Key words thrombectomy - stroke - stent retriever
Introduction
Mechanical thrombectomy with stent retrievers has emerged as an effective and safe
therapy in acute ischemic anterior circulation stroke with thromboembolic proximal
vessel occlusion [1 ]
[2 ]
[3 ]
[4 ]
[5 ]. Various stent retrievers are in use, differing in stent design and mechanical properties
[6 ]
[7 ]
[8 ]
[9 ]
[10 ]
[11 ]
[12 ]
[13 ]. The exact mechanisms of device-vessel wall and device-thrombus interaction are
not fully understood yet and probably depend on stent design and behavior during retraction,
vessel anatomy, histological and mechanical properties of the thrombus [14 ]
[15 ]
[16 ]
[17 ]. However, it is necessary to understand which stent retriever is most efficient
in a given clinical situation. It has to be taken into account that the average age
of patients with ischemic stroke is older than 70 years [18 ]
[19 ]. During thrombectomy, the neurointerventionalist often encounters tortuous intra-
and extracranial vessels, which are atherosclerotically altered and probably more
vulnerable. The goal of this study was to compare the behavior, efficacy and safety
of commercially available stent retrievers and 2 prototypes using an in vitro-flow
model with a curved M1 segment.
Materials and Methods
Flow model
The stent retrievers were evaluated in a self-constructed closed-circuit silicone
model. It comprises a tube with a 4 mm diameter to simulate the terminal carotid segment,
a carotid T with an angle of 150°, a tapered tube with a proximal diameter of 2.5 mm
and a distal diameter of 2 mm to simulate the M1 segment. The M1 segment was arranged
in a c-shape simulating a curved vessel anatomy. The system was filled with saline
at body temperature. Blood circulation was simulated by a pump producing a pulsatile
flow at a frequency of 70 pulses per minute maintaining a pressure of 70 mmHg. A filter
was placed in the distal outflow tract to retain thrombus fragments.
Thrombi
Standardized whole blood thrombi with a diameter of 3 mm were generated in a Chandler
Loop from human blood as described before [20 ]
[21 ] and cut into fragments with a length of 15 mm. The thrombi were aspirated together
with saline into a 20 ml syringe and injected into the internal carotid artery segment
of the silicone model. After that, the A1 segment was clamped and the pump engaged
until the thrombus had migrated into the M1 segment. The clamp was removed and the
thrombectomy maneuver started.
Stent retrievers
Six commercially available stent retrievers were compared:
Aperio (Acandis, Pforzheim, Germany) 4.5 × 40 mm, Separator 3 D (Penumbra, Almeda,
USA) 4.5 × 26 mm, pREset (Phenox, Bochum, Germany) 4 × 20 mm, Revive SE (Codman Neuro,
Raynham, USA), 4.5 × 22 mm, Solitaire FR (ev3, Irvine, USA), 4 × 20 mm and Trevo (Concentric
Medical, Mountain View, USA) 4 × 20 mm.
Prototypes
We tested two NiTinol stent retriever prototypes (Acquandas GmbH, Kiel, Germany) with
a generic cell design, which were manufactured using a unique fabrication based on
UV lithography, sacrificial layer and wet etching technology [22 ]. These devices were designed with special features: 1) tapered (2.0 – 3.5 × 35 mm);
2) thin struts (Prototype I – 40 µm, Prototype II – 30 µm) ([Fig. 1 ]).
Fig. 1 Prototype I (40 µm stent struts, tapered design, 2.0 – 3.5 × 35 mm) with thrombus.Abb. 1 Prototyp I (40 µm Streben, konisches Design, 2,0 – 3,5 × 35 mm) mit Thrombus.
Thrombectomy maneuver
A distal access catheter (DAC 057, Concentric Medical, Mountain View, USA) was advanced
through an introducer sheath (6F Radiofocus Introducer 2, Terumo, USA) to a position
3 cm proximal to the thrombus. A microcatheter containing the stent retriever was
navigated through the thrombus to a position with the tip 0.5 cm distal to the thrombus
and the stent retriever was deployed. After 3 minutes, the stent retriever and microcatheter
were retracted while applying negative flow through aspiration at the distal access
catheter.
Six thrombi were extracted with each stent retriever. If the thrombus was not removed
on the first attempt, two more attempts were allowed.
Contact of the stent struts with the wall of the curved M1 segment was observed during
retraction.
Retraction force measurements
The force that had to be applied to retract the stent retriever was measured with
a spring scale (Micro Line 20 001, Pesola, Schindellegi, Switzerland) in a separate
experiment using the same flow model without aspiration. The spring scale was connected
to the microcatheter via a crocodile clamp. The maximum force that was necessary to
retract the microcatheter and stent retriever through the M1 segment was documented.
Thrombus fragmentation and embolism
After each retraction maneuver, the filter was checked for embolized thrombus fragments.
Fragments > = 1 mm (macroemboli) and < 1 mm (microemboli) were separately counted.
Statistics
Results of retraction force measurements were tested for normal distribution with
the Shapiro-Wilk test, followed by an analysis of variance. Significance levels were
determined using the Scheffé procedure and significance was set at p < 0.05.
Results
Aperio, Separator and pREset lined up with the wall of the curved M1 segment along
their entire length ([Fig. 2 ]).
Fig. 2 Wall-stent apposition in the a Aperio, b Separator and c pREset. Complete contact of the stent retrievers with the tube wall is noted.Abb. 2 Wand-Stent-Apposition beim a Aperio, b Separator und c pREset. Es besteht ein vollständiger Kontakt des Stentes mit der Schlauchwand.
In contrast, the inner stent struts of the bent Revive were not in contact with the
wall at the concave aspect of the tube curvature.
Solitaire and Trevo both showed malapposition at the convex aspect of the vessel ([Fig. 3 ]).
Fig. 3 Wall-stent apposition in the a Revive, b Solitaire and c Trevo. There is incomplete apposition of the inner struts in the Revive a . Solitaire b and Trevo c show a gap between struts and the wall at the outer aspect of the tube curvature.Abb. 3 Wand-Stent-Apposition beim a Revive, b Solitaire und c Trevo. Es besteht lediglich ein partieller Kontakt der inneren Streben beim Revive
a . Solitaire b und Trevo c zeigen eine Lücke zwischen Streben und äußerer Kurvatur der Schlauchwand.
The prototypes apposed the wall only with the broader proximal part.
All stent retrievers including the prototypes were able to remove the clot after a
maximum of 2 attempts. Aperio, Separator and pREset always removed the thrombus in
the first pass.
Macroemboli did not occur with the Aperio and Separator stent retrievers. Most macroemboli
were observed using the Trevo. Revive (n = 1), Solitaire (n = 1) and Trevo (n = 2)
lost the complete thrombus without fragmentation during retraction resulting in embolization
to the original occlusion site.
The 2 prototypes did not differ in number of macroemboli (n = 1).
Microembolism occurred in all stent retrievers including the prototypes ([Table 1 ]).
Table 1
First pass recanalization rate, number of macro- and microemboli in commercial stent
retrievers and prototypes.Tab. 1 First pass-Rekanalisierungsrate, Anzahl der Makro- und Mikroemboli bei kommerziellen
Stentretrievern und Prototypen.
first pass recanalization (%)
macroemboli (n)
microemboli (n)
Aperio
100
0
2
Separator
100
0
1
pREset
100
1
4
Revive
83
1
3
Solitaire
83
1
2
Trevo
67
3
2
Prototype I
83
1
4
Prototype II
83
1
4
A significantly higher retraction force was necessary for pREset and Separator compared
to the other commercial stent retrievers. However, the highest retraction force was
observed with the 40 µm prototype. The 30 µm prototype ranked second compared to the
commercial stent retrievers ([Fig. 4 ]).
Fig. 4 Retraction force of the commercially available stent retrievers and prototypes. Retraction
force of the pREset and Separator was significantly higher than retraction force of
the Revive, Solitaire and Trevo stent retrievers (p < 0.05). Prototype II with thinner
struts had a lower retraction force than Prototype I.Abb. 4 Retraktionskraft bei den kommerziellen Stentretrievern und Prototypen. Die aufzubringende
Retraktionskraft bei pREset und Separator war signifikant höher als bei Revive, Solitaire
and Trevo (p < 0,05). Bei Prototyp II mit dünneren Stentstreben war eine geringere
Retraktionskraft notwendig als bei Prototyp I.
Discussion
Thrombectomy with NiTinol stent retrievers has undoubtedly brought about a significant
improvement in anterior circulation stroke therapy. Various stent retrievers are available
differing in strut arrangement, strut thickness and radial force. Moreover, recanalization
rates most probably not only depend on stent retriever design but on mechanical and
histologic properties of the thrombus, possibly altered by thrombolytic agents or
platelet inhibitors, thrombus interaction with the vessel wall, relapsed time since
occlusion onset, thrombus length, occlusion site and vessel anatomy. This leads to
a complex clinical situation and the demand for an individually adjusted therapy.
Few in vivo studies examining the mechanical properties, recanalization, and embolization
rate of stent retrievers have been conducted in the recent past [9 ]
[23 ]
[24 ].
However, these studies focused on clot-stent interaction, in contrast to our study,
where stent-vessel wall alignment was evaluated.
Machi et al. [25 ] conducted a comprehensive evaluation comprising mechanical and functional tests
of all available stent retrievers. They performed a visual assessment of strut alignment
with the vessel wall with the results differing slightly from the results of our study.
The authors described full adaptation of the stent to the vessel wall with the pREset
4 × 20 mm in accordance with our results. In contrast to that, they also described
full apposition of the Solitaire 4 × 20 mm, which our observations did not confirm.
The group used a rigid vascular model with a straight M1 segment and described stent
adaptation throughout the entire retraction maneuver covering also more proximal vessel
segments with larger diameters. In our model, however, we simulated a curved M1 segment
in a non-rigid vascular model and focused our observations of stent apposition solely
on the clot-bearing segment.
To our knowledge, there are no previous studies evaluating stent retriever prototypes.
Since they are generic designs, it is possible to fabricate stent retrievers with
any geometry and surface patterning. Using our own prototypes enabled us to study
the influence of strut thickness on the mechanical behavior of a thrombectomy device.
In this in vitro study, we isolated the effect of stent retriever design on efficacy
and safety in a curved M1 segment with a minimum diameter of 2 mm. The retrievers
with complete stent apposition had the highest first pass recanalization success rate
(Aperio, Separator and pREset) and did not cause macroemboli (Aperio and Separator).
Close contact of the entire stent structure to the vessel wall seems to be an important
factor to ensure complete thrombus retraction and prevent thrombus fragmentation.
On the other hand, the highest retraction force had to be applied using these stent
retrievers. The higher force needed for retraction relates to the higher radial force
of the stent and possibly the higher risk of wall damage. It has been suggested that
vessel wall damage occurs during thrombectomy with stent retrievers, which possibly
results in secondary stenosis [26 ]
[27 ]
[28 ]
[29 ]
[30 ]. However, another study could not find any evidence of significant wall trauma [31 ]. Still, the question of whether vessel wall injury leads to clinically significant
complications has not been fully explored yet [32 ]
[33 ]. A reduction of strut thickness in our prototype led to a significantly reduced
retraction force while maintaining an excellent recanalization and low embolization
rate, indicating the fact that modifications of the stent design resulting in reduced
radial force do not necessarily negatively affect recanalization efficiency.
A major drawback of this study is the in vitro nature. The flow model closely simulates
in vivo conditions but does not incorporate the possible influence of the endothelium
on stent-thrombus interaction. Furthermore, it remains uncertain whether the results
can be transferred to the occlusion of M2 or more distal segments with a smaller diameter.
Rapid development in the field of mechanical thrombectomy continuously provides newer
generation stent retrievers, and the results may not be applicable to these devices.
Various thrombectomy techniques are in clinical use today. In our study, we performed
thrombectomy with simultaneous aspiration through a distal access catheter. This technique
has been proven effective in in vitro and clinical studies [24 ]
[34 ]. We did not consider other techniques such as proximal flow arrest using a balloon
guide catheter.
In conclusion, the results show that choosing a stent retriever with complete vessel
wall-stent apposition even in a curved vessel segment maximizes the success rate and
minimizes the embolization rate, but correlates with a higher retraction force and
thus possibly a higher risk of endothelium damage. Modifications in stent retriever
prototype design leading to a reduced retraction force did not compromise efficacy
and embolism rate.
Clinical relevance of this study
Understanding the behavior of a stent retriever during deployment and retraction is
crucial for the choice of the appropriate device in a given clinical situation.
It is pivotal to elucidate the properties of different stent retrievers to assess
the risk of thrombectomy therapy.
Understanding the impact of changes to stent retriever design aids in the development
of new devices.
