tidal recruitment assessed by electrical impedance ... muders - tidal... · tidal recruitment...

9
Tidal recruitment assessed by electrical impedance tomography and computed tomography in a porcine model of lung injury* Thomas Muders, MD; Henning Luepschen, MSc, PhD; Jo ¨ rg Zinserling, MSc, PhD; Susanne Greschus, MD; Rolf Fimmers, MSc; Ulf Guenther, MD; Miriam Buchwald, Daniel Grigutsch, Steffen Leonhardt, MD, PhD; Christian Putensen, MD, PhD; Hermann Wrigge, MD, PhD I n patients with acute lung injury (ALI) or acute respiratory distress syndrome (ARDS), mechanical ventilation should avoid inspira- tory overdistension and tidal recruitment of lung units (1– 4) caused either by cy- clic opening and closing or by movement of fluid and foam in distal airways (5–7). In contrast to lowering tidal volumes (8, 9), the general use of high levels of pos- itive end-expiratory pressures (PEEP) has, so far, not been demonstrated to improve survival in mixed populations of ALI/ARDS patients (10 –13). This may be partially explained by individual differ- ences in the potential for alveolar recruit- ment in response to higher airway pres- sures (14, 15). Recent computed tomography (CT) studies suggest that pa- tients with increased potential for re- cruitment have a higher mortality (14) and that an increased amount of tidal recruitment is an independent risk factor for death (16). In these patients, higher PEEP levels decreased cyclic tidal recruit- ment, which prevailed over the effect of increasing alveolar strain and thereby im- *See also p. 1015. From the Department of Anesthesiology and Intensive Care Medicine (TM, HL, JZ, UG, MB, DG, CP, HW), Uni- versity of Bonn, Bonn, Germany; Philips Chair for Medical Information Technology (HL, SL), Helmholtz-Institute for Biomedical Engineering, RWTH Aachen University, Aachen, Germany; Department of Radiology (SG), Univer- sity of Bonn, Bonn, Germany; Institute for Medical Biom- etry, Informatics, and Epidemiology (RF), University of Bonn, Germany; Department of Anesthesiology and Inten- sive Care Medicine (HW), University of Leipzig, Germany. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s Web site (http://journals.lww.com/ccmjournal). Part of these data has been presented in form of abstracts at the ATS International Conference 2009 (San Diego) and at the EIT conference 2009 (Man- chester), and in the form of an extended abstract at the IFMBE World Congress on Medical Physics and Bio- medical Engineering 2009 (Munich). Supported, in part, by a grant of the German Re- search Council “Deutsche Forschungsgemeinschaft,” DFG (WR47-1-1). Draeger Medical provided an EIT device and the ventilator was supplied by GE Healthcare without any restrictions. The University Hospitals of Bonn and Leipzig were supported by research funding from Draeger Medical not related to this study. Prof. Leonhardt, Prof. Putensen, and Prof. Wrigge received honoraria/speaking fees and grants from Drager Medical AG. These grants were not related to the currently submitted study. Further, our study was supported by the DFG granted to Prof. Wrigge. The remaining authors have not disclosed any potential conflicts of interest. For information regarding this article, E-mail: [email protected] Copyright © 2012 by the Society of Critical Care Medicine and Lippincott Williams & Wilkins DOI: 10.1097/CCM.0b013e318236f452 Objectives: To determine the validity of electrical impedance tomography to detect and quantify the amount of tidal recruitment caused by different positive end-expiratory pressure levels in a porcine acute lung injury model. Design: Randomized, controlled, prospective experimental study. Setting: Academic research laboratory. Subjects: Twelve anesthetized and mechanically ventilated pigs. Interventions: Acute lung injury was induced by central venous oleic acid injection and abdominal hypertension in seven animals. Five healthy pigs served as control group. Animals were ventilated with positive end-expiratory pressure of 0, 5, 10, 15, 20, and 25 cm H 2 O, respectively, in a randomized order. Measurements and Main Results: At any positive end-expira- tory pressure level, electrical impedance tomography was ob- tained during a slow inflation of 12 mL/kg of body weight. Regional-ventilation-delay indices quantifying the time until a lung region reaches a certain amount of impedance change were calculated for lung quadrants and for every single electrical impedance tomography pixel, respectively. Pixel-wise calculated regional-ventilation-delay indices were plotted in a color-coded regional-ventilation-delay map. Regional-ventilation-delay inho- mogeneity that quantifies heterogeneity of ventilation time courses was evaluated by calculating the scatter of all pixel-wise calculated regional-ventilation-delay indices. End-expiratory and end-inspiratory computed tomography scans were performed at each positive end-expiratory pressure level to quantify tidal re- cruitment of the lung. Tidal recruitment showed a moderate inter-individual (r .54; p < .05) and intra-individual linear correlation (r .46 up to r .73 and p < .05, respectively) with regional-ventilation-delay obtained from lung quadrants. Region- al-ventilation-delay inhomogeneity was excellently correlated with tidal recruitment intra- (r .90 up to r .99 and p < .05, respectively) and inter-individually (r .90; p < .001). Conclusions: Regional-ventilation-delay can be noninvasively measured by electrical impedance tomography during a slow inflation of 12 mL/kg of body weight and visualized using venti- lation delay maps. Our experimental data suggest that the imped- ance tomography-based analysis of regional-ventilation-delay in- homogeneity provides a good estimate of the amount of tidal recruitment and may be useful to individualize ventilatory set- tings. (Crit Care Med 2012; 40:903–911) KEY WORDS: acute lung injury; electrical impedance tomography; mechanical ventilation; positive end-expiratory pressure; regional ventilation delay inhomogeneity; tidal recruitment 903 Crit Care Med 2012 Vol. 40, No. 3

Upload: nguyenkhanh

Post on 20-Aug-2018

227 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: Tidal recruitment assessed by electrical impedance ... Muders - Tidal... · Tidal recruitment assessed by electrical impedance tomography and computed tomography in a porcine model

Tidal recruitment assessed by electrical impedance tomographyand computed tomography in a porcine model of lung injury*

Thomas Muders, MD; Henning Luepschen, MSc, PhD; Jorg Zinserling, MSc, PhD; Susanne Greschus, MD;Rolf Fimmers, MSc; Ulf Guenther, MD; Miriam Buchwald, Daniel Grigutsch, Steffen Leonhardt, MD, PhD;Christian Putensen, MD, PhD; Hermann Wrigge, MD, PhD

I n patients with acute lung injury(ALI) or acute respiratory distresssyndrome (ARDS), mechanicalventilation should avoid inspira-

tory overdistension and tidal recruitmentof lung units (1–4) caused either by cy-clic opening and closing or by movementof fluid and foam in distal airways (5–7).In contrast to lowering tidal volumes (8,

9), the general use of high levels of pos-itive end-expiratory pressures (PEEP)has, so far, not been demonstrated toimprove survival in mixed populations ofALI/ARDS patients (10–13). This may bepartially explained by individual differ-ences in the potential for alveolar recruit-ment in response to higher airway pres-sures (14, 15). Recent computed

tomography (CT) studies suggest that pa-tients with increased potential for re-cruitment have a higher mortality (14)and that an increased amount of tidalrecruitment is an independent risk factorfor death (16). In these patients, higherPEEP levels decreased cyclic tidal recruit-ment, which prevailed over the effect ofincreasing alveolar strain and thereby im-

*See also p. 1015.From the Department of Anesthesiology and Intensive

Care Medicine (TM, HL, JZ, UG, MB, DG, CP, HW), Uni-versity of Bonn, Bonn, Germany; Philips Chair for MedicalInformation Technology (HL, SL), Helmholtz-Institute forBiomedical Engineering, RWTH Aachen University,Aachen, Germany; Department of Radiology (SG), Univer-sity of Bonn, Bonn, Germany; Institute for Medical Biom-etry, Informatics, and Epidemiology (RF), University ofBonn, Germany; Department of Anesthesiology and Inten-sive Care Medicine (HW), University of Leipzig, Germany.

Supplemental digital content is available for this article.Direct URL citations appear in the printed text and areprovided in the HTML and PDF versions of this article on

the journal’s Web site (http://journals.lww.com/ccmjournal).Part of these data has been presented in form of

abstracts at the ATS International Conference 2009(San Diego) and at the EIT conference 2009 (Man-chester), and in the form of an extended abstract at theIFMBE World Congress on Medical Physics and Bio-medical Engineering 2009 (Munich).

Supported, in part, by a grant of the German Re-search Council “Deutsche Forschungsgemeinschaft,”DFG (WR47-1-1). Draeger Medical provided an EIT deviceand the ventilator was supplied by GE Healthcare withoutany restrictions. The University Hospitals of Bonn andLeipzig were supported by research funding from DraegerMedical not related to this study.

Prof. Leonhardt, Prof. Putensen, and Prof. Wriggereceived honoraria/speaking fees and grants fromDrager Medical AG. These grants were not related tothe currently submitted study. Further, our study wassupported by the DFG granted to Prof. Wrigge. Theremaining authors have not disclosed any potentialconflicts of interest.

For information regarding this article, E-mail:[email protected]

Copyright © 2012 by the Society of Critical CareMedicine and Lippincott Williams & Wilkins

DOI: 10.1097/CCM.0b013e318236f452

Objectives: To determine the validity of electrical impedancetomography to detect and quantify the amount of tidal recruitmentcaused by different positive end-expiratory pressure levels in aporcine acute lung injury model.

Design: Randomized, controlled, prospective experimental study.Setting: Academic research laboratory.Subjects: Twelve anesthetized and mechanically ventilated pigs.Interventions: Acute lung injury was induced by central venous

oleic acid injection and abdominal hypertension in seven animals.Five healthy pigs served as control group. Animals were ventilatedwith positive end-expiratory pressure of 0, 5, 10, 15, 20, and 25cm H2O, respectively, in a randomized order.

Measurements and Main Results: At any positive end-expira-tory pressure level, electrical impedance tomography was ob-tained during a slow inflation of 12 mL/kg of body weight.Regional-ventilation-delay indices quantifying the time until alung region reaches a certain amount of impedance change werecalculated for lung quadrants and for every single electricalimpedance tomography pixel, respectively. Pixel-wise calculatedregional-ventilation-delay indices were plotted in a color-codedregional-ventilation-delay map. Regional-ventilation-delay inho-mogeneity that quantifies heterogeneity of ventilation time

courses was evaluated by calculating the scatter of all pixel-wisecalculated regional-ventilation-delay indices. End-expiratory andend-inspiratory computed tomography scans were performed ateach positive end-expiratory pressure level to quantify tidal re-cruitment of the lung. Tidal recruitment showed a moderateinter-individual (r � .54; p < .05) and intra-individual linearcorrelation (r � .46 up to r � .73 and p < .05, respectively) withregional-ventilation-delay obtained from lung quadrants. Region-al-ventilation-delay inhomogeneity was excellently correlatedwith tidal recruitment intra- (r � .90 up to r � .99 and p < .05,respectively) and inter-individually (r � .90; p < .001).

Conclusions: Regional-ventilation-delay can be noninvasivelymeasured by electrical impedance tomography during a slowinflation of 12 mL/kg of body weight and visualized using venti-lation delay maps. Our experimental data suggest that the imped-ance tomography-based analysis of regional-ventilation-delay in-homogeneity provides a good estimate of the amount of tidalrecruitment and may be useful to individualize ventilatory set-tings. (Crit Care Med 2012; 40:903–911)

KEY WORDS: acute lung injury; electrical impedance tomography;mechanical ventilation; positive end-expiratory pressure; regionalventilation delay inhomogeneity; tidal recruitment

903Crit Care Med 2012 Vol. 40, No. 3

Page 2: Tidal recruitment assessed by electrical impedance ... Muders - Tidal... · Tidal recruitment assessed by electrical impedance tomography and computed tomography in a porcine model

proved outcome (16). Thus, individualsetting of PEEP appears reasonable (17,18) and should aim to reduce tidal re-cruitment (15).

Whereas CT provides insights into re-gional aeration heterogeneities (7, 19, 20)but cannot be used at bedside, global in-dices of lung function such as blood gasesand respiratory mechanics that are usedto titrate ventilator settings do not allowconclusion of regional ventilation distri-bution and, hence, of regional tidal re-cruitment (21).

Electrical impedance tomography(EIT) of the lungs measures relative im-pedance changes in lung tissue duringtidal breathing and creates images of thelocal ventilation distribution at bedside(22–33). An EIT-based index (regional-ventilation-delay [RVD]) describing thescatter of impedance time courses hasrecently been shown to be able to detectrecruitment during a vital capacity infla-tion maneuver (30). In this experiment,we studied the hypotheses that RVD canbe measured by EIT during a slow infla-tion maneuver of only 12 mL/kg of bodyweight (BW) and can be used to estimatethe amount of tidal recruitment and itsPEEP-associated changes.

MATERIALS AND METHODS

Animals, Lung Injury and Ventilatory Set-tings. After approval of the local animal ethicscommittee, 12 healthy pigs (29–35 kg) wereanesthetized, tracheotomized, and instru-mented in supine position as previously de-scribed (34–36). Pigs were mechanically ven-tilated (Engstrom Carestation; GE Healthcare,Helsinki, Finland) in a volume-controlledmode with a tidal volume (VT) set to 6–8mL/kg BW, a respiratory rate of 15 min�1, aninspiratory:expiratory ratio of 1:1, FIO2 of 0.5,and PEEP of 5 cm H2O. After preparation, ALIwas induced in seven animals by intravenousinjection of oleic acid (0.1 mL/kg) in combi-nation with intra-abdominal hypertension of20 cm H2O caused by intraperitoneal infusionof saline (34). Five animals served as controls.Then, respiratory rate was increased up to 30min�1 to avoid hypercapnia (PaCO2 �50 torr),and FIO2 was increased to maintain a PaO2 of�80 torr. Intrinsic PEEP was excluded by ob-servation of end-expiratory zero flow patternsat the ventilator.

Measurement and Interventions. After 3hrs, animals were transferred to the CT scan-ner without interrupting ventilation. Then,PEEP was changed to 0, 5, 10, 15, 20, and 25cm H2O in a randomized order, whereas allother ventilatory settings remained un-changed. Lungs were derecruited (by discon-nection of the respirator) and then recruitedby a continuous positive airway pressure of 50

cm H2O for 40 secs before PEEP settings werechanged, respectively.

After 30 mins of ventilation on the selectedPEEP level blood gases, respiratory mechanicsand hemodynamics were measured (34–36).EIT measurements were performed during asingle slow inflation maneuver with a constantgas flow by setting respiratory rate to 4 min�1,which resulted in an inflation time of 7.5 secsand a VT of 12 mL/kg BW. Then, spiral CTscans were performed at end-expiration andend-inspiration (with a clamped tube) to mea-sure amount of gas, lung tissue aeration, andtidal recruitment.

EIT and CT Imaging. An EIT system (EITevaluation KIT II; Drager Medical GmbH, Lu-beck, Germany) was used. It reconstructs im-ages of ventilation distribution by comparingimpedance changes to a reference state apply-ing a modified Newton-Raphson algorithm(37). Regional and global time courses of im-pedance changes were recorded with a tempo-ral resolution of 20 Hz during a single slowinflation maneuver (Fig. 1). The global imped-ance time curve, �Z(t), was calculated as thesum over the impedance changes of all pixels.

RVD time (�tRVD) (30) was determined be-tween the start of inspiration defined as firstincrease of the global �Z(t) curve and the timewhen the respective regional curve �Zi(t)reaches a threshold of 40% of the maximallocal impedance change (Fig. 2A and 2B). Toaddress the fact that �tRVD depends on infla-tion time, it was normalized by dividing byinflation time (�tmax�min):

RVD � �tRVD/�tmax�min

Thus, RVD index describes the delay givenin percent of inflation time until the respec-tive regional impedance change exceeds a cer-tain threshold (30). RVDs were obtained forlung quadrants (Fig. 2A) (30) and for any sin-gle EIT pixel.

A color-coded map was plotted to visualizethe pixel RVD indices (Fig. 2B). To quantifyRVD inhomogeneity (SDRVD), the SD over allsingle-pixel RVDs (Fig. 2B) was calculated af-ter filtering and masking. A more detaileddescription of EIT data acquisition and pro-cessing is given in the Supplemental data(SDC-1, Supplemental Digital Content 1,http://links.lww.com/CCM/A361).

CT Scanning and Analysis. Spiral CT scans(120 kV, 120 mA) of the total lungs wereperformed during end-expiration and end-inspiration holds using a Brilliance 64 CTscanner (Philips Healthcare, Hamburg, Ger-many). Images were reconstructed in 8-mmslices using a standard filter.

We performed densitometric analysis of allpulmonary CT slices using a computer pro-gram (Osiris; University of Geneva, Geneva,Switzerland) as described previously (34–36).From four CT slices covering the EIT belt andfrom all CT slices, the amounts of hyperin-flated, normally, poorly, and nonaerated lungtissue and tidal recruitment (19, 38, 39) werecalculated for lung quadrants (30) and for thetotal lung, respectively (SDC-2 for details;Supplemental Digital Content 1, http://links.lww.com/CCM/A361).

Statistical Analysis. After confirming nor-mal distribution (Shapiro-Wilk W test), data

Figure 1. Scan protocol. At every positive end-expiratory pressure level, electrical impedance tomog-raphy (EIT) measurements were performed during a single slow inflation breath with a constant gasflow and a tidal volume of 12 mL/kg body weight (BW). Thereafter, spiral computed tomography (CT)scans were performed at end-expiration and end-inspiration (during ventilation with a tidal volume of6–8 mL/kg BW) to measure the amount of gas, nonaerated lung tissue, and regional tidal recruitment.SDRVD, regional ventilation delay inhomogeneity.

904 Crit Care Med 2012 Vol. 40, No. 3

Page 3: Tidal recruitment assessed by electrical impedance ... Muders - Tidal... · Tidal recruitment assessed by electrical impedance tomography and computed tomography in a porcine model

were compared between ALI and controlgroup (factor ALI) and between the differentPEEP levels (factor PEEP) using a two-wayrepeated measures analysis of variance (Statis-tica 6.0; StatSoft, Tulsa, OK). Differences wereseparated by post hoc tests (Newman-Keulstest) comparing three groups of low PEEP (0and 5 cm H2O), moderate PEEP (10 and 15 cmH2O), and high PEEP (20 and 25 cm H2O)within and between both groups, respectively.

Tidal recruitment (calculated from fourEIT-covering CT slices as well as from totallung CT) was linearly correlated with quadrantRVD and SDRVD, respectively. Linear correla-tions were calculated between these parame-ters for single animals (intra-individually) andbetween these parameters for all animals (in-ter-individually) by taking data points from all

PEEP levels. Data are given as mean � SD; p �.05 was considered significant.

RESULTS

Gas Exchange, Respiratory, andHemodynamic Parameters

Detailed tables of cardiorespiratoryand gas exchange variables are given inthe Supplemental data (SDC-3, TablesS1–S3; Supplemental Digital Content 1,http://links.lww.com/CCM/A361). Briefly,PaO2/FIO2 was significantly higher in thecontrol group when compared to injuredpigs (p � .05) and increased significantly

with higher PEEP (p � .001) (SDC-3, Fig.S3C; Supplemental Digital Content 1,http://links.lww.com/CCM/A361). Respi-ratory system compliance (SDC-3, Fig.S3A, Supplemental Digital Content 1,http://links.lww.com/CCM/A361) was al-ways lower (p � .05) in ALI pigs andincreased when PEEP was increased (p �.001) but decreased with highest PEEPlevels in control animals (p � .001). In-creasing PEEP decreased cardiac outputand mean arterial blood pressure in thecontrol group but not in the ALI group(p � .001; Table S1; Supplemental DigitalContent 1, http://links.lww.com/CCM/A361).

Lung Aeration, Gas Volume, andTidal Recruitment

In ALI animals, 52% � 29% of thelung was nonaerated at end-expiration atzero end-expiratory pressure (Table 1).Tidal recruitment amounted 13% �7%of the total lung (Table 1). When PEEPwas increased, end-expiratory lung vol-ume significantly increased up to four-fold (p � .001), whereas the amount ofnonaerated lung tissue and tidal recruit-ment decreased (p � .001, respectively;Table 1).

In control animals, end-expiratoryamount of nonaerated lung tissue andtidal recruitment were much lower whencompared to the ALI group (p � .05;Table 1) and not detectable at PEEP �5cm H2O (p � .001; Table 1). The increasein end-expiratory lung volume paralleledPEEP (p � .001). Compared to the ALIgroup, end-expiratory lung volume wasapproximately doubled at every PEEPstep (p � .001; Table 1). Hyperinflatedlung tissue amounted �1% of the lungs.

RVD Index and RVDInhomogeneity

RVD indices (Fig. 3) were higher independent quadrants when compared tonondependent quadrants (p � .001), andthey decreased with increasing PEEP(p � .001). At higher PEEP levels, differ-ences between dependent and nondepen-dent quadrants were reduced. Accord-ingly, RVD maps (Fig. 4A, fourth row)showed early aeration (i.e., low RVD val-ues) in the nondependent and delayedaeration (i.e., high RVD values) in thedependent lung regions at low PEEP.When PEEP increased, the number ofpixels with early and delayed aeration de-creased. At the highest PEEP levels, all

Figure 2. A, Calculation of regional ventilation delay (RVD) index, schematic description. Left,Functional image (ventilation map) recorded by electrical impedance tomography during a slowinflation maneuver. Right, The global impedance time curve (upper curve) that summarizes allregional impedance time curves increases with aeration during slow inflation. �tmin-max is the durationof the slow inflation. The lower curve is the regional impedance time curve of a single lung quadrantand shows an offset and a more curved pattern. The regional curve reaches its 40% maximum in agiven time (�tRVD). �tRVD is normalized by division by �tmin-max. Thus, the resulting RVD indexquantifies the delay (in % of inflation time) until the respective regional impedance time curve reaches40% of its regional maximal change. The onset of the inflation-dependent increase in impedance doesnot occur at identical time points in individual regional tracings. Thus, RVD is influenced by theregional offset of ventilation and by concavity of the regional ventilation time course. B, Calculationof RVD inhomogeneity (SDRVD), schematic description. Left, Functional image (ventilation map).Middle, Calculation of RVD was performed for regional impedance time curves of any single pixel ofthe electrical impedance tomography matrix. Here, three exemplary pixels are displayed. A color-codedRVD map (right) was plotted to visualize the RVD values of the pixels. The SD over all single pixel RVDvalues (SDRVD) was calculated to quantify RVD inhomogeneity, a measure of heterogeneity in regionalventilation time courses.

905Crit Care Med 2012 Vol. 40, No. 3

Page 4: Tidal recruitment assessed by electrical impedance ... Muders - Tidal... · Tidal recruitment assessed by electrical impedance tomography and computed tomography in a porcine model

lung regions showed a much more ho-mogeneous distribution of RVD indicesand, consequently, a decrease in RVDinhomogeneity (Fig. 4). In all ALI ani-mals, RVD inhomogeneity significantlydecreased with increasing PEEP (Fig.5A). In control animals showing no orminimal tidal recruitment at PEEP lev-els �5 cm H2O, RVD inhomogeneitywas always low (Fig. 5A and 5B). Be-cause tidal recruitment did not changerelevantly (p � .001; Table 1) in thecontrol animals, RVD inhomogeneitydid not change either. At high PEEPlevels, SDRVD was not different betweengroups (post hoc; Fig. 5A).

Correlation of EIT Parametersand Tidal Recruitment

Inter-individual and intra-individualRVD indices and tidal recruitment corre-lated only moderately when obtained inlung quadrants (Tables 2 and 3). In con-trast, for single animals, RVD inhomoge-neity (measured by EIT) and the amountof tidal recruitment (calculated by CT)with changes in PEEP showed similarcourses (Fig. 4). ALI pigs showed highRVD inhomogeneities when the amountof tidal recruitment was high at lowerPEEP levels, but reached small RVD in-homogeneities at highest PEEP levels,when tidal recruitment was minimized.Increasing PEEP from 20 to 25 cm H2Odid not further reduce tidal recruitmentand RVD inhomogeneity reached a pla-teau. Both parameters showed good toexcellent linear within-subject correla-tion (Fig. 4, Table 2) and for all animals(Fig. 5B, Table 2). In control animalsshowing no or minimal tidal recruit-ment, RVD inhomogeneity was alwayslow (Fig. 5A and 5B) and, thus, datashowed a moderate but still significantcorrelation (Fig. 4, Table 2). At the high-est PEEP levels, when tidal recruitmentwas comparable for control and ALI pigs(Table 1), both groups could not be fur-ther distinguished by means of RVD in-homogeneity (Fig. 5A). Calculating tidalrecruitment from only four CT slices cov-ering the EIT belt did not relevantlychange these results (Table 3).

DISCUSSION

The main findings of our animal studyare that RVD indices can be measured by

EIT during a 12-mL/kg slow inflation ma-neuver and can be regionally visualizedby RVD maps. The SD of pixel-wise–defined RVD values quantifies RVD inho-mogeneity (SDRVD) that is highly linearlycorrelated to tidal recruitment and itsPEEP-associated changes.

Detection of Tidal Recruitment

It is worth noting that our study wasnot designed to compare different strate-gies of PEEP titration, but rather to de-termine the value of EIT in measuringchanges in tidal recruitment at differentPEEP levels. Our data show that theamount of tidal recruitment can be reli-ably estimated by RVD inhomogeneity, assuggested by a good intra-individual andinter-individual linear correlation of bothparameters (Table 2, Figs. 4 and 5B).When compared to global parameters oflung mechanics and gas exchange, RVDinhomogeneity was the most robust pa-rameter quantifying tidal recruitment inour model (Table 2, Fig. 5, SDC-3 FigureS3 [Supplemental Digital Content 1,http://links.lww.com/CCM/A361]). Al-though measurement of lung and chestwall compliance using an esophagealpressure catheter (17) and determinationof dynamic compliance (40) may be use-ful to individually titrate PEEP, our find-ings confirm the theoretical model oflung mechanics in ARDS proposed byHickling (21), demonstrating that thepresence of different opening pressures ofventilator units might remain undetectedwhen focusing on the global pressure/volume curve only. In this respect, themethod presented here seems to be fea-

Figure 3. Regional ventilation delay (RVD) indexmeasured in lung quadrants by electrical imped-ance tomography during 12 mL/kg of bodyweight slow inflation at positive end-expiratorypressure (PEEP) levels of 0 to 25 cm H2O. Filledtriangles indicate right dorsal quadrant; outlinedtriangles, left dorsal quadrant; filled circles, rightventral quadrant; outlined circles, left ventralquadrant. Statistics: repeated-measures analysisof variance with factors “quadrant” (for differ-ences between lung quadrants) and “PEEP” (fordifferences between PEEP levels). Data are givenas mean and SE.

Table 1. Lung aeration, recruitment, and ventilation delay inhomogeneity

Parameter Group

Positive End-Expiratory Pressure in �cm H2O�

Analysis of Variance

PositiveEnd-Expiratory

Pressure Group Interaction0 5 10 15 20 25

End-expiratory lung volume Control 131 � 37 174 � 35 267 � 35 357 � 51 402 � 82 477 � 82 b b b

(mL) ALI 72 � 25 93 � 32 121 � 35 165 � 43 218 � 57 277 � 88Poorly aerated lung tissue Control 40 � 11 37 � 10 24 � 7 13 � 7 8 � 7 6 � 10 a a a

(% lung) ALI 37 � 11 36 � 10 29 � 6 40 � 9 39 � 15 30 � 20Nonaerated lung tissue Control 2 � 1 2 � 3 1 � 1 1 � 1 1 � 1 1 � 1 b a b

(% lung) ALI 52 � 29 45 � 42 35 � 34 12 � 10 8 � 6 8 � 8Tidal recruitment Control 3 � 3 2 � 1 0 � 0 0 0 0 b a b

(% lung) ALI 13 � 7 12 � 8 9 � 3 5 � 4 3 � 2 1 � 2Hyperinflated lung tissue Control �1 �1 �1 �1 �1 �1(% lung) ALI �1 �1 �1 �1 �1 �1SDRVD Control 4 � 1 5 � 1 4 � 1 4 � 1 3 � 1 3 � 1 b a b

(%) ALI 14 � 4 13 � 5 11 � 4 8 � 2 6 � 2 5 � 3

ALI, acute lung injury.

906 Crit Care Med 2012 Vol. 40, No. 3

Page 5: Tidal recruitment assessed by electrical impedance ... Muders - Tidal... · Tidal recruitment assessed by electrical impedance tomography and computed tomography in a porcine model

sible and may complement global mea-surement of respiratory mechanics andgas exchange to individually improveventilatory settings.

Methodologic Aspects

We have recently shown that the RVDindex is able to detect lung recruitmentduring a vital capacity inflation maneuver(30). The present study aimed to refinethis method by reducing the slow infla-

Figure 4. Example of images (upper four rows) for a lung-injured pig (intravenous oleic acid injection andintra-abdominal hypertension). Spiral computed tomography (CT) scans at end-inspiration (first row) andend-expiration (second row) during volume-controlled ventilation (tidal volume � 6–8 mL/kg of bodyweight) at positive end-expiratory pressure (PEEP) levels from 0 to 25 cm H2O. Ventilation maps (third row,relative distribution of tidal impedance changes �Z) and regional ventilation delay (RVD) index maps(fourth row) during low-flow breath of 12 mL/kg of body weight with a positive end-expiratory pressure(PEEP) of 0 to 25 cm H2O. Lower four rows, Example of images from a healthy pig. Spiral CT scans atend-expiration and end-inspiration during volume controlled ventilation (tidal volume � 6–8 mL/kg ofbody weight) at PEEP levels from 0 to 25 cm H2O. Ventilation maps (relative distribution of tidal impedancechanges �Z) and RVD index maps during low-flow breath of 12 mL/kg of body weight with PEEP of 0 to25 cm H2O. Amount of tidal recruitment given in percent of lung in acute lung injury pig (gray bars) andhealthy pig (black bars) obtained by CT and RVD inhomogeneity (SDRVD) measured by electrical impedancetomography (EIT) with changes in PEEP from 0 to 25 cm H2O in acute lung injury pig (gray boxes) andhealthy animal (black boxes). Amount of cyclic tidal recruitment and SDRVD showed an excellent linearcorrelation (acute lung injury animal: r � .99, p � .001; healthy animal: r � .85, p � .05).

Figure 5. A, Course of regional ventilation delay(RVD) inhomogeneity (SDRVD) with changes inpositive end-expiratory pressure (PEEP) from 0to 25 cm H2O in healthy (control: outlined boxes)and lung-injured (acute lung injury, filled boxes)pigs. B, Correlation between amount of tidal re-cruitment in percent of lung and SDRVD in per-cent of inflation time. Outlined symbols indicatecontrol pigs; filled symbols, acute lung injurypigs. Different markers indicate different levels ofPEEP: circles, 0 cm H2O; downward triangles, 5cm H2O; upwards triangles, 10 cm H2O; squares,15 cm H2O; diamonds, 20 cm H2O; hexagons, 25cm H2O. A linear regression analysis shows mean(solid line) and 95% confidence intervals. r �Pearson linear correlation coefficient. Statistics:repeated-measures analysis of variance with fac-tors “acute lung injury” (for differences betweenhealthy and lung-injured animals) and “PEEP”(for differences between PEEP levels). Post hoctests (Newman-Keuls test) to separate differencesbetween low (0 and 5 cm H2O), moderate (10 and15 cm H2O), and high (20 and 25 cm H2O) PEEPwithin one group (a: p � .05 vs. low PEEP; b: p �.05 vs. middle PEEP; c: p � .05 vs. high PEEP)and between both groups (d: p � .05 vs. controllow PEEP; e: p � .05 vs. control moderate PEEP;f: p � .05 vs. control high PEEP). Box plots showmedian and quartiles.

907Crit Care Med 2012 Vol. 40, No. 3

Page 6: Tidal recruitment assessed by electrical impedance ... Muders - Tidal... · Tidal recruitment assessed by electrical impedance tomography and computed tomography in a porcine model

tion breath from 40 mL/kg BW to 12mL/kg BW to ensure feasibility of re-peated RVD measurements in patientsduring PEEP titration. Further, we stud-ied whether RVD measurements are suit-able to estimate changes in tidal recruit-ment at different PEEP levels. After thismodification, correlation between RVDindices and tidal recruitment was merelymoderate when calculated from lungquadrants (Tables 2 and 3), because re-ducing inflation volume decreased thesignal to be measured and thereby dis-proved signal-to-noise ratio of this mea-sure (SDC-4, Supplemental Digital Con-tent 1, http://links.lww.com/CCM/A361).

However, these data suggested thatchanges in heterogeneity of differentRVD indices contain more information

regarding tidal recruitment than theactual RVD value per se (Fig. 3). Thissounds reasonable because tidal re-cruitment causes a delay in regionalaeration when compared to aeration inalready opened regions. After this con-sideration and to further improve spa-tial resolution, RVD indices were calcu-lated for every single pixel, and thescatter of single pixel values was ex-pressed calculating their SD (SDRVD),which allows accessing changes in theinhomogeneity of ventilation timecourses. Even though measurement ofRVD inhomogeneity provides one single“global” number, it derives from a plen-titude of regionally measured RVDs andthus still reflects regional information.Additionally, RVD maps may help to

regionally visualize delayed alveolaraeration (Fig. 6).

The methodology used aimed atmaximizing the strength of the linearrelationship between tidal recruitmentassessed by CT and RVD obtained fromEIT. Using different thresholds or sev-eral numerical curve-fitting methodswith different model functions to quan-tify the ventilation delay did not im-prove our results (SDC-5 and Table S4;Supplemental Digital Content 1, http://links.lww.com/CCM/A361).

It is important to note that EIT scanswere obtained during 12 mL/kg BW slowinflation breath, whereas tidal recruit-ment was measured by CT during subse-quent “regular” tidal volume breaths of6–8 mL/kg BW (Fig. 1). We also calcu-lated RVD indices and inhomegeneityfrom EIT signals of regular 6–8 mL/kgbreaths that were recorded immediatelybefore the slow inflation was performed(SDC-4, Supplemental Digital Content 1,http://links.lww.com/CCM/A361). Thesewere not well-correlated with tidal re-cruitment, always underestimated RVDinhomogeneity when compared to slowinflation, and did not discriminate be-tween different PEEP levels, suggestingthat slow inflations with slightly in-creased tidal volume are required(SDC-4, Supplemental Digital Content 1,http://links.lww.com/CCM/A361). Ourstudy aimed at the development of a di-agnostic maneuver to characterize the re-gional mechanical behavior of the lung.Slow inflation maneuvers, however, arecommonly used in respiratory mechanicswith even higher inflation volumes andpressures (41). If PEEP titration usingthis EIT-based method will be proven inpatients in the future, then a possiblebenefit from that may surely exceed thepotential harm of a single low-flow breathwith a limited inflation volume. Hence,although obtained during an artificialmaneuver with different flow and infla-tion volume, measurement of RVD in-homogeneity delineates heterogeneityin regional lung mechanics at certainPEEP levels that, in turn, are affectingtidal recruitment during regularbreaths. This regional heterogeneitymay mainly reflect differences in alve-olar opening pressures (21), resultingin different aeration times during slowinflation. Therefore, RVD inhomogene-ity is not a direct measure of tidal re-cruitment, but rather a surrogate thatwas strongly related to tidal recruit-ment in our experiment.

Table 2. Linear correlation coefficients, global lung

Group

r (Linear Correlation Coefficient)Tidal Recruitment vs.

Quadrant RegionalVentilation Delay SDRVD

Control Acute Lung Injury Control Acute Lung Injury

Inter- individuallyAll 0.54a 0.90a

Within group 0.49a 0.49a 0.50b 0.92a

Pig Pig Pig Pig

Intra- individually 8 0.63a 1 0.48b 8 0.50b 1 0.99a

9 0.46b 2 0.55b 9 0.70 2 0.90a

10 0.73a 3 0.68a 10 0.33 3 0.98a

11 0.64a 4 0.48b 11 0.71 4 0.98b

12 0.73a 5 0.40b 12 0.79 5 0.91b

13 0.50b 6 0.79a 13 0.85b 6 0.93b

7 0.01 7 0.95b

Table 3. Linear correlation coefficients, electrical impedance tomography-covered lung region

Group

r (Linear Correlation Coefficient)Tidal Recruitment vs.

Quadrant RegionalVentilation Delay SDRVD

Control Acute Lung Injury Control Acute Lung Injury

Inter- individuallyAll 0.67a 0.87a

Within group 0.55a 0.66a 0.43b 0.88a

Pig Pig Pig Pig

Intra- individually 8 0.70a 1 0.72a 8 0.71 1 0.99a

9 0.50b 2 0.81a 9 0.76 2 0.93a

10 0.66a 3 0.66a 10 0.19 3 0.98a

11 0.68a 4 0.80a 11 0.71 4 0.88b

12 0.65a 5 0.44b 12 0.82b 5 0.69b

13 0.41b 6 0.79a 13 0.88b 6 0.88b

7 0.68 7 0.97a

908 Crit Care Med 2012 Vol. 40, No. 3

Page 7: Tidal recruitment assessed by electrical impedance ... Muders - Tidal... · Tidal recruitment assessed by electrical impedance tomography and computed tomography in a porcine model

Other EIT-Based Approaches

Recently, further EIT-based ap-proaches have been conducted to individ-ually optimize ventilatory settings (42).Changes in distribution and homogeneityof end-expiratory lung volume (43), tidalvolume (25, 26, 31, 33, 44, 45), and col-lapse (33) were quantified at differentPEEP levels.

A major advantage of EIT is its hightemporal resolution, allowing analysis ofaeration and ventilation time courses. Toour knowledge, our study is the first fo-cusing on regional distribution and inho-mogeneity of ventilation time courses (asopposed to the amount and distributionof ventilation) to quantify tidal recruit-ment and the influence of different PEEPlevels with EIT and CT as the gold stan-dard. Victorino et al (28) already de-scribed “inflation delay” in dependentzones during mechanical ventilation.Regional filling characteristics of differ-ent lung regions were found to be in-homogeneous in patients with ARDS(46) or in animal models (47–50). Othergroups found a large heterogeneity ofinflection points (51) derived from

global pressure-volume and regionalpressure-impedance curves.

However, these studies are limited byseveral restrictions (42) because no refer-ence methods were used (31, 44, 47, 48,50) or because combined measurements ofEIT and lung mechanics during vital capac-ity maneuvers were required (33, 51). Nev-ertheless, our results support the findingsthat global parameters differ from regionalimpedance curves (46, 48–50).

Limitations

Our study clearly has several limita-tions. The EIT belt only covers a limitedsection of the lung. However, calculatingtidal recruitment only from the EIT beltcovering CT slices did not affect our re-sults, indicating that the lung section an-alyzed by the EIT was representative.

Even though the oleic acid injectionmodel of lung injury is widely used (52),as this model mimics the pathologic fea-tures of ARDS (53), comparisons withpatients with ALI or ARDS should bemade with caution. The amount ofnonaerated lung tissue found in our ani-mals (Table 1) was comparable to that

described in a mixed population of pa-tients with ARDS (16). The amount ofrecruitable lung tissue (Table 1), how-ever, was higher in our animals, whichmight be explained by the increased ab-dominal pressure used in our model. Inpatients with a more focal loss of lungaeration and low potential for alveolarrecruitment, regional mechanical char-acteristics may differ from those in ourmodel. Reanalyzes of previously pub-lished data showed that adding a pulmo-nary ALI group would not expand ourfindings (SDC-6, Supplemental DigitalContent 1, http://links.lww.com/CCM/A361). Finally, the value of our methodhas yet to be proven in clinical studies.

From our data, we cannot define aSDRVD value that should be targeted dur-ing PEEP titration. However, a certainRVD inhomogeneity is present in evenhealthy and fully recruited lungs, andclinical studies have to identify a normalrange in patients. It is supposable that apredefined SDRVD number does not haveto be reached, but individually minimiz-ing this parameter might be appropriate.

Currently, a scientific debate concen-trates on the possible underlying mecha-nisms of tidal recruitment (5–7, 54, 55),which goes far beyond the scope of thisarticle. We have intensively discussed thisissue previously (56). However, withoutdoubt, these processes on the alveolarscale will lead to the development of venti-lator-induced lung injury (1–4) and, re-gardless of the mechanisms involved, willlead to impaired patient outcome (16).Consequently, the quantification of tidal re-cruitment by EIT at the bedside may helpto individualize ventilator management.

Changes in lung fluid and blood vol-ume might also influence impedance, es-pecially when comparing impedance in-formation over a longer period. Ourmethod, however, only takes tidal changesin impedance into account, minimizing theinfluence of fluid change (33).

In single pigs, we found RVD values ofventral regions to increase again by afurther increase in PEEP (as deduciblefrom the yellow to orange pixels in ven-tral regions at higher PEEP levels; Fig. 4,third row). Higher RVD values may alsoresult from delayed filling attributable toalveolar hyperdistention that is also sup-posed to cause ventilator-associated lunginjury. Even though the current use ofCT in detecting hyperdistention might bematter of discussion (SDC-2, Supplemen-talDigitalContent1,http://links.lww.com/CCM/A361) (57), we did not observe CT

Figure 6. Exemplary end-expiratory computed tomography (CT) scans (left) and regional ventilationdelay (RVD) maps (right) obtained from electrical impedance tomography (EIT) from a single pigduring ventilation with positive end-expiratory pressure (PEEP) levels of 20 (upper panels) and 25(lower panels) cm H2O. The marked lung region (circle) was not fully aerated when a PEEP of 20 cmH2O was used. Delayed aeration during inflation is notably visible by the red pixels of the respectivearea in the RVD map. During ventilation with a PEEP of 25 cm H2O, the marked region is fully aeratedat end-expiration. Consequently, minimal delay in aeration is visible on the corresponding RVD map.

909Crit Care Med 2012 Vol. 40, No. 3

Page 8: Tidal recruitment assessed by electrical impedance ... Muders - Tidal... · Tidal recruitment assessed by electrical impedance tomography and computed tomography in a porcine model

densities ��900, which should repre-sent hyperinflated regions; thus, based onour data, we were unable to systemati-cally examine this phenomenon. There-fore, conclusions have to be restricted todetection of tidal recruitment. However,it recently has been shown that the ben-efit of reducing tidal recruitment by in-creasing PEEP prevails over the effect ofincreasing alveolar strain in patients withhigh lung recruitability (16), suggestingthat minimizing cyclic tidal recruitmentis a critical matter.

CONCLUSION

In conclusion, our study suggests thatanalyzing the scatter of aeration times ofregional ventilation during a slow infla-tion maneuver as measured by EIT allowsassessing temporal inhomogeneity oftidal ventilation. Delayed regional aera-tion of the lung can be visualized by RVDmaps and quantified by RVD inhomoge-neity (SDRVD). SDRVD is a feasible estimateof the amount of tidal recruitment whenchanging PEEP in an animal model. Fur-ther studies are necessary to investigate ifEIT-based PEEP titration leads to reduc-tion in ventilator-induced lung injuryand is feasible in patients with ALI orARDS. If so, then individually adjustingthe lowest PEEP that minimizes tidal re-cruitment might also decrease injuriouseffects of mechanical ventilation.

ACKNOWLEDGMENTS

We thank Dr. Dirk Varelmann, Boston,MA, and Dr. Andreas Reske, Leipzig, Ger-many, for their helpful review of themanuscript and stimulating discussions.

REFERENCES

1. International consensus conferences in in-tensive care medicine. Ventilator-associatedlung injury in ARDS. American Thoracic So-ciety, European Society of Intensive CareMedicine, Societe de Reanimation LangueFrancaise. Intensive Care Med 1999; 25:1444–1452

2. Ranieri VM, Suter PM, Tortorella C, et al:Effect of mechanical ventilation on inflam-matory mediators in patients with acute re-spiratory distress syndrome: A randomizedcontrolled trial. JAMA 1999; 282:54–61

3. Slutsky AS: Lung injury caused by mechan-ical ventilation. Chest 1999; 116:9S–15S

4. Uhlig S: Ventilation-induced lung injury andmechanotransduction: Stretching it too far?Am J Physiol Lung Cell Mol Physiol 2002;282:L892–L896

5. Hubmayr RD: Another look at the opening

and collapse story. Crit Care Med 2009; 37:2667–2668

6. Hubmayr RD: Perspective on lung injury andrecruitment: A skeptical look at the openingand collapse story. Am J Respir Crit Care Med2002; 165:1647–1653

7. Rouby JJ, Puybasset L, Nieszkowska A, et al:Acute respiratory distress syndrome: lessonsfrom computed tomography of the wholelung. Crit Care Med 2003; 31:S285–S295

8. Ventilation with lower tidal volumes as com-pared with traditional tidal volumes for acutelung injury and the acute respiratory distresssyndrome. The Acute Respiratory DistressSyndrome Network. N Engl J Med 2000; 342:1301–1308

9. Amato MB, Barbas CS, Medeiros DM, et al:Effect of a protective-ventilation strategy onmortality in the acute respiratory distresssyndrome. N Engl J Med 1998; 338:347–354

10. Brower RG, Lanken PN, MacIntyre N, et al:Higher versus lower positive end-expiratorypressures in patients with the acute respira-tory distress syndrome. N Engl J Med 2004;351:327–336

11. Meade MO, Cook DJ, Guyatt GH, et al: Ven-tilation strategy using low tidal volumes, re-cruitment maneuvers, and high positive end-expiratory pressure for acute lung injury andacute respiratory distress syndrome: A ran-domized controlled trial. JAMA 2008; 299:637–645

12. Mercat A, Richard JC, Vielle B, et al: Positiveend-expiratory pressure setting in adultswith acute lung injury and acute respiratorydistress syndrome: A randomized controlledtrial. JAMA 2008; 299:646–655

13. Putensen C, Theuerkauf N, Zinserling J, et al:Meta-analysis: ventilation strategies and out-comes of the acute respiratory distress syn-drome and acute lung injury. Ann InternMed 2009; 151:566–576

14. Gattinoni L, Caironi P, Cressoni M, et al:Lung recruitment in patients with the acuterespiratory distress syndrome. N Engl J Med2006; 354:1775–1786

15. Slutsky AS, Hudson LD: PEEP or noPEEP–lung recruitment may be the solution.N Engl J Med 2006; 354:1839–1841

16. Caironi P, Cressoni M, Chiumello D, et al:Lung opening and closing during ventilationof acute respiratory distress syndrome. Am JRespir Crit Care Med 2010; 181:578–586

17. Talmor D, Sarge T, Malhotra A, et al: Me-chanical ventilation guided by esophagealpressure in acute lung injury. N Engl J Med2008; 359:2095–2104

18. Suter PM, Fairley B, Isenberg MD: Optimumend-expiratory airway pressure in patientswith acute pulmonary failure. N Engl J Med1975; 292:284–289

19. Gattinoni L, Pesenti A, Bombino M, et al:Relationships between lung computed tomo-graphic.density, gas exchange, and PEEP inacute respiratory failure. Anesthesiology1988; 69:824–832

20. Gattinoni L, Caironi P, Pelosi P, et al: Whathas computed tomography taught us about

the acute respiratory distress syndrome?Am J Respir Crit Care Med 2001; 164:1701–1711

21. Hickling KG: The pressure-volume curve isgreatly modified by recruitment. A mathe-matical model of ARDS lungs. Am J RespirCrit Care Med 1998; 158:194–202

22. Brown BH, Barber DC, Seagar AD: Appliedpotential tomography: possible clinical appli-cations. Clin Phys Physiol Meas 1985;6:109–121

23. Frerichs I, Hinz J, Herrmann P, et al: Detec-tion of local lung air content by electricalimpedance tomography compared with elec-tron beam CT. J Appl Physiol 2002; 93:660–666

24. Frerichs I, Hahn G, Schiffmann H, et al:Monitoring regional lung ventilation byfunctional electrical impedance tomographyduring assisted ventilation. Ann N Y Acad Sci1999; 873:493–505

25. Luepschen H, Meier T, Grossherr M, et al:Protective ventilation using electrical imped-ance tomography. Physiol Meas 2007; 28:S247–S260

26. Meier T, Luepschen H, Karsten J, et al: As-sessment of regional lung recruitment andderecruitment during a PEEP trial based onelectrical impedance tomography. IntensiveCare Med 2008; 34:543–550

27. Pulletz S, Elke G, Zick G, et al: Performanceof electrical impedance tomography in de-tecting regional tidal volumes during one-lung ventilation. Acta Anaesthesiol Scand2008; 52:1131–1139

28. Victorino JA, Borges JB, Okamoto VN, et al:Imbalances in regional lung ventilation: Avalidation study on electrical impedance to-mography. Am J Respir Crit Care Med 2004;169:791–800

29. Wolf GK, Arnold JH: Noninvasive assessmentof lung volume: Respiratory inductanceplethysmography and electrical impedancetomography. Crit Care Med 2005; 33:S163–S169

30. Wrigge H, Zinserling J, Muders T, et al: Elec-trical impedance tomography compared withthoracic computed tomography during aslow inflation maneuver in experimentalmodels of lung injury. Crit Care Med 2008;36:903–909

31. Zhao Z, Moller K, Steinmann D, et al: Eval-uation of an electrical impedance tomogra-phy-based global inhomogeneity index forpulmonary ventilation distribution. Inten-sive Care Med 2009; 35:1900–1906

32. Costa EL, Lima RG, Amato MB: Electricalimpedance tomography. Curr Opin Crit Care2009; 15:18–24

33. Costa EL, Borges JB, Melo A, et al: Bedsideestimation of recruitable alveolar collapseand hyperdistension by electrical impedancetomography. Intensive Care Med 2009; 35:1132–1137

34. Varelmann D, Muders T, Zinserling J, et al:Cardiorespiratory effects of spontaneousbreathing in two different models of experi-

910 Crit Care Med 2012 Vol. 40, No. 3

Page 9: Tidal recruitment assessed by electrical impedance ... Muders - Tidal... · Tidal recruitment assessed by electrical impedance tomography and computed tomography in a porcine model

mental lung injury: A randomized controlledtrial. Crit Care 2008; 12:R135–R142

35. Wrigge H, Zinserling J, Neumann P, et al:Spontaneous breathing improves lung aera-tion in oleic acid-induced lung injury. Anes-thesiology 2003; 99:376–384

36. Wrigge H, Zinserling J, Neumann P, et al:Spontaneous breathing with airway pressurerelease ventilation favors ventilation in de-pendent lung regions and counters cyclicalveolar collapse in oleic-acid-induced lunginjury: A randomized controlled computedtomography trial. Crit Care 2005;9:R780–R789

37. Lionheart WR: EIT reconstruction algo-rithms: pitfalls, challenges and recent devel-opments. Physiol Meas 2004; 25:125–142

38. Lundquist H, Hedenstierna G, Strandberg A,et al: CT-assessment of dependent lung den-sities in man during general anaesthesia.Acta Radiol 1995; 36:626–632

39. Vieira SR, Puybasset L, Richecoeur J, et al: Alung computed tomographic assessment ofpositive end-expiratory pressure-inducedlung overdistension. Am J Respir Crit CareMed 1998; 158:1571–1577

40. Suarez-Sipmann F, Bohm SH, Tusman G, etal: Use of dynamic compliance for open lungpositive end-expiratory pressure titration inan experimental study. Crit Care Med 2007;35:214–221

41. Gattinoni L, Eleonora C, Caironi P: Monitor-ing of pulmonary mechanics in acute respi-ratory distress syndrome to titrate therapy.Curr Opin Crit Care 2005; 11:252–258

42. Muders T, Luepschen H, Putensen C: Imped-

ance tomography as a new monitoring tech-nique. Curr Opin Crit Care 2010; 16:269–275

43. Bikker IG, Leonhardt S, Bakker J, et al: Lungvolume calculated from electrical impedancetomography in ICU patients at differentPEEP levels. Intensive Care Med 2009; 35:1362–1367

44. Frerichs I, Dargaville PA, van GenderingenH, et al: Lung volume recruitment after sur-factant administration modifies spatial distri-bution of ventilation. Am J Respir Crit CareMed 2006; 174:772–779

45. Dargaville PA, Rimensberger PC, Frerichs I:Regional tidal ventilation and complianceduring a stepwise vital capacity manoeuvre.Intensive Care Med 2010; 36:1953–1961

46. Hinz J, Gehoff A, Moerer O, et al: Regionalfilling characteristics of the lungs in me-chanically ventilated patients with acutelung injury. Eur J Anaesthesiol 2007; 24:414–424

47. van Genderingen HR, van Vught AJ, JansenJR: Regional lung volume during high-frequency oscillatory ventilation by electricalimpedance tomography. Crit Care Med 2004;32:787–794

48. Kunst PW, Bohm SH, Vazquez de Anda G, etal: Regional pressure volume curves by elec-trical impedance tomography in a model ofacute lung injury. Crit Care Med 2000; 28:178–183

49. Grant CA, Fraser JF, Dunster KR, et al: Theassessment of regional lung mechanics withelectrical impedance tomography: A pilot

study during recruitment manoeuvres. In-tensive Care Med 2009; 35:166–170

50. Rooney D, Friese M, Fraser JF, et al: Gravity-dependent ventilation distribution in ratsmeasured with electrical impedance tomog-raphy. Physiol Meas 2009; 30:1075–1085

51. Grychtol B, Wolf GK, Arnold JH: Differencesin regional pulmonary pressure-impedancecurves before and after lung injury assessedwith a novel algorithm. Physiol Meas 2009;30:S137–S148

52. Matute-Bello G, Frevert CW, Martin TR: An-imal models of acute lung injury. Am JPhysiol Lung Cell Mol Physiol 2008; 295:L379–L399

53. Wang HM, Bodenstein M, Markstaller K:Overview of the pathology of three widelyused animal models of acute lung injury. EurSurg Res 2008; 40:305–316

54. Bilek AM, Dee KC, Gaver DP III: Mechanismsof surface-tension-induced epithelial celldamage in a model of pulmonary airway re-opening. J Appl Physiol 2003; 94:770–783

55. Tsuchida S, Engelberts D, Peltekova V, et al:Atelectasis causes alveolar injury in nonat-electatic lung regions. Am J Respir Crit CareMed 2006; 174:279–289

56. Muders T, Wrigge H: New insights into ex-perimental evidence on atelectasis andcauses of lung injury. Best Practice and Re-search Clinical Anaesthesiology 2010; 24:171–182

57. Reske AW, Busse H, Amato MB, et al: Imagereconstruction affects computer tomo-graphic assessment of lung hyperinflation.Intensive Care Med 2008; 34:2044–2053

911Crit Care Med 2012 Vol. 40, No. 3