Investigation of Experimental Endovascular Air Embolisms Using a New Model for the Generation and Detection of Highly Calibrated Micro Air Bubbles

Abstract

Background:

Air embolism (AE), especially when affecting the brain, is an underrated and potentially life-threatening complication in various endovascular interventions. This study aims to investigate experimental AEs using a new model to generate micro air bubbles (MAB), to assess the impact of a catheter on these MAB, and to demonstrate the applicability of this model in vivo.

Materials and Methods:

Micro air bubbles were created using a system based on microfluidic channels. The MAB were detected and analyzed automatically. Micro air bubbles, with a target size of 85 µm, were generated and injected through a microcatheter. The MAB diameters proximal and distal to the catheter were assessed and compared. In a subsequent in vivo application, 2000 MAB were injected into the aorta (at the aortic valve) and into the common carotid artery (CCA) of a rat, respectively, using a microcatheter, resembling AE occurring during cardiovascular interventions.

Results:

Micro air bubbles with a highly calibrated size could be successfully generated (median: 85.5 µm, SD 1.9 µm). After passage of the microcatheter, the MAB were similar in diameter (median: 86.6 µm) but at a lower number (60.1% of the injected MAB) and a substantially higher scattering of diameters (SD 29.6 µm). In vivo injection of MAB into the aorta resulted in cerebral microinfarctions in both hemispheres, whereas injection into the CCA caused exclusively ipsilateral microinfarctions.

Conclusion:

Using this new AE model, MAB can be generated precisely and reproducibly, resulting in cerebral microinfarctions. This model is feasible for further studies on the pathophysiology and prevention of AE in cardiovascular procedures.

Introduction

Iatrogenic air embolism (AE) is an underrated and potentially life-threatening complication in various endovascular procedures. ¹ Micro AE is frequent and mostly clinically silent, whereas the injection of larger air bubbles may result in severe clinical sequelae.^2–4

Because of the poor collateralization of the terminal vasculature and the short ischemic tolerance, arterial AE to the brain is of particular importance as air obstruction will result in ischemic brain infarction. Depending on the size and amount of air bubbles, AE can have various effects, reaching from microinfarction without any clinical sequalae, or only subtle neurocognitive deficits, to potentially fatal cerebral ischemic strokes.^1,5–7 So far, it is unknown whether there is a critical size of air bubbles causing irreversible brain injury. Besides the size of bubbles with subsequent brain ischemia, the localization on a critical, eloquent brain area is crucial for an overt neurological deficit. ⁴

Arterial AE with potential embolization to the brain can occur during any endovascular catheter-based diagnostic or therapeutic intervention on the ascending aorta or the aortic arch, such as thoracic endovascular aneurysm repair (TEVAR),^8–11 on the vessels supplying the heart (eg, heart valve interventions),^12,13 or the brain (eg, treatment of ischemic stroke, intracranial aneurysms, and vascular malformations).^4,14,15 Similarly, open surgical interventions can cause arterial embolism to the brain, with cardiopulmonary bypass surgery being one of the riskiest procedures.^16,17 For TEVAR, the risk of stroke has been reported to be 2% to 6%, and up to 10% in some studies, which led to the investigation of techniques for the reduction of AE, which is one of the suspected reasons for these strokes.^18,19 It was shown that novel flushing techniques, mainly by using carbon dioxide, can effectively reduce the amount of AE during TEVAR.^18,20–22

Several studies investigated the occurrence of cerebral AE during cardiac and craniocervical endovascular interventions, mostly by using transcranial doppler ultrasound, but the size of the air bubbles was only assessed in one clinical study.^13,23–26 Chung et al ¹³ measured the bubbles and demonstrated that, during various cardiosurgical interventions, most bubbles that enter the brain are so-called micro air bubbles (MAB), which are defined as air bubbles with a diameter smaller than 100 µm. These results could be confirmed by Makaloski et al ²⁶ who investigated air bubbles during thoracic stent-graft deployment in a flow model and observed that the majority of air bubbles are MAB with diameters between 10 and 100 µm.

Despite the frequency and high clinical impact of cerebral AE caused by cardiac or craniocervical vascular procedures, only a few studies addressed the underlying pathomechanisms and experimental studies on the mechanisms of the formation of the air bubbles and their effect on the brain on the microvascular and histopathological level are rare.^2,12,14,27 Similarly, only a few reports are available on MAB generated manually or by a bubble generator injected to the brain vessels and the correlation between MAB size and number as well as size of brain infarctions.^28–32 The impact of calibrated MAB injected through catheter in the vasculature, resembling the clinical situation, was not addressed before.

The aim of this study was to establish a new model for the generation of calibrated MAB, to assess the impact of a catheter on the number and size of the generated MAB and to demonstrate the principal applicability of this new model in an in vivo small animal model.

Materials and Methods

In Vitro Generation and Detection of MAB

The model and the experimental setting are illustrated in Figure 1. The model consists of a bubble generation part and a bubble detection part.

Figure 1.

Schematic illustration of the model and the experimental setting. Blue arrows: diluted contrast agent, black arrows: air, dashed blue/black arrows: air bubbles in diluted contrast agent, and green arrow: microcatheter.

The bubble generation part includes a microfluidic flow controller (OB1 MK3+ microfluidic flow controller, Elveflow, Paris, France), which works similar to a pump and is connected to a cylinder containing compressed air with an interconnected pressure reducer. The pump pressure of the respective channel can be manually adjusted using the software of the controller. The flow controller has 2 outflow channels that are connected to 2 falcon tubes, of which 1 contains air and 1 contains 80% iodinated contrast agent (Iohexol 350 mg/ml, ACCUPAQUE-350, GE Healthcare, Boston) diluted with 0.9% saline (NaCl 0.9%, Serumwerk Bernburg Ag, Bernburg, Germany). We used a mixture of saline and contrast agent as medium instead of saline only in contrast to most other AE studies.^29–31 This resembles the clinical situation much better and due to the high viscosity of the iodine contrast agent, the number of bubbles per second can be reduced and can thus be controlled much more precisely. Moreover, the bubbles are also more stable in diluted contrast agent. The falcon tubes are each equipped with an outflow channel that is connected to a microfluidic channel system (MCS) in which the air bubbles. The MCS was specifically designed and manufactured for this study by Sensific GmbH (Ulm, Germany). For the connection between the gas cylinder, microfluidic flow controller, and the falcons we used polytetrafluoroethylene (PTFE) tubes with an inner diameter of 6 mm and an outer diameter of 8 mm. The connection between the falcons and the MCS was established by PTFE tubes with an inner diameter of 0.5 mm and an outer diameter of 1.0 mm. The MCS consists of polydimethylsiloxane (PSMD) and contains one channel for the inlet of contrast agent, one channel for the inlet of air, and one channel for the outlet of the air bubbles. Both channels are connected through a rectangular intersection point, as shown in Figure 1. A video that shows the generation of the air bubbles is available as supplemental material (original speed, air bubble for demonstration purposes colorized in blue, and MCS displayed in the background). For the generation of MAB, the pressure of the microfluidic flow controller has to be adjusted, depending on the desired size of the bubbles. In our study, we aimed to generate MAB with a size of 85 µm, which can be achieved by a pressure of 600 mbar for the diluted contrast agent channel and 300 mbar for the air channel, using an MCS with a channel diameter of 64 × 11 µm. For MAB with another size, different pressures and/or a different MCS will be required.

The bubble detection part consists of a microscope (CKX53; Olympus, Hamburg, Germany) that is connected to ODIN (Sensific GmbH, Ulm, Germany), a special commercially available system that allows the automatic assessment of MAB and of particular agents. An integrated high-speed camera with an on-chip evaluation unit automatically detects droplets by analyzing the inflow and outflow of pixel brightness. Once a bubble is detected, the system autonomously starts recording. Thresholding and averaging are used to calculate the length and width of the droplet with subpixel resolution. For each bubble, the determined parameters and respective images are saved. A counter can be set for live counting of the generated bubbles. ³³ The microfluidic channel in which the bubbles are formed can be observed live during the generation of the MAB. The MAB are compressed in the MCS as shown in Figure 1 and in the Supplemental Video. The bubble relaxes when it arrives in the outlet tube and gets a round shape due to the surface tension. To determine the relaxed diameter of the elongated MAB, the volume V of the bubble can be calculated using the formula proposed by Musterd et al, ³⁴ with the aid of the height H and width W of the specific channel and the length L of the bubble, as determined by the ODIN system.

$$\begin{array}{l}Volumeofthe\\ relaxedMAB=\left[\begin{array}{l}H\cdot W-\left(4-\pi \right)\\ {\left(\frac{2}{H}+\frac{2}{W}\right)}^{-2}-c\cdot H^2\end{array}\right]\left(L-\frac{W}{3}\right)\end{array}$$

To evaluate the relaxed diameter, we used the radius-formula of a sphere using the volume.

$$DiameteroftherelaxedMAB=2x{\left(\frac{3}{4\pi }V\right)}^{\frac{1}{3}}$$

The Degree of Calibration of the MAB

To investigate the degree of calibration of the MAB, 8 MCS were used to generate 50 (n=3), 100 (n=3), and 2000 (n=2) MAB. The MAB were automatically counted and assessed using the ODIN system. The mean, median, SD, minimum, maximum, and the coefficient of variation of the diameters of the MAB were calculated.

Impact of a Catheter on the Number and Size of MAB

For the assessment of the impact of a catheter on the MAB, the outflow tube of the MCS (also a 0.5 mm PTFE tube) was connected to a commercially available microcatheter with a length of 156 cm and an inner diameter of 0.0165 inch / 0.419 mm, which is a commonly used catheter for neurointerventional procedures (Headway Duo; MicroVention, Aliso Viejo, California). To reduce the dead space of the system and to prevent an influence of surface transitions on the MAB, a manually modified syringe–catheter interface adapter (MicroVention) was positioned between the tube and the catheter. After the connection of the tube to the catheter, the proximal part of the tube was disconnected from the MCS and the MAB were manually injected into the catheter using a 1.0 mL syringe with a flow of 0.1 mL per second. The catheter tip was positioned within a petri dish filled with 80% contrast agent under the microscope camera system and a high-resolution video with 75 frames per second was recorded during the injection. Two PTFE tubes were positioned side by side parallel to the microcatheter, 3 mm proximal to the catheter tip, and were flushed with 80% contrast agent with a flow of 1.0 mL per second during the injection to simulate blood flow and to prevent an accumulation of air bubble at the catheter tip. The diameters of the MAB at the catheter tip were measured using the open-source software ImageJ (https://imagej.nih.gov) during the injection of 50 (n=3) and 100 (n=3) MAB (see the section “The Degree of Calibration of the MAB”). The same descriptive statistical parameters were calculated for the MAB distal to the catheter. The diameters of the MAB proximal and distal to the microcatheter were statistically compared using the Welch test, including an F test, to compare the variances. The p values of 0.05 were defined as the threshold for statistical significance and were not adjusted for multiple testing due to the fact that this is a hypothesis-generating study. GraphPad Prism (La Jolla, California; Version: 7.04) was used for this statistical analysis.

In Vivo Example

To demonstrate the general applicability of the model, a defined number of MAB were generated and injected into craniocervical circulation in an in vivo small animal model.

Animal Preparation and Application of the MAB

All experiments were performed in accordance with the Guide for the Care and Use of Laboratory Animals (Directive 2010/63/EU of the European Parliament). State animal care and ethics committee approval was obtained (registry number: 35-9185.81/G-183/20).

Two male Wistar rats (Janvier Laboratories, Le Genet-St-Isle, France) were anesthetized with an intraperitoneal injection of 100 mg/kg ketamin (Ketamin 10%, Pharmanovo, Austria) and 5 mg/kg xylazine (Xylariem, Ecuphar GmbH, Germany) and positioned on a heating pad. The left femoral artery was exposed and after inserting a 26 G venous catheter (BD Neoflon), a 156 cm/0.0165 inch Headway Duo microcatheter (MicroVention) was inserted using the Seldinger technique with a 0.014 microguidewire (Traxcess 14; MicroVention). Using the same guidewire, the microcatheter, which was connected to a continuous pressure flush with 0.9% saline, was then positioned at the level of the aortic valve and the left common carotid artery, respectively, under fluoroscopic guidance (SIREMOBIL Compact L; Siemens Healthineers, Erlangen, Germany; illustrated in Figure 2). Similar to the ex vivo experiments, the hub of the microcatheter was subsequently connected to the MCS and 2000 MAB were injected into the vascular system.

Figure 2.

Illustration of the location of air bubble injection. A conventional microcatheter was navigated into the aortic arch at the level of the aortic valve (left; green catheter) or into the left common carotid artery (right; blue catheter) through a femoral access in 2 different animals. Afterward, micro air bubbles (illustrated as white circles) were injected.

Magnetic Resonance Imaging (MRI)

One hour after the intervention, MRI was performed with a 9.4 Tesla small animal MR system (Bruker Biospec 94/20; Bruker Biospin, Ettlingen, Germany). During the MRI, the rats were sedated by using inhalation anesthesia with a small animal anesthetic mask with an initial dose of 4% isoflurane and a maintenance dose of 2% isoflurane. They were kept on a heating pad to maintain a constant body temperature. The animals were breathing spontaneously and were monitored during the MRI with a breathing surface pad checked by an in-house developed LabView program (National Instruments, Munich, Germany). The following MRI sequences were performed: time-of-flight (TOF) angiography, T2-weighted sequences (with 0.5 and 0.1 mm slice thickness, respectively), diffusion-weighted sequences, T1-weighted sequences before and after the application of contrast agent, and a susceptibility-weighted sequence. T2-weighted sequences were performed as the first and last sequence to monitor any change in size or demarcation of infarctions and to detect infarctions that might develop during the acquisition. For contrast-enhanced sequences, 0.2 mmol/kg of gadoteric acid (Dotarem; Guerbet, Istanbul, Turkey) was administered intravenously through a venous tail vein catheter. Further details regarding the MRI protocol are presented in the Supplemental Table. After the MRI, the animals were finalized by an intraperitoneal injection of a ketamin/xylazin overdose (200 mg/kg ketamin and 10 mg/kg xylazine).

Results

The Degree of Calibration of the MAB

The descriptive statistics and size distributions of the generated MAB are summarized in Table 1 and Figure 3, respectively. The size of the generated MAB was very close to the projected 85.0 µm, with an average diameter of all generated MAB of 85.3 µm (mean)/85.5 µm (median), ranging from 78.3 to 92.6 µm. The bubble size was highly reproducible, with an average SD of 1.9 µm (for all MAB), ranging from 0.7 to 2.8 µm.

Table 1.

Summary of the Number and Diameters of the Generated and Detected MAB.

No. of generated MAB	50			100			2000 a	All b
Generated MAB injected into the microcatheter
Mean	85.3	86.5	87.1	87.2	86.0	87.0	85.1	85.3
Median	85.4	87.6	87.6	86.5	85.8	86.2	85.4	85.5
SD	0.7	2.5	2.8	2.3	2.5	2.8	1.7	1.9
Minimum	83.1	80.0	80.9	84.3	78.3	81.2	79.1	78.3
Maximum	87.7	89.8	92.6	92.4	91.7	91.3	88.4	92.6
Coefficient of variation	0.01	0.03	0.03	0.03	0.03	0.03	0.02	0.02
MAB detected distal to the microcatheter
Detected MAB c	32 (64.0%)	26 (52.0%)	35 (70.0%)	75 (75.0%)	52 (52.0%)	51 (51.0%)		271 (60.2%)
Mean	93.8	97.1	91.5	83.6	96.6	95.5	—	91.9
Median	88.6	87.4	88.6	79.8	97.2	86.3	—	86.6
SD	16.6	27.2	23.8	30.1	32.0	35.7	—	29.6
Minimum	72.8	68.8	42.6	33.7	35.0	35.0	—	33.7
Maximum	140.2	192.3	178.1	197.2	174.3	232.4	—	232.4
Coefficient of variation	0.18	0.28	0.26	0.36	0.33	0.37	—	0.32
Comparative statistics
p value Welch test d	0.007	0.059	0.282	0.300	0.270	0.093	—	—
p value F test e	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001	—	—

The diameters are indicated in µm.

Abbreviation: MAB, micro air bubbles.

^aThe 2000 MAB were injected into the vasculature of a rat and could therefore not be analyzed distal to the catheter.

^bn=4450 for the generated MAB, n=271 for the detected MAB.

^cAbsolute number (relative frequency in %).

^dComparison of the diameters of the injected and detected MAB.

^eComparison of the variance of the diameters of the injected and detected MAB.

Figure 3.

Illustration of the size of the generated and detected air bubbles. Note the very high reproducibility of the MAB diameter (blue dots) in all experiments, visible as only minimal deviation from the median (red horizontal line). The calibration of the generated and detected MAB was similar regarding the numbers of generated MAB (n=50 (left row), n=100 (middle row), or n=2000 (right upper corner). The frequency distribution of the MAB diameters is additionally illustrated in a histogram analysis (right lower corner). After the injection of the MAB into a microcatheter, the scattering of the diameter of the MAB was substantially higher (red dots) and the number of detected MAB was lower. MAB, micro air bubbles.

Impact of a Catheter on the Number and Size of MAB

An example video (played at 50% speed) of the MAB that were detected distal to the catheter tip is available as supplemental online material. A snapshot of MAB distal to the catheter is shown in Figure 4. In Figure 3, the size of the MAB that were detected distal to the microcatheter (red dots), in comparison with the MAB that were injected (blue dots), is also graphically illustrated. The results of the comparative statistics are summarized in Table 1.

Figure 4.

Snapshot of micro air bubbles (MAB) distal to the microcatheter. The distal tip of the microcatheter was filmed during injection of the MAB using a high-speed camera (75 frames per second). MAB of various sizes were detected distal to the microcatheter, with a substantially higher degree of scattering, compared with the highly calibrated MAB that were injected into the microcatheter.

Compared with the highly calibrated MAB that were injected into the microcatheter, a lower number of MAB was detected distal to the microcatheter (average: 60.2%). According to the Welch test, comparing the diameters of the MAB proximal and distal with the catheter, the size of the MAB was not significantly different for all but one experiment. However, there was a significantly higher degree of scattering of the diameters of the MAB, which were detected distal to the microcatheter, compared with the highly calibrated MAB that were injected proximally. Accordingly, the F test revealed a statistically significant difference between the variances of the MAB proximal and distal to the microcatheter for all experiments (p<0.001, respectively).

In Vivo Example

Both animals received exactly 2000 air bubbles with a total volume of 0.643 µl (MAB injected at the level of the aortic valve) and 0.676 µl (MAB injected into the left common carotid artery). The imaging findings of the in vivo injection of MAB are summarized in Figure 5. The MRI of the animal with application of the MAB at the level of the aortic valve showed a total number of 6 acute ischemic infarctions of the brain in the following regions: cerebral cortex (both sides), striatum (both sides), and hippocampus (left side). These infarctions varied in size, ranging from 0.6 to 1.8 mm in diameter, with the smallest infarction being located in the left striatum and the largest one being located in the right cortex. The MRI of the animal with application of the MAB into the left common carotid artery showed 6 acute ischemic infarctions in the ipsilateral hemisphere. They were also distributed in the cerebral cortex, striatum, and hippocampus, and ranged in size from 0.7 to 1.6 mm in diameter. Similarly, the smallest infarction was located in the striatum and the largest one in the cortex. These infarctions were visible as well-demarcated, T2-hyperintense regions, with increased signal in the diffusion-weighted imaging sequence, indicating acute ischemic tissue injury. In the TOF angiography, no macroscopically visible occlusion of a brain-supplying artery was observed. In the susceptibility-weighted sequence, no macroscopically visible air bubble, which would be visible as a hypointense lesion, was detected.

Figure 5.

Illustration of the magnetic resonance imaging (MRI) findings. The MRI findings after injection of air bubbles into the aorta (A–D) and into the carotid artery (E–H). In the T2-weighted sequences (A: 3D sequence, E: 2D sequence), several hyperintense lesions were detected in the hemispheres, in cortical and subcortical location, and in the basal ganglia (white arrows). Diffusion-weighted imaging (B, F) showed corresponding hyperintensities (white arrows; in B on this slice only clearly visible for the lesion in the right hemisphere). After injection of air bubble into the aorta, acute infarctions (illustrated as colored volume renderings) were detected in both hemispheres (C), whereas air bubbles injected into the carotid artery caused only ipsilateral infarctions (G). In both animals, time-of-flight angiography (D) did not show any obvious vascular occlusions and, in susceptibility-weighted imaging (H), no circumscribed areas of signal loss, consistent with air bubbles, were seen.

Discussion

AE is a frequent, yet poorly understood complication of vascular medical procedures, ranging from micro AE with diffuse tissue damage to macro AE with large vessel occlusion. So far, little is known on the exact pathomechanisms of vessel obstruction and pathophysiology of tissue damage, which is required for prevention and treatment of AE.

In this study, a new system for the generation and detection of calibrated MAB was established and tested in an in vitro catheter model and in vivo small animal model. Based on microfluidic channels and an automated analysis of air bubbles, MAB with the homogeneous size were generated and detected with high precision and reproducibility. We could show that a catheter has a major impact on the number and the size of MAB that are injected through it. The feasibility of this new system was successfully demonstrated in an exemplary in vivo experiment by injecting MAB at the level of the aortic valve and into the common carotid artery using a conventional microcatheter, simulating micro AE occurring during cardiovascular and neurovascular interventions.

Although macroscopically invisible MAB are frequently injected in the patient unintentionally during interventional procedures, there are few systematic data in literature on their formation, vessel occlusion, and subsequent tissue damage.^30–32

For a reliable and conclusive investigation of cerebral AE in experimental studies, it is crucial that air bubbles have a defined number and size. The lack of adequate systems for the generation and detection of air bubbles, and a suitable experimental model, is the most important obstacle for dedicated research in AE. So far, the only study that used a bubble generator that automatically generated a defined number of air bubbles with a specific size is the work by Gerriets et al. ³¹ All other reports, which were mainly experimental studies in small and large animals, showed that air bubbles were generated by manually mixing liquid and air in syringes.^28–30,32 Using this method, the volume of air is known, but it produces an unknown number of air bubbles of various sizes, which substantially limits the conclusions that can be made on studies with such bubbles. It is known that not only the volume of air influences the effects of the AE on the brain, but also the size of the bubbles. ³¹ It can be assumed that the infarction pattern is influenced by the bubble size because bubbles of different size will lodge in vessels of different diameter. A further drawback of most of the existing experimental studies is that macro air bubbles instead of MAB were used, whereas during catheter-based interventions, it was shown that the majority of the bubbles that occur are MAB with a diameter <100 µm.^8,13,30–32 In this regard, it has to be mentioned that there may be substantial differences in the size of air bubbles between the different procedures, such as TEVAR, cardiac interventions, and neurointerventional procedures.^8,13

This is the first study in which the femoral artery was used for access in a small animal model and a conventional microcatheter was used for the injection of the MAB, according to clinical practice. Previous studies that investigated AE in small animals used a direct surgical access to the internal carotid artery, mostly through the external carotid artery.^29,31 This open surgical intervention bears a risk for the generation of AE to the brain and can distort the study’s findings. Only a few large animal studies injected air bubbles using conventional catheters through a transfemoral access into the vasculature, simulating catheter-based interventions in humans.^30,32

In addition, until now, no study analyzed what happens to the air bubbles that flow through a catheter. It was expected that the air bubbles that are injected into the catheter leave it distally without any change. We could demonstrate that air bubbles that are injected through a catheter undergo a substantial change in number and size. We observed a lower number of MAB (60.2%) and a significantly greater scattering in the diameter of the bubbles reflected, for example, by the coefficient of variation that was 0.02 for the MAB that were injected into the microcatheter and 0.32 for the MAB that were detected distally. However, the average size of the MAB was not different (median: 85.5 vs. 86.6 µm). The most likely explanation for this phenomenon is the dispersion and coalescence of MAB on their way through the catheter. Every edge, such as the hub of the catheter, and any change in size can potentially lead to dispersion or to coalescence of MAB. These findings must be considered for the interpretation of the studies that were performed using this technique for the design and interpretation of future studies, and for endovascular interventions in clinical practice.

In an exemplary in vivo experiment, we demonstrated that the system that was investigated in this study is feasible for the research of cerebral micro AE. Various small cerebral infarctions of different sizes in both hemispheres were observed after the injection of MAB at the level of the aortic valve, whereas after injection into the carotid artery only infarctions in the ipsilateral hemisphere were detected. This model can thus be used for the systematic investigation of cerebral AE using highly calibrated MAB with conditions resembling catheter-based interventions in clinical practice. Further studies using this model are warranted to investigate further aspects of AE, such as procedural parameters, physical features, and histopathological findings.

We acknowledge that this study has some limitations. Only one size of MAB was investigated. Air bubbles of smaller or larger size may lead to different results and will be the topic of future studies. Furthermore, the number of performed experiments was relatively small. However, the results were highly consistent, when comparing the different experiments. Another drawback is that only one microcatheter was used for this study. A catheter with a different size or length could have different influences on the air bubbles. Only 2 exemplary in vivo experiments were included into this study to show that the calibrated MAB actually result in cerebral infarction. However, the focus of this proof of concept study was to establish and investigate this new model for future systemic research of AE.

Conclusion

Using a system based on microfluidic channels and an automated air bubble analysis, MAB with the same size can be generated precisely and reproducibly. These MAB can be used for the investigation of AE in experimental in vitro and in vivo models. When such MAB are injected through a microcatheter, this catheter has a major influence on the number and size of the bubbles.