INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The use of a lung protective ventilatory strategy (LPVS) during mechanical ventilation of patients with adult respiratory distress syndrome (ARDS) has been recommended by many investigators (1). Essentially, this strategy relies on the avoidance of tidal volumes (VT) that result in lung overdistension, and the setting of PEEP at a level sufficient to avoid derecruitment of unstable lung units with each ventilatory cycle (1). This strategy can promote adequate oxygenation and prevent ventilator-induced lung injury in animal models (5). In addition, Amato and colleagues (2) showed improved outcome in patients with ARDS during conventional mechanical ventilation when this strategy was compared with an approach using VT set at 12 ml/kg and PEEP set solely based on oxygenation response.

Partial liquid ventilation (PLV) has been shown to support gas exchange in several animal models of acute respiratory failure (9, 10). Recent studies comparing PLV and gas ventilation (GV) reveal improved gas exchange and hemodynamics in animals treated with PLV (11). In addition, a number of recent comparisons in animal models demonstrated anti- inflammatory effects of PLV resulting in greater lung protection (15). However, none of these studies used a LPVS during conventional mechanical ventilation (9). Low levels of PEEP (5 cm H₂O) and/or high peak alveolar pressures were used in most studies. Whether application of a LPVS would result in similar gas exchange during PLV and GV is unknown.

We therefore conducted this study to compare the effects of a LPVS during PLV and GV on gas exchange, hemodynamics, pulmonary mechanics, and lung histology in a sheep saline lung lavage model of ARDS. We hypothesized that no differences would exist between the two groups since a LPVS was applied to all animals.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The following protocol was approved by the Subcommittee on Research Animal Care of the Massachusetts General Hospital.

Preparation

Fifteen fasted Hampshire sheep weighing 28.1 ± 6.2 kg were anesthetized with halothane and orally intubated using a 9-mm inner diameter endotracheal tube (Hi-Lo Tube; Mallinckrodt Medical, Inc., St. Louis, MO). A 14-Fr orogastric tube (Mallinckrodt Laboratories Ltd., Athlone, Ireland) was inserted to ensure gastric drainage. After cannulating the left jugular vein, anesthesia was induced using fentanyl (0.3 mg) and sodium pentobarbital (200 mg). Anesthesia was then maintained using fentanyl (3 µg/kg/h), sodium pentobarbital (5 mg/ kg/h), and a continuous infusion of pancuronium bromide (2 mg/h) for paralysis. Mechanical ventilation was provided using a PB 7200ae ventilator (Nellcor-Puritan-Bennett, Carlsbad, CA) with the following ventilatory settings: volume-controlled ventilation, tidal volume (VT), 12 ml/kg; respiratory rate (RR), 15 breaths/min; inspiratory to expiratory (I:E) ratio, 1:1; inspiratory plateau time (TI_plat) 0.6 s, the fraction of inspired oxygen (FI_O2), 0.5; and PEEP, 5 cm H₂O. The right femoral artery was cannulated for arterial blood pressure monitoring and blood sampling. A pulmonary artery catheter (Edwards Swan Ganz, 7.0 Fr; Baxter Healthcare Corp., Irvine, CA) was inserted via the right jugular vein and connected to a cardiac output computer (9520A; American Edwards Laboratories, Santa Ana, CA). Lactated Ringer's solution was administered intravenously at a constant rate of 20 ml/kg/h.

Experimental Protocol

After the surgical procedure and a stabilization period of 30 min, baseline parameters (preinjury) were obtained. After baseline measurements lung injury was established with warm (39° C) saline lung lavage, repeated every 15 min until Pa_O2 fell below 80 mm Hg and remained stable with the baseline ventilator settings. Each lavage was preformed with a 30 ml/kg volume. The fluid was held about 36 inches above the animal during filling, with a 60-s procedure (fill, dwell, drain) time. The number of lavages per group is listed in Table 1. A stable lung injury was defined as a Pa_O2 change of less than 10% for more than 30 min. Another set of measurements was obtained along with the measurement of the pressure-volume (P-V) curve of the respiratory system after establishing lung injury (Injury), then animals were randomly assigned to be ventilated in the supine position using gas ventilation (GV group) or partial liquid ventilation (PLV group). Animals in the PLV group were filled over approximately 1 h with Perflubron (PFB) (LiquiVent; Alliance Pharmaceutical Corp., San Diego, CA) until a meniscus at the teeth was identified in the endotracheal tube during disconnection from the ventilator. Animals in the GV group were ventilated for another hour with the same ventilator settings to insure equal total time of baseline ventilation. After a stabilization period of 15 min, another set of measurements was performed (postfill) along with the measurement of the P-V curve of the respiratory system. Then ventilatory settings were changed in both groups to pressure control mode (PCV); RR, 20/min, and inspiratory time, 1.5 s. PEEP was adjusted to 1 cm H₂O above the lower corner pressure (Plc) on the inspiratory P-V curve. Peak inspiratory pressure (PIP) was titrated to establish normocapnia, but was limited to 35 cm H₂O or the upper corner pressure (Puc) on the inspiratory P-V curve. To replace the PFB evaporated, 1 ml/kg/h of PFB was added in the PLV group. Animals from both groups were then ventilated for 5 h, with hourly assessment of gas exchange, lung mechanics, and hemodymanics, after which animals were killed and their lungs were carefully excised for histologic examination.

Measurements

Hemodynamics, gas exchange, and pulmonary mechanics were measured after initial stabilization (preinjury), after lung injury (Injury), after filling with PFB (postfill), and hourly during the following 5-h ventilation period in both groups.

Hemodynamics

Mean arterial blood pressure, mean pulmonary artery (), pulmonary artery occlusion (Ppao), and central venous (PCV) pressures were monitored using pressure transducers (Argon; Maxxim Medical, Athens, TX) calibrated at 50 mm Hg, and zeroed at midthorax in the supine position. Cardiac output (CO) was measured by thermodilution (9520A computer; American Edwards Laboratories) using three 5-ml injections of saline.

Gas Exchange

Arterial and mixed venous blood samples were drawn at each measurement period and PO₂, PCO₂, pH, oxygen saturation, and hemoglobin content were assessed by a blood gas analyzer (Model 238; Ciba Corning Diagnostics Corp., Norwood, MA) and co-oximeter (Model 282; Instrumentation Laboratories, Lexington, MA). Shunt ratio (S/ T) was calculated using the equation: S/T = (Cc_O2  CA_O2)/(Cc_O2  Cv_O2), where Cc_O2 (oxygen content of alveolar capillary blood) was calculated assuming capillary oxygen tension to be equal to the alveolar oxygen tension. To calculate dead space/tidal volume ratio (VD/ VT), expired gas was collected for 3 min in a Douglas bag (Vacu-Med 25 L balloon; Vacumedics Inc., Ventura, CA). The mean expired CO₂ concentration was measured using a capnometer (Model 2200; Traverse Medical Monitors, Saline, MI), and VD/VT was calculated according to Enghoff's modification of the Bohr equation (18).

Lung Mechanics

Gas flow was monitored at the airway opening using a heated pneumotachometer (3700A; Hans-Rudolph Inc., Kansas City, MO) with differential pressure transducer (Model 45-14-871 ± 2; Validyne, Northridge, CA) calibrated with a precision rotometer. Airway pressure was monitored using a pressure transducer (Model 45-32-871 ± 100) calibrated with a water manometer. Tidal volume was integrated using the flow signal and reconfirmed with a 500 ml calibrated syringe. We also measured total PEEP with a 5-s end-expiratory hold maneuver using the auto-PEEP function of the 7200 ventilator. Quasi-static lung compliance (CstL) was calculated with the following equation: CstL = VT/(Pplat  PEEP_t). All signals were amplified (Model 8805C; Hewlett-Packard, Waltham, MA) and recorded at 100 Hz using an analog-to-digital conversion system (WINDAQ/200 v1.36; Dataq Instruments, Hartfield, PA) and a personal computer. All devices were properly calibrated at the beginning of the experiment.

P-V Curve

An inspiratory P-V curve of the respiratory system was obtained in all animals at postinjury and at postfill using a calibrated syringe. Stepwise, inflations were stopped when airway pressure exceeded 40 cm H₂O. The injured lungs in both the GV and the PLV groups exhibited a nonlinear P-V relationship. The pressures associated with changes in the slope of the P-V curve were identified as Plc and Puc. Each was determined from the crossing of tangents drawn to the varying slopes of the curves.

Histologic Analysis

Sample preparation. The excised lungs were inflated to a pressure of 23 cm H₂O with Trump's fixative solution (4% formaldehyde with 1% glutaraldehyde) and then immersed in fixative for 60 h. After fixation, lungs were separated at the hilus, and each lung was cut horizontally into 1-cm slices from dependent to nondependent regions. Two tissue blocks (1 cm wide × 1 cm long × 1 cm thick) were cut from each slice in a randomized manner. Each tissue block was embedded in paraffin wax and 5-mm sections were cut, mounted, and stained with hematoxylin-eosin.

Analysis. Quantitative examination by light microscopy was performed by one of the investigators blinded to the experimental protocol and the region of sampling. Each sample was examined under both low and high power fields. At least four sections were obtained from each block, and 20 fields were randomly selected and analyzed for each section. The severity of histologic lesions was assessed using a score (HIS) based on six parameters: intra-alveolar edema, hyaline membrane formation, hemorrhage, recruitment of granulocytes into the air spaces, focal alveolar collapse or consolidation, and epithelial desquamation/necrosis of airways or alveoli. Each parameter was evaluated semiquantitatively using the following scale: 0 = absent, 1 = mild, 2 = moderate, 3 = prominent. In addition, the percentage of the involved area of each histologic specimen was estimated (0 to 100%) to quantify the histologic changes. For each sample a weighted histologic score (WIS) was computed from the product of HIS and the percentage of area involved. HIS and WIS scores were calculated for each animal for dependent and non-dependent lung regions.

Statistical Analysis

Data are expressed as mean ± SD. Both PLV and GV groups were compared using a two-way analysis of variance (ANOVA) between Preinjury and Injury and from Injury to the end of the 5-h ventilation period. When statistical significance was reached, it was followed by a post-hoc analysis (Tukey honest significant difference [HSD] test). Analysis of histopathologic data was performed by unpaired t test with the Bonferroni correction for multiple comparison. Chi-square analysis was used to analyze the distribution of histologic score. A statistics software package (STATISTICA 5.1; StatSoft Inc., Tulsa, OK) was used, considering a p value less than 0.05 as statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Three sheep died from hypoxemia after creation of lung injury and, consequently, were excluded from further analysis. Twelve sheep were randomly assigned to GV and PLV groups and survived the study period. Body weight was 27.4 ± 5.6 kg in the GV group and 28.8 ± 7.2 kg in the PLV group (p = 0.71).

Lung Injury

Saline lung lavages resulted in a similar pattern of injury in both groups (Table 1). Specifically, Pa_O2 decreased (p < 0.05) (Figure 1), Pplat increased (p < 0.05) (Figure 2), CstL decreased (p < 0.05) (Table 2), VD/VT increased (p < 0.05) (Figure 3), increased (p < 0.05) (Table 3), and S/T increased (p < 0.05) (Table 2). The Plc in the GV group did not differ postfill from postinjury. However, the Plc in the PLV group decreased from 17.8 ± 2.4 cm H₂O postinjury to 11.2 ± 2.5 cm H₂O (p < 0.01) postfill. A representative P-V curve from the PLV group before (injury) and after (postfill) filling with PLV is shown in Figure 4. The Plc was clearly identified on the inspiratory P-V curve in all animals but one. Applied PEEP was set at 20 cm H₂O in this sheep (randomized to GV). No sheep showed a Puc below 35 cm H₂O.

View larger version (14K): [in this window] [in a new window]   Figure 1.   PaO2 (mean ± SD) in the gas ventilation (GV) and partial liquid ventilation (PLV) groups over the course of the study. There was no difference between groups: p < 0.05 versus *preinjury, injury, and  postfill.

View larger version (12K): [in this window] [in a new window]   Figure 2.   End-inspiratory plateau pressure (Pplat) (mean ± SD) in the GV and the PLV groups over the course of the study: p < 0.05 versus *preinjury and versus #GV.

View larger version (13K): [in this window] [in a new window]   Figure 3.   Dead space/tidal volume ratio (VD/VT) (mean ± SD) in the GV and PLV groups over the course of the study: p < 0.05 versus *preinjury, injury, postfill, and #GV.

View larger version (8K): [in this window] [in a new window]   View larger version (8K): [in this window] [in a new window]   Figure 4.   Representative P-V curves from the PLV group before (injury) (top) and after (postfill) (bottom) filling with PFB. PCL decreased with filling from 17 cm H2O to 11 cm H2O.

Gas Exchange

Applied PEEP was set at 18.9 ± 1.4 cm H₂O in the GV group and 11.5 ± 2.5 cm H₂O in the PLV group throughout the 5-h ventilation period. Pa_O2 increased, but not significantly, from 58 ± 2 to 94 ± 64 mm Hg in the PLV group after filling (Figure 1). After increasing PEEP above Plc, Pa_O2 increased significantly in both the GV and the PLV groups. However, Pa_O2 did not differ significantly between groups at any time during the experiment. Changes in Pa_CO2 are presented in Figure 5. Pa_CO2 markedly increased throughout the course of the experiment in the GV group (p < 0.05 versus injury, postfill, and PLV), thereby resulting in severe respiratory acidosis (Table 2). Conversely, Pa_CO2 remained in the normal range throughout the study period during PLV (Figure 5). Sv_O2 and S/T did not change significantly throughout the study period (Table 2).

View larger version (13K): [in this window] [in a new window]   Figure 5.   PaCO2 (mean ± SD) in the GV and PLV groups over the course of the study: p < 0.05 versus injury, postfill, and #GV.

Pulmonary Mechanics

Partial liquid ventilation improved the mechanical properties of the respiratory system. After postfill and during the 5-h ventilation period, Pplat was significantly lower in the PLV group than in the GV group (p < 0.05) (Figure 2). All animals in the GV group were ventilated at a Pplat of 35 cm H₂O, whereas only one of the animals in the PLV group required this maximum pressure. Application of a LPVS significantly improved CstL in both groups (Table 2). However, CstL tended to improve sooner (22.7 ± 5 versus 14.3 ± 2.7 at 1 h, p < 0.05) and to a greater extent in the PLV group than in the GV group. After increasing PEEP above the Plc, VT decreased significantly from 1 to 5 h ventilation in the GV group, but not in the PLV group (p < 0.05 versus injury, postfill, and PLV) (Figure 6). Pressure control level was not changed during the 5-h protocol. Finally, VD/VT increased in the GV group from 1 to 5 h ventilation (p < 0.05 versus injury, postfill, and PLV), but it did not change throughout the study period in the PLV group (Figure 3).

View larger version (14K): [in this window] [in a new window]   Figure 6.   Tidal volume (mean ± SD) in the GV and PLV groups over the course of the study: p < 0.05 versus injury, postfill, and #GV.

Hemodynamics

Although mean blood pressure did not change throughout the study in the PLV group, during GV it significantly decreased at 4 and 5 h ventilation (p < 0.05) (Table 3). After PEEP was set above the Plc, PCV, and increased significantly in the GV group, but not in the PLV group (Table 3). Both PCV and reached significantly higher values during the ventilation period in the GV group when compared with the PLV group (Table 3). Similarly, Ppao increased from 1 to 5 h ventilation in the GV group (p < 0.05 versus injury) and was higher than that in the PLV group from 2 to 4 h ventilation (p < 0.05) (Table 3). Systemic vascular resistance (SVR) decreased significantly from 2 to 5 h ventilation in the GV group (p < 0.05 versus injury and postfill) but was unchanged in the PLV group (Table 3). Cardiac output, heart rate, and pulmonary vascular resistance (PVR) did not change significantly throughout the study period in either group (Table 3).

Histologic Evaluation

Results of histologic analysis are summarized in Table 4. The architecture of nondependent regions of the lungs was preserved in both groups. The extent and the severity of lung injury in the dependent regions (WIS) was significantly greater in the GV group (5.6 ± 0.9 versus 3.9 ± 1.2, p < 0.05) (Table 4). Analysis revealed greater alveolar collapse and consolidation in the GV group than in the PLV group (2.4 ± 0.8 versus 0.9 ± 0.6, p < 0.05), despite higher PEEP levels in the GV group (Table 4). Dependent regions from both GV and PLV are illustrated in Figure 7.

View larger version (108K): [in this window] [in a new window]   View larger version (86K): [in this window] [in a new window]   Figure 7.   Light photomicrograph of the dependent regions of lungs from the gas ventilation group (A) and PLV group (B). Original print magnification: ×100.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The major findings of this study are as follows. (1) There was no significant difference in oxygenation between PLV and GV when PEEP was set 1 cm H₂O above the Plc on the inspiratory P-V curve. (2) PLV resulted in better ventilation than GV. (3) Histologic architecture of the lung was better preserved with PLV than with GV.

Gas Exchange

Several studies have reported improved oxygenation during PLV as compared with GV in large animal models of ARDS (10, 15, 19, 20). However, all of these studies used relatively low PEEP levels during GV. Experimental (5) as well as clinical (2) evidence support the use of PEEP set above the Plc to improve oxygenation and prevent ventilator-induced lung injury during GV. On the other hand, we previously reported that the Plc of the P-V curve shifts to the left (Figure 4) with PFB and that setting PEEP 1 cm H₂O above the Plc improves oxygenation during PLV as compared with 5 cm H₂O PEEP (26). This synergistic effect of PEEP on oxygenation during PLV results from improved gas distribution as PFB distributes to the lung periphery, thereby minimizing impedance to ventilation and stabilizing nondependent lung regions (26, 27). Because of its high density (1.92 g/ml), PFB also exerts a PEEP-like effect during PLV and enhances recruitment of the dependent lung.

Ferreyra and colleagues (28) recently illustrated the effect PEEP has on impedance to gas delivery during PLV. During pressure ventilation, low PEEP results in a markedly decreased peak flow and VT. Similarly, with volume ventilation, peak airway pressure is markedly increased and the establishment of peak flow delayed at low PEEP. Because low PEEP allows PFB to exist in large airways at end exhalation, impedance to gas delivery is increased. High PEEP moves the PFB out of nondistending airways, reducing impedance to ventilation.

As observed in Figure 4, PFB has a marked effect on the inspiratory P-V curve. Perflubron shifts the curve to the left, decreasing the Plc and increasing compliance (19, 26) of the gas-fluid filled lung. Because PFB recruits lung and increases compliance, tidal recruitment is completed at a lower pressure with PLV (lower peak ventilating pressure) than with GV. This is consistent with the data from Hickling (29) and Jonson and colleagues (30), indicating that recruitment occurs throughout a P-V curve measurement and that compliance is decreased at high versus low PEEP during GV because derecruitment is prevented (30).

The significance of the Plc in both GV and PLV is not clear. Plc in gas ventilation may simply identify the minimum PEEP level that prevents marked derecruitment (29), but it may not be the optimal PEEP level maximizing oxygenation. In PLV it seems clear that the Plc represents the minimum PEEP level needed to distribute PFB to the lung periphery and minimize impedance to ventilation (19, 26, 28). However, the Plc may not be the optimal PEEP in PLV to maximize recruitment of both nondependent and dependent lung. PEEP levels higher than Plc may be needed to maximize gas exchange, but no data are currently available evaluating PEEP levels greater than 13-14 cm H₂O (26).

PEEP also recruits and stabilizes nondependent lung and enhances recruitment in dependent lung. It is clear in the GV group that a PEEP of 18.9 cm H₂O was needed to maintain recruitment and oxygenation. In the PLV group the use of PFB at 22 ml/kg fill volume plus PEEP at 11.5 cm H₂O was needed to maintain the same level of recruitment and oxygenation. In addition, PEEP can be expected to improve / matching in nondependent lung. Doctor and colleagues (31) and Lowe and Shaffer (32) have reported that PLV shifts pulmonary blood flow to nondependent lung as a result decreasing the transpulmonary pressure stabilizing these lung units. PEEP at Plc + 1 cm H₂O during PLV improves oxygenation by stabilizing these lung units.

In the present study, we compared PLV with GV after applying PEEP based on lung mechanics during PLV and GV. Although oxygenation improved in both groups, we were unable to identify any significant difference between groups. In addition, filling with PFB did not significantly improve oxygenation in our study. The severity of the lung injury (Pa_O2/ FI_O2 60 ± 12 mm Hg in GV and 58 ± 2 mm Hg in PLV) created in this study may explain the discrepancy between this result and previously published studies (20). The increase in Pa_O2 with the application of high PEEP, however, is consistent with our previous data (19, 26) and that of others (27).

The most surprising and unexpected finding of this study was the dramatic difference in Pa_CO2 levels between the two groups (Figure 5). Part of this difference can be attributed to improved lung mechanics during PLV. Setting PEEP at 19 cm H₂O and plateau pressure at 35 cm H₂O resulted in a 16 cm H₂O driving pressure in the GV group, whereas PEEP was about 11.5 cm H₂O and plateau pressure 30 cm H₂O, resulting in a driving pressure of 18.5 cm H₂O during PLV. This 2.5 cm H₂O driving pressure difference was, however, associated with a 35 to 40% reduction in VT in the GV group between postfill and the 5-h time points, whereas VT remained stable in the PLV group (Figure 6). This reflects a clinically relevant improvement of lung mechanics during PLV that certainly contributed to improved CO₂ removal. In addition, VD/VT was also much greater in the GV group (Figure 3). Better / matching because of the shifting of pulmonary blood flow to nondependent lung (29, 30) during PLV at high PEEP than during GV is the most likely explanation for this finding. Because ventilation during PLV at a FRC fill is preferentially distributed to nondependent lung, the VD/VT and efficiency of CO₂ elimination is improved compared with GV.

Pulmonary Mechanics

In contrast to our previous study in sheep with less severe injury (19), CstL did not significantly increase after filling with a 22-ml/kg dose of PFB. However, CstL did improve after 1 h in the PLV group, whereas improvement in CstL during GV did not occur until after 4 h. Although CstL only significantly differed between PLV and GV at 1 h, there was a trend toward greater CstL in PLV throughout the 5-h ventilation period. Improved compliance and greater driving pressures (18.5 versus 16 cm H₂O) resulted in a greater VT and CO₂ elimination during PLV. The improved compliance during PLV is most probably a result of improved surface tension as a result of the PFB and the application of PEEP > Plc (19, 26). As previously shown by Hickling (29) and Jonson and colleagues (30), the application of PEEP in ARDS does not necessarily increase compliance even though derecruitment is prevented because recruitment may continue throughout the entire inspiratory phase. In this model of ARDS we would not expect compliance to increase unless surface tension improved.

Hemodynamics

In the GV group, mean blood pressure and SVR decreased, whereas Pa, Ppao, and PCV increased during the 5-h ventilation period. Increased mean airway pressures and hypercapnia both contribute to these hemodynamic changes. Acute increases in Pa_CO2 increase vascular filling by constricting capacitance vessels (33, 34), reducing left ventricular contractility and improving left heart ejection by reducing arteriolar tone and afterload (35). However, severe respiratory acidosis may have blunted myocardial and vascular responsiveness to catecholamines and therefore may have resulted in the compromised hemodynamics during GV.

Lung Protection

Experimental (5) and clinical (2) studies have demonstrated a lung protective effect of ventilatory strategies based on high PEEP and lung recruitment during GV. Lung protection has also been reported during PLV in several animal models of acute lung injury (15, 21, 36). In many studies, this lung protective effect was observed despite the use of low PEEP levels during PLV (15, 21, 36). The role of lung recruitment in improving pulmonary structure and function during PLV is therefore difficult to ascertain. In addition, several investigators have reported that PFB may also have anti-inflammatory properties that could modulate the severity and the extent of lung injury during PLV (36, 37). In our study, the only histologic difference that we could detect between groups was a reduction in alveolar collapse and consolidation in dependent lung regions of PLV animals, indicating improved recruitment with PLV. The method that was chosen for tissue fixation may have lowered differences that, although small, remained significant. Essentially, the application of a LPVS during both PLV and GV preserved lung architecture.

Clinical Implication

It has been well established that both PLV and PEEP can recruit lung and improve ventilation. However, the cost of improved oxygenation at "high" PEEP in GV with limited end-inspiratory alveolar pressure is hypercarbia, as illustrated in this study. The use of PLV with PEEP > Plc and a low end-inspiratory plateau pressure can achieve the same oxygenation as a GV LPVS but without the cost of hypercarbia. In those patients where hypercarbia and its associated acidosis are unacceptable, PLV may be the treatment of choice.

Limitation

The lung injury model used in this study was a surfactant depletion model that does differ from adult ARDS. Surfactant depletion models are mechanically stable and suitable to evaluate pulmonary mechanics, but they do not demonstrate the permeability-type edema and inflammation observed in ARDS. In addition, this model could favor PLV because of the surfactant-like properties of the PFB. Both groups were compared for only 5 h and it is possible that results would differ with a longer observation period. We also acknowledge that the small number of animals in each group (n = 5) does not rule out the possibility of beta error. Finally, it is difficult to maintain a constant level of PFB in the lung. Shaffer and colleagues (39) developed an apparatus to evaluate the evaporation loss of PFB. They showed that evaporation of PFB was influenced by minute ventilation and postural positioning. Because there is no standard method to compensate for the evaporative loss during PLV, we provided 1 ml/kg/h supplemental dose of PFB. This supplemental dose has been used in other preclinical studies and tested in preclinical and clinical studies to compensate for evaporative loss (personal communications with Alliance). Our results in terms of oxygenation may have differed depending upon the dose of PFB.

In conclusion, PLV and GV both improved oxygenation in our ovine model of acute lung injury. However, PLV resulted in better ventilation than GV while preserving lung histologic architecture when PEEP was set 1 cm H₂O above the Plc on inspiratory P-V curve and peak alveolar pressure was limited to 35 cm H₂O in both GV and PLV. These results suggest that PLV may have a role in the treatment of patients with ARDS when limitation of VT and peak alveolar pressure combined with a high PEEP strategy results in severe hypercapnia and respiratory acidosis.