INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The mechanical properties of the lungs of patients suffering from acute respiratory distress syndrome (ARDS) are not uniform (1), with diseased and nondistensible zones existing alongside quasinormal parenchyma that may be open and ventilated or collapsed and potentially recruitable. Mechanical ventilation of these lungs may result in overinflation of the aerated zones, with the risk of worsening injury. Experimental studies have shown that mechanical ventilation with high airway pressures and high tidal volumes may be deleterious (2- 4). Overinflation changes capillary permeability, causes edema, and causes diffuse alveolar damage (4, 5).

The risk of ventilator-induced lung injury (VILI) has substantially influenced the way in which mechanical ventilation is used in patients with acute lung diseases (6, 7). Several strategies have been proposed to reduce the risk of VILI. One such strategy limits the pressure and volume of gas delivered to the lungs, at the expense of producing hypercapnia (8, 9). Another ("open lung approach" [10]) tries to limit VILI by increasing the amount of aerated lung, which may reduce the risk of regional overinflation. This is accomplished by stabilizing terminal units with an appropriate level of positive end-expiratory pressure (PEEP) (11) or by instilling tensioactive material. Encouraging results have been obtained with surfactant replacement therapy in premature babies (14), but its beneficial effect during adult ARDS is less certain (15). Liquid ventilation with perfluorocarbons (PFCs) has also been suggested to reduce surface tension during ventilation (16). Partial liquid ventilation (PLV) with PFCs (at a dose equivalent to the FRC) may reopen closed units and improve lung mechanics because of the tensioactive properties of PFCs (17). Many studies have shown that PLV (also referred to as perfluorocarbon-associated gas exchange) markedly improves lung mechanics and gas exchange in animal models of hypoxemic respiratory failure, such as after surfactant depletion by repeated saline lavage, or in lung injury caused by oleic acid or hydrochloric acid (17). Overinflation edema may result at least in part from surfactant inactivation, and the lung injury seen after long periods of high volume ventilation has been ascribed to the interaction of edema, local overinflation, and repeated opening and closing of terminal units (4). Existing lung abnormalities may potentiate the deleterious effects of high volume ventilation (20). PLV may protect against VILI by reducing the nonuniformity of the lung and improving its mechanical properties. This was not assessed in previous studies, which mainly focused on the effects of PLV on gas exchange. Documenting such an effect would have important clinical implications: ARDS patients seldom die from hypoxemia, whereas VILI may contribute to a poor outcome in this syndrome.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Experimental Preparations

Male Wistar rats (Iffa-Credo, Oncins, France) weighing 320 ± 5 g were anesthetized by intraperitoneal injection of thiobarbital (50 mg/kg; Inactin, Promonta, Hamburg, Germany) and a tracheostomy was performed. The tracheotomy tube was tightly secured to the trachea to preclude leakage, and the rats were allowed to breathe spontaneously before being given mechanical ventilation. They were then paralyzed with succinylcholine (5 mg/kg) and connected to a Harvard volume ventilator for small animals (Ealing; Les Ulis, France). The airway pressure was continuously monitored via a side port on the inspiratory tubing, which was connected to a Validyne MP45 ± 50 cm H₂O transducer (Northridge, CA) that was referenced to the atmospheric pressure. A displacement transducer (DCDT 7000; Hewlett-Packard, Rockville, MD) was connected to the piston of the respirator to monitor the inspired volume. Airway pressure and volume signals were fed to an analog-to-digital converter (Lab-PC; National Instruments, Austin, TX), sampled, and stored on files in an IBM PC-compatible computer (Elonex; Genneviliers, France). Pressure-volume curves for the respiratory system were drawn to determine the pressure value of the lower inflection point (opening pressure) and the end-inspiratory pressure.

Experimental Protocol

Two sets of experiments were performed. The first set evaluated the severity of the injury produced by 10 min of mechanical ventilation as aerated lung volume was reduced. The volume was reduced by instilling 6 ± 0.1 ml/kg normal saline into the trachea to mimic alveolar edema. Several values of tidal volumes (VT) (7 ± 0.2, 16 ± 0.7, 24 ± 0.4, and 32 ± 0.7 ml/kg) were tested. The number of animals (ordered according to VT) were four, five, four, and five in the control groups, and seven, five, four, and six in the instilled groups. The respiratory frequency was 60, 32, 25, and 25 breaths/min for animals ventilated at a VT of 7, 16, 24, and 32 ml/kg, respectively. These experiments were done to determine the type of ventilation that did significantly more damage to lungs with reduced aerated volume than to normal lungs, as indicated by the indices of permeability edema described in detail subsequently. Respiratory system compliance was not measured in these animals. The second part of the study was conducted to determine whether instillation of a small volume (3.3 ± 0.02 ml/kg) of PFC (Liquivent; Alliance Pharmaceutical, San Diego, CA) protected lungs whose aerated volume was reduced by instilling 6.7 ± 0.03 ml/kg saline (n = 19) against the deleterious effects of mechanical ventilation with 33 ± 0.2 ml/kg VT for 10 min. There were seven control (uninstilled) rats. Three strategies for PFC administration were evaluated: in the first two, the PFC was instilled as a bolus (n = 8, Modality a) or over a 10-min period with a motorized syringe (n = 9, Modality b) before saline was instilled; in the third strategy, the PFC was given as a bolus after saline instillation (n = 16, Modality c).

Indices of Pulmonary Microvascular Injury

The rats were injected with radiolabeled tracers (see below) 30 min before being killed (i.e., 20 min before the period of mechanical ventilation). This interval was necessary to ensure time for mixing and exchange of the tracers. We estimated the extravascular lung water content (in animals that were not instilled with saline) and bloodless dry lung weight, using red blood cells labeled with ^99mTc to determine the lung blood content, as previously reported (5, 20). Changes in microvascular permeability were quantified by the distribution space in lungs of rats injected intravenously with ¹²⁵I-labeled serum albumin (radioisotopic serum albumin [RISA]) (5, 20). After the animals were killed with a lethal dose of pentobarbital, the lungs and a blood sample obtained by cardiac puncture were weighed, and their tracer activities were determined by gamma counting (Wizard Wallac, Turku, Finland), with proper corrections for the overlap between channels. Extravascular lung water (Qwl), bloodless dry lung weight (DLW), and the extravascular albumin distribution space in lungs (Alb. Sp.) were calculated according to usual formulas (5, 20). In animals instilled with saline, DLW was corrected by subtracting the dry weight of the instillate (9 mg/ml × 2 ml = 18 mg).

Data Presentation

Data were standardized to rat body weight. All results are expressed as means ± SEM. Comparisons between experiments were made by analysis of variance (ANOVA). Logarithmic transformations were made to homogenize variances when necessary. Regression analyses were done with the unweighted least-squares method. Segmental regression was done according to Gallant and Fuller (21). Statistical significance was accepted at the level of p < 0.05. 

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Ventilating intact rats at a VT of 6, 16, or 24 ml/kg for 10 min did not significantly modify Qwl, DLW, or Alb. Sp., and thus did not produce pulmonary edema (Figures 1 and 2). There was a small but significant increase in Qwl and Alb. Sp. with a VT of 32 ml/kg (Figures 1 and 2), indicating mild permeability edema. Rats instilled with 2 ml saline were more sensitive to the harmful effects of ventilation. Extravasation of plasma protein was accelerated with a VT of 8 ml, as indicated by the increase in Alb. Sp. (Figure 2b), but there was no significant increase in DLW (Figure 2a). By contrast, DLW was significantly increased in rats instilled with saline and ventilated with a VT of 32 ml/kg (p < 0.05). The increase in Alb. Sp. was impressive (p < 0.001). A two-factor interaction between ventilation and alveolar flooding was disclosed by two-way ANOVA (F = 35, p < 0.001), indicating a synergy between the two insults.

View larger version (34K): [in this window] [in a new window]   Figure 1.   Effect of increasing VT during mechanical ventilation for 10 min on the amount of extravascular lung water (Qwl) in intact rats. Pulmonary edema occurred only with the largest VT (*p < 0.05 as compared with the other groups).

View larger version (19K): [in this window] [in a new window]   View larger version (17K): [in this window] [in a new window]   Figure 2.   Effect of increasing VT during mechanical ventilation for 10 min on lung permeability indices: bloodless DLW (a) and extravascular albumin distribution space in lungs (b) of rats with intact lungs (open bars) or with alveolar flooding (closed bars) produced by saline instillation. There was a moderate increase in the Alb. Sp. in intact rats at the larger VT. Lung flooding did not produce significant increases of Alb. Sp. or DLW when the VT was normal or moderately increased. A VT of 24 ml/kg increased only the Alb. Sp. A VT of 32 ml/kg increased both parameters significantly. The increase in Alb. Sp. greatly exceeded additivity, indicating a positive interaction between the two insults (see text). (+p < 0.05 as compared with other groups of intact rats; *p < 0.05 and ***p < 0.001, respectively, as compared with intact animals).

The effect of instillation of PFC on the indices of permeability pulmonary edema in rats ventilated with a VT of 33 ml/ kg is shown in Figure 3. Flooding in the absence of PFC produced the same increases in dry lung weight (Figure 3a) and albumin space (Figure 3b) as in the first set of experiments. PFC instillation, by all the means tested, significantly decreased DLW (p < 0.05, Figure 3a) and Alb. Sp. (p < 0.001, Figure 3b). However this response was not uniform. Instillation of PFC antagonized the effects of alveolar flooding in most rats. There was no reduction in the permeability changes in a few instances. As a result, Alb. Sp. remained significantly different from that of controls (p < 0.05).

View larger version (15K): [in this window] [in a new window]   View larger version (15K): [in this window] [in a new window]   Figure 3.   Effect of PFC instillation on indices of permeability pulmonary edema in rats ventilated with a VT of 33 ml/kg. Flooding increased both DLW (a) (p < 0.05) and Alb. Sp. (b) (p < 0.001). PFC given as a bolus dose before flooding (a), by slow infusion before flooding (b), or as a bolus dose after flooding (c) resulted in significant decreases in both DLW (p < 0.05) (a), whose values were no longer different from controls, and Alb. Sp. (p < 0.001), whose values remained higher than in controls (p < 0.05). Closed circles with error bars indicate means ± SEM.

Examples of pressure-volume curves of control rats and rats instilled with saline and instilled with saline + PFC are shown in Figure 4. These curves were obtained under quasistatic conditions, in view of the low respiratory frequency (25 breaths/min) used when VT was 32 ml/kg (20). Saline instillation markedly modified the shape of the pressure-volume curve. The initial part of the curve, with a very low slope, corresponded to the increase in airway pressure needed to open the lung. After the opening pressure was reached (lower inflection point), the lungs filled abruptly at almost constant pressure and, later, more progressively. End-inspiratory pressure was displaced to the right. Instillation of PFC reduced the lower inflection point pressure and shifted the whole curve to the left. The combination of these two effects resulted in a decrease in end-inspiratory pressure identical to that observed in controls. The changes in opening pressure and end-inspiratory pressure in rats given saline and PFC are shown in Figure 5. Opening pressure decreased significantly (p < 0.001) but remained higher than in controls (p < 0.01). The end-inspiratory pressure was significantly decreased, to become identical with that of controls. The pattern of distribution of these pressures was similar to that of the indices of permeability edema (Figures 3a and 3b).

View larger version (18K): [in this window] [in a new window]   Figure 4.   Representative pressure-volume curves of the respiratory system of rats ventilated with a VT of 33 ml/kg. Intact animals (controls) had no discernible lower inflection point. Flooding produced an inflection point; additionally, the slope of the linear part of the curve was reduced and end-inspiratory airway pressure increased markedly. PFC reduced the pressure at which the lower inflection point occurred, and normalized the end-inspiratory pressure.

View larger version (35K): [in this window] [in a new window]   View larger version (32K): [in this window] [in a new window]   Figure 5.   Effect of PFC instillation on the lower inflection point of the pressure-volume curve (a) and maximum airway pressure (b) in rats ventilated with a VT of 33 ml/kg. Flooding increased both pressure values (p < 0.001). PFC given as a bolus dose before flooding (a), by slow infusion before flooding (b), or as a bolus dose after flooding (c) significantly decreased both the pressure at the lower inflection point (p < 0.001) and the maximum airway pressure (p < 0.001). The lower inflection point remained higher than in controls (p < 0.01), whereas the maximum airway pressure was normalized by PFC. (***p < 0.001 as compared with flooded rats).

Figure 6 shows the index of microvascular permeability to albumin (Alb. Sp.) plotted against the airway pressure at the lower inflection point in control and test rats. There was a significant positive correlation between Alb. Sp. and this pressure (r = 0.74, n = 59, p < 0.001). Because the relationship appeared to be nonlinear (p < 0.01 by the run test), a segmented correlation was performed. The fit obtained with two joined linear segments was significantly better than that with a single line (F = 7.0, p < 0.01). The slope of the first segment was essentially zero (t = 0.68, p = NS); the two segments joined at a pressure of 14.8 cm H₂O. Among animals given PFC, those in which Alb. Sp. was high had higher inflection-point pressures. Alb. Sp. was also significantly correlated with the end-inspiratory pressure (r = 0.69, p < 0.001), without any departure from linearity.

View larger version (16K): [in this window] [in a new window]   Figure 6.   Correlation between the lower inflection point pressure and Alb. Sp. in controls (open circles), flooded animals (closed circles), and animals given a bolus dose of perflubron before flooding (open squares), a slow infusion before flooding (open triangles), or a bolus dose after flooding (open diamonds). All animals were ventilated with a VT of 33 ml/kg. Segmented regression analysis (see text) revealed that the best fit was obtained with two joined linear segments. The slope of the first segment is essentially zero. Animals in which perflubron reduced the lower inflection point pressure had normal or near-normal values for Alb. Sp. There was a threshold for a pressure value of around 15 cm H2O.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Ventilatory strategy for ARDS patients often takes into account the severity of lung mechanical abnormalities, in order to limit the putative deleterious effects of mechanical ventilation (22). Another approach tries to improve the mechanical properties of the lung. Surfactant administration gave unrewarding results during ARDS (15). PLV with PFC may restore ventilation and gas exchange in collapsed areas by eliminating the gas-liquid interface in the lungs (17). PFCs readily dissolve gas and have low surface tension properties, and may therefore facilitate airway reopening by removing fluid from small to larger bronchi (25). Most studies of PLV have focused on gas exchange in experimentally altered lungs (17). Some have found that PLV reduced lung injury caused by oleic acid (19). This may have been be due to a reduction of the stress applied to the lungs during ventilation by preventing the repeated opening and closing of distal airways (11, 12), or by providing more uniform ventilation and thus reducing local overinflation. This latter effect has not, to our knowledge, been the subject of a specific study.

We have shown in this study that ventilation that is not (24 ml/kg VT), or is minimally (32 ml/kg VT) deleterious to the lungs results in severe changes in microvascular permeability when the lungs are instilled with saline. We evaluated lung microvascular permeability by two means: DLW and Alb. Sp. The reduction of Alb. Sp. by PFC was considerable, indicating that lung injury was markedly reduced. Changes in DLW (in both directions; Figures 3a and 3b) were not as impressive as those of Alb. Sp. because even in the case of pure plasma extravasation, the increase in DLW will be minor with respect to the normal DLW. Saline instillation is a simple way of mimicking severe alveolar flooding. It may also produce other effects, such as surfactant inactivation. Surfactant inactivation has been reported during alveolar edema (26), and may facilitate the development of pulmonary edema (27). However, reduction of ventilable lung volume, with resulting overinflation of open zones, was probably the predominant primary alteration under our study conditions, because saline instillation did not affect microvascular permeability or the development of edema when VT was within the normal range (Figure 2). The changes in the pressure-volume curve were consistent with this interpretation. The lower inflection point indicated the presence of a pressure threshold above which zones excluded from ventilation, because of collapse or airway closure, were massively recruited (Figures 4 and 5). The reduction of aerated lung volume probably persisted at higher volumes, increasing the end-inspiratory pressure (Figures 4 and 5).

We previously described a positive interaction between mechanical ventilation and existing mild lung abnormalities in an -naphthylthiourea (ANTU)-induced model of lung injury (20). In the present study, we used a model in which aerated lung volume was undoubtedly decreased (by alveolar flooding or atelectasis). Saline instillation and mechanical ventilation interacted positively, as before (20), but at substantially lower tidal volumes. Tidal volumes of 16 ml/kg or less did not influence Alb. Sp. in rats instilled with saline. Alb. Sp. increased slightly when VT was 24 ml/kg, but a VT of 32 ml/kg caused marked changes in microvascular permeability. The two-way ANOVA disclosed a very significant interaction between flooding and VT, which confirmed that lungs with reduced recruitable volume are more prone to edema induced by mechanical ventilation. The extreme susceptibility of such lungs to a VT of 32 ml/kg suggested that this model may be very sensitive to overinflation, and suitable for evaluating the effect of PFC administration in the presence of alveolar edema. Short periods of mechanical ventilation were studied to minimize interanimal differences. Indeed, during longer periods, the amount of saline present in air spaces might have differed between animals because of different rates of saline absorption by the alveolar epithelium (28). This might have produced additional and confounding variations.

We speculated that instillation of a small volume (3.3 ml/ kg) of PFC might allow easier opening of small bronchi (e.g., by replacing saline by PFC menisci), or might redistribute alveolar fluid (17), thus counteracting the reduction of aerated lung volume resulting from alveolar flooding. We found that instillation of PFC produced a significant decrease in microvascular permeability and in the severity of edema, as indicated by the changes in Alb. Sp. and DLW. However, values of Alb. Sp. were scattered; some were identical to those of controls, whereas others remained as high as in animals that were flooded but not given PFC (Figure 3b). Analysis of the respiratory system pressure-volume curve explained these results.

There were remarkable changes in the airway-pressure volume curves in most of the rats instilled with PFC (Figures 4 and 5). The pressure at which the lower inflection point occurred decreased, reflecting easier opening of the lungs. End-inspiratory pressure was reduced to such an extent that it did not differ significantly from that of controls. This suggests that PFC instillation recruited the lung along the entire tidal volume, or that there was a marked improvement in distal lung mechanical properties, perhaps through the replacement of air-water interfaces having a high surface tension by PFC-air interfaces having a lower surface tension (17). A reduction of surface tension may affect lung recoil by changing lung microarchitecture (29) in such a way that tissue stretch is less for a given lung volume. This reduction of tissue stress may reduce the severity of ventilator-induced lung injury (30).

In this study, the alleviation of alterations in permeability was correlated with the decrease in the opening and end- inspiratory airway pressures. Whereas this improvement appeared to be proportional to the reduction of end-inspiratory pressure (probably an index of tissue stretch under these conditions), there was a clear threshold of opening pressure (about 15 cm H₂O) below which no significant change in permeability occurred (Figure 6). In some instances, instillation of PFC was not successful in reducing the opening pressure below the threshold value. The alterations in permeability were then not corrected. This failure in some instances of PFC to counteract the effect of saline instillation did not depend on the manner of instillation of PFC or on whether PFC was given before or after saline. The distribution of PFC in the lung may have been uneven, and this may correspond to the failure, as just mentioned of instillation of PFC to reduce the airway opening pressure below the threshold. By contrast, when PFC produced a decrease in the opening pressure, the alterations in permeability caused by saline instillation were almost completely corrected (the cluster of flooding + PFC data with opening pressures below the threshold did not differ from the Alb. Sp. of animals that were not instilled with saline). The DLWs in animals that were given PFC remained slightly increased but did not differ significantly from those of controls (without saline instillation). The Alb. Sp. in animals ventilated at a VT of 32 ml/kg and instilled with PFC (Figure 3b) was similar to that of rats instilled with saline and ventilated with a VT of 16 to 24 ml/kg (Figure 2b). Thus, in terms of reducing alterations in permeability, instillation of 3.3 ml/kg PFC was equivalent to a reduction in VT of at least 6 ml/kg. Instillation of PFC reduced the mechanical nonuniformity of flooded lungs and probably opposed overinflation of the more compliant, aerated zones. The severity of VILI was therefore reduced.

The appropriate dose of PFC during PLV remains to be determined. Improvement of lung mechanics by low doses of PFC has been reported (18). Moreover, a recent experimental study by Cox and colleagues (31) clearly showed that excessive tidal volumes and large doses of PFC may result in air and PFC leaks. In contrast, a recent experimental study showed that high doses of perflubron (perfluorooctyl bromide) resulted in better lung mechanics and less evidence of barotrauma than did lower doses when it was administered as a rescue agent after surfactant treatment during mechanical ventilation of preterm lambs (32). Our study shows that small doses of PFC can considerably reduce the harmful effects of mechanical ventilation of lungs with reduced aerated volume. This may be important in clinical practice for several reasons. First, it may help in designing safer clinical studies. Indeed, although encouraging results have been obtained in newborns during preliminary clinical trials (33, 34), there were many air leaks in early clinical trials with PLV in adults (35), precluding the demonstration of a beneficial effect of this technique. Second, the efficacy of small doses of PFC suggests that the cost of PLV may be diminished by reducing the doses of PFC requested. Additionally, the effectiveness of PLV in reducing VILI may be predicted by analyzing the respiratory pressure- volume curve, which is easy to do at the bedside. However, caution is advisable when dealing with the clinical correlates of experimental studies, and definite conclusions cannot be drawn until ongoing clinical trials of PFC are completed.