INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

LIQUID VENTILATION has long been studied as a way of improving lung mechanics in the face of lung injury and/or prematurity (17, 20, 21, 23, 27, 28). There are two distinct forms of liquid ventilation, i.e., total (tidal) liquid ventilation (TLV) and partial liquid ventilation (PLV). In both forms of liquid ventilation, a liquid that has high carrying capacity for oxygen and carbon dioxide and that does not disrupt the surfactant layer is administered into the lungs. The most commonly used liquids are the relatively inert family of compounds known as perfluorochemicals (PFCs). In TLV (23), the lung is first completely filled with PFC, and liquid tidal breaths of PFC (Fig. 1) are given by using a custom liquid ventilator (22). In PLV, the lung is partially filled with PFC, and gas tidal breaths (Fig. 1) are given by using a conventional ventilator as described by Tutuncu et al. (26) and Fuhrman et al. (4). PLV was first used in the clinical setting in 1989 by Greenspan et al. (6, 7). PLV with PFC is undergoing clinical trials in patients with severe diffuse lung disease (8, 12). The mechanisms by which PLV improves gas exchange in the setting of acute (adult) or idiopathic (infant) respiratory distress syndrome are poorly understood. One proposed mechanism is improved compliance by alveolar recruitment in the settings of surfactant deficiency or surfactant dysfunctional states (4, 15, 16, 26). However, little is known about PFC-surfactant interaction. Leach et al. (13) demonstrated little additional effect on compliance when combining an artificial surfactant with PLV in premature lambs with severe respiratory distress syndrome, but Bachofen et al. (2) and Gladstone et al. (5) demonstrated pronounced effects in different lung models.

We previously reported data on the changes in lung compliance in excised preterm lamb lungs during TLV, with and without surfactant, and developed a methodology to estimate the magnitude of surface tension during inflation and deflation with gas and PFC (25). Although additional studies during PLV were performed in these same animals, it was impractical to report and interpret data from these additional studies in that paper. In the present paper, we report measurements of pressure and volume during PLV (after draining all but 4 ml/kg PFC) in excised lungs from surfactant-deficient animals either pretreated or not with exogenous surfactant. These measurements were performed to determine the effects of PLV on lung compliance, compared with air-filled and TLV lungs. These data also allowed us to draw conclusions regarding the geometric configuration of residual PFC in the lung and changes in the droplet configuration during the respiratory cycle.

We hypothesized that a potential explanation for the improvement in compliance reported with PLV might be the filming of PFC droplets when gas-lung interfacial tension is high. In the film state, the combined interfacial tension of gas-PFC and PFC-lung might be less than that of bare lung, thereby reducing net interfacial tension. We further hypothesized that compliance or recoil would be worse during PLV than during TLV because of the addition of a gas-lung and gas-PFC interface during PLV, compared with the single PFC-lung interface during TLV (Fig. 1). Based on force-balance equations governing the behavior of immiscible droplets on a liquid surface (Fig. 2), we predicted that during PLV the PFC droplets would spread to a film near end inspiration, even in normal lungs, because of the significant increase in gas-lung interfacial tension. Finally, we hypothesized that surfactant would offer additional benefit to PLV alone, since previous work had shown that surfactant reduces both gas-lung interfacial tension (during gas inflation) and PFC-lung interfacial tension (during TLV) (25, 27).

View larger version (17K): [in this window] [in a new window]   Fig. 1.   A: gas ventilation; B: total liquid ventilation (TLV); C: partial liquid ventilation (PLV). During air inflation, there is an air-lung interface (more precisely, an air-surfactant-subphase-lung interface) and thus an air-lung interfacial tension ("surface tension") that contributes to lung recoil. Similarly, during liquid inflation, there is a perfluorochemical (PFC)-lung interface (more precisely, a PFC-surfactant-subphase-lung interface) and thus a PFC-lung interfacial tension contributing to lung recoil. Situation during PLV is more complex. At a minimum, there are 2 interfacial tensions contributing to lung recoil pressure: a PFC-lung interface and a gas-PFC interface. There is potentially also a gas-lung interface, as shown for droplet configuration. When droplet spreads into a film covering the entire alveolar surface, 2 interfaces act in series (gas-PFC and PFC-lung), resulting in a summation of tensions.

View larger version (20K): [in this window] [in a new window]   Fig. 2.   Combined effects of 3 interfacial tension vectors (PFC-lung, gas-PFC, gas-lung) determine the shape of a droplet of PFC at lung lining layer. As gas-lung becomes high relative to gas-PFC (which is constant at 18 mN/m) and to PFC-lung, droplet (A) flattens into a lens shape (B). If gas-lung exceeds the sum of PFC-lung and gas-PFC, then lens spreads further to a film. , contact angle.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Specimen preparation and materials. The overall temporal sequence, described in detail below, was as follows: 1) saline curves, 2) baseline gas curves, 3) degassing 1, 4) randomization to exogenous surfactant or control, 5) gas curves, 6) degassing 2, 7) TLV curves, 8) draining of PFC, 9) PLV curves, and 10) leak testing.

As previously described (25), lungs from 10 consecutive timed-gestation lamb fetuses [2.29 ± 0.15 kg body wt, 125 ± 1 (SD) days gestation, term = 145 days] were excised without introducing air into the trachea. After being weighed, the lungs were immersed in saline in a plethysmograph and allowed to drain retained fluid passively to their resting volumes. Saline curves were obtained. After baseline gas curves, the lungs were degassed by applying vacuum simultaneously to the trachea and plethysmograph for 30 s (slow boiling at room temperature with ambient pressure less than vapor pressure of water). One-half of the animals had their lungs then treated with exogenous surfactant (see below), and the other half were sham treated with equal volumes of air. After a second degassing, the central airways were filled with PFC before initiating liquid pressure-volume (P-V) loops. A few residual air bubbles were removed during the first TLV P-V loop, but on subsequent cycles no bubbles were visible. After the last TLV P-V loop, the PLV studies followed immediately. Lungs were drained of all but 4 ml/kg of PFC, based on measurement of infusion and withdrawal volumes. The amount of 4 ml/kg of PFC left in the lung was chosen to match the residual volume of gas during air inflation after surfactant treatment (24). The PFC-filled syringes were replaced with air-filled syringes, and another set of air P-V curves was then obtained (PLV condition). Because the PLV protocol (lasting 30 min) immediately followed the TLV protocol and because the syringe/plethysmograph system is a closed system, there was no loss of PFC due to volatilization.

P-V curves. Under each experimental condition, we attempted a family of P-V curves from residual (resting) volume up to target volumes of 20, 30, and 40 ml/kg body wt at 22 ± 1°C (1, 18). For measurements during gas inflation, the lungs were surrounded by air, during TLV the lungs were immersed in PFC, and during PLV the lungs were partially immersed in PFC (the buoyancy of the air and tissue causing part of the lung to float above the surface of the PFC). The volume after initial excision and draining was taken as the zero reference, and all subsequent lung volumes were expressed relative to this one. For PLV curves, "volume" was the sum of the residual 4 ml/kg of PFC plus the superimposed air volume. Quasi-static (stop-flow) pressure readings were obtained at volume increments of 10% of the target volume. To reduce the danger of lung rupture, dynamic airway pressures were limited to 30 cmH₂O for TLV P-V curves and to 35 cmH₂O for gas and PLV curves. Three curves were obtained to each target volume, the third being used for analysis. The pressure limit for TLV was set lower than for gas and PLV curves based on pilot data that showed an unacceptable rate of lung rupture during TLV with dynamic pressures exceeding 30 cmH₂O. As a consequence, not all families of curves are complete (i.e., 40 ml/kg was not always possible without exceeding these pressure limits) (25). Because only the 20 ml/kg family of curves is complete for all conditions, only these results are presented. The rate of inflation and deflation for liquid P-V curves was low (15 ml/min) to minimize flow-resistive gradients and fluid trapping on expiration. Gas (including PLV) curves were obtained at flow rates of 100 ml/min [Bachofen et al. (3)]. Fluid and gas trapping were assessed from the syringe volume not recovered at the end of deflation [defined as transpulmonary pressure (0.1 cmH₂O)]. Our pilot studies like those of Bachofen et al. (1) showed no significant difference in gas P-V characteristics between a rate of 15 and 100 ml/min, and thus the higher rate was chosen.

Automated quasi-static P-V curves were obtained by using a custom computer-driven precision stepper-motor syringe pump with a rapid-response feedback loop based on dynamic airway pressure. Real time P-V curves were generated with an XY plotter, and interval-selected data were stored on the hard drive. Pressure from a transducer was sampled by the computer by using a precision (12-bit) analog-to-digital converter. Volume was calculated based on syringe displacement and corrected for system compliance. Quasi-static pressure readings were obtained by waiting 5 s after each volume change in both the liquid- and air-filled lungs to allow for stress adaptation and for equilibration of proximal airway pressure with distal alveolar pressures (3). Transpulmonary pressures were measured as previously described (25).

After the measurement of all P-V curves, the lungs were leak tested by water immersion during inflation with air to the same maximum volumes achieved during acquisition of P-V curves. Any bubbling of air from the pleural surface was defined as a leak. Transpulmonary pressures were not monitored during this process.

Surfactant administration and PFC. In one-half of the lungs, 4 ml/kg body wt of surfactant suspension (Survanta, Abbott Laboratories, Columbus, OH; 100 mg/kg total dose of phospholipids, 44-62 mg/kg diphosphatidylcholine) were administered manually by a syringe via a side port in the tracheal cannula and dispersed by repeated infusion and withdrawal (15-20 cycles over 2 min, 4 ml/kg tidal volume) finishing with an infusion. To distribute the surfactant, the lungs were inflated and deflated with air six times over 10 min between pressure limits of 0 and 35 cmH₂O. The control group of lungs received the same protocol, except that 4 ml/kg of air was substituted for the surfactant, analogously to clinical trials (9). Administration and distribution of the surfactant took place over 20 min to allow time for surfactant adsorption. The PFC used in these TLV and PLV studies was perflubron (LiquiVent, Alliance Pharmaceutical, San Diego, CA). Perflubron is a slightly lipophilic PFC (C₈F₁₇Br, density 1.918 g/cm³, surface tension 18.0 mN/m) in use in human clinical trials (11, 12).

Analysis. Compliance per kilogram was defined as a straight-line slope: peak minus minimal volume difference (normalized to body weight), divided by the corresponding quasi-static pressure difference at those points. Average P-V curves were determined by taking the mean of transpulmonary pressure readings at isovolumetric points on inflation and deflation.

Estimates of in situ interfacial tensions () at various volumes of inflation were used in conjunction with a force-balance equation to predict alveolar PFC droplet configuration and spreading during lung inflation and deflation.

As previously described (25), we estimated  from P-V curves by using data from Bachofen et al. (3). Their data consisted of  vs. volume (-V) loops and simultaneously measured P-V loops. From these two relations, the -P relationship can be obtained, i.e., between airway pressure (Paw) and  at isovolumetric points:   f(Paw), where f is a function of Paw. To correct for possible differences in tissue elastic recoil between the adult rabbits studied by Bachofen et al. and our preterm lambs, we calculated the "interfacial" recoil (P) by subtracting saline pressure (P_sal) (using our own normal saline data) from Paw to obtain an empiric relation:   f(Paw  P_sal)  f(P). We found that, unlike the relatively wide hysteresis in both the -V and P-V relationships, the -P relationship appears to be a single-valued function.

Prediction of alveolar PFC configuration during PLV. The shape of a droplet of PFC resting on the lung lining layer is determined by the relative magnitude of three vectors representing interfacial tension (Fig. 2): gas-PFC, gas-lung, and PFC-lung. This property of PFC to form droplets of various diameters and shapes was used by Schürch and colleagues (19) to determine gas-lung interfacial tension in situ.

Conceptually, the shape of the droplet can be predicted as follows. When the _gas-lung vector to the right is almost as large as the sum of the horizontal components of the vectors to the left (_PFC-lung and _gas-PFC), then the PFC drop spreads into a lens shape as shown in Fig. 2. If _gas-lung exceeds the sum of the _PFC-lung and _gas-PFC components, then the forces become unbalanced and the PFC spreads further to become a thin film covering the lung lining layer. This would occur whenever the lung surface tension is high, as in surfactant deficiency or at the end of a deep inspiration.

To estimate the PFC droplet configuration at the lung-PFC interface more systematically, we applied a force-balance equation. The contact angle () of the droplet with the lung is determined by the relative magnitude of the interfacial force vectors (_PFC-lung, _gas-PFC, and _gas-lung) (14, 19). Horizontal force balance requires that _PFC-lung + _gas-PFC cos = _gas-lung. By rearranging,  = cos¹ [(_{gas-lung  PFC-lung})/_gas-PFC].

The magnitude of the three  vectors is estimated as follows. The _gas-lung results from the interfacial tension at the gas-lung interface (Fig. 1) and is the interfacial tension responsible for lung recoil during gas inflation. Thus, using our estimates of in situ interfacial tension for gas inflation of the preterm lung with and without exogenous surfactant (ES) we were able to determine at each stop-flow point (2 ml/kg increments) on inflation and deflation an estimate of _gas-lung (25) [gas and ES+gas curves in Figs. 3 and 4, respectively]. At the same stop-flow points, we used the TLV estimated in situ interfacial tension to estimate _PFC-lung (25) [TLV and ES+TLV curves in Figs. 3 and 4, respectively]. Due to the fact that the control gas inflation condition (Fig. 3) could not achieve volumes >11.4 ml/kg, we estimated the _gas-lung vector to be 60 mN/m for all volumes 11.4 ml/kg. For example, as shown in Fig. 3, on deflation to 8 ml/kg, _gas-lung is 31 mN/m and _PFC-lung is 14 mN/m. The _gas-PFC vector is constant at 18.0 mN/m for perflubron (11). These three numbers can be substituted in the equation above to give an estimate of the PFC-lung  in the surfactant-deficient lung at 8 ml/kg on deflation during PLV. This substitution is possible because a droplet of PFC on the lung lining layer does not disrupt the surfactant layer, nor does the surfactant spread over the droplet (19). Using this approach, we were able to estimate the relationship between lung volume and  at points along the inflation and deflation limb of the PLV P-V curves in both surfactant-deficient and surfactant-treated groups.

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Transformation of our previous pressure-volume (P-V) data during gas and TLV studies of preterm lamb lungs into interfacial tension vs. volume data (surfactant-deficient group) (see Ref. 25). These derived interfacial tension data are used to determine magnitude of vectors shown in Fig. 2 to solve the force-balance equation. For example, at 8 ml/kg, interfacial tension on deflation during TLV (PFC-lung) is 14 mN/m and interfacial tension during gas deflation (gas-lung) is 31 mN/m. Force-balance equation predicts that the PFC-lung  will be 19°. Inset shows force-balance vectors schematically along with the calculated , predicting a lens configuration.

View larger version (15K): [in this window] [in a new window]   Fig. 4.   Transformation of previous P-V data during gas and TLV studies of preterm lamb lungs into interfacial tension vs. volume [exogenous surfactant (ES)-treated group] (see Ref. 25). These derived interfacial tension data are used to determine magnitude of vectors shown in Fig. 2 to solve the force-balance equation. For example, at 20 ml/kg, interfacial tension during ES+TLV (PFC-lung) is 27 mN/m, and interfacial tension during gas deflation (gas-lung) is 51 mN/m. Solution to force-balance equation is indeterminate. Analysis of the magnitude of vectors, however, reveals that gas-lung (to right, 51 mN/m) exceeds the sum of PFC-lung and gas-PFC (to left, 18 + 27 = 45 mN/m), thus pulling the PFC droplet into a film covering lung lining layer. Inset shows force-balance vectors schematically along with 0° .

Statistical analysis. Two-sample t-tests not assuming equal variances were used to compare compliances and inflation pressures. Data are reported as the means ± SE. Results were considered significant at P < 0.05.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

P-V relationships during PLV compared with TLV and gas. In both the control surfactant-deficient group (Fig. 5) and the surfactant-treated preterm lambs (Fig. 6), total compliance for the PLV condition is intermediate between the gas inflation condition and the TLV. Without surfactant pretreatment, the compliance for PLV inflation is 0.57 ± 0.04 ml/cmH₂O, representing a 63% improvement (P < 0.005) compared with baseline for gas inflation (0.35 ± 0.01), although not as great an improvement (P < 0.02) as seen with TLV (128%, 0.80 ± 0.05 ml/cmH₂O). Similarly, in the group treated with ES, the compliance for PLV inflation is 0.72 ± 0.03 ml/cmH₂O, representing a 106% improvement (P < 0.001) compared with baseline for gas inflation (0.35 ± 0.01) or a 35% improvement (P < 0.002) compared with gas inflation after surfactant (0.53 ± 0.02 for ES+gas). Even with surfactant the improvement in compliance with PLV is not as good as for ES+TLV (1.06 ± 0.03), which represents a 203% improvement over baseline for gas inflation (P < 0.0001).

View larger version (13K): [in this window] [in a new window]   Fig. 5.   Static P-V relationship obtained during the PLV condition in surfactant-deficient preterm lamb lung (PLV) is shown compared with previously published data on same lungs during gas and TLV. PLV is as effective at recruiting lung volume as TLV (20 ml/kg) and more effective than gas inflation. Except at low lung volumes, PLV requires higher pressures to achieve same lung volumes than TLV. Error bars represent means ± SE.

View larger version (13K): [in this window] [in a new window]   Fig. 6.   Static P-V relationship obtained during the PLV condition in ES-treated group (ES+PLV) is shown compared with previously published data on same lungs during gas (ES+Gas) and TLV (ES+TLV). ES+PLV is more effective at reducing compliance than ES+Gas and almost as effective as TLV at reducing pressures on inflation. On deflation, PLV is less effective at retaining lung volume compared with TLV and to ES after surfactant. Error bars represent means ± SE.

In the surfactant-deficient state (Fig. 5), PLV is more effective at recruiting lung volume than is gas inflation. During most of inflation, higher pressures are required with PLV to achieve the same lung volumes as TLV. Similarly, during most of deflation, the PLV pressures are higher compared with TLV. In the surfactant-replacement group (Fig. 6), the situation is more complex. During inflation, PLV reduces inflation pressures, compared with gas inflation at isovolumetric points. At lower lung volumes, it is indistinguishable from TLV, but at higher volumes, inflation pressures are higher for PLV. On deflation PLV always requires higher pressures to maintain the same lung volumes compared with TLV. Gas deflation stability after ES is improved to the point that, at lung volumes <14 ml/kg, PLV actually requires higher pressures to maintain lung volume than does the gas condition. Figure 7 combines the PLV data from Figs. 5 and 6 and shows that at all volumes on inflation and deflation the ES+PLV curve is shifted to lower pressures than in the PLV alone group. The 26% improvement in compliance of ES+PLV compared with PLV alone (0.72 ± 0.03 vs. 0.57 ± 0.04 ml/cmH₂O) is statistically significant (P < 0.02).

View larger version (12K): [in this window] [in a new window]   Fig. 7.   Combination of PLV data from Figs. 5 and 6, comparing the surfactant-deficient preterm lamb lung P-V relationship (PLV) and surfactant-treated group (ES+PLV). At all volumes on inflation and deflation, the ES+PLV curve is shifted to left of the PLV alone group.

Predicted PFC droplet configuration. Solving the force-balance equation (Fig. 2) with the data from Figs. 3 and 4 allows us to compare the  for PFC-lung and thus droplet configuration during PLV between the PLV and ES+PLV groups. In the surfactant-deficient PLV group (Fig. 8A), the PFC is drawn out into a film during most of the inflation and deflation cycle (from ~40 to 100% maximal lung volume) because _gas-lung (lung surface tension) is high. In this range, the PFC acts like a surfactant that covers the alveolar surface and reduces net interfacial forces (Fig. 5). However, in the ES replacement ES+PLV group (Fig. 8B), the PFC remains in a droplet/lens configuration until near-maximal inflation volumes where _gas-lung is sufficiently high to film the PFC over the alveolar lining layer.

View larger version (23K): [in this window] [in a new window]   Fig. 8.   A: predicted PFC-lung  during PLV in surfactant-deficient preterm lamb lung derived from data in Fig. 5 and solving the force-balance equation. B: predicted PFC-lung  during PLV in ES-treated preterm lamb lung derived from data in Fig. 6 and solving the force-balance equation. Nos. represent  values. When the solution to force-balance equation is indeterminate ( = 0), PFC lens is stretched into a film configuration. Regions of the P-V loops where PFC is in the film configuration are represented by a thin solid line, and regions where PFC would assume the droplet/lens configurations are represented by a thicker solid line. Transitions are represented by dashed lines.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

This study extends previous work on lung mechanics and interfacial tension during TLV to the PLV condition. We found that total compliance during PLV is intermediate between gas inflation and TLV. The addition of ES before treatment with PFC improves compliance for all modes of inflation (gas, PLV, TLV), but TLV plus surfactant remains the most effective at reducing compliance. In an effort to understand why the PLV curve is intermediate between TLV and gas inflation we used a force-balance equation to predict the  between the alveolar PFC and the lung lining layer. Our estimates of  predict that, during PLV at low lung volumes, the PFC is in a droplet/lens configuration with exposed alveolar surface. As lung volume and/or lung surface tension rises, the PFC is pulled into a lens shape with a smaller , until the PFC spreads out into a film covering the alveolar surface.

Our results are complementary to those published by other investigators studying the effects on lung mechanics of gas inflation of lungs after treatment with liquid surface tension-reducing agents. Bachofen et al. (2) used a similar excised-lung model to study the effects of disturbances of the alveolar lining layer on lung microarchitecture. The model used adult rabbit lungs and estimated in situ interfacial tension based on morphometric techniques. The hexadecane-filled ("TLV") P-V loops were shifted to the left compared with the hexadecane-rinsed ("PLV") P-V loops, consistent with our data. Findings by Bachofen et al. confirm our findings that gas inflation after PFC treatment is more effective at reducing interfacial tension than gas inflation alone but is not as efficient as total liquid filling. Our data extend the findings by Bachofen et al. to the surfactant-deficient preterm model and to surfactant pretreatment before PFC treatment.

Our findings differ from those reported by Bachofen et al. (2) for FC-77. We attribute this to the fact that perflubron has a different interaction with surfactant than FC-77, despite the fact that both are PFCs. We concur with these authors' conclusion that "the molecular structure of the test liquids and their interfacial tension with surfactant are as important as their air/liquid interfacial tensions in determining alveolar surface forces" (2). We believe that our specific finding should not be extrapolated beyond the specific PFC-surfactant combination we studied.

The work of Wolfson et al. (27), who used a similar but in vivo model, found virtually identical improvements in compliance and P-V curve shape despite differing residual volumes across experimental conditions. Specifically, their in vivo model TLV curves started at much higher residual volumes than their PLV curves, which, in turn, started at higher residual volumes than their gas curves. Thus, unlike our study, their comparisons of inflation pressures across conditions were not isovolumetric.

Our findings are at odds with the study by Leach et al. (13), which showed that PLV with surfactant was not more effective at improving compliance than PLV alone and which gave a different estimate for improvements in compliance. There are three important differences in the two studies that likely account for the differences. The most important is that a different surfactant was used (in our study, we used the natural surfactant Survanta, and in the study by Leach et al. the artificial surfactant Exosurf was used). Given that they showed no effect on compliance when comparing the baseline gas to the surfactant-treated lungs, it appears that their surfactant treatment was ineffective. Another important difference is that our studies were on excised lungs, and thus our compliance comparisons are at isovolumetric conditions, whereas their studies in intact animals likely compared compliances starting from different resting volumes. Finally, our curves are under stop-flow conditions, whereas theirs are dynamic curves. Gladstone et al. (5) measured compliance by using in situ preterm lamb lungs and compared treatment with surfactant (natural sheep surfactant) with a 3% solution of the PFC FC-100. Their findings were similar to ours in that the FC-100 "rinse" was more effective than surfactant at improving total lung compliance. Their baseline static compliance (0.11 ml · cmH₂O¹ · kg¹) was lower than in our study (0.35 ml · cmH₂O¹ · kg¹); however, they achieved similar compliances in the FC-100-treated group as we did in our PLV control group (0.54 vs. 0.53 ml · cmH₂O¹ · kg¹).

Limitations of the methodology used to estimate in situ interfacial tension based on P-V loops have been previously discussed (25). These limitations affect the magnitude of the estimate of the air-lung and PFC-lung interfacial tension vectors but do not change the relative rank ordering of the interfacial tensions. Thus, if at isovolumetric points the transpulmonary pressure is lower, then the magnitude of the interfacial tension must be lower. This does not affect our results regarding the raw P-V data. However, it does have an impact on our estimations of . Specifically, if the magnitude of the vectors in Fig. 2 is inaccurate, then the  values and the exact point during the P-V loop when the PFC spreads to cover the alveolus (Fig. 8, A and B) will be affected. Our prediction that the PFC does undergo conformational changes (droplet to film) during the respiratory cycle is unlikely to be affected.

During PLV, PFC is not uniformly distributed in the lung, as supported by our mathematical modeling work (24) with a tendency to "pool" in dependent regions (10, 29). Our previous work (25) and that of Bachofen et al. and Schürch et al. (2, 3, 18, 19) assume that the interfacial tension is uniform throughout the lung, whereas if the PFC is not uniformly distributed, this is unlikely to be the case. Our conclusions are drawn from the net lung recoil, including contributions by dependent and nondependent alveoli, and thus describe what occurs for a representative alveolus. The implications are that for one given alveolus the transition from droplet to film and back may occur at different points on the P-V relationship than for another alveolus. Nevertheless, our conclusion that there is a conformational transition is unaffected.

Our conclusion that the PFC undergoes a conformational change from droplet to film is supported by the preliminary calibration data used by Schürch et al. (19). They used those data for estimation of in situ interfacial tension during gas inflation based on the shape of PFC droplets on the alveolar lining layer. Specifically, they noted that, as surface tension rose, PFC droplets would change to film, as we have predicted on the basis of the force-balance equations. Furthermore, even if our estimates of interfacial tension are incorrect, the relative magnitude of PFC-lung and air-lung interfacial tension based on the P-V curves suggest that the interfacial tensions are higher in the surfactant-deficient lungs, thus favoring filming at lower lung volumes. Finally, our model assumes a solid planar alveolus rather than a deformable curved surface. Although this may affect the exact point at which the PFC films, it does not alter our conclusion that PFC does, in fact, undergo transition from droplet to film.

The implications of our results are as follows. The combination of ES+PLV is more effective at improving lung compliance than either modality alone. A caveat is that this is likely to be dependent on the surfactant preparation used [as seen in the discrepancy between our results and those of Leach et al. (13)] and dependent on the PFC used (2, 25). Our conclusions may only apply to the combination of Survanta and perflubron (2, 25). In terms of total compliance, TLV seems to offer advantages over PLV, but other factors must be considered such as efficiency of gas exchange and requirements for a liquid ventilator. The combination of ES+PLV is almost as effective as TLV alone at improving total compliance and may offer a reasonable compromise. Improving compliance may reduce lung inflation pressures, thus reducing barotrauma. On the other hand, even small changes in interfacial tension may have significant effects on overdistension of alveolar ducts, as discussed by Bachofen et al. (2) and our group (25). Thus, if reducing inflation pressures and alveolar duct stretch is more important than optimizing gas exchange, then our work and that of Wolfson et al. (27, 30) show that TLV is more effective than PLV at reducing inflation pressures, improving compliances, and reducing interfacial tension (25). Our work shows that surfactant pretreatment increases compliance and reduces inflation pressures during PLV and TLV but does not eliminate the advantage that TLV has over PLV in terms of minimizing inflation pressures. The conformational change of the PFC from droplet to film state during PLV will potentially affect gas exchange by creating a diffusion barrier to gas exchange when in the film state (15). The larger the PFC dose, the more marked this effect would be. These conformational changes occur with PLV and not with TLV; however, the therapeutic implications of this observation are unknown. The effects of the conformational change of the PFC on the alveolar lining layer and on alveolar wall and alveolar duct tensile and sheer forces are unknown and warrant further investigation.

In conclusion, in a preterm excised lamb lung model of surfactant deficiency, compliance during PLV with the PFC perflubron is intermediate between gas ventilation and TLV. This compliance remains intermediate after pretreatment with ES (Survanta), although the combination of ES+PLV is almost as effective as TLV alone at improving total compliance. Alveolar PFC is predicted to undergo conformational changes from a droplet to a film during inflation.