| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
From the Departments of *Anesthesiology and Surgical Intensive Care Unit, University Hospital, RWTH Aachen; Applied Mathematics II, RWTH Aachen, Aachen, Germany; and Surgical Intensive Care Unit, Klinikum Berlin Buch, Berlin, Germany.
Address correspondence to Dietrich Henzler, MD, PhD, Department of Anesthesiology, University Hospital, RWTH Aachen, Pauwelsstr. 30, D-52074 Aachen, Germany. Address e-mail to mail{at}d-henzler.de.
 |
Abstract |
|---|
 Top Abstract IntroductionMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
METHODS: Six anesthetized pigs with lung lavage-induced acute lung injury were ventilated with lung-protective settings. We measured the effects of a standard recruitment maneuver on hysteresis area and ratio obtained from the PV curve and on the stress index obtained from the Paw-t curve and correlated this with aerated and nonaerated lung volumes as measured by multislice computed tomography.
RESULTS: Hysteresis area and ratio correlated with aerated lung volume (r = 0.886). The recruitment maneuver resulted in an increase in aerated (+12%) and a decrease (–18%) in nonaerated lung. Hysteresis area correlated with alveolar recruitment, represented by an increase in aerated lung (r = 0.886) and a decrease in nonaerated lung (r = –0.829) during tidal ventilation. The stress index was always >1 and indicated tidal hyperinflation only. Values did not change after the recruitment maneuver and did not correlate with any other lung volume.
CONCLUSIONS: Parameters derived from the PV curve may help in characterizing the lung aeration of the lung and in indicating recruitment. In the presence of lung-protective ventilator settings, the stress index derived from the Paw-t curve was not able to indicate recruitment.
|
Introduction |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The role of recruitment maneuvers remains controversial, as no clinical study using oxygenation improvement as an indicator for successful recruitment has been able to show improved survival in patients with a deliberate recruitment strategy (5–7). Therefore, parameters of respiratory mechanics, which are influenced by recruitment and derecruitment, have been investigated in theoretical models (8) and clinical studies (9). Adjusting the inspiratory pressures above the lower inflection point of the respiratory system pressure–volume curve (PV curve) has resulted in improved outcome of patients with ARDS (10), although ventilation pressures ranging between the lower and upper inflection points do not minimize tidal recruitment or hyperinflation (9).
One parameter is the profile of the pressure–time curve (Paw-t curve) during constant-flow ventilation, which has been shown to predict lung-protective ventilation (11). This "stress index" is the coefficient of an exponential equation indicating the slope progression of the Paw-t curve. A convex Paw-t curve indicates increasing compliance, representing tidal recruitment, whereas an upward concave shape in the Paw-t curve indicates decreasing compliance, representing tidal hyperinflation. This has been confirmed by a CT study in a lavage model of ALI in pigs (12). Also, a change of the stress index after a standard recruitment maneuver was noted.
The aim of this study was to assess whether the parameters of dynamic respiratory mechanics, derived from the PV and Paw-t curves, could indicate changes in lung aeration as determined by lung CT.
|
METHODS |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Esophageal pressure was measured by a balloon catheter (International Medical, Kleve, Germany) connected to the ventilator (Galileo Gold, Hamilton Medical, Rhäzüns, Switzerland), and correct placement was verified using the occlusion technique (15). Gas flow and airway pressure were measured between the tracheal tube and the Y-piece of the ventilator circuit by a differential pressure transducer connected to the ventilator, and were recorded with a sampling rate of 70 Hz (Datalogger 3.27). Tidal volume (VT) was derived from the integrated flow signal. All expiratory volumes were corrected for gas composition, intrapulmonary humidification, and temperature, as proposed by Jonson et al. (16). A calibration factor for expiratory volumes was calculated so that, at steady-state, expired volumes became identical to inspired volumes. For validation, expiratory VT (VTEX) was corrected (VTCORR) on the basis of Boyle’s law as proposed by Gattinoni et al. (17):
|
|
for TA = ambient temperature, TB = body temperature, PB = barometric pressure, PH2O = vapor pressure at TA. VTCORR highly correlated with inspiratory VT (r = 0.921), exhibiting no difference (P = 0.346).
The PV curve, starting from zero end-expiratory pressure, was constructed by inflating the lungs with a constant pressure increase of 2 cm H2O per second, up to a maximum pressure of 45 cm H2O, keeping a plateau at 35 cm H2O for 1 s and deflating with the same rate with a 0.6 s pause every 5 cm H2O downwards to the preset positive end-expiratory pressure (PEEP). Total maneuver time was 54 s. Expiratory volumes were corrected as described. Inflation and deflation curves were separately analyzed by fitting the data to the equation (18).
|
|
The goodness of fit was excellent for inflation (range of R2 = 0.9915–1.0) and for deflation curves (range of R2 = 0.9188–1.0). Maximum volume (VMAX) was the corresponding volume at end-inflation pressure. We calculated the area between inflation and deflation curves, which is known as the hysteresis area (HA) (19,20):
|
|
By indexing to the PV product of the entire PV loop, the hysteresis ratio (HR) (Fig. 1) is obtained as
|
|
|
The HR has been described to be a constant if there is no change in pulmonary compliance (21).
For calculation of the stress index, the Paw-t curve was fitted to an exponential equation in the form of
|
|
where a, b, and c are the fitting parameters (11). a represents the slope of the curve at t = 1 s and c is the y-axis intercept. The coefficient b is a dimensionless number describing the curvature. For b > 1, it will take an upward, concave shape, indicating tidal over-distension; for b < 1, it will take a downward convex shape, indicating tidal recruitment. For b = 1, a straight line will result (Fig. 2). Fitting was done on airway and pressures to analyze respiratory system and lung profiles, respectively. To omit possible on–off flow transients, only data points obtained from 50 ms after beginning to 50 ms before end of inspiration (Fig. 2) were used. The fitting procedure was excellent with the range for R2 = 0.9935–1.0.
|
CT Scan Analysis
Multislice CT scans of the whole lung (Siemens Somatom Sensation 16, Erlangen, Germany) were obtained at end-expiratory and inspiratory occlusion (1,22). Scan parameters and image reconstruction were chosen as previously reported (23) using commercial software (Syngo, Siemens). Lung was characterized as hyperinflated (VHYP) from –1000 to –900 HU, as aerated (VAER) from –900 to –100 HU, and as nonaerated (VNON) from –100 to +100 HU (22). In this investigation, we did not differentiate between normally and poorly aerated lung to reduce the number of statistical procedures. The gas content was calculated as VGAS = [V x HUmean]/(–1000) (22). All end-inspiratory and end-expiratory volumes were calculated separately. The VT changes (
V) were calculated by subtracting end-expiratory from end-inspiratory values, with VNON representing tidal recruitment and VHYP representing tidal hyperinflation.
Data were gathered before a standard recruitment maneuver (23). The recruitment maneuver was performed by inflating the lungs three times with a continuous positive pressure of 45 cm H2O for 40 s each (24–27). After the recruitment maneuver, tidal ventilation was resumed with unchanged ventilatory settings and the measurements were repeated. After the experiment, the animals were killed.
Statistical Analysis
Univariate analysis was performed to calculate median and interquartile range by Tukey’s Hinges. Values before and after the recruitment maneuver were compared using Wilcoxon’s exact signed rank test, and the fractional change for any parameter X was calculated as FC(X) = XAFTER/XBEFORE – 1. Statistical significance was accepted for P values <0.05; the calculated P values were considered explorative.
Since the stress index is derived from an inspiratory curve and the HA is in large part determined by inspiration, the correlation between both parameters and the differently aerated CT-derived lung volumes during inspiratory hold was calculated using Spearman’s rank coefficient (r). Second, the VT changes within the lung were calculated for each compartment by the difference of inspiratory and expiratory CTs. Stress index and hysteresis were investigated to indicate and predict the changes in tidal over-distension or recruitment after the recruitment maneuver. Alveolar recruitment was defined as a decrease in nonaerated lung volume, and over-distension as an increase in VHYP. All calculations were done using SPSS WIN 14.0 (SPSS Corp., Chicago, IL).
|
RESULTS |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
VNON and VHYP indicated tidal over-distension and tidal recruitment happening in parallel in different compartments (Table 2). After the recruitment maneuver, tidal recruitment was reduced, but over-distension was not. No change was observed in ventilation to aerated lung or the tidal intrathoracic gas volume (Table 2).
|
|
Parameters Derived from the PV Curve
Both HA and VMAX increased after the recruitment maneuver (Table 3). A positive correlation was found between VAER and HA (r = 0.886) and HR (r = 0.886), before and after the recruitment maneuver (r = 0.771 for both). The increase in HA correlated with the decrease in VNON (r = –0.829) and also with the increase in VAER (r = 0.886). No correlations were found with VHYP.
|
VMAX was increased after the recruitment maneuver. The FC of VMAX correlated inversely with the decrease of VNON (r = –0.886).
Parameters Derived from the Paw-t Curve
The stress index was close to 1, regardless of whether it was calculated for the airway or the transpulmonary pressure gradient (Table 3). Since the value was always >1, it indicated tidal hyperinflation. The stress index correlated inversely with VHYP before (r = –0.9) but not after (r = –0.6) the recruitment maneuver. Even though the FC in stress index and VHYP was not significant, both parameters correlated inversely with each other (r = 1.0, Fig. 3). The stress index did not correlate with the decrease in VNON (r = –0.2). No correlations were found between the stress index and VT changes of any compartment.
|
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Usually, ventilatory settings are adjusted according to oxygenation parameters, since impaired oxygenation is the main symptom in ALI (31). However, the partial oxygen pressure is influenced by pulmonary blood flow (32), systemic circulation (33,34), and the distribution of ventilation and perfusion (35,36). We have shown previously that lung mechanics are superior to oxygenation parameters in indicating lung aeration and recruitment (14). Several studies in patients with ARDS have demonstrated increases in oxygenation using various recruitment strategies, but no improvement in outcome when the ventilatory settings were adjusted according to oxygenation parameters (5,7,37,38). In contrast, studies targeted at individual patients’ respiratory mechanics have shown promising results (10).
Information regarding the diseased lung has been obtained from parameters derived from PV curves (9,39–41). However, these simple interpretation models of the PV curve have lately been questioned, since mathematical modeling (8) and clinical studies (9) have shown evidence of recruitment above the lower inflection point. In one approach, the PV curve was used to center VTs for optimizing lung protective ventilation in dependence of mathematically derived parameters (42), but no difference in biotrauma was noted compared with the ARDS network settings. For mathematical analysis, the authors used the Venegas-Harris equation (18), which has been used for experimental (30,43) and clinical studies (44,45) but which might be erroneous if corrected for asymmetric data (46). We therefore used the uncorrected equation, considering the systematic error from fitting to nonsigmoid data.
To our knowledge, HA and HR have not been evaluated as indicators of recruitment. We hypothesized that these are representative of both inflation and deflation, and would thus enable a better characterization of the diseased lung. Hysteresis has been shown to be independent of lung size or species (20), and to represent the viscoelastic properties of the lungs in different pathologic conditions (21). The HA is determined by two major mechanisms: stress relaxation of tissue after deformation and viscoelastic forces at the gas–liquid interface. This is consistent with the correlation of both parameters with aerated lung, where surfactant-depleted nearly collapsed alveoli with high surface tension exist parallel with normally inflated alveoli. If the elastance of nonaerated lung regions is many times greater than the elastance in aerated lung regions, then the elastance of the respiratory system is determined primarily by the elastance of the nonaerated lung. If HA and HR depend on the viscoelastic properties of the lung, they must be sensitive to changes in the most determinant lung regions, the atelectatic ones.
The stress index has been used to indicate lung-protective ventilation, with a reduced release of inflammatory mediators in excised lungs if ventilation resulted in b values close to 1. This has been attributed to the ability of the stress index to characterize continuing recruitment, resembling alveolar collapse and reopening during tidal ventilation. On the other hand, over-distension at end-inflation can be indicated by an upward concave curve, resulting in b values >1 (11). We hypothesized that, if a recruitment maneuver reduced tidal recruitment and hyperinflation, the stress index could be used to identify recruitment at the bedside.
Grasso et al. (12) have elegantly shown, in a porcine lavage model, that ventilation with a very high VT and low PEEP produces tidal recruitment, and ventilation with high VT and high PEEP produces significant hyperinflation. This resulted in b values <0.8 and >1.3, respectively. The authors used a deliberate strategy to attain certain b values, which were then correlated with radiographic findings.
In contrast, we chose one type of ventilation, which was not changed during the experiment, and looked at possible changes in the stress index induced by a recruitment maneuver. Throughout the experiment, lung-protective ventilator settings were used, a situation similar to a clinical setting. We found that the profile of the Paw-t curve did not change significantly after lung recruitment, presumably because the initial ventilator settings produced b values already close to 1. Although tidal hyperinflation and recruitment were present in parallel, the stress index indicated only over-distension (b values >1). It completely failed to indicate the reduced amount of nonaerated lung. The fraction of alveolar collapse and inspiratory reopening was reduced by 23%, but was not reflected by changes in the stress index. Surprisingly, an inverse correlation was found for the FCs with tidal ventilation. In those cases where VHYP increased, the b values decreased. This was caused by a redistribution of ventilation from nonaerated to hyperinflated lung regions, thereby still increasing end-inspiratory compliance. We speculate that the stress index offers little if any additional information about the diseased lung if principles of lung-protective ventilation are used.
The difference between hysteresis and the stress index might be attributed to the fact that hysteresis is a two-dimensional term, whereas the stress index is a one-dimensional term, which is not able to characterize tidal hyperinflation and tidal recruitment happening at the same time. The stress index offers the advantage that it can be obtained during normal ventilation, whereas hysteresis can only be measured by a special maneuver. However, neither parameter has been evaluated during assisted spontaneous breathing; thus, it is possible that both are helpful only during controlled ventilation.
|
CONCLUSIONS |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
Special thanks are due to Diana Bauer, MSc, Institute for Medical Statistics, University Hospital, RWTH Aachen, Germany, for statistical review of the manuscript.
|
Footnotes |
|---|
This study has been supported in part by an unrestricted research grant from Hamilton Medical, Rhazüns, Switzerland.
None of the authors has a financial relationship with a commercial entity, which has a vested interest in the outcome of the study.
Preliminary results of this study were presented at the 18th Annual Congress of the European Society of Intensive Care Medicine, September 27, 2005, Amsterdam, The Netherlands.
Reprints will not be available from the author.
|
REFERENCES |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
