flow-controlled expiration: a novel ventilation mode to attenuate

RESPIRATION AND THE AIRWAY

Flow-controlled expiration: a novel ventilation modeto attenuate experimental porcine lung injuryU. Goebel1, J. Haberstroh2, K. Foerster3, C. Dassow1, H.-J. Priebe1, J. Guttmann1 and S. Schumann1*

1 Division for Experimental Anaesthesiology, Department of Anaesthesiology,2 Experimental Surgery, CEMT-FR and 3 Department ofCardiovascular Surgery, University Medical Centre Freiburg, Freiburg, Germany

* Corresponding author. E-mail: [email protected]

Editor’s key points

† The effects offlow-controlled expiration(FLEX) were studied in aporcine model of lunginjury.

† Addition of FLEX andvolume-controlledventilation improved lungmechanics and function,and reduced lung injury.

† Further studies arerequired to determinewhether FLEX mightimprove lung-protectiveventilation in humans.

Background. Whereas the effects of various inspiratory ventilatory modifications in lung injuryhave extensively been studied, those of expiratory ventilatory modifications are less wellknown. We hypothesized that the newly developed flow-controlled expiration (FLEX) modeprovides a means of attenuating experimental lung injury.

Methods. Experimental acute respiratory distress syndrome was induced by i.v. injection ofoleic acid in 15 anaesthetized and mechanically ventilated pigs. After established lunginjury (PaO2/FIO2 ratio ,27 kPa), animals were randomized to either a control group receivingvolume-controlled ventilation (VCV) or a treatment group receiving VCV with additionalFLEX (VCV+FLEX). At predefined times, lung mechanics and oxygenation were assessed. Atthe end of the experiment, the pigs were killed, and bronchoalveolar fluid and lung biopsieswere taken. Expression of inflammatory cytokines was analysed in lung tissue andbronchoalveolar fluid. Lung injuryscorewasdetermined on the basis of stained tissue samples.

Results. Compared with the control group (VCV; n¼8), the VCV+FLEX group (n¼7)demonstrated greater dynamic lung compliance and required less PEEP at comparable FIO2

(both P,0.05), had lower regional lung wet-to-dry ratios and lung injury scores (bothP,0.001), and showed less thickening of alveolar walls (an indicator of interstitial oedema)and de novo migration of macrophages into lung tissue (both P,0.001).

Conclusions. The newly developed FLEX mode is able to attenuate experimental lung injury.FLEX could provide a novel means of lung-protective ventilation.

Keywords: acute respiratory distress syndrome; oleic acid; positive pressure ventilation;pulmonary oedema

Accepted for publication: 2 December 2013

Positive pressure ventilation consists of active insufflation ofthe lungs followed by passive exhalation as a result of elasticrecoil forces of the respiratory system.1 Whereas the effectsof various modifications of the inspiratory phase (e.g. byvarying end-inspiratory volume, peak inspiratory pressure,and flow) have been investigated,2 – 5 with the exception ofPEEP,6 7 modifications of the expiratory phase have receivedlittle attention. Consequently, in routine mechanical ventila-tion, approximately half of the respiratory cycle (i.e. the expira-tory phase) is not utilized for active ventilatory management.

We studied the effects of modification of the expiratoryphase of mechanical ventilation byapplying a newlydevelopedmode of ventilation, flow-controlled expiration (FLEX). FLEXslows the expiratory peak flow rate and maintains decreasedflow throughout expiration, thereby prolonging the non-zeroflow phase (and, in turn, total expiratory flow time) andincreasing mean airway pressure at otherwise unchanged

ventilatory settings. This is expected to reduce airway collapseand oedema formation, especially in injured lungs. Wehypothesized that FLEX would attenuate experimental lunginjury.

MethodsThe study was approved by the Animal Welfare Committees ofthe University of Freiburg, Germany (Registration No: G-09/17),and was carried out in accordance with the German law foranimal protection and the animal care guidelines of the Euro-pean Community (86/609/EC).

Surgical preparation

Sixteen healthy German Landrace Hybrid pigs of either sex {bodyweight 62.5 (5.0) kg [mean (SD)]} were starved for 8 h and preme-dicated with i.m. 0.5 mg kg21 midazolam (Dormicumw, Roche,

British Journal of Anaesthesia 113 (3): 474–83 (2014)Advance Access publication 2 April 2014 . doi:10.1093/bja/aeu058

& The Author [2014]. Published by Oxford University Press on behalf of the British Journal of Anaesthesia. All rights reserved.For Permissions, please email: [email protected]

Downloaded from https://academic.oup.com/bja/article-abstract/113/3/474/2919885by gueston 13 February 2018

mailto:[email protected]



Grenzach-Wyhlen, Germany) and 20 mg kg21 ketamine hydro-chloride (Ketaminw 10%, Intervet, Unterschleißheim, Germany).Anaesthesia was induced with i.v. 2–4 mg kg21 propofol (Propo-folw 1%,Fresenius Kabi, Bad Homburg,Germany) and maintainedby infusions of 1–2 mg kg21 h21 midazolam, 4–6 mg kg21 h21

ketamine hydrochloride, and 10 mg kg21 h21 fentanyl citrate(Fentanyl Janssenw, Janssen-Cilag, Neuss, Germany). Musclerelaxation was maintained by i.v. 0.5 mg kg21 h21 vecuronium(Vecuronium-Inresaw, Inresa, Freiburg, Germany). After trachealintubation, the lungs were ventilated (Evita 4, Drager Medical,Lubeck, Germany) in the volume-controlled mode at a respiratoryrate of 15 bpm, a tidal volume of 7–8 ml kg21, an I:E ratio of 1:1.5,and PEEP of 8 cm H2O. Inspired oxygen fraction (FIO2 ) was main-tained at 0.21. Respiratory rate and I:E ratio were kept constant.Ringer’s solution (B. Braun Melsungen AG, Melsungen, Germany)was infused at 10 mg kg21 h21.

The animals were kept in the supine position. The rightcarotid artery was cannulated to monitor mean arterial pres-sure (MAP) and obtain blood samples (arterial blood sampler,Pico 50, Radiometer, Brønshøj, Denmark) for blood gas analysisand haemoximetry (Cobas B 121, Roche Diagnostics, Stuttgart,Germany). The left external jugular vein was cannulated for in-sertion of a pulmonary artery thermodilution catheter (7 Fr,Edwards, Irvine, CA, USA) via an introducer sheath (8.5 Fr,Arrow, Reading, PA, USA) for measurements of mean pulmon-ary artery pressure (MPAP), cardiac output (CO), pulmonary ca-pillary wedge pressure (PCWP), and central venous pressure. Asecond introducer sheath (7 Fr Prelude, Merit Medical, UT, USA)was inserted into the right external jugular vein for administra-tion of oleic acid. A suprapubic catheter was inserted into thebladder for urine collection.

Expiration control

We slowed the expiratory peak flow rate by inserting an expira-tory mechanical resistor in the expiratory limb of the ventila-tor.8 The resistor contained an aperture that could variablybe occluded by a cone that was connected to a computer-controlled linear motor (PS01-23Sx80, LinMot, Spreitenbach,Switzerland) for computer-controlled positioning. Flow datawere continuously sampled and analysed by a personal com-puter that controlled the linear motor system. Once start of in-spiration was detected by the flow signal, the cone was movedto a position occluding the aperture. At the start of expiration,the cone was pulled back at constant speed, thereby graduallyopening the aperture and decreasing expiratory resistance.

Induction of lung injury

Oleic acid (Oleic acid PhEur, 75096, Sigma Aldrich, Munich,Germany) was emulsified with an equal volume of 5%glucose (Glucose 5%, B. Braun Melsungen AG). Lung injurywas induced by repeated i.v. boli of 1 ml of oleic acid emulsionuntil PaO2

/FIO2ratio was ,27 kPa at an FIO2

of 1.0. Subsequent-ly, FIO2 and PEEP were adjusted to maintain PaO2 .8 kPa inaccordance with the ARDSnet recommendations.9 First, FIO2

was reduced to 0.8 to minimize absorption atelectasis. Subse-quently, PEEP was reduced stepwise (2 cm H2O per step) to

minimize peak pressure. The response of PaO2to these inter-

ventions was monitored by blood gas analysis. If PaO2 was.8 kPa for .30 min, FIO2

and PEEP were further reduced. IfPaO2

was ,8 kPa, FIO2was increased first to 0.8, followed by

stepwise increases of PEEP to a maximum of 15 cm H2O.

Experimental protocol

After established lung injury, lungs were recruited by applyingPEEP of 20 cm H2O for 15 s by end-inspiratory hold. Using acomputer-generated randomization sequence, animals wereallocated to either the control group receiving volume-controlled ventilation only (VCV, n¼8) or to the treatmentgroup receiving VCV plus flow-controlled expiration (VCV+FLEX; n¼8). The observation period lasted 6 h. At the end ofthe study, sternotomy was performed for right lung lobectomy.The animals were then killed by intracardiac potassiumchloride. Bronchoalveolar fluid was collected post-mortem.

Respiratory system mechanics

Airway pressure and flow were read off the ventilator at a sam-pling rate of 125 Hz. Dynamic compliance was calculated by mul-tiple regression analysis of pressure, volume, and flow curves.

Pulmonary markers of inflammatory response

Pro-inflammatory markers [interleukin (IL)-1b, IL-6, IL-8,tumour necrosis factor (TNF)-a] were measured in serum,bronchoalveolar fluid, and lung tissue using ELISA kits(DuoSet, R&D Systems, Wiesbaden, Germany) according tothe manufacturer’s instructions. Protein content was deter-mined by BCA assay (ThermoScientific, Rockford, IL, USA).IL-1b, IL-6, and IL-8 mRNA concentrations were determinedin lung tissue. One microgram of RNA was transcribed tocDNA using a reverse transcription kit (iScript, Bio Rad Labora-tories, Munchen, Germany). Quantitative real-time reversetranscriptase–polymerase chain reaction (qRT–PCR) was per-formed with a mastermix (ABsolute SYBR Green, ThermoScien-tific) monitored with an iCycler (Bio Rad Laboratories). Datawere normalized to b-actin.

Lung histopathology and lung injury score

Immediately before killing the animals, four lung biopsies weretaken from the ventral and dorsal portions of the apical andbasal lobes. Samples were either fixed in 4% formalin for histo-pathology, snap frozen, and stored at 2808C for molecularanalysis, or weighed for determination of the wet/dry ratio.Slices of 4 mm thickness were obtained by microtome andstained with haematoxylin/eosin and an antibody [mousemonoclonal (MAC-387) against macrophage-expressed cal-protectin; #ab22506, Cambridge, UK] for microscopic examin-ation. This included measurement of alveolar wall thicknessand count of de novo migrated macrophages in 10 independ-ent fields of vision in each sample. Two independent outsideexaminers experienced in lung histopathological assessmentand blinded to group assignment analysed the lung tissue sec-tions according to the scoring system of the American ThoracicSociety.10

Flow-controlled expiration BJA

475


Indicators of organ function

Cardiac index (CI), systemic vascular resistance index (SVRI),and pulmonary vascular resistance index (PVRI) were calcu-lated by standard formulae. CO values were measured in tripli-cate using ice-cold normal saline. Various indicators of organfunction (e.g. creatine kinase, NT-pro-BNP, and cystatin C)were analysed in the venous blood. Haemodynamic, haemato-logical, blood chemistry, blood gas, and ventilatory variableswere determined immediately before injection of oleic acid(pre-lung injury), after established lung injury (0 h), and then2 hourly until the end of the 6 h observation period.

Quantitative and statistical analysis

An a priori performed power calculation showed that a samplesize of n¼14 would be required to detect a change in respira-tory system compliance .15% at b¼0.2. Respiratory datawere analysed by a computerized statistical program (Graph-Pad Prism 5, GraphPad Software, San Diego, CA, USA). As wedid not find statistically relevant changes in the 2 hourly

determined PaO2/FIO2

ratios, lung mechanics, and bloodgases throughout the 6 h observation period after inductionof lung injury, we treated the 6 h observation period as onephase. Serial measurements were averaged by the method ofsummary measures.11 These summary data were treated asraw data and analysed by the Mann–Whitney U-test forbetween-group comparison. Serum cytokines were analysedby repeated-measures analysis of variance (ANOVA). If not indi-cated otherwise, values are expressed as mean (SD). A P-valueof ,0.05 is considered statistically significant.

ResultsAll animals survived throughout the observation period. In oneanimal of the VCV+FLEX group, wewere unable to achieve lunginjury; data from this animal were excluded from analysis.Before induction of lung injury, the PaO2

/FIO2ratio was compar-

able between the groups [VCV, 53.9 (5.6) kPa; VCV+FLEX: 54.9(3.9) kPa]. Induction of lung injury caused a sustained decreasein PaO2

/FIO2ratio by �80% in both groups [VCV, 13.9 (5.1) kPa;

VCV

500

0

Flo

w (

ml s

–1)

Vol

ume

(ml)

Paw

(cm

H2O

)

–500

–1000

500

400

300

200

100

40

30

20

10

0

40

30

20

10

0

0

500

400

300

200

100

0

0 1 2 3

500

0

–500

–10000 1 2 3

0 1 2 3 0 1 2 3

0

Inspiration Expiration Inspiration Expiration

1 2Time (s) Time (s)

3 0 1 2 3

VCV+FLEXA B

C D

E F

Fig 1 Representative diagrams of flow (A and B), volume (C and D), and airway pressure curves (E and F) in one animal each during ventilation with VCV(A, C, and E) and VCV+flow-controlled expiration (VCV+FLEX) (B, D, and F). Solid vertical lines indicate end-inspiration. Inspiratory curves were com-parable during both conditions. During expiration, decreases in peak flow (B), volume (D), and airway pressure (F) were slower during ventilation withVCV+FLEX compared with VCV.

BJA Goebel et al.

476


VCV+FLEX: 14.9 (8.5) kPa; each P,0.001]. Introduction of FLEXcaused characteristic changes in expiratory flow, pressure, andvolume curves (Fig. 1). Compared with VCV, VCV+FLEX wasassociated with slower decreases in flow (Fig. 1B), volume(Fig. 1D), and airway pressure (Fig. 1F) during expiration. Tidalvolume [VCV, 6.8 (0.5) ml kg21 BW; VCV+FLEX, 6.8 (0.7) mlkg21 BW; P¼0.89] and mean airway pressure [VCV, 20 (2) cmH2O; VCV+FLEX, 19 (3) cm H2O; P¼0.68] were comparablebetween the groups.

Compared with VCV, expiratory peak flow [VCV, 1109 (88)ml s21; VCV+FLEX, 331 (33) ml s21; P¼0.0015] and plateaupressures [VCV, 35 (3) cm H2O; VCV+FLEX, 30 (6) cm H2O;P¼0.024] were lower during VCV+FLEX. FLEX allowed main-taining comparable PaO2 [VCV, 11.2 (3.3) kPa; VCV+FLEX, 10.4(1.7) kPa; P¼0.121] and FIO2

(Fig. 2) at �20% lower PEEP[VCV, 11.5 (1.8) cm H2O; VCV+FLEX, 9.3 (1.6) cm H2O;P¼0.013]. This was accompanied by a decrease in PCO2 [VCV,8.3 (0.9) kPa; VCV+FLEX: 7.2 (0.5) kPa; P¼0.0289] and an in-crease in dynamic lung compliance in the FLEX-treatedanimals [VCV, 18.4 (1.7) ml cm H2O21; VCV+FLEX: 22.4 (3.8)ml cm H2O21; P¼0.017]. Flow at end-expiration was similar inboth groups [VCV, 26 (16) ml s21; VCV+FLEX, 0 (1) ml s21].Wet-to-dry ratios of lung tissue from both ventral and dorsalsegments were lower in the VCV+FLEX group than in the VCVgroup (P,0.001; Fig. 3). Within lung segments, differences inwet-to-dry ratios between the groups tended to be more pro-nounced in the ventral (�60%) than in the dorsal segments(�35%). Lung injury scores were lower in the VCV+FLEXgroup than in the VCV group (P,0.001; Fig. 3). FLEX was asso-ciated with lesser alveolar wall thickness (P,0.001; Fig. 4A

and B) and lower macrophage count (P,0.001; Fig. 5A and B).Pulmonary mRNA and cytokine expression of IL-1b, IL-6,

IL-8, and TNF-a did not differ significantly between thegroups and lung segments (Table 1). Haematological variablesand serum concentrations of various indicators of organfunction remained comparable in both groups throughoutthe observation period (Tables 2 and 3).

While heart rate, MAP, CI, SVRI, and PVRI did not show anydifferences between the groups, MPAP (P¼0.021) and PCWP(P¼0.012) were significantly lower during lung injury in theVCV+FLEX group than in the VCV group (Table 4).

DiscussionCompared with conventional VCV, active modification ofthe expiratory phase by FLEX mode was associated with(i) improved dynamic lung compliance and oxygenation(reflected by the need for lower PEEP to maintain comparablePaO2

), and (ii) ameliorated lung injury (reflected by lowerwet-to-dry ratios of lung tissue, lower lung injury score, lesseralveolar wall thickness, and lower macrophage count).

Apart from applying PEEP, only two approaches have tar-geted the expiratory phase of mechanical ventilation. One isacceleration of expiration by applying pressures below PEEPin early expiration to shorten expiration time, thereby reducingthe risk of dynamic hyperinflation. Automatic tube compensa-tion (ATC) utilizes this approach.12 – 16 Another approach of tar-geting the expiratory phase is the slowing of expiration. Thishad previously been achieved by adding constant resistancesto the expiratory limb of the ventilator tubing system.17 18

However, clinically relevant benefits could not be demon-strated,18 19 and the technique is no longer available. Although

1.0

0.8

0.6

FI O

2

0.4

0.2

0.0VCV VCV+FLEX

Fig 2 Inspired oxygen fraction (FIO2 ). Data are presented as boxplots, displaying median, 25–75 percentiles, and full range. VCV,volume-controlled ventilation; FLEX, flow-controlled expiration.There was no significant difference between the groups (P¼0.978).

20A

B

Ventral Dorsal

***

***

***

15

Wet

-to-

dry

ratio

Lung

inju

ry s

core

10

5

VCV VCVVCV+FLEX

VCV VCV+FLEX

VCV+FLEX0

4

2

0

6

Fig 3 (A) Wet-to-dry ratios in ventral and dorsal parts of the lungs.(B) Lung injury scores according to the American Thoracic Societyscoring system. Data are presented as box plots, displayingmedian, 25–75 percentiles, and full range. VCV, volume-controlledventilation; FLEX, flow-controlled expiration. ***P,0.001.


477


the concept of slowing expiration is similar to that of FLEX, thereason for the lack of proven benefit of this technique mightbe that constant expiratory resistance increases the risk ofdynamic hyperinflation (auto-PEEP/intrinsic PEEP) by uniformlyincreasing the expiratory time constant. This can lead to incom-plete expiration if not compensated for by prolonged expirationtime. In contrast, the FLEX mode actively modifies the pattern ofexpiration. Its main characteristic is reduced expiratory peakflow rate during the early phase of expiration. This results in alonger duration of elevated airway pressure during expirationwhich, in turn, increases expiratory gas flow during late expir-ation. Both attenuate end-expiratory closure of distal airwaysand alveolar collapse, thereby facilitating lung emptying.

The shape of the inspiratory flow curve is determined by theventilator mode. Looking at the ventilator as either a flowsource (volume-controlled ventilation) or a pressure source(pressure-controlled ventilation), FLEX can be viewed asvolume-controlled expiration with constant flow rate in con-trast to the conventional ventilation with pressure-controlledpassive expiration. These characteristics of FLEX have the po-tential of being lung protective by improving airflow dynamics,lung compliance, alveolar stability, and oxygenation.

Oleic acid-induced lung injury caused a significant decreasein compliance. In general, low compliance is the predominantcause for high inspiratory peak pressure. Under such condi-tions, the driving pressure for passive expiration is high. If

VCVA

Ven

tral

Dor

sal

VCV+FLEX

B Ventral25

20

15

Alv

eola

r w

all t

hick

ness

(µm

)

10

5

0

Dorsal

***

***

VCV VCV+FLEX VCV VCV+FLEX

50 µm 50 µm

50 µm 50 µm

Fig 4 (A) Representative lung tissue staining for measurement of alveolar wall thickness in ventral and dorsal lung areas. (B) Summary data of al-veolar wall thickness. Data are presented as box plots, displaying median, 25–75 percentiles, and full range. VCV, volume-controlled ventilation;FLEX, flow-controlled expiration. ***P,0.001.

BJA Goebel et al.

478


passive expiration is viewed as unloading of a pneumatic cap-acitor, the process is fast if the capacitance (i.e. compliance) islow. The unloading dynamics are characterized by the expira-tory time constant tex which reflects the time period for expo-nential unloading from expiratory peak flow to 36.8% (¼1/e) ofthis flow. Complete expiration requires about 3tex.20 Exponen-tial decay of expiratory peak flow within three time constantswill lead to unphysiological acceleration of lung emptying.This is expected to cause considerable shear forces andairway collapse with resultant adverse effects on oxygenationand lung morphology, especially during underlying lung injury.As the injured lung is characterized by increased collapsibility,

lung recruitment can be expected to have only short-livedrather than sustained beneficial effect on lung function.Slowing of expiration and shortening of the zero-flow phaseduring expiration should stabilize the lung by maintainingairway pressure during expiration and leaving less time forlung collapse. Interventions that favourably affect the dynam-ics of exponential unloading can thus be expected to be ofbenefit. In the presence of a pronounced oleic acid-eliciteddecrease in pulmonary compliance, FLEX reduced the highexpiratory peak flow observed during conventional VCV by ap-proximately two-thirds, thereby eliminating tex as the main de-terminant of expiratory airflow dynamics. These findings and

VCVA

Ventral200

150

100

Mac

roph

age

coun

t(p

er fi

eld

of v

isio

n)

50

0

Dorsal

VCV VCV+FLEX

***

***

VCV VCV+FLEX

B

Ven

tral

Dor

sal

VCV+FLEX

200 µm 200 µm

200 µm 200 µm

Fig 5 (A) Representative lung tissue staining for alveolar macrophage count in ventral and dorsal lung areas. (B) Summary data of alveolar macro-phage counts were derived from counts in 10 randomly chosen fields of vision in each of eight lung biopsies in ventral and dorsal areas. Data arepresented as box plots, displaying median, 25–75 percentiles, and full range. VCV, volume-controlled ventilation; FLEX, flow-controlled expiration.***P,0.001.


479


data derived in a mathematical model21 emphasize the import-ance of time-dependent factors during mechanical ventilation.

Benefits similar to those observed in our study were previ-ously demonstrated for reduced inspiratory peak flow rates inrabbits.5 The adverse pulmonary effects of mechanical ventila-tion with high tidal volumes were ameliorated if peak inspira-tory flow was reduced. The associated reduced parenchymalshear forces22 might have been caused by a more homoge-neous distribution of air during reduced inspiratory flow. Thissuggests that limiting peak airway flows during mechanicalventilation might per se be beneficial for the lungs.

It could be argued that the beneficial effects observedduring FLEX were secondary to an increase in mean airwaypressure. However, as ventilation with FLEX allowed loweringof PEEP in the VCV+FLEX group, plateau pressure was lowerthan in the VCV group resulting in comparable mean airwaypressures in both groups at comparable tidal volumes. Thus,the beneficial effects of FLEX cannot be attributed to highermean airway pressures. As cessation of end-expiratory flowrates indicated complete expiration in both groups, neithercan intrinsic PEEP have contributed to the findings. Rather,FLEX likely improved lung performance by delaying alveolarclosure during expiration, resulting in maintained alveolar PO2

and facilitating gas exchange until the late phase of expiration.Improved oxygenation and lower PCO2 in the FLEX group aresuggestive evidence of such a mechanism.

Improved ventilatory characteristics during FLEX wereaccompanied by morphological improvements. Lesser alveolar

wall thickening and lower alveolar macrophage count in theVCV+FLEX group than in the VCV group are consistent with at-tenuation of alveolar and interstitial oedema formation. Thismight have been related to increased intrapulmonary pressurecaused by the FLEX-induced retarded expiration, resulting in alower transalveolar filtration pressure through lower MPAP andPCWP on the vascular (up-stream pressure) side, and a longerduration of elevated airway pressure above PEEP level duringexpiration on the airway (down-stream pressure) side.

Airway inflammation is a concomitant phenomenon of bothexperimental and human lung injury, and is possibly involved inthe pathogenesis of airway remodelling in lung injury.23 FLEX-associated improved airflow dynamics and respiratory systemcompliance, and the likely resultant reduced mechanical stresson the small airways, might further reduce the stimuli for par-enchymal remodelling. Interestingly, functional and morpho-logical improvements were not accompanied by changes inpulmonary mRNA and cytokine expression. This is in accord-ance with previous studies,24 25 suggesting that lack of agree-ment between morphological and molecular effects might betypical of oleic acid-induced lung injury.

Critique of methods

I.V. injection of oleic acid is an established method of inducingstable experimental lung injury.26 It is characterized byelevatedairway pressure, reduced lung compliance, alveolar oedema,impaired gas exchange and pulmonary hypertension.26 I.V.

Table 1 mRNA and protein expression of inflammatory markers in lung tissue. VCV, volume-controlled ventilation; VCV+FLEX, VCV plusflow-controlled expiration; IL, interleukin; TNF, tumour necrosis factor. Values are given as mean (SD). Statistical tests were performed using theMann–Whitney U-test. No significant differences between the groups were identified for any variable by repeated-measures ANOVA

Lungarea

Group IL-6 mRNA(arbitrary units)

IL-6 protein(pg mg21)

IL-8 mRNA(arbitrary units)

IL-8 protein(pg mg21)

IL-10 mRNA(arbitrary units)

IL-1b protein(pg mg21)

TNFa protein(pg mg21)

Overall VCV 0.8 (0.6) 216 (51) 17 (12) 107 (71) 0.20 (0.18) 90 (28) 28 (5)VCV+FLEX 0.7 (0.9) 243 (64) 13 (8) 86 (98) 0.15 (0.05) 91 (23) 31 (9)

Ventral VCV 1.1 (0.6) 226 (42) 11 (4) 122 (84) 0.15 (0.07) 88 (26) 26 (4)VCV+FLEX 0.9 (1.2) 250 (75) 10 (7) 95 (119) 0.16 (0.06) 91 (25) 30 (9)

Dorsal VCV 0.6 (0.4) 206 (61) 22 (16) 91 (57) 0.25 (0.24) 93 (30) 29 (6)VCV+FLEX 0.5 (0.4) 235 (56) 16 (8) 77 (81) 0.15 (0.03) 91 (23) 31 (9)

Table 2 Protein expression of inflammatory markers in serum and bronchoalveolar fluid. VCV, volume-controlled ventilation; VCV+FLEX, VCV plusflow-controlled expiration; IL, interleukin; TNF, tumour necrosis factor; BALF, bronchoalveolar fluid. Values are given as mean (SD). No significantdifferences between the groups were identified for any variable by repeated-measures ANOVA

Variable (pg mg21) Treatment group Time 21 h Time 0 h Time 2 h Time 4 h Time 6 h BALF

IL-6 VCV 268 (236) 256 (209) 366 (291) 356 (240) 344 (234) 134 (43)VCV+FLEX 275 (102) 229 (115) 463 (280) 369 (145) 345 (114) 143 (95)

IL-8 VCV 121 (75) 193 (83) 200 (109) 222 (98) 190 (125) 24 (13)VCV+FLEX 194 (316) 162 (111) 196 (217) 72 (74) 72 (95) 34 (19)

IL-1b VCV 16 (19) 22 (15) 20 (20) 27 (21) 23 (16) 9 (4)VCV+FLEX 7 (11) 9 (14) 10 (11) 13 (14) 12 (16) 7 (3)

TNFa VCV 26 (31) 53 (38) 53 (38) 33 (31) 53 (84) 5 (3)VCV+FLEX 89 (164) 130 (165) 145 (198) 115 (226) 84 (179) 4 (2)

BJA Goebel et al.

480


administration of oleic acid promotes alveolar, interstitial, orboth flooding by increasing endothelial permeability and block-ing active sodium transport.27 It might cause structural disrup-tions in the hydrophobic lipid bilayer core, and disturbmembrane fluidity and lipid–protein interactions.28 Compar-able PaO2

/FIO2ratios in both groups throughout the observation

period are indicative of comparable and sustained lung injury.In the presence of oleic acid-induced lung injury, addition of

FLEX to VCVdid not adversely affect gene expression of inflam-matory mediators or their release into lung tissue, serum, andbronchoalveolar fluid, or various biochemical indicators oforgan function when compared with VCV alone. Despiteincreased intrathoracic pressure caused by FLEX-associated re-tardation of exhalation, there were no significant differencesbetween the groups in most haemodynamic variables. MPAPand PCWP were even lower in the VCV+FLEX group comparedwith the VCV group.

We assessed the effects of FLEX during the early phase ofacute, oleic acid-induced experimental lung injury. Thus, ourfindings cannot automatically be translated to treatment ofhuman lung injury. We used FLEX in combination with VCV.Although it can be applied in any controlled ventilatory mode(it merely requires adjustment of expiratoryflowto avoid intrin-sic PEEP), the effects of FLEX-associated increased expiratory

Table 3 Serum concentrations of variables reflecting various organ functions. VCV, volume-controlled ventilation; VCV+FLEX, VCV plusflow-controlled expiration; AST, aspartate aminotransferase; ALT, alanine aminotransferase; AP, alcaline phosphatase; CK, creatine kinase; CK-MB,creatine kinase muscle and brain subunit; LDH, lactate dehydrogenase; NT-proBNP, N-terminal pro-brain natriuretic peptide. Values are given asmean (SD). No significant differences between the groups were identified for any variable by repeated-measures ANOVA

Variable Treatment group Time 21 h Time 0 h Time 2 h Time 4 h Time 6 h

Leucocytes (×1000 ml21) VCV 3.5 (0.7) 10.2 (3.8) 7.2 (4.5) 7.9 (3.9) 9.6 (4.3)VCV+FLEX 3.8 (1.0) 11.4 (4.3) 7.9 (3.7) 8.3 (3.5) 8.3 (4.1)

Platelets (106 ml21) VCV 124 (25) 168 (57) 213 (63) 221 (78) 199 (93)VCV+FLEX 117 (32) 132 (69) 197 (72) 241 (51) 218 (76)

Creatinine (mg dl21) VCV 0.87 (0.11) 0.77 (0.13) 0.77 (0.16) 0.76 (0.14) 0.78 (0.13)VCV+FLEX 0.83 (0.15) 0.72 (0.15) 0.81 (0.21) 0.78 (0.12) 0.81 (0.19)

Uric acid (mg dl21) VCV 12 (5) 14 (5) 15 (3) 12 (5) 16 (4)VCV+FLEX 14 (3) 13 (7) 13 (6) 13 (5) 13 (5)

Cystatin C (mg litre21) VCV 0.7 (0.1) 0.73 (0.1) 0.81 (0.14) 0.75 (0.1) 0.77 (0.1)VCV+FLEX 0.7 (0.1) 0.79 (0.2) 0.76 (0.2) 0.77 (0.14) 0.76 (0.12)

AST (U litre21) VCV 27 (4) 27 (4) 38 (8) 41 (13) 35 (12)VCV+FLEX 27 (6) 26 (5) 35 (12) 38 (12) 34 (8)

ALT (U litre21) VCV 40 (2) 41 (3) 44 (6) 54 (12) 47 (13)VCV+FLEX 38 (2) 40 (2) 41 (5) 48 (13) 45 (7)

AP (U litre21) VCV 86 (13) 88 (16) 90 (13) 82 (10) 83 (14)VCV+FLEX 73 (15) 81 (10) 83 (12) 81 (7) 79 (7)

g-GT (unit litre21) VCV 28 (13) 30 (12) 29 (10) 28 (9) 29 (12)VCV+FLEX 32 (16) 33 (8) 31 (13) 30 (8) 28 (8)

CK (unit litre21) VCV 1437 (122) 3341 (452) 3145 (298) 2978 (301) 3012 (396)VCV+FLEX 1618 (141) 3093 (510) 3412 (534) 3125 (464) 3225 (485)

CK-MB (unit litre21) VCV 250 (87) 306 (65) 341 (54) 278 (58) 257 (50)VCV+FLEX 298 (59) 343 (86) 365 (67) 301 (73) 299 (88)

LDH (unit litre21) VCV 421 (76) 1134 (211) 1265 (542) 1054 (356) 867 (564)VCV+FLEX 402 (63) 1345 (368) 1096 (593) 1104 (502) 945 (437)

NT-proBNP (pg ml21) VCV ,5 7 (1) 7 (2) 7 (2) 7 (2)VCV+FLEX ,5 8 (2) 6 (2) 7 (2) 6 (1)

Table 4 Haemodynamic variables in the VCVand VCV+FLEX groupsbefore (t¼21 h) and during (t¼0–6 h) lung injury. VCV,volume-controlled ventilation; VCV+FLEX, VCV plus flow-controlledexpiration; HR, heart rate; MAP, mean arterial pressure; CVP, centralvenous pressure; CI, cardiac index; MPAP, mean pulmonary arterypressure; PCWP, pulmonary capillary wedge pressure, SVRI,systemic vascular resistance index; PVRI, pulmonary vascularresistance index. Values are means (SD). During lung injury, data areaveraged from 2 hourly measurements using the method ofsummary measures. *P,0.05 compared with preceding VCV value

Variable VCV VCV1FLEX VCV VCV1FLEXBefore (21 h) lunginjury

During (0–6 h) lunginjury

HR (beats min21) 90 (22) 93 (26) 104 (28) 102 (29)

MAP (kPa) 10.3 (2.4) 9.6 (0.9) 11.4 (1.9) 10.9 (2.0)

CVP (kPa) 0.8 (0.4) 0.7 (0.4) 1.2 (0.3) 1.2 (0.7)

CI (ml min21 kg21) 97 (23) 96 (25) 75 (23) 61 (17)

MPAP (kPa) 2.5 (0.4) 2.5 (0.5) 5.2 (0.4) 4.7 (0.5)*

PCWP (kPa) 1.2 (0.3) 0.9 (0.1) 1.5 (0.1) 1.2 (0.1)*

SVRI(dyn s cm21 kg21)

63 (25) 60 (19) 86 (29) 101 (23)

PVRI(dyn s cm25 kg21)

9 (2) 11 (4) 34 (10) 36 (11)


481


resistance during ventilator modes supporting spontaneousbreathing remain to be investigated.

ConclusionsCompared with conventional mechanical ventilation, additionof our newlydeveloped FLEX attenuated lung injury in a porcinemodel of oleic acid-induced lung injury. This novel treatmentmode has the potential to improve the therapeutic effective-ness of respiratory support in human lung injury.

Authors’ contributionsAll authors drafted the article and added substantial intellec-tual content; U.G.: experimental procedure and analysis ofhistological data; J.H.: experimental procedure and data ana-lysis; K.F.: experimental procedure; C.D.: experimental assist-ance and molecular biological data analysis; H.-J.P.: studydesign and data analysis; J.G.: study design and experimentalassistance; S.S.: study design, technical developments, experi-mental assistance, and data analysis.

AcknowledgementsThe authors gratefully acknowledge the expert skills ofMatthias Schneider in developing and manufacturing theexpiratory resistor device and Torsten Weich and Jens Neumannfor analysing the lung tissue sections.

Declaration of interestNone declared.

FundingThis work was supported by the Deutsche Forschungsge-meinschaft (DFG GU 691/6-2) and by departmental funding.

References1 Soni N, Williams P. Positive pressure ventilation: what is the real

cost? Br J Anaesth 2008; 101: 446–57

2 The Acute Respiratory Distress Syndrome Network. Ventilation withlower tidal volumes as compared with traditional tidal volumes foracute lung injury and the acute respiratory distress syndrome. TheAcute Respiratory Distress Syndrome Network. N Engl J Med 2000;342: 1301–8

3 Campbell RS, Davis BR. Pressure-controlled versus volume-controlled ventilation: does it matter? Respir Care 2002; 47: 416–26

4 Lessard MR, Guerot E, Lorino H, Lemaire F, Brochard L. Effects ofpressure-controlled with different I:E ratios versus volume-controlled ventilation on respiratory mechanics, gas exchange,and hemodynamics in patients with adult respiratory distress syn-drome. Anesthesiology 1994; 80: 983–91

5 Maeda Y, Fujino Y, Uchiyama A, Matsuura N, Mashimo T,Nishimura M. Effects of peak inspiratory flow on development ofventilator-induced lung injury in rabbits. Anesthesiology 2004;101: 722–8

6 Brower RG, Lanken PN, MacIntyre N, et al. Higher versus lower posi-tive end-expiratory pressures in patients with the acute respiratorydistress syndrome. N Engl J Med 2004; 351: 327–36

7 Grasso S, Fanelli V, Cafarelli A, et al. Effects of high versus low posi-tive end-expiratory pressures in acute respiratory distress syn-drome. Am J Respir Crit Care Med 2005; 171: 1002–8

8 Schumann S, Goebel U, Haberstroh J, et al. Determination of respira-tory system mechanics during inspiration and expiration by flow-controlled expiration (FLEX): a pilot study in anesthetized pigs.Minerva Anestesiol 2014; 80: 19–28

9 NIH NHLBI Clinical Network. Ventilator Protocol Card. Available fromwww.ardsnet.org/system/files/Ventilator%20Protocol%20Card.pdf (accessed 4 November 2013)

10 Matute-Bello G, Downey G, Moore BB, et al. An official AmericanThoracic Society workshop report: features and measurements ofexperimental acute lung injury in animals. Am J Respir Cell MolBiol 2011; 44: 725–38

11 Matthews JN, Altman DG, Campbell MJ, Royston P. Analysis of serialmeasurements in medical research. Br Med J 1990; 300: 230–5

12 Haberthur C, Mehlig A, Stover JF, et al. Expiratory automatic endo-tracheal tube compensation reduces dynamic hyperinflation in aphysical lung model. Crit Care 2009; 13: R4

13 Uchiyama A, Nishimura M, Amata M, Mashimo T, Fujino Y. A new ex-piratory support system for resolving air trapping in lungs duringmechanical ventilation: a lung model study. Technol Health Care2007; 15: 213–20

14 Mols G, Rohr E, Benzing A, Haberthur C, Geiger K, Guttmann J.Breathing pattern associated with respiratory comfort during auto-matic tube compensation and pressure support ventilation innormal subjects. Acta Anaesthesiol Scand 2000; 44: 223–30

15 Guttmann J, Bernhard H, Mols G, et al. Respiratory comfort of auto-matic tube compensation and inspiratory pressure support in con-scious humans. Intensive Care Med 1997; 23: 1119–24

16 Haberthur C, Mols G, Elsasser S, Bingisser R, Stocker R, Guttmann J.Extubation after breathing trials with automatic tube compensa-tion, T-tube, or pressure support ventilation. Acta AnaesthesiolScand 2002; 46: 973–9

17 Aerts JG, van den Berg B, Bogaard JM. Controlled expiration inmechanically-ventilated patients with chronic obstructive pulmon-ary disease (COPD). Eur Respir J 1997; 10: 550–6

18 Georgopoulos D, Mitrouska I, Markopoulou K, Patakas D,Anthonisen NR. Effects of breathing patterns on mechanically ven-tilated patients with chronic obstructive pulmonary disease anddynamic hyperinflation. Intensive Care Med 1995; 21: 880–6

19 Gultuna I, Huygen PE, Ince C, Strijdhorst H, Bogaard JM, Bruining HA.Clinical evaluation of diminished early expiratory flow (DEEF) venti-lation in mechanically ventilated COPDpatients. Intensive Care Med1996; 22: 539–45

20 Guttmann J, Eberhard L, Fabry B, et al. Time constant/volume rela-tionship of passive expiration in mechanically ventilated ARDSpatients. Eur Respir J 1995; 8: 114–20

21 Bates JH, Irvin CG. Time dependence of recruitment and derecruit-ment in the lung: a theoretical model. J Appl Physiol 2002; 93:705–13

22 Mead J, Takishima T, Leith D. Stress distribution in lungs: a model ofpulmonary elasticity. J Appl Physiol 1970; 28: 596–608

23 Morales MM, Pires-Neto RC, Inforsato N, et al. Small airway remod-eling in acute respiratory distress syndrome: a study in autopsy lungtissue. Crit Care 2011; 15: R4

24 Heuer JF, Pelosi P, Hermann P, et al. Acute effects of intracranialhypertension and ARDS on pulmonary and neuronal damage: arandomized experimental study in pigs. Intensive Care Med 2011;37: 1182–91

25 Rosenthal C, CaroniaC,Quinn C, Lugo N,SagyM. Acomparison amonganimal models of acute lung injury. Crit Care Med 1998; 26: 912–6

BJA Goebel et al.

482


www.ardsnet.org/system/files/Ventilator%20Protocol%20Card.pdf




26 Matute-Bello G, Frevert CW, Martin TR. Animal models of acute lunginjury. Am J Physiol Lung Cell Mol Physiol 2008; 295: L379–99

27 Vadasz I, Morty RE, Kohstall MG, et al. Oleic acid inhibits alveolarfluid reabsorption: a role in acute respiratory distress syndrome?Am J Respir Crit Care Med 2005; 171: 469–79

28 Zwissler B, Forst H, Ishii K, Messmer K. A new experimental model ofARDS and pulmonary hypertension in the dog. Res Exp Med (Berl)1989; 189: 427–38

Handling editor: H. C. Hemmings


483


flow-controlled expiration: a novel ventilation mode to attenuate

Documents