Assessment of distribution of ventilation and regional lung compliance by electrical impedance tomography in anaesthetized horses undergoing alveolar recruitment manoeuvres

Accepted Manuscript Assessment of distribution of ventilation and regional lung compliance by electrical impedance tomography in anaesthetized horses undergoing alveolar recruitment manoeuvres Tamas D. Ambrisko, Johannes Schramel, Klaus Hopster, Sabine Kästner, Yves Moens PII:

S1467-2987(17)30004-1

DOI:

10.1016/j.vaa.2016.03.001

Reference:

VAA 32

To appear in:

Veterinary Anaesthesia and Analgesia

Received Date: 22 September 2015 Revised Date:

18 January 2016

Accepted Date: 4 March 2016

Please cite this article as: Ambrisko TD, Schramel J, Hopster K, Kästner S, Moens Y, Assessment of distribution of ventilation and regional lung compliance by electrical impedance tomography in anaesthetized horses undergoing alveolar recruitment manoeuvres, Veterinary Anaesthesia and Analgesia (2017), doi: 10.1016/j.vaa.2016.03.001. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

RESEARCH PAPER

2

T Ambrisko et al.

3

Alveolar recruitment manoeuvre in horses

4

Assessment of distribution of ventilation and regional lung compliance by electrical

5

impedance tomography in anaesthetized horses undergoing alveolar recruitment

6

manoeuvres

7

Tamas D Ambrisko*, Johannes Schramel*, Klaus Hopster†, Sabine Kästner† & Yves Moens*

8

*Anaesthesiology and Perioperative Intensive Care Medicine, Department for Companion

9

Animals and Horses, University of Veterinary Medicine, Vienna, Austria

SC

M AN U

10

RI PT

1

†Clinic for Horses, University of Veterinary Medicine, Hannover, Germany

11

Correspondence: Tamas D Ambrisko, Anaesthesiology and Perioperative Intensive Care

13

Medicine, Department for Companion Animals and Horses, University of Veterinary Medicine,

14

Veterinärplatz 1, A-1210 Vienna, Austria. E-mail: [email protected]

TE D

12

15

Abstract

17

Objective To examine changes in the distribution of ventilation and regional lung compliances in

18

anaesthetized horses during the alveolar recruitment manoeuvre (ARM).

19

Study design Experimental study in which a series of treatments were administered in a fixed

20

order on one occasion.

21

Animals Five adult Warmblood horses.

22

Methods Animals were anaesthetized (xylazine, midazolam–ketamine, isoflurane), placed in

23

dorsal recumbency and ventilated with 100% oxygen using peak inspiratory pressure (PIP) and

24

positive end-expiratory pressure (PEEP) of 20 cmH2O and 0 cmH2O, respectively. Thoracic

AC C

EP

16

ACCEPTED MANUSCRIPT

electrical impedance tomography (EIT), spirometry and routine anaesthesia monitoring were

26

performed. At 90 minutes after induction of anaesthesia, PIP and PEEP were increased in steps of

27

5 cmH2O to 50 cmH2O and 30 cmH2O, respectively, and then decreased to baseline values. Each

28

step lasted 10 minutes. Data were recorded and functional EIT images were created using three

29

breaths at the end of each step. Arterial blood samples were analysed. Values for left-to-right and

30

sternal-to-dorsal centre of ventilation (COV), lung compliances and Bohr dead space were

31

calculated.

32

Results Distribution of ventilation drifted leftward and dorsally during recruitment.

33

Mean ± standard deviation (SD) values at baseline and highest airway pressures, respectively,

34

were 49.9 ± 0.7% and 48.0 ± 0.6% for left-to-right COV (p = 0.009), and 46.3 ± 2.0% and

35

54.6 ± 2.0% for sternal-to-dorsal COV (p = 0.0001). Compliance of dependent lung regions and

36

PaO2 increased, whereas compliance of non-dependent lung regions decreased during ARM and

37

then returned to baseline (p < 0.001). Bohr dead space decreased after ARM (p = 0.007).

38

Interestingly, PaO2 correlated to the compliance of the dependent lung (r2 = 0.71, p < 0.001).

39

Conclusions and clinical relevance The proportion of tidal volume distributed to dependent and

40

left lung regions increased during ARM, presumably as a result of opening atelectasis.

41

Monitoring compliance of the dependent lung with EIT may substitute PaO2 measurements

42

during ARM to identify an optimal PEEP.

43

Keywords distribution of ventilation, electrical impedance tomography, horse, lung compliance,

44

respiratory.

SC

M AN U

TE D

EP

AC C

45

RI PT

25

ACCEPTED MANUSCRIPT

Introduction

47

Anaesthetized horses often develop pronounced disturbance of gas exchange (Nyman &

48

Hedenstierna 1989). This is typically represented by a large intrapulmonary shunt fraction with

49

little or no perfusion of low ventilation/perfusion (V/Q) regions, irrespective of body positioning

50

(dorsal or lateral) and method of ventilation (spontaneous or mechanical) (Nyman &

51

Hedenstierna 1989). Nevertheless, high inspired oxygen concentrations increase shunt fraction

52

(Marntell et al. 2005). Pulmonary atelectasis seems to be the chief mechanism in the development

53

of intrapulmonary shunt in anaesthetized horses (Nyman & Hedenstierna 1989; Nyman et al.

54

1990).

55

To date, the only proven method of reversing pulmonary atelectasis and improving gas exchange

56

during anaesthesia in human subjects is to perform an alveolar recruitment manoeuvre (ARM)

57

and continue mechanical ventilation with an individually defined positive end-expiratory pressure

58

(PEEP) to prevent the re-collapse of alveoli (open lung concept) (Lachmann 1992; Tusman &

59

Bohm 2010). Such an ARM typically consists of two consecutive PEEP titrations (Tusman &

60

Bohm 2010). During the first PEEP titration the ‘optimal’ PEEP is determined. During the

61

second titration the lungs are reopened and subsequently ventilated at the previously determined

62

PEEP.

63

Although a large number of papers have been published on the subject in human medicine, only a

64

few report the effects of ARM in horses. For example, a series of sustained lung inflation

65

manoeuvres were performed in 15 horses during colic surgery using stepwise increasing airway

66

pressures lasting 10 seconds at each pressure step (Schürmann et al. 2008). This study concluded

67

that most horses needed airway pressure as high as 60–80 cmH2O in order to increase arterial

68

partial pressure of oxygen (PaO2) to > 30 mmHg (4 kPa). Another study reported that horses

69

ventilated according to a modified open-lung concept had higher PaO2 values and lower blood

AC C

EP

TE D

M AN U

SC

RI PT

46

ACCEPTED MANUSCRIPT

pressures (Bringewatt et al. 2010). One of the first studies to show improvement in a clinically

71

important outcome variable reported that horses receiving ARM during anaesthesia for colic

72

surgery stood up faster after surgery than those that did not receive ARM (Hopster et al. 2011).

73

Another study showed improved PaO2 values after ARM using volume-controlled ventilation and

74

PEEP titration (Ambrosio et al. 2013). However, neither sustained inflation manoeuvres nor

75

volume-controlled ventilation is ideal for performing an ARM; instead, a pressure-controlled

76

ventilation mode using PEEP titration to allow for determination of the optimal PEEP should be

77

applied (Tusman & Bohm 2010). Two studies satisfying these conditions in horses have been

78

published. One of these was performed in a single laterally recumbent horse using electrical

79

impedance tomography (EIT) (Moens et al. 2014) and the other was conducted in six dorsally

80

recumbent ponies without EIT monitoring (Wettstein et al. 2006).

81

A variety of methods that can be used to identify the optimal PEEP required after an ARM in

82

human subjects have been reported (Caramez et al. 2009). However, monitoring the compliance

83

of the dependent lung using a non-invasive and clinically applicable tool, such as EIT, is a

84

promising new option (Gomez-Laberge et al. 2013). Therefore, combining the advantages of

85

PEEP titration using a pressure-controlled mode of ventilation and continuous monitoring with a

86

non-invasive device such as EIT during the performance of ARM in horses is reasonable and

87

deserves investigation. The feasibility of this approach was demonstrated in the pilot study

88

conducted in a single laterally recumbent horse (Moens et al. 2014).

89

The aims of the current study were to demonstrate changes in the distribution of ventilation and

90

regional lung compliances using EIT during ARM (using PEEP titration and a pressure-

91

controlled ventilation mode) in dorsally recumbent horses.

AC C

EP

TE D

M AN U

SC

RI PT

70

ACCEPTED MANUSCRIPT

Materials and methods

93

This study was reviewed by the Ethics Committee for Animal Experiments of Lower Saxony,

94

Germany, and approved (no. 33.14-42502-04- 11/0572) according to the German Animal

95

Welfare Act (Tierschutzgesetz).

96

Animals

97

Eight Warmblood horses were enrolled in the study. Sample size calculations

98

(http://powerandsamplesize.com/Calculators/Compare-k-Means/1-Way-ANOVA-Pairwise)

99

indicated that four subjects would be sufficient to demonstrate changes of expected magnitude in

SC

RI PT

92

sternal-to-dorsal centre of ventilation (COV) assuming α = 0.05 and β = 0.2. The COV values

101

(46.3% and 52.9%) and the standard deviations (SDs) of such values (2.5%) used in the sample

102

size calculation were extracted from other studies conducted in horses (Moens et al. 2014;

103

Ambrisko et al. 2015).

104

The horses were considered systemically healthy based on physical examination and routine

105

haematological and biochemical blood work. All horses had severe and untreatable orthopaedic

106

diseases or were blind and their owners had requested euthanasia. With the permission of their

107

owners, several experiments were performed on these animals under general anaesthesia, after

108

which, while still under anaesthesia, the horses were euthanized by the administration of

109

intravenous (IV) pentobarbital 70 mg kg−1 (Euthadorm 400; CP-Pharma Handelsgesellschaft

110

mbH, Germany).

111

Anaesthesia and instrumentation

112

Animals were fasted overnight before the experiment but given access to water. Before

113

anaesthesia, a 12 gauge catheter was placed into the left jugular vein and two balloon-tipped

114

catheters were separately placed in the right jugular vein to facilitate cardiac output

115

measurements for an unrelated study (Hopster et al. 2015). At this time, an area of skin 5 cm

AC C

EP

TE D

M AN U

100

ACCEPTED MANUSCRIPT

wide was clipped circumferentially around the thorax immediately behind the olecranon for

117

placement of the EIT electrode belt. The horses were premedicated with xylazine 0.8 mg kg−1 IV

118

(Xylapan; Vetoquinol GmbH, Germany) and anaesthesia was induced with midazolam

119

0.05 mg kg−1 (Midazolam-ratiopharm; Ratiopharm GmbH, Germany) and ketamine 2.2 mg kg−1

120

(Narketan; Vetoquinol GmbH) IV. Anaesthesia was maintained with isoflurane (IsofluranCP;

121

CP-Pharma Handelsgesellschaft mbH) in oxygen. Intravenous crystalloids (lactated Ringer’s

122

solution; B. Braun Melsungen AG, Germany) and dobutamine (Dobutamin-ratiopharm;

123

Ratiopharm GmbH) were administered at rates of 10 mL kg−1 hour−1 and 0.5 µg kg−1 minute−1,

124

respectively. Following induction of anaesthesia and tracheal intubation, the horses were

125

positioned in dorsal recumbency and ventilated immediately with a pressure-limited and

126

pressure-cycled large animal ventilator (Vet-Tec model JAVC 2000; JD Medical Distributing Co.,

127

AZ, USA) equipped with a custom-made PEEP valve. The transverse facial artery was

128

cannulated with a 20 gauge catheter for invasive blood pressure monitoring and arterial blood

129

sampling. Arterial blood pressures, pulmonary artery pressure, heart rate, respiratory rate,

130

inspiratory oxygen fraction (FIO2), expiratory isoflurane concentration and thermodilution

131

cardiac output were monitored with a Cardiocap 5 monitor (Datex-Ohmeda GmbH, Germany)

132

and manually recorded. Arterial pH, PaO2 and partial pressure of carbon dioxide (PaCO2) were

133

measured immediately after anaerobic blood sampling (AVL995; AVL Medizintechnik GmbH,

134

Germany).

135

A custom-made flow partitioning device (Ambrisko et al. 2014; Schramel et al. 2014) was

136

connected between the Y-piece and the endotracheal tube adapter to facilitate measurements of

137

tidal volume (VT), airway pressures and mainstream capnography. This device was equipped with

138

four identical flow sensors, one of which was connected to a dedicated spirometry monitor

139

(NICO2 respiratory profile monitor, model 7600; Respironics California, Inc., CA, USA). Thirty-

AC C

EP

TE D

M AN U

SC

RI PT

116

ACCEPTED MANUSCRIPT

two EIT electrodes, attached to a custom-made rubber belt, were equidistantly placed around the

141

thorax over the previously clipped area. Contact resistance was minimized using electrode gel

142

(Henry Schein, Inc., NY, USA). The electrodes were connected to an EIT device (model DX

143

1800; Timpel SA [formerly Dixtal Biomédica Indústria e Comércio Ltda], SP, Brazil). Detailed

144

descriptions of this EIT system are available elsewhere (Costa et al. 2008, 2009). The EIT and

145

spirometry data were downloaded simultaneously and saved on a dedicated computer at a

146

sampling frequency of 50 Hz.

147

Experiment protocol

148

During the first 2 hours after the induction of anaesthesia (stabilization period), the horses were

149

ventilated with a peak inspiratory pressure (PIP) of 20 cmH2O and PEEP of 0 cmH2O. The

150

respiratory rate was adjusted to maintain PaCO2 at 35–45 mmHg (4.7–6.0 kPa). This was

151

followed by an experimental ARM conducted by making stepwise increases in PIP and PEEP of

152

5 cmH2O at each step until PIP and PEEP values of 50 cmH2O and 30 cmH2O, respectively, were

153

reached. Thereafter, the pressures were decreased similarly in steps of 5 cmH2O until PIP and

154

PEEP values of 20 cmH2O and 0 cmH2O, respectively, were achieved. Each step of recruitment

155

was maintained for 10 minutes to allow multiple measurements of cardiac output and other

156

variables. EIT and spirometry data were recorded during each step of recruitment. Cardiac output

157

measurements, using cold saline thermodilution and lithium dilution techniques, were regularly

158

performed throughout the procedure in an experiment for which the results have been published

159

elsewhere (Hopster et al. 2015). However, the cardiac indices (measured by the thermodilution

160

technique) of those horses in which the ARM was completed are also presented here. During the

161

entire study process, laparotomy and intestinal surgery were performed for an unrelated study.

162

The study described in this manuscript was completed at the end of the ARM.

163

Post hoc analysis of EIT and spirometry data

AC C

EP

TE D

M AN U

SC

RI PT

140

ACCEPTED MANUSCRIPT

During the creation of the EIT images, the thoracic shape of a horse was projected into a

165

rectangular image of 32 × 32 pixels while an intermediate Gaussian spatial filter and a low-pass

166

temporal filter at 1 Hz were applied (EIT Analysis Tools Beta 7.4.57; Timpel SA). In this image,

167

418 pixels were within the thoracic boundary and contained impedance signals. Each single pixel

168

signal consisted of the relative impedance change (∆Z) over time compared with a common

169

reference value of that individual pixel. The reference impedance was calculated, for each pixel,

170

as the average Z-value for the first 300 frames of each EIT recording. Three similar breaths were

171

selected from these images at the end of each recruitment step for further analysis.

172

Two types of functional EIT image were created from each selection: SD images containing the

173

pixel-wise SDs of the impedance signals, and R images containing Pearson correlation

174

coefficients (linear regression analysis of each pixel and a reference respiratory signal located in

175

the middle of a lung) (Ambrisko et al. 2015).

176

During the analysis of SD images, the only pixel SD values used were those for which R values

177

were positive (respiratory signals). The following variables were calculated from each SD image:

178

left-to-right and sternal-to-dorsal COV (Frerichs et al. 2006); global inhomogeneity (GI) index

179

(Zhao et al. 2009), and dependent and non-dependent fractions of VT. Left-to-right COV

180

expressed the distance from the left side of the thorax as a percentage of thoracic width, and

181

sternal-to-dorsal COV represented the distance from the sternum as a percentage of thoracic

182

height. In order to define dependent and non-dependent lung regions, each image was separated

183

by a horizontal line exactly in the middle. Therefore, the original functional EIT images (32 × 32

184

pixels) were divided into two smaller images (16 × 32 pixels) representing dependent and non-

185

dependent lung regions, respectively. The location of this separating line was determined after

186

examining the regional distribution of pixel compliances in all animals. The sum of SD values in

187

the dorsal (dependent) part divided by the sum of SD values in the full image resulted in the

AC C

EP

TE D

M AN U

SC

RI PT

164

ACCEPTED MANUSCRIPT

fraction of ventilation to dependent lung regions. Similarly, the sum of SD values in the sternal

189

(non-dependent) part divided by the sum of SD values in the full image provided the fraction of

190

ventilation to non-dependent lung regions. These fractions were multiplied by VT values

191

measured by the NICO monitor and divided by the pressure difference (PIP minus PEEP) to

192

result in (quasi-static) compliance of the dependent and non-dependent lung regions, respectively.

193

Similarly, total lung compliance was calculated as VT/(PIP − PEEP). The existence of no gas

194

flow in the airways for > 0.5 seconds was confirmed at the end of both inspiration and expiration;

195

therefore the PIP and PEEP values reported here reflect alveolar pressures. Bohr dead space was

196

calculated from the NICO data using a published equation (Tusman et al. 2009).

197

Global inhomogeneity indices were also calculated for the R images for positive and negative

198

pixel R values separately. In the case of negative R values, the GI indices were multiplied by − 1.

199

After the calculations were complete, the SD and R images were normalized to 1 and averaged

200

across steps of recruitment. MATLAB Version 7.7.0.471 (R2008b) (MathWorks, Inc., MA, USA)

201

was used for calculations.

202

Statistical analysis

203

Normality of the data was examined with the Shapiro–Wilk test. The data were compared across

204

steps of recruitment using repeated-measures analysis of variance (RM-ANOVA) with

205

Greenhouse–Geisser corrections. Correlations were examined using linear regression analysis. A

206

p-value of < 0.05 was accepted as indicating statistical significance. Analyses were performed

207

using IBM SPSS Statistics for Windows Version 20.0 (IBM Corp., NY, USA).

208

Results

209

Five horses completed the experiment (one stallion, one gelding, three mares). These horses had a

210

mean ± SD age of 7.7 ± 5.2 years and body weight of 558 ± 54 kg.

AC C

EP

TE D

M AN U

SC

RI PT

188

ACCEPTED MANUSCRIPT

In three horses in this study (mean ± SD age: 19.3 ± 3.2 years; mean ± SD body weight:

212

623 ± 34 kg), recruitment manoeuvres were not completed and data collected from these animals

213

were excluded from analysis. One animal, a mare aged 23 years and weighing 586 kg, died

214

suddenly during the ramp-down phase of recruitment at a PEEP of 20 cmH2O. The last values

215

recorded in this animal were cardiac output of 5 L minute−1, mean arterial blood pressure of

216

47 mmHg (6.3 kPa), heart rate of 71 beats minute−1 and PaO2 of 354 mmHg (47.2 kPa).

217

Recruitment manoeuvres were aborted in the two other animals (aged 17 years and 18 years,

218

respectively) because blood pressure decreased during an early phase of recruitment and did not

219

respond to volume therapy with 2 L of hypertonic saline IV. These three animals were older than

220

the five in which the experiment was completed; age was the only obvious factor that could be

221

associated with failure to complete the ARM in this study.

222

The averaged SD and R images are shown in Fig. 1. Sternal-to-dorsal COV gradually increased

223

(shifted dorsally) at high airway pressures and then decreased to baseline values (p < 0.001)

224

(Fig. 2). Left-to-right COV decreased at high pressures (shifted to the left) and then returned to

225

baseline (p = 0.02) (Fig. 2). Total lung compliance was highest during the ramp-down phase of

226

recruitment at PEEP of 20 cmH2O, 15 cmH2O and 10 cmH2O (p = 0.001) (Fig. 3). The

227

compliance of the dependent lung increased during recruitment and was highest during the ramp-

228

down phase at PEEP of 25 cmH2O and 20 cmH2O (p < 0.001) (Fig. 3). The compliance of the

229

non-dependent lung decreased during recruitment and was lowest at the highest airway pressures

230

(p < 0.001) (Fig. 3). Bohr dead space decreased during the ramp-down phase of recruitment

231

(p = 0.007) (Fig. 4). Almost inversely, PaO2 started to increase at the highest airway pressures

232

and reached its highest values during the ramp-down phase at PEEP of 25 cmH2O and 20 cmH2O

233

before returning to baseline (p < 0.001) (Fig. 4). Interestingly, PaO2 correlated to the compliance

234

of the dependent lung (second-degree polynomial regression, r2 = 0.71, p < 0.001).

AC C

EP

TE D

M AN U

SC

RI PT

211

ACCEPTED MANUSCRIPT

The GI indices for the SD and positive R images did not change (p = 0.136 and p = 0.082,

236

respectively; data not shown), but increased for the negative R images at the highest PEEP values

237

during recruitment (Table 1). Mean values of PaCO2, VT, respiratory rate, cardiac index and heart

238

rate are shown in Table 1.

AC C

EP

TE D

M AN U

SC

RI PT

235

ACCEPTED MANUSCRIPT

Discussion

240

Horses commonly develop atelectasis in the dependent lung regions at an early phase of

241

anaesthesia (Nyman et al. 1990). The resulting shunt fraction is larger in dorsal than in lateral

242

recumbency (Nyman & Hedenstierna 1989). Atelectasis in these studies was successfully treated

243

with deep insufflation of the lungs (Nyman et al. 1990) or selective ventilation of the dependent

244

lung regions using a PEEP of 30 cmH2O (Nyman & Hedenstierna 1989). Similar observations

245

have been noted in many species (Duggan & Kavanagh 2005). These observations were

246

compatible with the results of this study because the distribution of VT (COV) shifted towards the

247

dependent (dorsal) lung regions during ARM (Figs 1 and 2), possibly as a result of opening

248

atelectasis at those sites. Although this is not new information, the current study represents one of

249

the first demonstrations of this phenomenon in horses using a continuous, non-invasive

250

monitoring method, such as EIT.

251

During ARM in this study, the compliance of the dependent lung increased and that of the non-

252

dependent lung decreased. After ARM the compliance of the dependent lung returned to baseline,

253

whereas that of the non-dependent lung remained high. A similar phenomenon has been

254

described in a porcine model of acute lung injury and was confirmed by histological findings that

255

dependent regions were mainly atelectatic and non-dependent regions were mainly over-

256

distended (Gomez-Laberge et al. 2013). Therefore, changes in the compliance of the dependent

257

lung may represent recruitment and de-recruitment of atelectasis, and increasing non-dependent

258

lung compliance values may represent relief from over-distension.

259

The total lung compliance, data for which would represent the only compliance data available

260

without EIT, is the result of the interaction between dependent and non-dependent lung

261

compliances and on its own has little physiological meaning (Gomez-Laberge et al. 2013). A

262

previous study conducted in anaesthetized ponies undergoing a similar ARM reported that during

AC C

EP

TE D

M AN U

SC

RI PT

239

ACCEPTED MANUSCRIPT

the ramp-down phase of PEEP titration, PaO2 decreased earlier than total lung compliance

264

(Wettstein et al. 2006); this observation was confirmed in the present study. Interestingly, in the

265

current study, PaO2 increased and decreased in a pattern similar to that of the compliance of the

266

dependent lung and these two variables significantly correlated. This is a physiologically

267

meaningful observation because the changing compliance of the dependent lung relates mostly to

268

lung collapse, which is the main reason for shunt formation and low PaO2 in healthy

269

anaesthetized horses (Nyman & Hedenstierna 1989). Bohr dead space decreased at the ramp-

270

down phase of the ARM, indicating improved gas exchange possibly as the result of the relief of

271

the non-dependent lung from over-distension. The observation that the compliance of the non-

272

dependent lung and Bohr dead space showed similar (inverse) tendencies after ARM in this study

273

confirms this assumption.

274

The possible deleterious effects of mechanical ventilation on injured lungs are well recognised

275

and chiefly associated with cyclic recruitment and over-distension of alveoli (Lachmann 1992;

276

Tusman et al. 2012). Prevention of both phenomena could be attempted by the application of the

277

‘open lung concept’ during mechanical ventilation (Lachmann 1992; Tusman & Bohm 2010). In

278

order to keep the lungs open, an individually defined optimal PEEP must be determined during

279

the ramp-down phase of PEEP titration. Of the numerous published methods of identifying the

280

optimal PEEP (Tusman & Bohm 2010), the two most commonly used are based on monitoring of

281

PaO2 or total lung compliance (Tusman et al. 2012). Airway pressures at which these values

282

decrease correspond to alveolar closing pressures and the optimal PEEPs should be set 1–

283

2 cmH2O higher than these values (Tusman et al. 2012). However, the use of EIT makes it

284

possible to specifically monitor the compliance of the more physiologically meaningful

285

dependent lung, rather than the total lung. In this study, using total lung compliance for this

286

purpose would result in an optimal PEEP of 10 cmH2O, although PaO2 decreased and atelectasis

AC C

EP

TE D

M AN U

SC

RI PT

263

ACCEPTED MANUSCRIPT

formation is expected at this PEEP level. When the compliance of the dependent lung or PaO2

288

was used to determine optimal PEEP, the result was 20 cmH2O with both methods. Therefore, the

289

current authors conclude that the compliance of the dependent lung and PaO2 are, at least in the

290

present study, equally suitable for identifying optimal PEEP, but that the compliance of the total

291

lung is not suitable.

292

Before the ARM, the right lung received a larger proportion of VT than the left, which agrees

293

with the findings of a previous study conducted in standing horses (Ambrisko et al. 2015). This

294

may be explained by the fact that the location of the heart extends further to the left than the right

295

of the thorax and therefore the right lung is larger. VT was distributed more equally (i.e. VT

296

shifted towards the left) during recruitment at high airway pressures. The present authors

297

hypothesize that most of the lung regions that were ventilated at baseline pressures (on the right

298

more than on the left side) became over-distended and received only a lower proportion of VT at

299

higher airway pressures. Subsequently, those dependent (dorsal) lung regions that received most

300

of the VT at higher pressures may have been similarly sized and therefore may have received a

301

similar proportion of VT.

302

The existence of inverse respiratory signals (negative R values) in this study was also observed in

303

a previous study in standing horses, the authors of which hypothesized that these signals may be

304

caused by gas pockets in the large colon (Ambrisko et al. 2015). In the current study, GI indices

305

increased for the negative R images during ARM, indicating that the quality of inverse

306

respiratory signals worsened. This can be observed on the R images (Fig. 1) in which the blue

307

area near the sternum becomes lighter (i.e. negative R values increase) at high airway pressures.

308

It is expected that the diaphragm and hence abdominal content shift caudally during high airway

309

pressures and, consequently, that abdominal gas compartments become more distant from the

310

EIT electrode belt; this may explain the worsening inverse correlation. The relationship between

AC C

EP

TE D

M AN U

SC

RI PT

287

ACCEPTED MANUSCRIPT

the GI indices of the negative R images and the position of the diaphragm requires further

312

examination.

313

With regard to the animals in which the study could not be completed, it should be noted that the

314

experimental recruitment manoeuvre performed in this study was conducted over a period of

315

approximately 2 hours. Clinical recruitment manoeuvres are therapeutic interventions that aim to

316

open atelectasis and increase oxygenation but pose only minimal cardiovascular risk to the

317

patient (Tusman & Bohm 2010). For this goal to be achieved, the duration of recruitment must be

318

restricted to the minimum. Further research is required to identify other risk factors and

319

predictive monitoring parameters (e.g. lowest acceptable values for blood pressure or cardiac

320

output) to aid in the establishment of safer methods of performing recruitment manoeuvres in

321

horses.

322

In conclusion, this study showed that VT was distributed more towards dependent lung regions

323

during recruitment and that dependent lung compliance and PaO2 were similarly useful in

324

determining optimal PEEP.

325

Acknowledgements

326

This work was supported by the University of Veterinary Medicine, Hannover, Germany and the

327

University of Veterinary Medicine, Vienna, Austria.

328

Authors’ contributions

329

TDA contributed to the study design, and the collection, analysis and interpretation of data, and

330

wrote the first draft of the manuscript. JS and KH contributed to the study design, the collection

331

and interpretation of data, and the critical revision of the paper. SK and YM contributed to the

332

study design, interpretation of data, and the critical revision of the paper. All authors approved

333

the final manuscript for publication.

AC C

EP

TE D

M AN U

SC

RI PT

311

ACCEPTED MANUSCRIPT

References

335

Ambrisko TD, Lammer V, Schramel JP et al. (2014) In vitro and in vivo evaluation of a new

336

large animal spirometry device using mainstream CO2 flow sensors. Equine Vet J 46, 507–511.

337

Ambrisko TD, Schramel JP, Adler A et al. (2015) Assessment of distribution of ventilation by

338

electrical impedance tomography in standing horses. Physiol Meas 37, 175–186.

339

Ambrosio AM, Ida KK, Souto MT et al. (2013) Effects of positive end-expiratory pressure

340

titration on gas exchange, respiratory mechanics and hemodynamics in anesthetized horses. Vet

341

Anaesth Analg 40, 564–572.

342

Bringewatt T, Hopster K, Kastner SB et al. (2010) Influence of modified open lung concept

343

ventilation on the cardiovascular and pulmonary function of horses during total intravenous

344

anaesthesia. Vet Rec 167, 1002–1006.

345

Caramez MP, Kacmarek RM, Helmy M et al. (2009) A comparison of methods to identify open-

346

lung PEEP. Intensive Care Med 35, 740–747.

347

Costa EL, Chaves CN, Gomes S et al. (2008) Real-time detection of pneumothorax using

348

electrical impedance tomography. Crit Care Med 36, 1230–1238.

349

Costa EL, Borges JB, Melo A et al. (2009) Bedside estimation of recruitable alveolar collapse

350

and hyperdistension by electrical impedance tomography. Intensive Care Med 35, 1132–1137.

351

Duggan M, Kavanagh BP (2005) Pulmonary atelectasis: a pathogenic perioperative entity.

352

Anesthesiology 102, 838–854.

353

Frerichs I, Dargaville PA, van Genderingen H et al. (2006) Lung volume recruitment after

354

surfactant administration modifies spatial distribution of ventilation. Am J Respir Crit Care Med

355

174, 772–779.

AC C

EP

TE D

M AN U

SC

RI PT

334

ACCEPTED MANUSCRIPT

Gomez-Laberge C, Rettig JS, Smallwood CD et al. (2013) Interaction of dependent and non-

357

dependent regions of the acutely injured lung during a stepwise recruitment manoeuvre. Physiol

358

Meas 34, 163–177.

359

Hopster K, Kastner SB, Rohn K et al. (2011) Intermittent positive pressure ventilation with

360

constant positive end-expiratory pressure and alveolar recruitment manoeuvre during inhalation

361

anaesthesia in horses undergoing surgery for colic, and its influence on the early recovery period.

362

Vet Anaesth Analg 38, 169–177.

363

Hopster K, Ambrisko TD, Stahl J et al. (2015) Influence of xylazine on the function of the

364

LiDCO sensor in isoflurane anaesthetized horses. Vet Anaesth Analg 42, 142–149.

365

Lachmann B (1992) Open up the lung and keep the lung open. Intensive Care Med 18, 319–321.

366

Marntell S, Nyman G, Hedenstierna G (2005) High inspired oxygen concentrations increase

367

intrapulmonary shunt in anaesthetized horses. Vet Anaesth Analg 32, 338–347.

368

Moens Y, Schramel JP, Tusman G et al. (2014) Variety of non-invasive continuous monitoring

369

methodologies including electrical impedance tomography provides novel insights into the

370

physiology of lung collapse and recruitment – case report of an anaesthetized horse. Vet Anaesth

371

Analg 41, 196–204.

372

Nyman G, Hedenstierna G (1989) Ventilation–perfusion relationships in the anaesthetised horse.

373

Equine Vet J 21, 274–281.

374

Nyman G, Funkquist B, Kvart C et al. (1990) Atelectasis causes gas exchange impairment in the

375

anaesthetised horse. Equine Vet J 22, 317–324.

376

Schramel JP, Wimmer K, Ambrisko TD et al. (2014) A novel flow partition device for

377

spirometry during large animal anaesthesia. Vet Anaesth Analg 41, 191–195.

AC C

EP

TE D

M AN U

SC

RI PT

356

ACCEPTED MANUSCRIPT

Schürmann P, Hopster K, Rohn K et al. (2008) Optimierung des pulmonalen Gasaustausches

379

während der Pferdenarkose durch Beatmung nach dem ‘Open Lung Concept’. Pferdeheilkunde

380

Fundus 24, 236–242.

381

Tusman G, Bohm SH (2010) Prevention and reversal of lung collapse during the intra-operative

382

period. Best Pract Res Clin Anaesthesiol 24, 183–197.

383

Tusman G, Scandurra A, Bohm SH et al. (2009) Model fitting of volumetric capnograms

384

improves calculations of airway dead space and slope of phase III. J Clin Monit Comput 23, 197–

385

206.

386

Tusman G, Bohm SH, Warner DO et al. (2012) Atelectasis and perioperative pulmonary

387

complications in high-risk patients. Curr Opin Anaesthesiol 25, 1–10.

388

Wettstein D, Moens Y, Jaeggin-Schmucker N et al. (2006) Effects of an alveolar recruitment

389

maneuver on cardiovascular and respiratory parameters during total intravenous anesthesia in

390

ponies. Am J Vet Res 67, 152–159.

391

Zhao Z, Moller K, Steinmann D et al. (2009) Evaluation of an electrical impedance tomography-

392

based Global Inhomogeneity Index for pulmonary ventilation distribution. Intensive Care Med 35,

393

1900–1906.

SC

M AN U

TE D

EP AC C

394

RI PT

378

ACCEPTED MANUSCRIPT

Figure 1 Averaged functional electrical impedance tomography (EIT) images in five

396

anaesthetized horses during an experimental alveolar recruitment manoeuvre. Numbers over the

397

images indicate the steps of recruitment manoeuvre (peak inspiratory pressure/positive end-

398

expiratory pressure [PIP/PEEP] in cmH2O). (a) Pixel-wise standard deviation (SD) of impedance

399

changes. (b) Pixel-wise regression coefficient (Pearson’s r) for each pixel that significantly

400

correlated with the reference respiratory signal. The top and right sides of each image indicate the

401

sternal and left sides of the horse contour.

SC

RI PT

395

402

Figure 2 Mean ± standard deviation centre of ventilation for left-to-right (open circles) and

404

sternal-to-dorsal (closed circles) directions in five anaesthetized horses during an experimental

405

alveolar recruitment manoeuvre (repeated-measures analysis of variance [RM-ANOVA], p = 0.02

406

and p < 0.001, respectively). PIP, peak inspiratory pressure; PEEP, positive end-expiratory

407

pressure.

TE D

408

M AN U

403

Figure 3 Mean ± standard deviation quasi-static compliance of the total (closed triangles),

410

dependent (closed circles) and non-dependent (open circles) lung regions in five anaesthetized

411

horses during an experimental alveolar recruitment manoeuvre (repeated-measures analysis of

412

variance [RM-ANOVA], p ≤ 0.001 for each variable). PIP, peak inspiratory pressure; PEEP,

413

positive end-expiratory pressure.

AC C

414

EP

409

415

Figure 4 Mean ± standard deviation Bohr dead space (open circles) and arterial partial pressure

416

of oxygen (PaO2) (closed circles) in five anaesthetized horses during an experimental alveolar

417

recruitment manoeuvre (repeated-measures analysis of variance [RM-ANOVA], p = 0.007 and

ACCEPTED MANUSCRIPT

418

p < 0.001, respectively). PIP, peak inspiratory pressure; PEEP, positive end-expiratory pressure;

419

Vd, dead space volume; VT, tidal volume.

AC C

EP

TE D

M AN U

SC

RI PT

420

ACCEPTED MANUSCRIPT

421

Table 1 Mean ± standard deviation data collected from five anaesthetized horses during an

422

experimental alveolar recruitment manoeuvre GI for R−

(cmH2O)

PaCO2

PaCO2

VT

fR

(mmHg)

(kPa)

(mL kg−1)

(breaths min−1) (mL min−1 kg−1) (beat

0.23 ± 0.06

49 ± 6

6.5 ± 0.8

10 ± 2

25/5

0.22 ± 0.04

48 ± 11

6.4 ± 1.5

10 ± 2

30/10

0.21 ± 0.07

51 ± 9

6.8 ± 1.2

10 ± 2

35/15

0.24 ± 0.08

52 ± 8

6.9 ± 1.1

11 ± 3

40/20

0.29 ± 0.04

51 ± 8

6.8 ± 1.1

45/25

0.41 ± 0.07

43 ± 7

50/30

0.40 ± 0.05

42 ± 9

45/25

0.44 ± 0.12

39 ± 6

40/20

0.33 ± 0.11

35 ± 5

35/15

0.34 ± 0.09

30/10

0.26 ± 0.06

25/5

0.26 ± 0.07

20/0

0.31 ± 0.08

p-value

0.007

Heart

12 ± 1

69 ± 18

43 ±

12 ± 2

66 ± 22

46 ±

12 ± 1

69 ± 21

48 ±

12 ± 1

59 ± 20

47 ±

12 ± 2

11 ± 1

50 ± 20

48 ±

5.7 ± 0.9

13 ± 3

10 ± 1

42 ± 17

53 ±

5.6 ± 1.2

13 ± 2

10 ± 1

34 ± 13

54 ±

5.2 ± 0.8

15 ± 2

10 ± 1

27 ± 15

55 ±

4.7 ± 0.7

17 ± 2

9±1

29 ± 13

53 ±

TE D

M AN U

SC

20/0

Cardiac index

RI PT

PIP/PEEP

4.4 ± 0.7

17 ± 2

8±1

34 ± 16

48 ±

34 ± 6

4.5 ± 0.8

17 ± 3

8±1

41 ± 13

45 ±

36 ± 5

4.8 ± 0.7

15 ± 2

8±1

49 ± 14

43 ±

5.3 ± 0.8

13 ± 2

8±1

65 ± 15

41 ±

0.006

0.001

< 0.001

0.001

0.160

EP

33 ± 5

AC C

40 ± 6

423

PIP, peak inspiratory pressure; PEEP, positive end-expiratory pressure; GI for R−, global

424

inhomogeneity index for inverse respiratory signals; PaCO2, arterial partial pressure of carbon

425

dioxide; VT, tidal volume; fR, respiratory rate; p-value, results of repeated-measures analysis of

426

variance (RM-ANOVA) with Greenhouse–Geisser corrections across PIP/PEEP.

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT