Atelectasis

Study objective: To study the dynamics of lung compartments by dynamic CT (dCT) imaging during uninterrupted pressure-controlled ventilation (PCV) and different positive end-expiratory pressure (PEEP) settings in healthy and damaged lungs.

Design: Experimental animal investigation.

Setting: Experimental animal facility of a university department.

Interventions: In seven anesthetized pigs, static inspiratory pressure volume curves were obtained to identify the individual lower infection point (LIP) before and after saline solution lung lavage. During PCV, PEEP was adjusted 5 millibars (mbar) below the individually determined LIP (LIP - 5), at the LIP, and 5 mbar above the LIP (LIP + 5).

Measurements and results: Measurements were repeated before and after induction of lung damage. Hemodynamics, arterial and mixed venous blood gases, and dCT imaging in one juxtadiaphragmatic slice (effective temporal resolution of 100 ms) were assessed during uninterrupted PCV in series of three successive respiratory eyeles. The mean fractional area (FA) of the hyperinflated lung (FA-H), mean FA of ventilated lung, mean FA of poorly ventilated lung, and mean FA of nonventilated lung (FA-NV), and the change in FA of the whole lung area ([deltaFA) were compared at different PEEP settings. Calculated pulmonary shunt (Qs/Qt) was compared to FA-NV. LIP + 5 decreased the amount of atelectasis (FA-NV) and increased hyperinflation (FA-H) in healthy and injured lungs. Cyclic changes of atelectasls ([delta]FA-NV) and hyperinflation ([delta]FA-H) were observed in both healthy and injured lungs. In the injured but not in the healthy lungs, the amount of cyclic changes of atelectasis and hyperinflation were independent from the adjusted PEEP level. FA-NV correlated with the calculated Qs/Qt, with a slight overestimation (mean [+ or -] SEM, 2.1 [+ or -] 4.1%).

Conclusions: dCT imaging allows the following: (1) the quantification of the extent of atelectasis, ventilated, poorly ventilated, and hyperinflated lung parenchyma during ongoing mechanical ventilation; (2) the detection and quantification of repeated recruitment and derecruitment, as well as hyperinflation; and (3) an estimation of Qs/Qt. dCT adds promising functional information for the respiratory treatment of early ARDS. Key words: ARDS; atelectasis; CT; ventilation

Abbreviations: Cdyn = dynamic lung compliance; CI = confidence interval; CVP = central venous pressure; dCT = dynamic CT; [delta]FA = change in fractional area of the total lung area; FA = fractional area; FA-H = fractional area of hyperinflated lung; FA-NV = fractional area of nonventilated lung; FA-PV = fractional area of poorly ventilated lung; FA-V = fractional area of ventilated lung; FI[O.sub.2 = fraction of inspired oxygen; HR = heart rate; HU = Hounsfield units; IQR = interquartile range; LIP = lower inflection point; LIP + 5 = five millibars above the lower inflection point; LIP - 5 = five millibars below the lower inflection point; MAP = mean arterial pressure; mbar = millibar; MPAP = mean pulmonary artery pressure; PCV = pressure-controlled ventilation; PEEP = positive end-expiratory pressure; PV = pressure volume; Qs/Qt = pulmonary shunt fraction; R/D = recruitment/derecruitment RR = respiratory rate; VT = tidal volume

**********

The current strategy of mechanical ventilation in patients with ARDS is to prevent further lung injury. High positive end-expiratory pressure (PEEP) is applied to avoid end-expiratory lung collapse and low tidal volumes (VTs) to avoid lung overdistention and to minimize repeated collapse and reopening of the lung. PEEP in patients with ARDS receiving mechanical ventilation increases pulmonary functional residual capacity and improves oxygenation by reducing alveolar collapse and pulmonary shunt fraction (Qs/Qt). (1) Static pressure volume (PV) curves are recommended to determine the lower inflection point (LIP) and upper inflection point, defining the margins of PEEP level and peak pressure that might be safe in a respective patient. (2) It has been shown that both recruitment and derecruitment occur over the entire PV curve and the LIP is not able to predict optimum PEEP accurately. (3-10) The influence of PEEP on expiratory alveolar patency is well known, as shown by quantification of atelectasis using CT. (11) However, ventilation is a cyclic process, and Crotti et al (9) reported a wide range of opening pressures for recruitment and closing pressures for derecruitment in injured lungs. Real-time identification and quantification of the dynamic behavior of time-dependent recruitment, derecruitment, and hyperinflation during ongoing mechanical ventilation is difficult. In this context, significant shortcomings of static CT imaging maneuvers of the lung typically performed at end-inspiration and end-inspiration have to be noticed. In contrast to static CT, dynamic CT (dCT) allows continuous visualization of the lung during the complete respiratory cycle in mechanical ventilation and allows capture of regional effects of lung recruitment, derecruitment, and hyperinflation. (12) However, this technique offers additionally the identification of time-dependent changes of lung compartments defined by CT densitometry with high temporal resolution within the single respiratory cycle. (13) We hypothesized that dCT is able to verify repeated recruitment/derecruitment (R/D) of the lung at different PEEP levels.

We therefore compared pigs receiving mechanical ventilation with healthy vs surfactant-depleted lungs to characterize the effects of different PEEP settings (PEEP at the LIP, PEEP at 5 millibars [mbar] above the LIP [LIP + 5], and PEEP at 5 mbar below the LIP [LIP - 5]) derived from the static PV curve on different lung compartments (defined by CT densitometry) during the whole respiratory cycle with pressure-controlled ventilation (PCV). We simultaneously captured three successive respiratory cycles by dCT and calculated the following measures from these images: (1) the mean area of nonventilated, poorly ventilated, ventilated, and hyperinflated lung of these three respiratory cycles; (2) the maximum cyclic changes of nonventilated, poorly ventilated, ventilated, and hyperinflated lung within these three respiratory cycles; and (3) correlation of the mean Qs/Qt calculated from blood gas analysis to the mean nonventilated lung area.

MATERIALS AND METHODS

Animal Preparation

The study protocol was approved by the state animal care committee. Seven pigs (median weight, 25 kg [minimum to maximum, 23 to 27 kg]) were anesthetized with fentanyl, 0.005 mg/kg IV, and thiopental, 5 mg/kg IV, followed by continuous IV infusion of fentanyl, 5 [micro]g/kg/h, and thiopental, 10 mg/kg/h, and positioned supine during the entire experiment. Neuromuscular blockade was obtained with IV bolus of pancuronium, 0.1 mg/kg. The lungs were intubated via an endotracheal tube (internal diameter, 7.0 mm) and ventilated mechanically in volume-controlled mode (fraction of inspired oxygen [FI[O.sub.2]] of 0.5) [Servo 900C; Siemens Elema; Solna, Sweden]. PEEP was set to 3 mbar, and the inspiratory time was set to 50% of the whole respiratory cycle. VT was set to 12 mL/kg, and respiratory, rate (RR) was adjusted to achieve an end-tidal carbon dioxide concentration from 4.7 to 5.3 kPa. A continuous IV infusion of Ringers solution at a rate of 5 mL/kg/h was administered during the entire experiment. After exposure of the femoral vessels, arterial, central venous, and pulmonary artery catheters were inserted. Intravascular pressures, airway pressures, minute volume, and VT (S/5 Monitoring; Datex-Ohmeda; Duisburg, Germany), arterial blood gases, and acid-base status (Paratrend 7; Diametrics Medical; Buckinghamshire, UK) were monitored continuously. Intermittent mixed venous and arterial blood gas levels were determined (ABL 500/OSM 3; Radiolneter; Copenhagen, Denmark), and the samples were used to calibrate the paratrend monitor for continuous blood gas measurement.

Lung Lavage Model

A surfactant-depletion model was induced by repetitive lung lavages until a Pa[O.sub.2]/FI[O.sub.2] ratio < 13.3 kPa was achieved. The endotracheal tube was therefore disconnected from the ventilator, and warmed isotonic Ringers solution (20 mL/kg, 38 [degrees] C) was instilled (height of 70 cm above the endotracheal tube) until an air-fluid level was seen in the endotracheal tube. After 30 s of apnea, the fluid was retrieved by gravity drainage. To maintain stable hemodynamics following lung lavage, a continuous infusion of 4 [micro]g/kg/h (range, 2 to 6 [micro]g/kg) of epinephrine was administered and was unchanged during the experiment. After lung lavage, lung injury was progressed by ventilating the animals at volume-constant mode and a PEEP of 5 mbar for 2 h (FI[O.sub.2] 1.0; VT, 20 mL/kg; inspiratory time, 50%; RR was set to achieve normocapnia).

Static PV Curves

After a recruitment maneuver (continuous positive airway pressure of 40 mbar for 30 s), the tube was disconnected from the ventilator and the inspiratory PV loop was obtained. The airway pressure was measured at the proximal end of the endotracheal tube by a water column and was zeroed to the atmosphere. Inspiratory static PV curves were obtained as follows: the lung was stepwise inflated in 100-mL increments with a calibrated syringe (1,000 mL). The maneuver was terminated in case the airway pressure rose > 50 mbar. The static airway pressure was recorded after a delay of 4 s after each volume increment. The pressure level associated with a change in the upward slope of the first part of the PV curve was identified as the LIP. The LIP was evaluated graphically from the crossing of tangents applied to the slopes of the PV curve as demonstrated in Figure 1. (14,15) Changes in gas temperature and humidity of the gas were not taken into account.

dCT Imaging and Image Analysis

dCT imaging (multiscan technique) was performed in one juxtadiaphragmatic slice with the following CT settings: tube voltage, 120 kilovolts; tube current, 110 mA; matrix, 512 X 512 mm; and slice thickness, 1.0 mm, resulting in a voxel size of 0.34 x 0.34 x 1.0 mm (Somatom Plus 4; Siemens; Erlangen, Germany). Images were reconstructed using a high-resolution reconstruction algorithm. An effective temporal resolution of 100 ms was achieved using an overlapping temporal increment (sliding window technique), with a total radiograph tube rotation time of 750 ms. (13) In each CT image, the lung parenchyma was detected and differentiated from nonpulmonary tissues using a dedicated software tool. (16) The cross-sectional total lung area was divided into fractional areas (FAs) of predefined density ranges that differentiate atelectatic from ventilated lung parenchyma. (12) A density range of -300 to 200 Hounsfield units (HU) was used to define atelectasis, whereas a density range of -900 to -600 HU reflected ventilated lung parenchyma. The density range of -600 to -300 HU was used to define poorly ventilated lung parenchyma, and the range of -1,024 to -900 HU was used to quantify hyperinflated lung parenchyma. The four density ranges were expressed as fraction of the total cross-sectional lung area (FA in percentage). For each measurement, the FA data of all individual 100-ms CT scans were averaged over three respiratory cycles, and expressed as mean FA in percentage of total lung area (FA of nonventilated lung [FA-NV], FA of poorly ventilated lung [FA-PV]; FA of ventilated lung [FA-V], and FA of hyperinflated lung [FA-H]). Recruitment was defined as the decrease of FA-NV, and derecruitment was the increase of FA-NV. An experienced radiologist examined the three successive respiratory cycles visually to rule out any software error. The difference between end-expiratory and end-inspiratory FA of nonventilated ([delta]FA-NV), ventilated ([delta]FA-V), poorly ventilated ([delta]FA-PV), and hyperinflated ([delta]FA-H) lung parenchyma was calculated to describe the cyclic changes of FA during uninterrupted PCV at different PEEP levels.

Experimental Protocol

After instrumentation, the pigs were transferred to the CT unit and positioned supine in the CT scanner. Each animal was studied before and 2 h after lung lavage. The inspiratory PV curve was obtained as described above (healthy and injured lungs), and the LIP was identified graphically. According to the study protocol, the ventilator was switched to PCV (FI[O.sub.2] of 1.0) and the PEEP was varied through the experiment to the identified LIP, LIP + 5, and LIP - 5. To achieve standardized conditions, we performed a recruitment maneuver (continuous positive airway pressure of 40 mbar for 30 s) before each measurement. The inspiratory pressure was adjusted to 20 mbar above the PEEP, and the inspiratory time was set to 50% of the whole respiratory cycle during the experiment. The RR was targeted to achieve normocapnia. The sequences of LIP, LIP + 5, and LIP - 5 in healthy and injured lungs were randomized by the methods of blocks using statistical software (BIAS Version 7.40; Epsilon-Verlag; Hochheim-Dannstadt, Germany). The following measurements were obtained approximately 15 min after each new adjustment of the ventilatory pattern.

Measurements

At each PEEP setting (LIP, LIP - 5, and LIP + 5), parameters were obtained after a stabilization period of 15 mm: hemodynamics (heart rate [HR], mean arterial pressure [MAP], mean pulmonary artery pressure [MPAP], central venous pressure [CVP]), arterial and mixed venous blood gases (P[O.sub.2], P[O.sub.2], pH, oxygen saturation, hemoglobin, oxygen content), and a dCT scan series over 15 s.

Statistical Analysis

Values are given as median and 25th and 75th percentiles (interquartile range [IQR]) unless otherwise specified. Intraindividual effects of different PEEP (LIP, LIP + 5, and LIP - 5) settings during PCV in healthy and in lung-injured animals on FA-NV, FA-PV, FA-V, FA-H, [delta]FA-NV, [delta]FA-PV, [delta]FA-V, [delta]FA-H, Pa[O.sub.2], PaC[O.sub.2] arterial pH, Qs/Qt, and hemodynamics were analyzed using nonparametric testing (Friedman analysis of variance and Wilcoxon signed-rank test with Bonferroni correction for multiple testing [BIAS Version 7.40; Epsilon-Verlag]); p [greater than or equal to] 0.05 was considered statistically significant. After induction of lung injury, we determined the following: (1) the relationship of the mean FA (FA-NV, FA-PV, FA-V, FA-H) was compared to VT per kilogram of body weight, RR, PEEP, dynamic lung compliance (Cdyn), and Qs/Qt; and (2) the differences between end-expiratory and end-inspiratory FA ([delta]FA-NV, [delta]FA-PV, [delta]FA-V, [delta]FA-H) were compared to VT per kilogram of body weight, RR, PEEP, and Cdyn using linear regression analysis. The correlation between Qs/Qt and FA-NV was expressed graphically by a Bland-Altman plot. (17)

RESULTS

The study procedure was completed in all seven pigs before lung lavage. One pig was hemodynamically unstable after lung lavage, and therefore measurements after lung lavage were excluded from the evaluation. The average ([+ or -] SEM) number of lung lavages to induce lung injury was 2.7 [+ or -] 0.5 at a mean volume of Ringers solution of 1,333 [+ or -] 258 mL. The inspiratory PV curves before and after lung lavage are shown in Figure 2. The inspiratory PV curve before lung lavage (n = 7) revealed a median LIP of 7 mbar (IQR, 4 to 9 mbar) and after lung lavage (n = 6) of 13 mbar (IQR, 10 to 15 mbar).

Hemodynamics and Gas Exchange

Data obtained before and after lung lavage are given in Table 1. Before and after lung lavage, CVP and HR were comparable at all PEEP levels. After lung lavage, MAP was higher at PEEP at LIP - 5 compared to PEEP levels at LIP and LIP + 5. MPAP increased at higher PEEP levels (LIP + 5) in injured animals (p = 0.02). In healthy animals, increasing the PEEP to LIP + 5 led to no further increase of Pa[O.sub.2] (p = 0.30) but to a reduction of Qs/Qt (p = 0.03). In injured lungs, ventilation at PEEP of LIP-5 led to a decrease in Pa[O.sub.2] (p = 0.02) and an increase in Qs/Qt (p = 0.02). Increasing the PEEP to LIP + 5 after lung lavage was followed by an increase in Pa[O.sub.2] (p = 0.003) and a decrease of Qs/Qt (p = 0.02). Pac[O.sub.2] and arterial pH were unchanged during the experiment before and after lung lavage.

Mean Nonventilated, Poorly Ventilated, Ventilated, and Hyperinflated Lung Area Measured by dCT at Different PEEP Settings During PCV

Figure 3 shows the four lung compartments defined by CT densitometry before and after lung lavage during PCV at different PEEP levels. Before lung lavage, dCT showed atelectasis ranging from 10 to 33% of the total lung area. In healthy lungs, increasing the PEEP to LIP + 5 reduced atelectasis and poorly ventilated lung, whereas PEEP levels less than the LIP had no influence on atelectasis and poorly ventilated lung. This effect was paralleled by unchanged (LIP - 5) and increased (LIP + 5) ventilated lung area. The hyperinflated lung area increased\ with higher PEEP levels and decreased with lower PEEP when compared to a PEEP setting at the LIP.

After lung lavage, ventilation at a PEEP of LIP - 5 led to an increased amount of atelectasis and was paralleled by an increase of Qs/Qt (Table 1). The amount of poorly ventilated lung in an injured lung was unaffected when PEEP was above or below the LIP. The ventilated lung area in damaged lungs increased with a higher PEEP (LIP + 5) and decreased when PEEP was set to LIP - 5. Setting the PEEP to LIP + 5 increased the amount of hyperinflated lung area.

Cyclic Changes of Nonventilated, Poorly Ventilated, Ventilated, and Hyperinflated Lung Area During the Respiratory Cycle at Different PEEP Settings With PCV

Figure 4 illustrates the measured cyclic changes of the four CT-defined functional lung compartments (nonventilated, poorly ventilated, ventilated, and hyperinflated lung parenchyma) at different PEEP levels after lung lavage within the respiratory cycle during uninterrupted PCV (pig 2). In this representative case, the mean FA-NV decreased, and vice versa the total FA-V increased when PEEP increased. However, the analysis of the time course of each functional lung compartment showed that the increase of PEEP alone had only a minor influence on repeated R/D hyperinflation. Figure 5 shows the [DELTA]FA-NV, [DELTA]FA-PV, [DELTA]FA-V, and [DELTA]FA-H before and after lung lavage of all pigs recorded within three successive respiratory cycles.

[FIGURE 5 OMITTED]

In healthy animals, [DELTA]FA-NV increased when PEEP was set below the LIP during PCV. PEEP above the LIP tended to decrease [DELTA]FA-NV but reached no statistical significance. The cyclic changes of ventilated lung area increased, but decreased for hyperinflated lung at a PEEP below the LIP. A higher PEEP (LIP + 5) had no influence on the amount of cyclic changes of ventilated and hyperinflated lung, whereas the amount of cyclic changes of poorly ventilated lung decreased. After induction of lung injury, PEEP settings of LIP - 5 or LIP + 5 had no influence on [DELTA]FA-NV, [DELTA]FA-PV, [DELTA]FA-V, and [DELTA]FA-H during PCV with a pressure amplitude of 20 mbar above PEEP.

Correlation Between Mean FA, [DELTA]FA, and Respiratory Parameters During PCV After Induction of Lung Injury

Mean FA-NV as well as mean FA-V showed a good correlation with PEEP (Fig 6, 7). However, there was no correlation between mean FA of the lung compartments (FA-NV, FA-PV, FA-V, FA-H) and VT per kilogram of body weight, RR, or Cdyn (Table 2; Fig 6, 7).

[FIGURES 6-7 OMITTED]

The correlation analysis between end-expiratory and end-inspiratory differences of the FA ([DELTA]FA-NV, [DELTA]FA-PV, [DELTA]FA-V, [DELTA]FA-H) during the respiratory cycle and the applied VT per kilogram of body weight, RR, PEEP, and calculated Cdyn are shown in Table 2 and Figures 8 and 9. We found a close linear correlation when the [DELTA]FA of ventilated and nonventilated lung was compared to VT per kilogram of body weight and calculated Cdyn (Fig 8, 9). The [DELTA]FA of poorly ventilated and hyperinflated lung vs VT per kilogram of body weight, RR, and calculated Cdyn showed only a moderate linear correlation (Table 2). However, the comparison of PEEP vs the [DELTA]FA of all four different lung compartments was weak in each case (Table 2; Fig 8, 9). We concluded that the [DELTA]FA of nonventilated and ventilated lung during ongoing PCV is influenced from resulted VTS and reflects repeated regional lung R/D.

[FIGURES 8-9 OMITTED]

Correlation of the Mean Nonventilated Lung Area and the Calculated Qs/Qt

Analysis of the mean FA-NV vs the calculated venous admixture showed a linear correlation in healthy (r = 0.89, p < 0.0001; 95% confidence interval [CI], 0.77 to 0.95 [p = 0.95]) and injured lungs (r = 0.93, p < 0.0001; 95% CI, 0.83 to 0.96 [p = 0.95]). The relationship between Qs/Qt and the mean FA-NV of healthy and injured lungs showed also a linear correlation (Fig 10 left, a; r = 0.91, p < 0.0001; 95% CI, 0.80 to 0.96 [p = 0.95]). At the Bland-Altman transformation (Fig 10, right, b), mean FA-NV received from dCT imaging overestimated Qs/Qt approximately 2.1 [+ or -] 4.1%.

[FIGURE 10 OMITTED]

DISCUSSION

This experimental study investigates changes of functional lung compartments defined by CT densitometry and derived by dynamic imaging technique during uninterrupted PCV at different PEEP settings with high temporal resolution. Simultaneously, within three successive respiratory cycles, this approach enabled the quantification of changes in respiratory settings on the net effect of lung recruitment; the identification and quantification of repeated recruitment, derecruitment, and hyperinflation depending on ventilator settings; and the noninvasive evaluation of Qs/Qt. In experimental injured lungs, cyclic recruitment, derecruitment, and hyperinflation were evident in each ease.

Cyclic collapse and reopening of lung units and overdistention of open lung units during each respiratory cycle induce stress and strain to the lung fibroskeleton and may activate an inflammatory cascade that leads to lung edema and lung inflammation. (18-21) R/D is thought to be governed by critical opening and dosing pressures. Recruitment of a closed lung unit occurs when the pressure acting to distend the unit reaches a critical value, at which point the subtended unit suddenly is ventilated. Consequently, it is thought that during deflation, the unit suddenly closes when its distending pressure drops below a certain critical value, which is lower compared to the critical opening pressure for that unit. The amount of alveolar recruitment can be measured as reaeration of previously nonventilated lung regions (and vice versa for derecruitment) with CT imaging according to the reopening and collapse hypothesis. (22,23) However, previous studies (3-10) demonstrated that recruitment and derecruitment proceed along the entire limb of the PV curve and that in injured lungs a wide range of opening and closing pressures coexist. Alveolar recruitment was observed up to end-inspiratory pressure or end of tidal insufflation of the PV curve. Because of the wide range of opening and closing pressures along the PV curve in heterogeneous injured lung, it is still under debate whether defined PEEP levels are able to prevent the whole lung from repeated R/D within the respiratory cycle. (8,9) However, the measurement of repeated R/D phenomenon during ongoing mechanical ventilation is difficult. (18-23) One possible experimental technique is reported by Baumgardner et al, (24) who demonstrated in an experimental study large oscillations in Pa[O.sub.2] dependent from ventilator settings. They suggest that a substantial fraction of collapsed or flooded alveoli are reopened with every respiratory cycle, and that various respirator settings (not a single PEEP increase) influence the amount of cyclic Pa[O.sub.2] oscillations. Schiller et al (25) demonstrated repeated R/D within the respiratory cycle by in vivo video microscopy. They documented alveolar instability after induction of acute lung injury and suggested that alveoli exhibited different types of mechanical behavior, and the altered mechanics of alveoli resulted in repeated R/D. Although PEEP has been proposed to limit cyclic alveolar collapse, Schiller et al (25) as well as Baumgardner et al (24) were unable to document a close correlation between PEEP levels and alveolar stability or Pa[O.sub.2] oscillations. The in vivo microscopy is limited to a subset of subpleural alveoli, and Schiller et al (25) may have observed lung regions that recruit at lower peak pressures than those required to recruit dependent lung regions. Neumann et al (26) investigated the dynamics of R/D during mechanical ventilation with repeated static CT scans. The data suggested that cyclic change of atelectasis during the respiratory cycle was present, but their temporal resolution was too low (1.25 Hz), and they reported no quantification of this behavior. (26) Our data support the findings obtained from any of these techniques; in addition to that, dCT offers an approach in a possible clinical context. The dCT technique used in the present study investigated a representative single slice of the lung, and we were able to detect repeated WD during uninterrupted mechanical ventilation by calculating the [DELTA]FA. The repeated R/D after lung lavage was dependent from the Cdyn and end-tidal volume but was nearly independent from an increase of PEEP alone.

Temporal dynamics also have to be taken into account when pulmonary R/D is observed. Bates and Irvin (27) investigated the assumption that R/D is a time-dependent phenomenon using a lung model. They considered that recruitment and derecruitment is not instantaneous, but rather has a time scale associated with it that may extend from seconds to minutes; they concluded that the time can influence the extent of WD. Neumann et al (26) reported that in the lavage-induced lung injury, a short delay occurred before atelectasis started to increase rapidly at the beginning of an expiration. During the first 4 s, only 30% of R/D occurred in the lavage model. We also demonstrated in previous studies (28,29) the existence of different pulmonary time constants in the same animal model as used in the present study. Markstaller et al (28) reported that slow inspiratory time constants of the recruitable lung after lung lavage ranged from 8 to 16.8 s, and slow expiratory time constants ranged from 7.1 to 34.2 s. In respect to the above-discussed factors, there are significant shortcomings of static CT imaging maneuvers when the dynamic behavior of R/D, especially of variations in RR, time ratio, or different flow patterns have to be taken into account. Breath-hold maneuvers at end-expiration and end-inspiration with static CT scanning may introduce a significant source of bias into the findings: the time is longer when compared to the respiratory cycle during ongoing ventilation. Nonventilated lung may be underestimated at end-inspiration and overestimated at end-expiration. Detecting hyperinflated lung during static CT scanning also poses problems: lung lavage led to an inhomogeneous injury that was mainly confined to the dependent two thirds of the lung. (30,31) As the imaging procedure took up to > 10 s, redistribution of gas between areas with different time constants has to be taken into account. Therefore, static CT scanning compared with uninterrupted positive pressure ventilation may underestimate lung areas classified as overinflated.

Markstaller et al (13) reported a close linear relationship between the amount of atelectasis (detected by dCT) and calculated Qs/Qt but overestimated them approximately 3.4 [+ or -] 7.5%. Our study confirms these findings, as we found an even closer relationship between the mean FA-NV and calculated Qs/Qt, and underscore the utility of the mean FA of atelectasis obtained by dCT as a global parameter of gas exchange. However, Qs/Qt was also overestimated by dCT.

Limitations of the Study

In the present study, atelectatic lung parenchyma was found in healthy lungs before lung lavage. The amount of atelectasis in healthy lungs varied; however, this phenomenon is already known in ventilated animals as well as in anesthetized patients. (14,26,30-32) The presence of atelectasis might be pronounced in animals receiving ventilation at FI[O.sub.2] of 1.0, low PEEP, and in supine position for several hours.

When extrinsic PEEP was set below the LIP after lung lavage, RRs were increased to achieve normocapnia, and this may have led to an increase of total PEEP due to an incomplete expiration. However, we had no data if the expired flow did not return to zero at end-expiration at the PEEP of LIP--5, and measurement of intrinsic PEEP was not part of the protocol.

PEEP at LIP + 5 decreased the total amount of atelectasis but had no influence on the repeated R/D during ongoing ventilation. The applied VTs between the PEEP at LIP and at LIP + 5 were kept constant. It remains unclear whether further increases in PEEP above LIP + 5 would have led to a decrease of the repeated WD and the VTs. However, in the present study we found no correlation between PEEP and repeated WD, whereas Cdyn and VT had a significant influence on R/D.

Another limitation is the lack of total end-expiratory and end-inspiratory lung volumes during the PEEP interventions. Measurement of the total lung volume by CT imaging required a static spiral CT technique in apnea at end-expiration and end-inspiration and excluded dynamic analyses during ongoing mechanical ventilation as well as influences of respirator settings (inspiratory/expiratory ratio, RRs, flow rates).

A possible reason for the overestimation of Qs/Qt is the lack of differentiation between atelectatie lung area, blood, and interstitial lung water with CT imaging, therefore an overestimation of atelectatic lung parenchyma. Lu and coworkers (32) reported that a single supradiaphragmatic CT section often underestimates recruitment of the whole lung, and Bletz et al (33) reported that atelectatic lung area derived from single juxtadiaphragmatic dCT imaging overestimated atelectasis approximately 9.8% when compared to spiral CT imaging of the whole lung.

CONCLUSION

The presented diagnostic approach using dCT on a clinically available CT scanner offers a functional investigation during uninterrupted mechanical ventilation: (1) the quantification of various respirator settings on the total amount of recruitment, derecruitment, and hyperinflation of injured lungs; (2) a simple in vivo detection and quantification of repeated WD during mechanical ventilation in a possible clinical setup; and (3) a good prediction of Qs/Qt without invasive catheter placement. The detection and quantification of repeated WD and hyperinflation defined as the difference between end-expiration and end-inspiration ([DELTA]FA) represents a new and simple method for lung CT interpretation. Theoretically, the amount of this dynamic component of WD may reflect pulmonary stress and strain, which has to be further investigated. Thus, dCT appears as a promising functional tool to adapt respiratory treatment without the necessity for new hardware.

REFERENCES

(1) Kumar A, Falke KJ, Geffin B, et al. Continuous positive-pressure ventilation in acute respiratory failure. N Engl J Med 1970; 238:1430-1436

(2) Ranieri VM, Eissa NT, Corbeil C, et al. Effects of positive end-expiratory pressure on alveolar recruitment and gas exchange in patients with the adult respiratory distress syndrome. Am Rev Respir Dis 1991; 144:544-551

(3) Mergoni M, Volpi A, Bricchi C, et al. Lower inflection point and recruitment with PEEP in ventilated patients with acute respiratory failure. J Appl Physiol 2001; 91:441-450

(4) Hickling KG. Best compliance during decremental, but not incremental, positive end-expiratory pressure trial is related to open-lung positive end-expiratory pressure: a mathematical model of acute respiratory distress syndrome lungs. Am J Respir Crit Care Med 2001; 163:69-78

(5) David M, Bletz C, David S, et al. Analysis of the static pressure volume curve of the lung in experimentally induced pulmonary damage by CT-densitometry. Fortschr Rontgenstr 2005; 177:751-757

(6) Downie JM, Nam A, Simon B. Pressure-volume curve does not predict steady-state lung volume in canine lavage lung injury. Am J Respir Crit Care Med 2004; 169:957-962

(7) Jonson B, Richard JC, Straus C, et al. Pressure-volume curves and compliance in acute lung injury: evidence of recruitment above the lower inflection point. Am J Respir Crit Care Med 1999; 159:1172-1178

(8) Pelosi P, Goldner M, McKibben A, et al. Recruitment and derecruitment during acute respiratory failure: an experimental study. Am J Bespir Crit Care Med 2001; 164:122-130

(9) Crotti S, Mascheroni D, Caironi P, et al. Recruitment and derecruitment during acute respiratory failure: a clinical study. Am J Respir Crit Care Med 2001; 164:131-140

(10) Luecke T, Meinhardt J, Hermann P, et al. Setting mean airway pressure during high-frequency oscillatory ventilation according to the static pressure-volume curve in surfactant-deficient lung injury: a computed tomography study. Anesthesiology 2003; 99:1313-1322

(11) Malbouisson LM, Muller JC, Constantin JM, et al. CT scan ARDS study group: computed tomography assessment of positive end-expiratory pressure-induced 'alveolar recruitment in patients with acute respiratory, distress syndrome. Am J Respir Crit Care Med 2001; 163:1444-1450

(12) Markstaller K, Kauczor HU, Eberle B, et al. Multi-rotation CT during continuous ventilation: comparison of different density areas in healthy lungs and in the ARDS lavage model. Fortschr Rontgenstr 1999; 170:575-580

(13) Markstaller K, Kauczor HU, Weiler N, et al. Lung density distribution in dynamic CT imaging correlates with oxygenation in ventilated pigs with lavage ARDS. Br J Anaesth 2003; 91:699-708

(14) Maggiore SM, Brochard L. Pressure-volume curve in the critically ill. Curr Opin Crit Care 2000; 6:1-10

(15) Slinger PD, Kruger M, McRae, et al. Relation of the static compliance curve and positive end-expiratory pressure on oxygenation during one-lung ventilation. Anesthesiology 2001:95:1096-1102

(16) Markstaller K, Arnold M, Doebrich M, et al. A software tool for automatic image-based ventilation analysis using dynamic chest CT-scanning in healthy and in ARDS lungs. Fortschr Rontgenstr 2001; 173:830-835

(17) Altman DG, Bland JM. Measurements in medicine: the analysis of method comparison studies. Statistician 1983; 32:307-318

(18) Muscedere JG, Mullen JBM, Gan K, et al. Tidal ventilation at low airway pressures can augment lung injury. Am J Respir Crit Care Med 1994; 149:1327-1334

(19) Gattinoni L, Pelosi P, Crotti S, et al. Effects of positive end-expiratory pressure on regional distribution of tidal volume and recruitment in adult respiratory distress syndrome. Am J Respir Crit Care Med 1995; 151:1807-1814

(20) Kolobow T, Moretti MP, Fumagalli R, et al. Severe impairment in lung function induced by high peak airway pressure during mechanical ventilation: an experimental study. Am Rev Respir Dis 1987; 135:312-315

(21) Gattinoni L, Carlesso E, Cadringher P, et al. Physical and biological triggers of ventilator-induced lung injury and its prevention. Eur Respir J Suppl 2003; 47:15-25

(22) Gattinoni L, D'Andrea L, Pelosi P, et 'al. Regional effects and mechanism of positive end-expiratory pressure in early adult respiratory distress syndrome. JAMA 1993; 269:2122-2127

(23) Gattinoni L, Pesenti A, Bombino M, et al. Relationships between lung computed tomographic density, gas exchange, and PEEP in acute respiratory failure. Anesthesiology 1988; 69:824-832

(24) Baumgardner JE, Markstaller K, Pfeiffer B, et al. Effects of respiratory rate, plateau pressure, and positive end-expiratory pressure on Pa[O.sub.2] oscillations after saline lavage. Am J Respir Crit Care Med 2002; 166:1556-1562

(25) Schiller H, McCann U, Carney D, et al. Altered alveolar mechanics in the acutely injured lung. Crit Care Med 2001; 29:1049-1055

(26) Neumann P, Berglund JE, Mondejar EF, et al. Effect of different pressure levels on the dynamics of lung collapse and recruitment in oleic-acid-induced lung injury. Am J Respir Crit Care Med 1998; 158:1636-1643

(27) Bates J, Irvin C. Time dependence of recruitment and derecruitment in the lung: a theoretical model. J Appl Physiol 2002; 93:705-713

(28) Markstaller K, Eberle B, Kauczor HU, et al. Temporal dynamics of lung aeration determined by dynamic CT imaging in a porcine model of ARDS. Br J Anaesth 2001; 87:459-768

(29) Doebrich M, Markstaller K, Karmrodt J, et al. Analysis of discrete and continuous distributions of ventilatory time constants from dynamic computed tomography. Phys Med Biol 2005; 50:1659-1673

(30) Neumann P, Berglund JE, Fernandez E, et al. Dynamics of lung collapse and recruitment during prolonged breathing in porcine lung injury. J Appl Physiol 1998; 85:1533-1543

(31) Luecke T, Both H, Herrmann P, et al. PEEP decreases atelectasis and extravascular lung water but not lung tissue volume in surfactant-washout lung injury. Intensive Care Med 2003; 29:2026-2033

(32) Lu Q, Malbouisson LM, Mourgeon E, et al. Assessment of PEEP-induced reopening of collapsed lung regions in acute lung injury: are one or three CT sections representative of the entire lung? Intensive Care Med 2001; 27:1504-1510

(33) Bletz C, Markstaller K, Karmrodt J, et al. Quantification of atelectases in artificial respiration: spiral versus dynamic single-slice CT. Fortschr Rontgenstr 2004; 176:409-416

* From the Departments of Anesthesiology (Drs. M. David, Karmrodt, and Herweling) and Radiology (Drs. Bletz and S. David) Johannes Gutenberg-University, Mainz, Germany; Department of Radiology (Dr. Kanezor), German Cancer Research Center, Heidelberg Germany. and Department of Anesthesiology (Dr Markstaller) Inselspital, University of Berne, Berne, Switzerland.

This study was funded by a German Research Council (DFG)

Grant: FOR 474/1//Ma 2398/13-1.

Manuscript received February 11, 2005; revision accepted June 25, 2005.

Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (www.chestjournal. org/misc/reprints.shtml).

Correspondence to: Matthias David, MD, Department of Anesthesiology, Johannes Gutenberg-University, Langenbeckstr. 1, D-55131 Mainz, Germany. e-mail: david@mail.uni-mainz.de

COPYRIGHT 2005 American College of Chest Physicians
COPYRIGHT 2005 Gale Group