New Embolization Coil Containing a Nitinol

Wire Core: Preliminary In Vitro and In Vivo

Experiences
András Kónya, MD, PhD,2 Michael Maxin, MD,1 and Kenneth C. Wright, PhD2

PURPOSE: To improve the stability of the standard stainless steel embolization coil by adding a nitinol wire core.
MATERIALS AND METHODS: With use of one coil with a nitinol wire core and one without one, stability and
resistance to elongation were measured in vitro. Thirty-one factory-made stainless-steel macrocoils equipped with
preshaped nitinol wire cores were acutely deployed into branches of the superior mesenteric artery (SMA) in five pigs
(part I of the study). Fifteen homemade retrievable coils with thermal shape memory (TSM) nitinol wire cores were
acutely deployed in the abdominal aortae or inferior venae cavae of four pigs (part II). Coils with a superelastic (SE)
nitinol wire core (n ⴝ 9), a TSM nitinel wire core (n ⴝ 5), and without a core (n ⴝ 5) were compared in carotid
embolization (part III).
RESULTS: In vitro, the expansile strength of the reinforced coils was significantly greater and elongation was
significantly less than the standard coils (P < .01). In part I, coils were easily deployed via diagnostic catheters. In part
II, ability to reposition the coils facilitated optimal coil configuration, which resulted in effective self-anchoring and
occlusion. In part III, no coils with a SE core migrated, whereas one of five with a TSM core and three of five without
reinforcement migrated immediately. The core significantly increased coil stability and postplacement configuration
(P < .01).
CONCLUSIONS: Addition of a nitinol wire core increases the intravascular stability of the standard embolization coil
by significantly enhancing expansile force and postplacement configuration. The wire core does not adversely affect
the handling of the coil. The ability to reposition/retrieve the coil is a desirable feature.

Index terms: Arteries, embolization • Embolization

J Vasc Interv Radiol 2001; 12:869 – 877

Abbreviations: IVC ⫽ inferior vena cava, SE ⫽ superelasticity, SMA ⫽ superior mesenteric artery, TSM ⫽ thermal shape memory

THE Gianturco stainless steel coil occluding device since the mid-1970s eliminate the risk of coil migration
(Cook, Bloomington, IN) has been the (1). The original design has undergone (11–13). For instance, conventional
most widely used mechanical vascular several modifications in size, configu- and retrievable spring coils have been
ration, thrombogenicity, and delivery successfully used for transcatheter clo-
(2–5). Recently, a retrievable macro- sure of patent ductus arteriosus, but
coil, the Jackson coil (William Cook their use has largely been limited to
From the Department of Diagnostic Radiology, the
University of Texas M.D. Anderson Cancer Center, Europe, Bjaeverskov, Denmark), has small and moderate-sized patent duc-
1515 Holcombe Blvd., Box 057, Houston, Texas become available for vessel occlusions tus arteriosus because of the physical
77030. Received October 20, 2000; revision requested throughout the body (6 –10). The re- limitation of stability (9,14 –17).
January 3, 2001; revision received February 7;
trievable version of the coil provides Several attempts have been made to
accepted February 12. Supported in part by a grant
from the John S. Dunn Research Foundation, the the ability to reposition or remove the prevent possible migration of the stain-
Cesare Gianturco Fund, and by grant NIH-NCI CA- coil if its position is not optimal, and less steel spring coil emboli. The “coil-
16672 from the National Cancer Institute. From the the risk of coil migration or distal em- in-coil” deployment technique uses a
2000 SCVIR Annual Meeting. Address correspon-
bolization is thereby minimized. How- smaller coil released into the lumen of a
dence to A.K., E-mail: [email protected]
1
Current address: Group Health Cooperative of ever, in nontapered vascular struc- larger coil to decrease the length of the
Puget Sound, Seattle, Washington. tures and veins, especially those with occluded vessel segment and prevent
2
This author has disclosed the existence of a poten- high-flow conditions, such as arterio- coil migration (18). A single barb has
tial conflict of interest.
venous malformations and fistulas, been attached to the coil to prevent mi-
© SCVIR, 2001 even optimal coil placement does not gration when placed on the venous side

869
870 • New Embolization Coil Containing a Nitinol Core July 2001 JVIR

(19), and the separately deployed Am- the homemade coils. Factory-manufac-
platz spider (Cook, Bloomington, IN) tured coils were placed back into their
was designed to prevent coil movement original metal loading cartridges,
in both the arterial (20) and venous sys- whereas the homemade coils were
tems (21). Recently, an anchoring sys- loaded into appropriately sized peel-
tem for stainless-steel Gianturco macro- away Teflon sheaths.
coils has been developed and tested in a
high-flow arterial model (22), and some In Vitro Studies
authors have proposed the use of a niti-
nol snare (23–25) or a balloon wedge Two homemade 0.032-inch coils 13
catheter (26) to prevent coil migration cm in length and 6 mm in diameter,
during patent ductus arteriosus one with a 0.011-inch SE nitinol wire
occlusion. core and one without it, were used to
This article reports the development test radial force in an in vitro model.
and initial testing of stainless-steel mac- To make the coil retrievable, a delivery
rocoils equipped with a nitinol wire core cable with a push-button release
to increase the coils’ self-anchoring ca- mechanism, designed for the deploy-
pability and decrease the incidence of ment of a Bird’s Nest filter (Cook), was
coil elongation after placement. In the- attached to the coil (Fig 1). The distal
ory, the presence of a programmed niti- end of the delivery cable consisted of a
nol wire core within the lumen of a movable notched cylinder covered by
stainless steel coil would enhance the a short cylindrical housing. To attach
coil’s expansile force and postplacement the coil to the delivery cable, a 2–3-mm
configuration. Improvement of these space was created within the last one
characteristics would increase the coil’s or two turns at the proximal end of the
intravascular stability. Standard stain- coil, and this space is where the
less steel coils were equipped with a notched cylinder at the distal end of
preprogrammed nitinol wire core and the delivery cable was attached. The
evaluated in vitro and in vivo with re- coil was then placed within the lumen
gard to handling, stability, occlusion ca- of a 6-mm internal diameter transpar-
pacity, and ability to resume their pre- ent polyvinyl chloride tube through a
programmed shape after deployment. 7-F Lumax guiding catheter (Cook).
Before each test, an attempt was made
to achieve the tightest possible coil
MATERIALS AND METHODS configuration by taking advantage of
Device Construction the ability to reposition the coils. After Figure 1. Schematic drawing showing the
placement, the delivery cable re- proximal end (push-button) and distal end
Factory-manufactured (Cook) and mained attached to the coil and the (notched cylinder with a short cylindrical
homemade stainless steel coils of differ- polyvinyl chloride tube was posi- housing) of the delivery cable. Activating
ent sizes were equipped with preshaped tioned vertically. A small plastic bas- the button exposes the notched cylinder
nitinol wire cores of varying size. The ket was attached to the delivery cable (left). When the button is released, a spring
nitinol wires (nickel-titanium alloy; Ni 15 cm from the coil, and lead weights housed in the proximal end of the cable
55% and Ti 45%; Shape Memory Appli- (1.8 g) were placed into the basket one pulls the notched cylinder with the proxi-
cations, Santa Clara, CA) were coiled at a time until the coil dislodged from mal end of the coil back into the short
cylindrical housing (right).
around cylindrical templates of differ- the tube (Fig 2). The total weight (in
ent diameters to form spirals. The wires grams) needed to dislodge the coil was
were restrained in the desired shape recorded. This evaluation was re-
and annealed at 500°C for 12–15 min- peated 20 times with each coil. Results were compared statistically
utes to achieve non–temperature-de- In a similar set of experiments, the with use of a two-sample Student t test
pendent superelasticity (SE) or 90 –120 coils were suspended from a metal bar assuming equal variances, and P val-
minutes to impart thermal shape mem- and the length of the coil was mea- ues ⱕ.05 were considered to be statis-
ory (TSM) with a transition temperature sured and recorded (Fig 3). A weight tically significant.
equal to body temperature. After cool- (1.8 g) was attached to the free end of
ing to room temperature, the nitinol the coil, and the length of the coil was Animal Studies
wire was placed in the lumen of a fac- again measured and recorded. This
tory-manufactured or homemade stain- evaluation was repeated 10 times with All experimentation involving ani-
less-steel coil. The nitinol core was se- each coil. Coil elongation was deter- mals was approved by the Institutional
cured to both ends of the stainless-steel mined by subtracting the pre-weight Animal Care and Use Committee of our
coil with silver solder. No modifications measurement from the post-weight institution. Animals were maintained in
were made to the polyester fibers on the measurement. Coil elongation was facilities approved by the Association
factory-made coils; polyester threads in used to evaluate the coil’s ability to for Assessment and Accreditation of
varying configurations were attached to regain its preprogrammed shape. Laboratory Animal Care International
Volume 12 Number 7 Kónya et al • 871

eter from 2.5 mm to 6 mm. Two and

three coils were also placed in the
main trunk of the SMA (7– 8-mm di-
ameter) in two pigs. Of the 31 coils, 10
were deployed into branches with a
diameter larger than 4.5 mm, 15 were
placed in vessels with a diameter of
3– 4 mm, and six were inserted into
branches that were smaller than 3 mm
in diameter. Hand injections of con-
trast material under fluoroscopic ob-
servation and repeat angiography
were used to document the position of
the coils. Within 1 hour after the last
coil placement, animals were killed
with an intravenous overdose (1.0
mL/lb body weight) of Beuthanasia
(Schering-Plough, Kenilworth, NJ)
and a partial necropsy was performed.

Part II: Occlusion Capability and

“Retrievability” under High Flow
Conditions
Figures 2, 3. (2) Measuring apparatus used for evaluating the self-anchoring ability of the
coils in vitro. After achieving the tightest possible coil configuration in a transparent polyvinyl A total of 15 0.032-inch homemade
chloride tube, the delivery cable remained attached to the coil and the tube was positioned coils 24 –34 cm in length (mean, 28 cm)
vertically. A small plastic basket was attached to the delivery cable and lead weights were were fabricated with 0.009-inch nitinol
placed into the basket one at a time until the coil dislodged from the tube. (3) Measuring wire cores with TSM. Six coils were
apparatus used for evaluating coil elongation. Two coils, one with nitinol reinforcement
(right) and one without reinforcement (left), which were identical in length and diameter,
produced in a tornado shape that
were suspended from a metal bar. A weight (1.8 g) was attached to the free end of each coil, gradually increased in diameter from
and the elongation of the coils was measured. 8 mm to 11 mm or from 16 mm to 24
mm. Nine additional coils were made
in a slightly irregular helix with
tighter loops that gradually increased
and in accordance with current U.S. De- equipped with a preshaped nitinol in diameter from 6.5 mm to 10 mm.
partment of Agriculture, Department of wire core of varying size (Table 1) and The six tornado-shaped coils and three
Health and Human Services, and Na- tested acutely in five pigs for handling of the nine tight-loop coils were
tional Institutes of Health regulations within a standard diagnostic catheter equipped with 10-mm-long polyester
and standards. and shape memory properties. Of 16 fibers positioned along the entire
Eighteen domestic pigs were used to coils with the 0.006-inch nitinol cores, length of the coils. Three other tight-
evaluate the new devices. The animals eight contained SE wire and eight con- loop coils had 10-mm long fibers only
were given no food for 18 hours before tained TSM nitinol wire. in their middle portion, and the three
each procedure. Each pig was sedated The right femoral artery was surgi- remaining tight-loop coils were
with an intramuscular injection of solu- cally isolated and a 6-F introducer equipped with 7–10 bundles of 4 –5-
tion containing ketamine hydrochloride sheath was inserted into the artery. A mm-long polyester fibers that were
(15 mg/kg), acepromazine (0.15 mg/ 6-F cobra catheter (0.035-inch ID) was positioned along the entire length of
kg), and atropine sulfate (0.04 mg/kg). advanced through the sheath and po- the coil and separated by bear
Anesthesia was induced with isoflurane sitioned in the main trunk of the supe- segments. All coils were placed in 7-F
(5%) administered via a mask. When the rior mesenteric artery (SMA) under Teflon peel-away sheaths.
pig was anesthetized, an endotracheal fluoroscopic monitoring, and angiog- The 15 0.032-inch homemade coils
tube was inserted and anesthesia was raphy was performed. The catheter were tested acutely in four pigs. Thir-
maintained with isoflurane (1.5%) and was then advanced selectively into the teen of the coils were placed in the
oxygen (0.8 L/min). All animals re- ileocolic branch of the SMA. After abdominal aortas, which had diame-
ceived heparin (100 IU/kg) immedi- achieving the desired superselective ters of 9 –12 mm, and two coils were
ately after placement of a vascular intro- catheter position, a coil was pushed placed in one inferior vena cava (IVC).
ducer sheath. into the lumen of the 6-F cobra cathe- An 8-F Lumax guiding catheter (Cook)
ter from the metal loading cartridge. was advanced into the abdominal
Part I: Handling of the Reinforced With use of a standard 0.035-inch aorta from a right carotid approach,
Coils in a Diagnostic Catheter guide wire for advancement, five to and an aortogram was obtained. After
nine coils were deployed consecu- connecting the coil to a delivery cable
Thirty-one factory-manufactured tively in the branch of each pig’s SMA with a push-button release mechanism
stainless steel coils (Cook) were each (Fig 4). These vessels ranged in diam- as described in the first in vitro study,
872 • New Embolization Coil Containing a Nitinol Core July 2001 JVIR

was placed, a randomly selected coil

Table 1
Number of Coils Equipped with Each Size of Nitinol Wire Core and Tested
of different design was placed in the
Acutely in Pigs (part I) contralateral common carotid artery in
the same manner. The coils were mon-
Size of Nitinol Wire itored for as long as 1 hour for vessel
occlusion and coil migration. Vascular
Coil Type 0.005-inch 0.006-inch 0.009-inch
occlusion was documented by hand
MWCE 35-3-5 (n ⫽ 10) 2 8 – injection of contrast material under
MWCE 35-5-5 (n ⫽ 8) 5 3 – fluoroscopic monitoring and postoc-
MWCE 35-5-8 (n ⫽ 13) 1 5 7 clusion arteriograms.
Five of the nine pigs were sacrificed
within 1 hour of the last coil placement
and the remaining four pigs were fol-
junction of the IVC. Postembolization lowed for 1 week and killed after fol-
cavography was performed to docu- low-up arteriography. A partial nec-
ment the position of the coils and their ropsy was performed on all animals
vaso-occlusive effect. after euthanasia with an intravenous
After placement, the coils were ob- overdose (1.0 mL/lb body weight) of
served for 1 hour for vessel occlusion Beuthanasia (Schering-Plough).
and coil migration, and then the ani- Results (migration and length of
mal was killed with an intravenous coil measured on the angiograms)
overdose (1.0 mL/lb body weight) of were compared with use of the two-
Beuthanasia (Schering-Plough) and a sample Student t test assuming equal
partial necropsy was performed. variances, and P values ⱕ.05 were
considered significant.
Part III—Comparison of Coils with
and without Reinforcement RESULTS
Figure 4. Selective angiogram obtained
after sequential placement of nine rein- Bilateral carotid coil placements In Vitro Study
forced factory-manufactured coils in the il- were performed to evaluate and com-
eocolic branch of the SMA. The arrange- pare coils with regard to postplace- In the first in vitro experiment, the
ment of two coils was not optimal (arrows); ment configuration and stability. expansile strength of the reinforced
the coil turns were not uniform. The SMA Nineteen Gianturco coils (MWCE 35– coil was significantly (P ⬍ .01) greater
occluded slightly below the origin of the 5-5; Cook) were placed bilaterally in than the standard coil. An average
right colic artery. the carotid arteries (4.5–5 mm) of nine force of 63.2 g (range, 49 –76 g) was
pigs (two coils were placed in one ca- needed to dislodge the standard coil
rotid artery). Nine coils were from the polyvinyl chloride tubing,
the coil was inserted in the guiding equipped with a SE wire (0.005-inch) whereas the coil containing the SE
catheter through the 7-F peel-away Tef- core, five with a TSM nitinol wire wire core required an average force of
lon sheath. The tip of the catheter was (0.005-inch) core, and five coils were 127.2 g (range, 96 –185 g).
then placed in the infrarenal or the used without a wire core. In the second in vitro experiment,
suprarenal portion of the aorta or just A 7-F Lumax guiding catheter with the average elongation of the standard
above the iliac bifurcation. Each coil a coaxially placed 5-F cobra catheter coil was significantly greater (94.9
was positioned in the recipient vessel was advanced over a guide wire into a mm, range, 94 –96 mm) than the niti-
and then completely or partially re- randomly selected common carotid ar- nol reinforced coil (12.3 mm; range,
trieved 3–10 times (mean, eight times). tery via the femoral approach. After 11–13 mm) (P ⬍ .01). The elongation of
After the last retrieval maneuver, the performing selective carotid arteriog- the standard coil was 7.7 times greater
coil was released from the delivery ca- raphy and measuring the diameter of than that of the coil with the nitinol
ble when an optimally tight coil mesh the artery, the cobra catheter was core after loading the same weight on
was achieved within the vascular removed. A randomly selected coil the free end of the suspended coil.
lumen. Contrast material was injected was attached to a delivery cable with a
by hand under fluoroscopic monitor- push-button release mechanism as de- Animal Studies
ing to determine the time needed for scribed in the first in vitro study and
complete occlusion. Postembolization loaded into the lumen of the guiding Part I: handling of the reinforced coils
aortography was then performed. catheter via a check-flow adapter in a diagnostic catheter.—All coils were
In one pig, an 8-F Lumax guiding (Cook). Although the coils were de- delivered through the catheter without
catheter was advanced into the IVC ployed with the button release deliv- difficulty. There was no discernible dif-
via an 8-F femoral sheath. Preemboli- ery cable, no coil retrieval was ference in resistance encountered dur-
zation vena cavograms were obtained. attempted. Instead, the mechanism of ing deployment of either type of coil.
With use of the delivery cable with the delivery consisted of simply unsheath- The tip of the delivery catheter re-
button release mechanism, two coils ing the coil while the delivery cable mained in place during passage of the
were deployed at the level of the iliac was held stationary. After the first coil coils.
Volume 12 Number 7 Kónya et al • 873

Figure 5. Retrievable homemade coils with TSM nitinol wire cores were tested in the IVC of the same pig. (a) A 34-cm-long, 0.032-inch
slightly tornado-shaped coil (20 –23 mm in diameter) with a 0.009-inch nitinol core was deployed at the iliocaval junction. Fifteen
minutes after deployment, partial occlusion was achieved. Note that the left ascending lumbar vein is now visualized (arrow). (b)
Another 24-cm-long tornado coil (8 –20 mm in diameter) with a 0.009-inch core was deployed with some overlap with the previous one.
Ten minutes later, complete IVC occlusion occurred. There was no coil migration.

Eight of the 31 coils showed subop- All six tornado-shaped coils with pulled back into the catheter several
timal arrangement in the vessel. Six polyester fibers (10 mm) along their times while attempting to achieve the
coils with a TSM nitinol wire core entire length resulted in complete aor- tightest possible coil configuration.
(0.005-inch, n ⫽ 1; 0.006-inch, n ⫽ 3; tic occlusion in 3–13 minutes (mean, 8 Complete occlusion of the IVC oc-
0.009-inch, n ⫽ 2) and one coil with a min). The three tight-loop coils with curred 10 minutes after deployment of
SE core (0.009-inch) remained partially fibers only in their middle portion pro- the second coil.
elongated. The elongation affected no duced complete occlusion in 12, 15, Part III: comparison of coils with and
more than one turn at one or both ends and 27 minutes, respectively. One of without reinforcement.—Tables 2 and
of the coil. Five of these six coils were the coils with separated polyester fiber 3 summarize the results with regard
placed in a vessel with a diameter of 4 (4 –5 mm) bundles caused complete to the coils’ self-anchoring capability
mm or smaller. One coil remained occlusion in 12 minutes; only partial and configuration after placement.
completely elongated. This coil con- occlusion was achieved with the other Six of the nine coils with a SE wire
tained a 0.009-inch TSM nitinol wire two. The tight-loop coils equipped core assumed their tight prepro-
core and was placed in a vessel with a with contiguous fibers (10 mm) cre- grammed shape immediately after
2.5-mm diameter. ated complete occlusion in 20 minutes deployment. These coils measured
Part II: occlusion capability and “re- in two cases; only partial occlusion 8 –14 mm (mean, 11.3 mm) in length.
trievability” under high flow condi- was achieved with one 25 minutes af- The other three coils of this type
tions.—A tight coil conglomerate was ter coil placement. showed partial elongation (each coil
achieved in all cases before release In the IVC, partial occlusion was was 17 mm long). No migration of the
from the delivery cable. No coil mi- achieved 15 minutes after deployment coils reinforced with a SE nitinol wire
gration was noted in any animal. of the first tornado-shaped coil (Fig 5). core occurred. The time required to
In the aortae, complete occlusion Therefore, another coil of the same produce occlusion with these coils was
occurred in 3–27 minutes with 12 coils; type was deployed within the previ- 5–15 minutes (mean, 7 min).
partial occlusion was achieved 20 – 42 ously deployed coil. During place- Only one coil with a TSM nitinol
minutes after placement of three coils. ment, the second coil was partially wire core spontaneously assumed its
874 • New Embolization Coil Containing a Nitinol Core July 2001 JVIR

wall. No intramural or subendothelial

Table 2
Comparative Results of Carotid Coil Placement with Regard to Stability (part III)
disruption or hemorrhage was noted.
In part III, complete occlusion of
SE Core TSM Core w/o Core the carotid artery was achieved irre-
(n ⫽ 9) (n ⫽ 5) (n ⫽ 5) spective of coil configuration. How-
ever, the configuration of the coil plug
Number of coils that migrated 0 1 3
did determine the length of the oc-
Note.—With core vs without core: P ⬍ .01. SE core vs without core: P ⬍ .01. cluded segment.

DISCUSSION
ond coil did not migrate and assumed Ideally, a vascular occluding coil
Table 3
Comparative Results of Carotid Coil
an elongated configuration (Fig 6). should possess certain characteristics. It
Placements with Regard to Of the eight coils that were placed must be flexible enough to be easily
Postplacement Configuration (part III) in four pigs and followed for 1 week, pushed through a delivery catheter, yet
two were standard (no core) and six capable of returning to its complex three-
Length of coil plugs contained a nitinol wire core (SE, n ⫽ dimensional shape after placement. In ad-
(mm) measured on 4; TSM, n ⫽ 2). None of these coils dition, the radial force of the coil must be
Stent Specification radiographs showed any change in position or con- strong enough to withstand hemody-
SE core 8 figuration during the follow-up, and namic forces that could result in deforma-
17 the embolized arteries remained com- tion and/or migration of the coil without
14 pletely occluded (Fig 7). causing damage to the vascular wall.
16 To produce a stainless steel embo-
18 lization macrocoil to fit a 0.035-inch
9 Necropsy catheter lumen, a thin (eg, 0.004 –
17
12 0.005-inch) wire is wound into a pri-
Twenty-three of the 31 smaller coils
11 mary coil with a 0.021-inch outer
(part I) assumed their preprogrammed diameter. The primary coil is then
TSM core 18 configuration, and all the large coils
32 given a secondary configuration (eg,
12
(part II) showed a compact conglom- simple or double helix, pretzel-shape,
34 erate at the site of deployment. The tornado-shape, etc.).
No core 27 lumen of the aorta and IVC was com- To prevent migration of conven-
19 pletely filled with coil turns in all but tional Gianturco coils, Chuang and
two cases. In two aortae, the turns of a Szwarc (27) used regularly and irreg-
Note.—SE core vs TSM core: P ⬍ .01. tight-loop coil placed at the iliac bifur-
SE core vs w/o core: P ⬍ .01. ularly shaped stainless-steel baffles in
cation were arranged in a single plane, a high-flow animal model. The baffles
which resulted in a gap between the were successful in preventing migra-
coil plug and the vascular wall. Only tion of the conventional coil emboli,
preprogrammed shape and formed a partial occlusion was achieved in these but the baffle/coils complex migrated
tight coil conglomerate after deploy- two cases. distally in some cases, indicating that
ment (12 mm). One coil of this type Blood clots were found adherent to insufficient friction existed between
showed partial elongation (18 mm), the polyesters fibers on the homemade the baffle and the vessel wall.
two remained substantially elongated coils used in part II. The polyester fi- The stability of the stainless-steel
(32 mm and 34 mm), and one migrated bers with the intertwined blood clots spring coils is greatly influenced by
immediately. The four coils that re- formed a compact plug in all cases. the size of the turns comprising the
mained in place after release from the The least amount of clot (mostly lim- coil’s secondary configuration. When
delivery catheter produced carotid oc- ited to the fibers) was found around coils are too small for a given vessel,
clusion in 4 –10 minutes (mean, 6 min). the coil plugs equipped with the sep- they tend to migrate. If they are too
Two of the five coils without a niti- arated, short (4 –5 mm) polyester bun- large, they tend to remain elongated,
nol wire core remained substantially dles, whereas the most abundant clot which limits their cross-sectional area
or partially elongated (27 mm and 17 (filling the entire occluded segment) and affects both their self-anchoring
mm) after deployment, whereas the was observed around the tornado- and occlusive abilities (28). By improv-
other three immediately migrated shaped coils. Vessel occlusions were ing the self-anchoring capability of the
from the common carotid artery into limited to the site of coil placement coil, the scope of its application and its
the external carotid artery. The two and, in some vessels, additional clot of safety may be increased.
coils that remained in place produced limited size (1–2 cm long) was also The reported studies show that the
complete arterial occlusion within 10 observed adjacent to the cephalad stability of the standard stainless-steel
minutes. and/or caudad ends of the coil embolization coils can be improved
In one of the acute pigs, a coil with conglomerate. Multiple indentations significantly by simply adding a niti-
a TSM nitinol wire core was deployed were seen on the luminal surface of all nol wire core. Although only one coil
at the same location as a standard coil vessels indicating firm contact be- of each type was tested, the in vitro
that had migrated cephalad. The sec- tween the coil plugs and the vessel studies showed that the expansile
Volume 12 Number 7 Kónya et al • 875

Figure 6. Comparison of coils with SE cores, TSM cores, and without core reinforcement
in the same pig. (a) Semiselective carotid angiogram after placement of a noncored coil in
the mid-portion of the left carotid artery. The coil immediately migrated and was
entrapped by the initial segment of the external carotid artery (hardly discernible, short
open arrow). Note that the flow was not compromised in the left common carotid artery.
(b) A coil of the same size with a TSM nitinol core was then deployed into the same
portion of the left carotid artery. The coil, which remained elongated and did not migrate, Figure 7. Comparison of coils with TSM
caused complete occlusion in 7 minutes. Another coil of the same size reinforced with an and SE nitinol wire core in the same
SE nitinol wire core was placed in the mid-portion of the right carotid artery. It assumed animal. A coil with a core programmed
a tighter configuration and occluded the artery in 4 minutes. with TSM was deployed into the left ca-
rotid artery. The coil assumed its coil shape
at both ends but remained elongated at its
strength and ability to regain the pre- regard to expansile force. In the part I mid-portion. Another coil with SE wire re-
programmed shape was significantly animal study, there was no apparent inforcement was placed in the right carotid
greater (P ⬍ .01) for coils containing a difference between the wire types. The artery. The coil assumed its programmed
nitinol wire core. The part III animal coils with TSM reinforcement showed a tight coil shape immediately. A plain ra-
study also revealed a significant (P ⬍ slight tendency to remain elongated in diograph obtained after bilateral coil place-
.01) improvement in postplacement the lumen of the artery. However, this ments showed complete stasis in both com-
configuration of the coils reinforced may have been largely caused by a mis- mon carotid arteries. The right and left
with a SE core versus nonreinforced match in coil and vessel diameter rather carotid arteries occluded in 6 and 8 min-
standard coils (Table 3). In addition, than wire type. In the part III animal utes, respectively. One week after coil em-
bolization, both vessels remained occluded
there was a striking difference be- study, the only reinforced coil that mi-
and there was no change in configuration
tween the self-anchoring capability of grated contained a TSM core. Although or position of the coils.
the coils with and without core rein- the coils with an SE core showed im-
forcement in the in vivo studies. Three proved intraluminal stability versus
of five coils (60%) without reinforce- nonreinforced coils (P ⬍ .01), there was strained coil and that of the recipient
ment migrated immediately upon re- no significant difference between the vessel. Because the exact size of the recip-
lease in the common carotid artery, self-anchoring capability of the two ient vessels was not known at the time the
whereas only one of 14 (7%) coils with types of reinforced coils. In addition, of coils were fabricated, the selection of coil
core reinforcement (SE or TSM) mi- the four coils with a TSM core that did sizes to be fitted with a nitinol wire core
grated (Table 2). These results indicate not migrate, only one obtained its tight was based on an estimate of the recipient
that the nitinol wire core significantly preprogrammed configuration, whereas vessel size.
(P ⬍ .01) increased the expansile force six of the nine (67%) coils with a SE core Coil elongation may be prevented
and, in turn, the self-anchoring ability formed a tight plug immediately upon with use of detachable coils. In re-
of the standard stainless-steel coils. placement (Table 3). ported studies, the ability to retrieve/
Although the reported studies were The reason why some coils containing reposition the coils facilitated optimal
not designed specifically to compare the nitinol wire cores failed to completely as- configuration and precise placement.
type of nitinol wire (SE vs TSM) used for sume their preprogrammed shape after Rotating the delivery cable and/or
the core, results suggest a difference ex- being placed was probably a suboptimal changing the position of the catheter
ists between the two wire types with match between the diameter of the uncon- tip during coil deployment facilitated
876 • New Embolization Coil Containing a Nitinol Core July 2001 JVIR

a more compact arrangement of the strained shape of the secondary coil, 3. Hawkins J, Quisling RG, Mickle JP,
coil conglomerate. Although the cable and (iv) the length of the coil. Hawkins IF. Retrievable Gianturco-coil
used in these studies was considered In an in vitro study, forces required to introducer. Radiology 1986; 158:262–264.
oversized, no problems were encoun- advance and retract microcoils (0.015-in 4. Graves VB, Partington CR, Rufenacht DA,
tered with retrieval and/or release of outer diameter) made of different materi- Rappe AH, Strother C. Treatment of ca-
als in varying shapes and lengths in mi- rotid artery aneurysms with platinum
the coils. Construction of a smaller de- coils: an experimental study in dogs. Am J
crocatheters (0.018-in internal diameter)
livery cable would enable use of a Neuroradiol 1990; 11:249–252.
were measured (5). The relative stiffness
smaller coil and delivery catheter. of the primary coil was also measured by 5. Marks MP, Tsai C, Chee H. In vitro eval-
The unconstrained shape of the coil the force that deflected the primary coil to uation of coils for endovascular therapy.
as well as the polyester fiber content a certain angle. The data showed that the Am J Neuroradiol 1996; 17:29–34.
and distribution seemed to be crucial stiffness of the primary coil as well as the 6. Dutton JAE, Jackson JE, Hughes JMB,
in achieving quick vessel occlusion. secondary shape of the coil had a signifi- et al. Pulmonary arteriovenous mal-
The tornado-shaped coils with a diam- cant effect on the friction within the cath- formations: results of treatment with
eter larger than that of the recipient eter but the overall length of the coil was coil embolization in 53 patients. AJR
relatively unimportant in this regard. Am J Roentgenol 1995; 165:1119 –1125.
vessel functioned more reliably than
The reported studies have shown that 7. Uzun O, Hancock S, Parsons JM, Dick-
the coils with a tight-loop shape. The inson DF, Gibbs JL. Transcatheter oc-
tornado coils with turns with diame- addition of a nitinol wire core increases
clusion of the arterial duct with Cook
ters of 8 –11 mm and 16 –24 mm more the stability of the standard embolization
detachable coils: early experience.
consistently formed a compact mesh coil by significantly increasing its expan- Heart 1996; 76:269 –273.
sile force and ability to form a compact
and filled the vascular lumen more 8. Tometzki AJ, Arnold R, Peart I, et al.
conglomerate. Increased coil stability may Transcatheter occlusion of the patent
completely than the coils with diame- virtually eliminate the possibility of coil
ters of 6.5–10 mm. The two coils that ductus arteriosus with Cook detach-
migration. This, in turn, could expand the able coils. Heart 1996; 76:531–535.
failed to fill the vascular lumen com- scope of coil embolization in veins 9. Akagi T, Hasahino K, Sugimura T, Ishii M,
pletely, resulting in partial occlusion (10,29,30), high-flow vascular lesions such Eyo G, Kato H. Coil occlusion of patent
were both 6.5–10 mm in diameter. as arteriovenous malformations and fistu- ductus arteriosus with detachable coils.
Although the friction within the cathe- las in the extremities and the lungs (11– Am Heart J 1997; 134:538–543.
ter caused by core reinforcement of the 13), and nontapering vascular structures. 10. Coley SC, Jackson JE. Endovascular
coil was not measured, neither the SE nor In addition, the nitinol wire core may en- occlusion with a new mechanical de-
TSM nitinol wire core noticeably changed able production of coils with different tachable coil. AJR Am J Roentgenol
the coil’s ability to be pushed. This sug- configurations that can be positioned in 1998; 171:1075–1079.
gests that any decrease in coil “pushabil- an optimal three-dimensional arrange- 11. Remy-Jardin M, Wattinne L, Remy J.
ity” caused by the enhanced radial force ment, resulting in faster occlusion. Transcatheter occlusion of pulmonary ar-
of the coil was offset by the increased stiff- In addition, the nitinol wire core does terial circulation and collateral supply: fail-
not alter the coil’s ability to be pushed, ures, incidents, and complications. Radiol-
ness of the coil when elongated in the which allows the coil to be delivered ogy 1991; 180:699–705.
delivery catheter. In addition, it was noted through the same catheters used for de- 12. White R, Lynch-Nyhan A, Terry P, et
that the ability of the coil to be pushed ployment of standard stainless-steel coils. al. Pulmonary arteriovenous malfor-
could be increased by leaving a segment The coil can also be deployed superselec- mations: techniques and long-term
(5–10 mm) of the proximal end of the coil tively without displacement of the deliv- outcome of embolotherapy. Radiology
straight and/or without a nitinol core. ery catheter tip. Ability to reposition and 1988; 169:663– 669.
An ideal match between the diam- retrieve is a desirable feature of a coil of 13. Keller FS, Rösch J, Barker AF, Nath PH.
eter of the coil and the luminal diam- this type because it adds to the self-an- Pulmonary arteriovenous fistulas oc-
eter of the delivery catheter is required choring and occlusive capabilities of the cluded by percutaneous introduction
device. Although the mechanical proper- of coil springs. Radiology 1984; 152:
for good coil “pushability.” In general, 373–375.
the outer diameter of the stainless- ties and behavior of the coil reinforced
with a nitinol wire core was found to be 14. Rothman A, Lucas vW, Sklansky MS,
steel spring macrocoil is considerably Cocalis MW, Kashani IA. Percutane-
less than the inner diameter of the favorable in limited acute and chronic an-
ous coil occlusion of patent ductus
imal models including high-flow condi-
catheter. For example, coils manufac- arteriosus. J Pediatr 1997; 130:447– 454.
tions, further studies are warranted to fur-
tured by Cook that are designed to be ther evaluate the coil’s use and safety.
15. Owada CY, Teitel DF, Moore P. Eval-
used in a 0.035-inch lumen are con- uation of Gianturco coils for closure of
structed from a 0.021-inch primary large (⬎ or ⫽ 3.5 mm) patent ductus
coil, and a short segment of the prox- Acknowledgments: The authors wish to arteriosus. J Am Coll Cardiol 1997; 30:
acknowledge Raquel Collins and Irene 1856 –1862.
imal end of the coil is widened so it Szwarc for their expert technical assistance. 16. Johnston TA, Stern HJ, O’Laughlin MP.
fills more of the catheter lumen. Apart Transcatheter occlusion of the patent
from the physical properties of the ductus arteriosus: use of the retrievable
References
stainless steel material, the thrombo- 1. Gianturco C, Anderson JH, Wallace S. coil device. Cathet Cardiovasc Inter-
genic fibers, and the catheter material, Mechanical device for arterial occlusion. vent 1999; 46:434 – 437.
there are four factors that appear to AJR Am J Roentgenol 1975; 124:428–435. 17. Radtke WA. Current therapy of the
have the greatest impact on the coil’s 2. Chuang VP, Wallace S, Gianturco C. patent ductus arteriosus. Curr Opin
“pushability:” (i) The relative stiffness A new improved coil for tapered-tip Cardiol 1998; 13:59 – 65.
of the primary coil, (ii) the spring force catheter for arterial occlusion. Radiol- 18. Butto F, Hunter DW, Castaneda-
of the secondary coil, (iii) the uncon- ogy 1980; 135:507–509. Zuniga W, Amplatz K. Coil-in-coil
Volume 12 Number 7 Kónya et al • 877

technique for vascular embolization. sisted technique for transcatheter coil shunts, complications, and technique.
Radiology 1986; 161:554 –555. occlusion of moderate to large patent Am Heart J 1999; 138:1176 –1183.
19. Castenada-Zuniga WR, Tadavarthy SM, ductus arteriosus: immediate and in- 27. Chuang VP, Szwarc I. The coil baffle
Gonzalez R, Rysavy J, Amplatz K. Single termediate results. J Am Coll Cardiol in the experimental occlusion of large
barbed stainless steel coils for venous 1999; 33:1710 –1718. vascular structures. Radiology 1982;
occlusion. A simple but useful modifi- 24. Sommer RJ, Gutierrez A, Lai WW, Parness 143:25–28.
cation. Invest Radiol 1982; 17:186–188. IA. Use of preformed nitinol snare to im- 28. Nancarrow PA, Fellows KE, Lock JE.
20. Lund G, Cragg AH, Rysavy JA, Cas- prove transcatheter coil delivery in occlu- Stability of coil emboli: an in vitro
taneda F, Castenada-Zuniga WR, Am- sion of patent ductus arteriosus. Am J Car- study. Cardiovasc Intervent Radiol
platz K. Detachable stainless-steel
diol 1994; 74:836–839. 1987; 10:226 –229.
spider. Radiology 1983; 148:567–568.
25. Weber HS, Cyran SE. Transvenous 29. Chomyn JJ, Craven WM, Groves BM,
21. Castenada-Zuniga WR, Tadavarthy
SM, Galliani CA, Laerum F, Schwarten “snare-assisted” coil occlusion of Durham JD. Percutaneous removal of a
DE, Amplatz K. Experimental venous patent ductus arteriosus. Am J Cardiol Gianturco coil from the pulmonary artery
occlusion with stainless-steel spiders. 1998; 82:248 –251. with use of flexible intravascular forceps. J
Radiology 1981; 141:238 –242. 26. Janorkar S, Goh T, Wilkinson J. Vasc Interv Radiol 1991; 2:105–106.
22. Kónya A, Wright KC, Wallace S. An- Transcatheter closure of patent ductus 30. Smith DC, Kohne RE, Taylor FC.
choring coil embolization in a high- arteriosus with the use of Rashkind oc- Steel coil embolization supplementing
flow arterial model: a pilot study. J cluders and/or Gianturco coils: long filter placement in a patient with a du-
Vasc Interv Radiol 1998; 9:249 –254. term follow-up in 123 patients and spe- plicated inferior vena cava. J Vasc In-
23. Ing FF, Sommer RJ. The snare-as- cial reference to comparison, residual terv Radiol 1992; 3:577–580.