Acta Anaesthesiol Scand 2014; 58: 1032–1039 Printed in Singapore. All rights reserved
© 2014 The Acta Anaesthesiologica Scandinavica Foundation. Published by John Wiley & Sons Ltd ACTA ANAESTHESIOLOGICA SCANDINAVICA
doi: 10.1111/aas.12378
Ventilation/perfusion ratios measured by multiple inert gas elimination during experimental cardiopulmonary resuscitation E. K. Hartmann1, B. Duenges1, S. Boehme1,2, M. Szczyrba1, T. Liu1, K. U. Klein1,2, J. E. Baumgardner3, K. Markstaller1,2 and M. David1
1 Department of Anaesthesiology, Medical Centre of the Johannes Gutenberg-University, Mainz, Germany, 2Department of Anaesthesia, General Critical Care Medicine and Pain Therapy, Medical University of Vienna, Vienna, Austria and 3Oscillogy LLC, Folsom, PA, USA
Background: During cardiopulmonary resuscitation (CPR) the ventilation/perfusion distribution (VA/Q) within the lung is difficult to assess. This experimental study examines the capability of multiple inert gas elimination (MIGET) to determine VA/Q under CPR conditions in a pig model. Methods: Twenty-one anaesthetised pigs were randomised to three fractions of inspired oxygen (1.0, 0.7 or 0.21). VA/Q by micropore membrane inlet mass spectrometry-derived MIGET was determined at baseline and during CPR following induction of ventricular fibrillation. Haemodynamics, blood gases, ventilation distribution by electrical impedance tomography and return of spontaneous circulation were assessed. Intergroup differences were analysed by non-parametric testing. Results: MIGET measurements were feasible in all animals with an excellent correlation of measured and predicted arterial oxygen partial pressure (R2 = 0.96, n = 21 for baseline; R2 = 0.82, n = 21 for CPR). CPR induces a significant shift from normal VA/Q ratios to the high VA/Q range. Electrical impedance
tomography indicates a dorsal to ventral shift of the ventilation distribution. Diverging pulmonary shunt fractions induced by the three inspired oxygen levels considerably increased during CPR and were traceable by MIGET, while 100% oxygen most negatively influenced the VA/Q. Return of spontaneous circulation were achieved in 52% of the animals. Conclusions: VA/Q assessment by MIGET is feasible during CPR and provides a novel tool for experimental purposes. Changes in VA/Q caused by different oxygen fractions are traceable during CPR. Beyond pulmonary perfusion deficits, these data imply an influence of the inspired oxygen level on VA/Q. Higher oxygen levels significantly increase shunt fractions and impair the normal VA/Q ratio.
T
tribution (VA/Q) mismatching. CPR further increases the complexity because of deteriorated perfusion going along with impaired ventilation, which results from interaction of chest compressions and mechanical ventilation. Few experimental technologies are available to assess the impact of these changes during CPR. The multiple inert gas elimination technique (MIGET) is the experimental gold standard for assessment of the VA/Q.6,7 The micropore membrane inlet mass spectrometry (MMIMS) is a novel method of MIGET that allows for rapid and standardised analysis with relatively small blood samples, and was recently referenced to the standard gas chromatographic method.6 The MMIMS-MIGET has been further validated against the calculated venous admixture and
he role of mechanical ventilation during cardiopulmonary resuscitation (CPR) is still controversially discussed.1 Several experimental models have suggested a benefit for intermittent positive pressure ventilation during CPR. Beneficial effects include improved systemic and tissue oxygenation as well as less atelectasis formation.2,3 Furthermore, improvement of neurological outcome and overall pulmonary function after return of spontaneous circulation (ROSC) has been reported.4,5 There is no clear evidence, however, regarding different ventilator modes and patterns. Gas exchange in general is assured by the matching of adequate ventilation and perfusion. In some primary pulmonary diseases, for example asthma, alterations in ventilation represent the main determinate of ventilation/perfusion dis-
1032 bs_bs_banner
Accepted for publication 01 July 2014 © 2014 The Acta Anaesthesiologica Scandinavica Foundation. Published by John Wiley & Sons Ltd
VA/Q during cardiopulmonary resuscitation
blood gas analysis.8,9 MIGET, however, has not been used under severely abnormal haemodynamic conditions like CPR. We hypothesised that the MMIMS-MIGET detects alterations of pulmonary gas exchange and shunt variations induced by CPR under different fractions of inspired oxygen (FiO2). The present study therefore examines two main aspects under experimental CPR: (1) feasibility and face validity of MMIMSMIGET as tool to assess VA/Q and (2) the influence of different FiO2 levels on the VA/Q in a pig model.
Methods Following approval of the state and institutional animal care committee (Landesuntersuchungsamt Rheinland-Pfalz, approval number 23-177-07/G1001-021) a convenience sample of 21 juvenile pigs (25–27 kg) was examined in a prospective, randomised setting.
Anaesthesia and instrumentation The animals were delivered following sedation by intramuscular injection of ketamine and midazolam. After vascular access, anaesthesia was induced and continuously maintained (propofol 8–12 mg/kg/h, fentanyl 0.1–0.2 mg/kg/h). A single dose of atracurium (0.5 mg/kg) was administered for endotracheal intubation. The animals were ventilated in volume-controlled mode targeted to tidal volume of 10 ml/kg, positive endexpiratory pressure of 5 cmH2O, FiO2 of 40% and a variable ventilator frequency adjusted to an endtidal carbon dioxide level < 6 kPa. Vascular catheters were placed by surgical cut-down: an arterial line, a pulse contour cardiac output catheter (PiCCO, Pulsion Medical, Munich, Germany) and a central venous line via femoral access. A pulmonary artery catheter and an introducer for a cardiac pacing electrode (Osypka Medical GmbH, Rheinfelden-Herten, Germany) were placed via the jugular veins. All animals received a balanced electrolyte solution with an infusion rate of 5 ml/kg/h. Haemodynamics and ventilatory data were continuously monitored and stored (S/5 Collect, Datex Ohmeda, GE Healthcare, Munich, Germany). Transpulmonary thermodilution by PiCCO was performed in triplicate according the manufacturer’s instructions.
VA/Q measurements and ventilation distribution The MIGET technology by MMIMS (Oscillogy LLC, Folsom, PA, USA) was applied after continuous
infusion of a saline solution containing the six inert gases (sulphur hexafluoride, krypton, desflurane, enflurane, diethyl ether and acetone) in subclinical and non-toxic dosages starting 30 min before cardiac arrest. Sampling procedure, data postprocessing and prediction of the arterial partial pressure of oxygen (PaO2) by MMIMS-MIGET have been described in detail by our group.8–10 Measurements were carried out in a healthy state and during ongoing CPR. Additionally, a 16-electrode electrical impedance tomography (EIT) system (Goe-MF II, CareFusion, San Diego, CA, USA) was used for analysis of thoracic bioimpedance changes associated with regional pulmonary aeration. The electrodes were placed on a transverse lung section approximately 10 cm above the diaphragm. Recordings were performed during baseline and CPR. The regional ventilation distribution was examined for the ventral, middle and dorsal lung regions.11 Each region’s relative impedance change is expressed as percentage of the global tidal amplitude.
Experimental protocol After instrumentation, a non-participant randomised the animals into three groups after drawing from a pool of 21 prepared envelopes containing the particular FiO2: (1) FiO2 100% during CPR (n = 7) (2) FiO2 70% during CPR (n = 7) (3) FiO2 21% during CPR (n = 7) Volume-constant ventilation was maintained using the previous settings. The FiO2 was set according to the randomisation 10 min before cardiac arrest induction. Baseline measurements were then performed and recorded. Ventricular fibrillation was induced by 13.8 volt current via the cardiac pacing catheter according to the manufacturer’s recommendations and verified by electrocardiogram and invasive blood pressure. The pacing catheter was removed subsequently. After 1 min of untreated cardiac arrest, chest compressions were started with an external device (Thumper 12715 Programmable CPR Controller, Michigan Instruments, Grand Rapids, MI, USA; 100 compressions/min, setup as reported by Markstaller et al.3) and nonsynchronised mechanical ventilation was restarted. Haemodynamics were recorded continuously during cardiac arrest. Arterial and mixed venous blood gas analysis (Rapidlab 248, Bayer Healthcare, Leverkusen, Germany) as well as blood withdrawal for the MMIMS-MIGET were performed synchronously within 6–8 min after the cardiac arrest. Eight
1033
E. K. Hartmann et al.
minutes after cardiac arrest (after 7 min of CPR), the heart rhythm was analysed. If ventricular fibrillation persisted, a single biphasic shock at 200 J was delivered (LIFEPAK 20, Physio-Control, Neuss, Germany) subsequently followed by chest compressions. Epinephrine administration was started by a single intravenous bolus of 40 μg/kg via the central venous line and followed by a continuous infusion of epinephrine at a rate of 10 μg/kg/min. Rhythm analyses and biphasic defibrillation were then repeated every 2 min until ROSC was achieved or a total of three shocks were applied. The experiment was stopped in animals without ROSC. Animals that achieved ROSC remained anaesthetised and after finishing the protocol were euthanised in deep general anaesthesia by potassium chloride administration.
Statistics Data are reported as median and interquartile range. The association between predicted PaO2 by MMIMS-MIGET and the measured value was analysed using linear regression and the Bland–Altman method. MIGET and EIT data during CPR were compared with their baseline values by the Wilcoxon rank test with correction for multiple comparisons using the Bonferroni method. Three-group comparisons were drawn by Kruskal–Wallis test with post-hoc Student–Newman–Keuls test. P values < 0.05 were regarded as significant.
Results A total of 21 animal experiments were conducted. The MMIMS-MIGET measurements in CPR were technically feasible in all animals. The overall linear correlation between the measured and the predicted PaO2 by MMIMS-MIGET was excellent (R2 = 0.930, P < 0.001, n = 42 measurements). Overall, the MMIMS-MIGET slightly underestimated the measured PaO2 at high values (10.29 + 0.88 × measured PaO2), an effect that was most pronounced in the CPR analysis. Figure 1 shows the separate correlations during baseline and CPR (Baseline: predicted = 9.30 + 0.91 × measured PaO2; CPR: predicted = 25.95 + 0.73 × measured PaO2). A Bland–Altman analysis yielded a mean difference of 21 mmHg (limits of agreement −108 to 66 mmHg) in the baseline and 18 mmHg (limits of agreement −140 to 103 mmHg) during CPR. Residual sum of squares (RSS) as an indicator of experimental error were 1.0 (0.3–2.2) during baseline and 2.1 (0.1–10) during CPR. Eleven of the
1034
21 animals achieved ROSC criteria (FiO2 0.21: 43%; FiO2 0.7: 71%; FiO2 1.0: 43%). Table 1 shows the key haemodynamic and blood gas data and indicates considerable differences in PaO2 as induced by the different FiO2 levels. Figure 2 summarises the group-wise VA/Q ratios. The baseline data depict the healthy conditions with a small shunt fraction in anaesthetised subjects. During CPR, a significant increase of the pulmonary shunt fraction occurs, while the normal VA/Q ratios considerably shift to high VA/Q. During baseline, significantly lower shunt and corresponding higher normal VA/Q fractions occurred in the FiO2 0.21– group, while in the intergroup comparisons, the highest shunt fractions were found under FiO2 1.0 conditions. In CPR, the shift to high VA/Q ratios was similar in all groups. However, shunt fractions were significantly higher when FiO2 1.0 was compared with the two other groups. Only minimal low VA/Q ratios occurred in all groups and time points (data not shown). Due to technical reasons in the challenging setting of CPR and chest compressions, exploitable EIT recordings were available from five animals per group each. Accordingly, we did not draw intergroup comparisons. The summarised data (n = 15, Fig. 3) show a significant shift of the regional distribution as measured by the amplitudes of thoracic bioimpedance variations from the dorsal and middle to the ventral lung region. In an additional subanalysis, the summed impedance changes from the middle and dorsal region significantly decreased, corresponding to the increase in the ventral region.
Discussion Methodical aspects The present study analyses the feasibility and face validity of MIGET-derived VA/Q measurements under highly deteriorated CPR conditions. In general, few technologies allow for repetitive assessment of VA/Q in experimental research or clinical routine. The MMIMS is a relatively novel MIGET device, which was developed to simplify and enhance VA/Q measurements. As a perfusiondependent technology, the MIGET, based on pulmonary elimination or retention of the applied inert gases, requires adequate haemodynamics and pulmonary perfusion. Loeckinger et al. conducted MIGET-derived VA/Q measurements in early stable ROSC in a porcine model and reported persisting impairment in terms of increased shunt and low VA/Q units over 2 h.12 To our knowledge, MIGET
VA/Q during cardiopulmonary resuscitation
Fig. 1. Partial pressure of oxygen (PaO2) prediction by micropore membrane inlet mass spectrometry-multiple inert gas elimination (MMIMS-MIGET): linear regression between measured PaO2 by blood gas analysis and the predicted value by MIGET (upper graphs), and corresponding Bland–Altman plots (lower graphs) during baseline and cardiopulmonary resuscitation (CPR) (n = 21 each).
never has been applied under CPR conditions. In our study, blood withdrawal and analysis was feasible in all animals. Comparison with an external reference by means of measured and predicted PaO26,9 during CPR yielded a strong linear correlation (Fig. 1). At baseline, there was a small underestimation of measured PaO2 at high PaO2, a trend that was more pronounced during CPR. At low values, pre-
dicted PaO2 was slightly higher than measured PaO2 during CPR, similar to baseline. The small difference at baseline between predicted and measured PaO2 at low PaO2 is similar to results from prior studies.6,13 Furthermore, this difference increased only slightly with CPR, which suggests that CPR does not lead to a considerable diffusion limitation of oxygen transfer.7 The time span to achieve a new
1035
E. K. Hartmann et al. Table 1 Haemodynamics and oxygenation data. Parameter
FiO2
Baseline
Arrest
CPR
MAP (mmHg)
0.21 0.7 1.0 0.21 0.7 1.0 0.21 0.7 1.0 0.21 0.7 1.0 0.21 0.7 1.0 0.21 0.7 1.0
106 (33) 92 (35) 110 (16) 5.1 (2.0)* 3.9 (2.1) 3.9 (1.4) 100 (24) 89 (39) 85 (13) 107 (16)* 374 (96)* 594 (211)* 510 (78) 534 (137) 594 (211) 73 (5)* 78 (6) 84 (10)
28 (9) 23 (9) 30 (9) – – – – – – – – – – – – – – –
44 (5) 44 (16) 47 (8) 1.0 (0.4) 0.8 (0.5) 0.8 (0.3) 100 (1) 100 (7) 100 (21) 63 (17)† 94 (261)† 175 (194)† 299 (80) 134 (373) 175 (194) 20 (6) 23 (19) 32 (24)
CO (l/min) HR (per minute) PaO2 (mmHg) PaO2/FiO2 (mmHg) SvO2 (%)
*P < 0.05 in three-group comparison at baseline. †P < 0.05 in three-group comparison during CPR, n = 7 per group. MAP, mean arterial pressure; CO, cardiac output; HR, heart rate or chest compression frequency; PaO2, arterial partial pressure of oxygen; FiO2, fraction of inspired oxygen; SvO2, mixed venous oxygen saturation.
steady state of ventilation and perfusion following cardiac arrest and initiation of CPR is not known and influenced by multiple factors. Traditional calculations suggest that 95% equilibrium to a new steady state adjusts within 2 min.14,15 The chosen interval of 6–8 min following cardiac arrest should be adequate in this context. Nevertheless, the exact time constants during CPR need to examined in further studies. RSS values as indicators of experimental failure were well below the thresholds that were defined for gas chromatographic MIGET measurements (RSS < 5.3 in 50%, < 10.6 in 90% and < 16.8 in 99% of the measurements16). The baseline VA/Q ratios further correspond well to previous data from healthy, comparable-sized pigs (mean normal VA/Q fraction 90–96%10,17). A high FiO2 leads to the formation of atelectasis in anaesthetised subjects.18,19 The small but significant shunt variations between the three groups with the lowest shunt during FiO2 0.21 could be explained by this concept. However, the measured changes did not lead to differences in oxygenation in healthy pigs with intact hypoxic pulmonary vasoconstriction.
Impact of CPR on VA/Q ratios A considerable shift from normal to high VA/Q is induced by CPR, which approximately equals the
1036
Fig. 2. Group-wise ventilation/perfusion distribution (VA/Q) ratios during baseline and cardiopulmonary resuscitation (CPR). Only minimal low VA/Q ratios (data not shown). Standard VA/Q ratios: shunt (VA/Q < 0.005), low VA/Q (0.005 < VA/Q < 0.1), normal VA/Q (0.1 < VA/Q < 10) and high VA/Q (10 < VA/ Q > 100). #P < 0.05 vs. baseline; *P < 0.05 in intergroup comparison, n = 7 per group. FiO2, fractions of inspired oxygen.
amount of persisting normal VA/Q ratios. A similar shift of the indirectly estimated global VA/Q in CPR has been previously reported.20 High VA/Q ratios imply two possible explanations: pulmonary hyperventilation or impaired perfusion. The overall VA/Q ratio of the main mode reflects the ratio of ventilation to effective perfusion, particularly the ventila-
VA/Q during cardiopulmonary resuscitation
Fig. 3. Regional ventilation distribution by electrical impedance tomography during baseline and cardiopulmonary resuscitation (CPR) (upper graph). Addition of the middle and dorsal lung region (lower graph). *P < 0.05.
tion to non-shunting blood flow. During CPR, pulmonary perfusion is severely impaired, although not annihilated (mean arterial pressure > 40 mmHg in our model). A high VA/Q shift related to perfusion deficits was also found in pig model of haemorrhagic shock that featured similarly low mean arterial pressure values.21 EIT measurements yielded significantly increased amplitudes in the ventral lung regions and a dorsal to ventral shift. In general, EIT data can be post-processed to discriminate ventilation from perfusion-related variations by band-pass filtering and Fourier analysis.22 Compression frequencies in a range of 100/min, however, clearly overlap with the perfusion signals. Also, the respiratory pattern is compromised by nonsynchronised ventilation.23 Chest compressions cause active compression/decompression-related changes in lung aeration and ventilation
inhomogeneity leading to ventral pulmonary hyperinflation and atelectasis formation in the dependent lung areas.3 In this context, Markstaller et al. reported up to 7% of overdistended lung tissue in a computer tomography-based analysis during CPR in pigs.3 These findings also influence the high VA/Q and shunt fractions; however, the ventilation to non-effectively perfused lung areas mainly seem to determine the mode of VA/Q mismatch.
Effect of FiO2 In CPR, significantly higher shunt fractions were observed during FiO2 0.7 and particularly FiO2 1.0 ventilation, which therefore seems to aggravate the pre-existing atelectatic lung areas. Gas exchange during CPR is influenced by the applied mechanical ventilation as well as chest compressions. Considerable impairment of the pulmonary function in
1037
E. K. Hartmann et al.
ROSC has been reported by several studies.4,12,24 Our data, in this context, indicate a relevant impact of FiO2 level on atelectasis formation already during CPR. Ongoing CPR, which is characterised by minimal perfusion and oxygen delivery conditions, should be differentiated from the ROSC-induced reperfusion phase, in which it is widely accepted to avoid FiO2 levels leading to hyperoxia.1,25,26 FiO2related differences of the shunt, but not of the high VA/Q ratios, the distribution of bioimpedance variations by EIT and comparable haemodynamics in the three groups (Table 1) imply that beyond perfusion, the FiO2 may considerably contribute to the VA/Q deterioration. Despite the highest overall PaO2 values during CPR, the FiO2 1.0 ventilation most negatively influences the VA/Q pattern. Furthermore, hypoxic pulmonary vasoconstriction may influence these findings. A prior study assumed that during CPR pure oxygen but not air ventilation may inhibit hypoxic pulmonary vasoconstriction.27 As noted, signs of lung injury following CPR have been reported in several studies, although these aspects lie beyond scope of this study. Nevertheless, our data indicate that impaired lung function immediately occurs during ongoing CPR and is not exclusively related to the reperfusion phase.
Limitations of the study The present study has several limitations. Because there were no preceding data for a power analysis, we chose a convenience sample of seven animals per group, which was oriented on previous publications from our group.24 The particular variant of MIGET measurement used in our laboratory requires cardiac output data, which is difficult during CPR. The PiCCO technology was used for this purpose, which was previously reported to provide reliable measurements and significant correlations to blood pressure during CPR in pigs.28,29 This pilot CPR study is limited to face validity, as assessed by successful comparison with an external reference (PaO2) and acceptable ranges of experimental error (RSS values). We did neither perform repeated measurements to examine retest reliability nor analysis of responses after a change in cardiac output. Hence, these distinct aspects of validation need to be addressed in future experiments. The study was neither designed nor powered to provide long-term ROSC or post-mortem data, but strictly focused on the feasibility of VA/Q measurements and the acute or FiO2-related changes during CPR. Additionally, in our experimental setting, the different FiO2 levels were initiated 10 min before induc-
1038
tion of arrest rather than on initiation of resuscitation. A model inevitably differs from the clinical scenario regarding brief and uniform performance of CPR, the defined intervals of ventilation and the standardised measurement procedure. Nevertheless, the reported overall ROSC rates (52%) imply a realistic scenario.
Conclusion The assessment of VA/Q by MIGET is feasible even during CPR and shows face validity. VA/Q measurements correlate well to the conventional blood gas analysis and FiO2-related changes were traceable in this setting. MMIMS-MIGET represents a novel tool for experimental purposes and provides further insights in the complex interactions of ventilation and perfusion and their influence upon the pulmonary function during CPR. Additional studies might therefore focus on the best-suited oxygen fraction regarding the VA/Q ratio and its impact on pulmonary as well as global outcome.
Acknowledgements All experiments were performed in the laboratories of the Department of Anaesthesiology, Medical Centre of the Johannes Gutenberg-University Mainz. Conflicts of interest: J. E. Baumgardner, the owner of Oscillogy LLC, which commercially distributes the MMIMS-MIGET, participated in the study setup and provided technical support, although he had no influence on performance of the experiments and data analysis. None of the other authors declares a conflict of interest. Funding: The study was funded in part by the German Research Council DFG PAK 415-1.
References 1. Kill C, Dersch W, Wulf H. Advanced life support and mechanical ventilation. Curr Opin Crit Care 2012; 18: 251–5. 2. Kill C, Torossian A, Freisburger C, Dworok S, Massmann M, Nohl T, Henning R, Wallot P, Gockel A, Steinfeldt T, Graf J, Eberhart L, Wulf H. Basic life support with four different compression/ventilation ratios in a pig model: the need for ventilation. Resuscitation 2009; 80: 1060–5. 3. Markstaller K, Karmrodt J, Doebrich M, Wolcke B, Gervais H, Weiler N, Thelen M, Dick W, Kauczor HU, Eberle B. Dynamic computed tomography: a novel technique to study lung aeration and atelectasis formation during experimental CPR. Resuscitation 2002; 53: 307–13. 4. Wang S, Wu JY, Guo ZJ, Li CS. Effect of rescue breathing during cardiopulmonary resuscitation on lung function after restoration of spontaneous circulation in a porcine model of prolonged cardiac arrest. Crit Care Med 2013; 41: 102–10. 5. Yannopoulos D, Matsuura T, McKnite S, Goodman N, Idris A, Tang W, Aufderheide TP, Lurie KG. No assisted ventilation cardiopulmonary resuscitation and 24-hour neurological outcomes in a porcine model of cardiac arrest. Crit Care Med 2010; 38: 254–60.
VA/Q during cardiopulmonary resuscitation 6. Kretzschmar M, Schilling T, Vogt A, Rothen HU, Borges JB, Hachenberg T, Larsson A, Baumgardner JE, Hedenstierna G. Multiple inert gas elimination technique by micropore membrane inlet mass spectrometry – a comparison with reference gas chromatography. J Appl Physiol 2013; 115: 1107–18. 7. Wagner PD. The multiple inert gas elimination technique (MIGET). Intensive Care Med 2008; 34: 994–1001. 8. Duenges B, Vogt A, Bodenstein M, Wang H, Bohme S, Rohrig B, Baumgardner JE, Markstaller K. A comparison of micropore membrane inlet mass spectrometry-derived pulmonary shunt measurement with Riley shunt in a porcine model. Anesth Analg 2009; 109: 1831–5. 9. Hartmann EK, Duenges B, Baumgardner JE, Markstaller K, David M. Correlation of thermodilution-derived extravascular lung water and ventilation/perfusioncompartments in a porcine model. Intensive Care Med 2013; 39: 1313–7. 10. Hartmann EK, Boehme S, Bentley A, Duenges B, Klein KU, Elsaesser A, Baumgardner JE, David M, Markstaller K. Influence of respiratory rate and end-expiratory pressure variation on cyclic alveolar recruitment in an experimental lung injury model. Crit Care 2012; 16: R8. 11. Bodenstein M, Wang H, Boehme S, Vogt A, Kwiecien R, David M, Markstaller K. Influence of crystalloid and colloid fluid infusion and blood withdrawal on pulmonary bioimpedance in an animal model of mechanical ventilation. Physiol Meas 2012; 33: 1225–36. 12. Loeckinger A, Kleinsasser A, Wenzel V, Mair V, Keller C, Kolbitsch C, Recheis W, Schuster A, Lindner KH. Pulmonary gas exchange after cardiopulmonary resuscitation with either vasopressin or epinephrine. Crit Care Med 2002; 30: 2059–62. 13. Wagner PD, Gale GE, Moon RE, Torre-Bueno JR, Stolp BW, Saltzman HA. Pulmonary gas exchange in humans exercising at sea level and simulated altitude. J Appl Physiol 1986; 61: 260–70. 14. Hopkins SR, Gavin TP, Siafakas NM, Haseler LJ, Olfert IM, Wagner H, Wagner PD. Effect of prolonged, heavy exercise on pulmonary gas exchange in athletes. J Appl Physiol 1998; 85: 1523–32. 15. Hopkins SR, Bayly WM, Slocombe RF, Wagner H, Wagner PD. Effect of prolonged heavy exercise on pulmonary gas exchange in horses. J Appl Physiol 1998; 84: 1723–30. 16. Roca J, Wagner PD. Contribution of multiple inert gas elimination technique to pulmonary medicine. 1. Principles and information content of the multiple inert gas elimination technique. Thorax 1994; 49: 815–24. 17. Neumann P, Hedenstierna G. Ventilation-perfusion distributions in different porcine lung injury models. Acta Anaesthesiol Scand 2001; 45: 78–86. 18. Dantzker DR, Wagner PD, West JB. Instability of poorly ventilated lung units during oxygen breathing. J Appl Physiol 1975; 38: 886–95. 19. Rothen HU, Sporre B, Engberg G, Wegenius G, Reber A, Hedenstierna G. Prevention of atelectasis during general anaesthesia. Lancet 1995; 345: 1387–91.
20. Wang S, Li C, Ji X, Yang L, Su Z, Wu J. Effect of continuous compressions and 30 : 2 cardiopulmonary resuscitation on global ventilation/perfusion values during resuscitation in a porcine model. Crit Care Med 2010; 38: 2024–30. 21. Robinson NB, Chi EY, Robertson HT. Ventilation-perfusion relationships after hemorrhage and resuscitation: an inert gas analysis. J Appl Physiol Respir Environ Exerc Physiol 1983; 54: 1131–40. 22. Bodenstein M, David M, Markstaller K. Principles of electrical impedance tomography and its clinical application. Crit Care Med 2009; 37: 713–24. 23. Kill C, Hahn O, Dietz F, Neuhaus C, Schwarz S, Mahling R, Wallot P, Jerrentrup A, Steinfeldt T, Wulf H, Dersch W. Mechanical ventilation during cardiopulmonary resuscitation with intermittent positive-pressure ventilation, bilevel ventilation, or chest compression synchronized ventilation in a pig model. Crit Care Med 2014; 42: e89–95. 24. Markstaller K, Rudolph A, Karmrodt J, Gervais HW, Goetz R, Becher A, David M, Kempski OS, Kauczor HU, Dick WF, Eberle B. Effect of chest compressions only during experimental basic life support on alveolar collapse and recruitment. Resuscitation 2008; 79: 125–32. 25. Kilgannon JH, Jones AE, Parrillo JE, Dellinger RP, Milcarek B, Hunter K, Shapiro NI, Trzeciak S. Relationship between supranormal oxygen tension and outcome after resuscitation from cardiac arrest. Circulation 2011; 123: 2717–22. 26. Kilgannon JH, Jones AE, Shapiro NI, Angelos MG, Milcarek B, Hunter K, Parrillo JE, Trzeciak S. Association between arterial hyperoxia following resuscitation from cardiac arrest and in-hospital mortality. JAMA 2010; 303: 2165–71. 27. Rubertsson S, Karlsson T, Wiklund L. Systemic oxygen uptake during experimental closed-chest cardiopulmonary resuscitation using air or pure oxygen ventilation. Acta Anaesthesiol Scand 1998; 42: 32–8. 28. Lopez-Herce J, Fernandez B, Urbano J, Mencia S, Solana MJ, Del Castillo J, Rodriguez-Nunez A, Bellon JM, Carrillo A. Correlations between hemodynamic, oxygenation and tissue perfusion parameters during asphyxial cardiac arrest and resuscitation in a pediatric animal model. Resuscitation 2011; 82: 755–9. 29. Lopez-Herce J, Fernandez B, Urbano J, Mencia S, Solana MJ, Rodriguez-Nunez A, Bellon JM, Carrillo A. Hemodynamic, respiratory, and perfusion parameters during asphyxia, resuscitation, and post-resuscitation in a pediatric model of cardiac arrest. Intensive Care Med 2011; 37: 147–55.
Address: Erik K. Hartmann Department of Anaesthesiology Medical Centre of the Johannes Gutenberg-University Langenbeckstraße 1 55131 Mainz Germany e-mail: [email protected]
1039
© 2014 The Acta Anaesthesiologica Scandinavica Foundation. Published by John Wiley & Sons Ltd ACTA ANAESTHESIOLOGICA SCANDINAVICA
doi: 10.1111/aas.12378
Ventilation/perfusion ratios measured by multiple inert gas elimination during experimental cardiopulmonary resuscitation E. K. Hartmann1, B. Duenges1, S. Boehme1,2, M. Szczyrba1, T. Liu1, K. U. Klein1,2, J. E. Baumgardner3, K. Markstaller1,2 and M. David1
1 Department of Anaesthesiology, Medical Centre of the Johannes Gutenberg-University, Mainz, Germany, 2Department of Anaesthesia, General Critical Care Medicine and Pain Therapy, Medical University of Vienna, Vienna, Austria and 3Oscillogy LLC, Folsom, PA, USA
Background: During cardiopulmonary resuscitation (CPR) the ventilation/perfusion distribution (VA/Q) within the lung is difficult to assess. This experimental study examines the capability of multiple inert gas elimination (MIGET) to determine VA/Q under CPR conditions in a pig model. Methods: Twenty-one anaesthetised pigs were randomised to three fractions of inspired oxygen (1.0, 0.7 or 0.21). VA/Q by micropore membrane inlet mass spectrometry-derived MIGET was determined at baseline and during CPR following induction of ventricular fibrillation. Haemodynamics, blood gases, ventilation distribution by electrical impedance tomography and return of spontaneous circulation were assessed. Intergroup differences were analysed by non-parametric testing. Results: MIGET measurements were feasible in all animals with an excellent correlation of measured and predicted arterial oxygen partial pressure (R2 = 0.96, n = 21 for baseline; R2 = 0.82, n = 21 for CPR). CPR induces a significant shift from normal VA/Q ratios to the high VA/Q range. Electrical impedance
tomography indicates a dorsal to ventral shift of the ventilation distribution. Diverging pulmonary shunt fractions induced by the three inspired oxygen levels considerably increased during CPR and were traceable by MIGET, while 100% oxygen most negatively influenced the VA/Q. Return of spontaneous circulation were achieved in 52% of the animals. Conclusions: VA/Q assessment by MIGET is feasible during CPR and provides a novel tool for experimental purposes. Changes in VA/Q caused by different oxygen fractions are traceable during CPR. Beyond pulmonary perfusion deficits, these data imply an influence of the inspired oxygen level on VA/Q. Higher oxygen levels significantly increase shunt fractions and impair the normal VA/Q ratio.
T
tribution (VA/Q) mismatching. CPR further increases the complexity because of deteriorated perfusion going along with impaired ventilation, which results from interaction of chest compressions and mechanical ventilation. Few experimental technologies are available to assess the impact of these changes during CPR. The multiple inert gas elimination technique (MIGET) is the experimental gold standard for assessment of the VA/Q.6,7 The micropore membrane inlet mass spectrometry (MMIMS) is a novel method of MIGET that allows for rapid and standardised analysis with relatively small blood samples, and was recently referenced to the standard gas chromatographic method.6 The MMIMS-MIGET has been further validated against the calculated venous admixture and
he role of mechanical ventilation during cardiopulmonary resuscitation (CPR) is still controversially discussed.1 Several experimental models have suggested a benefit for intermittent positive pressure ventilation during CPR. Beneficial effects include improved systemic and tissue oxygenation as well as less atelectasis formation.2,3 Furthermore, improvement of neurological outcome and overall pulmonary function after return of spontaneous circulation (ROSC) has been reported.4,5 There is no clear evidence, however, regarding different ventilator modes and patterns. Gas exchange in general is assured by the matching of adequate ventilation and perfusion. In some primary pulmonary diseases, for example asthma, alterations in ventilation represent the main determinate of ventilation/perfusion dis-
1032 bs_bs_banner
Accepted for publication 01 July 2014 © 2014 The Acta Anaesthesiologica Scandinavica Foundation. Published by John Wiley & Sons Ltd
VA/Q during cardiopulmonary resuscitation
blood gas analysis.8,9 MIGET, however, has not been used under severely abnormal haemodynamic conditions like CPR. We hypothesised that the MMIMS-MIGET detects alterations of pulmonary gas exchange and shunt variations induced by CPR under different fractions of inspired oxygen (FiO2). The present study therefore examines two main aspects under experimental CPR: (1) feasibility and face validity of MMIMSMIGET as tool to assess VA/Q and (2) the influence of different FiO2 levels on the VA/Q in a pig model.
Methods Following approval of the state and institutional animal care committee (Landesuntersuchungsamt Rheinland-Pfalz, approval number 23-177-07/G1001-021) a convenience sample of 21 juvenile pigs (25–27 kg) was examined in a prospective, randomised setting.
Anaesthesia and instrumentation The animals were delivered following sedation by intramuscular injection of ketamine and midazolam. After vascular access, anaesthesia was induced and continuously maintained (propofol 8–12 mg/kg/h, fentanyl 0.1–0.2 mg/kg/h). A single dose of atracurium (0.5 mg/kg) was administered for endotracheal intubation. The animals were ventilated in volume-controlled mode targeted to tidal volume of 10 ml/kg, positive endexpiratory pressure of 5 cmH2O, FiO2 of 40% and a variable ventilator frequency adjusted to an endtidal carbon dioxide level < 6 kPa. Vascular catheters were placed by surgical cut-down: an arterial line, a pulse contour cardiac output catheter (PiCCO, Pulsion Medical, Munich, Germany) and a central venous line via femoral access. A pulmonary artery catheter and an introducer for a cardiac pacing electrode (Osypka Medical GmbH, Rheinfelden-Herten, Germany) were placed via the jugular veins. All animals received a balanced electrolyte solution with an infusion rate of 5 ml/kg/h. Haemodynamics and ventilatory data were continuously monitored and stored (S/5 Collect, Datex Ohmeda, GE Healthcare, Munich, Germany). Transpulmonary thermodilution by PiCCO was performed in triplicate according the manufacturer’s instructions.
VA/Q measurements and ventilation distribution The MIGET technology by MMIMS (Oscillogy LLC, Folsom, PA, USA) was applied after continuous
infusion of a saline solution containing the six inert gases (sulphur hexafluoride, krypton, desflurane, enflurane, diethyl ether and acetone) in subclinical and non-toxic dosages starting 30 min before cardiac arrest. Sampling procedure, data postprocessing and prediction of the arterial partial pressure of oxygen (PaO2) by MMIMS-MIGET have been described in detail by our group.8–10 Measurements were carried out in a healthy state and during ongoing CPR. Additionally, a 16-electrode electrical impedance tomography (EIT) system (Goe-MF II, CareFusion, San Diego, CA, USA) was used for analysis of thoracic bioimpedance changes associated with regional pulmonary aeration. The electrodes were placed on a transverse lung section approximately 10 cm above the diaphragm. Recordings were performed during baseline and CPR. The regional ventilation distribution was examined for the ventral, middle and dorsal lung regions.11 Each region’s relative impedance change is expressed as percentage of the global tidal amplitude.
Experimental protocol After instrumentation, a non-participant randomised the animals into three groups after drawing from a pool of 21 prepared envelopes containing the particular FiO2: (1) FiO2 100% during CPR (n = 7) (2) FiO2 70% during CPR (n = 7) (3) FiO2 21% during CPR (n = 7) Volume-constant ventilation was maintained using the previous settings. The FiO2 was set according to the randomisation 10 min before cardiac arrest induction. Baseline measurements were then performed and recorded. Ventricular fibrillation was induced by 13.8 volt current via the cardiac pacing catheter according to the manufacturer’s recommendations and verified by electrocardiogram and invasive blood pressure. The pacing catheter was removed subsequently. After 1 min of untreated cardiac arrest, chest compressions were started with an external device (Thumper 12715 Programmable CPR Controller, Michigan Instruments, Grand Rapids, MI, USA; 100 compressions/min, setup as reported by Markstaller et al.3) and nonsynchronised mechanical ventilation was restarted. Haemodynamics were recorded continuously during cardiac arrest. Arterial and mixed venous blood gas analysis (Rapidlab 248, Bayer Healthcare, Leverkusen, Germany) as well as blood withdrawal for the MMIMS-MIGET were performed synchronously within 6–8 min after the cardiac arrest. Eight
1033
E. K. Hartmann et al.
minutes after cardiac arrest (after 7 min of CPR), the heart rhythm was analysed. If ventricular fibrillation persisted, a single biphasic shock at 200 J was delivered (LIFEPAK 20, Physio-Control, Neuss, Germany) subsequently followed by chest compressions. Epinephrine administration was started by a single intravenous bolus of 40 μg/kg via the central venous line and followed by a continuous infusion of epinephrine at a rate of 10 μg/kg/min. Rhythm analyses and biphasic defibrillation were then repeated every 2 min until ROSC was achieved or a total of three shocks were applied. The experiment was stopped in animals without ROSC. Animals that achieved ROSC remained anaesthetised and after finishing the protocol were euthanised in deep general anaesthesia by potassium chloride administration.
Statistics Data are reported as median and interquartile range. The association between predicted PaO2 by MMIMS-MIGET and the measured value was analysed using linear regression and the Bland–Altman method. MIGET and EIT data during CPR were compared with their baseline values by the Wilcoxon rank test with correction for multiple comparisons using the Bonferroni method. Three-group comparisons were drawn by Kruskal–Wallis test with post-hoc Student–Newman–Keuls test. P values < 0.05 were regarded as significant.
Results A total of 21 animal experiments were conducted. The MMIMS-MIGET measurements in CPR were technically feasible in all animals. The overall linear correlation between the measured and the predicted PaO2 by MMIMS-MIGET was excellent (R2 = 0.930, P < 0.001, n = 42 measurements). Overall, the MMIMS-MIGET slightly underestimated the measured PaO2 at high values (10.29 + 0.88 × measured PaO2), an effect that was most pronounced in the CPR analysis. Figure 1 shows the separate correlations during baseline and CPR (Baseline: predicted = 9.30 + 0.91 × measured PaO2; CPR: predicted = 25.95 + 0.73 × measured PaO2). A Bland–Altman analysis yielded a mean difference of 21 mmHg (limits of agreement −108 to 66 mmHg) in the baseline and 18 mmHg (limits of agreement −140 to 103 mmHg) during CPR. Residual sum of squares (RSS) as an indicator of experimental error were 1.0 (0.3–2.2) during baseline and 2.1 (0.1–10) during CPR. Eleven of the
1034
21 animals achieved ROSC criteria (FiO2 0.21: 43%; FiO2 0.7: 71%; FiO2 1.0: 43%). Table 1 shows the key haemodynamic and blood gas data and indicates considerable differences in PaO2 as induced by the different FiO2 levels. Figure 2 summarises the group-wise VA/Q ratios. The baseline data depict the healthy conditions with a small shunt fraction in anaesthetised subjects. During CPR, a significant increase of the pulmonary shunt fraction occurs, while the normal VA/Q ratios considerably shift to high VA/Q. During baseline, significantly lower shunt and corresponding higher normal VA/Q fractions occurred in the FiO2 0.21– group, while in the intergroup comparisons, the highest shunt fractions were found under FiO2 1.0 conditions. In CPR, the shift to high VA/Q ratios was similar in all groups. However, shunt fractions were significantly higher when FiO2 1.0 was compared with the two other groups. Only minimal low VA/Q ratios occurred in all groups and time points (data not shown). Due to technical reasons in the challenging setting of CPR and chest compressions, exploitable EIT recordings were available from five animals per group each. Accordingly, we did not draw intergroup comparisons. The summarised data (n = 15, Fig. 3) show a significant shift of the regional distribution as measured by the amplitudes of thoracic bioimpedance variations from the dorsal and middle to the ventral lung region. In an additional subanalysis, the summed impedance changes from the middle and dorsal region significantly decreased, corresponding to the increase in the ventral region.
Discussion Methodical aspects The present study analyses the feasibility and face validity of MIGET-derived VA/Q measurements under highly deteriorated CPR conditions. In general, few technologies allow for repetitive assessment of VA/Q in experimental research or clinical routine. The MMIMS is a relatively novel MIGET device, which was developed to simplify and enhance VA/Q measurements. As a perfusiondependent technology, the MIGET, based on pulmonary elimination or retention of the applied inert gases, requires adequate haemodynamics and pulmonary perfusion. Loeckinger et al. conducted MIGET-derived VA/Q measurements in early stable ROSC in a porcine model and reported persisting impairment in terms of increased shunt and low VA/Q units over 2 h.12 To our knowledge, MIGET
VA/Q during cardiopulmonary resuscitation
Fig. 1. Partial pressure of oxygen (PaO2) prediction by micropore membrane inlet mass spectrometry-multiple inert gas elimination (MMIMS-MIGET): linear regression between measured PaO2 by blood gas analysis and the predicted value by MIGET (upper graphs), and corresponding Bland–Altman plots (lower graphs) during baseline and cardiopulmonary resuscitation (CPR) (n = 21 each).
never has been applied under CPR conditions. In our study, blood withdrawal and analysis was feasible in all animals. Comparison with an external reference by means of measured and predicted PaO26,9 during CPR yielded a strong linear correlation (Fig. 1). At baseline, there was a small underestimation of measured PaO2 at high PaO2, a trend that was more pronounced during CPR. At low values, pre-
dicted PaO2 was slightly higher than measured PaO2 during CPR, similar to baseline. The small difference at baseline between predicted and measured PaO2 at low PaO2 is similar to results from prior studies.6,13 Furthermore, this difference increased only slightly with CPR, which suggests that CPR does not lead to a considerable diffusion limitation of oxygen transfer.7 The time span to achieve a new
1035
E. K. Hartmann et al. Table 1 Haemodynamics and oxygenation data. Parameter
FiO2
Baseline
Arrest
CPR
MAP (mmHg)
0.21 0.7 1.0 0.21 0.7 1.0 0.21 0.7 1.0 0.21 0.7 1.0 0.21 0.7 1.0 0.21 0.7 1.0
106 (33) 92 (35) 110 (16) 5.1 (2.0)* 3.9 (2.1) 3.9 (1.4) 100 (24) 89 (39) 85 (13) 107 (16)* 374 (96)* 594 (211)* 510 (78) 534 (137) 594 (211) 73 (5)* 78 (6) 84 (10)
28 (9) 23 (9) 30 (9) – – – – – – – – – – – – – – –
44 (5) 44 (16) 47 (8) 1.0 (0.4) 0.8 (0.5) 0.8 (0.3) 100 (1) 100 (7) 100 (21) 63 (17)† 94 (261)† 175 (194)† 299 (80) 134 (373) 175 (194) 20 (6) 23 (19) 32 (24)
CO (l/min) HR (per minute) PaO2 (mmHg) PaO2/FiO2 (mmHg) SvO2 (%)
*P < 0.05 in three-group comparison at baseline. †P < 0.05 in three-group comparison during CPR, n = 7 per group. MAP, mean arterial pressure; CO, cardiac output; HR, heart rate or chest compression frequency; PaO2, arterial partial pressure of oxygen; FiO2, fraction of inspired oxygen; SvO2, mixed venous oxygen saturation.
steady state of ventilation and perfusion following cardiac arrest and initiation of CPR is not known and influenced by multiple factors. Traditional calculations suggest that 95% equilibrium to a new steady state adjusts within 2 min.14,15 The chosen interval of 6–8 min following cardiac arrest should be adequate in this context. Nevertheless, the exact time constants during CPR need to examined in further studies. RSS values as indicators of experimental failure were well below the thresholds that were defined for gas chromatographic MIGET measurements (RSS < 5.3 in 50%, < 10.6 in 90% and < 16.8 in 99% of the measurements16). The baseline VA/Q ratios further correspond well to previous data from healthy, comparable-sized pigs (mean normal VA/Q fraction 90–96%10,17). A high FiO2 leads to the formation of atelectasis in anaesthetised subjects.18,19 The small but significant shunt variations between the three groups with the lowest shunt during FiO2 0.21 could be explained by this concept. However, the measured changes did not lead to differences in oxygenation in healthy pigs with intact hypoxic pulmonary vasoconstriction.
Impact of CPR on VA/Q ratios A considerable shift from normal to high VA/Q is induced by CPR, which approximately equals the
1036
Fig. 2. Group-wise ventilation/perfusion distribution (VA/Q) ratios during baseline and cardiopulmonary resuscitation (CPR). Only minimal low VA/Q ratios (data not shown). Standard VA/Q ratios: shunt (VA/Q < 0.005), low VA/Q (0.005 < VA/Q < 0.1), normal VA/Q (0.1 < VA/Q < 10) and high VA/Q (10 < VA/ Q > 100). #P < 0.05 vs. baseline; *P < 0.05 in intergroup comparison, n = 7 per group. FiO2, fractions of inspired oxygen.
amount of persisting normal VA/Q ratios. A similar shift of the indirectly estimated global VA/Q in CPR has been previously reported.20 High VA/Q ratios imply two possible explanations: pulmonary hyperventilation or impaired perfusion. The overall VA/Q ratio of the main mode reflects the ratio of ventilation to effective perfusion, particularly the ventila-
VA/Q during cardiopulmonary resuscitation
Fig. 3. Regional ventilation distribution by electrical impedance tomography during baseline and cardiopulmonary resuscitation (CPR) (upper graph). Addition of the middle and dorsal lung region (lower graph). *P < 0.05.
tion to non-shunting blood flow. During CPR, pulmonary perfusion is severely impaired, although not annihilated (mean arterial pressure > 40 mmHg in our model). A high VA/Q shift related to perfusion deficits was also found in pig model of haemorrhagic shock that featured similarly low mean arterial pressure values.21 EIT measurements yielded significantly increased amplitudes in the ventral lung regions and a dorsal to ventral shift. In general, EIT data can be post-processed to discriminate ventilation from perfusion-related variations by band-pass filtering and Fourier analysis.22 Compression frequencies in a range of 100/min, however, clearly overlap with the perfusion signals. Also, the respiratory pattern is compromised by nonsynchronised ventilation.23 Chest compressions cause active compression/decompression-related changes in lung aeration and ventilation
inhomogeneity leading to ventral pulmonary hyperinflation and atelectasis formation in the dependent lung areas.3 In this context, Markstaller et al. reported up to 7% of overdistended lung tissue in a computer tomography-based analysis during CPR in pigs.3 These findings also influence the high VA/Q and shunt fractions; however, the ventilation to non-effectively perfused lung areas mainly seem to determine the mode of VA/Q mismatch.
Effect of FiO2 In CPR, significantly higher shunt fractions were observed during FiO2 0.7 and particularly FiO2 1.0 ventilation, which therefore seems to aggravate the pre-existing atelectatic lung areas. Gas exchange during CPR is influenced by the applied mechanical ventilation as well as chest compressions. Considerable impairment of the pulmonary function in
1037
E. K. Hartmann et al.
ROSC has been reported by several studies.4,12,24 Our data, in this context, indicate a relevant impact of FiO2 level on atelectasis formation already during CPR. Ongoing CPR, which is characterised by minimal perfusion and oxygen delivery conditions, should be differentiated from the ROSC-induced reperfusion phase, in which it is widely accepted to avoid FiO2 levels leading to hyperoxia.1,25,26 FiO2related differences of the shunt, but not of the high VA/Q ratios, the distribution of bioimpedance variations by EIT and comparable haemodynamics in the three groups (Table 1) imply that beyond perfusion, the FiO2 may considerably contribute to the VA/Q deterioration. Despite the highest overall PaO2 values during CPR, the FiO2 1.0 ventilation most negatively influences the VA/Q pattern. Furthermore, hypoxic pulmonary vasoconstriction may influence these findings. A prior study assumed that during CPR pure oxygen but not air ventilation may inhibit hypoxic pulmonary vasoconstriction.27 As noted, signs of lung injury following CPR have been reported in several studies, although these aspects lie beyond scope of this study. Nevertheless, our data indicate that impaired lung function immediately occurs during ongoing CPR and is not exclusively related to the reperfusion phase.
Limitations of the study The present study has several limitations. Because there were no preceding data for a power analysis, we chose a convenience sample of seven animals per group, which was oriented on previous publications from our group.24 The particular variant of MIGET measurement used in our laboratory requires cardiac output data, which is difficult during CPR. The PiCCO technology was used for this purpose, which was previously reported to provide reliable measurements and significant correlations to blood pressure during CPR in pigs.28,29 This pilot CPR study is limited to face validity, as assessed by successful comparison with an external reference (PaO2) and acceptable ranges of experimental error (RSS values). We did neither perform repeated measurements to examine retest reliability nor analysis of responses after a change in cardiac output. Hence, these distinct aspects of validation need to be addressed in future experiments. The study was neither designed nor powered to provide long-term ROSC or post-mortem data, but strictly focused on the feasibility of VA/Q measurements and the acute or FiO2-related changes during CPR. Additionally, in our experimental setting, the different FiO2 levels were initiated 10 min before induc-
1038
tion of arrest rather than on initiation of resuscitation. A model inevitably differs from the clinical scenario regarding brief and uniform performance of CPR, the defined intervals of ventilation and the standardised measurement procedure. Nevertheless, the reported overall ROSC rates (52%) imply a realistic scenario.
Conclusion The assessment of VA/Q by MIGET is feasible even during CPR and shows face validity. VA/Q measurements correlate well to the conventional blood gas analysis and FiO2-related changes were traceable in this setting. MMIMS-MIGET represents a novel tool for experimental purposes and provides further insights in the complex interactions of ventilation and perfusion and their influence upon the pulmonary function during CPR. Additional studies might therefore focus on the best-suited oxygen fraction regarding the VA/Q ratio and its impact on pulmonary as well as global outcome.
Acknowledgements All experiments were performed in the laboratories of the Department of Anaesthesiology, Medical Centre of the Johannes Gutenberg-University Mainz. Conflicts of interest: J. E. Baumgardner, the owner of Oscillogy LLC, which commercially distributes the MMIMS-MIGET, participated in the study setup and provided technical support, although he had no influence on performance of the experiments and data analysis. None of the other authors declares a conflict of interest. Funding: The study was funded in part by the German Research Council DFG PAK 415-1.
References 1. Kill C, Dersch W, Wulf H. Advanced life support and mechanical ventilation. Curr Opin Crit Care 2012; 18: 251–5. 2. Kill C, Torossian A, Freisburger C, Dworok S, Massmann M, Nohl T, Henning R, Wallot P, Gockel A, Steinfeldt T, Graf J, Eberhart L, Wulf H. Basic life support with four different compression/ventilation ratios in a pig model: the need for ventilation. Resuscitation 2009; 80: 1060–5. 3. Markstaller K, Karmrodt J, Doebrich M, Wolcke B, Gervais H, Weiler N, Thelen M, Dick W, Kauczor HU, Eberle B. Dynamic computed tomography: a novel technique to study lung aeration and atelectasis formation during experimental CPR. Resuscitation 2002; 53: 307–13. 4. Wang S, Wu JY, Guo ZJ, Li CS. Effect of rescue breathing during cardiopulmonary resuscitation on lung function after restoration of spontaneous circulation in a porcine model of prolonged cardiac arrest. Crit Care Med 2013; 41: 102–10. 5. Yannopoulos D, Matsuura T, McKnite S, Goodman N, Idris A, Tang W, Aufderheide TP, Lurie KG. No assisted ventilation cardiopulmonary resuscitation and 24-hour neurological outcomes in a porcine model of cardiac arrest. Crit Care Med 2010; 38: 254–60.
VA/Q during cardiopulmonary resuscitation 6. Kretzschmar M, Schilling T, Vogt A, Rothen HU, Borges JB, Hachenberg T, Larsson A, Baumgardner JE, Hedenstierna G. Multiple inert gas elimination technique by micropore membrane inlet mass spectrometry – a comparison with reference gas chromatography. J Appl Physiol 2013; 115: 1107–18. 7. Wagner PD. The multiple inert gas elimination technique (MIGET). Intensive Care Med 2008; 34: 994–1001. 8. Duenges B, Vogt A, Bodenstein M, Wang H, Bohme S, Rohrig B, Baumgardner JE, Markstaller K. A comparison of micropore membrane inlet mass spectrometry-derived pulmonary shunt measurement with Riley shunt in a porcine model. Anesth Analg 2009; 109: 1831–5. 9. Hartmann EK, Duenges B, Baumgardner JE, Markstaller K, David M. Correlation of thermodilution-derived extravascular lung water and ventilation/perfusioncompartments in a porcine model. Intensive Care Med 2013; 39: 1313–7. 10. Hartmann EK, Boehme S, Bentley A, Duenges B, Klein KU, Elsaesser A, Baumgardner JE, David M, Markstaller K. Influence of respiratory rate and end-expiratory pressure variation on cyclic alveolar recruitment in an experimental lung injury model. Crit Care 2012; 16: R8. 11. Bodenstein M, Wang H, Boehme S, Vogt A, Kwiecien R, David M, Markstaller K. Influence of crystalloid and colloid fluid infusion and blood withdrawal on pulmonary bioimpedance in an animal model of mechanical ventilation. Physiol Meas 2012; 33: 1225–36. 12. Loeckinger A, Kleinsasser A, Wenzel V, Mair V, Keller C, Kolbitsch C, Recheis W, Schuster A, Lindner KH. Pulmonary gas exchange after cardiopulmonary resuscitation with either vasopressin or epinephrine. Crit Care Med 2002; 30: 2059–62. 13. Wagner PD, Gale GE, Moon RE, Torre-Bueno JR, Stolp BW, Saltzman HA. Pulmonary gas exchange in humans exercising at sea level and simulated altitude. J Appl Physiol 1986; 61: 260–70. 14. Hopkins SR, Gavin TP, Siafakas NM, Haseler LJ, Olfert IM, Wagner H, Wagner PD. Effect of prolonged, heavy exercise on pulmonary gas exchange in athletes. J Appl Physiol 1998; 85: 1523–32. 15. Hopkins SR, Bayly WM, Slocombe RF, Wagner H, Wagner PD. Effect of prolonged heavy exercise on pulmonary gas exchange in horses. J Appl Physiol 1998; 84: 1723–30. 16. Roca J, Wagner PD. Contribution of multiple inert gas elimination technique to pulmonary medicine. 1. Principles and information content of the multiple inert gas elimination technique. Thorax 1994; 49: 815–24. 17. Neumann P, Hedenstierna G. Ventilation-perfusion distributions in different porcine lung injury models. Acta Anaesthesiol Scand 2001; 45: 78–86. 18. Dantzker DR, Wagner PD, West JB. Instability of poorly ventilated lung units during oxygen breathing. J Appl Physiol 1975; 38: 886–95. 19. Rothen HU, Sporre B, Engberg G, Wegenius G, Reber A, Hedenstierna G. Prevention of atelectasis during general anaesthesia. Lancet 1995; 345: 1387–91.
20. Wang S, Li C, Ji X, Yang L, Su Z, Wu J. Effect of continuous compressions and 30 : 2 cardiopulmonary resuscitation on global ventilation/perfusion values during resuscitation in a porcine model. Crit Care Med 2010; 38: 2024–30. 21. Robinson NB, Chi EY, Robertson HT. Ventilation-perfusion relationships after hemorrhage and resuscitation: an inert gas analysis. J Appl Physiol Respir Environ Exerc Physiol 1983; 54: 1131–40. 22. Bodenstein M, David M, Markstaller K. Principles of electrical impedance tomography and its clinical application. Crit Care Med 2009; 37: 713–24. 23. Kill C, Hahn O, Dietz F, Neuhaus C, Schwarz S, Mahling R, Wallot P, Jerrentrup A, Steinfeldt T, Wulf H, Dersch W. Mechanical ventilation during cardiopulmonary resuscitation with intermittent positive-pressure ventilation, bilevel ventilation, or chest compression synchronized ventilation in a pig model. Crit Care Med 2014; 42: e89–95. 24. Markstaller K, Rudolph A, Karmrodt J, Gervais HW, Goetz R, Becher A, David M, Kempski OS, Kauczor HU, Dick WF, Eberle B. Effect of chest compressions only during experimental basic life support on alveolar collapse and recruitment. Resuscitation 2008; 79: 125–32. 25. Kilgannon JH, Jones AE, Parrillo JE, Dellinger RP, Milcarek B, Hunter K, Shapiro NI, Trzeciak S. Relationship between supranormal oxygen tension and outcome after resuscitation from cardiac arrest. Circulation 2011; 123: 2717–22. 26. Kilgannon JH, Jones AE, Shapiro NI, Angelos MG, Milcarek B, Hunter K, Parrillo JE, Trzeciak S. Association between arterial hyperoxia following resuscitation from cardiac arrest and in-hospital mortality. JAMA 2010; 303: 2165–71. 27. Rubertsson S, Karlsson T, Wiklund L. Systemic oxygen uptake during experimental closed-chest cardiopulmonary resuscitation using air or pure oxygen ventilation. Acta Anaesthesiol Scand 1998; 42: 32–8. 28. Lopez-Herce J, Fernandez B, Urbano J, Mencia S, Solana MJ, Del Castillo J, Rodriguez-Nunez A, Bellon JM, Carrillo A. Correlations between hemodynamic, oxygenation and tissue perfusion parameters during asphyxial cardiac arrest and resuscitation in a pediatric animal model. Resuscitation 2011; 82: 755–9. 29. Lopez-Herce J, Fernandez B, Urbano J, Mencia S, Solana MJ, Rodriguez-Nunez A, Bellon JM, Carrillo A. Hemodynamic, respiratory, and perfusion parameters during asphyxia, resuscitation, and post-resuscitation in a pediatric model of cardiac arrest. Intensive Care Med 2011; 37: 147–55.
Address: Erik K. Hartmann Department of Anaesthesiology Medical Centre of the Johannes Gutenberg-University Langenbeckstraße 1 55131 Mainz Germany e-mail: [email protected]
1039
