Treatment of Pulmonary Hypertension and Hypoxia Due to Oleic Acid Induced Lung Injury with Intratracheal Prostaglandin E¹during Partial Liquid Ventilation

The experiments were reviewed and approved by the Animal Care and Use Committee of Tokyo Medical and Dental University and were carried out according to the National Institutes of Health guidelines. Fifty-five mature Japanese white rabbits, weighing 3.1 +/- 0.1 kg, were anesthetized with 30 mg/kg ketamine and 0.3 mg/kg xylazine given intramuscularly. With the animals supine, a midline cervical incision was made after subcutaneous infiltration with 0.5% (wt/vol) lidocaine and a tracheostomy was established. The trachea was intubated with a 4-mm (inner diameter) endotracheal tube. A 22-gauge polyethylene catheter was secured inside the lumen of the tracheal tube so that its tip was positioned at the distal end of the endotracheal tube. Mechanical ventilation (tidal volume, 15 ml/kg; respiratory frequency, 30/min; inspiratory:expiratory ratio, 1:2; fractional concentration of oxygen in inspired gas, 1.0) was initiated (SN-480-6, Shinano Co., Tokyo, Japan). A 4-French double-lumen central venous catheter (CS-15402, Arrow International Inc., Reading, PA) was introduced through a jugular vein to measure central venous pressure and to infuse fluids and drugs. Hydroxyethylstarch (6% wt/vol) in lactated Ringer's solution (Hespander: Kyorin Pharmaceutical Co., Tokyo, Japan) was infused intravenously at a rate of 10 ml [middle dot] kg (^-1) [middle dot] h^-1throughout the study. Anesthesia was maintained by administering 2 mg [middle dot] kg^-1[middle dot] h^-1ketamine, 6 mg [middle dot] kg^-1[middle dot] h^-1propofol, and 0.05 mg [middle dot] kg^-1[middle dot] h^-1pancuronium through the central venous line. The incremental infusion rate of ketamine and propofol was given as necessary when inadequate anesthesia was observed. Inotropic support was not instituted in any of the animals. The left carotid artery was cannulated to measure the mean arterial pressure and to sample blood. A median sternotomy after local subcutaneous infiltration with 0.5% (wt/vol) lidocaine was performed without damaging the pleura, and catheters were inserted directly into the left atrium to measure the left atrial pressure, and into the pulmonary arterial trunk to measure the PAP and to sample blood. An electromagnetic flow probe (6-mm inner diameter; model MFV 1100; Nihon Kohden Co., Tokyo, Japan) was attached around the ascending aorta to measure the cardiac output. The flow probe was calibrated before use, and the expected error of the probe was within +/- 15% (published data from Nihon Kohden Co.). The arterial pressure, PAP, left atrial pressure, central venous pressure, cardiac output, and airway pressure were recorded simultaneously using a polygraph (142-8; San-ei Instrument Co., Tokyo, Japan), and blood gases were analyzed using a blood gas analyzer (1306A, Instrumentation Laboratory, Milan, Italy). Blood hemoglobin levels were measured using a blood cell analyzer for animals (MEK-6108, Nihon Kohden Co.). Each animal's chest was kept open and its body temperature was maintained at approximately 37 [degree sign]C throughout the study.

Approximately 60 min after the experimental preparation, baseline measurements, including hemodynamics, blood gas parameters, and pulmonary compliance were obtained after which 0.08 ml/kg oleic acid mixed with 5 ml heparinized blood was infused through the central venous catheter during a period of 20 min. Sixty minutes after completion of the oleic acid infusion, control measurements were taken, after which the animals were assigned to one of four groups. Control group animals (group 1) were ventilated at the aforementioned settings throughout the experiment. Group 2 received aerosolized PGE¹(Prostandin 500; Ono Pharmaceutical Co., Osaka, Japan) by an ultrasonic nebulizer (Soniclizer 305, Atom Co., Tokyo, Japan) that was connected to the inspiratory limb of the ventilator circuit. The diameter of aerosolized particle was 1 or 2 [micro sign]m (published data from Atom Co.). The PGE¹was diluted with normal saline (100 [micro sign]g/25 ml) and introduced into the bottom of the nebulizer chamber through a catheter, using a syringe pump (model 1235N; Atom Co.). The nebulizer chamber was filled with 5 [micro sign]g PGE¹beforehand, and PGE¹was administered at a dosage of 0.1 [micro sign]g [middle dot] kg^-1[middle dot] min^-1. The ventilator settings were not changed except that 5 cm H²O positive end-expiratory pressure was added after the initiation of aerosolized PGE¹administration. Groups 3 and 4 received 15 ml/kg perflubron (CF³(CF (²))⁶CF²Br, Nippon Mektron Ltd., Tokyo, Japan) intratracheally in combination with mechanical ventilation. The perflubron was instilled in three divided doses. First the animals were tilted to the left decubitus position and 5 ml/kg perflubron was administered into the trachea through the endotracheal tube, followed by administration of the same dose with the animals in the right lateral decubitus and then in the supine position. Positive end-expiratory pressure was not applied in groups 3 and 4. After the perflubron administration was complete, group 4 received a 5-[micro sign]g bolus followed by 0.1 [micro sign]g [middle dot] kg^-1[middle dot] min (^-1) PGE¹continuously instilled intratracheally through the catheter attached to the tip of the endotracheal tube. PGE¹was diluted with normal saline (100 [micro sign]g/5 ml).

Measurements were repeated every 30 min for 2 h after recording control data. Additional doses of perflubron were not administered during the 2 h of the study period after the initial dose. If animals died or experienced pneumothorax before the experiment was completed, they were excluded from the data analysis. The experiments were done until 10 complete sets of data were obtained for each group.

The physiologic shunt fraction during gas ventilation (Qs/Qt) was calculated using the arterial oxygen content (Ca^O2), mixed venous oxygen content (Cv^O2), and alveolar capillary oxygen content (Cc (^O)²) and the following Equation 1: where Cc^O2= 1.34 x Hb + 0.003 x Pa^O2, Ca^O2= 1.34 x Hb x Sa^O2+ 0.003 x PA^O2, Cv^O2= 1.34 x Hb x Sv^O2+ 0.003 x Pv (^O)², PA^O2= (barometric pressure - 47) x 1 - Pa^CO2. The vapor pressure of perflubron used in this experiment was 10.5 mmHg at 37 [degree sign]C, so when we calculated Qs/Qt during partial liquid ventilation, Pa^O2was calculated as (barometric pressure - 47 - 10.5) x 1.0 - Pa^CO2. Dynamic compliance was calculated using the formula tidal volume/end-inspiratory pressure - end-expiratory pressure.

The data are expressed as mean +/- SD. All statistical analyses on recorded data were performed using the statistical software package StatView (J 4.5, Abacus Concepts, Berkeley, CA). The intragroup comparisons of control data and data obtained at 30, 60, 90 and 120 min, and the intergroup comparisons at each time interval were performed using repeated-measures analysis of variance. When a significant difference was noted, post hoc analysis using Bonferroni's method was performed within and between groups. Overall statistical significance was assumed at P < 0.05.

Administration of oleic caused hypoxia and hypercapnia, decreased compliance, increased PAP, and decreased cardiac output in each group (Table 1). Fifteen of 55 animals were excluded from the study: 12 died of severe hypoxia or right ventricular failure (7 before control measurements, 3 in group 1, and 2 in group 2), and 3 developed pneumothorax (2 in group 1 and 1 in group 3). The control data after lung injury were not different among the groups.

(Table 1) shows hemodynamic changes. The mean PAP was not altered in groups 1 and 2 after injury. The mean PAP of group 3 decreased 30 min after partial liquid ventilation (P = 0.0012 vs. control data), but the difference was no longer significant after 60 min. The mean PAP of group 4 was reduced significantly after partial liquid ventilation with PGE¹(P < 0.0001 at 30 min and P = 0.0026 at 120 min vs. control data). The mean PAP values of group 4 after partial liquid ventilation and PGE¹were significantly less than the corresponding group 1 values (P = 0.0014 at 30 min and P = 0.0025 at 120 min). They were also less than those of group 2 30, 60, and 90 min after partial liquid ventilation and PGE¹(P = 0.004 at 30 min and P = 0.0063 at 90 min).

End-inspiratory pressure and compliance were not altered after lung injury was established in all groups (Table 2). The increase in Qs/Qt and decrease in Pa^O2were sustained throughout the study in groups 1 and 2. The Qs/Qt values decreased after initiation of partial liquid ventilation in groups 3 and 4 (P < 0.005 vs. control data) and they were less than those of groups 1 and 2 (P < 0.006). The Pa^O2of group 3 increased 30 min after partial liquid ventilation (P = 0.003 vs. control data), but the difference was no longer significant 60 min after partial liquid ventilation, whereas the increase in Pa^O2was sustained in group 4 (P < 0.0001 at 30 min, P = 0.0024 at 120 min vs. control data; Figure 1). Compared with groups 1 and 2, the Pa^O2values were significantly greater in group 3 at 30 and 60 min (p < 0.006) and also higher in group 4 throughout the experiment after partial liquid ventilation and PGE (¹) (P < 0.0001). The Pa^O2values of group 4 at 30, 60, and 120 min after partial liquid ventilation and PGE¹was higher than the corresponding value of group 3 (P = 0.005, P = 0.0033, and P = 0.0016, respectively).

Intratracheal administration of PGE¹combined with partial liquid ventilation improved gas exchange and pulmonary circulation without causing systemic hypotension or reducing cardiac output, whereas aerosolized PGE¹combined with conventional gas ventilation failed to improve oxygenation or reduce PAP. The results suggest that PGE¹delivery during liquid ventilation augments the improvement in oxygenation and decreased PAP seen with liquid ventilation alone. Treatment of severe acute respiratory distress syndrome should be directed to recruitment of atelectasis, improvement of oxygenation without compromising the systemic circulation, and reduction of pulmonary vascular resistance. Among the four treatments in the current study, intratracheal PGE¹combined with partial liquid ventilation achieved those purposes most successfully.

The administration doses of PGE¹in those two groups were not equivalent. The dose of PGE¹in group 4 was larger than the intravenous dose used clinically in acute respiratory distress syndrome, [2,3]whereas in group 2 it was probably much less because some aerosolized PGE¹was lost through expired gas. In addition, it is possible that our aerosol delivery system failed to deliver particles to the lung, especially in small animals. Because no rabbit developed hypotension or low cardiac output in response to intratracheal PGE¹during partial liquid ventilation, it did not seem that the dose of PGE¹in group 4 was extremely large. There may be some other possible mechanisms by which aerosolized PGE¹failed to improve oxygenation in our lung injury model. We assume that the distribution of PGE¹during partial liquid ventilation and gas ventilation was different. Aerosolized PGE¹may not be delivered to the alveoli because the positive end-expiratory pressure level might be inadequate to recruit collapsed alveoli in our study. Some of the aerosolized PGE¹seemed to dissolve in edema fluid, which subsequently increased the shunt fraction. Those two factors might offset any favorable effects of PGE¹on gas exchange and circulation in lung injury. On the other hand, intratracheally administered PGE¹during partial liquid ventilation could be delivered to alveolar units by conductive transport of perflubron. Partial liquid ventilation improves gas exchange by several mechanisms: pulmonary blood flow redistribution, recruitment of atelectatic alveoli, and displacement of alveolar edema fluid by perflubron. [11,13]A recent study using isolated piglet lungs showed that partial liquid ventilation abated oleic acid-induced elevation in pulmonary vascular resistance when given therapeutically after lung injury. [14]We also showed a transient reduction of PAP without decreasing cardiac output soon after partial liquid ventilation. Increased vasodilation of well-ventilated areas that received both oxygen-rich perflubron and PGE¹may account for improved oxygenation and decreased pulmonary vascular resistance in group 4.

Few reports describe the effect of aerosolized PGE¹on oxygenation and pulmonary circulation in acute respiratory distress syndrome. Keenan et al. [15]reported that inhaled PGE¹improved the Pa^O(^2/FI)^O2ratio from 78 to 107 in seven patients with severe acute respiratory distress syndrome. However, there were no significant changes in PAP or cardiac output. It seems that aerosolized drugs are more efficiently delivered to the distal airways in large animals or humans than they are in small animals. From the conflicting results of Keenan et al.'s and our current investigation, more studies are needed to determine the effects of inhalational PGE¹on gas exchange and pulmonary circulation in acute respiratory distress syndrome.

In the current study, we used an open-chest preparation, which obviously differs physiologically from a closed-chest preparation. First, in an open-chest model under mechanical ventilation, the formation of atelectasis may be more extensive than in the closed chest, especially without suitable positive end-expiratory pressure. Second, larger doses of perflubron (15 ml/kg) were needed in this open-chest model to observe the presence of a perflubron meniscus in the tracheal tube compared with previous studies using rabbits. [12,16]Tutuncu et al. [12]investigated the effects of lesser doses (3, 6, 9, and 12 ml/kg) of perflubron in rabbits with acute lung injury induced by alveolar lavage. They observed that the groups treated with 9 or 12 ml/kg perflubron showed a significantly higher Pa^O(²) level at the end of a 6-h study (without additional doses) compared with pretreatment values. On the other hand, the improvement of Pa^O2by partial liquid ventilation in our study was only transitory. However, the improvement of Pa^O2in the experiments of Tutuncu et al. [12]also seemed time dependent. According to a recent study by Mates et al. [17]using healthy piglets, estimation of evaporative losses of perflubron during partial liquid ventilation were calculated to be approximately 2 ml [middle dot] kg^-1[middle dot] h^-1(assuming a ventilatory rate of 20 breaths/min, a tidal volume of 15 ml/kg, and a dead space of 3 ml/kg). In the current study, the loss of perflubron was estimated to be about 9 ml/h, provided that the ventilatory rate was 30 breaths/min, the rabbit weighed 3 kg, and the expired gas was fully saturated with perflubron (10.5 mmHg at 37 [degree sign]C), and this might affect the gas exchange.

Recently, the combination of perflubron and nitric oxide for ventilating surfactant-depleted lungs [18]or oleic acid-injured lungs [19]was found to have an additive effect, improving gas exchange and reducing PAP. It was proposed that the amount of perflubron used for partial liquid ventilation could be reduced by concomitant administration of nitric oxide. Other authors have investigated the efficacy of pulmonary administration of drugs during liquid ventilation. [20]Wolfson et al. [20]showed the effects of pulmonary administration of vasoactive drugs during total liquid ventilation using preterm and neonatal lambs. Bolus injections of acetylcholine, epinephrine, and priscoline exerted dose-dependent effects on the systemic vasculature, but priscoline resulted in a significant decrease in the PAP relative to the systemic arterial pressure. They also investigated the pulmonary distribution of drugs administered through the airway during total liquid ventilation using intratracheal¹⁴C-dipalmitoylphosphatidylcholine, which was found to be distributed relatively homogeneously. From the results of their and our studies, perflubron would appear to be a good vehicle for drug delivery to the lung during both total and partial liquid ventilation.

We wonder whether the results of this study on rabbits can be extrapolated to other species or other circumstances. The perflubron dose required to improve gas exchange and oxygen delivery may be much larger than that for rabbits, especially for large animals. [18,21]Incremental dosing of perflubron may increase the PAP, and a few studies have shown this phenomenon. The PAP tended to increase in response to cumulative doses of perflubron in large-animal models of acute respiratory failure, although the increases were not significant. [18]Studies that evaluate the effects of intratracheal vasodilators during partial liquid ventilation on gas exchange and pulmonary circulation in large-animal models should be performed before this technique is used clinically.

In conclusion, intratracheal PGE¹combined with partial liquid ventilation not only improved Pa^O2and cardiac output but also reduced PAP in a rabbit model of acute respiratory distress syndrome. Intratracheal administration of PGE¹during partial liquid ventilation may offer an alternative treatment for acute respiratory distress syndrome with pulmonary hypertension.