Galley Proof 28/07/2016; 14:14 File: thc–1-thc1248.tex; BOKCTP/xhs p.

Technology and Health Care -1 (2016) 1–10 1

DOI 10.3233/THC-161248
IOS Press

1 Potential impact of oxygenators with venous

2 air trap on air embolism in veno-arterial
3 extracorporeal life support

4 Frank Born1 , Nawid Khaladj1 , Maximilian Pichlmaier, René Schramm, Christian Hagl and

n
5 Sabina P.W. Guenther∗

o
6 Department of Cardiac Surgery, University Hospital Munich, Ludwig-Maximilian-University, Munich,

si
7 Germany

8 Received 15 June 2016

er
fv
9 Accepted 13 July 2016
oo

10 Abstract.
11 BACKGROUND: Air embolism is a potentially fatal but underrecognized complication in Extracorporeal Life Support
12 (ECLS). Oxygenators containing venous air traps have been developed to minimize the risk of air embolism in daily care.
pr

13 OBJECTIVE: We reproduced air embolism as occurring via a central venous catheter in an experimental setting to test the
14 potential of oxygenators with and without venous bubble trap (VBT) to withhold air.
ed

15 METHODS: An in vitro ECLS-circuit was created and a central venous catheter with a 3-way stopcock and a perforated male
16 luer cap was inserted into the inflow line. Three different oxygenators with and without VBT and their capability to withhold
17 air were examined. After 60 seconds of stable ECLS-flow, the stopcock was opened towards the atmosphere for 3 minutes.
ct

18 Afterwards, air accumulation within the oxygenator was determined.

19 RESULTS: Comparison of the total air entrapment showed a significant superiority of the oxygenators with VBT (p < 0.001).
rre

20 All oxygenators were able to partly withhold macro air boli, however, the capacity of oxygenators with VBT was higher. Pass-
21 ing the oxygenator resulted in a reduction of microbubbles in all cases.
22 CONCLUSIONS: Macro air emboli can be substantially reduced by usage of oxygenators that contain a VBT, whereas the
co

23 capability to withhold microbubbles to a vast extent seems to depend on the intrinsic oxygenator’s membrane.
un

24 Keywords: Air embolism, extracorporeal life support, oxygenator, venous air trap

25 1. Introduction

26 Veno-arterial Extracorporeal Life Support (ECLS) provides immediate cardiopulmonary stabilization

27 and has become increasingly widespread used in both patients with refractory cardiogenic shock due to
28 multifaceted reasons as well as in postcardiotomy patients when weaning from cardiopulmonary bypass
29 (CPB) had failed [1–5].

1
These authors contributed equally.
∗
Corresponding author: Sabina Guenther, Department of Cardiac Surgery, University Hospital Munich, Ludwig-Maximilian-
University, Marchioninistr. 15, 81377 Munich, Germany. Tel.: +49 089/4400 73457; Fax: +49 089/4400 78873; E-mail: sabina.
guenther@med.uni-muenchen.de.

0928-7329/16/$35.00 
c 2016 – IOS Press and the authors. All rights reserved
 Galley Proof 28/07/2016; 14:14 File: thc–1-thc1248.tex; BOKCTP/xhs p. 2

2 F. Born et al. / Potential impact of oxygenators with venous air trap on air embolism in veno-arterial ECLS

30 Whilst tolerable to some extent in the veno-venous setting, air embolism in veno-arterial ECLS is
31 potentially fatal. Besides gaseous microembolism (GME), macro emboli may occur and impose a life-
32 threatening risk on the patient [6–8]. Guidelines and standards of care have been established by national
33 and international committees including the Extracorporeal Life Support Organization (ELSO) [4,5,9].
34 However, even though air embolism in conventional cardiopulmonary bypass has been studied to some
35 extent, research on air embolism in ECLS and especially on its prevention so far is scarce.
36 Negative pressure in ECLS theoretically may rise up to −300 mmHg. Oxygenators designed for long
37 duration use contain plasma-tight membranes with reduced permeability and therefore a reduced filtering
38 capacity for bubbles. Air embolism may occur in standard clinical settings such as on the intensive
39 care unit, during interventions or during cardiac surgery. Potential sources include incompletely de-
40 aired circuits, open central venous catheter lines, interventions and fracturing of circuit components.
41 Electronic clamps to prevent from air entry exist; however, sudden discontinuation of ECLS in critically
42 ill patients may be fatal.

n
43 Oxygenators containing venous air traps have been developed to minimize the risk of air embolism

o
44 in daily care. In this experimental analysis, we thus aimed to reproduce air embolism as occurring via a

si
45 central venous catheter to test the potential of oxygenators with and without venous air trap to withhold

er
46 air. An experimental setup derived from conventional CPB according to Hogetveit et al. was modified
47 for application in ECLS and used [7].
fv
oo

48 2. Materials and methods

pr

49 2.1. Bubble detector GAMPT BCC 200, software version 3.4

Using an ultrasound-based bubble detector, non-invasive measurement of the air bubble activity in the
ed

50

51 blood during extracorporeal circulation (ECC) is feasible. Likewise, the device is suitable to analyze the
52 air filtering capacity of ECC components such as oxygenators.
ct

53 According to the manufacturer, the GAMPT BCC 200 (GAMPT, Merseburg, Germany) is able to
rre

54 determine number and size of microbubbles during ECC. The measurement is based on a self-calibrating
55 pulsed ultrasound Doppler (transmission frequency 2 MHz). It is thus supposed to be independent of tube
co

56 material, blood hemoglobin concentration and blood flow velocity as far as possible. Acoustic coupling
57 is controlled by the system during the analysis and used for the correction of the signals.
Micro bubbles from 5 or 10 µm up to 250 or 500 µm of size will be detected and displayed in a his-
un

58

59 togram according to preselection of the desired measuring range. Clamp-on probes are attached to the
60 circuit’s lines. The device provides the examiner with the number of bubbles detected within the measur-
61 ing range and calculates the corresponding bubble volume. Bubbles of sizes exceeding the preselected
62 measuring range will be valued as “over range”. As the bubble detector will not be able to classify these
63 bubbles, for volume calculation of such over range bubbles the software assumes the maximum value of
64 the measuring range (i.e. 250 or 500 µm). If a larger air amount, i.e. gross or macro air, occurs within
65 the circuit, the system will not be able to individually distinguish bubbles anymore but will determine
66 the air amount as a total volume (“bolus volume”) instead.

67 2.2. Experimental setup

68 For the experimental setup, the SCPC (Stöckert Centrifugal Pump Console) ECLS system with a Rev-
69 olution 5 centrifugal pump (Sorin Group, Munich, Germany) was used. A 40% glycerol/saline-mixture
 Galley Proof 28/07/2016; 14:14 File: thc–1-thc1248.tex; BOKCTP/xhs p. 3

F. Born et al. / Potential impact of oxygenators with venous air trap on air embolism in veno-arterial ECLS 3

o n
si
er
fv
oo
pr

Fig. 1. Experimental setup. VBT venous bubble trap.

ed

70 severed as priming solution since rheological characteristics resemble human blood [10,11]. To maintain
ct

71 stable rheology, the temperature was monitored and kept at 37◦ C. The ECLS system was primed with
rre

72 650 ml and the outflow line was attached to a prefilled cardiotomy reservoir (4500 ml). The level within
73 the cardiotomy reservoir was kept at 4500 ml at all times. At the outlet of the reservoir a y-shaped con-
74 nector (3/8 × 3/8 × 1/4 inch) was installed and the inflow line was connected to the 3/8 port. A 35 cm
co

75 1/4 inch tube was attached to the 1/4 inch ending of the connector. Altogether, the tube lengths were
similar to standard clinical settings. At the end of the latter tube a ThruPort Introducer Sheath IS 19 A
un

76

77 with a hemostasis valve (Edwards Lifesciences, Unterschleissheim, Germany) was installed into which a
78 commonly used de-aired trilumen central venous catheter (18 Gauge, Arrow R
International Inc., Read-
79 ing, USA) was inserted. At its end, a three-way stopcock was mounted and closed towards the inflow line
80 with the security clip being additionally sealed. A male luer cap with a 1 mm perforation (Fig. 1) was
81 attached to one of the stopcock’s endings. This cap is the standard cap in several commercially available
82 central venous catheterization kits.
83 Clamp-on probes of an ultrasound-based microbubble detector (GAMPT BCC 200) were attached to
84 the outflow and inflow line. 500 µm was defined as the range up to which bubbles should be classified
85 during the measurement. Bubbles exceeding this size will be valued as over range.
86 The detailed experimental setup is depicted in Fig. 1. Figure 2 shows the close proximity of a central
87 venous catheter and the tip of the venous cannula in a patient with femoral ECLS implantation for
88 cardiogenic shock.
89 Three diffusion membrane oxygenators and their capability to withhold air were examined (Oxy1:
 Galley Proof 28/07/2016; 14:14 File: thc–1-thc1248.tex; BOKCTP/xhs p. 4

4 F. Born et al. / Potential impact of oxygenators with venous air trap on air embolism in veno-arterial ECLS

Table 1
Description of the tested oxygenators. Oxy oxygenator, VBT venous bubble trap
Oxy1 (ECC.O 5) Oxy2 (Hilite
R
LT) Oxy3 (ECCO)
Membrane material Polymethylpentene fiber Polymethylpentene fiber Polymethylpentene fiber
Membrane surface area [m2 ] 1.2 1.9 1.8
Priming volume [ml] 390 320 225
Coating Ph.i.s.i.o. RheoparinR
Phosphorylcholine
VBT + + −

o n
si
er
fv
oo

Fig. 2. Inverted chest x-ray of a patient with femoral ECLS implantation for cardiogenic shock. Close proximity of the central
pr

venous catheter (white arrow) and the tip of the venous ECLS cannula (black arrow).

ECC.O 5, Sorin Group, Munich, Germany; Oxy2: Hilite R

LT, Medos R
ed

90 , Stollberg, Germany; Oxy3:

TM
91 ECCO, Eurosets , Medolla, Italy), detailed information on the oxygenators is given in Table 1. Two
contain a venous bubble trap (Oxy1: ECC.O 5, Sorin Group; Oxy2: Hilite R
LT, Medos R
). Currently,
ct

92

93 for ECLS these are the only diffusion membrane oxygenators with integrated VBT available on the
rre

94 European market. Both oxygenators with VBT are equipped with a prime line for de-airing. Via that
95 port – which is located at the highest point of the VBT – air can also be manually aspirated out of the
co

96 oxygenator. The VBT was de-aired according to manufacturer’s instructions. Afterwards, the port was
97 closed for the whole test run and used for suctioning the air out of the oxygenator after termination of
un

98 the experiment. All tested oxygenators are used routinely in ECLS and in our department. The setup was
99 identical in all cases with only the oxygenator varying. The oxygenator line pressure, as measured after
100 the oxygenator, was kept between 260 and 280 mmHg at all times.
101 ECLS flow was started with a pump flow of 4.3 l/min and kept stable for 60 seconds in order to
102 create a steady state. During this period, the central venous catheter remained closed towards the in-
103 flow line of the ECLS system. After 60 seconds of stable ECLS flow, the stopcock was opened towards
104 the atmosphere representing a clinical setting in which the standard perforated cap accidentally has not
105 been switched to an unperforated cap or a setting in which a cap is not properly mounted and thus
106 does not completely seal the central venous line’s entry. For the whole observation period, revolutions
107 per minute [U/min], pump flow [l/min] as well as negative pressure intensity within the venous line
108 [mmHg] were registered. After opening the three-way stopcock, observation and measurements were
109 continued for 3 minutes. Simultaneously, parameters as registered by the bubble detector were collected.
110 Three minutes mimic a typical time span for identification of a relevant problem and taking appropri-
111 ate countermeasures in daily clinical practice. After 3 minutes, the experiments were terminated and
 Galley Proof 28/07/2016; 14:14 File: thc–1-thc1248.tex; BOKCTP/xhs p. 5

F. Born et al. / Potential impact of oxygenators with venous air trap on air embolism in veno-arterial ECLS 5

Table 2
ECLS parameters before (0–60 seconds; part A) and after ( 80 seconds; part B) air inlet. Oxy oxygenator, VBT venous bubble
trap, sec seconds, RPM revolutions per minute, Flow pump flow, Suction negative pressure intensity within the venous line
Oxy1 (ECC.O 5, with VBT) Oxy2 (HiliteR
LT, with VBT) Oxy3 (ECCO, no VBT)
Time RPM Flow Suction RPM Flow Suction RPM Flow Suction
[sec] [l/min] [mmHg] [l/min] [mmHg] [l/min] [mmHg]
20 2776 ± 3 4.30 ± 0.01 93.20 ± 1.17 2801 ± 2 4.29 ± 0.01 96.80 ± 0.98 2635 ± 2 4.29 ± 0.02 88.80 ± 0.40
40 2770 ± 4 4.27 ± 0.01 88.20 ± 0.74 2807 ± 2 4.27 ± 0.00 92.40 ± 1.36 2631 ± 2 4.27 ± 0.01 91.20 ± 0.75
60 2767 ± 0 4.28 ± 0.01 91.80 ± 0.98 2811 ± 1 4.28 ± 0.01 93.40 ± 1.02 2630 ± 1 4.30 ± 0.00 92.80 ± 0.40
80 2901 ± 115 3.24 ± 0.25 51.00 ± 10.20 2795 ± 16 3.23 ± 0.10 49.00 ± 2.28 2592 ± 29 3.20 ± 0.18 42.00 ± 4.60
100 2901 ± 116 3.34 ± 0.18 47.80 ± 2.14 2795 ± 15 3.20 ± 0.08 49.60 ± 3.77 2591 ± 29 3.25 ± 0.17 41.40 ± 4.50
120 2900 ± 115 3.39 ± 0.14 49.20 ± 1.47 2796 ± 16 3.27 ± 0.10 51.40 ± 4.08 2591 ± 30 3.24 ± 0.18 42.20 ± 3.49
140 2899 ± 115 3.37 ± 0.16 48.60 ± 1.02 2795 ± 15 3.23 ± 0.06 48.00 ± 3.16 2590 ± 29 3.27 ± 0.17 43.20 ± 3.71
160 2900 ± 117 3.38 ± 0.15 49.20 ± 1.47 2794 ± 15 3.19 ± 0.09 47.80 ± 2.40 2591 ± 29 3.37 ± 0.11 45.60 ± 2.58
180 2901 ± 116 3.35 ± 0.15 48.00 ± 1.26 2797 ± 14 3.19 ± 0.05 47.20 ± 1.72 2590 ± 29 3.21 ± 0.19 40.40 ± 4.03
200 2901 ± 116 3.37 ± 0.15 48.20 ± 0.75 2793 ± 18 3.13 ± 0.02 45.60 ± 1.85 2591 ± 28 3.30 ± 0.15 44.60 ± 3.01

n
220 2902 ± 116 3.37 ± 0.15 48.40 ± 1.02 2796 ± 15 3.17 ± 0.05 45.60 ± 1.02 2591 ± 29 3.32 ± 0.11 44.20 ± 3.82
2901 ± 116 3.37 ± 0.15 48.00 ± 0.89 2794 ± 16 3.16 ± 0.05 45.40 ± 0.49 2591 ± 30 3.28 ± 0.12 43.80 ± 3.66

o
240

si
112 air accumulation within the oxygenator was measured by aspirating it using conventional perfusor sy-

er
113 ringes (50 ml). The influence on ECLS flow parameters was determined via comparison of RPM, pump
fv
114 flow and negative pressure intensity within the venous line at 60 seconds (i.e. before) and 80 seconds
115 (i.e. after stopcock-opening). For each oxygenator, the whole experiment was repeated five times under
oo

116 standardized conditions as explained above to increase measuring accuracy.

pr

117 2.3. Statistics

118 Categorical variables are given as numbers and percentages. Data concerning continuous variables are
ed

119 expressed as mean ± standard deviation (SD). Statistical comparison was performed using the paired t-
120 test. ECLS flow parameters were compared at 60 seconds (i.e. before) and 80 seconds (i.e. after stopcock
ct

121 opening). Air bubble activity was collected for 3 minutes after air inlet until termination of the experi-
rre

122 ment. After termination, total air bubble counts as well as air accumulation within the oxygenators were
123 compared. IBM SPSS Statistics software, Version 20 (IBM Corp. Released 2011. IBM SPSS Statistics
co

124 for Windows, Version 20.0. Armonk, NY: IBM Corp.) was used for statistical analysis.
un

125 3. Results

126 Within the initial 60 seconds of steady state, ECLS parameters (RPM, pump flow, negative pressure
127 intensity within the venous line) were kept stable (see Table 2, part A). Part B of Table 2 depicts ECLS
128 flow parameters after opening of the stopcock, Table 3 shows the registered air amounts and bubbles.
129 The total air volume withheld by the oxygenators as determined by aspiration using a conventional
130 syringe after termination of the experiments was higher for the oxygenators with bubble trap (Oxy1
131 (ECC.O 5, with VBT): 71.6 ± 2.7 ml, Oxy2 (Hilite R
LT, with VBT): 50.0 ± 3.0 ml and Oxy3 (ECCO,
132 no VBT): 6.4 ± 1.5 ml, respectively). Statistical comparison confirmed these findings (p < 0.001 for
133 Oxy1 vs. Oxy3 and for Oxy2 vs. Oxy3).
134 Since the experimental setup was designed to evaluate the effects of macro air entering an ECLS
135 circuit large bolus volumes were seen in all cases on the venous side. All oxygenators were able to
136 withhold parts of this bolus air volume whereas the capacity of the oxygenators with air trap was higher
 Galley Proof 28/07/2016; 14:14 File: thc–1-thc1248.tex; BOKCTP/xhs p. 6

6 F. Born et al. / Potential impact of oxygenators with venous air trap on air embolism in veno-arterial ECLS

Table 3
Registered bubbles and air amounts. Oxy oxygenator, VBT venous bubble trap, No number
Oxy1 (ECC.O 5, Oxy2 (HiliteR
Oxy3 (ECCO,
with VBT) LT, with VBT) no VBT)
Bubbles detected in the arterial line [No] 1938.6 ± 432.3 2947.2 ± 456.1 2984.8 ± 443.6
Bubbles detected in the venous line [No] 6830.4 ± 154.9 7232.4 ± 944.1 11641.4 ± 556.5
Bubbles detected in the arterial line over range [No] 22349.6 ± 68.3 21696.6 ± 551.2 23820.8 ± 383.1
Bubbles detected in the venous line over range [No] 12814.2 ± 466.1 12546.4 ± 870.5 12014.0 ± 892.6
Bubble volume detected in the arterial line [µl] 184.2 ± 0.7 185.6 ± 2.2 198.0 ± 1.6
Bubble volume detected in the venous line [µl] 114.6 ± 10.3 104.3 ± 6.9 113.2 ± 10.0
Bolus volume detected in the arterial line [µl] 6.0 ± 7.4 4.6 ± 5.4 1307.6 ± 84.4
Bolus volume detected in the venous line [µl] 12770.8 ± 1306.9 14090.0 ± 832.5 18311.2 ± 1303.1

137 (Oxy1 (ECC.O 5, with VBT): 99.95%, Oxy2 (Hilite R

LT, with VBT): 99.97%, Oxy3 (ECCO, no VBT):
138 92.86%, respectively, Table 3).

n
139 Analysis of the registered microbubbles showed a reduction of in-range bubbles pre- versus post-
oxygenator in all cases (Oxy1 (ECC.O 5, with VBT): 71.62%, Oxy2 (Hilite

o
R
140 LT, with VBT): 59.25%,

si
141 Oxy3 (ECCO, no VBT): 74.36%, respectively; Table 3), whereas the numbers of range-exceeding bub-
142 bles increased which resulted in an increase of the overall bubble volumes registered after the oxygenator

er
143 in comparison to those registered pre-oxygenator. fv
144 In all cases, a reduction of pump flow was seen after opening the stopcock and thus air inlet. Compari-
145 son of values before opening the stopcock at 60 seconds and afterwards at 80 seconds showed a decrease
in all cases (p = 0.001 for Oxy1 (ECC.O 5, with VBT), p < 0.001 for Oxy2 (Hilite
oo
R
146 LT, with VBT)
147 and Oxy3 (ECCO, no VBT), respectively, Table 2). Likewise, capability to maintain negative pressure
within the venous line was reduced in all cases as a result of decreased pump flow (p = 0.001 for Oxy1,
pr

148

149 p < 0.001 for Oxy2 and Oxy3, Table 2). Revolutions per minute did not or only slightly change after
air inlet (p = 0.081 for Oxy1 with, however, varying RPMs as illustrated by a comparably high standard
ed

150

151 deviation, p = 0.114 for Oxy2 and p = 0.057 for Oxy3, Table 2).
ct
rre

152 4. Discussion

Air embolism during extracorporeal circulation may occur in standard clinical settings with potentially
co

153

154 fatal consequences [7]. It is very often underrecognized or overlooked in daily practice and current
literature to a vast extend lacks information on gas embolism and its prevention especially in veno-
un

155

156 arterial ECLS [12,13]. Diagnostic means for the detection of air emboli such as transcranial Doppler
157 (TCD), precordial Doppler and transesophageal echocardiography (TEE) detect the emboli post factum,
158 i.e. after having entered the body, and are technically demanding. The dilemma is further aggravated
159 by the fact that no reversing causal therapy exists and only symptomatic care, such as optimization of
160 hemodynamics and perfusion including microcirculation, can be provided.
161 Potential sources of air embolism in ECLS include human error (incompletely de-aired circuits, air en-
162 trapped after connecting cannulas and tubes, disconnected or incompletely closed central venous catheter
163 lines), interventions (taking blood gases, drug administration, insertion of central venous catheters, car-
164 diac catheterization or other invasive procedures, respectively) as well as technical failure (fracturing
165 of venous and arterial catheter or ECLS-system lines) [6,7,14,15]. As shown in our analysis, perforated
166 luer-lock caps besides imposing infectious hazards are also a potential air entry site if not immediately
167 switched to a conventional cap. Temperature as well as blood flow velocity are key factors influencing
168 the risk of air embolism [15]. Zanatta et al. saw a direct correlation in ECLS between air entry via central
 Galley Proof 28/07/2016; 14:14 File: thc–1-thc1248.tex; BOKCTP/xhs p. 7

F. Born et al. / Potential impact of oxygenators with venous air trap on air embolism in veno-arterial ECLS 7

169 venous infusion lines and cerebral microembolic signals as registered by TCD. They demonstrated gas
170 embolism to be triggered by microbubbles in continuous infusion pumps, in infusion solutions or by
171 unlocked catheter taps [6]. Hogetveit et al. recently verified that closest proximity of a central venous
172 catheter and the venous cannula of a heart-lung-machine may lead to massive air entry if the catheter
173 is accidentally open [7]. In contrast to this setting, in closed ECLS systems with transvenous cannula-
174 tion a higher negative pressure intensity may occur depending on the venous cannula’s size [16]. Due
175 to the increasingly widespread use of peripheral ECLS systems precautions to prevent from devastating
176 consequences are necessary [17].
177 In addition to macro air entering the circuit also microbubbles entering or arising within the system
178 may result in dismal outcome. Besides other adverse effects especially neuropsychological deficits and
179 complications have been related to gaseous microembolism (GME) during cardiopulmonary bypass aris-
180 ing from diffuse microischemia [15,18]. Additionally to vascular obstruction GME causes blood stasis,
181 damage of the endothelium as well as inflammation with subsequent complement and leucocyte activa-

n
182 tion, platelet adhesion, reduced nitric oxide production and changes of pressure within the microcircu-

o
183 lation and interstitium around the affected vessel, which further aggravate impaired cerebral blood flow.

si
184 Clinical outcome, however, varies depending on the size of the bubble, its gas composition, the location,

er
185 general status of the patient, co-morbidity as well as several other and partly unknown factors [12,15].
186 Altogether, there is a spectrum of scenarios ranging from a single exposure to a very large air volume
fv
187 to (recurrent) small volumes of microbubbles [12]. Especially during several days of ECLS therapy,
188 the overall embolic load can be enormous [6]. Air bubbles in veno-venous extracorporeal membrane
oo

189 oxygenation (ECMO) or on the venous side of veno-arterial extracorporeal life support (ECLS) can be
190 tolerated to some extent with the lung serving as a filtering capacity as defined by 0.3–0.35 ml/kg/min
pr

191 air at maximum with, however, adverse effects including downstream pulmonary and systemic alter-
192 ations [7,12,13,19]. Nonetheless, in worst-case scenarios large air accumulations may lead to cardiocir-
ed

193 culatory failure due to right ventricular outflow obstruction or an air lock of the extracorporeal circuit.
194 Paradoxical embolism may not only occur via a patent foramen ovale but also through arteriovenous
ct

195 pulmonary malformations or other right-to-left shunts [7]. On the arterial side of a veno-arterial ECLS
rre

196 circuit, however, even small amounts of air may have disastrous consequences imposing high morbidity
197 and mortality on the patient including stroke, myocardial ischemia and infarction of other organs. The
198 embolus causes end-artery occlusion with subsequent ischemia and necrosis. Whilst central, i.e. aortic,
co

199 or subclavian arterial cannulation amongst others provide the benefit of antegrade body perfusion air
200 within the arterial ECLS system may in these cases directly enter the coronary or brain feeding arter-
un

201 ies with subsequent vital consequences. In peripheral (i.e. femoral) arterial cannulation, a watershed
202 depending on the intrinsic ejection of the left ventricle is created [20].
203 Total volume and rate of air entrainment influence outcome [13]. In contrast to conventional cardiopul-
204 monary bypass systems in ECLS nowadays mainly centrifugal pumps are used. The negative pressure
205 within these systems can theoretically rise up to −300 mmHg which may on the one hand result in an
206 entry of even larger volumes of air into the circuit with the subsequent consequences. On the other hand,
207 massive negative pressure can result in an outfall of physically solved gases with subsequent occurrence
208 of gaseous microbubbles.
209 Despite early diagnosis and subsequent treatment catastrophic cardiovascular collapse may occur.
210 Therefore, adequate prevention of air embolism is of utmost importance.
211 In our analyses, all tested oxygenators were able to withhold small air bubbles to a certain extent.
212 Numerically, the oxygenator without air trap even showed a slightly greater potential to withhold these
213 microbubbles whereas – for unclear reasons – in this case substantially more bubbles had been detected
 Galley Proof 28/07/2016; 14:14 File: thc–1-thc1248.tex; BOKCTP/xhs p. 8

8 F. Born et al. / Potential impact of oxygenators with venous air trap on air embolism in veno-arterial ECLS

214 on the venous side when comparing with the other measurements. Since this capacity to withhold mi-
215 crobubbles thus not seems to depend on the existence of an air trap, rather the oxygenator’s membrane
216 itself accounts for this filtering aspect. It remains however unclear whether the numerically higher ca-
217 pacity of Oxy3 (ECCO, no VBT) to withhold in-range microbubbles is a true finding or whether the level
218 of remnant microbubbles (which was comparable in all cases) simply presents the maximum value of an
219 oxygenator’s membrane filtering capacity. The fact that a membrane oxygenator, although not designed
220 for it, is capable of removing air bubbles has also been pointed out by De Somer et al. in CPB [18]. In
221 case of gross macro air entering the venous line of the ECLS circuit, the oxygenator itself is not capable
222 of withholding the total bolus, however, passage of the pump and the membrane results in a breakup
223 or defragmentation of macro air with smaller boli or bubbles being able to pass the system onto the
224 arterial side. De Somer et al. highlighted that macro air does get fragmented within a CBP circuit before
225 reaching the arterial line and that consecutively venous line filtration might be a more effective way
226 than arterial line filtration [18]. In our experimental ECLS setup we thus saw more over range bubbles

n
227 on the arterial side in comparison to the number registered on the venous side and subsequently also

o
228 noted an increase of the total bubble volume in all cases. The bolus volume, in contrast, decreased in all

si
229 cases with an obvious superiority of oxygenators containing air traps. This is underlined by the signifi-
230 cant difference in the total air volume entrapped by oxygenators with in comparison to the one without

er
231 air trap. Of note, the bubble counter was not able to discriminate single bubbles in such a “showering
fv
232 situation”. Similarly, other ultrasound-based devices such as transcranial Doppler are pushed to their
233 diagnostic limits in these situations [6]. Likewise, recent testing showed that both, the BCC 200 – which
oo

234 was used in our setting – as well as the Emboli Detection and Classification (EDAC) quantifier (Luna
235 Technologies, Blacksburg, VA, USA), have limitations especially in extreme situations even though the
pr

236 latter tests had been performed in vivo and rollerpumps were used [18].
237 Our results clearly demonstrate that oxygenators with venous bubble trap at least in trapping macro
air show an obvious and significant advantage.
ed

238

239 Nonetheless, the fact that total filtering of air with the devices used obviously is not feasible is clini-
240 cally alarming. Even small amounts of air and microemboli may result in negative sequelae due to diffuse
ct

241 microischemia. This clearly underlines the importance of adequate monitoring and – moreover – of ef-
rre

242 ficient prevention. Bubble counters are able to detect air and can be combined with electronic clamps.
243 In contrast to other devices that have so far been used in extracorporeal circulation such as transcranial
co

244 Doppler the bubble counter detects potential deleterious air before entering the human body and thus
245 prior to potential harm having been caused [6]. Additionally, the bubble detector device can be used
un

246 constantly throughout the whole treatment period and is easy to handle in routine clinical settings with-
247 out additional expertise being necessary. If combined with an electronic clamp, air may be hindered
248 from entering the body’s circulation right after detection. Sudden discontinuation of extracorporeal life
249 support, however, especially in critically ill patients may result in immediate cardiovascular collapse
250 and death limiting routine usage of these clamps. Even though reimbursement of add-on devices is an
251 issue, in selected cases reduction of morbidity and mortality can result in a reduction of costs for both,
252 hospitals and health care systems.
253 To the best of our knowledge, no clear guidelines or recommendations on optimal central venous
254 catheter positioning in relation to the venous ECLS cannula exist. Importantly, sufficient venous drainage
255 needs to be achieved. We do recommend, however, routine x-ray acquisition after ECLS cannula or
256 central venous catheter insertion in order to rule out closest or “kissing” proximity which could also
257 result in impaired or delayed drug effects [21].
258 In conventional cardiopulmonary bypass, substantial effort has already been made to reduce gas em-
259 bolism. Besides precautions during priming, cannulation, connection and initiation of CPB, usage of
 Galley Proof 28/07/2016; 14:14 File: thc–1-thc1248.tex; BOKCTP/xhs p. 9

F. Born et al. / Potential impact of oxygenators with venous air trap on air embolism in veno-arterial ECLS 9

260 arterial filters and dynamic bubble traps has been proposed [15,22,23]. Even though not available for
261 ECLS, Gerriets et al. showed in vivo that usage of a dynamic bubble traps (DBT) in conventional CPB
262 to reduce gaseous microembolism is associated with an improved cognitive outcome which underlines
263 the clinical importance of a reduction of air embolism [23].
264 Similar standards of care should be established in ECLS therapy. Maximum caution during initiation
265 of ECLS is of utmost importance. Besides, highest attention whilst setting and managing central venous
266 lines as well as during drug and infusion administration is required. Air filtering devices attached in
267 sequence to central venous catheters may provide additional security [6]. Standard setups should include
268 oxygenators containing air traps. They are easy to implement, safe and suitable for daily care and use.
269 Besides, bubble counters can detect air and further increase patient safety. Usage of electronic clamps is –
270 as explained above – limited to selected cases, for example if there is an increased risk of air embolism
271 due to surgical, procedure- or intervention-related manipulation.

n
5. Limitations

o
272

si
273 The bubble counting device is primarily designed to detect microbubbles and not to measure macro air.

er
274 In such settings, the bolus volume is the most reliable parameter, which showed significant differences
275 when comparing oxygenators with and without air trap. However, examinations that are more specific
fv
276 are warranted.
277 Additionally, different oxygenators from several manufacturers were used to test the hypothesis. Theo-
oo

278 retically, factors unique to the oxygenators or their construction might influence the results. Furthermore,
279 in this preliminary analysis only three oxygenators were examined. Confirming analyses are thus needed.
pr

280 The analysis presented was performed using an in vitro model. Even though fluid rheology highly
281 resembled blood bubble detection in vivo might be compromised due to blood being an opaque fluid
ed

282 containing microparticles. Verification will thus be necessary.

283 Furthermore, clinically important, so far no clear information is available on at what point the negative
ct

284 pressure starts be a considerably dangerous factor. Additionally, influencing factors such as temperature
and blood flow velocity need to be investigated. Future in vitro, animal and later-on clinical in vivo trials
rre

285

286 are thus required to address these issues.

co

287 6. Conclusion
un

288 The optimal management of air embolism is prevention. Macro air emboli can be substantially reduced
289 by usage of oxygenators containing air traps; the capability to withhold microbubbles to a vast extent
290 seems to depend on the intrinsic oxygenator’s membrane. Bubble counting devices are capable to detect
291 air before entering the body.
292 The increase in ECLS procedures inherits the risk of an increase in ECLS associated complications.
293 Specific precautions and standards of care on air embolism should be established and added to existing
294 national and international guidelines.

295 References
296 [1] Guenther SP, Brunner S, Born F, Fischer M, Schramm R, Pichlmaier M, et al. When all else fails: Extracorporeal life
297 support in therapy-refractory cardiogenic shock. European Journal of Cardio-Thoracic Surgery: Official Journal of the
298 European Association for Cardio-Thoracic Surgery. 2015. DOI 10.1093/ejcts/ezv212.
 Galley Proof 28/07/2016; 14:14 File: thc–1-thc1248.tex; BOKCTP/xhs p. 10

10 F. Born et al. / Potential impact of oxygenators with venous air trap on air embolism in veno-arterial ECLS

299 [2] Wu MY, Lin PJ, Tsai FC, Chu JJ, Chang YS, Haung YK, et al. Postcardiotomy extracorporeal life support in adults: the
300 optimal duration of bridging to recovery. ASAIO journal (American Society for Artificial Internal Organs: 1992). 2009;
301 55(6): 608-13. DOI 10.1097/MAT.0b013e3181b899c0.
302 [3] Welz A, Welz C, Winkler K, Schiller W, Theuerkauf N. Extrakorporale Unterstützungssysteme in Herzchirurgie. Inten-
303 sivund Notfallmedizin.
304 [4] Paden ML, Conrad SA, Rycus PT, Thiagarajan RR. Extracorporeal Life Support Organization Registry Report 2012.
305 ASAIO Journal (American Society for Artificial Internal Organs: 1992). 2013; 59(3): 202-10. DOI 10.1097/MAT.0b013e
306 3182904a52.
307 [5] Beckmann A, Benk C, Beyersdorf F, Haimerl G, Merkle F, Mestres C, et al. Position article for the use of extracorporeal
308 life support in adult patients. European Journal of Cardio-Thoracic Surgery: Official Journal of the European Association
309 for Cardio-Thoracic Surgery. 2011; 40(3): 676-80. DOI 10.1016/j.ejcts.2011.05.011.
310 [6] Zanatta P, Forti A, Bosco E, Salvador L, Borsato M, Baldanzi F, et al. Microembolic signals and strategy to prevent gas
311 embolism during extracorporeal membrane oxygenation. Journal of Cardiothoracic Surgery. 2010; 5: 5. DOI 10.1186/
312 1749-8090-5-5.
313 [7] Hogetveit JO, Saatvedt K, Norum H, Kristiansen F, Elvebakk O, Dahle G, et al. Central venous catheters may be a
314 potential source of massive air emboli during vascular procedures involving extracorporeal circulation: an experimental
315 study. Perfusion. 2011; 26(4): 341-6. DOI 10.1177/0267659111403644.

n
316 [8] Burnside J, Gomez D, Preston TJ, Olshove VF, Jr., Phillips A. In-vitro quantification of gaseous microemboli in two

o
317 extracorporeal life support circuits. The Journal of Extra-Corporeal Technology. 2011; 43(3): 123-9.

si
318 [9] ELSO ELSO. ELSO Guidelines for Cardiopulmonary Extracorporeal Life Support, Version 1.3, Ann Arbor, MI, USA.
319 wwwelsonetorg. 2013.

er
320 [10] Nguyen TT, Biadillah Y, Mongrain R, Brunette J, Tardif JC, Bertrand OF. A method for matching the refractive index and
321 kinematic viscosity of a blood analog for flow visualization in hydraulic cardiovascular models. Journal of Biomechanical
fv
322 Engineering. 2004; 126(4): 529-35.
323 [11] Lim WL, Chew YT, Chew TC, Low HT. Pulsatile flow studies of a porcine bioprosthetic aortic valve in vitro: PIV
324 measurements and shear-induced blood damage. Journal of Biomechanics. 2001; 34(11): 1417-27.
oo

325 [12] Barak M, Katz Y. Microbubbles: pathophysiology and clinical implications. Chest. 2005; 128(4): 2918-32. DOI 10.1378/
326 chest.128.4.2918.
pr

327 [13] Mirski MA, Lele AV, Fitzsimmons L, Toung TJ. Diagnosis and treatment of vascular air embolism. Anesthesiology.
328 2007; 106(1): 164-77.
329 [14] Orebaugh SL. Venous air embolism: clinical and experimental considerations. Critical Care Medicine. 1992; 20(8):
ed

330 1169-77.
331 [15] Lou S, Ji B, Liu J, Yu K, Long C. Generation, detection and prevention of gaseous microemboli during cardiopulmonary
332 bypass procedure. Int J Artif Organs. 2011; 34(11): 1039-51. DOI 10.5301/ijao.5000010.
ct

333 [16] von Segesser LK, Kalejs M, Ferrari E, Bommeli S, Maunz O, Horisberger J, et al. Superior flow for bridge to life
334 with self-expanding venous cannulas. European Journal of Cardio-Thoracic Surgery: official Journal of the European
rre

335 Association for Cardio-Thoracic Surgery. 2009; 36(4): 665-9. DOI 10.1016/j.ejcts.2009.05.026.
336 [17] Sutton RG, Salatich A, Jegier B, Chabot D. A 2007 survey of extracorporeal life support members: personnel and
337 equipment. The Journal of Extra-Corporeal Technology. 2009; 41(3): 172-9.
co

338 [18] De Somer FM, Vetrano MR, Van Beeck JP, Van Nooten GJ. Extracorporeal bubbles: a word of caution. Interactive
339 Cardiovascular and Thoracic Surgery. 2010; 10(6): 995-1001. DOI 10.1510/icvts.2009.229088.
un

340 [19] van Hulst RA, Klein J, Lachmann B. Gas embolism: pathophysiology and treatment. Clinical Physiology and Functional
341 Imaging. 2003; 23(5): 237-46.
342 [20] Hoeper MM, Tudorache I, Kuhn C, Marsch G, Hartung D, Wiesner O, et al. Extracorporeal membrane oxygenation
343 watershed. Circulation. 2014; 130(10): 864-5. DOI 10.1161/circulationaha.114.011677.
344 [21] Reichelt A, Pichlmaier M, Hagl C, Khaladj N. Malposition of a venous extracorporal life support cannula in a patent
345 foramen ovale. European Journal of Cardio-Thoracic Surgery: Official Journal of the European Association for Cardio-
346 Thoracic Surgery. 2013; 43(2): 441. DOI 10.1093/ejcts/ezs476.
347 [22] Johagen D, Appelblad M, Svenmarker S. Can the oxygenator screen filter reduce gaseous microemboli? The Journal of
348 Extra-Corporeal Technology. 2014; 46(1): 60-6.
349 [23] Gerriets T, Schwarz N, Sammer G, Baehr J, Stolz E, Kaps M, et al. Protecting the brain from gaseous and solid micro-
350 emboli during coronary artery bypass grafting: a randomized controlled trial. European Heart Journal. 2010; 31(3):
351 360-8. DOI 10.1093/eurheartj/ehp178.