-PLV or VT-PLV, respectively).

variance [ANOVA]). Post-hoc analysis showed a significant decrease in the injury index when comparing BT versus BT-PLV (P ⫽ .024) and VT versus VT-PLV (P ⫽ .014). When comparing the BT and BT-PLV groups, there was a significant difference in the tidal volume required to achieve a peak inspiratory pressure (PIP) of 45 cm H2O (mean TV, BT ⫽ 33.1 ⫾ 4.5 mL/kg, BT-PLV ⫽ 38.8 ⫾ 3.8 mL/kg; P ⫽ .014). A significantly lower PIP was observed in the VT-PLV group when compared with the VT group at a tidal volume of 30 mL/kg (mean PIP, VT ⫽ 37.1 ⫾ 5.8 cm H2O, VTPLV ⫽ 28.7 ⫾ 1.6 cm H2O; P ⫽ .003). DISCUSSION

The important findings from this study are that albumin leak is reduced during partial liquid ventilation with perflubron when compared with gas ventilation when elevated airway pressures or tidal volumes are applied in an uninjured rat model. A major conceptual shift has occurred over the last 5 to 10 years in the approach to mechanical ventilation of the patient with respiratory insufficiency since Kolobow et al8 showed that the acute respiratory distress syndrome (ARDS) could be induced in normal sheep by application of high pressure (50 cm H2O) mechanical ventilation over a 24-hour period. Subsequent studies by Dreyfuss et al6 showed, in normal rats, that albumin leak was increased after exposure to high-pressure ventilation for 30 minutes. Application of PEEP of 5 cm H2O appeared to be protective. Cilley et al,7 similarly observed an increase in wet-to-dry lung weight when rats were exposed to excess tidal volumes of 20 mL/kg when compared with those rats ventilated with reasonable volumes of 5 mL/kg. Additional studies by Parker et al9 measuring pulmonary capillary leak coefficients (Kf,c), documented an increase in the Kf,c in the lungs of normal animals that had been exposed to high-pressure mechanical ventilation (⬎ 55 cm H2O) for only 20 minutes.

PLV IN VENTILATOR-INDUCED LUNG INJURY

These laboratory studies have culminated in the concept of ventilator-induced lung injury, which is now accepted as a valid phenomenon by many critical care practitioners. Whether it is in the preterm neonate in whom bronchopulmonary dysplasia (BPD) develops or the adult who overcomes an acute respiratory disease process only to succumb to the development of fibrosis, a portion of the lung injury and inflammatory response is thought to be secondary to the application of highpressure ventilation for prolonged periods. Avoidance of high-pressure ventilation and “volutrauma” has been a tenet in the use of extracorporeal life support (ECLS) in the setting of respiratory insufficiency for many years.10 In addition, Hickling11 has been a proponent of the concept of “permissive hypercapnia” and has shown that elevated PaCO2 levels as high as 100 mmHg may be tolerated in the patient who is free of hemodynamic compromise and who has a pH greater than 7.10 to 7.20. The mortality rate among a series of 43 adult patients with ARDS who had an estimated mortality rate of 43% by APACHE was only 16% in their series of patients treated with permissive hypercapnia. Most recently, a randomized study treating a series of adult ARDS patients with standard high-pressure mechanical ventilation or low tidal volume ventilation with recruitment strategies showed a marked increase in pulmonary compliance, gas exchange, and survival rate in the lung protective strategy group.12 A decrease in 28-day mortality rate from 71% in the conventional to 38% in the low pressure lung protective strategy group was noted. As mentioned previously, Dreyfuss et al6 observed that lung injury in rats exposed to high tidal volume ventilation appeared to be reduced when PEEP was applied. An additional report suggests that outcome may be enhanced when high levels of PEEP are applied in adults with ARDS.13 Specifically, it has been suggested that application of an end-expiratory pressure equivalent to or greater than the pressure at the inflection point on the pressure-volume curve may maintain recruitable alveoli open rather than allowing them to collapse with each ventilatory cycle; it is thought that this inflationcollapse cycle is a source of lung injury in these marginally functional, recruitable lung regions.14 The goal of PEEP application may be then to maintain these at-risk alveoli open while minimizing the peak inspiratory pressure and, therefore, ventilator-induced lung injury. The data from this study suggest that albumin leak is reduced during high-pressure and high-volume ventilation with partial liquid ventilation when compared with gas ventilation. We have shown previously that albumin leak and neutrophil infiltration are reduced during partial liquid when compared with gas ventilation in a cobra venom factor rat lung injury model.4,5 In addition, these studies showed that neutrophil accumulation was re-

1335

duced during partial liquid ventilation when compared with gas ventilation in the noninjured lung. However, the mechanisms behind these observations and the protective effects of partial liquid ventilation observed in the current study remain unclear. Previous studies have suggested that neutrophil and macrophage superoxide and fluorescent product formation are reduced during exposure to perfluorocarbon fluids.15,16 In addition, systemic tumor necrosis factor (TNF)-␣, is reduced during partial liquid ventilation of HCl lung-injured animals.17 We have shown previously that neutrophil adhesion and cell killing are reduced only during, but not after, exposure to perflubron.18 These data would suggest that the protective pulmonary effects observed may be caused by an effect of perflubron upon leukocyte function or a mechanical effect that reduces leukocyte adhesion, activation, or amplification of the inflammatory process. Certainly, other mechanisms resulting in this phenomenon must be considered such as the lavage of exudate and inflammatory mediators from alveoli as the perflubron moves in and out of the central airways. As mentioned previously, alveolar recruitment may decrease the resultant lung injury associated with high-pressure mechanical ventilation.6 We have shown previously that expiratory lung volume, and, therefore, alveolar recruitment is enhanced during partial liquid when compared with gas ventilation in a sheep model of oleic acid lung injury.19 In other studies we have shown that the reduction in neutrophil accumulation associated with partial liquid ventilation is similar to that observed during application of PEEP in a cobra venom factor lung injury rat model.4 It may be, therefore, that the protective effects observed during partial liquid ventilation are mainly secondary to alveolar recruitment by mechanisms similar to those postulated for the protective effects of PEEP in the setting of high-pressure ventilation. Evaluation of the effects of PEEP in a similar model and comparison with the level of protection associated with partial liquid ventilation in the setting of high pressure and high volume ventilation might shed further light on this question. It is intriguing to consider that partial liquid ventilation could, as one of its benefits, provide a protective effect during mechanical ventilation. Certainly, this effect will need to be explored in larger animal models of respiratory insufficiency. Randomized clinical trials, which are now underway, should allow determination of whether a protective effect in both the setting of lung injury and performance of mechanical ventilation will result in a decrease in pulmonary morbidity and an enhancement in survival rate. In the meantime, these data suggest that administration of intratracheal perflubron is associated with a reduction in albumin leak in normal animals exposed to high levels of pressure and volume during mechanical ventilation.

1336

LEWIS ET AL

REFERENCES 1. 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 2. Hirschl RB, Tooley R, Parent A, et al: Improvement of gas exchange, pulmonary function, and lung injury with partial liquid ventilation. A study model in a setting of severe respiratory failure. Chest, 108:500-508, 1995 3. Quintel M, Heine M, Hirschl RB, et al: Effects of partial liquid ventilation (PLV) on lung injury in a model of acute respiratory failure a histologic and morphometric analysis. Crit Care Med 26:833-843, 1998 4. Colton DM, Hirschl RB, Till GO, et al: Neutrophil accumulation is reduced during partial liquid ventilation. Crit Care Med 26:17161724, 1998 5. Colton DM, Till GO, Johnson KJ, et al: Partial liquid ventilation decreases albumin leak in the setting of acute lung injury. J Crit Care 13:136-139, 1998 6. Dreyfuss D, Basset G, Soler P, et al: Intermittent positivepressure hyperventilation with high inflation pressures produces pulmonary microvascular injury in rats. Am Rev Respir Dis 132:880-884, 1985 7. Cilley RE, Wang JY, Coran AG: Lung injury produced by moderate lung overinflation in rats. J Pediatr Surg 28:488-495, 1993 8. Kolobow T, Moretti MP, Fumagalli R, et al: Severe impairment in lung function induced by high peak airway pressure during mechanical ventilation. Am Rev Respir Dis 135:312-315, 1987 9. Parker JC, Townsley MI, Rippe B, et al: Increased microvascular permeability in dog lungs due to high peak airway pressures. J Appl Physiol 57:1809-1816, 1984

10. Bartlett R: Extracorporeal life support for cardiopulmonary failure. Curr Probl Surg 27:621-705, 1990 11. Hickling KG: Low volume ventilation with permissive hypercapnea in the adult respiratory distress syndrome. Clin Int Care 3:6778, 1992 12. Amato MB, Barbas CS, Mederios DM, et al: Effect of a protective-ventilation strategy on mortality in the acute respiratory distress syndrome. N Engl J Med 338:347-354, 1998 13. DiRusso SM, Nelson LD, Safcsak K, et al: Survival in patients with severe adult respiratory distress syndrome treated with high-level positive end-expiratory pressure. Crit Care Med 23:1485-1496, 1995 14. Mancebo J: PEEP, ARDS, and alveolar recruitment. Int Care Med 18:383-385, 1992 15. Virmani R, Fink LM, Gunter K, et al: Effect of perflurochemical blood substitutes on human neutrophil function. Transfusion 24:343347, 1984 16. Smith T, Steinhorn D, Marcucci K, et al: Perflubron (PFB) decreases free radical (FR) production by alveolar macrophages (AM) in vitro. Crit Care Med 22:A196, 1994 17. Kawamae K, Pristine G, Chiumello D, et al: Partial liquid ventilation decreases serum tumor necrosis factor-␣ concentrations in a rat acid aspiration lung injury model. Crit Care Med 28:479-783, 2000 18. Varani J, Fligiel SEG, Till GO, et al: Pulmonary endothelial cell killing by human neutrophils, possible involvement of hydroxyl radical. Lab Invest 53:656-663, 1985 19. Gauger PG, Overbeck MC, Chambers SD, et al: Partial liquid ventilation improves gas exchange and increases end-expiratory lung volume in acute lung injury. J Appl Physiol 84:1566-1572, 1998