Effect of Prone Position on Regional Shunt, Aeration, and Perfusion in Experimental Acute Lung Injury
http://www.100md.com
美国呼吸和危急护理医学 2005年第8期
Department of Anesthesia and Critical Care and Department of Medicine (Pulmonary and Critical Care Unit), Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts
Clinic of Anesthesiology and Intensive Care Medicine, University Clinic Carl Gustav Carus, Dresden University of Technology, Dresden, Germany
ABSTRACT
Rationale: The prone position is used to improve gas exchange in patients with acute respiratory distress syndrome. However, the regional mechanism by which the prone position improves gas exchange in acutely injured lungs is still incompletely defined. Methods: We used positron emission tomography imaging of [13N]nitrogen to assess the regional distribution of pulmonary shunt, aeration, perfusion, and ventilation in seven surfactant-depleted sheep in supine and prone positions. Results: In the supine position, the dorsal lung regions had a high shunt fraction, high perfusion, and poor aeration. The prone position was associated with an increase in lung gas content and with a more uniform distribution of aeration, as the increase in aeration in dorsal lung regions was not offset by loss of aeration in ventral regions. Consequently, the shunt fraction decreased in dorsal regions in the prone position without a concomitant impairment of gas exchange in ventral regions, thus leading to a significant increase in the fraction of pulmonary perfusion participating in gas exchange. In addition, the vertical distribution of specific alveolar ventilation became more uniform in the prone position. A biphasic relation between regional shunt fraction and gas fraction showed low shunt for values of gas fraction higher than a threshold, and a steep linear increase in shunt for lower values of gas fraction. Conclusion: In a surfactant-deficient model of lung injury, the prone position improved gas exchange by restoring aeration and decreasing shunt while preserving perfusion in dorsal lung regions, and by making the distribution of ventilation more uniform.
Key Words: adult respiratory distress syndrome; emission-computed tomography; nitrogen isotopes; prone position; pulmonary gas exchange
Despite increasing use of the prone position as a means to improve gas exchange in patients with acute respiratory distress syndrome (ARDS) (1, 2), few studies have investigated the regional mechanism of this improvement in acutely injured lungs (3, 4). Using single-photon emission computed tomography in an oleic acid model of lung injury, Lamm and coworkers (4) showed that the prone position was associated with a narrower distribution of the ventilation-to-perfusion ratio and with an increase in the relative (i.e., mean-normalized) ventilation-to-perfusion ratio in dorsal lung regions. However, because regional shunt could not be measured directly by single-photon emission computed tomography, whether the increase in relative ventilation-to-perfusion ratio in dorsal regions corresponded to reversal of shunt and, more importantly, the magnitude of the associated improvement of regional gas exchange could not be determined. Furthermore, single-photon emission computed tomography did not allow assessment of regional aeration. Combining measurement of regional aeration with measurement of regional shunt and perfusion is important for two reasons. First, it allows determination of whether the improvement of gas exchange in the prone position is due to restored aeration to regions of shunt or to redistribution of perfusion away from these regions. In this respect, it has been suggested that the regional mechanism by which the prone position improves gas exchange may bear prognostic significance and that the prone position increases survival of patients with ARDS if the improvement of gas exchange on turning prone is associated with restored aeration to regions of shunt rather than with redistribution of perfusion away from these regions (5). Therefore, further insights into the effect of the prone position on regional shunt, aeration, and perfusion can have important clinical implications. Second, by combining measurements of regional aeration and regional shunt, it is possible to determine the functional consequence of the change in distribution of lung density associated with prone positioning (6).
Positron emission tomography (PET) imaging of infused [13N]nitrogen (13N2) is a noninvasive technique that allows the in vivo assessment of regional gas exchange and the direct quantification and localization of intrapulmonary shunt (7eC9). In this study, we used PET imaging of infused and inhaled 13N2 to assess the effect of the prone position on the regional distribution of pulmonary perfusion, shunt, aeration, and ventilation in a saline lavage model of acute lung injury. We investigated whether, in acute lung injury caused by surfactant depletion, the prone position improves gas exchange by restoring aeration and decreasing shunt in dorsal lung regions or by redistributing perfusion away from these regions. A preliminary analysis of some of the results of this study was previously reported in the form of an abstract (10).
METHODS
Experimental Procedures
Seven sheep (22.4 ± 2.9 kg) were anesthetized, paralyzed, and mechanically ventilated (tidal volume, 8 ml/kg; inspiratory-to-expiratory ratio, 1:2; fraction of inspired oxygen, 1.0; and positive end-expiratory pressure, 5 cm H2O). Respiratory rate was adjusted to normocapnia (19.7 ± 2 breaths/minute; PaCO2 = 39.9 ± 3.9 mm Hg) and was not modified during the study. An arterial line and a Swan-Ganz catheter were inserted and a jugular line was placed for infusion of 13N2 in saline solution. The animals were then suspended in a cylindric tube, which allowed them to remain positioned within the field of view of the PET scanner while body position was changed by rotation of the tube. This suspension system allowed the abdomen to be unsupported in the prone position.
After collection of control physiologic data in both positions, lung lavage was performed by instillation and drainage of warm saline (30 ml/kg) and repeated every 15 minutes until the PaO2 fell below 100 mm Hg in the supine position (Figure 1). After stable injury, defined as a PaO2 below 200 mm Hg that changed by less than 10% at two consecutive 15-minute intervals, was obtained, a set of physiologic data was collected in the supine position. The animals were then transported to the PET suite, where a recruitment maneuver (airway pressure of 40 cm H2O for 30 seconds) was performed and the previous mechanical ventilation settings were resumed. After 15 minutes, a 10-minute PET transmission scan was taken, followed by a 4-minute 13N2eCsaline infusion and 4-minute 13N2 inhalation PET emission scans. Physiologic data were collected during the last 2 minutes of the transmission scan. After the last PET scan, the animal was turned to the opposite body position, and the recruitment maneuver, imaging sequence, and data collection were repeated.
PET Imaging
The PET scanner imaged 15 contiguous, 6.5-mm-thick slices of thorax.
13N2eCsaline bolus infusion (perfusion) scan.
A bolus of 13N2 dissolved in saline was infused through the jugular catheter at the start of a 60-second apnea performed at a constant airway pressure corresponding to the mean airway pressure measured during positive-pressure ventilation (7, 9, 11, 12). Because of the low partition coefficient of nitrogen (water/air, 0.015), the kinetics of 13N2 during apnea differs between regions that are perfused and aerated, in which virtually all 13N2 diffuses into the alveolar airspace at first pass and accumulates in proportion to regional perfusion (12), and regions that contain alveolar units that are perfused but not aerated (i.e., "shunting"). In the latter regions, the 13N2 kinetics (Figure 2) shows a peak, related to total regional perfusion, followed by a plateau proportional to perfusion to aerated units within the region, because shunting units do not retain 13N2 during apnea (Figure 3). For each of eight horizontal regions of interest (ROIs) of lung (i = 1, ..., 8), regional perfusion (i), expressed as a fraction of total perfusion to the imaged lung, and regional shunt fraction (FS,PETi) were derived by curve fitting the kinetics of 13N2 with a model (7). The fraction of imaged pulmonary perfusion shunted within each ROI was calculated as Si = i · FS,PETi.
Regional specific alveolar ventilation of perfused units (sVAi, i.e., alveolar ventilation per unit of lung gas volume) was calculated as the reciprocal of the time constant derived from the initial 30 seconds of the tracer washout curve, obtained after mechanical ventilation was resumed (13).
13N2 inhalation (gas fraction) scan.
After equilibration of the alveolar gas with inhaled 13N2 (Figure 3), regional gas fraction (FGASi) was calculated by dividing mean regional specific tracer activity (i.e., activity per unit of lung volume) by the activity of a 1-ml sample of inhaled gas (9). Regional lung volume (VLi) was calculated as the product of the number of voxels within each ROI by the volume of a voxel. Regional gas volume was calculated as VGASi = VLi · FGASi and regional tissue volume was calculated as VTISi = VLi eC VGASi. An estimate of regional absolute alveolar ventilation was calculated as VAi = sVAi · VGASi.
Statistical Analysis
Data were analyzed by two-way analysis of variance. Statistical significance is reported only for comparisons between prone and supine positions after injury. Linear regression was used to estimate associations between variables and to compute vertical gradients (positive gradients indicate an increase from the dependent to the nondependent regions). Data are presented as means ± SD.
RESULTS
Global Physiologic Variables
Cardiovascular and respiratory variables measured in the supine position after injury and during imaging were not significantly different (see Table E1 in the online supplement), thus supporting the physiologic stability of the injury model throughout the experiment.
After injury, PaO2 was markedly lower, PaCO2 higher, and pH lower in the supine than in the prone position (Table 1). Turning prone was associated with a significant decrease in mean pulmonary arterial pressure, while cardiac output, heart rate, and mean arterial pressure were not significantly affected by the body position change. Pulmonary capillary wedge pressure was slightly lower in the prone position. Inspiratory plateau airway pressure was substantially lower in the prone than in the supine position (Table 1).
Regional Gas Content
The total, voxel-by-voxel mean-normalized variance of gas fraction, COVtot2 (see the online supplement), was almost fivefold greater in the supine than in the prone position (0.60 ± 0.24 vs. 0.13 ± 0.09; p < 0.001). The systematic component of the variance in the vertical direction, COVv2, accounted for 62 ± 21 and 29 ± 14% of COVtot2 in the supine and prone positions, respectively. In contrast, the systematic component of the variance in the craniocaudal direction, COVcc2, accounted for a small fraction of the total variance in both body positions (4 ± 4% in the supine position and 6 ± 12% in the prone position).
Both the gas volume and the gas fraction of the imaged lung were higher in the prone than in the supine position (384 ± 102 vs. 286 ± 88 ml, p < 0.05; and 0.45 ± 0.09 vs. 0.32 ± 0.10, p < 0.001), whereas the volume of imaged lung was not different between the two positions. There was significant interregional variation in gas fraction (effect of ROI: p < 0.01), and the anatomic (i.e., dorsal to ventral) distribution profile of regional gas fraction (Figure 4) was different between the supine and prone positions (interaction between ROI and body position: p < 0.001). Accordingly, regional gas fraction showed a steeper vertical gradient favoring nondependent lung regions in the supine position (FGASi gradient, 0.035 ± 0.011/cm in the supine position vs. 0.011 ± 0.014/cm in the prone position; p < 0.05).
Taken together, these results are consistent with a more uniform distribution of aeration in the prone than in the supine position.
Global and Regional Shunt Fraction
Consistent with the reduction of PaO2, shunt fraction measured from blood gases (FSO2) was significantly higher in the supine than in the prone position (Table 1). The global shunt fraction of the imaged lung, estimated by PET (FS,PET), was also higher in the supine than in the prone position (0.50 ± 0.13 vs. 0.14 ± 0.14; p < 0.001). In neither position was the difference between FS,PET and FSO2 statistically significant, and FS,PET showed good correlation with FSO2 (FS,PET = 0.83 · FSO2 + 0.06; r = 0.87, p < 0.001).
Analysis of variance showed a significant difference in regional shunt fraction (FS,PETi) between body positions (p < 0.001) and between ROIs (p < 0.001). In both positions the regional shunt fraction tended to decrease from dependent to nondependent ROIs (i.e., FS,PETi showed a negative vertical gradient). However, the vertical gradient of regional shunt fraction was significantly steeper in the supine than in the prone position (FS,PETi gradient, eC0.054 ± 0.018/cm in the supine position vs. eC0.013 ± 0.006/cm in the prone position; p < 0.01), indicating that the vertical dependence of shunt fraction was more pronounced in the supine than in the prone position. Consistent with these gradients, analysis of variance showed a significant interaction between the effect of body position and the effect of ROI (p < 0.001), which indicates that the anatomic distribution profile of regional shunt was different between the two body positions.
Taken together, these results indicate that the prone position was associated with an overall reduction of regional shunt and with a more uniform distribution of the residual shunt among the ROIs.
Regional Perfusion
The anatomic distributions of regional perfusion (i) and shunt flow were also affected by body position (Figure 5). There was significant interregional variation of perfusion (effect of ROI: p < 0.001), and the anatomic distribution profile of perfusion was different between the supine and prone positions (interaction between ROI and body position: p < 0.001). In the supine position, the dorsal (i.e., dependent) half of the ROIs received the greatest fraction of perfusion. In the prone position, the dorsal half of the ROIs still received most of the perfusion. These results were consistent with measurements of the perfusion gradient, which showed reversal of the nondependent-to-dependent gradient in the prone position, as evidenced by the opposite signs of the gradients in the prone position (positive gradient) compared with the supine position (negative gradient; see STATISTICAL ANALYSIS, in METHODS, in the online supplement for interpretation of the sign of the gradient). However, the absolute value of the vertical gradient of regional perfusion normalized by relative regional tissue volume, which represents a measure of the steepness of the vertical gradient independent of its direction and of the relative amount of lung tissue within each ROI, was significantly lower in the prone position (0.040 ± 0.024/cm) than in the supine position (0.098 ± 0.030/cm; p < 0.01), indicating some redistribution of perfusion toward ventral regions in the prone position.
As a result of the high shunt fraction and high perfusion in dependent regions in the supine position, fractional pulmonary shunt flow (Si) was also highest in the dorsal half of the ROIs (i.e., a large fraction of pulmonary perfusion went to shunting units in dorsal, dependent ROIs). In contrast, in the prone position, the dependent (i.e., ventral) half of the ROIs contributed only modestly to shunt flow (Figure 5). Accordingly, fractional shunt flow was significantly lower in the prone position (effect of body position: p < 0.001) and its anatomic distribution was different between the supine and prone positions (interaction between ROI and body position: p < 0.001), as fractional shunt flow decreased markedly in dorsal regions in the prone position.
Taken together, these results suggest that, despite some degree of perfusion redistribution toward ventral regions in the prone position, a substantial fraction of perfusion was maintained in dorsal regions. Because in the prone position the dorsal half of the ROIs had a major reduction in regional shunt fraction, while still receiving most of the perfusion, this position was associated with a considerable increase in the fraction of perfusion contributing to gas exchange.
Relation between Regional Shunt and Gas Fraction
When regional data from each animal for both body positions were pooled, an inverse relation between average regional values of FS,PETi and FGASi became apparent (Figure 6A). The main effect of the prone position was to shift the data points corresponding to the most dorsal ROIs from the low-gas and high-shunt fraction to the high-gas and low-shunt fraction portion of the relation. When plots of FS,PETi versus FGASi were analyzed for each animal, biphasic behavior was identified in five of them: values of shunt fraction were low for regional gas fraction higher than a threshold (FGASi 0.3) and increased linearly for values of gas fraction lower than that threshold (Figure 6B). In these animals, when the equation FS,PET = 1 eC FGAS/F0 was fitted to regional data points with FS,PETi > 0.2, the intercept on the FGAS axis was F0 = 0.35 ± 0.10. In the remaining two animals, biphasic behavior could not be identified (Figure 6C).
Regional Specific Alveolar Ventilation
Specific alveolar ventilation of the imaged lung was similar in the supine and prone positions (0.040 ± 0.012 and 0.039 ± 0.012 secondeC1, respectively). Regional specific alveolar ventilation (sVAi) showed significant interregional variation (effect of ROI: p < 0.001). Although the linear component of the vertical variation of specific ventilation was not significant in either position, the vertical distribution profile of specific ventilation was more uniform in the prone than in the supine position (Figure 7). Accordingly, when the ROIs were ordered by vertical location (i.e., nondependent to dependent), analysis of variance showed a significant interaction between the effect of ROI and body position (p < 0.01).
DISCUSSION
The main finding of this study is that the improvement of gas exchange in the prone position was associated with restored aeration and reduced shunt in dorsal lung regions, without a concomitant reduction of aeration or increase in shunt in ventral regions. The improvement of gas exchange was further enhanced by the fact that perfusion to dorsal regions remained substantial in the prone position, thus allowing these regions to take advantage of their restored aeration and leading to an increase in dorsal, and total, gas-exchanging pulmonary blood flow.
Limitations of the Study
The field of view of the PET scanner did not encompass the entire lung. In sheep with similar weight to those of this study, we previously estimated that the imaged lung corresponded to about 70% of the entire lung (14). Although the field of view did not include the most cranial and the most caudal, subdiaphragmatic regions, the internal consistency between values of shunt fraction of the imaged lung, derived by PET, and values of global shunt fraction suggests that the behavior of the imaged lung was representative of the behavior of the entire lung. We also acknowledge that imaging the same cross-section of thorax in both positions did not necessarily ensure that the same cross-section of lung was imaged, as the lung could expand or compress along the craniocaudal axis due to shift of the diaphragm. Although the limitation of PET to resolve fine anatomic markers did not allow spatial coregistration of the supine and prone images, the length of the imaged cross-section of lung ( 10 cm) was much larger than the possible lung displacement along the craniocaudal axis. This, as well as the small contribution of the craniocaudal direction to the variance of gas fraction, suggest that the effect of this potential artifact was probably small.
We used the saline lavage model, which represents a highly recruitable model of acute lung injury (15). After lung lavage, the damage to alveolar epithelial, endothelial, and perivascular cells is probably less than that of other lung injury models, such as the oleic acid or the endotoxin infusion models (16, 17), in which the primary insult is delivered from the pulmonary vascular side. Nonetheless, lung lavage creates interstitial and intraalveolar edema and causes impaired gas exchange by its detrimental effect on surfactant (18eC20). Although the pathogenetic mechanism is different, the severity of the impairment of gas exchange that was obtained after lung lavage is comparable to that of other experimental models of acute lung injury and is not the result of a specific pattern of disruption of ventilation-to-perfusion distribution peculiar to this model (21). Therefore, this model reflects the pathophysiology of those cases of ARDS that are characterized predominantly by alveolar instability and collapse, with high potential for alveolar recruitment, rather than by alveolar consolidation or flooding.
We analyzed the regional data in eight horizontal ROIs. We recognize that such an approach neglects intraregional heterogeneity. However, in this model of lung injury, the contribution of the systematic component of the variance in the vertical direction to the total variance of gas fraction was substantially greater than the contribution of the component in the craniocaudal direction. Therefore, analyzing the data by horizontal ROIs allowed us to capture the main systematic trend in regional aeration.
Effect of the Prone Position on Global Gas Exchange and Hemodynamics
After injury, the prone position was associated with improved gas exchange compared with the supine position, in line with the results of previous studies (3, 4, 22, 23). Whereas all these studies consistently reported an improvement of PaO2 and a decrease in global shunt fraction in the prone position, our results also showed an improvement of PaCO2. Because body temperature was not different between the two positions and tidal volume and respiratory rate were kept constant, the improvement of PaCO2 suggests that ventilation was more efficient in the prone position. This is consistent with our finding that regional specific alveolar ventilation was more uniform in the prone position, indicating a more homogeneous turnover of alveolar gas of perfused units in this position. Furthermore, in the prone position, the vertical distribution profile of absolute alveolar ventilation, estimated from the product of specific ventilation and gas volume, followed closely the vertical profile of relative perfusion (compare Figure 5 with Figure E1 in the online supplement). This finding suggests improved ventilation-to-perfusion matching in the prone position (4) and, because minute ventilation and cardiac output were not different between the supine and prone positions, it would explain the reduction of PaCO2. It has been shown that patients with ARDS who responded to a trial of prone positioning with a reduction of PaCO2 had increased survival at 28 days, and it has been hypothesized that in these patients the improvement of PaO2 in the prone position was due to alveolar recruitment rather than to redistribution of perfusion away from regions of shunt (5). Our data are compatible with this hypothesis by showing that, when the main effect of the prone position was to restore aeration, while preserving perfusion, to shunting units in dorsal regions, the increase in PaO2 was accompanied by a significant decrease in PaCO2 and by evidence of improved ventilation-to-perfusion matching. However, this reduction of PaCO2 cannot be considered a specific marker of alveolar recruitment, because PaCO2 should also decrease if the improvement of ventilation-to-perfusion matching were due to redistribution of perfusion away from regions of shunt.
The finding that pulmonary arterial pressure was higher in the supine than in the prone position after injury is consistent with the effect of higher PaCO2, lower pH, and lower PaO2 on pulmonary vascular tone. Pulmonary capillary wedge pressure was slightly lower in the prone than in the supine position. It is possible, however, that by setting the zero value of the pressure transducer at midchest level, we slightly underestimated left ventricular filling pressure in the prone position, as the prone position displaces the left ventricle inferiorly (23).
Effect of the Prone Position on Regional Perfusion, Shunt, and Aeration
The observation that substantial perfusion was preserved in dorsal lung regions in the prone position is consistent with previous observations in oleic acideCinduced lung injury (3, 24, 25) and in healthy animals (26eC30), and it could be due to higher vascular conductance in these regions (31, 32). However, the fact that the vertical perfusion gradient was not fully reversed in the prone position suggests some degree of perfusion redistribution toward ventral regions, consistent with results obtained in oleic acideCinduced lung injury (25). We previously reported marked interindividual variability in the extent to which perfusion redistributed toward ventral regions in prone normal humans (12). Whether this variability is present also in patients with ARDS and whether patients who maintain substantial dorsal perfusion are those who respond favorably to prone positioning remains to be investigated. In this respect, it is important to emphasize that the extent to which inferences on the importance of structural factors in favoring perfusion to dorsal regions can be extrapolated to humans is unclear. Indeed, whereas it would appear teleologically meaningful for quadrupeds, who are normally prone, to develop mechanisms that, by favoring perfusion to dorsal lung regions, render perfusion more homogeneous, we cannot envision, a priori, a physiologic advantage of diverting perfusion toward dorsal regions in bipeds.
Critical to the improvement of gas exchange was the restored aeration to dorsal lung regions. This led to an increase in lung gas volume and to a more uniform distribution of gas fraction, as the gain in aeration in dorsal regions was not offset by loss of aeration in ventral regions. This finding is consistent with the presence of a smaller vertical pleural pressure gradient in the prone position (33). Although improvement of gas exchange in the prone position can occur in the absence of an increase in functional residual capacity (22, 33), the increase in lung gas volume likely contributed to the dramatic improvement of oxygenation that we observed. A potential contributor to the improved aeration of dorsal regions was the lack of abdominal support in the prone position. Unsupporting the abdomen likely decreased intraabdominal pressure (potentially creating negative intraabdominal pressure), reduced the pressure gradient across the diaphragm, and promoted alveolar recruitment preferentially in the dorsal regions of this surfactant-deficient model of lung injury. However, the prone position has also been shown to improve gas exchange when the abdomen was resting on a surface (33eC35). Furthermore, the effect of unsupporting the abdomen in the prone position may depend on the pathophysiology of the injury: unsupporting the abdomen may promote alveolar recruitment and more uniform lung aeration in the presence of atelectasis and alveolar collapse, but it may be ineffective in the presence of alveolar consolidation or flooding, as in the oleic acid model (36). Although our method of suspending the animal in the prone position is not applicable to humans, and decreased intraabdominal pressure may be difficult to achieve when patients with ARDS are positioned prone, technical developments have led to critical care beds that can be used to pronate patients and that could potentially incorporate features to leave the abdomen unsupported in the prone position. In this respect, studies in anesthetized humans undergoing surgical procedures in the prone position suggest that respiratory mechanics are optimized when the abdomen is unsupported (37, 38). Future studies, however, will be needed to clarify the relative contribution of unsupporting the abdomen to the improvement of gas exchange in the prone position.
The beneficial effect that the improvement of regional aeration had on regional gas exchange was apparent from the relation between regional shunt fraction and regional gas fraction. This relation showed that the effect of the prone position was to restore aeration to dorsal regions, thus shifting these regions to a more favorable portion of the shunt-versus-gas fraction relation. Interestingly, in five animals we observed biphasic behavior, with low shunt fraction in regions with gas fraction higher than a critical threshold and a steep linear increase in shunt in regions with gas fraction lower than this threshold. This threshold (F0 = 0.35) was lower than the gas fraction of the normal lung ( 0.6 [27, 39]). This finding suggests that, in these animals, even poorly aerated regions maintained their ability to exchange gas. Assuming that the shunt fraction is proportional to the fraction of derecruited units, the linear increase in shunt for a gas fraction lower than this threshold is consistent with a "quantal" loss of gas-exchanging units, so that the decrease in aeration of a region is absorbed only by a certain number of its units that become atelectatic and lose the ability to exchange gas, whereas the remaining units preserve their aeration and, accordingly, their gas exchange. Therefore, the biphasic relation between regional shunt fraction and gas fraction would suggest the presence of a gas fraction threshold for the development of subregions that collapse or flood and begin to shunt. The lack of biphasic behavior in the other two animals could have been related to intraregional differences in the degree of surfactant depletion, so that alveolar collapse started to occur at a higher gas fraction in units with more severe impairment of surfactant. Supporting this hypothesis is the fact that these two animals exhibited the highest contribution of the component of variance in the craniocaudal direction to the total variance of gas fraction, suggesting that the horizontal ROIs used to analyze their data included units with different conditions.
Importantly, the improved aeration and gas exchange in dorsal regions in the prone position were obtained at a lower airway pressure than in the supine position. The prone position might thus be an effective means to promote alveolar recruitment in dorsal regions, while limiting airway pressure and ventilator-associated lung injury (17, 40).
In summary, we demonstrated that, in a surfactant-deficient model of ARDS, the prone position reduced shunt in dorsal lung regions. The reduction of shunt was associated with an increase in pulmonary gas content and with a more homogeneous distribution of aeration, which placed dorsal lung regions on a more favorable portion of the shunt-versus-gas fraction relation without a concomitant, substantial increase in shunt in ventral regions. Preservation of perfusion to dorsal regions contributed to the improvement of global shunt fraction in the prone position. The prone position with unsupported abdomen appears to be a useful means to promote alveolar recruitment and improve global and regional gas exchange, while limiting airway pressure, when the main pathophysiologic alteration is loss of surfactant and the associated alveolar derecruitment.
Acknowledgments
The authors thank E. Kuhlisch, for statistical advice; S. B. Weise, for technical assistance with image acquisition and processing; J. A. Correia, Ph.D., W. M. Bucelewicz, and D. F. Lee, B.S., for preparation of the radioisotope; O. Syrkina, M.D., for assistance with animal preparation; and T. Schroeder, M.Eng., for insightful comments.
This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org
In only two animals, COVv2 accounted for less than 60% of COVtot2 in the supine position. Excluding them from this analysis raised the contribution of COVv2 to 73 ± 8% in the supine position and to 30 ± 12% in the prone position, and lowered the contribution of COVcc2 to 2.4 ± 1% (supine) and 1.7 ± 1% (prone).
REFERENCES
Gattinoni L, Tognoni G, Pesenti A, Taccone P, Mascheroni D, Labarta V, Malacrida R, Di Giulio P, Fumagalli R, Pelosi P, et al. Effect of prone positioning on the survival of patients with acute respiratory failure. N Engl J Med 2001;345:568eC573.
Papazian L, Bregeon F, Gaillat F, Thirion X, Gainnier M, Gregoire R, Saux P, Gouin F, Jammes Y, Auffray JP. Respective and combined effects of prone position and inhaled nitric oxide in patients with acute respiratory distress syndrome. Am J Respir Crit Care Med 1998;157:580eC585.
Wiener CM, Kirk W, Albert RK. Prone position reverses gravitational distribution of perfusion in dog lungs with oleic acid-induced lung injury. J Appl Physiol 1990;68:1386eC1392.
Lamm WJ, Graham MM, Albert RK. Mechanism by which the prone position improves oxygenation in acute lung injury. Am J Respir Crit Care Med 1994;150:184eC193.
Gattinoni L, Vagginelli F, Carlesso E, Taccone P, Conte V, Chiumello D, Valenza F, Caironi P, Pesenti A. Decrease in PaCO2 with prone position is predictive of improved outcome in acute respiratory distress syndrome. Crit Care Med 2003;31:2727eC2733.
Gattinoni L, Pelosi P, Vitale G, Pesenti A, D'Andrea L, Mascheroni D. Body position changes redistribute lung computed-tomographic density in patients with acute respiratory failure. Anesthesiology 1991;74:15eC23.
Galletti GG, Venegas JG. Tracer kinetic model of regional pulmonary function using positron emission tomography. J Appl Physiol 2002;93:1104eC1114.
Willey-Courand DB, Harris RS, Galletti GG, Hales CA, Fischman A, Venegas JG. Alterations in regional ventilation, perfusion, and shunt after smoke inhalation measured by PET. J Appl Physiol 2002;93:1115eC1122.
Musch G, Harris RS, Vidal Melo MF, O'Neill KR, Layfield JDH, Winkler T, Venegas JG. Mechanism by which a sustained inflation can worsen oxygenation in acute lung injury. Anesthesiology 2004;100:323eC330.
Richter T, Harris RS, Vidal Melo MF, Musch G, Layfield JDH, Venegas JG. Prone position virtually eliminates regions of shunt in surfactant depleted sheep . Am J Respir Crit Care Med 2002;165:A784.
O'Neill K, Venegas JG, Richter T, Harris RS, Layfield JDH, Musch G, Winkler T, Vidal Melo MF. Modeling kinetics of infused 13NNeCsaline in acute lung injury. J Appl Physiol 2003;95:2471eC2484.
Musch G, Layfield JDH, Harris RS, Vidal Melo MF, Winkler T, Callahan RJ, Fischman AJ, Venegas JG. Topographical distribution of pulmonary perfusion and ventilation, assessed by PET in supine and prone humans. J Appl Physiol 2002;93:1841eC1851.
Vidal Melo MF, Harris RS, Layfield D, Musch G, Venegas JG. Changes in regional ventilation after autologous blood clot pulmonary embolism. Anesthesiology 2002;97:671eC681.
Vidal Melo MF, Layfield D, Harris RS, O'Neill K, Musch G, Richter T, Winkler T, Fischman AJ, Venegas JG. Quantification of regional ventilationeCperfusion ratios with positron emission tomography. J Nucl Med 2003;44:1982eC1991.
Downie JM, Nam AJ, Simon BA. PressureeCvolume curve does not predict steady-state lung volume in canine lavage lung injury. Am J Respir Crit Care Med 2004;169:957eC962.
Schuster DP. ARDS: clinical lessons from the oleic acid model of acute lung injury. Am J Respir Crit Care Med 1994;149:245eC260.
Brigham KL, Meyrick B. Endotoxin and lung injury. Am Rev Respir Dis 1986;133:913eC927.
Berggren P, Lachmann B, Curstedt T, Grossmann G, Robertson B. Gas exchange and lung morphology after surfactant replacement in experimental adult respiratory distress syndrome induced by repeated lung lavage. Acta Anaesthesiol Scand 1986;30:321eC328.
Lachmann B, Jonson B, Lindroth M, Robertson B. Modes of artificial ventilation in severe respiratory distress syndrome: lung function and morphology in rabbits after wash-out of alveolar surfactant. Crit Care Med 1982;10:724eC732.
Lachmann B, Robertson B, Vogel J. In vivo lung lavage as an experimental model of the respiratory distress syndrome. Acta Anaesthesiol Scand 1980;24:231eC236.
Neumann P, Hedenstierna G. VentilationeCperfusion distributions in different porcine lung injury models. Acta Anaesthesiol Scand 2001;45:78eC86.
Albert RK, Leasa D, Sanderson M, Robertson HT, Hlastala MP. The prone position improves arterial oxygenation and reduces shunt in oleic-acid-induced acute lung injury. Am Rev Respir Dis 1987;135:628eC633.
Broccard AF, Shapiro RS, Schmitz LL, Ravenscraft SA, Marini JJ. Influence of prone position on the extent and distribution of lung injury in a high tidal volume oleic acid model of acute respiratory distress syndrome. Crit Care Med 1997;25:16eC27.
Richard JC, Janier M, Lavenne F, Berthier V, Lebars D, Annat G, Decailliot F, Guerin C. Effect of position, nitric oxide, and almitrine on lung perfusion in a porcine model of acute lung injury. J Appl Physiol 2002;93:2181eC2191.
Schuster DP, Haller J. Effects of body position on regional pulmonary blood flow during acute pulmonary edema in dogs: a positron emission tomography study. J Crit Care 1991;6:19eC28.
Glenny RW, Lamm WJ, Albert RK, Robertson HT. Gravity is a minor determinant of pulmonary blood flow distribution. J Appl Physiol 1991;71:620eC629.
Treppo S, Mijailovich SM, Venegas JG. Contributions of pulmonary perfusion and ventilation to heterogeneity in A/ measured by PET. J Appl Physiol 1997;82:1163eC1176.
Walther SM, Domino KB, Glenny RW, Hlastala MP. Pulmonary blood flow distribution in sheep: effects of anesthesia, mechanical ventilation, and change in posture. Anesthesiology 1997;87:335eC342.
Beck KC, Vetterman J, Rehder K. Gas exchange in dogs in the prone and supine positions. J Appl Physiol 1992;72:2292eC2297.
Won C, Chon D, Tajik J, Tran BQ, Robinswood GB, Beck KC, Hoffman EA. CT-based assessment of regional pulmonary microvascular blood flow parameters. J Appl Physiol 2003;94:2483eC2493.
Beck KC, Rehder K. Differences in regional vascular conductances in isolated dog lungs. J Appl Physiol 1986;61:530eC538.
Pelletier N, Robinson NE, Kaiser L, Derksen FJ. Regional differences in endothelial function in horse lungs: possible role in blood flow distribution J Appl Physiol 1998;85:537eC542.
Mutoh T, Guest RJ, Lamm WJE, Albert RK. Prone position alters the effect of volume overload on regional pleural pressures and improves hypoxemia in pigs in vivo. Am Rev Respir Dis 1992;146:300eC306.
Mure M, Glenny RW, Domino KB, Hlastala MP. Pulmonary gas exchange improves in the prone position with abdominal distension. Am J Respir Crit Care Med 1998;157:1785eC1790.
Hering R, Wrigge H, Vorwerk R, Brensing K, Schrder S, Zinserling J, Hoeft A, Spiegel TV, Putensen C. The effects of prone positioning on intraabdominal pressure and cardiovascular and renal function in patients with acute lung injury. Anesth Analg 2001;92:1226eC1231.
Colmenero-Ruiz M, Pola-Gallego de Guzman D, del Mar Jimenez-Quintana M, Fernandez-Mondejar E. Abdomen release in prone position does not improve oxygenation in an experimental model of acute lung injury. Intensive Care Med 2001;27:566eC573.
Palmon SC, Kirsch JR, Depper JA, Toung TJK. The effect of the prone position on pulmonary mechanics is frame-dependent. Anesth Analg 1998;87:1175eC1180.
Tanskanen P, Kytt J, Randell T. The effect of patient positioning on dynamic lung compliance. Acta Anaesthesiol Scand 1997;40:602eC606.
Hoffman E. Effect of body orientation on regional lung expansion: a computed tomographic approach. J Appl Physiol 1985;59:468eC480.
Valenza F, Guglielmi M, Maffioletti M, Tedesco C, Maccagni P, Fossali T, Aletti G, Porro GA, Irace M, Carlesso E, et al. Prone position delays the progression of ventilator-induced lung injury in rats: does lung strain distribution play a role Crit Care Med 2005;33:361eC367(Torsten Richter, Giacomo )
Clinic of Anesthesiology and Intensive Care Medicine, University Clinic Carl Gustav Carus, Dresden University of Technology, Dresden, Germany
ABSTRACT
Rationale: The prone position is used to improve gas exchange in patients with acute respiratory distress syndrome. However, the regional mechanism by which the prone position improves gas exchange in acutely injured lungs is still incompletely defined. Methods: We used positron emission tomography imaging of [13N]nitrogen to assess the regional distribution of pulmonary shunt, aeration, perfusion, and ventilation in seven surfactant-depleted sheep in supine and prone positions. Results: In the supine position, the dorsal lung regions had a high shunt fraction, high perfusion, and poor aeration. The prone position was associated with an increase in lung gas content and with a more uniform distribution of aeration, as the increase in aeration in dorsal lung regions was not offset by loss of aeration in ventral regions. Consequently, the shunt fraction decreased in dorsal regions in the prone position without a concomitant impairment of gas exchange in ventral regions, thus leading to a significant increase in the fraction of pulmonary perfusion participating in gas exchange. In addition, the vertical distribution of specific alveolar ventilation became more uniform in the prone position. A biphasic relation between regional shunt fraction and gas fraction showed low shunt for values of gas fraction higher than a threshold, and a steep linear increase in shunt for lower values of gas fraction. Conclusion: In a surfactant-deficient model of lung injury, the prone position improved gas exchange by restoring aeration and decreasing shunt while preserving perfusion in dorsal lung regions, and by making the distribution of ventilation more uniform.
Key Words: adult respiratory distress syndrome; emission-computed tomography; nitrogen isotopes; prone position; pulmonary gas exchange
Despite increasing use of the prone position as a means to improve gas exchange in patients with acute respiratory distress syndrome (ARDS) (1, 2), few studies have investigated the regional mechanism of this improvement in acutely injured lungs (3, 4). Using single-photon emission computed tomography in an oleic acid model of lung injury, Lamm and coworkers (4) showed that the prone position was associated with a narrower distribution of the ventilation-to-perfusion ratio and with an increase in the relative (i.e., mean-normalized) ventilation-to-perfusion ratio in dorsal lung regions. However, because regional shunt could not be measured directly by single-photon emission computed tomography, whether the increase in relative ventilation-to-perfusion ratio in dorsal regions corresponded to reversal of shunt and, more importantly, the magnitude of the associated improvement of regional gas exchange could not be determined. Furthermore, single-photon emission computed tomography did not allow assessment of regional aeration. Combining measurement of regional aeration with measurement of regional shunt and perfusion is important for two reasons. First, it allows determination of whether the improvement of gas exchange in the prone position is due to restored aeration to regions of shunt or to redistribution of perfusion away from these regions. In this respect, it has been suggested that the regional mechanism by which the prone position improves gas exchange may bear prognostic significance and that the prone position increases survival of patients with ARDS if the improvement of gas exchange on turning prone is associated with restored aeration to regions of shunt rather than with redistribution of perfusion away from these regions (5). Therefore, further insights into the effect of the prone position on regional shunt, aeration, and perfusion can have important clinical implications. Second, by combining measurements of regional aeration and regional shunt, it is possible to determine the functional consequence of the change in distribution of lung density associated with prone positioning (6).
Positron emission tomography (PET) imaging of infused [13N]nitrogen (13N2) is a noninvasive technique that allows the in vivo assessment of regional gas exchange and the direct quantification and localization of intrapulmonary shunt (7eC9). In this study, we used PET imaging of infused and inhaled 13N2 to assess the effect of the prone position on the regional distribution of pulmonary perfusion, shunt, aeration, and ventilation in a saline lavage model of acute lung injury. We investigated whether, in acute lung injury caused by surfactant depletion, the prone position improves gas exchange by restoring aeration and decreasing shunt in dorsal lung regions or by redistributing perfusion away from these regions. A preliminary analysis of some of the results of this study was previously reported in the form of an abstract (10).
METHODS
Experimental Procedures
Seven sheep (22.4 ± 2.9 kg) were anesthetized, paralyzed, and mechanically ventilated (tidal volume, 8 ml/kg; inspiratory-to-expiratory ratio, 1:2; fraction of inspired oxygen, 1.0; and positive end-expiratory pressure, 5 cm H2O). Respiratory rate was adjusted to normocapnia (19.7 ± 2 breaths/minute; PaCO2 = 39.9 ± 3.9 mm Hg) and was not modified during the study. An arterial line and a Swan-Ganz catheter were inserted and a jugular line was placed for infusion of 13N2 in saline solution. The animals were then suspended in a cylindric tube, which allowed them to remain positioned within the field of view of the PET scanner while body position was changed by rotation of the tube. This suspension system allowed the abdomen to be unsupported in the prone position.
After collection of control physiologic data in both positions, lung lavage was performed by instillation and drainage of warm saline (30 ml/kg) and repeated every 15 minutes until the PaO2 fell below 100 mm Hg in the supine position (Figure 1). After stable injury, defined as a PaO2 below 200 mm Hg that changed by less than 10% at two consecutive 15-minute intervals, was obtained, a set of physiologic data was collected in the supine position. The animals were then transported to the PET suite, where a recruitment maneuver (airway pressure of 40 cm H2O for 30 seconds) was performed and the previous mechanical ventilation settings were resumed. After 15 minutes, a 10-minute PET transmission scan was taken, followed by a 4-minute 13N2eCsaline infusion and 4-minute 13N2 inhalation PET emission scans. Physiologic data were collected during the last 2 minutes of the transmission scan. After the last PET scan, the animal was turned to the opposite body position, and the recruitment maneuver, imaging sequence, and data collection were repeated.
PET Imaging
The PET scanner imaged 15 contiguous, 6.5-mm-thick slices of thorax.
13N2eCsaline bolus infusion (perfusion) scan.
A bolus of 13N2 dissolved in saline was infused through the jugular catheter at the start of a 60-second apnea performed at a constant airway pressure corresponding to the mean airway pressure measured during positive-pressure ventilation (7, 9, 11, 12). Because of the low partition coefficient of nitrogen (water/air, 0.015), the kinetics of 13N2 during apnea differs between regions that are perfused and aerated, in which virtually all 13N2 diffuses into the alveolar airspace at first pass and accumulates in proportion to regional perfusion (12), and regions that contain alveolar units that are perfused but not aerated (i.e., "shunting"). In the latter regions, the 13N2 kinetics (Figure 2) shows a peak, related to total regional perfusion, followed by a plateau proportional to perfusion to aerated units within the region, because shunting units do not retain 13N2 during apnea (Figure 3). For each of eight horizontal regions of interest (ROIs) of lung (i = 1, ..., 8), regional perfusion (i), expressed as a fraction of total perfusion to the imaged lung, and regional shunt fraction (FS,PETi) were derived by curve fitting the kinetics of 13N2 with a model (7). The fraction of imaged pulmonary perfusion shunted within each ROI was calculated as Si = i · FS,PETi.
Regional specific alveolar ventilation of perfused units (sVAi, i.e., alveolar ventilation per unit of lung gas volume) was calculated as the reciprocal of the time constant derived from the initial 30 seconds of the tracer washout curve, obtained after mechanical ventilation was resumed (13).
13N2 inhalation (gas fraction) scan.
After equilibration of the alveolar gas with inhaled 13N2 (Figure 3), regional gas fraction (FGASi) was calculated by dividing mean regional specific tracer activity (i.e., activity per unit of lung volume) by the activity of a 1-ml sample of inhaled gas (9). Regional lung volume (VLi) was calculated as the product of the number of voxels within each ROI by the volume of a voxel. Regional gas volume was calculated as VGASi = VLi · FGASi and regional tissue volume was calculated as VTISi = VLi eC VGASi. An estimate of regional absolute alveolar ventilation was calculated as VAi = sVAi · VGASi.
Statistical Analysis
Data were analyzed by two-way analysis of variance. Statistical significance is reported only for comparisons between prone and supine positions after injury. Linear regression was used to estimate associations between variables and to compute vertical gradients (positive gradients indicate an increase from the dependent to the nondependent regions). Data are presented as means ± SD.
RESULTS
Global Physiologic Variables
Cardiovascular and respiratory variables measured in the supine position after injury and during imaging were not significantly different (see Table E1 in the online supplement), thus supporting the physiologic stability of the injury model throughout the experiment.
After injury, PaO2 was markedly lower, PaCO2 higher, and pH lower in the supine than in the prone position (Table 1). Turning prone was associated with a significant decrease in mean pulmonary arterial pressure, while cardiac output, heart rate, and mean arterial pressure were not significantly affected by the body position change. Pulmonary capillary wedge pressure was slightly lower in the prone position. Inspiratory plateau airway pressure was substantially lower in the prone than in the supine position (Table 1).
Regional Gas Content
The total, voxel-by-voxel mean-normalized variance of gas fraction, COVtot2 (see the online supplement), was almost fivefold greater in the supine than in the prone position (0.60 ± 0.24 vs. 0.13 ± 0.09; p < 0.001). The systematic component of the variance in the vertical direction, COVv2, accounted for 62 ± 21 and 29 ± 14% of COVtot2 in the supine and prone positions, respectively. In contrast, the systematic component of the variance in the craniocaudal direction, COVcc2, accounted for a small fraction of the total variance in both body positions (4 ± 4% in the supine position and 6 ± 12% in the prone position).
Both the gas volume and the gas fraction of the imaged lung were higher in the prone than in the supine position (384 ± 102 vs. 286 ± 88 ml, p < 0.05; and 0.45 ± 0.09 vs. 0.32 ± 0.10, p < 0.001), whereas the volume of imaged lung was not different between the two positions. There was significant interregional variation in gas fraction (effect of ROI: p < 0.01), and the anatomic (i.e., dorsal to ventral) distribution profile of regional gas fraction (Figure 4) was different between the supine and prone positions (interaction between ROI and body position: p < 0.001). Accordingly, regional gas fraction showed a steeper vertical gradient favoring nondependent lung regions in the supine position (FGASi gradient, 0.035 ± 0.011/cm in the supine position vs. 0.011 ± 0.014/cm in the prone position; p < 0.05).
Taken together, these results are consistent with a more uniform distribution of aeration in the prone than in the supine position.
Global and Regional Shunt Fraction
Consistent with the reduction of PaO2, shunt fraction measured from blood gases (FSO2) was significantly higher in the supine than in the prone position (Table 1). The global shunt fraction of the imaged lung, estimated by PET (FS,PET), was also higher in the supine than in the prone position (0.50 ± 0.13 vs. 0.14 ± 0.14; p < 0.001). In neither position was the difference between FS,PET and FSO2 statistically significant, and FS,PET showed good correlation with FSO2 (FS,PET = 0.83 · FSO2 + 0.06; r = 0.87, p < 0.001).
Analysis of variance showed a significant difference in regional shunt fraction (FS,PETi) between body positions (p < 0.001) and between ROIs (p < 0.001). In both positions the regional shunt fraction tended to decrease from dependent to nondependent ROIs (i.e., FS,PETi showed a negative vertical gradient). However, the vertical gradient of regional shunt fraction was significantly steeper in the supine than in the prone position (FS,PETi gradient, eC0.054 ± 0.018/cm in the supine position vs. eC0.013 ± 0.006/cm in the prone position; p < 0.01), indicating that the vertical dependence of shunt fraction was more pronounced in the supine than in the prone position. Consistent with these gradients, analysis of variance showed a significant interaction between the effect of body position and the effect of ROI (p < 0.001), which indicates that the anatomic distribution profile of regional shunt was different between the two body positions.
Taken together, these results indicate that the prone position was associated with an overall reduction of regional shunt and with a more uniform distribution of the residual shunt among the ROIs.
Regional Perfusion
The anatomic distributions of regional perfusion (i) and shunt flow were also affected by body position (Figure 5). There was significant interregional variation of perfusion (effect of ROI: p < 0.001), and the anatomic distribution profile of perfusion was different between the supine and prone positions (interaction between ROI and body position: p < 0.001). In the supine position, the dorsal (i.e., dependent) half of the ROIs received the greatest fraction of perfusion. In the prone position, the dorsal half of the ROIs still received most of the perfusion. These results were consistent with measurements of the perfusion gradient, which showed reversal of the nondependent-to-dependent gradient in the prone position, as evidenced by the opposite signs of the gradients in the prone position (positive gradient) compared with the supine position (negative gradient; see STATISTICAL ANALYSIS, in METHODS, in the online supplement for interpretation of the sign of the gradient). However, the absolute value of the vertical gradient of regional perfusion normalized by relative regional tissue volume, which represents a measure of the steepness of the vertical gradient independent of its direction and of the relative amount of lung tissue within each ROI, was significantly lower in the prone position (0.040 ± 0.024/cm) than in the supine position (0.098 ± 0.030/cm; p < 0.01), indicating some redistribution of perfusion toward ventral regions in the prone position.
As a result of the high shunt fraction and high perfusion in dependent regions in the supine position, fractional pulmonary shunt flow (Si) was also highest in the dorsal half of the ROIs (i.e., a large fraction of pulmonary perfusion went to shunting units in dorsal, dependent ROIs). In contrast, in the prone position, the dependent (i.e., ventral) half of the ROIs contributed only modestly to shunt flow (Figure 5). Accordingly, fractional shunt flow was significantly lower in the prone position (effect of body position: p < 0.001) and its anatomic distribution was different between the supine and prone positions (interaction between ROI and body position: p < 0.001), as fractional shunt flow decreased markedly in dorsal regions in the prone position.
Taken together, these results suggest that, despite some degree of perfusion redistribution toward ventral regions in the prone position, a substantial fraction of perfusion was maintained in dorsal regions. Because in the prone position the dorsal half of the ROIs had a major reduction in regional shunt fraction, while still receiving most of the perfusion, this position was associated with a considerable increase in the fraction of perfusion contributing to gas exchange.
Relation between Regional Shunt and Gas Fraction
When regional data from each animal for both body positions were pooled, an inverse relation between average regional values of FS,PETi and FGASi became apparent (Figure 6A). The main effect of the prone position was to shift the data points corresponding to the most dorsal ROIs from the low-gas and high-shunt fraction to the high-gas and low-shunt fraction portion of the relation. When plots of FS,PETi versus FGASi were analyzed for each animal, biphasic behavior was identified in five of them: values of shunt fraction were low for regional gas fraction higher than a threshold (FGASi 0.3) and increased linearly for values of gas fraction lower than that threshold (Figure 6B). In these animals, when the equation FS,PET = 1 eC FGAS/F0 was fitted to regional data points with FS,PETi > 0.2, the intercept on the FGAS axis was F0 = 0.35 ± 0.10. In the remaining two animals, biphasic behavior could not be identified (Figure 6C).
Regional Specific Alveolar Ventilation
Specific alveolar ventilation of the imaged lung was similar in the supine and prone positions (0.040 ± 0.012 and 0.039 ± 0.012 secondeC1, respectively). Regional specific alveolar ventilation (sVAi) showed significant interregional variation (effect of ROI: p < 0.001). Although the linear component of the vertical variation of specific ventilation was not significant in either position, the vertical distribution profile of specific ventilation was more uniform in the prone than in the supine position (Figure 7). Accordingly, when the ROIs were ordered by vertical location (i.e., nondependent to dependent), analysis of variance showed a significant interaction between the effect of ROI and body position (p < 0.01).
DISCUSSION
The main finding of this study is that the improvement of gas exchange in the prone position was associated with restored aeration and reduced shunt in dorsal lung regions, without a concomitant reduction of aeration or increase in shunt in ventral regions. The improvement of gas exchange was further enhanced by the fact that perfusion to dorsal regions remained substantial in the prone position, thus allowing these regions to take advantage of their restored aeration and leading to an increase in dorsal, and total, gas-exchanging pulmonary blood flow.
Limitations of the Study
The field of view of the PET scanner did not encompass the entire lung. In sheep with similar weight to those of this study, we previously estimated that the imaged lung corresponded to about 70% of the entire lung (14). Although the field of view did not include the most cranial and the most caudal, subdiaphragmatic regions, the internal consistency between values of shunt fraction of the imaged lung, derived by PET, and values of global shunt fraction suggests that the behavior of the imaged lung was representative of the behavior of the entire lung. We also acknowledge that imaging the same cross-section of thorax in both positions did not necessarily ensure that the same cross-section of lung was imaged, as the lung could expand or compress along the craniocaudal axis due to shift of the diaphragm. Although the limitation of PET to resolve fine anatomic markers did not allow spatial coregistration of the supine and prone images, the length of the imaged cross-section of lung ( 10 cm) was much larger than the possible lung displacement along the craniocaudal axis. This, as well as the small contribution of the craniocaudal direction to the variance of gas fraction, suggest that the effect of this potential artifact was probably small.
We used the saline lavage model, which represents a highly recruitable model of acute lung injury (15). After lung lavage, the damage to alveolar epithelial, endothelial, and perivascular cells is probably less than that of other lung injury models, such as the oleic acid or the endotoxin infusion models (16, 17), in which the primary insult is delivered from the pulmonary vascular side. Nonetheless, lung lavage creates interstitial and intraalveolar edema and causes impaired gas exchange by its detrimental effect on surfactant (18eC20). Although the pathogenetic mechanism is different, the severity of the impairment of gas exchange that was obtained after lung lavage is comparable to that of other experimental models of acute lung injury and is not the result of a specific pattern of disruption of ventilation-to-perfusion distribution peculiar to this model (21). Therefore, this model reflects the pathophysiology of those cases of ARDS that are characterized predominantly by alveolar instability and collapse, with high potential for alveolar recruitment, rather than by alveolar consolidation or flooding.
We analyzed the regional data in eight horizontal ROIs. We recognize that such an approach neglects intraregional heterogeneity. However, in this model of lung injury, the contribution of the systematic component of the variance in the vertical direction to the total variance of gas fraction was substantially greater than the contribution of the component in the craniocaudal direction. Therefore, analyzing the data by horizontal ROIs allowed us to capture the main systematic trend in regional aeration.
Effect of the Prone Position on Global Gas Exchange and Hemodynamics
After injury, the prone position was associated with improved gas exchange compared with the supine position, in line with the results of previous studies (3, 4, 22, 23). Whereas all these studies consistently reported an improvement of PaO2 and a decrease in global shunt fraction in the prone position, our results also showed an improvement of PaCO2. Because body temperature was not different between the two positions and tidal volume and respiratory rate were kept constant, the improvement of PaCO2 suggests that ventilation was more efficient in the prone position. This is consistent with our finding that regional specific alveolar ventilation was more uniform in the prone position, indicating a more homogeneous turnover of alveolar gas of perfused units in this position. Furthermore, in the prone position, the vertical distribution profile of absolute alveolar ventilation, estimated from the product of specific ventilation and gas volume, followed closely the vertical profile of relative perfusion (compare Figure 5 with Figure E1 in the online supplement). This finding suggests improved ventilation-to-perfusion matching in the prone position (4) and, because minute ventilation and cardiac output were not different between the supine and prone positions, it would explain the reduction of PaCO2. It has been shown that patients with ARDS who responded to a trial of prone positioning with a reduction of PaCO2 had increased survival at 28 days, and it has been hypothesized that in these patients the improvement of PaO2 in the prone position was due to alveolar recruitment rather than to redistribution of perfusion away from regions of shunt (5). Our data are compatible with this hypothesis by showing that, when the main effect of the prone position was to restore aeration, while preserving perfusion, to shunting units in dorsal regions, the increase in PaO2 was accompanied by a significant decrease in PaCO2 and by evidence of improved ventilation-to-perfusion matching. However, this reduction of PaCO2 cannot be considered a specific marker of alveolar recruitment, because PaCO2 should also decrease if the improvement of ventilation-to-perfusion matching were due to redistribution of perfusion away from regions of shunt.
The finding that pulmonary arterial pressure was higher in the supine than in the prone position after injury is consistent with the effect of higher PaCO2, lower pH, and lower PaO2 on pulmonary vascular tone. Pulmonary capillary wedge pressure was slightly lower in the prone than in the supine position. It is possible, however, that by setting the zero value of the pressure transducer at midchest level, we slightly underestimated left ventricular filling pressure in the prone position, as the prone position displaces the left ventricle inferiorly (23).
Effect of the Prone Position on Regional Perfusion, Shunt, and Aeration
The observation that substantial perfusion was preserved in dorsal lung regions in the prone position is consistent with previous observations in oleic acideCinduced lung injury (3, 24, 25) and in healthy animals (26eC30), and it could be due to higher vascular conductance in these regions (31, 32). However, the fact that the vertical perfusion gradient was not fully reversed in the prone position suggests some degree of perfusion redistribution toward ventral regions, consistent with results obtained in oleic acideCinduced lung injury (25). We previously reported marked interindividual variability in the extent to which perfusion redistributed toward ventral regions in prone normal humans (12). Whether this variability is present also in patients with ARDS and whether patients who maintain substantial dorsal perfusion are those who respond favorably to prone positioning remains to be investigated. In this respect, it is important to emphasize that the extent to which inferences on the importance of structural factors in favoring perfusion to dorsal regions can be extrapolated to humans is unclear. Indeed, whereas it would appear teleologically meaningful for quadrupeds, who are normally prone, to develop mechanisms that, by favoring perfusion to dorsal lung regions, render perfusion more homogeneous, we cannot envision, a priori, a physiologic advantage of diverting perfusion toward dorsal regions in bipeds.
Critical to the improvement of gas exchange was the restored aeration to dorsal lung regions. This led to an increase in lung gas volume and to a more uniform distribution of gas fraction, as the gain in aeration in dorsal regions was not offset by loss of aeration in ventral regions. This finding is consistent with the presence of a smaller vertical pleural pressure gradient in the prone position (33). Although improvement of gas exchange in the prone position can occur in the absence of an increase in functional residual capacity (22, 33), the increase in lung gas volume likely contributed to the dramatic improvement of oxygenation that we observed. A potential contributor to the improved aeration of dorsal regions was the lack of abdominal support in the prone position. Unsupporting the abdomen likely decreased intraabdominal pressure (potentially creating negative intraabdominal pressure), reduced the pressure gradient across the diaphragm, and promoted alveolar recruitment preferentially in the dorsal regions of this surfactant-deficient model of lung injury. However, the prone position has also been shown to improve gas exchange when the abdomen was resting on a surface (33eC35). Furthermore, the effect of unsupporting the abdomen in the prone position may depend on the pathophysiology of the injury: unsupporting the abdomen may promote alveolar recruitment and more uniform lung aeration in the presence of atelectasis and alveolar collapse, but it may be ineffective in the presence of alveolar consolidation or flooding, as in the oleic acid model (36). Although our method of suspending the animal in the prone position is not applicable to humans, and decreased intraabdominal pressure may be difficult to achieve when patients with ARDS are positioned prone, technical developments have led to critical care beds that can be used to pronate patients and that could potentially incorporate features to leave the abdomen unsupported in the prone position. In this respect, studies in anesthetized humans undergoing surgical procedures in the prone position suggest that respiratory mechanics are optimized when the abdomen is unsupported (37, 38). Future studies, however, will be needed to clarify the relative contribution of unsupporting the abdomen to the improvement of gas exchange in the prone position.
The beneficial effect that the improvement of regional aeration had on regional gas exchange was apparent from the relation between regional shunt fraction and regional gas fraction. This relation showed that the effect of the prone position was to restore aeration to dorsal regions, thus shifting these regions to a more favorable portion of the shunt-versus-gas fraction relation. Interestingly, in five animals we observed biphasic behavior, with low shunt fraction in regions with gas fraction higher than a critical threshold and a steep linear increase in shunt in regions with gas fraction lower than this threshold. This threshold (F0 = 0.35) was lower than the gas fraction of the normal lung ( 0.6 [27, 39]). This finding suggests that, in these animals, even poorly aerated regions maintained their ability to exchange gas. Assuming that the shunt fraction is proportional to the fraction of derecruited units, the linear increase in shunt for a gas fraction lower than this threshold is consistent with a "quantal" loss of gas-exchanging units, so that the decrease in aeration of a region is absorbed only by a certain number of its units that become atelectatic and lose the ability to exchange gas, whereas the remaining units preserve their aeration and, accordingly, their gas exchange. Therefore, the biphasic relation between regional shunt fraction and gas fraction would suggest the presence of a gas fraction threshold for the development of subregions that collapse or flood and begin to shunt. The lack of biphasic behavior in the other two animals could have been related to intraregional differences in the degree of surfactant depletion, so that alveolar collapse started to occur at a higher gas fraction in units with more severe impairment of surfactant. Supporting this hypothesis is the fact that these two animals exhibited the highest contribution of the component of variance in the craniocaudal direction to the total variance of gas fraction, suggesting that the horizontal ROIs used to analyze their data included units with different conditions.
Importantly, the improved aeration and gas exchange in dorsal regions in the prone position were obtained at a lower airway pressure than in the supine position. The prone position might thus be an effective means to promote alveolar recruitment in dorsal regions, while limiting airway pressure and ventilator-associated lung injury (17, 40).
In summary, we demonstrated that, in a surfactant-deficient model of ARDS, the prone position reduced shunt in dorsal lung regions. The reduction of shunt was associated with an increase in pulmonary gas content and with a more homogeneous distribution of aeration, which placed dorsal lung regions on a more favorable portion of the shunt-versus-gas fraction relation without a concomitant, substantial increase in shunt in ventral regions. Preservation of perfusion to dorsal regions contributed to the improvement of global shunt fraction in the prone position. The prone position with unsupported abdomen appears to be a useful means to promote alveolar recruitment and improve global and regional gas exchange, while limiting airway pressure, when the main pathophysiologic alteration is loss of surfactant and the associated alveolar derecruitment.
Acknowledgments
The authors thank E. Kuhlisch, for statistical advice; S. B. Weise, for technical assistance with image acquisition and processing; J. A. Correia, Ph.D., W. M. Bucelewicz, and D. F. Lee, B.S., for preparation of the radioisotope; O. Syrkina, M.D., for assistance with animal preparation; and T. Schroeder, M.Eng., for insightful comments.
This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org
In only two animals, COVv2 accounted for less than 60% of COVtot2 in the supine position. Excluding them from this analysis raised the contribution of COVv2 to 73 ± 8% in the supine position and to 30 ± 12% in the prone position, and lowered the contribution of COVcc2 to 2.4 ± 1% (supine) and 1.7 ± 1% (prone).
REFERENCES
Gattinoni L, Tognoni G, Pesenti A, Taccone P, Mascheroni D, Labarta V, Malacrida R, Di Giulio P, Fumagalli R, Pelosi P, et al. Effect of prone positioning on the survival of patients with acute respiratory failure. N Engl J Med 2001;345:568eC573.
Papazian L, Bregeon F, Gaillat F, Thirion X, Gainnier M, Gregoire R, Saux P, Gouin F, Jammes Y, Auffray JP. Respective and combined effects of prone position and inhaled nitric oxide in patients with acute respiratory distress syndrome. Am J Respir Crit Care Med 1998;157:580eC585.
Wiener CM, Kirk W, Albert RK. Prone position reverses gravitational distribution of perfusion in dog lungs with oleic acid-induced lung injury. J Appl Physiol 1990;68:1386eC1392.
Lamm WJ, Graham MM, Albert RK. Mechanism by which the prone position improves oxygenation in acute lung injury. Am J Respir Crit Care Med 1994;150:184eC193.
Gattinoni L, Vagginelli F, Carlesso E, Taccone P, Conte V, Chiumello D, Valenza F, Caironi P, Pesenti A. Decrease in PaCO2 with prone position is predictive of improved outcome in acute respiratory distress syndrome. Crit Care Med 2003;31:2727eC2733.
Gattinoni L, Pelosi P, Vitale G, Pesenti A, D'Andrea L, Mascheroni D. Body position changes redistribute lung computed-tomographic density in patients with acute respiratory failure. Anesthesiology 1991;74:15eC23.
Galletti GG, Venegas JG. Tracer kinetic model of regional pulmonary function using positron emission tomography. J Appl Physiol 2002;93:1104eC1114.
Willey-Courand DB, Harris RS, Galletti GG, Hales CA, Fischman A, Venegas JG. Alterations in regional ventilation, perfusion, and shunt after smoke inhalation measured by PET. J Appl Physiol 2002;93:1115eC1122.
Musch G, Harris RS, Vidal Melo MF, O'Neill KR, Layfield JDH, Winkler T, Venegas JG. Mechanism by which a sustained inflation can worsen oxygenation in acute lung injury. Anesthesiology 2004;100:323eC330.
Richter T, Harris RS, Vidal Melo MF, Musch G, Layfield JDH, Venegas JG. Prone position virtually eliminates regions of shunt in surfactant depleted sheep . Am J Respir Crit Care Med 2002;165:A784.
O'Neill K, Venegas JG, Richter T, Harris RS, Layfield JDH, Musch G, Winkler T, Vidal Melo MF. Modeling kinetics of infused 13NNeCsaline in acute lung injury. J Appl Physiol 2003;95:2471eC2484.
Musch G, Layfield JDH, Harris RS, Vidal Melo MF, Winkler T, Callahan RJ, Fischman AJ, Venegas JG. Topographical distribution of pulmonary perfusion and ventilation, assessed by PET in supine and prone humans. J Appl Physiol 2002;93:1841eC1851.
Vidal Melo MF, Harris RS, Layfield D, Musch G, Venegas JG. Changes in regional ventilation after autologous blood clot pulmonary embolism. Anesthesiology 2002;97:671eC681.
Vidal Melo MF, Layfield D, Harris RS, O'Neill K, Musch G, Richter T, Winkler T, Fischman AJ, Venegas JG. Quantification of regional ventilationeCperfusion ratios with positron emission tomography. J Nucl Med 2003;44:1982eC1991.
Downie JM, Nam AJ, Simon BA. PressureeCvolume curve does not predict steady-state lung volume in canine lavage lung injury. Am J Respir Crit Care Med 2004;169:957eC962.
Schuster DP. ARDS: clinical lessons from the oleic acid model of acute lung injury. Am J Respir Crit Care Med 1994;149:245eC260.
Brigham KL, Meyrick B. Endotoxin and lung injury. Am Rev Respir Dis 1986;133:913eC927.
Berggren P, Lachmann B, Curstedt T, Grossmann G, Robertson B. Gas exchange and lung morphology after surfactant replacement in experimental adult respiratory distress syndrome induced by repeated lung lavage. Acta Anaesthesiol Scand 1986;30:321eC328.
Lachmann B, Jonson B, Lindroth M, Robertson B. Modes of artificial ventilation in severe respiratory distress syndrome: lung function and morphology in rabbits after wash-out of alveolar surfactant. Crit Care Med 1982;10:724eC732.
Lachmann B, Robertson B, Vogel J. In vivo lung lavage as an experimental model of the respiratory distress syndrome. Acta Anaesthesiol Scand 1980;24:231eC236.
Neumann P, Hedenstierna G. VentilationeCperfusion distributions in different porcine lung injury models. Acta Anaesthesiol Scand 2001;45:78eC86.
Albert RK, Leasa D, Sanderson M, Robertson HT, Hlastala MP. The prone position improves arterial oxygenation and reduces shunt in oleic-acid-induced acute lung injury. Am Rev Respir Dis 1987;135:628eC633.
Broccard AF, Shapiro RS, Schmitz LL, Ravenscraft SA, Marini JJ. Influence of prone position on the extent and distribution of lung injury in a high tidal volume oleic acid model of acute respiratory distress syndrome. Crit Care Med 1997;25:16eC27.
Richard JC, Janier M, Lavenne F, Berthier V, Lebars D, Annat G, Decailliot F, Guerin C. Effect of position, nitric oxide, and almitrine on lung perfusion in a porcine model of acute lung injury. J Appl Physiol 2002;93:2181eC2191.
Schuster DP, Haller J. Effects of body position on regional pulmonary blood flow during acute pulmonary edema in dogs: a positron emission tomography study. J Crit Care 1991;6:19eC28.
Glenny RW, Lamm WJ, Albert RK, Robertson HT. Gravity is a minor determinant of pulmonary blood flow distribution. J Appl Physiol 1991;71:620eC629.
Treppo S, Mijailovich SM, Venegas JG. Contributions of pulmonary perfusion and ventilation to heterogeneity in A/ measured by PET. J Appl Physiol 1997;82:1163eC1176.
Walther SM, Domino KB, Glenny RW, Hlastala MP. Pulmonary blood flow distribution in sheep: effects of anesthesia, mechanical ventilation, and change in posture. Anesthesiology 1997;87:335eC342.
Beck KC, Vetterman J, Rehder K. Gas exchange in dogs in the prone and supine positions. J Appl Physiol 1992;72:2292eC2297.
Won C, Chon D, Tajik J, Tran BQ, Robinswood GB, Beck KC, Hoffman EA. CT-based assessment of regional pulmonary microvascular blood flow parameters. J Appl Physiol 2003;94:2483eC2493.
Beck KC, Rehder K. Differences in regional vascular conductances in isolated dog lungs. J Appl Physiol 1986;61:530eC538.
Pelletier N, Robinson NE, Kaiser L, Derksen FJ. Regional differences in endothelial function in horse lungs: possible role in blood flow distribution J Appl Physiol 1998;85:537eC542.
Mutoh T, Guest RJ, Lamm WJE, Albert RK. Prone position alters the effect of volume overload on regional pleural pressures and improves hypoxemia in pigs in vivo. Am Rev Respir Dis 1992;146:300eC306.
Mure M, Glenny RW, Domino KB, Hlastala MP. Pulmonary gas exchange improves in the prone position with abdominal distension. Am J Respir Crit Care Med 1998;157:1785eC1790.
Hering R, Wrigge H, Vorwerk R, Brensing K, Schrder S, Zinserling J, Hoeft A, Spiegel TV, Putensen C. The effects of prone positioning on intraabdominal pressure and cardiovascular and renal function in patients with acute lung injury. Anesth Analg 2001;92:1226eC1231.
Colmenero-Ruiz M, Pola-Gallego de Guzman D, del Mar Jimenez-Quintana M, Fernandez-Mondejar E. Abdomen release in prone position does not improve oxygenation in an experimental model of acute lung injury. Intensive Care Med 2001;27:566eC573.
Palmon SC, Kirsch JR, Depper JA, Toung TJK. The effect of the prone position on pulmonary mechanics is frame-dependent. Anesth Analg 1998;87:1175eC1180.
Tanskanen P, Kytt J, Randell T. The effect of patient positioning on dynamic lung compliance. Acta Anaesthesiol Scand 1997;40:602eC606.
Hoffman E. Effect of body orientation on regional lung expansion: a computed tomographic approach. J Appl Physiol 1985;59:468eC480.
Valenza F, Guglielmi M, Maffioletti M, Tedesco C, Maccagni P, Fossali T, Aletti G, Porro GA, Irace M, Carlesso E, et al. Prone position delays the progression of ventilator-induced lung injury in rats: does lung strain distribution play a role Crit Care Med 2005;33:361eC367(Torsten Richter, Giacomo )