- Email: [email protected]
Relationship Between Left Ventricle Position and Haemodynamic Parameters During Cardiopulmonary Resuscitation in a Pig Model
HLC 2481 1–9
Heart, Lung and Circulation (2017) xx, 1–9 1443-9506/04/$36.00 http://dx.doi.org/10.1016/j.hlc.2017.08.020
1 2 3
Q1
4 5
6 7 8 9 10 11 12 13 14
Q2
ORIGINAL ARTICLE
Relationship Between Left Ventricle Position and Haemodynamic Parameters During Cardiopulmonary Resuscitation in a Pig Model [TD$FIRSNAME]Yong Hun[TD$FIRSNAME.] [TD$SURNAME]Jung[,TD$SURNAME.] MD a[7_TD$IF], [TD$FIRSNAME]Kyung Woon[TD$FIRSNAME.] [TD$SURNAME]Jeung[TD$SURNAME.], MD, PhD a*, [TD$FIRSNAME]Dong Hun[TD$FIRSNAME.] [TD$SURNAME]Lee[TD$SURNAME.], MD a, [TD$FIRSNAME]Young Won[TD$FIRSNAME.] [TD$SURNAME]Jeong[TD$SURNAME.], EMTP a, [TD$FIRSNAME]Sung Min[TD$FIRSNAME.] [TD$SURNAME]Lee[TD$SURNAME.], MD a, [TD$FIRSNAME] Byung Kook[TD$FIRSNAME.] [TD$SURNAME]Lee[TD$SURNAME.], MD, PhD a, [TD$FIRSNAME]In Seok[TD$FIRSNAME.] [TD$SURNAME]Jeong[TD$SURNAME.], MD, PhD b, [TD$FIRSNAME]Sang-Kwon[TD$FIRSNAME.] [TD$SURNAME]Lee[TD$SURNAME.], DVM c, [TD$FIRSNAME]Jihye[TD$FIRSNAME.] [TD$SURNAME]Choi[TD$SURNAME.], DVM PhD c a
Department of Emergency Medicine, Chonnam National University Hospital, Gwangju, Republic of Korea Department of Thoracic and Cardiovascular Surgery, Chonnam National University Hospital, Gwangju, Republic of Korea c Veterinary Medical Imaging, College of Veterinary Medicine, Chonnam National University, Gwangju, Republic of Korea b
Received 14 June 2017; received in revised form 14 August 2017; accepted 22 August 2017; online published-ahead-of-print xxx
Background
From the viewpoint of cardiac pump theory, the area of the left ventricle (LV) subjected to compression increases as the LV lies closer to the sternum, possibly resulting in higher blood flow in patients with LV closer to the sternum. However, no study has evaluated LV position during cardiac arrest or its relationship with haemodynamic parameters during cardiopulmonary resuscitation (CPR). The objectives of this study were to determine whether the position of the LV relative to the anterior-posterior axis representing the direction of chest compression shifts during cardiac arrest and to examine the relationship between LV position and haemodynamic parameters during CPR.
Methods
Subcostal view echocardiograms were obtained from 15 pigs with the transducer parallel to the long axis of the sternum before inducing ventricular fibrillation (VF) and during cardiac arrest. Computed tomography was performed in three pigs to objectively observe LV position during cardiac arrest. Left ventricular position parameters including the shortest distance between the anterior-posterior axis and the mid-point of the LV chamber (DAP-MidLV), the shortest distance between the anterior-posterior axis and the LV apex (DAP-Apex), and the area fraction of the LV located on the right side of the anterior-posterior axis (LVARight/ LVATotal) were measured.
Results
DAP-MidLV, DAP-Apex, and LVARight/LVATotal decreased progressively during untreated VF and basic life support (BLS), and then increased during advanced cardiovascular life support (ACLS). A repeated measures analysis of variance revealed significant time effects for these parameters. During BLS, the end-tidal carbon dioxide and systolic right atrial pressure were significantly correlated with the LV position parameters. During ACLS, systolic arterial pressure and systolic right atrial pressure were significantly correlated with DAP-MidLV and DAP-Apex.
Conclusions
Left ventricular position changed significantly during cardiac arrest compared to the pre-arrest baseline. LV position during CPR had significant correlations with haemodynamic parameters.
Keywords
Heart arrest Cardiopulmonary resuscitation Heart ventricles Haemodynamics
Q3
15 *Corresponding author at: Department of Emergency Medicine, Chonnam National University Hospital, 42 Jebong-ro, Donggu, Gwangju, Republic of Korea. Fax: +82 62 228 7417., Email: [email protected] © 2017 Australian and New Zealand Society of Cardiac and Thoracic Surgeons (ANZSCTS) and the Cardiac Society of Australia and New Zealand (CSANZ). Published by Elsevier B.V. All rights reserved.
Please cite this article in press as: Jung YH, et al. Relationship Between Left Ventricle Position and Haemodynamic Parameters During Cardiopulmonary Resuscitation in a Pig Model. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j. hlc.2017.08.020
HLC 2481 1–9
2
Y.H. Jung et al.
16 17
Introduction
18
Cardiopulmonary resuscitation (CPR) has been the cornerstone of treatment for cardiac arrest since 1960 [1]. Over the past 50 years, considerable efforts have been made to expand the understanding of the physiology behind CPR [2–5]. However, many questions remain unexplored or incompletely understood. Blood flow during CPR is thought to be provided by direct cardiac compression (cardiac pump) and/or increased intrathoracic pressure (thoracic pump) [2–5]. According to the first theory, blood flow is generated as the heart is squeezed between the sternum and the spine at the level of the ventricles. From this point-of-view, the area of the left ventricle (LV) subjected to compression increases as the LV lies closer to the sternum, possibly resulting in higher blood flow in patients with LV closer to the sternum [3]. A number of studies have investigated the spatial relationship between the sternum and anatomical structures using magnetic resonance imaging or computed tomography (CT) [6–11]. However, most of the studies have been focussed on the vertical relationship between the sternum and the inner viscera. Until now, few studies have explored the spatial relationship in the transverse direction [11,12], and, to our knowledge, no study has evaluated the relationship between LV position and haemodynamic parameters during CPR. Furthermore, all of these studies have used images obtained in patients not in cardiac arrest [6–11]. However, the position of the cardiac chamber may shift during cardiac arrest. In this study, we investigated the spatial relationship of the LV in a transverse direction, in relation to the anterior-posterior axis, in a swine model of cardiac arrest to determine whether LV position changes during cardiac arrest. We also examined the relationship between LV position and haemodynamic parameters during CPR. We hypothesised that LV position would change during cardiac arrest compared to the pre-arrest baseline and that LV position would be related to haemodynamic parameters during CPR.
19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53
Q4
54
Material and Methods
55
65
A total of 18 Yorkshire/Landrace-cross pigs weighing 23.9 2.2 kg were used for this study. This study consisted of two sub-studies: an echocardiography study to examine the relationship between LV position and haemodynamic parameters during CPR (n = 15) and a CT study for the more objective observation of LV position during cardiac arrest (n = 3). The Animal Care and Use Committee of Chonnam National University approved the protocol for this study (CNU IACUC-H-2016-22). Animal care and experiments were in accord with the National Institutes of Health Guide for the Care and Use of Laboratory Animals.
66
Animal Preparation
67
After premedication (ketamine, 20 mg/kg; xylazine, 2.2 mg/kg; atropine, 0.04 mg/kg, intramuscular), the
56 57 58 59 60 61 62 63 64
animals were placed supine in a u-shaped trough with the limbs secured to prevent lateral displacement of the chest during CPR. After tracheal intubation, anaesthesia was provided with 70%:30% N2O:O2 and 0.5–2% sevoflurane, which was titrated to prevent signs of pain (reactive wide pupils, tachycardia, and hypertension). A doublelumen catheter was inserted via the right femoral artery for monitoring of aortic pressure. The right external jugular vein was cannulated with an 8-French introducer sheath to monitor right atrial (RA) pressure, and to insert a right ventricle (RV) pacing catheter. An end-tidal carbon dioxide (ETCO2) sample line was connected to the ventilator circuit. During the preparation phase, saline was administered to maintain an RA pressure of 10–12 mmHg.
68
Echocardiography Study
82
After baseline measurements, ventricular fibrillation (VF) was induced by passing 60 Hz of AC current at 30 mA through the RV pacing catheter. The animals were then disconnected from ventilatory support. After 14 mins of untreated VF ([1_TD$IF]Figure 1), basic life support (BLS) was started, using cycles of 30 chest compressions followed by two ventilations with ambient air. Closed-chest compressions were administered by two experienced investigators (DHL and YWJ) who were unaware of the study’s aims, at a rate of 100/ min and a compression depth of 25% of the anterior-posterior diameter of the chest wall. The depth of the chest compressions was monitored by another investigator (KWJ). After 8 mins of BLS, advanced cardiovascular life support (ACLS) based on recent resuscitation guidelines was started [13]. Asynchronous positive-pressure ventilations with high flow O2 (14 l/min) were provided at a rate of 8/min, using a volume-marked bag devised by Cho et al. [14]. At the start of ACLS, all animals received 0.6 U/kg of vasopressin intravenously. After 4 mins of ACLS, 0.02 mg/kg of epinephrine was administered every 3 mins if required. If sustained restoration of spontaneous circulation (ROSC) was not achieved within 12 mins of ACLS [15], resuscitation efforts were discontinued. Animals that achieved ROSC underwent a 4-h period of intensive care for another study not reported here. Subcostal view echocardiograms (UST-9130, Hitachi Aloka Medical Ltd., Tokyo, Japan) were obtained by a single researcher at the pre-arrest baseline (5, 10, and 15 mins before VF induction), during untreated VF (1, 7, and 13 mins after VF induction), and every 2 mins during CPR until 6 mins after the start of ACLS. To obtain consistent orientation of the transducer in relation to the sternum, the following protocol was employed. A straight line was drawn through the middle of the transducer and a vertical line was drawn along the sternum down to the xiphoid process ([1_TD$IF]Figure 2). The transducer was placed just below the xiphoid process (with the line on the transducer aligned with the line drawn on the sternum) and gently pushed inward. The transducer was angled upward until the maximum obtainable chamber size of the LV was visualised. The subcostal view was then maintained and stored for 5 secs. The straight line on the
83
Please cite this article in press as: Jung YH, et al. Relationship Between Left Ventricle Position and Haemodynamic Parameters During Cardiopulmonary Resuscitation in a Pig Model. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j. hlc.2017.08.020
69 70 71 72 73 74 75 76 77 78 79 80 81
84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123
HLC 2481 1–9 Relationship Between Left Ventricle Position and Haemodynamic Parameters
3
[1_TD$IF]Figure 1 Experimental timelines of the study using echocardiography (A) and the study using CT (B). The lightning marks indicate the onset of the 10-sec pause in chest compressions for rhythm analysis and a 150-J shock if indicated.
124 125 126 127 128
transducer was maintained in a parallel orientation to the line drawn along the sternum throughout the procedure. The echocardiograms were obtained at end-expiratory phase, and those during resuscitation were obtained during a 10sec pause in chest compression for rhythm analysis.
CT Study The VF induction and resuscitation procedures were identical to those described above, except that the durations of the noflow period, BLS, and ACLS were shorter. In addition, no defibrillation was attempted ([1_TD$IF]Figure 1). Each animal
Figure 2 Subcostal view echocardiography. (A) Transducer placement. (B) Measurements of left ventricle position parameters. Please cite this article in press as: Jung YH, et al. Relationship Between Left Ventricle Position and Haemodynamic Parameters During Cardiopulmonary Resuscitation in a Pig Model. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j. hlc.2017.08.020
129 130 131 132
HLC 2481 1–9
4
133
Y.H. Jung et al.
144
underwent CT scanning (SOMATOM Emotion 16, Siemens, Forchheim, Germany) four times: 2 mins before VF induction, 7 mins after VF induction, immediately after the BLS, and immediately after the ACLS. We used the following settings: slice thickness, 1 mm; pitch, 0.8; rotation duration, 600 ms; tube voltage, 120 kV, and tube current, 120 mA. All scans were performed after injecting a dose of 30 mL Iohexol (Ominpaque 300, GE healthcare, Oslo, Norway) at a flow rate of 3 ml/sec. Contrast material injection was started 15 secs before each CT scan. In order to exclude the effects of breathing on LV position, vecuronium bromide was infused at a rate of 100 mg/min, and no ventilation was provided during the scan.
145
Measurements
134 135 136 137 138 139 140 141 142 143
146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163
Aortic pressure and RA pressure were continuously monitored (CS/3 CCM, Datex-Ohmeda, Helsinki, Finland) and transferred to a personal computer using S/5 Collect software (Datex-Ohmeda, Helsinki, Finland). Systolic arterial pressure was defined as the peak aortic pressure during the chest compression phase. Coronary perfusion pressure (CPP) was calculated by subtracting RA end-diastolic pressure from simultaneous aortic end-diastolic pressure. Systolic arterial pressure, systolic RA pressure, and CPP were sampled at 2-min intervals by averaging pressures from five consecutive compressions immediately before the 10-sec pause in chest compression for rhythm analysis. ETCO2 values were determined every 2 mins by averaging ETCO2 values for the preceding 30-sec interval. A single observer blinded to the haemodynamic data analysed the echocardiographic images. A line (LineMitral) was drawn between the points of mural attachment of the mitral valve leaflets ([1_TD$IF]Figure 2), and a straight line representing the LV axis was drawn between the mid-point of the LineMitral and the LV apex.
PointMidLV, which represented the centre of the LV chamber, was defined as the mid-point of the LV axis. The anteriorposterior axis was defined as a line passing through the middle of the image. With these definitions, the following LV position parameters were measured: DAP-MidLV (shortest distance between the anterior-posterior axis and PointMidLV) and DAP-Apex (shortest distance between the anterior-posterior axis and the LV apex). In addition, LV area (LVATotal) was measured and the area fraction of the LV located on the right side of the anterior-posterior axis (LVARight/LVATotal) was calculated by dividing the LV area located on the right side of the anterior-posterior axis (LVARight) by LVATotal. The average LV position parameter values were obtained during each period and were used for analysis. The CT images were analysed using a picture archiving and communication system (MAROSIS Maroview, Marotech Inc., Seoul Korea). An axial image at the level with the maximal LV area was chosen among the images obtained at the prearrest baseline. Axial images at the same level were then selected from the images obtained from each scan and LV position parameters were measured using the definitions described above.
164
Statistical Analysis
186
Sample size calculation was based on DAP-MidLV data from a pilot study [9_TD$IF]( 5.3 3.0 mm at pre-arrest baseline vs. 10.1 4.4 mm at BLS), yielding a minimum requirement of 10 animals with a set at 0.05 and a power of 0.90. Considering the result of the sample size calculation for the other study conducted simultaneously, 15 animals were used for the echocardiography study. We used three animals for the CT study due to our limited resources. Normally distributed variables were presented as
Table 1 Baseline characteristics. Total (N = 15)
ROSC (N = 8)
No ROSC (N = 7)
p Value
Systolic arterial pressure (mmHg)
117.4 (11.7)
117.9 (13.1)
116.9 (10.8)
0.873
Diastolic arterial pressure (mmHg)
79.7 (10.2)
81.0 (11.7)
78.1 (8.8)
0.606
Mean arterial pressure (mmHg)
94.8 (11.4)
96.3 (12.9)
93.1 (10.1)
0.616
Systolic RA pressure (mmHg)
14.5 (1.5)
13.9 (1.2)
15.1 (1.6)
0.105
Diastolic RA pressure (mmHg)
11.2 (1.5)
11.0 (10.0–11.3)
12.0 (11.5–12.5)
0.193
Mean RA pressure (mmHg)
12.0 (12.0–14.0)
12.0 (12.0–12.5)
14.0 (12.5–14.0)
0.240
Heart rate (/min)
98 (85–107)
102 (19)
100 (26)
0.870
ETCO2 (mmHg) pH
39.4 (2.4) 7.483 (0.046)
39.6 (2.8) 7.490 (0.047)
39.1 (2.0) 7.476 (0.046)
0.711 0.576
PaCO2 (mmHg)
39.5 (2.7)
39.8 (2.8)
39.1 (2.7)
0.635
PaO2 (mmHg)
142.5 (23.4)
146.8 (30.1)
137.6 (13.0)
0.468
HCO3 (mmol/l)
29.1 (2.1)
29.8 (2.2)
28.3 (1.7)
0.173
SaO2 (%)
99.5 (98.9–99.8)
99.7 (98.9–100)
99.5 (99.1–99.5)
0.483
Lactate (mmol/l)
0.8 (0.7–1.0)
0.8 (0.2)
1.0 (0.5)
0.348
Data are presented as mean (standard deviation) or as median (interquartile ranges). Abbreviations: RA, right atrial; ETCO2, end-tidal carbon dioxide; ROSC, Return of spontaneous circulation, ETCO2, end-tidal carbon dioxide.
Please cite this article in press as: Jung YH, et al. Relationship Between Left Ventricle Position and Haemodynamic Parameters During Cardiopulmonary Resuscitation in a Pig Model. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j. hlc.2017.08.020
165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185
187 188 189 190 191 192 193 194
HLC 2481 1–9 Relationship Between Left Ventricle Position and Haemodynamic Parameters
207
mean standard deviation (SD) and an independent t-test was performed, whereas non-normally distributed variables were presented as medians with interquartile ranges and a Mann-Whitney U test was conducted. Correlations between variables were determined using Pearson or Spearman correlation tests where appropriate. Repeated measures analysis of variance (ANOVA) was used for serial measurements. Pair-wise comparison with Bonferroni adjustment was performed for post-hoc analysis. To evaluate intraobserver variability in the echocardiographic measurements, intraclass correlation coefficient was calculated using the data obtained at the pre-arrest baseline. A p value of [10_TD$IF]<0.05 was considered significant.
208
Results
209
Echocardiography Study
210
240
Baseline characteristics are shown in Table 1. Eight (53.3%) of the 15 animals studied achieved sustained ROSC. An overview of LV position parameters in the echocardiography study is shown in Supplementary material 1. The intraclass correlation coefficient for the echocardiographic measurements was 0.922 (95% confidence interval, 0.876–0.953). DAP-MidLV, DAP-Apex, and LVARight/LVATotal decreased progressively during untreated VF and BLS, and then increased during ACLS ([1_TD$IF]Figure 3). The repeated measures ANOVA revealed significant time effects for all of these parameters (p = 0.001 for DAP-MidLV, <0.001 for DAP-Apex, and <0.001 for LVARight/LVATotal). The repeated measures ANOVA for comparisons of parameters between animals that achieved Return of spontaneous circulation (ROSC) and those that did not showed no significant inter-group differences for any of the LV position parameters (p = 0.287 for DAP-MidLV, 0.338 for DAP-Apex, and 0.362 for LVARight/LVATotal). Correlations between the LV position parameters and haemodynamic parameters during CPR are summarised in [1_TD$IF]Figures 4 and 5. During BLS, ETCO2 and systolic RA pressure had significant correlations with all LV position parameters. Systolic arterial pressure and CPP were not significantly correlated with the LV position parameters. During ACLS, the correlations between the systolic RA pressure and LV position parameters remained significant, but the correlations between ETCO2 and LV position parameters were no longer significant. Systolic arterial pressure during ACLS was significantly correlated with DAP-MidLV and DAP-Apex. Systolic arterial pressure during ACLS increased as the LVARight/LVATotal increased, although the correlations did not reach statistical significance (p = 0.071).
241
CT Study
242
Axial images at the level with the maximal LV area at the prearrest baseline are shown in [1_TD$IF]Figure 6. The LV position parameters in the CT study are displayed in Table 2. Consistent with the findings of the echocardiography study, DAP-MidLV, DAP-Apex, and LVARight/LVATotal decreased progressively during cardiac arrest in all pigs.
195 196 197 198 199 200 201 202 203 204 205 206
211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239
243 244 245 246 247
Q5
5
Figure 3 Changes in DAP-MidLV (A), DAP-Apex (B), and LVARight/LVATotal (C) during the pre-arrest baseline, untreated VF, BLS, and ACLS. A repeated measures analysis of variance revealed significant time effects for all of these parameters (p = 0.001 for DAP-MidLV, <0.001 for DAP-Apex, and <0.001 for LVARight/LVATotal). The average of the values obtained during each period was used for analysis. Positive DAP-MidLV and DAP-Apex values indicate right positions and negative values indicate left positions. *p < 0.05 versus baseline value. Abbreviations;: VF = ventricular fraction; BLS = basic life support; ACLS = advanced cardiovascular life support.
Discussion
248
In the present study, LV position changed significantly during cardiac arrest compared with the pre-arrest baseline. More specifically, the LV chamber progressively moved toward the left during untreated VF and BLS, and then moved toward the right during ACLS. Left ventricular position had significant correlations with haemodynamic parameters. To our knowledge, the transverse relationship between the sternum and the LV has been investigated in two studies [11,12], both of which indicate that only a small proportion of the LV is directly compressed by the sternum during CPR. Consistent with these studies, the PointMidLV was situated 7.0 6.5 mm left of the anterior-posterior axis, and more than 75% of the LV area was situated on the left side of the anterior-posterior axis at the pre-arrest baseline. To our knowledge, our study is the first to evaluate LV position
249
Please cite this article in press as: Jung YH, et al. Relationship Between Left Ventricle Position and Haemodynamic Parameters During Cardiopulmonary Resuscitation in a Pig Model. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j. hlc.2017.08.020
250 251 252 253 254 255 256 257 258 259 260 261 262 263 264
HLC 2481 1–9
6
Y.H. Jung et al.
Figure 4 Correlations between left ventricular position parameters and haemodynamic parameters during BLS. Echocardiograms and haemodynamic data were obtained four times at 2-min interval during BLS, resulting in 60 measurements for each parameter. Abbreviations;: BLS = basic life support.
265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289
during cardiac arrest. In our study, the LV progressively moved toward the left while assuming a more vertical position during cardiac arrest. The mechanisms underlying the shift in LV position are not readily apparent, but CT images from our study indicate that RV distension during cardiac arrest might have caused the shift in LV position. However, further studies are required to clarify the mechanisms underlying the shift in LV position during cardiac arrest. ETCO2 has been shown to correlate well with cardiac output during CPR [16,17]. In the present study, ETCO2 was significantly correlated with the LV position parameters during BLS. This finding indicates that cardiac output during CPR is related to the LV position. Consistent with this finding, Hwang et al. investigated the most prominently compressed cardiovascular structure during CPR using transoesophageal echocardiography and reported that LV stroke volume increased as the area of maximal compression was located closer to the LV [18]. Meanwhile, arterial pressure was not correlated with the LV position parameters during BLS. In contrast to ETCO2, the systemic blood pressure during CPR depends not only on cardiac output but more critically on vasomotor tone. Therefore, systemic blood pressure during CPR does not necessarily correlate with cardiac output [19]. During ACLS, the significance of the correlations between ETCO2 and LV position parameters disappeared. One explanation for this finding is the use of high-dose vasopressin
during ACLS, which might have influenced ETCO2 values by decreasing cardiac output [20,21]. Meanwhile, the correlations between systolic arterial pressure and LV position parameters (DAP-MidLV and DAP-Apex) became significant during ACLS. Systolic arterial pressure during ACLS increased as the LVARight/LVATotal increased, although the correlations did not reach statistical significance. These findings may indirectly indicate that the haemodynamic efficacy of vasopressor was greater in pigs with LV chamber closer to the midline. In a pig model of cardiac arrest, epinephrine improved haemodynamic parameters when administered during good-quality CPR, but it did not do so when administered during relatively poor-quality CPR [22]. In our study, no significant differences in the LV position parameters were observed between animals that achieved ROSC and those that did not. This study was not designed to evaluate the effects of LV position on resuscitation rate, but rather to evaluate changes in LV position during cardiac arrest. Thus, the sample size here may have been too small to allow us to detect potential differences in the resuscitation rate. On the other hand, the LV position might have been inconsequential for ROSC. In the present study, the correlations between the LV position parameters and haemodynamic parameters were weak. This may indicate that several factors, besides LV position, affect haemodynamic parameters during CPR.
Please cite this article in press as: Jung YH, et al. Relationship Between Left Ventricle Position and Haemodynamic Parameters During Cardiopulmonary Resuscitation in a Pig Model. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j. hlc.2017.08.020
290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315
HLC 2481 1–9 Relationship Between Left Ventricle Position and Haemodynamic Parameters
7
Figure 5 Correlations between left ventricular position parameters and haemodynamic parameters during ACLS. In order to determine correlations between left ventricular position parameters and haemodynamic parameters during ACLS, data obtained at 2 mins after the start of ACLS were used. Abbreviations;: ACLS = advanced cardiovascular life support.
316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338
The relatively high standard deviation (SD) values of the LV position parameters in the present study indicate that there is inter-individual variation in LV position. Consistent with this finding, several studies have reported wide inter-individual variations in thoracic anatomy [7,11,23]. Tømte et al. observed cardiac chambers using transoesophageal echocardiography in pigs that underwent mechanical CPR [23]. They reported that, despite consistent sternal piston positioning, some animals showed exclusive left chamber compression while others had a more symmetrical two-chamber compression or mainly right chamber compression. The inter-individual variation in thoracic anatomy, together with the findings of our study, indicate that adjusting resuscitation efforts (for example, the positioning of the hands during CPR) according to LV position may improve the haemodynamic efficacy of CPR by optimising the cardiac pump. This concept is supported by a recent study indicating that a resuscitation strategy that takes into account LV position during CPR may improve resuscitation outcomes [24]. Anderson et al. compared chest compressions directed over the LV and standard chest compressions in a pig model of cardiac arrest. They reported that chest compressions directed over the LV resulted in improved haemodynamics and a higher rate of ROSC compared to standard chest compressions [24].
Our study has several limitations. First, the findings of this study cannot be directly extrapolated to humans because of differences in chest configuration and cardiac anatomy between swine and humans. Second, echocardiograms and CT images were obtained during chest compression pauses because it was not possible to obtain satisfactory images during chest compression. Left ventricular position during the chest compression pause might not be exactly the same as that during the chest compression. Third, chest compressions were performed manually. Despite our efforts to deliver consistent chest compressions, their variability might have influenced the results. Fourth, echocardiographic measurement is affected by the operator’s technical skill. Although the echocardiograms were conducted by a single researcher, observer-dependent bias might have occurred in our study. Fifth, the anterior-posterior axis was chosen to represent the direction of sternal displacement by chest compression. However, chest compression transfers pressure to a wider area than this thin line. Sixth, the RA pressure at the prearrest baseline was slightly high in the present study, presumably because of the saline administration during the preparation phase. This might have affected the results. In conclusion, LV position changed significantly during cardiac arrest compared to the pre-arrest baseline in a pig
Please cite this article in press as: Jung YH, et al. Relationship Between Left Ventricle Position and Haemodynamic Parameters During Cardiopulmonary Resuscitation in a Pig Model. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j. hlc.2017.08.020
339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362
HLC 2481 1–9
8
Y.H. Jung et al.
Figure 6 Axial images at the level with the maximal LV area at the pre-arrest baseline in the three pigs (A, B, C) that underwent CT scanning 2 mins before the induction of VF, 7 mins after the initiation of VF, immediately after the 2-min BLS, and immediately after the 2-min ACLS. Abbreviations;: VF = ventricular fraction; BLS = basic life support; ACLS = advanced cardiovascular life support; CT = computed tomography.
Table 2 Left ventricular position parameters in the three animals that underwent computed tomography. Baseline
DAP-MidLV (mm)
-3.5,
14.7,
DAP-Apex (mm)
10.5,
6.9, 7.9
4.1
LVARight/LVATotal (%)
31.1, 0, 26.9
7 mins after the
2 mins after the
2 mins after the
initiation of VF
start of BLS
start of ACLS
-5.8, 3.1,
16.3, 12.7,
18.7, 0, 6.7
9.4 5.6
-9.4, 1.0,
17.5, 13.9,
13.4, 0, 1.0
16.3 11.6
-7.8, 1.9,
17.5, 13.5,
15.4 10.5
10.2, 0, 0
Positive DAP-MidLV and DAP-Apex values indicate right positions and negative values indicate left positions. Abbreviations[8_TD$IF]: VF, ventricular fibrillation; BLS, basic life support; ACLS, advanced cardiovascular life support.
363
2015R1D1A1A09057248]. The funder had no role in the study design, data collection, analysis, decision to publish, or preparation of the manuscript. We thank Mr. Suh Kang-Suk and Mrs. Choi Young-Ok for their generous support.
372
365
model. Left ventricular position during the chest compression pause had significant correlations with haemodynamic parameters during CPR.
366
Competing Interests
367
None.
368
Acknowledgements
Appendix A. Supplementary data
377
369
This work was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education [NRF-
Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.hlc. 2017.08.020.
378
364
370 371
Please cite this article in press as: Jung YH, et al. Relationship Between Left Ventricle Position and Haemodynamic Parameters During Cardiopulmonary Resuscitation in a Pig Model. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j. hlc.2017.08.020
373 374 375 376
379 380
HLC 2481 1–9 Relationship Between Left Ventricle Position and Haemodynamic Parameters
381 382
References
383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424
[1] Kouwenhoven WB, Jude JR, Knickerbocker GG. Closed-chest cardiac massage. JAMA 1960;173:1064–7. [2] Deshmukh HG, Weil MH, Gudipati CV, Trevino RP, Bisera J, Rackow EC. Mechanism of blood flow generated by precordial compression during CPR. I. Studies on closed chest precordial compression. Chest 1989;95: 1092–9. [3] Kim H, Hwang SO, Lee CC, Lee KH, Kim JY, Yoo BS, et al. Direction of blood flow from the left ventricle during cardiopulmonary resuscitation in humans: its implications for mechanism of blood flow. Am Heart J 2008;156(1222):e1–7. [4] Rudikoff MT, Maughan WL, Effron M, Freund P, Weisfeldt ML. Mechanisms of blood flow during cardiopulmonary resuscitation. Circulation 1980;61:345–52. [5] Niemann JT, Rosborough JP, Hausknecht M, Garner D, Criley JM. Pressure-synchronized cineangiography during experimental cardiopulmonary resuscitation. Circulation 1981;64:985–91. [6] Papadimitriou P, Chalkias A, Mastrokostopoulos A, Kapniari I, Xanthos T. Anatomical structures underneath the sternum in healthy adults and implications for chest compressions. Am J Emerg Med 2013;31:549–55. [7] Pickard A, Darby M, Soar J. Radiological assessment of the adult chest: implications for chest compressions. Resuscitation 2006;71:387–90. [8] Park YS, Park I, Kim YJ, Chung TN, Kim SW, Kim MJ, et al. Estimation of anatomical structures underneath the chest compression landmarks in children by using computed tomography. Resuscitation 2011;82:1030–5. [9] Cha KC, Kim YJ, Shin HJ, Cha YS, Kim H, Lee KH, et al. Optimal position for external chest compression during cardiopulmonary resuscitation: an analysis based on chest CT in patients resuscitated from cardiac arrest. Emerg Med J 2013;30:615–9. [10] Shin J, Rhee JE, Kim K. Is the inter-nipple line the correct hand position for effective chest compression in adult cardiopulmonary resuscitation? Resuscitation 2007;75:305–10. [11] Nestaas S, Stensæth KH, Rosseland V, Kramer-Johansen J. Radiological assessment of chest compression point and achievable compression depth in cardiac patients. Scand J Trauma Resusc Emerg Med 2016; 24:54. [12] Yun JG, Lee BK. Spatial relationship of the left ventricle in the supine position and the left lateral tilt position: Implication for cardiopulmonary resuscitation in pregnant patients. Fire Sci Eng 2013;27:75–9. [13] Neumar RW, Otto CW, Link MS, Kronick SL, Shuster M, Callaway CW, et al. Part 8: adult advanced cardiovascular life support: 2010 American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Circulation 2010;122:S729–67.
9
[14] Cho YC, Cho SW, Chung SP, Yu K, Kwon OY, Kim SW. How can a single rescuer adequately deliver tidal volume with a manual resuscitator? An improved device for delivering regular tidal volume. Emerg Med J 2011;28:40–3. [15] Idris AH, Becker LB, Ornato JP, Hedges JR, Bircher NG, Chandra NC, et al. Utstein-style guidelines for uniform reporting of laboratory CPR research A statement for healthcare professionals from a Task Force of the American Heart Association, the American College of Emergency Physicians, the American College of Cardiology, the European Resuscitation Council, the Heart and Stroke Foundation of Canada, the Institute of Critical Care Medicine, the Safar Center for Resuscitation Research, and the Society for Academic Emergency Medicine. Resuscitation 1996;33:69–84. [16] Weil MH, Bisera J, Trevino RP, Rackow EC. Cardiac output and end-tidal carbon dioxide. Crit Care Med 1985;13:907–9. [17] Ryu SJ, Lee SJ, Park CH, Lee SM, Lee DH, Cho YS, et al. Arterial pressure, end-tidal carbon dioxide, and central venous oxygen saturation in reflecting compression depth. Acta Anaesthesiol Scand 2016;60:1012–23. [18] Hwang SO, Zhao PG, Choi HJ, Park KH, Cha KC, Park SM, et al. Compression of the left ventricular outflow tract during cardiopulmonary resuscitation. Acad Emerg Med 2009;16:928–33. [19] Ornato JP, Levine RL, Young DS, Racht EM, Garnett AR, Gonzalez ER. The effect of applied chest compression force on systemic arterial pressure and end-tidal carbon dioxide concentration during CPR in human beings. Ann Emerg Med 1989;18:732–7. [20] Lindner KH, Brinkmann A, Pfenninger EG, Lurie KG, Goertz A, Lindner IM. Effect of vasopressin on hemodynamic variables, organ blood flow, and acid-base status in a pig model of cardiopulmonary resuscitation. Anesth Analg 1993;77:427–35. [21] Lindner KH, Prengel AW, Pfenninger EG, Lindner IM, Strohmenger HU, Georgieff M, et al. Vasopressin improves vital organ blood flow during closed-chest cardiopulmonary resuscitation in pigs. Circulation 1995;91:215–21. [22] Pytte M, Kramer-Johansen J, Eilevstjønn J, Eriksen M, Strømme TA, Godang K, et al. Haemodynamic effects of adrenaline (epinephrine) depend on chest compression quality during cardiopulmonary resuscitation in pigs. Resuscitation 2006;71:369–78. [23] Tømte Ø, Sjaastad I, Wik L, Kuzovlev A, Eriksen M, Norseng PA, et al. Discriminating the effect of accelerated compression from accelerated decompression during high-impulse CPR in a porcine model of cardiac arrest. Resuscitation 2010;81:488–92. [24] Anderson KL, Castaneda MG, Boudreau SM, Sharon DJ, Bebarta VS. Left ventricular compressions improve hemodynamics in a swine model of out-of-hospital cardiac arrest. Prehosp Emerg Care 2017;21:272–80.
Please cite this article in press as: Jung YH, et al. Relationship Between Left Ventricle Position and Haemodynamic Parameters During Cardiopulmonary Resuscitation in a Pig Model. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j. hlc.2017.08.020
425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468
Heart, Lung and Circulation (2017) xx, 1–9 1443-9506/04/$36.00 http://dx.doi.org/10.1016/j.hlc.2017.08.020
1 2 3
Q1
4 5
6 7 8 9 10 11 12 13 14
Q2
ORIGINAL ARTICLE
Relationship Between Left Ventricle Position and Haemodynamic Parameters During Cardiopulmonary Resuscitation in a Pig Model [TD$FIRSNAME]Yong Hun[TD$FIRSNAME.] [TD$SURNAME]Jung[,TD$SURNAME.] MD a[7_TD$IF], [TD$FIRSNAME]Kyung Woon[TD$FIRSNAME.] [TD$SURNAME]Jeung[TD$SURNAME.], MD, PhD a*, [TD$FIRSNAME]Dong Hun[TD$FIRSNAME.] [TD$SURNAME]Lee[TD$SURNAME.], MD a, [TD$FIRSNAME]Young Won[TD$FIRSNAME.] [TD$SURNAME]Jeong[TD$SURNAME.], EMTP a, [TD$FIRSNAME]Sung Min[TD$FIRSNAME.] [TD$SURNAME]Lee[TD$SURNAME.], MD a, [TD$FIRSNAME] Byung Kook[TD$FIRSNAME.] [TD$SURNAME]Lee[TD$SURNAME.], MD, PhD a, [TD$FIRSNAME]In Seok[TD$FIRSNAME.] [TD$SURNAME]Jeong[TD$SURNAME.], MD, PhD b, [TD$FIRSNAME]Sang-Kwon[TD$FIRSNAME.] [TD$SURNAME]Lee[TD$SURNAME.], DVM c, [TD$FIRSNAME]Jihye[TD$FIRSNAME.] [TD$SURNAME]Choi[TD$SURNAME.], DVM PhD c a
Department of Emergency Medicine, Chonnam National University Hospital, Gwangju, Republic of Korea Department of Thoracic and Cardiovascular Surgery, Chonnam National University Hospital, Gwangju, Republic of Korea c Veterinary Medical Imaging, College of Veterinary Medicine, Chonnam National University, Gwangju, Republic of Korea b
Received 14 June 2017; received in revised form 14 August 2017; accepted 22 August 2017; online published-ahead-of-print xxx
Background
From the viewpoint of cardiac pump theory, the area of the left ventricle (LV) subjected to compression increases as the LV lies closer to the sternum, possibly resulting in higher blood flow in patients with LV closer to the sternum. However, no study has evaluated LV position during cardiac arrest or its relationship with haemodynamic parameters during cardiopulmonary resuscitation (CPR). The objectives of this study were to determine whether the position of the LV relative to the anterior-posterior axis representing the direction of chest compression shifts during cardiac arrest and to examine the relationship between LV position and haemodynamic parameters during CPR.
Methods
Subcostal view echocardiograms were obtained from 15 pigs with the transducer parallel to the long axis of the sternum before inducing ventricular fibrillation (VF) and during cardiac arrest. Computed tomography was performed in three pigs to objectively observe LV position during cardiac arrest. Left ventricular position parameters including the shortest distance between the anterior-posterior axis and the mid-point of the LV chamber (DAP-MidLV), the shortest distance between the anterior-posterior axis and the LV apex (DAP-Apex), and the area fraction of the LV located on the right side of the anterior-posterior axis (LVARight/ LVATotal) were measured.
Results
DAP-MidLV, DAP-Apex, and LVARight/LVATotal decreased progressively during untreated VF and basic life support (BLS), and then increased during advanced cardiovascular life support (ACLS). A repeated measures analysis of variance revealed significant time effects for these parameters. During BLS, the end-tidal carbon dioxide and systolic right atrial pressure were significantly correlated with the LV position parameters. During ACLS, systolic arterial pressure and systolic right atrial pressure were significantly correlated with DAP-MidLV and DAP-Apex.
Conclusions
Left ventricular position changed significantly during cardiac arrest compared to the pre-arrest baseline. LV position during CPR had significant correlations with haemodynamic parameters.
Keywords
Heart arrest Cardiopulmonary resuscitation Heart ventricles Haemodynamics
Q3
15 *Corresponding author at: Department of Emergency Medicine, Chonnam National University Hospital, 42 Jebong-ro, Donggu, Gwangju, Republic of Korea. Fax: +82 62 228 7417., Email: [email protected] © 2017 Australian and New Zealand Society of Cardiac and Thoracic Surgeons (ANZSCTS) and the Cardiac Society of Australia and New Zealand (CSANZ). Published by Elsevier B.V. All rights reserved.
Please cite this article in press as: Jung YH, et al. Relationship Between Left Ventricle Position and Haemodynamic Parameters During Cardiopulmonary Resuscitation in a Pig Model. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j. hlc.2017.08.020
HLC 2481 1–9
2
Y.H. Jung et al.
16 17
Introduction
18
Cardiopulmonary resuscitation (CPR) has been the cornerstone of treatment for cardiac arrest since 1960 [1]. Over the past 50 years, considerable efforts have been made to expand the understanding of the physiology behind CPR [2–5]. However, many questions remain unexplored or incompletely understood. Blood flow during CPR is thought to be provided by direct cardiac compression (cardiac pump) and/or increased intrathoracic pressure (thoracic pump) [2–5]. According to the first theory, blood flow is generated as the heart is squeezed between the sternum and the spine at the level of the ventricles. From this point-of-view, the area of the left ventricle (LV) subjected to compression increases as the LV lies closer to the sternum, possibly resulting in higher blood flow in patients with LV closer to the sternum [3]. A number of studies have investigated the spatial relationship between the sternum and anatomical structures using magnetic resonance imaging or computed tomography (CT) [6–11]. However, most of the studies have been focussed on the vertical relationship between the sternum and the inner viscera. Until now, few studies have explored the spatial relationship in the transverse direction [11,12], and, to our knowledge, no study has evaluated the relationship between LV position and haemodynamic parameters during CPR. Furthermore, all of these studies have used images obtained in patients not in cardiac arrest [6–11]. However, the position of the cardiac chamber may shift during cardiac arrest. In this study, we investigated the spatial relationship of the LV in a transverse direction, in relation to the anterior-posterior axis, in a swine model of cardiac arrest to determine whether LV position changes during cardiac arrest. We also examined the relationship between LV position and haemodynamic parameters during CPR. We hypothesised that LV position would change during cardiac arrest compared to the pre-arrest baseline and that LV position would be related to haemodynamic parameters during CPR.
19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53
Q4
54
Material and Methods
55
65
A total of 18 Yorkshire/Landrace-cross pigs weighing 23.9 2.2 kg were used for this study. This study consisted of two sub-studies: an echocardiography study to examine the relationship between LV position and haemodynamic parameters during CPR (n = 15) and a CT study for the more objective observation of LV position during cardiac arrest (n = 3). The Animal Care and Use Committee of Chonnam National University approved the protocol for this study (CNU IACUC-H-2016-22). Animal care and experiments were in accord with the National Institutes of Health Guide for the Care and Use of Laboratory Animals.
66
Animal Preparation
67
After premedication (ketamine, 20 mg/kg; xylazine, 2.2 mg/kg; atropine, 0.04 mg/kg, intramuscular), the
56 57 58 59 60 61 62 63 64
animals were placed supine in a u-shaped trough with the limbs secured to prevent lateral displacement of the chest during CPR. After tracheal intubation, anaesthesia was provided with 70%:30% N2O:O2 and 0.5–2% sevoflurane, which was titrated to prevent signs of pain (reactive wide pupils, tachycardia, and hypertension). A doublelumen catheter was inserted via the right femoral artery for monitoring of aortic pressure. The right external jugular vein was cannulated with an 8-French introducer sheath to monitor right atrial (RA) pressure, and to insert a right ventricle (RV) pacing catheter. An end-tidal carbon dioxide (ETCO2) sample line was connected to the ventilator circuit. During the preparation phase, saline was administered to maintain an RA pressure of 10–12 mmHg.
68
Echocardiography Study
82
After baseline measurements, ventricular fibrillation (VF) was induced by passing 60 Hz of AC current at 30 mA through the RV pacing catheter. The animals were then disconnected from ventilatory support. After 14 mins of untreated VF ([1_TD$IF]Figure 1), basic life support (BLS) was started, using cycles of 30 chest compressions followed by two ventilations with ambient air. Closed-chest compressions were administered by two experienced investigators (DHL and YWJ) who were unaware of the study’s aims, at a rate of 100/ min and a compression depth of 25% of the anterior-posterior diameter of the chest wall. The depth of the chest compressions was monitored by another investigator (KWJ). After 8 mins of BLS, advanced cardiovascular life support (ACLS) based on recent resuscitation guidelines was started [13]. Asynchronous positive-pressure ventilations with high flow O2 (14 l/min) were provided at a rate of 8/min, using a volume-marked bag devised by Cho et al. [14]. At the start of ACLS, all animals received 0.6 U/kg of vasopressin intravenously. After 4 mins of ACLS, 0.02 mg/kg of epinephrine was administered every 3 mins if required. If sustained restoration of spontaneous circulation (ROSC) was not achieved within 12 mins of ACLS [15], resuscitation efforts were discontinued. Animals that achieved ROSC underwent a 4-h period of intensive care for another study not reported here. Subcostal view echocardiograms (UST-9130, Hitachi Aloka Medical Ltd., Tokyo, Japan) were obtained by a single researcher at the pre-arrest baseline (5, 10, and 15 mins before VF induction), during untreated VF (1, 7, and 13 mins after VF induction), and every 2 mins during CPR until 6 mins after the start of ACLS. To obtain consistent orientation of the transducer in relation to the sternum, the following protocol was employed. A straight line was drawn through the middle of the transducer and a vertical line was drawn along the sternum down to the xiphoid process ([1_TD$IF]Figure 2). The transducer was placed just below the xiphoid process (with the line on the transducer aligned with the line drawn on the sternum) and gently pushed inward. The transducer was angled upward until the maximum obtainable chamber size of the LV was visualised. The subcostal view was then maintained and stored for 5 secs. The straight line on the
83
Please cite this article in press as: Jung YH, et al. Relationship Between Left Ventricle Position and Haemodynamic Parameters During Cardiopulmonary Resuscitation in a Pig Model. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j. hlc.2017.08.020
69 70 71 72 73 74 75 76 77 78 79 80 81
84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123
HLC 2481 1–9 Relationship Between Left Ventricle Position and Haemodynamic Parameters
3
[1_TD$IF]Figure 1 Experimental timelines of the study using echocardiography (A) and the study using CT (B). The lightning marks indicate the onset of the 10-sec pause in chest compressions for rhythm analysis and a 150-J shock if indicated.
124 125 126 127 128
transducer was maintained in a parallel orientation to the line drawn along the sternum throughout the procedure. The echocardiograms were obtained at end-expiratory phase, and those during resuscitation were obtained during a 10sec pause in chest compression for rhythm analysis.
CT Study The VF induction and resuscitation procedures were identical to those described above, except that the durations of the noflow period, BLS, and ACLS were shorter. In addition, no defibrillation was attempted ([1_TD$IF]Figure 1). Each animal
Figure 2 Subcostal view echocardiography. (A) Transducer placement. (B) Measurements of left ventricle position parameters. Please cite this article in press as: Jung YH, et al. Relationship Between Left Ventricle Position and Haemodynamic Parameters During Cardiopulmonary Resuscitation in a Pig Model. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j. hlc.2017.08.020
129 130 131 132
HLC 2481 1–9
4
133
Y.H. Jung et al.
144
underwent CT scanning (SOMATOM Emotion 16, Siemens, Forchheim, Germany) four times: 2 mins before VF induction, 7 mins after VF induction, immediately after the BLS, and immediately after the ACLS. We used the following settings: slice thickness, 1 mm; pitch, 0.8; rotation duration, 600 ms; tube voltage, 120 kV, and tube current, 120 mA. All scans were performed after injecting a dose of 30 mL Iohexol (Ominpaque 300, GE healthcare, Oslo, Norway) at a flow rate of 3 ml/sec. Contrast material injection was started 15 secs before each CT scan. In order to exclude the effects of breathing on LV position, vecuronium bromide was infused at a rate of 100 mg/min, and no ventilation was provided during the scan.
145
Measurements
134 135 136 137 138 139 140 141 142 143
146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163
Aortic pressure and RA pressure were continuously monitored (CS/3 CCM, Datex-Ohmeda, Helsinki, Finland) and transferred to a personal computer using S/5 Collect software (Datex-Ohmeda, Helsinki, Finland). Systolic arterial pressure was defined as the peak aortic pressure during the chest compression phase. Coronary perfusion pressure (CPP) was calculated by subtracting RA end-diastolic pressure from simultaneous aortic end-diastolic pressure. Systolic arterial pressure, systolic RA pressure, and CPP were sampled at 2-min intervals by averaging pressures from five consecutive compressions immediately before the 10-sec pause in chest compression for rhythm analysis. ETCO2 values were determined every 2 mins by averaging ETCO2 values for the preceding 30-sec interval. A single observer blinded to the haemodynamic data analysed the echocardiographic images. A line (LineMitral) was drawn between the points of mural attachment of the mitral valve leaflets ([1_TD$IF]Figure 2), and a straight line representing the LV axis was drawn between the mid-point of the LineMitral and the LV apex.
PointMidLV, which represented the centre of the LV chamber, was defined as the mid-point of the LV axis. The anteriorposterior axis was defined as a line passing through the middle of the image. With these definitions, the following LV position parameters were measured: DAP-MidLV (shortest distance between the anterior-posterior axis and PointMidLV) and DAP-Apex (shortest distance between the anterior-posterior axis and the LV apex). In addition, LV area (LVATotal) was measured and the area fraction of the LV located on the right side of the anterior-posterior axis (LVARight/LVATotal) was calculated by dividing the LV area located on the right side of the anterior-posterior axis (LVARight) by LVATotal. The average LV position parameter values were obtained during each period and were used for analysis. The CT images were analysed using a picture archiving and communication system (MAROSIS Maroview, Marotech Inc., Seoul Korea). An axial image at the level with the maximal LV area was chosen among the images obtained at the prearrest baseline. Axial images at the same level were then selected from the images obtained from each scan and LV position parameters were measured using the definitions described above.
164
Statistical Analysis
186
Sample size calculation was based on DAP-MidLV data from a pilot study [9_TD$IF]( 5.3 3.0 mm at pre-arrest baseline vs. 10.1 4.4 mm at BLS), yielding a minimum requirement of 10 animals with a set at 0.05 and a power of 0.90. Considering the result of the sample size calculation for the other study conducted simultaneously, 15 animals were used for the echocardiography study. We used three animals for the CT study due to our limited resources. Normally distributed variables were presented as
Table 1 Baseline characteristics. Total (N = 15)
ROSC (N = 8)
No ROSC (N = 7)
p Value
Systolic arterial pressure (mmHg)
117.4 (11.7)
117.9 (13.1)
116.9 (10.8)
0.873
Diastolic arterial pressure (mmHg)
79.7 (10.2)
81.0 (11.7)
78.1 (8.8)
0.606
Mean arterial pressure (mmHg)
94.8 (11.4)
96.3 (12.9)
93.1 (10.1)
0.616
Systolic RA pressure (mmHg)
14.5 (1.5)
13.9 (1.2)
15.1 (1.6)
0.105
Diastolic RA pressure (mmHg)
11.2 (1.5)
11.0 (10.0–11.3)
12.0 (11.5–12.5)
0.193
Mean RA pressure (mmHg)
12.0 (12.0–14.0)
12.0 (12.0–12.5)
14.0 (12.5–14.0)
0.240
Heart rate (/min)
98 (85–107)
102 (19)
100 (26)
0.870
ETCO2 (mmHg) pH
39.4 (2.4) 7.483 (0.046)
39.6 (2.8) 7.490 (0.047)
39.1 (2.0) 7.476 (0.046)
0.711 0.576
PaCO2 (mmHg)
39.5 (2.7)
39.8 (2.8)
39.1 (2.7)
0.635
PaO2 (mmHg)
142.5 (23.4)
146.8 (30.1)
137.6 (13.0)
0.468
HCO3 (mmol/l)
29.1 (2.1)
29.8 (2.2)
28.3 (1.7)
0.173
SaO2 (%)
99.5 (98.9–99.8)
99.7 (98.9–100)
99.5 (99.1–99.5)
0.483
Lactate (mmol/l)
0.8 (0.7–1.0)
0.8 (0.2)
1.0 (0.5)
0.348
Data are presented as mean (standard deviation) or as median (interquartile ranges). Abbreviations: RA, right atrial; ETCO2, end-tidal carbon dioxide; ROSC, Return of spontaneous circulation, ETCO2, end-tidal carbon dioxide.
Please cite this article in press as: Jung YH, et al. Relationship Between Left Ventricle Position and Haemodynamic Parameters During Cardiopulmonary Resuscitation in a Pig Model. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j. hlc.2017.08.020
165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185
187 188 189 190 191 192 193 194
HLC 2481 1–9 Relationship Between Left Ventricle Position and Haemodynamic Parameters
207
mean standard deviation (SD) and an independent t-test was performed, whereas non-normally distributed variables were presented as medians with interquartile ranges and a Mann-Whitney U test was conducted. Correlations between variables were determined using Pearson or Spearman correlation tests where appropriate. Repeated measures analysis of variance (ANOVA) was used for serial measurements. Pair-wise comparison with Bonferroni adjustment was performed for post-hoc analysis. To evaluate intraobserver variability in the echocardiographic measurements, intraclass correlation coefficient was calculated using the data obtained at the pre-arrest baseline. A p value of [10_TD$IF]<0.05 was considered significant.
208
Results
209
Echocardiography Study
210
240
Baseline characteristics are shown in Table 1. Eight (53.3%) of the 15 animals studied achieved sustained ROSC. An overview of LV position parameters in the echocardiography study is shown in Supplementary material 1. The intraclass correlation coefficient for the echocardiographic measurements was 0.922 (95% confidence interval, 0.876–0.953). DAP-MidLV, DAP-Apex, and LVARight/LVATotal decreased progressively during untreated VF and BLS, and then increased during ACLS ([1_TD$IF]Figure 3). The repeated measures ANOVA revealed significant time effects for all of these parameters (p = 0.001 for DAP-MidLV, <0.001 for DAP-Apex, and <0.001 for LVARight/LVATotal). The repeated measures ANOVA for comparisons of parameters between animals that achieved Return of spontaneous circulation (ROSC) and those that did not showed no significant inter-group differences for any of the LV position parameters (p = 0.287 for DAP-MidLV, 0.338 for DAP-Apex, and 0.362 for LVARight/LVATotal). Correlations between the LV position parameters and haemodynamic parameters during CPR are summarised in [1_TD$IF]Figures 4 and 5. During BLS, ETCO2 and systolic RA pressure had significant correlations with all LV position parameters. Systolic arterial pressure and CPP were not significantly correlated with the LV position parameters. During ACLS, the correlations between the systolic RA pressure and LV position parameters remained significant, but the correlations between ETCO2 and LV position parameters were no longer significant. Systolic arterial pressure during ACLS was significantly correlated with DAP-MidLV and DAP-Apex. Systolic arterial pressure during ACLS increased as the LVARight/LVATotal increased, although the correlations did not reach statistical significance (p = 0.071).
241
CT Study
242
Axial images at the level with the maximal LV area at the prearrest baseline are shown in [1_TD$IF]Figure 6. The LV position parameters in the CT study are displayed in Table 2. Consistent with the findings of the echocardiography study, DAP-MidLV, DAP-Apex, and LVARight/LVATotal decreased progressively during cardiac arrest in all pigs.
195 196 197 198 199 200 201 202 203 204 205 206
211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239
243 244 245 246 247
Q5
5
Figure 3 Changes in DAP-MidLV (A), DAP-Apex (B), and LVARight/LVATotal (C) during the pre-arrest baseline, untreated VF, BLS, and ACLS. A repeated measures analysis of variance revealed significant time effects for all of these parameters (p = 0.001 for DAP-MidLV, <0.001 for DAP-Apex, and <0.001 for LVARight/LVATotal). The average of the values obtained during each period was used for analysis. Positive DAP-MidLV and DAP-Apex values indicate right positions and negative values indicate left positions. *p < 0.05 versus baseline value. Abbreviations;: VF = ventricular fraction; BLS = basic life support; ACLS = advanced cardiovascular life support.
Discussion
248
In the present study, LV position changed significantly during cardiac arrest compared with the pre-arrest baseline. More specifically, the LV chamber progressively moved toward the left during untreated VF and BLS, and then moved toward the right during ACLS. Left ventricular position had significant correlations with haemodynamic parameters. To our knowledge, the transverse relationship between the sternum and the LV has been investigated in two studies [11,12], both of which indicate that only a small proportion of the LV is directly compressed by the sternum during CPR. Consistent with these studies, the PointMidLV was situated 7.0 6.5 mm left of the anterior-posterior axis, and more than 75% of the LV area was situated on the left side of the anterior-posterior axis at the pre-arrest baseline. To our knowledge, our study is the first to evaluate LV position
249
Please cite this article in press as: Jung YH, et al. Relationship Between Left Ventricle Position and Haemodynamic Parameters During Cardiopulmonary Resuscitation in a Pig Model. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j. hlc.2017.08.020
250 251 252 253 254 255 256 257 258 259 260 261 262 263 264
HLC 2481 1–9
6
Y.H. Jung et al.
Figure 4 Correlations between left ventricular position parameters and haemodynamic parameters during BLS. Echocardiograms and haemodynamic data were obtained four times at 2-min interval during BLS, resulting in 60 measurements for each parameter. Abbreviations;: BLS = basic life support.
265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289
during cardiac arrest. In our study, the LV progressively moved toward the left while assuming a more vertical position during cardiac arrest. The mechanisms underlying the shift in LV position are not readily apparent, but CT images from our study indicate that RV distension during cardiac arrest might have caused the shift in LV position. However, further studies are required to clarify the mechanisms underlying the shift in LV position during cardiac arrest. ETCO2 has been shown to correlate well with cardiac output during CPR [16,17]. In the present study, ETCO2 was significantly correlated with the LV position parameters during BLS. This finding indicates that cardiac output during CPR is related to the LV position. Consistent with this finding, Hwang et al. investigated the most prominently compressed cardiovascular structure during CPR using transoesophageal echocardiography and reported that LV stroke volume increased as the area of maximal compression was located closer to the LV [18]. Meanwhile, arterial pressure was not correlated with the LV position parameters during BLS. In contrast to ETCO2, the systemic blood pressure during CPR depends not only on cardiac output but more critically on vasomotor tone. Therefore, systemic blood pressure during CPR does not necessarily correlate with cardiac output [19]. During ACLS, the significance of the correlations between ETCO2 and LV position parameters disappeared. One explanation for this finding is the use of high-dose vasopressin
during ACLS, which might have influenced ETCO2 values by decreasing cardiac output [20,21]. Meanwhile, the correlations between systolic arterial pressure and LV position parameters (DAP-MidLV and DAP-Apex) became significant during ACLS. Systolic arterial pressure during ACLS increased as the LVARight/LVATotal increased, although the correlations did not reach statistical significance. These findings may indirectly indicate that the haemodynamic efficacy of vasopressor was greater in pigs with LV chamber closer to the midline. In a pig model of cardiac arrest, epinephrine improved haemodynamic parameters when administered during good-quality CPR, but it did not do so when administered during relatively poor-quality CPR [22]. In our study, no significant differences in the LV position parameters were observed between animals that achieved ROSC and those that did not. This study was not designed to evaluate the effects of LV position on resuscitation rate, but rather to evaluate changes in LV position during cardiac arrest. Thus, the sample size here may have been too small to allow us to detect potential differences in the resuscitation rate. On the other hand, the LV position might have been inconsequential for ROSC. In the present study, the correlations between the LV position parameters and haemodynamic parameters were weak. This may indicate that several factors, besides LV position, affect haemodynamic parameters during CPR.
Please cite this article in press as: Jung YH, et al. Relationship Between Left Ventricle Position and Haemodynamic Parameters During Cardiopulmonary Resuscitation in a Pig Model. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j. hlc.2017.08.020
290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315
HLC 2481 1–9 Relationship Between Left Ventricle Position and Haemodynamic Parameters
7
Figure 5 Correlations between left ventricular position parameters and haemodynamic parameters during ACLS. In order to determine correlations between left ventricular position parameters and haemodynamic parameters during ACLS, data obtained at 2 mins after the start of ACLS were used. Abbreviations;: ACLS = advanced cardiovascular life support.
316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338
The relatively high standard deviation (SD) values of the LV position parameters in the present study indicate that there is inter-individual variation in LV position. Consistent with this finding, several studies have reported wide inter-individual variations in thoracic anatomy [7,11,23]. Tømte et al. observed cardiac chambers using transoesophageal echocardiography in pigs that underwent mechanical CPR [23]. They reported that, despite consistent sternal piston positioning, some animals showed exclusive left chamber compression while others had a more symmetrical two-chamber compression or mainly right chamber compression. The inter-individual variation in thoracic anatomy, together with the findings of our study, indicate that adjusting resuscitation efforts (for example, the positioning of the hands during CPR) according to LV position may improve the haemodynamic efficacy of CPR by optimising the cardiac pump. This concept is supported by a recent study indicating that a resuscitation strategy that takes into account LV position during CPR may improve resuscitation outcomes [24]. Anderson et al. compared chest compressions directed over the LV and standard chest compressions in a pig model of cardiac arrest. They reported that chest compressions directed over the LV resulted in improved haemodynamics and a higher rate of ROSC compared to standard chest compressions [24].
Our study has several limitations. First, the findings of this study cannot be directly extrapolated to humans because of differences in chest configuration and cardiac anatomy between swine and humans. Second, echocardiograms and CT images were obtained during chest compression pauses because it was not possible to obtain satisfactory images during chest compression. Left ventricular position during the chest compression pause might not be exactly the same as that during the chest compression. Third, chest compressions were performed manually. Despite our efforts to deliver consistent chest compressions, their variability might have influenced the results. Fourth, echocardiographic measurement is affected by the operator’s technical skill. Although the echocardiograms were conducted by a single researcher, observer-dependent bias might have occurred in our study. Fifth, the anterior-posterior axis was chosen to represent the direction of sternal displacement by chest compression. However, chest compression transfers pressure to a wider area than this thin line. Sixth, the RA pressure at the prearrest baseline was slightly high in the present study, presumably because of the saline administration during the preparation phase. This might have affected the results. In conclusion, LV position changed significantly during cardiac arrest compared to the pre-arrest baseline in a pig
Please cite this article in press as: Jung YH, et al. Relationship Between Left Ventricle Position and Haemodynamic Parameters During Cardiopulmonary Resuscitation in a Pig Model. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j. hlc.2017.08.020
339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362
HLC 2481 1–9
8
Y.H. Jung et al.
Figure 6 Axial images at the level with the maximal LV area at the pre-arrest baseline in the three pigs (A, B, C) that underwent CT scanning 2 mins before the induction of VF, 7 mins after the initiation of VF, immediately after the 2-min BLS, and immediately after the 2-min ACLS. Abbreviations;: VF = ventricular fraction; BLS = basic life support; ACLS = advanced cardiovascular life support; CT = computed tomography.
Table 2 Left ventricular position parameters in the three animals that underwent computed tomography. Baseline
DAP-MidLV (mm)
-3.5,
14.7,
DAP-Apex (mm)
10.5,
6.9, 7.9
4.1
LVARight/LVATotal (%)
31.1, 0, 26.9
7 mins after the
2 mins after the
2 mins after the
initiation of VF
start of BLS
start of ACLS
-5.8, 3.1,
16.3, 12.7,
18.7, 0, 6.7
9.4 5.6
-9.4, 1.0,
17.5, 13.9,
13.4, 0, 1.0
16.3 11.6
-7.8, 1.9,
17.5, 13.5,
15.4 10.5
10.2, 0, 0
Positive DAP-MidLV and DAP-Apex values indicate right positions and negative values indicate left positions. Abbreviations[8_TD$IF]: VF, ventricular fibrillation; BLS, basic life support; ACLS, advanced cardiovascular life support.
363
2015R1D1A1A09057248]. The funder had no role in the study design, data collection, analysis, decision to publish, or preparation of the manuscript. We thank Mr. Suh Kang-Suk and Mrs. Choi Young-Ok for their generous support.
372
365
model. Left ventricular position during the chest compression pause had significant correlations with haemodynamic parameters during CPR.
366
Competing Interests
367
None.
368
Acknowledgements
Appendix A. Supplementary data
377
369
This work was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education [NRF-
Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.hlc. 2017.08.020.
378
364
370 371
Please cite this article in press as: Jung YH, et al. Relationship Between Left Ventricle Position and Haemodynamic Parameters During Cardiopulmonary Resuscitation in a Pig Model. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j. hlc.2017.08.020
373 374 375 376
379 380
HLC 2481 1–9 Relationship Between Left Ventricle Position and Haemodynamic Parameters
381 382
References
383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424
[1] Kouwenhoven WB, Jude JR, Knickerbocker GG. Closed-chest cardiac massage. JAMA 1960;173:1064–7. [2] Deshmukh HG, Weil MH, Gudipati CV, Trevino RP, Bisera J, Rackow EC. Mechanism of blood flow generated by precordial compression during CPR. I. Studies on closed chest precordial compression. Chest 1989;95: 1092–9. [3] Kim H, Hwang SO, Lee CC, Lee KH, Kim JY, Yoo BS, et al. Direction of blood flow from the left ventricle during cardiopulmonary resuscitation in humans: its implications for mechanism of blood flow. Am Heart J 2008;156(1222):e1–7. [4] Rudikoff MT, Maughan WL, Effron M, Freund P, Weisfeldt ML. Mechanisms of blood flow during cardiopulmonary resuscitation. Circulation 1980;61:345–52. [5] Niemann JT, Rosborough JP, Hausknecht M, Garner D, Criley JM. Pressure-synchronized cineangiography during experimental cardiopulmonary resuscitation. Circulation 1981;64:985–91. [6] Papadimitriou P, Chalkias A, Mastrokostopoulos A, Kapniari I, Xanthos T. Anatomical structures underneath the sternum in healthy adults and implications for chest compressions. Am J Emerg Med 2013;31:549–55. [7] Pickard A, Darby M, Soar J. Radiological assessment of the adult chest: implications for chest compressions. Resuscitation 2006;71:387–90. [8] Park YS, Park I, Kim YJ, Chung TN, Kim SW, Kim MJ, et al. Estimation of anatomical structures underneath the chest compression landmarks in children by using computed tomography. Resuscitation 2011;82:1030–5. [9] Cha KC, Kim YJ, Shin HJ, Cha YS, Kim H, Lee KH, et al. Optimal position for external chest compression during cardiopulmonary resuscitation: an analysis based on chest CT in patients resuscitated from cardiac arrest. Emerg Med J 2013;30:615–9. [10] Shin J, Rhee JE, Kim K. Is the inter-nipple line the correct hand position for effective chest compression in adult cardiopulmonary resuscitation? Resuscitation 2007;75:305–10. [11] Nestaas S, Stensæth KH, Rosseland V, Kramer-Johansen J. Radiological assessment of chest compression point and achievable compression depth in cardiac patients. Scand J Trauma Resusc Emerg Med 2016; 24:54. [12] Yun JG, Lee BK. Spatial relationship of the left ventricle in the supine position and the left lateral tilt position: Implication for cardiopulmonary resuscitation in pregnant patients. Fire Sci Eng 2013;27:75–9. [13] Neumar RW, Otto CW, Link MS, Kronick SL, Shuster M, Callaway CW, et al. Part 8: adult advanced cardiovascular life support: 2010 American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Circulation 2010;122:S729–67.
9
[14] Cho YC, Cho SW, Chung SP, Yu K, Kwon OY, Kim SW. How can a single rescuer adequately deliver tidal volume with a manual resuscitator? An improved device for delivering regular tidal volume. Emerg Med J 2011;28:40–3. [15] Idris AH, Becker LB, Ornato JP, Hedges JR, Bircher NG, Chandra NC, et al. Utstein-style guidelines for uniform reporting of laboratory CPR research A statement for healthcare professionals from a Task Force of the American Heart Association, the American College of Emergency Physicians, the American College of Cardiology, the European Resuscitation Council, the Heart and Stroke Foundation of Canada, the Institute of Critical Care Medicine, the Safar Center for Resuscitation Research, and the Society for Academic Emergency Medicine. Resuscitation 1996;33:69–84. [16] Weil MH, Bisera J, Trevino RP, Rackow EC. Cardiac output and end-tidal carbon dioxide. Crit Care Med 1985;13:907–9. [17] Ryu SJ, Lee SJ, Park CH, Lee SM, Lee DH, Cho YS, et al. Arterial pressure, end-tidal carbon dioxide, and central venous oxygen saturation in reflecting compression depth. Acta Anaesthesiol Scand 2016;60:1012–23. [18] Hwang SO, Zhao PG, Choi HJ, Park KH, Cha KC, Park SM, et al. Compression of the left ventricular outflow tract during cardiopulmonary resuscitation. Acad Emerg Med 2009;16:928–33. [19] Ornato JP, Levine RL, Young DS, Racht EM, Garnett AR, Gonzalez ER. The effect of applied chest compression force on systemic arterial pressure and end-tidal carbon dioxide concentration during CPR in human beings. Ann Emerg Med 1989;18:732–7. [20] Lindner KH, Brinkmann A, Pfenninger EG, Lurie KG, Goertz A, Lindner IM. Effect of vasopressin on hemodynamic variables, organ blood flow, and acid-base status in a pig model of cardiopulmonary resuscitation. Anesth Analg 1993;77:427–35. [21] Lindner KH, Prengel AW, Pfenninger EG, Lindner IM, Strohmenger HU, Georgieff M, et al. Vasopressin improves vital organ blood flow during closed-chest cardiopulmonary resuscitation in pigs. Circulation 1995;91:215–21. [22] Pytte M, Kramer-Johansen J, Eilevstjønn J, Eriksen M, Strømme TA, Godang K, et al. Haemodynamic effects of adrenaline (epinephrine) depend on chest compression quality during cardiopulmonary resuscitation in pigs. Resuscitation 2006;71:369–78. [23] Tømte Ø, Sjaastad I, Wik L, Kuzovlev A, Eriksen M, Norseng PA, et al. Discriminating the effect of accelerated compression from accelerated decompression during high-impulse CPR in a porcine model of cardiac arrest. Resuscitation 2010;81:488–92. [24] Anderson KL, Castaneda MG, Boudreau SM, Sharon DJ, Bebarta VS. Left ventricular compressions improve hemodynamics in a swine model of out-of-hospital cardiac arrest. Prehosp Emerg Care 2017;21:272–80.
Please cite this article in press as: Jung YH, et al. Relationship Between Left Ventricle Position and Haemodynamic Parameters During Cardiopulmonary Resuscitation in a Pig Model. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j. hlc.2017.08.020
425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468