Lung volume assessments in normal and surfactant depleted lungs: agreement between bedside techniques and CT imaging

The history of medicine has been marked by a spectacular evolution of mechanical ventilation and its inevitable benefit in providing life support both in anaesthesia and critical care management. Nevertheless, one of the main consequences of mechanical ventilation with positive intermittent pressure in the airways is the loss of static lung volumes subsequent to atelectasis, intra-alveolar fluid accumulation, interstitial edema formation, and surfactant damage. In order to prevent these deleterious effects on the lung, new concepts of mechanical ventilation target an open lung ventilation strategy involving protective ventilation modes with low tidal volumes, regular alveolar recruitments and maintenance of optimal positive end-expiratory pressure (PEEP). In order to optimize ventilation strategy, it is essential to include lung volume assessment as part of ventilation monitoring.

Assessment of lung volumes is widespread in the respiratory medicine; however, there are only few techniques that can be implemented at bed side to optimize the ventilation strategy. Static lung volume, such as the end-expiratory lung volume (EELV) in ventilated patients is assessed by analyzing the concentration changes of an inert gas during wash-in/wash-out maneuvers [1‐4]. A recently developed alternative to EELV is the effective lung volume (ELV), which is based on a continuous breath-by-breath analysis of the carbon dioxide after alteration of the inspiratory pattern [5]. We showed that ELV detects lung derecruitment and recruitment [6] with the additional advantage of continuous monitoring of the ventilated patient.

Image processing to quantify lung volume from computed tomography CT scans has been largely used as a reference method in patients with acute lung injury [7‐9]. However, it is technically demanding, time consuming and involves radiation exposure. CTs are therefore inappropriate for routine clinical practice. Bedside monitoring of EELV may provide similar insight into the lung volume available for gas exchange. Comparative results for these different available approaches for quantification of the aerated lung areas are not available.

We aimed to relate lung volume indices obtained from bedside techniques to those computed by chest CT image analyses. Furthermore, we aimed to establish the relationships between static lung volume outcome variables obtained from healthy lungs and in a model of acute lung injury by surfactant depletion.

Lung volume indices obtained at different PEEP levels in healthy and surfactant-depleted lungs are demonstrated in Figure 2. As expected, increasing PEEP led to elevations in all end expiratory lung volume variables before (p < 0.001) and after whole lung lavage (p < 0.001). In the healthy lungs, no statistically significant differences were evident between EELV_MBW,He and EELV_CT at any PEEP levels. However, ELV was significantly greater than either EELV variable in healthy lung at low PEEP levels (p < 0.001) while this difference was not obvious during the maintenance of a PEEP of 9 cmH₂O. No statistically significant difference was detectable between the lung volume variables after lung lavage at any PEEP level.Figure 3 depicts changes in the LCI with PEEP under baseline conditions and after lung lavage. Surfactant depletion elevated LCI at all PEEP levels (p < 0.001). Increasing PEEP resulted in significant decreases in the LCI (p < 0.001); these reductions were more pronounced for the lungs after lavage.

The relationship between the lung volume indices obtained by the three different measurements techniques are demonstrated on Figure 4. There were statistically significant correlations between each pair of variables (p < 0.001) with strongest associations between EELV_MBW,He and EELV_CT (r = 0.93, p < 0.0001). The correlations between the ELV and the other two lung volume indices (EELV_CT and EELV_MBW,He) were somewhat weaker (r = 0.58 and r = 0.76), but still highly significant (p < 0.0001).

Figure 5 summarizes the degrees of agreements between each pair of lung volume variables measured on polar plots. The mean value of the variable pairs are plotted as a distance from the centre of the polar plot, while the angle between the radial vector and the horizontal axis represents the differences between each pair of variables. Excellent agreement was found between EELV_MBW,He and EELV_CT in the normal lungs with mean polar angle of 2.4 degrees and radial limits of agreement of -15.5 and 20.3 degrees (based on the 95% confidence interval limits). This close agreement remained after lavage with mean polar angle of 8.0 degrees and agreement limits of -19.6 and 35.6 degrees. There was a weaker agreement between the ELV and the other two lung volume variables, with the mean polar angles of -23.5 (EELV_MBW,He) and -21.1 (EELV_CT) degrees and radial limits of agreement of -58.9 and 11.8 (EELV_MBW,He), and -69.3 and 27.1 degrees (EELV_CT), respectively, under baseline condition. Surfactant depletion improved these agreements with mean polar angles of -2.1 (EELV_MBW,He) and 5.8 (EELV_CT) degrees and radial limits of agreement of -34.1 and 29.8 (EELV_MBW,He), and -45.6 and 57.4 degrees (EELV_CT), respectively.

The results of the present study demonstrate that bedside assessment of lung volume offers a meaningful alternative to those obtained by CT scan imaging. Both bedside lung volume assessments were able to follow the increased lung aeration with elevating PEEP and the development of airway closures following surfactant depletion. This study also evidenced an excellent value of the lung volume measurement performed by inert gas multiple breath washout technique both in normal lungs and in a model of acute lung injury. Conversely, effective lung volume available for gas exchange assessed from the analyses of the expired CO₂ exhibited greater scatters in the agreements with the other two indices.

We adopted the technique of surfactant depletion by performing whole lung lavage as a model of acute lung injury. The pathogenesis of this severe pulmonary disorder is far more complex then the loss of surfactant function. Nonetheless, the present model mimics lung collapsibility and the loss or regional lung ventilation with increased heterogeneity, particularly at low PEEP levels [17]. Atelectasis development was evidenced from the marked and significant decreases in EELV_MBW,He (-38 ± 5.3%, p < 0.005) and ELV (-65 ± 4.2%, p < 0.001) at PEEP 0. The lavage-induced changes in EELV_CT were less obvious, since lung volume variable includes all aerated structures including those with trapped air [17]. To avoid time effects in our findings, the order of PEEP was randomized within a set of recordings. However, the three different lung volume readings were always performed in the same sequence. None of the measurements techniques required altering the ventilation pressures beyond the normal range. The only change in the ventilation pattern was necessary to record ELV, but the short inspiratory pause did not affect the peak or mean airway pressures markedly and thus, it was unlikely to bias our findings.

An important finding of the present study is that while ELV follows the expected trend with increasing PEEP, it systematically overestimates the EELV at low PEEP levels in the healthy lungs (Figure 2). To provide a plausible explanation for this apparent discrepancy, the fundamental differences between ELV and EELV estimates warrant consideration. ELV reported in the present study reflects all compartments contributing to the CO₂ content in the lung including the alveoli, lung tissue and lung capillary blood. While the individual contribution of these compartments is not straightforward to assess in the present study, previous investigations established a correction factor of 55% at PEEP 0 in healthy lungs to adjust for both blood and lung tissue CO₂ content [18]. Indeed, implementation of this correction factor at PEEP 0 reveals excellent agreements between the corrected ELV (26.9 ± 2.4 ml), the EELV_MBW,He (26.2 ± 1.8 ml) and the EELV_CT data (24.1 ± 2.1 ml). Nevertheless, a systematic application of such a correction factor during the maintenance of different PEEP levels and in injured lungs would be erroneous, since both PEEP and surfactant depletion affect markedly the capillary blood in the alveolar wall and the parenchymal contribution to the total CO₂[19]. A PEEP-dependent correction factor taking into account the capillary blood volume and the lung tissue contribution in healthy and injured lungs may be a subject for further investigations.

When different lung volume indices and their changes are related, an important feature of their ability to access trapped (non-ventilated) lung regions should be taken into account. Since EELV and ELV are assessed by analyzing the dynamics of the expired gas, these measures provide information only about the size of the ventilated lung regions. Conversely, quantification of lung volume based on chest CT image analyses includes all low attenuation areas and thus, it cannot distinguish between ventilated or trapped aerated regions. This difference is unlikely to affect the lung volume estimates in a normal lungs, however may gain importance after lung lavage where the airway closures may lead to trapped air in the lung periphery. Using a similar experimental model where the regional lung ventilation was assessed by synchrotron imaging, we recently demonstrated that air trapping may be encountered at PEEP level of 3 cmH₂O [17]. Considering that the maximum amount of such lung regions with trapped air does not exceed 10%, the impact of this phenomenon on our results is minor. Accordingly, we observed no significant difference between the results of the three measurement techniques after lung lavage, despite a trend for a greater EELV_CT than the other two lung volume indices that may be explained by this factor (Figure 2, right).

To our knowledge, this is the first study relating lung volume variables obtained by a continuous assessment (ELV) to wash-in/wash-out and imaging techniques in ventilated subjects. A remarkable finding of the present study is the excellent concordance between the lung volume variables obtained by inert gas washout and chest CT imaging. This was manifested both in the excellent correlations between EELV_MBW,He and EELV_CT (Figure 4) and the good agreement (Figure 5). The importance of this finding stems from the fact that multiple breath washout technique by using ultrasonic flow meter has been applied commonly in experimental [6, 10, 17, 20] and clinical studies [2, 21‐26]. However, the relationship of EELV_MBW,He obtained by this approach was not related to that obtained by using a reference method in patients with acute lung injury based on CT imaging [7], and the validity of this variable has not been evaluated. Therefore, the results of the present study provides additional evidence on the validity of this ultrasonic tool for monitoring lung volume changes since it provides essentially identical information than the more cumbersome CT-imaging. Moreover, the application of the multiple breath wash-out technique gives quantitative information about the magnitude of ventilation heterogeneities, which may be another important variable to guide optimal lung ventilation without the need for the costly imaging involving radiation exposure [27, 28].

The advantage of ELV as opposed to the other two lung volume indices is that it allows real-time monitoring of the ventilated lung area available for gas exchange. The results of the present study agree with our previous findings demonstrating that ELV systematically overestimates lung volume measured by inert gas washout in normal lungs at low PEEP levels. This discrepancy disappeared at high PEEP level in normal lungs and was no longer detectable at any PEEP after lung lavage. This divergent relationship resulted in a poorer but significant correlation and worsened the agreement between ELV and EELV_MBW,He in normal and surfactant depleted lungs (Figures 4 and 5). The novelty of the present study in this regard is the demonstration of a similar relationship of ELV with changes in lung volume computed from chest CT scans. The systematic overestimation of static lung volumes by ELV is most plausibly explained by the contribution of lung capillary blood volume to ELV. Since ELV estimation is based on the analyses of the CO₂ as a tracer gas, it comprises the total CO₂ content of the lung, which is the sum of the CO₂ in the blood and in the alveolar space [6]. Thus, a larger lung volume is expected from the soluble CO₂ estimates as compared to a method based on non-soluble tracer gas, such as Helium, which solely measures the alveolar gas volume. Applying an elevated PEEP decreases the capillary blood volume in the pulmonary circulation [19], and this results in a better agreement between ELV and other lung volume indices when high positive pressure is applied in the airways. Furthermore, airway closures induced by surfactant depletion compromises CO₂ diffusion into the airspaces, which may explain the improved correlation and closer agreement between ELV and the other lung volume variables in the present of diminished surfactant function. These considerations highlight the potential susceptibility of ELV to alterations in ventilation/perfusion mismatch.

Lung volume assessment is required to monitor the development of atelectasis and alveolar recruitment and to guide optimal mechanical ventilation. However, lung volume indices that can be obtained bedside have not been related to reference values measured by chest CT scans. Therefore, this study provides meaningful information about the usefulness of bedside monitoring of lung volume to follow its changes with PEEP in normal lungs and in a model of acute lung injury. While lung volume variables obtained by multiple breath washout, CO₂ analyses and chest CT scans indicate lung expansion with increasing PEEP and alveolar derecruitment with lavage, their relationships are variable. Close correlation and excellent agreements were found between lung volumes obtained by He washout and chest CT analyses. However, the association between the effective lung volume achieved by analyzing the dynamics of expiratory CO₂ concentration and the other variables proved to be less clear-cut. This may be attributed to methodological differences with particular involvement of the capillary blood volume in the ELV estimates. In conclusion, both bedside techniques offer a valuable tool to follow lung volume changes without the need for cumbersome intervention. However, lung volume monitored by the CO₂ tracing provides continuous assessments that are subjected to pulmonary hemodynamical changes, while the EELV_MBW,He provides more specific measurements but at a cost of intermittent recordings.