Key words
interventional procedures - ischemia/infarction - stent retriever - ADAPT - balloon
guide catheter
Purpose
In ischemic stroke caused by large-vessel occlusion, mechanical thrombectomy with
stent retrievers (SR) is the standard therapy, in addition to systemic lysis [1]
[2]
[3]
[4]
[5]. The different SRs seem to perform equally well [6]
[7]
[8]
[9]. Additional aspiration is thought to improve the effectiveness of thrombectomy [6]
[9]
[10]. Aspiration can be performed distally using an intermediate catheter or proximally
using a balloon guide catheter (BGC). Moreover, new intermediate catheters (i. e.,
SOFIA (Microvention, USA), ACE (Penumbra, USA)) have shown promising results in a
direct aspiration, first-pass technique (ADAPT) [11]
[12]. The ASTER trial showed that ADAPT was not superior to stent thrombectomy. The clinical
results did not show any significant differences [13].
Experimental studies on ADAPT and BGC under standardized conditions with a transparent
flow model may help to provide more objective results, while a special focus should
be placed on the influence of the clot composition and the interaction between the
vessel and the devices during each thrombectomy maneuver.
Materials and Methods
Thrombus models
Two different clot models were used for the experimental studies [14]
[15]
[16]:
-
An erythrocyte-rich (red) clot: fresh human blood was put in a Chandler loop system
to create a thrombus under dynamic conditions. The clots were cut into pieces 10 × 3.5 mm
in size.
-
A fibrin-rich (white) clot: citrated human blood was stored standing for 24 h. The
plasma, the buffy coat, and a few erythrocytes were aspirated. The sample was recalcified
and incubated for a further 72 h. The clots were cut into pieces 10 × 3.5 mm in size.
Flow model
The experiments were performed under standardized and physiological hemodynamic conditions.
The flow model consists of a custom-made, transparent silicon phantom (Elastrat, Switzerland)
of the right anterior circulation (common carotid artery, internal carotid artery,
external carotid artery, middle cerebral artery, and anterior cerebral artery), which
has a physiological curvature and inner diameters, including the proximal M2 and A2
segments. To decrease friction, the inside of the phantom was coated with Slippery-Liquid
(Elastrat Sarl, Switzerland) before each use. Glycerol-saline solution was used to
mimic blood (60/40 by volume saline/glycerin; 37 °C). A precise, programmable piston
pump (CompuFlow 1000, Shelley Medical Imaging Technologies, Canada) produced a physiological
flow velocity and flow profile, which was monitored by Doppler sonography (carotid
profile, velocity of 100 cm/s, pulse 60/min). Interposed resistances guaranteed physiological
pressure, which was monitored invasively (middle pressure 100 mmHg). The clots were
injected into the model and flowed into the M1 segment of the MCA. All thrombectomy
procedures were recorded on video (Full High Definition, 60 frames per second, Panasonic
HC-V250, Japan). Simultaneously, the video signal was broadcast live to a monitor
simulating an angio-suite-like setup.
Advanced Thrombectomy Techniques
The following recanalization techniques were used:
-
ADAPT: An intermediate aspiration catheter (SOFIA 5F, Microvention, USA) was placed
directly proximal to the occlusion site and continuous machine aspiration was performed
(Penumbra Aspiration Pump, USA). If the tip was blocked by the clot, the intermediate
catheter was pulled back into the long sheath (Arrow-Flex 6F, 90 cm, Teleflex, Ireland)
in the ICA. While withdrawing the distal catheter, we performed an additional aspiration
via the sheath using a 20-ml Luer-Lock syringe. If necessary, subsequent passes followed.
-
An SR with proximal flow arrest and aspiration (SR+BGC): A SR (4 × 20 mm Trevo ProVue,
Stryker, USA) was placed at the occlusion site over a microcatheter (Trevo 0.021,
Stryker, USA) such that the clot was in the proximal part of the working zone. A BGC
(Cello 7F, Covidien, USA) was placed in the subpetrous part of the ICA. Just before
retrieving the clot, the balloon of the BGC was inflated, creating complete proximal
flow arrest. Proximal machine aspiration was performed simultaneously (Penumbra Aspiration
Pump, USA) when the SR was retracted into the BGC. After the SR was pulled out of
the BGC, the balloon was deflated. If necessary, subsequent passes followed.
Analysis
The videos of the thrombectomy procedures were analyzed after the experiments. The
number of passes and the occurrence of distal emboli and emboli in new territories
were documented. Distal emboli and emboli in the anterior cerebral artery were counted
on the videos.
Results
Both the red clots and the white clots were successfully produced as described above.
[Fig. 1] shows two specimens of the different clots.
Fig. 1 Thrombus models: (*) Erythrocyte-rich (red) thrombus; (**) fibrin-rich (white) thrombus.
Abb. 1 Thrombus-Modelle: (*) Erythrozyten-reicher (roter) Thrombus, (**) Fibrin-reicher
(weißer) Thrombus.
Six experiments were conducted per clot model and thrombectomy technique, resulting
in a total of n = 24 experiments. [Table 1] shows a summary of the results.
Table 1
Summary of the results.
Tab. 1 Zusammenfassung der Ergebnisse.
|
clot model
|
technique
|
median and interquartile range number of passes
|
rate of distal emboli
|
total number of distal emboli
|
rate of embolism in new territories
|
|
red (erythrocyte-rich)
|
BGC+SR
|
1 (1–1.75)
|
4/6
|
5
|
1/6
|
|
ADAPT
|
1.5 (1–2)
|
2/6
|
2
|
1/6
|
|
white (fibrin-rich)
|
BGC+SR
|
1 (1–1)
|
1/6
|
1
|
0/6
|
|
ADAPT
|
1 (1–1)
|
3/6
|
6
|
0/6
|
ADAPT = a direct aspiration, first-pass technique; BGC = balloon guide catheter; SR = stent
retriever.
ADAPT = direkte Aspirationstechnik; BGC = Ballon-Führungskatheter; SR = Stent-Retriever.
[Fig. 2], [3], [4] show the interaction between the techniques and the thrombus in consecutive pictures
of live videos. The size of distal emboli and emboli in new territories was max. 3 mm.
Fig. 2 Direct aspiration in white clot. The 5F aspiration catheter is not capable of aspirating
the whole clot. The proximal part of the clot is stuck at the tip of the aspiration
catheter. The catheter has to be retracted, while the major part of the thrombus is
unprotected (black arrow).
Abb. 2 Direkte Aspiration bei weißem Thrombus. Der 5F-Aspirationskatheter ist nicht in der
Lage, den gesamten Thrombus einzusaugen. Der proximale Anteil des Thrombus verstopft
die Spitze des Aspirationskatheters. Der Katheter muss so zurückgezogen werden, während
der Großteil des Thrombus ungeschützt ist (schwarzer Pfeil).
Fig. 3 Direct aspiration in red clot. The thrombus is aspirated directly at the occlusion
site, no clogging of the aspiration catheter.
Abb. 3 Direkte Aspiration bei rotem Thrombus. Der Thrombus wird direkt an der Okklusionsstelle
aspiriert, kein Verstopfen des Aspirationskatheters.
Fig. 4 Stent retriever with Balloon Guide Catheter in red thrombus. During retraction of
the stent retriever, the clot rolls towards the retriever tip and is fragmented (thin
black arrows). When the retriever is pulled back into the BGC, a large part of the
clot (white thick arrow) loses contact with the retriever tip (black thick arrow).
Due to flow arrest and flow reversal in the ICA, this part of the clot is aspirated
into the BGC. (*) = inflated balloon.
Abb. 4 Stent-Retriever mit Ballon-Führungskatheter bei rotem Thrombus. Während des Zurückziehens
des Retrievers rollt der Thrombus nach distal entlang des Retrievers und wird fragmentiert
(dünne schwarze Pfeile). Als der Retriever in den Ballon-Führungskatheter zurückgezogen
wird, löst sich ein großes Fragment des Thrombus vom Retriever (dicker weißer Pfeil).
Aufgrund des Flussarrestes und der Flussumkehr in der ACI wird das Fragment durch
den Ballon-Führungskatheter aspiriert. (*) = inflatierter Ballon.
Thrombectomy in white clots
For white clots, BGC and SR showed a lower risk of distal emboli than ADAPT (BGC 1/6,
n = 1; statistically not significant). Using ADAPT, white clots could not be aspirated
entirely ([Fig. 2]) because the catheter was clogged every time and had to be retracted into the sheath,
pulling the clot behind. Thus, the distal part of the clots loosened, causing distal
emboli (ADAPT 3/6, n = 6).
The stent-thrombus interaction in the BGC group was rather superficial. The thrombus
rolled between the retriever and the vessel wall distally, eventually losing the clot
shortly before entering the BGC every time. With additional flow arrest, the lost
clot could be saved in the petrous ICA and then finally aspirated in the BGC before
the balloon was deflated.
There were no emboli in the ACA. In both groups just one pass was necessary for full
recanalization.
Thrombectomy in red clots
For red clots, ADAPT produced fewer distal emboli than BGC. The aspiration catheter
was able to retrieve the clot just at the occlusion site without becoming clogged
in 4 of 6 cases ([Fig. 3]). As soon as the clogged aspiration catheter had to be retracted, distal emboli
developed (ADAPT 2/6, n = 2; not statistically significant).
The rolling clot phenomenon occurred with red clots as it did in white clots also
when using the BGC ([Fig. 4]). The red clot was more fragile than the white clot, which caused clot fragmentation
by the stent retriever in a more distal location (already in the M1 segment or the
carotid T). Due to the distal location of the fragmentation, those fragments could
not be caught by aspiration via the BGC and caused distal emboli (BGC 4/6, n = 5).
In both groups emboli developed in the ACA (1/6, respectively). A maximum of two passes
was necessary in both techniques, thus showing no relevant differences.
Discussion
In this in vitro study, a difference in the effectiveness of thrombectomy could be
found depending on the clot composition. RBC-rich thrombi were removed more effectively
by ADAPT, and fibrin-rich thrombi by SR/BGC.
Constant physiological vessel architecture, hemodynamics, and thrombus composition
were considered important factors to achieve realistic results. However, the lack
of endothelium, coagulation factors, and blood cells apart from the thrombus represent
limitations in this study. We deliberately selected a transparent vessel and blood
model to visualize the thrombectomy procedure throughout all stages. We chose to film
the experiments and analyze the direct interaction between the thrombus and the given
technique. Chueh et al. published a similar study in 2016 [17], but performed the experiments under fluoroscopic control. However, like in a clinical
setting under fluoroscopy, what is happening mechanically during the thrombectomy
maneuver is not visible. Liebeskind et al. [14] described that retrieved clots mainly consisted of fibrin and red blood cells (RBC).
They categorized thrombi as fibrin-dominant (“white”) and RBC-dominant (“red”). Nevertheless,
the clots were highly heterogeneous even within those groups, with smooth transitions
from one to the other. In this study, we chose representative fibrin-rich and RBC-rich
clots made of human blood. Artificial additives such as barium were not used to keep
the thrombus as natural as possible, in contrast to other studies [10]
[17]. We are aware that not all possible clot compositions were investigated in our experimental
study. However, using more extreme clot compositions, we expected to find differences
in the effectiveness of the given thrombectomy techniques. Beyond that, there is a
major issue concerning clot composition. The composition of acute clots in ischemic
stroke remains unclear because existing histopathological analyses have mostly dealt
with retrieved, processed clots. Indeed, it is only possible to approximately determine
the actual composition.
Using machine aspiration, a continuous vacuum could be maintained throughout the whole
procedure. Chueh et al. used a 20-ml syringe, which makes changing or emptying of
the syringe necessary and a loss of vacuum possible and probable [17]. The ADAPT technique can hardly be applied without a good vacuum. In contrast, the
stent retriever technique is probably more robust with respect to less effective aspiration.
A higher rate of distal emboli was considered negative in this study. The actual biological
effect of those emboli cannot be simulated in a flow model. Nor is it clear how to
interpret the size of distal emboli, e. g., very small emboli (< 100 µm) might be
harmful because the occluded microcirculation probably has no collaterals, or harmless
because they are more prone to react to medical or intrinsic lysis. Finally, it is
not possible to predict whether a certain embolus will produce an infarction, which
depends on the collateral flow. Nevertheless, a reduction in the number of emboli
is very likely to reduce the probability of infarctions.
The results of our in vitro study are in line with other recently published clinical
thrombectomy studies. Brinjikji et al. [18] suggested that BGC use during mechanical thrombectomy for acute ischemic stroke
is associated with superior clinical and angiographic outcomes. Teleb [19] concluded that the use of BGC was associated with a good first-pass effect and an
overall recanalization of TICI 2b/3 of 94 %. Maegerlein et al. [20] and Turk et al. [21] concluded that ADAPT is a very successful approach to revascularization and shows
advantages in comparison to the SR with respect to procedure time and safety, at a
similar clinical outcome. In the PROMISE study [22], the ACE68 / ACE64 catheters for aspiration thrombectomy were found to be safe and
showed similar efficacy to randomized trials using other revascularization techniques.
The ASTER study [13] showed very similar results with respect to both recanalization rate and clinical
outcome comparing ADAPT and SR combined with BGC. Since a randomized patient population
was treated and thus an average clot was retrieved, those results do not contradict
the findings in our study. Overall both techniques performed well in the flow model
and in the clinical setting.
In this study the occurrence of distal emboli using ADAPT depended strictly on whether
the aspiration catheter was clogged. If the catheter was clogged, the distal part
of the clot was pulled back into the guide catheter in the ICA with no protection
and parts of the thrombus were more likely to be lost. In a clinical setting, the
clogging can be easily recognized, checking whether there is any backflow during the
aspiration, and the results (occurrence of distal emboli) could be checked by post-treatment
MRI.
An important observation in the SR+BGC group was that the clot always rolled between
the SR and the vessel wall. The longer the distance over which the thrombus had to
be retracted, the more the clot moved towards the tip of the SR and was eventually
lost in many cases. These findings were already described in two previous independent
studies by Madjidyar et al. [6] and Machi et al. [23]. To prevent distal embolization and to increase the number of successful first-pass
recanalizations, an advanced thrombectomy technique with proximal flow arrest or an
additional aspiration seems to be crucial.
Conclusion
Although many physiological aspects were incorporated in the model, the experiments
were still performed in an artificial environment. This limitation must be considered
when interpreting our results. Nevertheless, this study sheds some light on the blind
side of mechanical thrombectomy, especially on device-thrombus interaction. Knowing
what happens between pre- and postintervention angiograms might reveal how thrombectomy
could be taken to a higher level. The selection of the device and technique might
depend on the thrombus composition. Thus, preinterventional clot imaging and categorization
could become a key for choosing the right recanalization technique. For the moment
the clot attenuation on non-enhanced CT (NECT) is promising. There is a link between
the density of the clot and the histological composition [24]
[25]. Since almost every patient undergoes NECT, the workflow would not be significantly
affected and no delay of therapy would occur. Aspiration catheters or balloon guide
catheters with larger lumina might increase the effectiveness.
