INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The management of acute lung injury and acute respiratory distress syndrome (ARDS) remains a challenge for clinicians. These heterogeneous lung injuries are characterized by atelectasis and hypoxemia with decreased compliance and increased intrapulmonary shunting (1). Historically, positive end-expiratory pressure (PEEP) has been used to stabilize recruited lung units improving hypoxemia, lung compliance, and intrapulmonary shunting (2). Recently, partial liquid ventilation (PLV) has been proposed as an alternative to conventional mechanical ventilation in the management of ARDS (3). PLV is conventional mechanical ventilation to a lung partially filled with a perfluorocarbon (PFC). Perflubron (PFB) (Liquivent; Alliance Pharmaceuticals, San Diego, CA) is a PFC developed for medical applications, with a high gas solubility (53 ml O₂/dl PFB; 210 ml CO₂/dl PFB at 25° C), a relatively high density (1.9 g/ml), and low surface tension (18 dyne/cm) (4). The low surface tension allows rapid distribution of PFB throughout the lung improving pulmonary compliance (5, 6). However, because of the high density of PFB, it is preferentially distributed to dependent lung regions, stabilizing recruited lung units. In contrast to homogeneously distributed applied PEEP this "liquid PEEP" primarily stabilizes gravity-dependent lung units (7).

The recommended amount of PFB filling the lung is either equal to the estimated functional residual capacity (30 ml/kg) or liquid is added until at end-expiration and zero PEEP a meniscus parallel to the anterior chest wall is seen in the endotracheal tube (7, 8). Low level applied PEEP (5 cm H₂O) has been recommended during PLV to avoid the high airway pressures caused by slow moving liquid in central airways and to optimize gas exchange (5, 8). However, in our own experience with patients this level of PEEP is insufficient to maximize gas exchange. Unlike the estimation of the minimal level of optimal applied PEEP using the pressure-volume (P-V) curve (9, 10), no method is in current practice to determine the optimal PEEP in combination with the PFB-filled lung.

Because the mechanisms of action proposed for PFB are similar to those of PEEP, we hypothesized that the degree of recruitment by the PFB added to the lung would be reflected by the P-V curve. Specifically, adding PFB would shift the lower inflection point (LIP) on the P-V curve to the left in a dose-dependent manner. We also hypothesized that regardless of the fill level with PFB, the inflection point would indicate the PEEP level required to optimize gas exchange by stabilization of nondependent lung units and movement of the PFB out of the nondistensible conducting airways, improving distribution of tidal ventilation. We tested the assumption that PEEP set above the LIP of the P-V curve during PLV would result in better gas exchange than PEEP set according to the current recommendations, at about 5 cm H₂O (5, 8).

	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

Twenty-three fasted Hampshire sheep (27.6 ± 3.5 body weight) were anesthetized with halothane and orally intubated using a 9-mm inner diameter cuffed endotracheal tube (Hi-Lo Tube; Mallinckrodt Medical, Inc., St. Louis, MO). To ensure gastric drainage an orogastric tube with esophageal balloon (151-14, 14 Fr; Mallinckrodt Laboratories Ltd., Athlone, Ireland) was also inserted. After cannulating the left jugular vein and a loading dose of fentanyl (0.3 mg) and diazepam (10 mg), anesthesia was maintained using fentanyl (3 µg/kg/h) and sodium pentobarbital (5 mg/kg/h), and paralysis was established using pancuronium bromide (2 mg/h). Mechanical ventilation was provided throughout the experiment using a PB 7200ae ventilator (Nellcor- Puritan-Bennett, Carlton, CA); volume control mode, tidal volume (VT) 15 ml/kg, respiratory rate 15 breaths/min, total inspiratory time 1.3 s, inspiratory plateau 0.6 s, and fraction of inspired oxygen (FI_O2) 1.0. PEEP was 5 cm H₂O, except when it was at LIP +1 cm H₂O as required by the protocol. The right femoral artery was cannulated and a pulmonary artery catheter (Edwards Swan Ganzcontinuous cardiac output/mixed venous saturation [CCO/S_O2] 8.0 Fr; Baxter Healthcare Corp., Irvine, CA) was placed into the right jugular vein and connected to a continuous cardiac output monitor (Vigilance; Baxter Healthcare Corp.). Lactated Ringer's solution was administered intravenously to maintain central venous pressure (CVP) between 5 and 7 mm Hg.

Lung Injury

Bilateral lung lavage was performed by instillation and passive drainage of isotonic saline (30 ml/kg body weight) warmed to 39° C (11, 12). The lavage was repeated every 15 min until the Pa_O2/FI_O2 ratio fell below 150 mm Hg. After lung lavage, the sheep were allowed to stabilize for at least 60 min, while the core temperature was maintained at 39° C using an electric heater and blankets.

Experimental Protocol

After the surgical procedures and a stabilization period of 30 min, baseline parameters (preinjury) were obtained as described in the measurement section. After lung injury and a stabilization period of 60 min the measurements were repeated (injury). Before and after lung lavage a static P-V curve was performed by stepwise 50 ml inflation of the lung followed by a  5 s hold (until the airway pressure graphic recording was stabilized) with a calibration syringe to a total inspiratory volume of 500 ml (13). Before the P-V curve determination, a volume history was established using 500 ml VT's. The lung injury resulted in a nonlinear pressure-volume relationship. The LIP was determined by identifying the point of intersection of the slopes of the initial flat and subsequently more vertical portions of the curve by an investigator blinded to the setting where the curve was measured.

Pressure-volume relationship. The first five sheep were used to investigate the influence of increasing doses of PFB on the P-V curve and the LIP. During volume-controlled ventilation, each lung-injured sheep was slowly filled with PFB in increments of 5 ml/kg body weight (ml/kg BW) via a sideport at the endotracheal tube. The total dose of PFB was 30 ml/kg BW, or until a meniscus parallel to the anterior chest wall was seen in the endotracheal tube when the ventilator was disconnected from the airway. After each dose and a stabilization period of 15 min an arterial blood gas value, pulmonary mechanics, and a P-V curve were obtained. Following the experiment a "blinded" investigator, not familiar with the actual animal nor PFB dose, evaluated the resulting pressure-volume relationships and determined the LIP from the inspiratory curve.

Perflubron dose and PEEP interaction. Sheep were randomly assigned to three groups of five animals each receiving different doses (7.5, 15, and 30 ml/kg BW) of PFB. After administering the PFB and a 30-min stabilization period another P-V curve was measured and the LIP determined. In random order two different PEEP levels, one at 1 cm H₂O above the LIP and one at 5 cm H₂O PEEP were applied. Each PEEP level was allowed to equilibrate for 30 min after which data were obtained. To ensure a stable model a third PEEP setting, with the PEEP level equal to the first randomized PEEP level was evaluated.

At the end of each experiment euthanasia was performed with an overdose of sodium pentobarbital and a rapid injection of saturated potassium chloride.

Measurements

Hemodynamics, gas exchange, and pulmonary mechanics were measured before lung injury, after lung injury, after each 5 ml/kg BW dose of PFB for the first five animals, and following stabilization at each of the three PEEP settings in the subsequent 15 animals. Mean arterial pressure, mean pulmonary artery pressure and CVP were monitored using pressure transducers (049924-507A, Argon; Maxxim Medical, Athens, TX) calibrated at 50 mm Hg, with the zero level at midthorax in the supine position. Pulmonary artery occlusion pressure (PAOP) and CVP were measured at end-expiration. Cardiac output and S_O2 were continuously monitored (Vigilance; Baxter Healthcare Corp.). At the end of each step arterial, and for the second part of the protocol mixed venous blood gases were drawn and PO₂, PCO₂, pH, oxygen saturation, and hemoglobin content assessed by a blood gas analyzer (model 238; Ciba Corning Diagnostics Corp., Norwood, MA) and co-oximeter (model 282; Instrumentation Laboratory, Lexington, MA). 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 in this bag was measured using a capnometer (Model 2200; Traverse Medical Monitors, Saline, MI) and VD/VT calculated according to the Enghoff's modification of the Bohr equation (14). Pulmonary mechanics, gas flow, and airway pressure were monitored at the airway opening using a pneumotachometer (3700A; Hans-Rudolph Inc., Kansas City, MO) and pressure transducers (model 45-32-871 ± 100; Validyne, Northridge, CA). Inspiratory tidal volume (no leak assured) was integrated using the flow signal and reconfirmed with a 500-ml calibration syringe (the presence of PFB vapor pressure did not affect VT measurement). At each setting we also measured total PEEP with an end-expiratory hold maneuver for 5 s using the function incorporated in the ventilator. Ventilator compliance (C_vent) was calculated from data obtained during mechanical ventilation as VT measured/(P_Plateau  PEEP_Total) for all sheep and static compliance (C_stat) was calculated from the volume/pressure change on the "linear portion" of the P-V curve above the LIP for the 15 sheep receiving 7.5, 15, or 30 ml/kg BW PFB doses. In no animal was an upper deflection point identified at the delivered volume of 500 ml. All signals were amplified (model 8805C; Hewlett-Packard, Waltham, MA) and recorded at 100 Hz using an A/D-conversion system (WINDAQ/200 V1.36; Dataq Instruments, Hartfield, PA) and a personal computer. All devices were calibrated at the beginning of the experiment.

Statistical Analysis

All data are expressed as mean ± SD, unless otherwise stated. For the pulmonary mechanics, three breaths were analyzed and averaged. Groups were compared at preinjury and injury using a one-way analysis of variance (ANOVA). To detect any variation over time, we also compared the repeated control PEEP settings in each group. A two-way ANOVA for repeated measures, comparing injury, PEEP below, and above the LIP at different PFB doses was performed. When statistical significance was reached it was followed by a post hoc analysis (Tukey honest significant difference [HSD] test). 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

Data from 20 of the 23 sheep investigated were analyzed. Two sheep died of hypoxemia after lung lavage. The third sheep, randomized to the 30 ml/kg BW filling group died of heart failure after filling and was excluded from the analysis.

Pressure-Volume Relationship

Lung lavage in the first five sheep (weight 29 ± 4) significantly decreased oxygenation, which improved with successive doses of PFB (Table 1). During this filling the lower inflection point decreased significantly with the first 10 ml/kg BW dosing, to a minimum for the following 20 ml/kg BW dosing of PFB (p < 0.05, Figure 1). Pulmonary mechanics also improved during the initial filling, reducing the peak inspiratory pressure, plateau pressure, and improving static compliance. However, exceeding 20 ml/kg BW dosing resulted in an increase in airway pressures, and a minor decrease in compliance. Plateau pressure and compliance improved significantly at filling levels of 5, 10, and 15 ml/kg BW (p < 0.05) when compared with pretreatment (0 ml filling). At 30 ml/kg BW peak inspiratory pressure reached its maximum, showing significant differences with the values observed during 5, 10, and 15 ml/kg BW PFB doses (p < 0.02, Table 1). Mean systemic and mean pulmonary artery pressures, and cardiac output were unaffected by increasing PFB doses, although PAOP increased with increasing doses resulting in significant differences between its lowest value at 5 ml/kg BW (9.1 ± 5.8) and values at 25 (11.4 ± 6.9, p = 0.021) and 30 ml/kg BW (11.6 ± 6.9, p < 0.01).

View larger version (13K): [in this window] [in a new window]   Figure 1.   Lower inflection point (LIP) (mean ± SEM) at each PFB dose. The LIP decreased to a minimum at the 10 ml/kg BW dose then remained stable (*p < 0.05 versus 0 ml/kg BW PFB dose).

Perflubron Dose and PEEP Interaction

After lung lavage each group of 5 sheep (27 ± 4 kg) was analyzed. No significant differences for the measured and calculated variables were observed between these groups.

Gas exchange. Arterial blood gases, shunt fraction, and VD/ VT are listed in Table 2. Lung injury significantly decreased Pa_O2 and oxyhemoglobin saturation (p < 0.01) and increased intrapulmonary shunt, and VD/VT (p < 0.05). Adding PFB led to a significant, dose-dependent increase in Pa_O2 (p < 0.01). However, increasing PEEP further improved oxygenation (p < 0.01), and had a greater impact on oxygenation than increasing the PFB dose (2-way ANOVA interaction: PFB dose versus PEEP level, p < 0.006). The Pa_O2 was highest for the PFB dose of 30 ml/kg BW and PEEP at LIP +1 cm H₂O (p < 0.05, Figure 2). This increase in oxygenation was significantly different from all other doses, except for the Pa_O2 at 15 ml/kg BW and PEEP at LIP +1 cm H₂O. Although increasing the PEEP to LIP +1 cm H₂O in the 15 ml dose group improved oxygenation by 34% compared with PEEP at 5 cm H₂O, it failed to reach significance.

View larger version (12K): [in this window] [in a new window]   Figure 2.   Arterial oxygen tension for the two PEEP levels (mean ± SEM) at 5 cm H2O (open bar) and at LIP +1 cm H2O (solid bar) at different doses of PFB. Increasing PEEP at a dose of 30 ml/kg BW increased oxygenation significantly (*post hoc, p < 0.05) compared with PEEP at 5 cm H2O. Between-group comparisons showed a significant increase versus 7.5 ml/kg BW (+post hoc, p < 0.05) and 15 ml/kg BW with PEEP at 5 cm H2O (#p < 0.05).

The effect of PEEP and PFB dose on intrapulmonary shunt fraction is summarized in Table 2. Whereas lung lavage increased the shunt fraction significantly, increasing PFB dose did not change this value. However, adding PEEP at LIP +1 cm H₂O resulted in a reduction of shunt fraction, reaching significance at the 30 ml/kg BW dose. Except for the setting at 15 ml/kg BW and PEEP at LIP +1 cm H₂O, this shunt fraction was significantly lower (p < 0.05) than settings with lower PFB doses, at both PEEP levels.

Mechanical ventilation with 15 ml/kg BW tidal volume led to a slight hypocapnia at preinjury levels. After lung lavage Pa_CO2 increased, and was further increased by PFB instillation. This resulted in the highest Pa_CO2 and VD/VT values for the 30 ml/kg BW group at the low PEEP level. Increasing PEEP to LIP +1 cm H₂O significantly improved VD/VT at the 30 ml/kg BW dose (p < 0.05, Table 2).

Airway pressures and volume. The main results are shown in Table 3. Pulmonary mechanics deteriorated significantly (p < 0.05) after lung lavage. Because of the constant tidal volume, peak inspiratory and end-inspiratory plateau pressures increased (p < 0.05). The administration of PFB with PEEP at 5 cm H₂O had no significant effect on peak inspiratory or plateau pressures, whereas increasing PEEP to LIP +1 cm H₂O resulted in significantly higher peak and plateau pressures for the 7.5 ml/kg BW group. At larger PFB doses peak inspiratory pressure trended upward, while plateau pressure trended downward. Both parameters did not reach significant levels compared with the injury level. However, the change in ventilator compliance differed significantly within the groups. Increasing the dose above 7.5 ml/kg BW and applying PEEP at LIP +1 cm H₂O improved ventilator compliance significantly (p < 0.01). Ventilator compliance for the 15 ml/kg BW dose was the same regardless of PEEP level. At the 30 ml/kg BW dose and PEEP at 5 cm H₂O, ventilator compliance significantly (p < 0.05) decreased. However, increasing PEEP to LIP +1 cm H₂O at the 30 ml/kg BW dose significantly improved ventilator compliance compared with 15 ml/kg BW and the 30 ml/kg BW dose with PEEP at 5 cm H₂O (p < 0.05, Figure 3). In addition, static compliance determined from the P-V curve at zero PEEP increased during the 15 ml and 30 ml/ kg BW doses (p < 0.05, Figure 3). As expected, the total PEEP and mean airway pressures were highest in the group with PEEP at LIP +1 cm H₂O (p < 0.05).

View larger version (21K): [in this window] [in a new window]   Figure 3.   Static lung compliance (Cstat) measured from the linear part of the P-V curve (top) and quasi-static compliance calculated from the airway pressures measured during volume control ventilation (Cvent = VT/PPlateau  PEEP) (bottom) for each group and experimental stage. Although absolute values of Cstat were larger than Cvent, the trends for both measurements were similar. Within groups the comparison of the lung injury settings showed significant differences to preinjury, as well as for intermediate to high fill levels of PFB ($p < 0.05). Increasing PEEP to LIP + 1 cm H2O at the 30 ml/kg dose increased Cvent (*versus 5 cm H2O PEEP 30 ml, p < 0.05). Comparisons across groups showed a better Cstat at preinjury for the 7.5 ml group (p < 0.05), whereas PFB doses above 7.5 ml/ kg resulted in improved compliances (+versus 7.5 ml, p < 0.05). The Cvent increase with PEEP at LIP +1 cm H2O in the 15 ml/kg group was greater than the compliance increase in the 30 ml/kg group at 5 cm H2O PEEP (versus LIP +1 cm H2O 15 ml, p < 0.05).

Hemodynamics. No statistically significant differences were observed for the mean arterial and pulmonary artery pressures (Table 4). Pulmonary artery occlusion pressures increased slightly with increasing doses of PFB or PEEP. For all doses the heart rate decreased significantly (p < 0.01) at both PEEP levels; more detailed analysis revealed only a significant decrease at the 7.5 ml/kg BW dose. No effect on cardiac output was observed, although stroke volume increased (p < 0.05) in the group with the 30 ml/kg BW PFB dose.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The most important findings of this study can be summarized as: (1) The LIP on the pressure-volume curve of the lung-thorax system in ARDS is shifted to the left with increasing doses of PFB. (2) The LIP plateaued at a dose of 10 ml/kg BW and was stable at this plateau with increasing doses of PFB. (3) The setting of PEEP at LIP +1 cm H₂O markedly increased Pa_O2 and decreased shunt fraction at moderate to high doses of PFB when compared with PEEP set at 5 cm H₂O. (4) At high doses of PFB, PEEP set at LIP +1 cm H₂O improved VD/ VT when compared with PEEP set at 5 cm H₂O. In addition, we observed a PFB dose-dependent improvement in oxygenation with PEEP set both at 5 cm H₂O and at LIP +1 cm H₂O as reported by others (5, 15).

Over the past 5 yr extensive research regarding the use of PLV in acute respiratory failure has been performed (5, 6, 15- 18). Many of these studies have shown a dose-dependent improvement in oxygenation in animal models of ARDS (5, 15). Although the proposed mechanisms for improving gas exchange are quite similar for PLV and PEEP (19, 20), little is known about the interaction of PEEP with a partially liquid-filled lung. With both PEEP and PLV gas exchange and lung mechanics are improved by the stabilization of recruited atelectatic lung units (19, 20). PEEP stabilizes recruited lung in a gravity-dependent manner (21). High levels of PEEP are required to stabilize the most dorsal lung units, and low levels of PEEP are needed for the most anterior lung units. As a result, PEEP because of its uniform distribution throughout the entire lung tends to overdistend anterior lung units when set high enough to stabilize the most dorsal lung regions. PLV, on the other hand, recruits the most dorsal lung units but may not affect anterior lung regions because of its distribution to gravity. The application of PEEP to the PFB-filled lung can be expected to stabilize nonrecruited lung units in anterior lung regions.

Lower Inflection Point

Numerous investigators have shown that the LIP on the P-V curve of the lung/thorax system is a useful estimate of the PEEP level needed to stabilize recruited lung and that there is a gravity-dependent distribution of LIPs (2, 9, 10, 21). We have shown that incremental doses of PFB decreased the LIP. This result is consistent with the proposed mechanism by which PLV counterbalances the gravity-dependent distribution of alveolar collapse in the ARDS lung. The nadir in the LIP we observed at 10 ml/kg BW PFB dose can be explained by stabilization of the most anterior lung regions and the need for PEEP to distribute PFB to the lung periphery. Because of the dependent distribution of PFB, regions in the middle and upper parts of the lung are not completely filled with liquid at low doses of PFB and remain unstable, opening and closing with each VT. As the PFB dose is further increased, liquid begins to fill the large airways and trachea, which must be displaced in order for the VT to distribute throughout the lung. At high doses of PFB the LIP may primarily represent the amount of pressure needed to displace the PFB from nondistensible conducting airways.

Gas Exchange

We and others have shown a dose-dependent improvement in oxygenation and shunt fraction with the administration of PFB (5, 15) in ARDS. This is attributed to increasing recruitment of atelectatic regions as the dose of PFB increases. We have also shown a marked further improvement in oxygenation as PEEP is set at LIP +1 cm H₂O. In fact, PEEP showed a greater improvement in oxygenation than PFB (p < 0.006). We believe this is a result of the stabilization of most anterior lung units and the clearance of PFB from central airways at end-expiration. The use of PEEP establishes a pressure barrier maintaining PFB in the distal lung, out of the endotracheal tube and large central airways. High PEEP levels improve the distribution of ventilation at high PFB doses (22). As shown by Hernan and coworkers (8) similar results can be achieved by large tidal volume ventilation. However, we were able to achieve oxygenation levels similar to Hernan while maintaining inspiratory plateau pressure at 27.9 ± 5.2 cm H₂O and VT at 15 ml/kg. Kaisers and coworkers (23) also observed an improvement in oxygenation by almost 200 mm Hg when PEEP was increased from 5 to 15 cm H₂O in a lung injury model. Hernan and coworkers (8) only observed minor increases in Pa_O2 as PEEP was increased from 6 to 14 cm H₂O; however, their observations were made in healthy sheep.

The mechanisms of PFB and PEEP are not limited to the improvement of lung mechanics and oxygenation due to the recruitment of lung units. Both strategies are also thought to improve ventilation-perfusion matching (24, 25) and redistribute pulmonary edema (26, 27). The more selective distribution to the dorsal parts of the lung is an advantage of PLV over PEEP. However, combining PEEP and PLV capitalizes on their different distributions. While the weight of the liquid plays an important role in redistributing blood flow to ventral lung units (24), PEEP ensures stable lung units in the anterior lung.

Pulmonary Mechanics

Increasing doses of PFB along with the setting of PEEP at LIP +1 cm H₂O improved compliance in our ARDS model. Although the improvement in compliance at doses well below FRC is consistent with data in small animal models (5, 15), other researchers (16, 17) have not observed this improvement in sheep models. Their data suggest a decrease in lung mechanics as filling of PFB progressed. We however, observed this improvement in compliance at higher doses of PFB especially when calculations were made from P-V curve data at zero PEEP. The fact that compliance measurements determined from P-V curve data were larger than those determined from ventilator data reflect the more static conditions during P-V determinations. At each volume delivered, the system was held constant for  5 s, whereas, during ventilator plateau pressure measurements the total holding period was 5 s.

One possible explanation for the discrepancy among reported compliance changes might be differences in the lung model used, saline lavage (5, 15) versus oleic acid (16, 17), implying that surfactant depletion injuries might better respond to PFB. On the other hand, it has been speculated that the distribution of gas ventilation during PLV might be less homogeneously distributed in the larger lung than the smaller lung (16). The inconsistent compliance change observed during PLV might also be explained by differences in methodology and PEEP level. The data provided by Tütüncü and coworkers (5, 15), are quasi-static compliance measurements but at 6 cm H₂O PEEP, whereas Overbeck and colleagues (16, 17) reported static compliance values at zero PEEP. Our data from P-V curves may have been determined at a more static state than others, because we maintained the lung volume at each step ( 5 s) and visually (graphic representation of airway pressure) ensured that a static condition was maintained. Because PFB has a relatively greater inertia than gas, a lengthy period is required to obtain a static state. As a result, technique may grossly affect the actual calculated compliance.

Clinical Implications

The clinical use of PFB requires careful attention not only to the dose and administration methodology but also to the precise method used to mechanically ventilate during PLV. We have shown that PEEP set at LIP +1 cm H₂O in an animal model of ARDS maximizes gas exchange and pulmonary mechanics. However, little is known about mode of ventilation or inspiratory/expiratory (I:E) ratio set during PLV nor is there clear data identifying the potential or lack of potential for lung injury during PLV. All of these areas require further investigation before specific clinical guidelines for the application of PLV can be defined.

Limitations

The LIP used to determine the setting of PEEP was obtained from the inspiratory P-V curve. As has been recently shown by Grunauer and coworkers (28) the ideal determination of the LIP should be made from the deflation limb of a complete P-V curve. However, in the PFB-filled lung lengthy equilibration times ( 5 s) were needed at each volume change to obtain static measurements because of the inertia of the PFB. Because these animals were severely injured, the time required to measure an inflation and deflation curve resulted in marked cardiopulmonary decompensation. As a result our determination of the LIP may have overestimated or underestimated the actual LIP.

Measurement of proximal airway pressure (peak inspiratory pressure [PIP], [P_Plateau], and PEEP) in the PFB-filled lung does not reflect the actual pressures in the dorsal lung under the weight of PFB. Compliance calculations as a result may not reflect regional overdistention most likely at high doses of PFB and PEEP. Because we did not measure absolute lung volume, it is difficult to interpret the compliance changes in this study. This experiment was not designed to examine the differences between anterior and dorsal lung mechanics. As a result concern regarding overdistention and possible volutrauma to dorsal lung units at PFB and PEEP levels resulting in maximal Pa_O2 cannot be ruled out.

The lung lavage injury model we used may not represent the same pathophysiology observed in most ARDS patients. However, this model is well established (5, 11, 12, 15) and was used because of its stability for up to 8 h. This was important, because the length of the experiment required a stable injury for approximately 6 h. Due to the nature of PLV it was impossible to obtain a postinjury control value at the end of the experiment. In spite of this limitation, we believe that this study clearly demonstrates a clinically relevant relationship between PLV and PEEP.

We did not include control groups without PLV in this study design because both PLV and PEEP have been independently shown to improve gas exchange and pulmonary mechanics (2, 5, 9, 10, 15). Our goal in this study was to examine the relationship between PLV and PEEP, not the independent effects of each.

In conclusion, the administration of PFB shifts the LIP of the pressure-volume curve to the left in a dose-dependent manner. The setting of PEEP at LIP +1 cm H₂O at moderate to high doses of PFB maximizes oxygenation, ventilation, and pulmonary mechanics without adversely affecting hemodynamics.