A Comparative
In Vi•ro
Analysis of thL iobin-Uddin,
-iJ&rayJGireenmicd and ;litinol Blood Clot Filters
by
Martin Raymond Prince
'I
Sumitted in
of the
Partial Fulfillment
tc.uirent.s for the
Degree of
C[laster of Science
at the
Massachusetts T.nstitute of Tch••ibloy
nOctober
c
l
iartin Raymond Prince
The author hereby gra nts to M.I.T. permission to rerroduce and
to distribute copies of this docu1ent in whole or in part.
Signature of Author
Departtment
f Mechanical Engineering
August 28, 1981
Certified by
.
Certc-
/
,ifiedby
Professor
Morris
Simon
&kesis
Supervisor
_--L__
U
Professor Robert M1ann
Thesis Suoervisor
cce-ted
by..
•-fi
rchve.S
Herbert B.
Chairman,
MASSACHUT..
,......
LOF3-,C;NO.9Y2
JUL 301982
LIB RTR1S
Mechanicali
Richardson
Engineering
A Comnparative In Vitro Analysis of the
:imray Greenfield, Mc•in-Uddin and Nitinol
Blood Clot Filters
by
Martin Raymond Prince
Submitted to the Department of Mechanical Engineering
on May 20, 1982 in partial fulfillment of the
requirements for the Degree of Master of Science in
Mechanical Engineering
Abstract
The Mobin-Uddin, Kimray Greenfield and Nitinol Inferior
Vena Cava (IVC) interruption devices were analyzed in an in
vitro simulation of the human inferior vena cava to evaluate the
correctness of filter positioning, the clot capturing
effectiveness, interference with blood flow and security of
filter anchoring. This simulation reproduced the temperature,
pressure, flow rate and vena cava sizes normally encountered in
the human.
It also included a movie camera and specially
oriented mirrors to document three simultahe-ous projections of
the filter delivery and the arrival of an embolus at the filter
for later slow notion analysis. The Kimray Greenfield filter
tended to assume a less effective tilted orientation; in large
cavae it allowed 7mm diameter emboli to pass through. The
vMobin-Uddin Umbrella captured emboli well but presented a
significant obstruction to flow and demonstrated a tendency to
dislodge in large vena cava and migrate with the fluid stream.
The most effective device was the Nitinol filter which was
well-oriented, captured clots effectively and shoýwed negligible
flow interference. The in vitro performance of the Kimray
Greenfield and Mobin-Uddin filters reflects observations
reported in the clinical literature. The Mitinol filter is
still experimental and has not yet been used in humans but its
superior in vitro performance is encouraging.
Thesis Supervisors:
Morris Simon
Professor of Radiology
Harvard Medical School
Robert Mann
Professor of Mechanical Engineering
Massachusetts Institute of Technology
r
Acknowl edgments
I wish to thank the members of the Nitinol filter research
team at Beth Israel Hospital for providing guidance and support
in this study. Professor Morris Simon, M.D., project director,
spent endless hours teaching me research and design skills. He
along with his colleague, Aubrey Palestrant, M.D., gave me
continual encouragement and helpful advice especially in the
clinical and radiologic aspects of pulmonary embolism.
Professor Robert Mann, Ph.D., my research advisor at M.I.T.,
afforded me invaluable direction in the conduct and reporting of
engineering research. Israel Soibelman was a reliable and
competent assistant. Finally, I wish to thank the
M.I.T./Harvard, Medical Engineering and Medical Physics program
in the Health Sciences and Technology Division for providing
fellowship support.
-4-
Introduction
ach year an estimated 630,000 Americans suffer
signifitant health problems when large clots that form in
the
veins o' the lower limbs or pelvis break loose and migrate to
the lunds (Figure 1).
For 200,000 Americans these clots, called
pulmona y emboli, will be fatal (1).
These emboli plug
pulmona y arteries blocking the flow of blood into the lungs.
This re ult is
so dangerous that clinicians are willing to take
substantial risks to vigorously treat patients for whom there isa high
robability of emboli migrating to the lungs
(Table 1).
The standard treatment for pulmonary embolism,
anticoa ulation therapy, effectively reduces the mortality but
carries a significant risk of bleeding complications (2).
Occasio ally these clots can be removed with surgery (3) or with
thrombo ytic (clot dissolving)
patient s life.
drugs (4),
in time to save a
An alternative form of treatment in patients
preclud d from receiving anticoagulation is to modify the shape
of the
en
ot
largest vein leading from the legs to the
lungs ( he inferior vena cava (IVC))
so that clots, which
migrate from the lower part of the body, are arrested in the IVC
and can rot reach the lungs.
This is presently accomplished by
three t chniques:
a)
direct abdominal surgery on the IVC to subdivide its lumen by
suture rraterial or external clips
b)
(5) (Figure 2A);
nec 2 or groin surgery to expose a major vein for insertion
-5-
of a transvenous filter (Figure 2B) via an incision in the vein
wall (6-6)
(Figure 3); or
c) percutaneous catheterization of the femoral or other
peripheral vein (without surgery) and insertion of a memory wire
(Nitinol) filter into the IVC via this catheter (9) (Figures 2B
& 4)
Each of the present clinical procedures involves
significant problems.
IVC interruption by the direct abdominal
surgical approach is effective in preventing pulmonary embolism
but there is a 14% mortality rate associated with using general
anesthesia on these acutely ill patients (10).
Surgery for the
indirect insertion of transvenous devices (see Figure 3) is much
less dangerous primarily because it involves only local
anesthesia.
Despite this advantage, complications such as
filter migration, vein wall perforation, filter misplacements,
infection, filter breakage, and the like have all been described
in the clinical literature (9,11-19).
Furthermore, clinicians
still await a well controlled analysis of the ability of these
filters
to capture emboli in the human.
The transcatheter, memory wire, Nitinol filter is still
experimental and has not yet been approved for clinical use.
However, the catheterization procedure is clearly much simpler
than neck,
groin,
transcatheter
it
or abdominal surgery.
If, in addition,
device proves to be effective in
the
capturing emboli,
would constitute a considerable advance in the treatment of
pulmonary embolism.
To provide more information on this issue
this research compares the clot capturing capability of an
experimental
transcatheter
devic1,
the R:itinol
filter,
with two
transcaval devices currently used clinically, the Greenfield
Filter and the Mobin-Uddin Umbrella.
In Vivo Studies Versus In Vitro Simulation
The first methodological consideration was the relative
value, at this stage of research/development, of in vivo versus
in vitro analysis.
In vivo studies using experimental animals
to evaluate these devices have not been carried out easily
(20-23).
Appropriate research animals are expensive and animal
experiments require substantial preparation time; thus, only a
limited amount of data can be collected in this way.
Another
problem is that the x-ray studies used to evaluate implanted
devices provide only a limited view of what's happening in the
IVC.
Further, each data point must be derived from a different
animal with different IVC anatomy, cava caliber, flow rate, and
other factors.
For these reasons, it is not possible to
adequately control experimental conditions nor to provide
appropriate data for statistical analysis.
The most important
limitation on experiments with dogs, the classic cardiovascular
research animal,
is
that they do not,
in
the case of the IVC,
provide a completely realistic simulation of human anatomy and
physiology.
For example,
when these devices are squeezed into a
13mm diameter dog IVC they may have very different shapes than
-7-
when they expand to fill a 28mm diameter human IVC (Figure 5 &
6)
Scientific evaluation of an IVC interruption device in human
subjects is not practical.
The number, size; shape and
clinical presentation of emboli found in humans are typically
unpredictable. Furthermore the clinical background of each
patient is unique. For obvious ethical reasons, the controlled
injection of emboli into human subjects is not acceptable. The
angiographic procedures which would be necessary to evaluate
these devices might generate other medical problems such as
allergic reactions to the X-ray contrast materials, infection,
internal bleeding, thrombus formation, and even loss of limbs.
These risks would be compounded by the poor general health of
most patients for whom IVC interruption would be appropriate.
This study overcame many of the problems just described by
evaluating the performance of IVC interruption devices in an in
vitro simulation of the human IVC described previously (24).
The simulated IVC was made completely of transparent materials
to allow clear visualization of the filters during
The caval tissue was simulated with cellulose
experimentation.
dialyzer tubing in three sizes within the range of observed
human IVC sizes. Supporting apparatus controlled temperature,
pressure and flow rate of the experimental
these
were
experiments).
set
A system of mirrors
up to document
performance
for
and a 16mam
three' simultaneous
slow motion analysis.
fluid (saline in
Filters
rmovie
camera
views of filter
and blood clot
-8-
were delivered to the simulated IVC via an access port.
7 and u are diagramatic illustrations of the simulation.
Figures
This
apparatus was used to conduct a comprehensive in vitro analysis
of the three previously mentioned filtering devices:
Greenfield,
r~.
[Mobin-Uddin and Nitinol Filters.
the Kimray
-9-
-Exoerimental
Overview:
Desiicn
This experiment examined four parameters of
filter performance relevant to clinical application:
orientation in the IVC,
i) filter
2) clot capturing ability, 3) filter
interference with blood flow and 4) security of filter
anchoring.
Filter orientation was recorded immediately after
delivery; clot capturing ability was determined by introducing
canine blood clots of standardized sizes into the test system;
filter interference with flow was measured as the pressure
gradient across the filter; filter anchoring strength was
measured by positioning the simulated vena cava vertically and
attaching weights to the filter within.
Before these parameters could be evaluated,
appropriate cava
sizes and experimental flow rates were established. The former
were determined by examining human venocavograms. Experimental
flow rate was determined by adjusting the flow rate until the
experimental emboli traveled at the same rate as emboli studied
in vivo.
Details of these two procedures and the experimental
methods will be described in the following sections.
Determination of Cava Size:
Since IVC interruption devices
tend to have a more compact shape in smaller cavae,
would be expected to influence the test parameters.
cava size
To
determine the mean and standard deviation of human caval
cross-sectional areas,
at the level of filter engagement
-10-
follow-up x-ray films of patients at Beth Israel Hospital and
MIiassachusetts General Hospital with KG filters
a)
cava diameter at site of hooks
magnification
measured filter
and
were examined for
b) filter length.
The
factor for the filter was calculated from the
length with equation (1):
measured filter length
magnification factor = ----------------------
(1)
actual filter length
Vena Cava diameter was calculated with equation (2)
actual
measured vena- cava diameter
= -----------------------------
vena cava
diameter
(2)
magnification factor
Thus since the actual length of filter limbs for the Greenfield
filter is known to be 4.6cm these equations can be combined to
form equation (3):
actual
vena cava
measured vena cava diameter
= --------------------------
diameter
X 4.6cm
(3)
measured filter length
Caval cross-sections were assumed to be nearly circular due to
the uniform pressure of filter hooks on the inner wall of the
vena cava.
Additional support for this assumption came from
computerized tomographic scans of human patients with vena cava
filters taken at the Massachusetts General Hospital which all
showed the vena cava to be circular in cross-section at the
level of the filter
hooks
(25).
Results are shown in Figure 9.
Selection of experimental vena cava diameters was limited to the
sizes of cellulose dialyzer tubing available.
acquired,
15nmm, 20rmm and 283mm are reasonably
The three sizes
representative of
-11the range of observed human vena cava sizes.
Determination of Flow P
Rate:
It is difficult to estimate
the average blood flow rate through the infrarenal vena cava in
the population of patients threatened by pulmonary emboli.
normal, healthy adult total cardiac output at rest is
In a
5 liters
per minute (although it may increase to 15 liters per minute
with exercise).
It is estimated that during the resting state
twenty percent of the cardiac output flows to the lower body
(below the kidney level)
(26).
Patients diagnosed as being at
high risk for pulmonary emboli are usually hospitalized and at
rest.
Therefore, their infrarenal IVC flow rate is one liter
per minute.
Unfortunately this flow rate in a 20 mm diameter cava was
insufficient to propel the experimental emboli to the filter at
a perceptible rate.
More realistic motion of emboli was
obtained by increasing the flow to two liters/minute or 100
cm/sec.
This clot motion was compared to cine films of
opacified blood clots delivered to dogs and seemed comparable.
The fluid velocity was standardized and appropriate volume flow
rates were calculated for the other sizes of cavae as shown in
Table 2.
: :.:::;:.::
-12-
Table 2
Tubing
Flat Width
(mm)
24
32
44
Cava
Diameter
(mm)
15
20
28
Area
(mm)
133
314
616
Volume
Flow Rate
(i/min)
1.1
2.0
3.6
Velocity
(mm/sec)
100
100
100
Evaluation of Filter Orientation and Anchring Securitv:•
For optimal clinical functioning the filters are designed to be
positioned along the central axis of the vena cava without tilt.
The filters must also be securely anchored to prevent migration
because a) movement toward the heart and lungs could be life
threatening and b) backward migration, away from the heart and
lungs, could perforate the vena cava,
and possibly damage
adjacent critical structures such as the aorta.
To evaluate filter orientation, MU and KG filters were
purchased from Beth Israel Hospital's clinical service and
loaded into appropriate delivery systems (Figure 3).
The
delivery capsule was then inserted into the simulated IVC.
The
filters were released by withdrawing the capsule from around the
filter as recommended in the clinical literature (27) .
Nitinol
filters were also delivered to the simulated IVC but via a
French 8 angiographic catheter
delivery the orientation
evaluate
filter
(see Figure 4).
(tilted or central)
After each
was noted.
To
anchoring security the simulated IVC was
positioned vertically and a cup was hung from the leading tip of
-13-
the filter
and filled slowly with water until the filter
sliding in
the simulated vena cava tubing.
began
The weight of the
cup and water was measured and recorded as the filter holding
force
(Table 3).
This procedure, including filter delivery and
holding force measurement,
was repeated ten times for each type
of filter in each of the three sizes of simulated venae cavae.
Evaluation of Clot Capture and Flow -Obstruction:
velocity was set to 100mm/sec as shown in Table 2.
The flow
Temperature
was adjusted to body temperature and the manometer measuring
fluid pressure just upstream from the filter was set to zero.
The filter
was delivered into the simulated vena cava. A
standardized blood clot, made previously from canine blood drawn
into glass tubing and left at room temperature for more than one
hour to ensure completion of the clotting process,
was then
introduced into the fluid flow 20 cm upstream from the filter.
Arrival of this clot at the filter was carefully observed and
recorded as follows:
+
+-
captured completely
clot is captured but more than 1 cm protrudes
through the filter
clot passed completely through the filter
The +- category was adopted because protruding fragments of
captured clots sometimes broke loose becoming,
at least in part,
an uncaptured embolus.
Occasionally the arrival of an embolus dislodged the filter
-14-
causing it to migrate downstream or obstruct flow through the
filter
so that no additional clots could reach the filter.
These events were recorded as:
M migration of filter
I
insufficient flow to bring clot to filter
Finally, pressure upstream from the filter was recorded.
Since the manometer was initially set to zero, this pressure was
the pressure gradient across the filter and thus reflected the
severity of flow obstruction due to the filter and contained
clot.
Two types of experiments were conducted.
One involved
delivery of a single 10 cm long clot to the filter.
The other
involved delivery of a set of five, two cm long clots in
succession.
Capture status and pressure changes were recorded
for each individual clot.
Determination of Test Clot Diameter:
The normal, healthy
human body has a tremendous excess of pulmonary capacity
utilising only a fraction of the vascular bed at rest (28).
Thus, for a clot to be lethal, it must block most of the
pulmonary arterial blood supply.
For this reason, many
clinicians feel that a clot is generally greater than 7mm in
diameter when fatal (29).
embolizing in
However several smaller clots,
rapid succession,
of a single large clot.
Also
,
can have the equivalent effect
a patient who already has
seriously compromised pulmonary function may have a reduced
·T1~r3s*a~
---
i
-15-
pulmonary reserve capacity. In such a case a single smaller
embolus could be fatal.
Indeed,
the p&tients most commonly
treated to prevent pulmonary embolism have prior heart or lung
a
disease and may have sustained a previous pulmonary embolus that
has further compromised their pulmonary function.
For these
reasons, capturing clots smaller than 7mm in diameter is
desirable. This does not mean that very small clots, those less
than 4mm in diameter,
have to be captured.
These are readily
cleared from the pulmonary arteries by normal biological
mechanisms in the lungs and do little or no harm (4).
To
summarize, then, an effective IVC filter should capture all
clots 7mmP
or larger, most 4-7mm clots, but may let smaller clots
through to avoid becoming unnecessarily occluded.
In line with
this reasoning this study is based upon testing the various
filters with 4mm and 7rmm diameter clots.
-16-
Results:
Orientlti
n and Anchoring Security
The three filters are designed to be oriented on the
central axis of the vena cava and
to lock into the caval wall.
Each of the filters was analyzed with regard to these two
parameters and is described by first defining its optimal
filtering orientation, second, examining the probability of
delivery into that orientation and third, by evaluating, the
filter anchoring security..
Greenfield Filter:
The optimal filtering orientation for
this device is with the apex located on the central axis of the
vena cava and the six limbs evenly spaced around the caval
circumference
(Figure 17).
Any deviation from this optimum
produces an asymmetry which increases the size of some spaces
and decreases others. This allow:,s more and larger clots to pass
through the filter. Delivery of the KG filter into the optimal
orientation occurred in only 8 of 30 deliveries while 22 were
tilted. (Table 3)
Three causes of filter tilt were observed:
I)
Capsule Weight:
containing
the
tended to lie
filter
During delivery the capsule
(see Figure
3)
was heavy
on the posterior surface
of
lumen of the supine patient (Figure 10A).
and
the caval
As the
-17-
filter was pushed out of the capsule, the apex emerged
last and thus tended to hug the bottom of the cava in a
tilted orientation (Figure 10).
2)
Filter spring energy:
As the filter emerged from
the capsule it reached a critical point, at which the
elastic energy stored in the filter was suddenly
released and the filter sprung out into the vena cava
(Figure 10C).
Even the most careful operator would
have difficulty controlling this.
3)
Filter Weight:
Even if the Greenfield filter was
delivered into the optimal central orientation, it was
not likely to remain so for more than a few moments
because the tilted orientation was thermodynamically
rore stable (Figure 10D).
This was due to the weight
of the filter itself, the lack of a centering support
and the larger circle formed by the legs when tilted.
It was also observed that the arrival of a clot would
knock the filter into a tilted orientation. Thus, the
filter's lowest energy state was in a tilted position.
During clot capture trials of the Greenfield filter in
a central orientation a support was necessary to hold
the filter in that orientation.
Once delivered, the Greenfield had substantial anchoring
-18security in the 15mm and 20=mm diameter cavae but this force
dropped off dramatically in the 20m'm cava.
It w"-as also observed
that all of the hooks engaged the cava in the smaller sizes as
compared with only a few hooks in the largest vena cava. Failure
to engage the wall seemed to be caused by an inappropriate hook
angle as shown in Figure 11.
An appropriate hook angle is
calculated in Figure 12 and was found to have about twice the
holding force as the existing hook in 28mm diameter cavae.
Ilobin-Uddin Umbrella:
Since this device incorporates a
silastic membrane between its tines, its performance was
unaffected by slight tilting.
Excessive tilting was undesirable
because the filter would fail to span the caval lumen and would
allow clots to slip through gaps between the filter and the vena
caval wall.
The sudden filter release from the capsule, the
capsule resting against the posterior caval wall, and
thermodynamic instability factors tended to result in a tilted
orientation after delivery similar to that observed in the case
of the KG filter. However, excessive tilting was prevented by
the centering effects of the attached guidewire.
After delivery of the filter, two significant problems were
noted.
First, the filter was not large enough to engage the
walls in the 28mrm vena cava.
It was therefore difficult to
unscrew, its delivery guidewire because the filter was not
secured;
the filter
rotated with the guidewire as it
and did not easily separate.
Once the filter
was turned
did release it
-19usually fell on its side and rolled downstream.
For clot
capture studies the hooks had to be engaged by hand.
Second,
in
the smaller cava, the silastic plastic sheet between the spokes
formed deep folds as the spokes came closer together to
accommodate the smaller cava size.
These folds constituted
channels larger than the holes in the membrane and permitted
experimental clots to pass.
The filtering qualities of the
device were thus determined by the channel size, rather than by
the size of holes in the mesh.
Holding forces for the MU filter were even larger than for
the KG filter in the 15mmr
and 20mm diameter cavae but there was
practically no anchoring security for the filter in the 28mm
cava. Thus, delivering the Liobin-Uddin Umbrella filter into a
28mm or larger vena cava would carry a high risk of embolizing
the filter itself.
Nitinol Filter:
As with the KG filter, this filter is
designed to function optimally with the apex centered and the
limbs evenly spaced.
However, the Nitinol filter has an
additional orienting element at its apex, the mesh of
overlapping loops, which centers the device automatically even
if the limbs were tilted slightly.
The Nitinol filter was
easily delivered into this optimal orientation through the lumen
of a French 8 angiographic catheter as shown in Figure 4.
the 30 deliveries, all resulted in appropriate orientations
although some were tilted slightly particularly in the 20mm
Of
-20-
diameter cava (see Table 3).
Nitinol filter holding forces in the 15mm and 20mm diameter
cavae were comparable to those of the Mobin-Uddin and Greenfield
filters but security was substantially higher in the 28mm cava.
More significantly, yield force resulted in the filter slipping
O*
along the cava as opposed to the caval tearing observed when
excessive force was applied to the other devices.
The
experimental Nitinol filter hooks were not needle sharp and
pressed into, but not through,
the vena cava wall.
In the
smaller vena cava sizes the mesh provided additional anchoring
security.
ResultSlCot Capturing _AIbii ty
Greenfield Filter:
effectively,
allowing
This filter captured clots the least
7rmm diameter clots through in all three
cava sizes (Figures 15-21).
In its most common orientation,
tilted, it let ~ 50% of the clots at least part way through.
Clots passing part way through were considered potentially as
dangerous as clots passing all the way through because the
downstream,
unprotected,
embolize to the lungs.
protruding portion could break off and
It
could also serve as a nidus for
further clot formation on the downstream side of the filter, a
possible source of lethal emboli.
-21-
This filter, when empty, captured small clots well since the
limbs guided them to the acex where the mesh was finer.
However,
as soon as the apex filled with clot the flow was
diverted to the wider peripheral spaces and all additional clots
passed through. The observation that the later clots tended to
pass through the filter is evident in Figures 17-21.
Sometimes, after its initial capture, a clot would be jarred
free by the impact of another clot arriving at the filter.
Similarly with a single long clot the leading end sometimes
filled the apex and forced the trailing end to somersault
through a larger peripheral space.
The momentum could pull the
rest of the clot through along with it.
This analysis suggests that if large peripheral holes exist,
clots may first plug all of the small central holes diverting
the flow and further clots through the remaining large openings
at the periphery.
Hgobin-Uddin Umbrella:
The IMobin-Uddin membrane was fairly
impervious to all test clots since its holes were only 3mm in
diameter.
Some 4mm diameter clots, however, passed through the
channels between the tines created by the infolding of the
silastic in accommodating smaller cavae.
Other clots forced the'
Mobin-Uddin to tilt excessively creating a space between the
membrane and the caval wall.
Occasionally large pressure
gradients across the filter for~ced 4mm clots through the 3mm
diameter holes.
Despite these problems, this device xias
-22effective in capturing clots in the two smaller sized venae
cavace
(F'igures 22 & 23)
In the 28mm vena cava, however,
the caval wall securely (Figure 24).
the filter did not engage
Its orientation frequently
changed with clot impact.
Nitinol Filter:
This filter has both a set of anchoring
limbs and an umbrella-like mesh, both of which function as clot
capturing elements. This combination would be expected to
capture clots as effectively than the other devices and, in
fact, the Nitinol filter captured nearly all clots in all three
vena cava sizes (Figures 25-27).
Paradoxically it captured
clots better in the 20mm cava than in the 15mm cava.
This
resulted because the overlapping loops formed mieore uniform hole
sizes at the 20mm diameter whereas some larger holes appeared in
the slightly less uniform mesh in the 15mni diameter vena cava.
Results:
Interference with flowý
Interference with Flow
reflected the tendency of the filter
to occlude the IVC and was measured as the pressure gradient
across the filter device.
The KG and KNitinol filters both
showed no measurable pressure gradidnt without clots and only
minimal gradients occurred after delivery of experimental clots
(Figures 17-21 & 25-27).
The MU filter, however,
even without
clots, created a measurable pressure gradient and after delivery
of clots the pressure gradient increased markedly and sometimes
exceeded the pressure limitations of the in vitro system
(Figures 22-24).
The vena cava upstream of the filter usually
distended due to this pressure gradient while the downstream
vena cava collapsed.
If the filter was poorly secured, this
often dislodged the filter and caused it to embolise.
Many
times flow through the Mobin-Uddin became so reduced that it was
insufficient to carry additional clots to the filter.
At other
times pressure gradients across the Mobin-Uddin became so large
that the vena cava tubing simply ripped apart.
-24-
Limitations of In
Vitiro
Simul•in
Obviously no model can perfectly simi.ulate the human body.
Failure to simulate tissue response,
immune mechanism,
the
reticulo-endothelial system, adjacent structures, and other
uniquely biological phenomena restrict use of this system to
evaluating mechanical effects only.
Analysis of these
mechanical effects is limited by some of the simplifications
the system design.
in
They can be enumerated as follows:
First, saline has different properties from blood, i.e.
their densities and viscosities are different.
These
differences may hve been compensated for by adjusting the flow
velocity but one cannot be absolutely sure.
Flow in the vena cava is known to be pulsatile (29) but the
simulation is strictly limited to uniform flow.
Pulsatile flow
might be expected to shake clots loose from the filter
increasing the probability they would pass through any of the
larger openings in the filtering devices.
The experimental emboli made by allowing dog blood to sit
stagnant in
a glass tube for one hour are probably not exact
replicae of human emboli which have bizarre shapes and are
formed over days or weeks developing substantial organization.
The simulating device was fixed in
a horizontal position Dut
patients might assume a variety of postures that could influence
filter
performance.
-25Caval tissue is
much tougher and more elastic than the
cellulose dialyzer tubing used in
tle simulation.
This lack of
perfect simulation probably had a marked effect on filter
holding force determination (Table 3).
Failure of the filter to
stay secure in the vena cava was usually due to tearing of the
cellulose dialyzer tubing.
Human caval tissue would tear only
when subjected to much higher forces.
Finally, humans also display much greater anatomical
variation than can be modeled with three sizes of cylindrical
tubing.
In fact,
body scans show the IVC before filter
insertion to be oval in cross-section as opposed to a circular
shape asuned by this model.
A further complication is that
Shuman subjects have curving vessels 'with numerous branches and
surrounding structures.
Many of these deviations from the real life situation were
purposely designed into the simulation to optimize visualization
and to generate a well controlled experiment.
could be improved.
Some aspects
Lethal human blood clots could be acquired
from autopsies and cut into standardized sizes.
pump could generate pulsatile flow.
A pulsatille
Gelatin could be added to
the saline to create a viscosity and density closer to that of
blood without sacrificing visualization.
importantly,
the circulation modelling parameters could be
studied during human surgical procedures
~
Perhaps most
or on ex:perimental
animals to determine optimal values fdr in vitro system
capacitance,
resistance,
inertia,
and pu•
up
characteristics.
-26-
Na-Wi 21aimn
Despite the problems enumerated above, this simulation goes
a long way toward evaluating the relative utility of three types
The validity of this comparison is
of intracaval devices.
verified by noting clinical ovservations that correlate with
observations made in the in vitro simulation.
The MU was observed to occlude readily and generate large
pressure gradients in vitro while the KG showed minimal pressure
gradients and never occluded.
Clinically the MU has a 73%
incidence of occlusion and venous stasis (a condition caused by
interference of venous flow out of a limb) while it is only 5%
for the KG (13)
In 28mm diameter cavae in vitro the 11U was observed to
dislodge and migrate.
Twenty-eight incidences of 1MU migration
have been reported clinically in 2215 deliveries (28).
Tilting of the KG filter seen in vitro has been reported
clinically (14,15) , was seen in 50% of Beth Israel hospital KG
filter patients and occurred in 63% of in vitro KG filter
Celiveries.
'The MIU and KG hooks are effective in preventing downstream
migration but offer little
awCay from the heart.
resistance to migration distally,
Distal rigration of the MU and KG filter
has been documented clinically (15,2).
These correlations with clinical observations suggest that
-27this si.mulation is a good model for predicting the clinical
properties of new IVC blood clot filters.
Clinical ImnlicatiQns
Filter clot capturing ability and anchoring security
diminished with increasing vena cava diameter.
In the 28mm
diameter cava the KG filter allowed most 7mm diameter clots to
pass and the MU migrated.
A decision to place a KG filter in a
patient with a cava 28 mm or more in diameter should therefore
be taken with caution and the MU is contraindicated.
Future
research might explore ways of retaining filter function in a
wider range of cava diameters or develop multiple sizes of
filters so appropriate filters will be available for every
patient threatened by pulmonary embolism.
Many clinicians report that a 7mm diameter embolus is
likely to be lethal or, at the vey least, clinically significant
(resulting in physiologic changes which can be detected by
routine clinical exam).
But the KG let 70% of 7mm diameter
clots through dispite its low reported recurrence rate for
pulmonary embolism clinically (3).
This observation can be
explained in three ways. First, the present in vitro simulation
may be unrealistic and 7mm diameter clots may not in fact pass
through the KG filter.
Second,
the clinical findings reported
may not reflect the true incidence of recurrent pulmonary
embolisms with the KG filter.
KG filter
Finally,
clinical trials
of the
may have yielded inaccurate results for the following
reasons:
(1) the patients selected for study did not have
m±grating emboli,
(2) more than one type of therapy was
employed, e.g. filter plus anticoagulation, (3) there was
inadequate patient follow-up, (4) there was insufficient
While
follow-up time for the clinical studies to be conclusive.
the present analysis of the KG filter was as precise as
possible,
the results of this study do not permit complete
resolution of these issues. But a filter that captures clots
better than the KG in vitro is likely to capture clots
effectively clinically.
Optimal filter design in terms of maximizing clot capturing
ability while minimizing filter pressure gradients was achieved
by the Nitinol filter which had the most uniform mesh.
Redesigning the MU and KG filters to have more uniform meshes
ýwould probably improve their performance significantly.
The MU
could have a fish net like structure in place of the silastic
membrane with occasional holes, thus reducing its obstructive
effects.
The KG could have the six tines branch to form 12 at
the periphery where most clots currently get through.
The most significant implication of this research is that
the favorable performance of the flitinol filter warrants its
further evaluation.
This study shoýwed it to anchor securely and
to capture emboli effectively without obstructing the IVC; all
this better than the two filters currently used clinically.
addition,
is
the Nitinol filter is
inserted without surgery.
shown also to be biocompatible
(30
),
it
should be
In
If
it
-29evaluated clinically.
-30-
Summary and ConclusioDns
An in vitro comparison was made of three filters designed to
prevent clinically significant and especially'potentially lethal
emboli from reaching the lungs.
Two of the filters, the Kimray
Greenfield and J1obin-Uddin, are currently used clinically; the
third, the Nitinol (a memory wire device), is experimental.
The
filters were delivered into an artificial vena cava where
measurements of clot capturing ability, flow interference
effects, anchoring security and orientation were obtained.
The -
Mlobin-Uddin captured clots well in small and average size cavae
but was very occlusive.
In large cavae the problem of
The Kimray Greenfield
Mobin-Uddin filter migration was severe.
was not very occlusive but tended to fall into a tilted
orientation and allowed large clots to pass, particularly after
occlusion of the small spaces in the apex.
captured clots well in all caval sizes,
The Nitinol filter
displayed the smallest
occlusive effect and was remarkably stable.
Since the Nitinol
filter can be inserted into the vena cava through the lumen of a
catheter, while the Greenfield and Mobin-Uddin require surgery,
it
is
clearly the easiest and safest of the three devices to
deliver.
While the present simulation of blood flow in the vena cava
did not completely parallel that of the human,
it
was
sufficiently close to permit a reasonable evhluation of the
three filters.
The excellent performance characteristics of the
-31-
Nitinol filter
warr-ants further research on its biologic aspects
through studies on both animals and selected human subjects.
-32-
1.
Dalen JE and Alpert JS: Natural
Embolism.
2.
History of Pulmonary
Prog Card Dis 17:259-270,
1975.
Kakkar VV: The Current Status of Low-Dose Hleparin in the
Prophylaxis of Thrombophlebitis and Pulmonary Embolism.
World J Surg 2:3-18,
3.
1978.
Greenfield LJ and Zocco JJ: "Intraluminal Management of
Acute Massive Pulmonary Thromboembolism," Journal of
Thoracic and Cardiovascular Surgery 77:402-410,
March,
4.
Sabiston DC et Al: Ann Surg 185:699-711,
5.
Bernstein EF: The Role of Operative Inferior Vena Caval
1979.
1977.
Interruption in the Management of Venous Thromboembolism.
Wiorld J Surg 2:61-71,
6.
1978.
Mobin-Uddin K, McClean R, Jude JR: A New Catheter Technique
of Interuption of Inferior Vena Cava for Prevention of
Pulmonary Embolism.
7.
1969.
liobin-Uddin K et Al: Transvenous Caval Interruption with
Umbrella Filter.
8.
Am Surg 35:889-894,
Greenfield,
LJ,
N Eng J 1led 286:55-58,
McCurdy,
JR,
Brown,
PP,
1972.
Elkins,
RC: A 1New
Intracaval Filter Permitting Continued Flow and Resolution
of• Emboli,"
9.
73:599-606,
April,
1973.
Simnon 8: A vena Cava Filter Using Thermal Shape Miemory
Alloy.
10.
Surgery,
Radiology 125:89-94,
1977.
Simon 8 and Palestrant AM: Transvenous Devices for the
--~··~·~p~g·
~ 7...7-7777-77-777----
-33lianagement of Pulmonary Embolism.
11.
Philli)s, HN,
idrich,
C, and Johnson,
C:
erforation o
the Inferior Vena Cava by the Kim-Ray Greenfield Filter.
Surgery 233-5, 1980.
12.
Sautter,
RD,
Myers,
7WO,Lawton,
Cava Filter 1Migration.
BR: Experience with Vena
(letter to the editor),
JAIA
219:1217, 1972.
13.
Cimochowski GE,
Evans RH,
Zarins CK, Lu CT, Demeester TR:
Greenfield filter versus Iiobin-Uddin umbrella
The
continuing wuest for the ideal method of vena caval
interruption. J Thor Cardiovasc Surg 79:353-365, 1980.
14.
Wingerd 1,
Bernhard VI,
Macaddison F,
Towne JB: Comparison of
Caval Filters in the iManagement of Venous Thromboembolism.
Arch Surg 113:1264-1268,1978.
15.
Berland LL,
Maddison FE,
Bernhard Vii: Radiologic Follow-up
of Vena Cava Filter Devices.
16.
AJR 134:1047-1052,
1980.
Mobin-Uddin K, Utley JR, Bryant LR: The Inferior Vena Cava
Umbrella Filter. Prog Card Dis 17:391-399, 1975.
17.
Sher
.MHP1: Complications in the Application of the Inferior
Vena Cava Umbrella Technique.
18.
Gaston EA:
JAd1A 214:2338,
1970.
Sauters RD: Experience with Vena Cava Filter Miigration.
219:1217,
20.
1971.
Incorrect Placement of Intracaval Prothesis for
Pulmonary Embolism.
19.
Arch Surg 103:688-690,
JAVA
1972.
••
Schroeder TM, Elkins RC, Gr-eenfield LJ: Entr apment
of Sized
Emboli by the Ki·A-Greenfield
Intracaval Filter. Surgery
-3483:435-439,
21.
Brown PP,
1978.
Peyton iID,
Elkins PC, Creenfield LJ:
Experimental
Comparison of a New Intracaval Filter with the M'obin-Uddin
Umbrella Device. Card Surg Suppl II to vols. 49 &
50:272-276,
1974.
22. Mobin-Uddin, K, Smith, PE, Hartinez,LD, Lombardo, CR, Jude,
JR: A Vena Cava Filter for the Prevention of Pulmonary
Embolus. Surg Forum, 18:209, 1967.
23.
Gianturco, C, Anderson, JH,
Filter:
Experimental Animal Evaluation. Radiology
137:835-837,
24.
Prince
and Wallace, S: A New Vena Cava
1980.
iDR: Simulation of the Human Inferior Vena Cava for
Evaluating IVC Interruption Devices. Bachelor's Thesis in
M•echanical
Engineering,
M.I.T.,
1980.
25. Novelline, PA: personal communication, Massachusetts General
Hospital, Boston, Oct.
26.
1, 1981.
Guyton AC: Textbook of Medical SchoL1 PthysiQ
vIgy,
B
Saunders Co. Philadelphia 1976, p.251.
27.
Correct technique for introducing filters
28.
West JB: Respiratory Physiology Williams & Wilkins,
29.
Gardner AR1I1:
Baltimore,
the essentials2nd edition,
1979.
Inferior Vena Caval Interruption in the
Prevention of Fatal Pulmonary Embolism.
Amer Heart J
95:679-G82, 1978.
30.
Castleman LS: Biocompatibility of Nitinol Alloy as an
Implant Miaterial.
J Biomed M!at Res 10:695-731,
1976.
E
v
H
O
o
0
cti
3
PO
Cd
Cd
0
S
*
H
Cd
O
.- t
.,
0
'H
42
O
U
'H
I ~
.H
0
4
S
rC
rd
0
H
0O
CD
0
Cd
0
rd
0
.H
~
__
OO
0
OL
Lo
C -H
0
"'
05
0
~-r Lr
CO
O
0
4-i
to
m
0
E
0
'H
coL
CO
to
Cd4J
'H O
O
S
ct
od
d 't 40
-H - .
0
Cd OH-
4- Cr, C
Cd 'H 'H
_~__~
_ ____C
4-J
to
ac D
CQdd
'H 42H
rd
tato
o0 '
'H
>54
UL-
0
i------I I-
F-4
Cd
00
4
U
00
4-i
0~
k
Cd
~
4-i
Cd
H
0Cd
H
4-J o
Cd
0
Cd
0
LH
r--4
CD
rd
O
0
H
0
0
Z
Cd
0
Cd
CO
Ici
U
..----------m1115·-·-
CO
rd
L)
0
-
4-1
0
4-J
Cd
0
4-J
ICd
H
S
4-J
Cd
0
0
-1
U
Q)
H
k
4
4-i
H
z0
i
0
4..
-H
0
U
tO
Cd
U0
Cd
Cd
0
0
U
4(
4-J
CO
,-
.,-
CO
to
C-)
-H
a
42
O
H0
4
o0
4
H
42
0
Et
a,ý4
4_1
O
H
I
0
CI
0
S
e o
0
QJ
H-{
Oo
co
CO
0
o
4-J
0
H0
-36-
Table 3
Filter
Deliveries
Mobin-Uddin
Greenfield
15
20
28
15
20
28
15
20
28
Orientation:
% tilted
60
90
*
60
80
50
40
70
0
% central
40
10
*
40
20
50
60
30
ý100
Holding Force
mean (gm)
83
117
0
77
55
34
91
55
49
20
33
0
12
6
22
25
21
37
6
5
6
4
2
0
0
0
IVC diam (cm)
standard
deviation
# hooks engaged
Failure mode
*
tear
0
tear
Nitinol
slip
the Mobin-Uddin Umbrella was not large enough to engage the cava
walls in the 28mm diameter cava.
For every delivery the filter
simply rolled on downstream.
-37PULMONARY EMBOLISM
Physiology
1)
Clot forms in
2)
lower body
Clot breaks loose and migrates toward lungs
3)
Clot plugs pulmonary artery
Anticoagulant therapy
prevents new clot formation
IVC interruption prevents
large clots from reaching
the lungs
Surgery removes clot
Thrombolytic drugs lyse clots
Figure 1
-Pulmonary
Embolism:
Physiology and Treatment.
-38-
sutures
DEWEESE FILTER
clips
ADAMS-DEWEESE
MORETZ
MILES
ai
Figure 2A
Direct surgical techniques for interrupting the IVC
intracaval devices
Mob in-Uddin
ad·iPd··bl-s
'"
MU delivery capsule
KG delivery capsule
delivery catheter
-L"-11 aýLNAA
4--.
FIGURE
2B
inside the-- IVC
Interruption devices that fit 1-
-39TECHNIQUE
A)
FOR INSERTING THE KINRAY GREENFIELD FILTER
Load filter
into capsule.
'
I
U
--B)
Expose internal jugular vein
.mmwmlmpý
(
Withdraw
-a
'-~---__
capsule.
-40INSERTION OF THE NITINOL FILTER
A)
/
Insert
and ini
i-••
C'
K
/
f
'
IL
I
1
'I
I
B)
Insert guidewire through ne
and into femoral vein.
- ·-------------
C)
-
---
- - --
----·--
Remove needle leaving guidewire
in vein.
Insert catheter into femoral
vein by sliding over guidewire.
I
r
r
r
r
I I
Figure 4a. Transcatheter approach to IVC interruption.
I
-41I
(
D)
Attach filter storage and
feeder device.
Advance Nitinol filter
into
catheter (in its straight
wire, low temperature form
-c~---~-~ ---
E)
Push filter
through catheter
and into the IVC.
(It recovers its high tempera
configuration instantl
filter
upon contact with blood at bo
temperature).
-------
F)
af
-
--
I
T-
-
-
Remove catheter and apply
compression at site of needle
puncture for 10 minutes to
prevent bleeding.
\ •
k
i
I
r
k
Figure 4b. Transcatheter approach to IVC interruption.
-42-
Large Dog
Average human
Large human
GREENFIELD
15mn5
t2Omm -4
,
i.
4- 28m.m
MOBIN-UDDIN
NITINOL
Figure 5
Variations in filter mesh with changing vena cava diameter
-43-
between hooks
s~
=
= ~spacing
f
d= cava diameter
2?~
L.-.&a
-isolate two adjacent limbs
-look at the largest circle
contained within the triangle
1-NT
fmInno'
Amfq"MA
sin(s)
sin() ab
cos()
=
c =2rb
r
c
2rb
a b
=
2r=- 2b+ab
r
2bba
b sin 2b
2bta
2r 2b+aCos
2r
cava diameter
(mm))
10
dog size
I
.range of
typical
human
very
large
human
I
12
14
16
18
20
22
24
26
28
30
32
34
36
38
Figure 6.
spacing between hooks
(mm)
5.2
6.3
7.3
8.4
9. 4
10.5
11.5
12.6
13.6
14,7
15.7
16.8
17.8
18.8
19.9
2abaj
cos(s)
S
diameter of largest hole
(mm)
3.6
4.3
5.0
5.6
6.3
7.0
7.6
8.3
8.9
9.5
10.1
10.7
11.3
11.9
12.5
KG filter hole size as a function of cava diameter.
r
Co
O
z
LL
0O•
CU
I
7I
-44-
m
,)
cc
ILl
co: w
LU
cc.
7,F
E
a
cU
in
U
a4
w
>-
C)
0
cc
LL
0
O
C)
OI
-45-
LC
0
4-J
O
o*
-46-
14
a Cava Diameters
12
patients with
L)
10
y Greenfield Filters
4-J
Cd
a) 8
44
0
6
a
15
4
wl
12
14
16
18
20
22
24
32
32
28r
26
30
28
Vena Cava Diameter (mm)
number
mean cava
diameter
Kimray Greenfield
with venogram
20
20
3.8
Kimray Greenfield
without venogram
45
19
3.2
total
65
20
3.4
.. .
Figure 9.
standard
deviation
,-
•,
....
. i.
...
.---
Infrarenal inferior vena cava diameters in KG filter patients.
-47-
,,
- ~- --'; ~ ---
i
--··
A)
Filter in
capsule.
(1-_;~~~---~--I I; l~-~kid
·-·I:---··-illF---_·A··
. L-6-
__
_·,,,
B)
-I
~
Capsule withdrawn
exposing filter.
~~
Mp
filter jumps out.
riia
on
<Ir--A
----
I
'
''-
i
D)
--
Figure 10.
-
KG filter delivery.
!ip
Capsule removed.
-48-
EXISTING HOOK DESIGN
hook not engaged
SUGGESTED HOOK DESIGN
both hooks engaged
note larger hook angle increases
the probability of hook engaging
Figure 11.
Analysis of Greenfield filter hooks
i
-49-
Calculation of Optimal Greenfield Hook Angle
7,
hook angle =
limb length
=
b
D= largest possible cava diameter
e
= angle between limb and cava wall
S= sin)1 D
if oC :•
then hook will not point into and engage cava
if -& > 90
hook will slip o'ut of cava wall
thus 90 >
optimal oC
let D = 40mm
90
>
Figure 12.
c.
and
>
b = 46rm
60
Determination of optimal KG filter hook angle.
-50-
12Q
Filter Holding Force
versus
- :
Cava Diameter
r
100
IF
8d
" Mobin-Uddin
h"
N*
N44
*N%
a 60
.i
I'
40o
%,Sreenfield
---
--
= Nitinol
'Til
*20
.....
20
.~.....
= Mobin-Uddin
--
-
= Greenfield
,1
,
3
15
20
28
Cava Diameter
Figure 13.
(nm)
Filter holding force versus cava diameter.
-51-
30.
Average Filter Pressure Gradient
\9
versus
Cava Diameter
cl
. Mobin-Uddin
.
r
r
r
~3220,
r
1
4-4
a)
to
k
0)
0
rH
·
S
t~
o
21
*
I--0...
. .. ...
Nitinol
.......- Mobin-Uddin
r
--
-&•
.
.Greenfield
C)
Greenfield
U
15
"two
20
Cava Diameter (mm)
Figure 14.- Average filter pressure gradients versus cava diameter.
28
-52-
100 -
..
..
obin-Uddin
Nitinol
90
% of Clots Capdtured Completely
ver ~us
80
Cava Diameter
70
H60
5Q
\ reenfield (cent ral orientation)
0
0
040
SGreenfield
-
,tilted
orientation)
30
Nitinol
.--20
-
--.
'
--- = Greenfield-central
7 --
-
-
=.Mobin-Uddin
= Greenfield-tilted
)c
0
Figure 15.
15
20
Cava Diameter (mrm)
Comparative clot capture performance.
28
-53-
I
nfl
Nitinol
1%
a
r.._i
ussF.runarru~~l-u4+rr~·l~'~cupe~i~i4~*
i+Ld~~··~;sulu~a~~YI,~
U.-...
r
-
".Mobin-Uddin
S.
Ile OF
ogoo °°r+°o
lot
90-.
BI'Ir
fp
.
op
\Greenfield
80-
70-
ý4
K
50 0
rH
% of 7mm diameter clots captured
Q
(Ci
versus
Cava Diameter
H40
0
30
-·.
-----
Nitinol
At..
Mobin-Uddin
-- J-- --
20
Greenfield
10
0
?/
~oaa~Ii
I
1•O
5
Cava Diameter
Figure 16.
i
uraqn
-·-----
28
20n
(nrm)
Comparative clot capture performance
(7mm clots only).
-54-----r~c-,----·~--·I
'i '
r
C~I~WCa I
--· r~r_)~~yc·C~
-i
GE.1EIE L5D
15mm
--------------2cm
10cm
Length
P
4
*
+
60
70
40
30
0
0
0
0
0
0
0
0
I
4. 5
3
2
1
Order
-
4mm
7mm
7mm 4mm
Diameter
2
3
4
5
*
*
*
*
*
*
*
3
3
5
100 80
70
50
90
78
80
80
30
40
20
50
20 $20
40
10
18
10
20
40
40
20
26
0
10
10
0
4
10
0
30
20
60
24
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
*
P pressure in
3
1
cm of H20
+ %clots captured completely
+-
%clots
captured partially
- %clots passing completely through filter
I %cases of insufficient flow
M %cases of filter migration
* pressure never exceeded 5cm of H 2 0
.
Figure 17.
.
..
. .
"-•
--
,,
• .
.... •
:_•
-:;
. . . : L
•
'": : .•
.
::,! :
~
• -'.
. .' '
...
_____ 1 ____
i -·_ -55-
- I'- ~---- c
~___
-·
-r
GAEELmEIRD
2,O.mm
^""L-~
Length
10cm
Diameter
2cm
7mm 4mm
Order
P
*
*
7mm
4rmm
1
2
3
4
5
1
*
*
*
2
2
1
1
2
3
4
5
Z
*
*
*
*
*
90
20
75
70
55
55
40
59
70
40
60
30
10
42
10
80
25
25
25
30
45
30
20
40
10
40
30
28
0
0
0
5
20
15
15
11
10
20
30
30
60
30
1
0
0
0
0
0
0
01
0
0
0
0
0
0
M
0
0
0
0
0
0
0
0 000
0
0
0
+-
P pressure in
0
0
cm of H2 0
+ %clots captured completely
+-
%clots captured partially
- %clots passing completely through filter
I %cases of insufficient flow
M %cases of filter migration
* pressure never exceeded 5cm of H20
II
Figure 18.
·-I
· ----~-------i
---
p
-56-
,--,----------·'"T~'~*-C~Ch~.~Lsrt~I·I
I
I
r
GREE NFI" I.
II
L
·
13
~·i·
--
TLLLED 2DmJLwn
2cm
0cm
Length
7ram 4mm
Diameter
Orde1
7mm
3. 4
2
+
80
5
1
2
0
80
20
30
20
30
36
0
0
" 20
90
20
70
30
70
50
48
40
0
10
0
10
40
10
20
16
I
0
0
0
0
0
0
0
M
0
0
0
0
0
0 0 0
4mm
4
3
5
Z
0
0
0
0
40
10
40
50 6 36
60
60
90
60
50
0
0
0
0
0
0
0
0
0
0
0
0
0
0
64
P pressure in cm of H20
+ %clots captured completely
+- %clots captured partially
- %clots passing completely through filter
I %cases of insufficient flow
M %cases of filter migration
_• _..
·-.----~-----c~
Figure 19.
• -
.
.
* pressure never exceeded 5cm of H20
"
'
.... '
...
. . .
..: ! ...
' :" . . <
"i
k_
•
• • i •-_-
v -.. .
I
: ._1
.
-57----. ;··--·------
----~-~r-~-----u
----
I
:i·
·
--
4
._2i mL
2cm
10cm
Length
7mram 4mm
Diameter
Order
4mm
7mm
1
2
3
4
5
1
2
3
4
5
+
0
0
20
20
30
10
10
18
10
10
0
0
0
4
+-
80
100
70
60
30
40
10
42
80
80
50
40 90
68
-
20
0
10
20
40
50
80
40
10
10
50
60 10
28
I
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
M
0
0
0
0
0
0
0
0
0
0
0
0
0
P pressure in cm of H2 0
+ %clots captured completely
+-
%clots captured partially
- %clots passing completely through filter
I %cases of insufficient flow
M %cases of filter migration
* pressure never exceeded 5cm of H20
2
C--
~----
Figure 20.
I
·;iiI
-58- ,_
I ~C ~`
----- -~e._~~-r~4
_-- --- ~.--- - -I ---
__
C
rr
r i
--
ri-mu
1GRIED.ll_D_28_mm_
--
·rr
2cm
10cm
Length
Diameter
7 **
0
+
3 .. 4
2
1
Order
P
4mm
7mm
7mm
5
J
*
010
10
0
*
1
2
*_
*
4
3
5
*1*'
*
4
0
0
0
0
0
0
1001100
50
30
10
20
20
26
100
90
50
30
50
64
0
0
50
70
80
70
80
70
0
10
50
70
50
36
S
0
0
0
0
0o0
0
0
0
0o
0
0
M
0
0
0
0
0
0
0
0
0
0
0
+-
0
0
0
P pressure in cm of H2 0
+ %clots captured completely
+- %clots captured partially
- %clots passing completely through filter
I %cases of insufficient flow
M %cases of filter migration
* pressure never exceeded 5cm of H 0
2
Figure 21.
'"'
.:':-~;
;.;::.';.'';::
1:·;_:
___~ ABW'MkqWWt'-59=._..__,,
II8M
15nmm
10cm
Length
2cm
7mm 4mm
Diameter
Order
4mm
7mm
__-I
Order1
2
3 .. 4
5
1
1
2
3
4
5
30
5
29
29
35
35
27
100 100 .20
20
0
48
P
33
20
16
30
33
35
35
+
90
100 100
80
20
10
0
42
+-
10
0
0
0
0
0
0
0 0
0
58
0
0
0 0
I
M
0
0
10
100
0
0 0 0
0
20
00
80
90
0
00
P pressure in
100
0
0
0
0
0'
0
0
0
0O
80
0 0
10
80
0
_
0
0
52
00
cm of H2 0 -
+ %clots captured completely
+-
%clots captured partially
- %clots passing completely through filter
I %cases of insufficient flow
M %cases of filter migration
* pressure never exceeded 5cm of H 2 0
--
-~
--
-
Figure 22.
-
--
-60I---~'-"
r-~~'
-
=L~L~-"L
-
-
~L-C --
LX
'
IcJ '
-
--
-
-~-s~
i··
-
---
-i
i'
1--
IN
B.INA-D.
20m_
-19.II3~
2cm
10cm
Length
+ 100 100
5
7
9
12
6
90
60
301
76
0
0
10
4
0
0
0 30
0
6
34
0
0
0
10
60
14
2
0
0
0
0
0
0
1
12
13
11
*
100100 90
10
10
62
0
0
2
0
0
2
9D
0
2
3..4
6
12
0
0
0
0
0
0
0
0
0 10
I
0
0
0
0
0
80
M
0
0
0
0 10
0
-
4
2
7
20
P
3
5
1
Order
4mrm
7mm
7mm 4mm
Diameter
10
2
*
100 100
P pressure in cm of H2 0.
+ %clots captured completely
+-
%clots captured partially
- %clots passing completely through filter
I %cases of insufficient flow
M %cases of filter
migration
* pressure never
. . .
-·C
I
Figure 23.
..... .
.
.... : .
• .:..
..
: •
_
..
. ·_
..•:. .
exceeded 5cm of H20
• . -: .. _. . >:; . .. .• u . -i '
i f -.
._• :• . i
'I
;-··--
-61I-
,,,,,
c~~=-~~---·---~-'PU------
-I
i
M.0 N-m
D~atP
98mm
0
-C
0Z
C
l
163
2cm
10cm
Length
7mm 4ramm
Diame ter
"b
1
Order
3
P
28
+
30 160
4mm
7mm
0
70
4
1
2
3
2
0
0
1
10
30
40
30
10
5
3
-4.
2
2
2
2
40
10
20
2
20
10
0
0
0
0
0
0
0
0
0
0
30
40
20
30
24
10
I
0
0
0
0
0
0
0
0
0
M
50
30
30
30
50
60
60
46
50
0
0
10
10
20.
0
-20 20
10
10
20
20
20
30
20
0
0
0
0
50 50
50
0
50 50
+ %clots captured completely
%clots captured partially
- %clots passing completely through filter
I %cases of insufficient flow
M %cases of filter migration
T-Figure 24.
x
1
P pressure in cm of H2 0
+-
5
-62-I
___--_ __ _ .----: -- ----i------- -~---;;~---
·
g
-
___
_I ,
---
Nff IN~L
2cm
10cm
Length
1
Order
P
+
+-
M,
2
3.
4
*
*
*
*4
1100
4mm
7mm
7mm 4mm
Diameter
5
2
1
2
4
2
*
*
90
80
100 100 100 100 100
100
4
5
I
*
*
*
*
90
70
100
86
0 10
0
2
20
0
12
3
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
10
20
0 o
'I
0
0o
0
0
0
0
0
0
0
0
0
j0
0
0
0
0
0
0
0
0
P pressure in
10
0 0
0
0
0
cm of H2 0
+ %clots captured completely
+-
%clots captured partially
- %clots passing completely through filter
I %cases of insufficient flow
M %cases of filter
migration
* pressure never exceeded 5cm of H20
`CI-C-I--I
Figure 25.
..
WN
W-..
_
-- ---_·
;--·- _~ ,,,,~~e -63-
-Omni
/7
2cm
10cm
Length
I
Order
+ 100 100
+-
*
*
*
P
0
0 o 0
2 3
X 1
3.4. 5
* 12 2.1 0.8 *
2
*
100 100 100 100 100 100
0
0
4mm
7mm
7rnnAmrmn
Diameter
*
*
100 100 .90
4 5
*
*
*
100 100 98
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
10
0
0
2
0
0
0
0
0
0
0
0
0
0
0
0
0
0
I
0
0
0
0
0
0 O-
M
0
0
0
0
0
0
0
P pressure in cm of H 2 0-
+ %clots captured completely
%clots captured partially
+-
- %clots passing completely through filter
I %cases of insufficient flow
M %cases of filter migration
~~IY9~~13LL~~~
-~
Figure 26.
-
I--
-
-
~
* pressure never exceeded 5cm of H20
IIr
~
~
-
.'--
-64I---I--
'-
7mm 4rmm
2
1
5
3.4
_
_
2
3
4
5
*
*
*
*
.*-
*
100 100 100 100 100 100
100
90 100 100
100
98
0
0
0
10
0
0
0
2
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
*
*
*
*
P
-*
*
+
100
80
+0
20
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
M
'I-ii--
4mm
71m
Order1
-
-
2cm
10cm
Diameter
-
T110 L
-2-8
m
~~NJ
Length
-
*
P pressure in cm of H2 0
+ %clots captured completely
+- %clots captured partially
- %clots passing completely through filter
I %cases of insufficient flow
M %cases
of filter
migration
* pressure never exceeded 5cm of H 0
-- -Figure 27.
2