Respiratory Physiology & Neurobiology 171 (2010) 225231

Contents lists available at ScienceDirect

Respiratory Physiology & Neurobiology

journal homepage: www.elsevier.com/locate/resphysiol

Gas distribution in a two-compartment model during volume or pressure ventilation: Role of elastic elements
Vittorio Antonaglia a, , Umberto Lucangelo a , Giuseppe Ristagno a , Simona Tantillo a , Massimo Ferluga a , Lorenzo Torelli a , Walter A. Zin b
a b

Department of Anesthesia and Intensive Care, Laboratory of Respiratory Biomechanics, University of Trieste, Cattinara Hospital, Strada di Fiume 447, I-34139 Trieste, Italy Carlos Chagas Filho Institute of Biophysics, Federal University of Rio de Janeiro, Rio de Janeiro, Brazil

a r t i c l e

i n f o

a b s t r a c t
The results of the studies on pulmonary gas distribution during constant-ow controlled-volume ination (VCV) and inspiratory constant pressure ination (PCV) in experimental studies are conicting. In a mathematical model, with the characteristics of two lung compartments including tissue viscoelastic properties, pulmonary gas distribution was tested by simulating PCV and VCV at same ination volumes. The compartmental distributions of the tidal volume were compared during CMV and PCV in different congurations obtained by changing the elastic and viscoelastic properties in each compartment, but maintaining the same total values of respiratory mechanics measured in patients. In all instances PCV resulted in a slightly higher air-trapping than in VCV mode. Heterogeneous elastic properties diverted most of the tidal volume towards the less compromised compartment. However, both ventilatory modes provided similar compartmental gas distribution, but during VCV compartmental peak pressures were higher in the sicker compartment respect to PCV. The use of PCV could grant a less remarkable pressure variability able to reduce the potential ventilator-associated lung injury. Moreover, the parameters measured during an end-inspiratory pause could not pinpoint unique characteristics for each conguration. 2010 Elsevier B.V. All rights reserved.

Article history: Accepted 16 March 2010 Keywords: Gas distribution Ventilation mode Ination Elastic and viscoelastic properties

1. Introduction Many studies and trials demonstrated that the ventilatory strategies recently designed to avoid exposing the lung to high pressure and volume might improve the outcome of patients with acute lung injury (The Acute Respiratory Distress Syndrome Network, 2000; Girard and Bernard, 2007). It is possible to provide pressurelimited ventilation either by pressure-targeted modes that limit the airway pressure to preset levels or by volume-cycled ventilation with precisely set values of pressure alarms and close monitoring of plateau pressure. Nevertheless, results regarding smaller morbidity and mortality of PCV in comparison to VCV are so far conicting (Young et al., 2004). Similar results were obtained in studies using respiratory mechanics and pulmonary CT to assess pulmonary gas distribution during PCV and VCV (Roth et al., 2004). Potential differences in volume distribution can be theoretically assessed by using respiratory models during constantow/pressure ventilation. In the past the end-inspiratory gas distribution was studied during VCV and PCV in a twocompartmental model by changing the values of resistance and

Corresponding author. Tel.: +39 040 3994471; fax: +39 040 912278. E-mail address: v.antonaglia@libero.it (V. Antonaglia). 1569-9048/$ see front matter 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.resp.2010.03.016

elastance in each compartment (Chatburn et al., 1994). However, the authors did not approach the volume distribution owing to viscoelastic properties. Indeed, an important component of pressure dissipation during breathing can be attributed to viscoelasticity (Mount, 1955; Sharp et al., 1967; Similowski et al., 1989; Jonson et al., 1993; Beydon et al., 1996). Recently, a two-compartment mathematical model demonstrated the importance of viscoelasticity on volume distribution both in normal subjects and in patients (Antonaglia et al., 2005). In the present investigation the gas distribution during PCV and VCV was studied using the same model (Antonaglia et al., 2005) and total resistance, elastance, and viscoelastic data corresponding to those actually measured in mechanically ventilated patients (Eissa et al., 1991). We hypothesized that, as the viscoelastic properties are ow-dependent, the different proles of inspiratory ow during VCV and PCV could modify gas distribution in the lung. To analyse the gas distribution during VCV and PCV four model congurations were tested: (a) homogeneous lesion determined by elastic and viscoelastic lung properties; (b) heterogeneous elastic properties; (c) heterogeneous elastic and viscoelastic properties; (d) heterogeneous viscoelastic properties. A different distribution of inspired gas during VCV and PCV could have a potential clinical impact, because it could affect ventilation/perfusion and, thus, gas exchange.

226

V. Antonaglia et al. / Respiratory Physiology & Neurobiology 171 (2010) 225231

Fig. 1. Upper panel: schematic representation of airway/lung mechanics. Lower panel: scheme of spring and dashpot two-compartmental model.

2. Methods 2.1. Mathematical model of the lung We used a viscoelastic two-compartmental mathematical model of the lung previously presented in detail and validated (Antonaglia et al., 2005). Briey, Fig. 1 depicts the mechanical analogue of the system, which includes a common branch and two compartments. The former is represented by a dashpot (Rtr ) reecting upper airway ow-dependent resistance. Each compartment (1 and 2) is made up of three elements: (i) a Newtonian dashpot (R1 or R2), that represents the central airway resistance; (ii) a spring (E1 or E2), representing static elastance of the system; and, (iii) a module (Maxwell body) comprising a dashpot (Rvisc 1 or Rvisc 2) and a spring (Evisc 1 or Evisc 2) that reects non-Newtonian behaviour due to tissue viscoelasticity and/or mechanical heterogeneities. The aforementioned parallel compartments are connected to Rtr that may correspond to any immediately proximal common pathway such as the trachea or a tracheal tube. The ow-dependence of Rtr was modelled using Rohrers equation (Rohrer, 1915; Behrakis et al., 1983). The compartmental elastances and resistances were considTable 1 Mechanical parameters used in modelling gas distribution. Conguration 1 2 3 4 Ri 12 cm H2 O/L/s 6.6 6.6 6.6 6.6 Re 12 cm H2 O/L/s 13.2 13.2 13.2 13.2 E1 cm H2 O/L/s 51.5 31 31 51.5

ered to be constant during the respiratory cycle, and an interaction between compartments 1 and 2 may occur at the branching point. In other words there was no interaction between the compartments except for the ow and pressure in the common pathway. Expiratory resistance doubled the inspiratory one. The model provided a point-by-point (at 100 Hz) numeric and graphic description of tracheal and compartmental ows, tracheal and alveolar pressures (elastic and viscoelastic components), and total and compartmental volumes during inspiration. An overall initial relaxation volume at the end expiration (Vr ) of 1 L was assumed and it was distributed according to each compartmental Vr , as determined by their elastic properties. A 4-s end-inspiratory pause was simulated by zeroing inspiratory ow. The same was done to generate a 4-s end-expiratory pause. Table 1 depicts the parameters used in the modelling of pulmonary mechanics (Eissa et al., 1991). These values can be considered as determining different theoretical conditions of disease degree, and elastic and viscoelastic behaviour in the two compartments. Arbitrarily, we adopted the following four congurations: (a) taking into account a uniform extension of illness, conguration 1 was modelled with the same values of elastic and viscoelastic parameters in both compartments; (b) in conguration 2 elastance in compartment 2 was 5-fold that in compartment 1, while the viscoelastic behaviour was equal in both compartments; (c) in conguration 3 both elastance and elastic component of viscoelasticity in compartment 2 were 5-fold those in compartment 1; and (d) nally, in conguration 4 elastance was the same in both compartments, while the elastic component of viscoelasticity in compartment 2 was 5-fold that in compartment 1. In all congurations solely the elastic properties and elastic component of the viscoelastic behaviour of the respiratory system were modied and the resistive properties were those previously reported (Eissa et al., 1991). To calculate the compartmental volume distribution the successive respiratory cycles reached a steady state in pressure and volume for each change in elastic and viscoelastic properties in all instances. We assumed that Vr in each compartment (Vr 1 and Vr 2) corresponded to the volume at end of the expiration. To start off they amounted to 0.5 L in conguration 1 and 4, and 0.83 L and 0.17 L in compartments 1 and 2, respectively, in congurations 2 and 3. Under these conditions the pressures in both compartments were null. 2.2. Data analysis In VCV mode the same ventilatory setting used by Eissa et al. (1991) were applied: constant ow (square waveform) using an inspiratory time (Ti ) of 0.7 s and a ow of 1 L/s with a respiratory rate of 15 breaths/min, obtaining tidal volume of 0.7 L. During PCV we used the same inspiratory and expiratory times used during VCV, and the inspiratory pressure was modulated to obtain the same tidal volume as in VCV mode. To calculate the compartmental volume distribution we used the same method previously reported (Antonaglia et al., 2005). During VCV and PCV the parameters in

E2 cm H2 O/L 51.5 155 155 51.5

Evisc 1 cm H2 O/L 22.2 22.2 13.5 13.5

Evisc 2 cm H2 O/L 22.2 22.2 67.5 67.5

Rvisc 12 cm H2 O/L/s 54.2 54.2 54.2 54.2

Values according to Eissa et al. (1991). Ri 12 and Re 12, inspiratory and expiratory resistances in compartments one and two, respectively; E1 and E2, static elastance in compartments 1 and 2, respectively; Evisc 1 and Evisc 2, viscoelastic elastances in compartments 1 and 2, respectively; Rvisc 1 and Rvisc 1, viscoelastic resistance in compartments 1 and 2, respectively.

V. Antonaglia et al. / Respiratory Physiology & Neurobiology 171 (2010) 225231

227

Table 1 were used to calculate the compartmental distribution of the inspired volume (V1 and V2), the ratio between V1 and V2, the ratio between end-expiratory lung volumes in the two compartments (Vr 1 and Vr 2), and the difference between end-expiratory volume and FRC in the model ( V). 1 and At the end of inspiration, the compartmental ows (V 2) and the maximal compartmental pressure were determined. V Moreover, we simulated, by zeroing ow, 4-s end-inspiratory and end-expiratory occlusions in upper airways and measured total Pmax (maximal pressure value when occlusion occurs), P1 (pressure value after the rapid drop due to the loss of Newtonian resistive pressure) and Pplateau (Pel = elastic recoil pressure of the lung) (Bates et al., 1988). Moreover, total Pvisc was considered as the difference between P1 and Pel , and total intrinsic positive end-expiratory pressure (PEEPi) as the value of pressure corresponding at that after 4-s end-expiratory occlusion; during this time the pressure values were balanced in the two compartments. Lung static elastance (Est ) was calculated as Pel minus PEEPi divided by Vt . During the occlusions, the compartmental elastic and viscoelastic pressures (Pel 1 and Pel 2, Pvisc 1 and Pvisc 2, respectively) were also determined. 2.3. Model limitations The present analysis is based on the assumption that the lung can be represented by two compartments, each one composed by a compliant and a resistive element, together with a Maxwell body (viscoelastic element), whose individual characteristics are constant at a given ow. The model was used assuming that in all instances compartmental resistances, elastances and viscoelastic parameters remained constant throughout the breath. The effect on the system of the non-linearity due to the variation of the alveolar and conducting tube compliances was considered negligible (Golden et al., 1973; Pride et al., 1967). Additionally, the results of the present investigation did not take into account the non-linear relationship between peripheral airway resistance and volume, expiratory ow-limitation, as well as changes in resistive properties. Finally, our analysis was based on the assumption that the lung behaves as a linear viscoelastic model, which clearly should not be regarded as a complete and perfect representation of the pulmonary mechanical prole. More complex non-linear viscoelastic (Suki and Bates, 1991) and viscoplastoelastic models (Hildebrandt, 1970) have been used to explain volume and time-dependence of energy dissipation within the respiratory system. 3. Results 3.1. Respiratory mechanics Table 2 lists respiratory mechanical parameters obtained by performing an end-inspiratory occlusion on the Ptr curves in the four-modelled congurations and, for comparison purposes, the mean values in patients (Eissa, 1991) at the same ventilatory set = 1 L/s). In all instances the modelled values ting (VCV, Ti = 0.7 s, V were similar to those in patients. Table 2 also lists the respiratory mechanical data gathered in PCV using Ti = 0.7 s, volume = 0.7 L, and inspiratory pressure ranging between 24 and 26 cm H2 O for the 4 congurations. While total Pmax and Pvisc were lower than the corresponding values in VCV, Pel , PEEPi, Est were similar. 3.2. Volume distribution in the four congurations In Table 3 the distribution of the inspiratory tidal volume and end-expiratory lung volume at steady state are expressed as ratios between the values found in the two compartments in VCV and PCV modes. The ideal ratios in congurations 1 and 4 are 1, while in congurations 2 and 3 (lung elastance in the compartment 2

Table 2 Respiratory mechanical parameters measured during ventilation with an inspiratory time of 0.7 s in the four congurations during volume- and pressure-controlled ventilations, and those reported by Eissa et al. (1991). Ventilation mode VCV Pmax (cm H2 O) Pvisc (cm H2 O) Pel (cm H2 O) PEEPi(cm H2 O) Est (cm H2 O/L) Rint (cm H2 O/L/s) PCV Pmax (cm H2 O) Pvisc (cm H2 O) Pel (cm H2 O) PEEPi (cm H2 O) Est (cm H2 O/L) Eissa 30.1 6.9 18.1 0 25.9 4.7 1 28.4 6.1 18.5 0.1 26.3 3.8 25.6 5.4 19.1 0.2 27.0 2 29.9 7.5 18.9 0.1 26.8 3.5 26.3 6.3 19.9 0.2 28.1 3 28.9 6.1 19 0.2 26.8 3.8 25.8 5 19.6 0.2 27.7 4 28.6 7.3 18.5 0.1 26.3 2.8 26.5 6.2 20.1 0.3 28.3

Eissa, 1, 2, 3 and 4, data by Eissa et al. (1991), and congurations 1: same values of elastic and viscoelastic parameters in both compartments; 2: elastance in compartment 2 was 3-fold that in compartment 1; 3: both elastance and elastic component of viscoelasticity in compartment 2 were 3 times higher than those in compartment 1; and 4: elastic component of viscoelasticity in compartment 2 was 3-fold that in compartment 1. VCV and PCV, volume- and pressure-controlled ventilation, respectively. Pmax , Pvisc , Pel , PEEPi, peak pressure, pressure used to overcome viscoelastic and heterogeneous mechanical components, static recoil pressure of the respiratory system and intrinsic positive end-expiratory pressure in the overall system, respectively. Est and Rint , respiratory system static elastance and Newtonian resistance, respectively.

is ve times higher than in compartment 1) they should equal 5. Vr 1/Vr 2 remained equal to the theoretical data, whereas Vt 1/Vt 2 was always smaller than the predicted values (i.e., 5) in congurations 2 and 3 and larger than the predicted value (i.e., 1) in conguration 4 in both ventilatory modes, although the values tend to be closer to the predicted ones in PCV. Additionally, lung volume above FRC ( V), reecting dynamic hyperination, is reported for each conguration. The values tended to be higher in PCV mode. Compartmental ows, Pel and Pvisc during VCV and PCV for each conguration are reported in Table 4. Under conditions of heterogeneity of either elastic or viscoelastic behaviours, as reproduced in congurations 2 and 4, different compartmental ow distributions were derived (Table 4). For the sake of clarity ow-, pressure- and volume-time proles in each conguration during VCV or PCV are depicted in
Table 3 Volume distributions and end-expiratory lung volume measured during ventilation with an inspiratory time of 0.7 s in the four congurations during volume- and pressure-controlled ventilation. Ventilation mode Conguration 1 VCV Vt 1/Vt 2 Vr 1/Vr 2 V (mL) PCV Vt 1/Vt 2 Vr 1/Vr 2 V (mL) 1 1 26 1 1 38 2 2.17 3.08 33 2.33 3.1 46 3 2.69 3.05 24 2.81 3.08 40 4 1.29 0.99 28 1.26 0.99 39

Conguration 1: same values of elastic and viscoelastic parameters in both compartments; 2: elastance in compartment 2 was 3-fold that in compartment 1; 3: both elastance and elastic component of viscoelasticity in compartment 2 were 3 times higher than those in compartment 1; and 4: elastic component of viscoelasticity in compartment 2 was 3-fold that in compartment 1. VCV and PCV, volumeand pressure-controlled ventilation, respectively. Vt 1/Vt 2, Vr 1/Vr 2 are, respectively, the ratios between tidal volumes in compartments 1 and 2, and residual volumes in compartments 1 and 2. V, difference between end-expiratory lung volume and FRC.

228

V. Antonaglia et al. / Respiratory Physiology & Neurobiology 171 (2010) 225231

Table 4 ) and alveolar pressure partitioned into viscoelastic and End-inspiratory ows (V elastic pressures (Pvisc and Pel ) measured at 0.7 s for the two compartments with four congurations. Ventilation mode Conguration 1 VCV 1 (L/s) V 2 (L/s) V Pvisc 1 (cm H2 O) Pvisc 2 (cm H2 O) Pel 1 (cm H2 O) Pel 2 (cm H2 O) PCV 1 (L/s) V 2 (L/s) V Pvisc 1 (cm H2 O) Pvisc 2 (cm H2 O) Pel 1 (cm H2 O) Pel 2 (cm H2 O) 0.5 0.5 6.4 6.4 19.2 19.2 2 0.7 0.3 8.8 4.1 18 24 3 0.734 0.266 6.4 6.3 18.9 20.7 4 0.55 0.45 5 10.1 21.6 16.9

Figs. 2 and 3. Left- and right-hand panels correspond to compartments 1 and 2, respectively. Considering the compartmental pressure curves, it can be seen that in compartment 1 during both VCV and PCV the curves presented a similar behaviour, while in compartment 2 higher maximal values were reached and a higher range (2.5 cm H2 O) was found in VCV (Figs. 2 and 3). The differences between the compartmental pressures were higher during VCV in congurations 2 and 3, while in PCV they were minimal. However, ows about of 2 L/s were reached in PCV model under high elastic loads in compartment 1 (Fig. 3A).

4. Discussion The results of the present investigation were not able to support the hypothesis that the different proles of inspiratory ow generated during VCV and PCV could signicantly modify gas distribution in the lungs modelled as heterogeneous elastic and/or viscoelastic properties. Interestingly enough, during VCV the highest variability in maximal compartmental pressure respect to the modication in elastic and viscoelastic properties was found in the sicker lung, while it remained substantially unchanged during PCV. A two-compartmental model with unequal time constants has been used to describe pulmonary heterogeneities and subsequent parallel gas distribution by sinusoidal pressure oscillation (Otis et al., 1956) and constant inspiratory ow (Bates et al., 1985), pointing out the contribution of viscoelasticity/mechanical heterogeneity to the overall mechanical behaviour of the respiratory

0.038 0.038 5.5 5.5 19.4 19.4

0.063 0.014 7.7 3.3 18.4 23

0.052 0.03 5.9 5 19.8 20.7

0.025 0.076 4.4 8.5 22 17.7

Conguration 1: same values of elastic and viscoelastic parameters in both compartments; 2: elastance in compartment 2 was 3-fold that in compartment 1; 3: both elastance and elastic component of viscoelasticity in compartment 2 were 3 times higher than those in compartment 1; and 4: elastic component of viscoelasticity in compartment 2 was 3-fold that in compartment 1. VCV and PCV, volume- and pressure-controlled ventilation, respectively.

Fig. 2. Flow-, pressure- and volume-time proles in each conguration during volume (VCV). In panels A and B the curves pertaining to compartments 1 and 2 are depicted, respectively. Please note that compartment 2 was modelled such that it is sicker than compartment 1.

V. Antonaglia et al. / Respiratory Physiology & Neurobiology 171 (2010) 225231

229

Fig. 3. Flow-, pressure- and volume-time proles in each conguration during volume (PCV). In panels A and B the curves pertaining to compartments 1 and 2 are depicted, respectively. Please note that compartment 2 was modelled such that it is sicker than compartment 1.

system. This approach has been frequently used (DAngelo et al., 1989, 1991; Jonson et al., 1993; Beydon et al., 1996). Moreover, gas distribution has been approached during mechanical ventilation in heterogeneous lung considering the frequency-dependent features of elastance and resistance (Similowski and Bates, 1991; Baird et al., 1994). Studies regarding pulmonary gas distribution during PCV versus VCV were based on static and dynamic CT scanning (Roth, 2004), and on different experimental lung injury models (Shardonofsky et al., 1991; Maeda et al., 2004). In an our previous investigation (Antonaglia et al., 2005), we employed a two-compartmental model with homogeneous viscoelastic properties modelling homogeneous resistive and elastic properties corresponding to normal subjects, and inhomogeneous resistive or elastic properties reecting severe COPD and ARDS patients, respectively. 4.1. Characteristics of the model In the present investigation we employed the same algorithms (Antonaglia et al., 2005), using mean values of respiratory parameters found in patients with elastic impairment (Eissa, 1991), and modelled four possible pathophysiological conditions. While Eissa et al. (1991) explained their results based on elastic heterogeneities only, our congurations were related to an overall heterogeneous impairment of elastic and elastic component of

viscoelastic properties. Resistance and the resistive viscoelastic constant were considered higher than normal values (DAngelo et al., 1989, 1991) but equal in each compartment. Also the elastic forces were increased due to the mechanical alterations of the lung parenchyma related to the modication of the potential interaction between strain and stress (Mead et al., 1970). An altered mechanical behaviour of the lung can be also due to an alteration of the bres in the lung extracellular matrix, reecting a pathological interaction between collagen and elastic bres (Bachofen and Schurch, 2001; Faffe and Zin, 2009) expressed in our model by the viscoelastic behaviours. In line with the analysis of Bates et al. (1985) that demonstrated the ow-dependence of resistances at the start of inspiration in a two-compartmental model, the value of initial inspiratory ow was the same in each conguration in PCV as well as in VCV (Figs. 2 and 3). 4.2. Overall homogeneous impairment of elastic and viscoelastic properties In conguration 1 it can be observed that the difference between end-expiratory lung volume and FRC ( V) was higher under PCV than VCV because the higher viscoelastic pressures at end inspiration in VCV (Tables 3 and 4) promote a larger lung emptying. This is due to the ow dependency of viscoelastic pressures, determining a higher value of pressure when the ow remains constant until the

230

V. Antonaglia et al. / Respiratory Physiology & Neurobiology 171 (2010) 225231

end of inspiration, respect to the decremental ow prole during PCV. This is in line with the ndings of Roth et al. (2004) obtained in experimental animal lung injury by a computed tomography study, demonstrating that overinated lung volumes were increased with PCV in comparison to VCV. 4.3. Heterogeneous elastic and homogeneous viscoelastic properties In conguration 2, tidal volume was predominantly directed to the sicker compartment during either VCV or PCV, as seen in Table 3 (Vt 1/Vt 2 are smaller than the expected value, i.e., 5). This results from an interaction between elastic and viscoelastic properties during lung ination (Antonaglia et al., 1998) and the different ows in each compartment alter the viscoelastic reaction in the lung (Similowski et al., 1989). During VCV the sicker compartment yields a higher Pel (25% higher than that generated in compartment 1), while the ow at end inspiration is 2.3 times higher in compartment 1. The generated ow-dependent viscoelastic pressure is more than double in compartment 1 (Table 4) and it is able to counterbalance the corresponding ow. The result is a limitation in terms of lung volume increase in compartment 1 and Vt 1/Vt 2 is smaller (Vt 1/Vt 2 = 2.17) than the expected value (Vt 1/Vt 2 = 5). It is presumable that a more remarkable impairment of viscoelastic properties, homogeneously distributed, determines a smaller difference in volume distribution respect to that due to the elastic heterogeneity alone. In PCV these effects are less evident (Vt 1/Vt 2 = 2.33) because the reduction of the ow at end inspiration decreases the values of the viscoelastic pressure (Table 4). 4.4. Heterogeneous elastic and viscoelastic properties When both worst elastic and viscoelastic properties are localized in compartment 2 (conguration 3), V, Pvisc and Pel were closer to the corresponding values found in condition 1 (Tables 3 and 4). It stems from the fact that the relationship between E1 in conditions 1 and 2 equals 1.6 and the same value results when Evisc 1 in condition 2 is divided by Evisc 1 in condition 3. In other words, the fall in Evisc 1 in condition 3 in relation to condition 2 balances the previous relative decay in E1 from condition 1 to condition 2. The contemporary presence of increased elastic and viscoelastic forces in compartment 2 causes a decrease in terms of delivered ow and a smoothing in terms of Pel and Pvisc in comparison to the values found in conguration 2, increasing the ow and volume distribution in compartment 1. Therefore, the impairment of the elastic and viscoelastic properties in the same compartment is the condition in which the volume distribution is more remarkable in the less sick compartment in both ventilatory modalities (Table 3). 4.5. Heterogeneous viscoelastic properties In conguration 4 the heterogeneity is due only to the viscoelastic forces, generated by a potential pathological interaction between collagen and elastic bres, conditioning a higher viscoelastic pressure in compartment 2 (Table 4). In comparison to conguration 1 the higher values of Pvisc are able to reduce in the same compartment during both VCV and PCV the volume (29 and 26%, respectively) and elastic pressure (12 and 9%, respectively), while ow (20%) during VCV. 4.6. At the bedside Facing a patient the physician is not aware ab initio of the mechanical characteristics of the diseased lung. Theoretically it

would be expected that conguration 3 would reect a severe heterogeneous global impairment of a lung. The measurements of Pel and Pvisc at the airway opening during end-inspiratory pause are not diagnostic of the affected component of lung mechanics in a mechanically heterogeneous lung (Table 2). The diverse values of Vt 1/Vt 2 (Table 3) are not associated with either Pel or Pvisc (Table 2). Under the present experimental conditions the changes in Pel of the overall system were minute among the four congurations (both VCV and PCV). Changes in Pvisc were not reected on the diverse Vt 1/Vt 2 either. Interestingly, if only either elastance or the elastic component of viscoelasticity were altered, Pvisc presented higher values than those expressed by increased either elastance or viscoelasticity. Furthermore, it can be seen that despite the similarity in Pvisc values, the distribution of Pvisc can vary importantly within the heterogeneous lung. Thus, the ndings of the present investigation were: (a) mechanical data of the overall lung obtained during end-inspiratory airway occlusion do not allow the identication of heterogeneous elastances and/or elastic component of viscoelasticity (Table 4); (b) our results were not able to demonstrate that gas distribution in the lungs can be modied during VCV and PCV. This is in line with the ndings of Prella et al. (2002) in patients with acute lung injury where the gas distribution was analysed using CT scan densitometric technique and identical lung density, dened by the total mean CT number, in both VCV and PCV was found. According with these ndings no differences in term of gas exchange were found; (c) during VCV maximal compartmental pressure in the sicker lung (compartment 2) presented higher values and variability (Fig. 2) respect to the modication of the elastic and viscoelastic properties, while it remained substantially unchanged during PCV. It must be stressed that in the maximal compartmental pressure value is included the resistive component due to Ri 12. The potential effects reecting in the alveolus can be evaluated adding the value of Pvisc and Pel obtained during end-inspiratory occlusion in each compartment (Table 4). It can be also observed that during VCV these values are higher in the sicker compartment respect to PCV. As the elastic and/or viscoelastic properties impairment is unidentiable at bedside the use of pressure-controlled ventilatory mode could grant a less remarkable pressure variability able to reduce the potential ventilator-associated lung injury. In patients with elastic load impairment, the ventilatory protective strategy, i.e. reduced tidal volume and adequate PEEP levels is demonstrated as able to avoid exposing the lung to high pressure and volume, while the results regarding smaller morbidity and mortality of PCV in comparison to VCV are so far conicting. According to the results of the present investigation, the choice of PCV or VCV can be supported by the aim to maintain a stable compartmental pressure instead of a diverse gas distribution in the overall lung in different elastic conditions. Thus, PCV could be able to satisfy this approach resulting under the present experimental conditions in a less important difference between the compartmental pressures. In conclusion, using different mechanical congurations of elastically heterogeneous lungs the volume distribution within the lung did not differ under VCV and PCV; during VCV the compartmental pressure was more variable in the sicker lung compartment depending on elastic and viscoelastic properties; nally, parameters measured during an end-inspiratory pause could not identify the characteristics of each conguration.

Acknowledgements The authors thank Dr. Giulio Torelli and DPDsoft (Padua) for the collaboration in writing the software used in this work. W.A. Zin was partially supported by CNPq and FAPERJ, Brazil.

V. Antonaglia et al. / Respiratory Physiology & Neurobiology 171 (2010) 225231

231

Appendix A. Differential equations for compartment 1: dP el1 P el2 E 1 P visc1 E 1 P visc2 E 1 P el1 E 1 + + = R1 + R2 R1 + R2 R1 + R2 R1 + R2 dt + R2 E 1 V R1 + R2

dP visc1 P el1 E 1 P el2 E 1 P visc1 E 1 P visc1 E 1 = + R1 + R2 R1 + R2 R1 + R2 Rvisc1 dt + P visc2 E 1 R2 E 1 V + R1 + R2 R1 + R2

is ow, R1 and R2 are Newtonian resistances; E1 is the where V static elastance of compartment 1; and, Pel and Pvisc represent elastic and viscoelastic pressures, respectively. Analogous equations were developed for compartment 2. References
Antonaglia, V., Grop, A., Demanins, P., Beltrame, F., Lucangelo, U., Peratoner, A., De Simoni, L., Gullo, A., Milic-Emili, J., 1998. Single-breath method for assessing the viscoelastic properties of the respiratory system. Eur. Respir. J. 12, 11911196. Antonaglia, V., Torelli, L., Zin, W.A., Gullo, A., 2005. Effects of viscoelasticity on volume distribution in a two-compartmental model of normal and sick lungs. Physiol. Meas. 26, 1328. Bachofen, H., Schurch, S., 2001. Alveolar surface forces and lung architecture. Comp. Biochem. Physiol. A: Mol. Integr. Physiol. 129, 183193. Baird, J.S., Johnson, J.L., Escudero, J., Powars, D.R., 1994. Combined pressure control/high frequency ventilation in adult distress respiratory syndrome and sickle cell anemia. Chest 106, 19131916. Bates, J.H.T., Rossi, A., Milic-Emili, J., 1985. Analysis of the behavior of the respiratory system with constant inspiratory ow. J. Appl. Physiol. 58, 18401848. Bates, J.H.T., Baconnier, P., Milic-Emili, J., 1988. A theoretical analysis of interrupter technique for measuring respiratory mechanics. J. Appl. Physiol. 64, 2204 2214. Behrakis, P.K., Higgs, B.O., Baydur, A., Zin, W.A., Milic-Emili, J., 1983. Respiratory mechanics during halothane anesthesia and anesthesiaparalysis in humans. J. Appl. Physiol. 55, 10851092. Beydon, L., Svantesson, C., Brauer, K., Lemaire, F., Jonson, B., 1996. Respiratory mechanics in patients ventilated for critical lung disease. Eur. Respir. J. 9, 262273. Chatburn, R.L., El Khatib, M.F., Smith, P.G., 1994. Respiratory system behavior during mechanical ination with constant inspiratory pressure and ow. Respir. Care 39, 979988. DAngelo, E., Calderini, E., Torri, G., Robatto, F.M., Bono, D., Milic-Emili, J., 1989. Respiratory mechanics in anesthetized paralyzed humans: effects of ow, volume, and time. J. Appl. Physiol. 67, 25562564. DAngelo, E., Robatto, F.M., Calderini, E., Tavola, M., Bono, D., Torri, G., Milic-Emili, J., 1991. Pulmonary and chest wall mechanics in anesthetized paralyzed humans. J. Appl. Physiol. 70, 26022610.

Eissa, N.T., Ranieri, V.M., Corbeil, C., Chass, M., Robatto, F.M., Braidy, J., Milic-Emili, J., 1991. Analysis of behavior of the respiratory system in ARDS patients: effects of ow, volume and time. J. Appl. Physiol. 70, 27192729. Faffe, D.S., Zin, W.A., 2009. Lung parenchymal mechanics in health and disease. Physiol. Rev. 89, 759775. Girard, T.D., Bernard, G.R., 2007. Mechanical ventilation in ARDS: a state-of-the-art review. Chest 131, 921929. Golden, J.F., Clark, J.W., Stevens, P.M., 1973. Mathematical modelling of pulmonary airway dynamics. IEEE Trans. Biomed. Eng. 20, 397404. Hildebrandt, J., 1970. Pressurevolume data of cat lung interpreted by a plastoelastic linear viscoelastic model. J. Appl. Physiol. 28, 365372. Jonson, B., Beydon, L., Brauer, K., Mansson, C., Valind, S., Grytzell, H., 1993. Mechanics of respiratory system in healthy anesthetized humans with emphasis on viscoelastic properties. J. Appl. Physiol. 75, 132140. Maeda, Y., Fujino, Y., Uchiyama, A., Matsuura, N., Mashimo, T., Nishimura, M., 2004. Effects of peak inspiratory ow on development of ventilator-induced lung injury in rabbits. Anesthesiology 101, 722728. Mead, J., Takishima, T., Leith, D., 1970. Stress distribution in lungs: a model of pulmonary elasticity. J. Appl. Physiol. 28, 596608. Mount, L.E., 1955. The ventilation ow-resistance and compliance of rat lungs. J. Physiol. Lond. 127, 157167. Otis, A.B., McKerrow, C.B., Bartlett, R.A., Mead, J., McIlroy, M.B., Selverstone, N.J., Radford Jr., E.P., 1956. Mechanical factors in distribution of pulmonary ventilation. J. Appl. Physiol. 8, 427443. Prella, M., Feihl, F., Domenighetti, G., 2002. Effects of short-term pressure-controlled ventilation on gas exchange, airway pressures, and gas distribution in patients with acute lung injury/ARDS. Chest 122, 13821388. Pride, N.B., Permut, S., Biley, R.L., Bromberger-Barnea, B., 1967. Determination of maximum expiratory ow from the lung. J. Appl. Physiol. 22, 10861089. Rohrer, F., 1915. Der Stroemungswiderstand in den menschlichen Atemwegen und der Einuss der Bronchialsystems auf den Atmungsverlauf in verschiedenen Lungenbezirken. Arch. Ges. Physiol. Mensch. Tiere. 162, 225299. Roth, H., Luecke, T., Deventer, B., Joachim, A., Herrmann, P., Quintel, M., 2004. Pulmonary gas distribution during ventilation with different inspiratory ow patterns in experimental lung injurya computed tomography study. Acta Anaesthesiol. Scand. 48, 851861. Sharp, J.T., Johnson, F.N., Goldberg, N.B., Van Lith, P., 1967. Hysteresis and stress adaptation in the human respiratory system. J. Appl. Physiol. 23, 487497. Similowski, T., Levy, P., Corbeil, C., Albala, M., Pariente, R., Derenne, J.-P., Bates, J.H.T., Jonson, B., Milic-Emili, J., 1989. Viscoelastic behavior of lung and chest wall in dogs determined by ow interruption. J. Appl. Physiol. 67, 22192229. Similowski, T., Bates, J.H.T., 1991. Two compartmental modelling of respiratory mechanics at low frequencies: gas redistribution or tissue rheology. Eur. Respir. Physiol. 4, 353358. Shardonofsky, F.R., Skaburskis, M., Sato, J., Zin, W.A., Milic-Emili, J., 1991. Effects of volume history and vagotomy on pulmonary and chest wall mechanics in cats. J. Appl. Physiol. 71, 498508. Suki, B., Bates, J.H.T., 1991. A non linear viscoelastic model of lung tissue mechanics. J. Appl. Physiol. 71, 826833. The Acute Respiratory Distress Syndrome Network, 2000. Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. N. Engl. J. Med. 342, 13011308. Young, M.P., Manning, H.L., Wilson, D.L., Mette, S.A., Riker, R.R., Leiter, J.C., Liu, S.K., Bates, J.T., Parsons, P.E., 2004. Ventilation of patients with acute lung injury and acute respiratory distress syndrome: has new evidence changed clinical practice? Crit. Care Med. 32, 12601265.