Gretchen
A. W. Gooding,
MD
Sandra
Perez,
#{149}
RN
Joseph
H. Rapp,
#{149}
MD
William
C. Krupski,
#{149}
MD
Lower-Extremity Vascular Grafts Placed for Peripheral Vascular Disease: Prospective Evaluation with Duplex Doppler Sonography’ Eighty-five men with 92 vascular grafts placed for peripheral vascular disease of the lower extremity underwent a total of 264 examinations with duplex Doppler over a 2.5-year penod. In 64 patients who underwent more than one examination, the total follow-up encompassed 740 months. In 220 native femoral arteries (96.0%) the peak systolic velocity (PSV) was higher than that in the graft. Arteriovenous shunting was associated with a normal PSV and a markedly elevated diastolic component at spectral analysis. Focal fluid collections were common initially near the graft and usually disappeared uneventfully. An average PSV of 32 cm/sec or less was always associated with impending occlusion. The sensitivity of an average PSV of 40 cm/sec or less to indicate impending graft occlusion by the next visit was only 33%; the specificity, 94%. At initial examination, stenoses were associated with high PSV focally in the graft or low PSv with absent diastolic flow. Index
terms:
Arteries,
teries, grafts and Arteries, stenosis teries, US studies,
supply,
92.7214
femoral,
prostheses,
or obstruction, 92.12984
92.7214
92.45,
#{149} Ar-
92.7214
92.7214 #{149}ArExtremities, blood (US), Doppler
#{149}
Ultrasound
#{149}
T
advent
HE
lows
of more
(US)
definition
of both
astolic
forward
flow
tients
who
had
of a vascular cular disease
pler
unit
(SPA
Calif). Their years
1000;
(average
age,
study
after
and
surgery.
Of the
85 patients underwent
the initial
Received
requested ceived
March
dress
reprint
C
November
February
RSNA,
29, 1989;
revision
20, 1990; final revision
25, 1991;
accepted
requests
to G.A.W.G.
1991
from April
April
re-
1. Ad-
one.
6 weeks, during
3
21
92
264
3 months,
6
what
actu-
64 pa-
examinations
than
In these
64 patients, the encompassed 740 months
total follow-up (average follow-up per graft, 11.6 months). The range of follow-up in these 64 patients
was
2 days
Imaging tient distal
the native
(with
at
the dorsalis
pedis
of these
specffic
were
performed:
artery.
levels,
seven
1. An image of the graft was generated. 2. A measurement of peak systolic velocity (PSV) in centimeters per second was obtained from this image. 3. A measurement of time-average yelodty (TAV) was also obtained from this image. The TAV measurement was generated with the range gate opened the width of the vessel on longitudinal scan. With range-gated black-and-white duplex which
was
used
in this
study,
abnormal sites of altered flow dynamics can be detected only by actual sampling of the flow velocity with analysis of the spectral data. It is possible that focal stenosis or other abnormalities were present but undetected in areas of the graft not specifically examined. Color-coded Doppler would changes
eliminate in color
this deficiency because indicate significant ab-
normalities in velocity or direction visually (3). The angle of insonation was 40.-60.
the pa-
vessel
to the
at specific areas along The graft was visualized its course with sonography,
anastomosis
the graft. throughout
from
and spectral only at desigfemoral artery;
to 27 months.
was performed
supine)
36
examination
was
Doppler,
to undergo
in Table 1. with 92 grafts,
more
1987 study,
the second
However,
is seen
imag-
lower-extrem-
patients
tients
of Radiology
7
calculations
separate
prospective
1 week,
Studies 23 42 60 52 30
At each
Milpitas,
grafts were performed to October 1989. In this
1 year,
Total No. of
the proximal, middle, and distal thigh; at the popliteal fossa; and, in more distal grafts, at the proximal, middle, and distal calf and at the anastomosis, which most
69.2 years).
in the
with
23 21 20 13 6 6
analysis were performed nated sites: at the native
from 56 to 92
of 92 peripheral
ally occurred
94121.
Diasonics,
sixty-four
enrolled
Graft
1 2 3 4 5 6
but Doppler
METHODS
ages ranged
No. of Graft Examinations
Total
Eighty-five men with peripheral grafts of the lower extremity underwent examination with a high-resolution (10-MHz imaging, 4.5-MHz Doppler) duplex Dop-
year
Departments
No. of US Studies
distally
months,
the
Infrainguinal
placement
AND
and in 85 Patients Grafts
a prognostic
ity
(G.A.W.G.) and Surgery (S.P., J.H.R., W.C.K.), San Francisco Veterans Administration Medical Center, 4150 Clement St. San Francisco, CA
of US Studies
Examinations
graft for peripheral vasof the lower extremity.
examination
From
are
1
Number
morphol-
undergone
PATIENTS
ing studies
180:379-386
the
Table
al-
indicator of poor graft survival (1,2). Our study prospectively examined the utility of duplex US to determine morphology and flow characteristics of 92 infrainguinal grafts in 85 pa-
Two hundred 1991;
technology
ogy and dynamics of flow in arteries and veins. The utility of this technology in the surveillance of infrainguinal vascular grafts placed for the treatment of peripheral vascular disease is a subject of great interest that holds promise in the delineation and detection of abnormal flow patterns to predict impending graft failure. Bandyk et al have shown that low graft blood velocities with absent di-
studies Radiology
sophisticated
ultrasound
Abbreviations: peak systolic
AV velocity,
=
arteriovenous, PSV = TAV = time-average ye-
locity.
379
Table 2 Average Values Value
of All Patent PSV
TAV
(crn/sec)
(cm/sec)
(cm)
(mL/min)
74.8
18.2 3-89
.48 .3-.8
175.6 23-646
Mean Range
34-118
Table 3 Average Values Figure 1. A longitudinal mal triphasic distal thigh
sonogram of a norsegment of graft
demonstrates a TAV measurement. The markers are on an initial peak systole and final peak systole. The range gate is open the entire vessel. DTG = distal thigh graft.
of G rafts
Diameter
0 cclusions
with
TAV
PSV
a to
Grafts
Value
(cm/sec)
(cm/sec)
Mean Range
61.4 21.7-120
9.4 2.3-27.7
this
diameter,
milliliters calculation
[cubic
a flow
measurement
was on
calculated. the formula:
per minute was based
centimeters
velocity
per
[centimeters
second] per
second]
(This flow
of arterial 6. Then
lumen [square the technologist
centimeters]).
cumference
of the graft
on a transverse
scan, and an area of the graft ously
defined
7. The
levels
machine
per minute)
liters
images
were
was
each
and on the the machine
flow
locity require
(Fig 1). TAV the range
compass
the
that the parabolic central
under and
adjacent
cluded.
The
of both
lower
to the
following
Germany) (n = 28); (n = 47);
wall
A specific diastole, but
whether present base, systolic
(millimeters
per
value was a notation
a diastolic or absent, or (c) elevated
en-
so
are
in-
was
used:
minute)
component compared
(4).
subject
to many
tamed to determine had any relevance
but
ours
The
same and
every
380
obtained
case
The in the
technician
did
Radiology
femoroanterior femoroposterior
Putzbrunn,
Area Flow (mL/min)
.22 .12-39
145.9 10-312
in situ
During
the
tients died unavailable
ob-
all the examiperiod. performed femo-
61-80
81100
1o2o1214:
PSV (in centimeters
per
second; values shown at bottom) in grafts that were occluded by the time of the subsequent examination (diagonal bars) (20 examinations) compared with average PSV in patent grafts (black bars) (233 examinations). In a total of 264 graft examinations, PSV was
in two examinations;
no PSV
was calculated and a 2#{176} occlusion was noted in nine examinations; grafts were patent in 233 examinations; and in 20 examinations, occlusion was shown in the next study.
tibial in situ tibial, reversed
saphenous pedis
in
phy,
grafts.
36 grafts;
in
study,
17 of the 85 pa-
(20%) and follow-up was in five patients. Thirty-
two graft examinations formed in those who
were perdied, 11 in the
five
follow-up;
patients
patient
was
without
lost
to follow-up
after
six
examinations over 1.5 years. At the conclusion of the study, 51 grafts were functioning (55%), 21 had occlusion (23%), and six had been re-
done
(6%).
performed
Thrombectomy in eight
grafts;
had angiogra-
been
angioplasty,
eight
This
grafts in which angioplasty had been performed. Of the eight grafts that required thrombectomy, seven continued to function; occlusion occurred
other.
Of the
six grafts
occlusion
that
occurred
had in
four, and two continued to function. In 230 of a possible 253 femoral arteries, the mean PSV was 124 cm/sec. Psv was not calculated in 23 femoral arteries wound,
one
and
left 50 patients with 51 grafts. At the end of the study, occlusion had occurred in all but two of the
in the
RESULTS
are
were
measurements
over a 30-45-minute 264 examinations were following types of grafts:
#{149}
polytetrafluoroethylene
41-60
Average
not calculated 28);
=
(n = 4); femoroposterior tibial, vein (n = 1); femoroposterior ribial in situ (n = 38); femoroperoneal in situ (n = 27); femoroperoneal, reversed saphenous (n = 3); femorodorsalis pedis, reversed saphenous (n = 6); femorodorsalis pedis in situ (n = 57); popliteal-posterior tibial, reversed saphenous (n = 3); popli-
the the
whether or not they to the diagnosis of graft
the
(n
21-40
3.
to be redone,
failure. nations
saphenous
Gore & Associates, (n = 14); femoropopliteal
was (a) or below with
in this study, errors,
of
for as to
measurements of graft obtained as meticulously
as the measurements
were ocat the top
of examinations.
teal-dorsalis pedis, reversed (n = 5); and popliteal-dorsalis situ (ti = 3).
component.
Such calculated flow, even those
PSV
cephalic
higher
not recorded was made
(b) above
Area (cm2)
saphenous
peripheral
formula
reversed
(Gore-Tex;
=
flow
nine grafts 2-7, numbers
femoropopliteal
TAV (centimeters per second) x the area of the vessel in cross section x 60 estimated
examinations
two
bar are the number
ropopliteal
from ye-
examination,
flows
velocities
velocities
systole, a mean
calculations initially gate to completely
vessel
In
-20
(in milli-
on film.
final peak calculates
132 8.7-327
Figure
grafts
measured.
A TAV is obtained from spectral analysis of a group of whole waveforms in series, with the cursor placed on the initial peak systole which
flow.
was not calculated; cluded. In Figures
cir-
on the basis of area. The
recorded
Flow
cmis#{149}c
Number of graft examinations by average PSV in the 253
that showed
the
PSV
Asr.g#{149}
Figure 2. categorized
at the previ-
calculated
157.5 26-816
S
x area
drew
(mL/min)
.18 .1-.6
(mL/min)
.5 .5
Area Flow
(cm2)
to the Last US Examination
Diameter
(cm)
Area
in
mean
=
Flow
S ubsequent
Diameter
4. A diameter of the graft at the previously defined levels was measured on a transverse image by means of a cursor placed by the technician. 5. After the computer calculated an area from
Diameter
because of occlusion, or incomplete data.
open Only
nine of 230 native femoral arteries (3.9%) had a PSV lower than that of the proximal graft. When the average of all the PSVs along the graft in the initial study was compared with the average at the last follow-up study 2 days to 27 months after imaging, the PSV fell in 168 femoral
arteries
(73.0%)
by
an
average
of
August
1991
4); femo1); femo= 4); femoropostibial in situ (n = 2); femoroperin situ (n = 1); femorodorsalis reversed saphenous (n = 1);
reversed
saphenous
ropopliteal ropopliteal
terior oneal pedis, femorodorsalis
a..
]
and (n
Area
Figure
4.
Flow
Average
In
area
flow
=
pedis
(in milliliters
per minute) in all grafts that had flow amination (n = 251). No area calculations were performed in four examinations; grafts were occluded.
Figure 6. Average diameter flow (in milliliters per minute) of all 252 grafts with flow at examination. In three examinations the di-
at exnine
ameter
was not calculated.
Nine
grafts
were
occluded.
patent
grafts.
Black
bars
examinations
=
showing patent grafts (,i = 231), diagonal bars = graft examinations performed before detection of occlusion in the next study ( n = 20). In one examination, no reading was performed because the artery was occluded at the initial study.
21 grafts,
(range,
femoral
2%-78%)
arteries
and
(27.0%)
rose
by
in 62
an
Figure 7. The distinguished basis of diameter tions of patent
Diameter
Flow
grafts
that
In
mi/mn
failed
from the patent
could
not
be
grafts on the
flow. Black bars = examinagrafts (n = 232), diagonal bars
= examinations of grafts performed before detection of occlusion in the next study (n = 20). In one examination, no reading was
performed because at the initial study.
the
artery
was
occluded
just
curred. whether revealed ure.
2 weeks
afterward,
tamed.
from
The
average
values
the last US examination
with subsequent shown in Table
erage
PSV
are
the numbers categorized
in the
grafts
an average the average
greater Graft
fined which
grafts
PSV
failure
in this
as graft
occlusion.
depicts
with
compared
group, confined
Volume
the
with
shows
that
study
de-
3, of PSV
eventually in the
this
group
particular
#{149} Number
but
0 to was
Figure
those
that
to any
180
cm/sec.
from
categories
flow
that
flow at the (77%) had
of 41-100
PSV ranged 141 cm/sec.
than
of by av-
examined
had flow. Most grafts with time of the US examination
failed patent
is not category.
2
the
nine
tamed
before
weeks
to 6 months
in four
3.
Figure 2 depicts graft examinations
for
Figure
of grafts
occlusions
in
the
before
the
grafts
occluded
subsequent
4 shows
at
these ob-
occlusion
2
later. the
average
area
examinations;
nine
grafts
were
occluded. In 168 examinations (67%), flow was 51- 200 mljmin. Figure 5 shows the relationship between grafts with flow that failed and the area flow of all the patent grafts; the group with impending graft occlusion could not be differentiated from the overall group on the basis of area flow alone. Figure
6 shows
the
average
diame-
occlusion
of oc-
In these cases, one wonders closer follow-up might have a cause for impending fail-
In patients
with
Increased
was
flow (in milliliters per minute) in all the grafts with flow at examination. No area calculations were performed
obtained
and
an
average
PSV
of
1-32 cm/sec. four graft examinations were done; all of the grafts had occlusion at the subsequent examination.
None
the time of their examination, studies were the last sonograms
of 101% (range, 1.5%-182%). Several of the grafts with a PSV that rose over time had developed marked stenosis. The average values obtained from the examinations of all the patent graft are shown in Table 2. Nine grafts were occluded at the time of the US examinations, one before any PSV values had been ob-
aver-
an occlusion after a norwas obtained, one pa-
sonography
Except
average
the
[38%]) and distal pseudoaneu(one graft [5%]). Five patients
flow
present
above
400
in 11 graft
of these
grafts
with occlusion, but ciated with increased 20%
7);
four other patients had an occlusion several months later; no indication potential failure was identified at
: the
=
in situ
or lower, and PSV was 45 cm/sec with no forward diastolic flow. Other contributing factors noted at sonography were stenosis (eight
tient
grafts with of all
(n
pedis
of these
(23.8%) had mal sonogram
between area flow
=
age PSV was 40 cm/sec in one graft the average
grafts rysm
Figure 5. Relationship flow that failed and
=
in situ
popliteal-dorsalis 1).
In seven
mI/mln
(n (n
Gore-Tex in situ (n
mL/min
examinations. was
associated
several were assodiastolic veloc-
ity, which was suspect for arteriovenous (AV) shunting. Findings in grafts with an average
PSV
of 40 cm/sec
or less are listed
in
Table 4. Only one occlusion occurred in grafts with an average PSV of 120 cm/sec or greater (Table 5). In the calculations of sensitivity,
true-positive cases were those grafts with occlusions that occurred within the criteria set forth herein. True-negative cases had no occlusion. Falsenegative cases were grafts with occlusion
that
occurred
outside
the
criteria,
and false-positive cases were grafts without occlusion but which would have been expected to have occlusion according to the previously defined criteria (average PSV of 40 cm/sec less) because they fell within the
ter flow (in milliliters per minute) of all the grafts with flow at examination; 158 grafts (63%) had flow of 51-
guidelines
200
sion occurred in 11 of these grafts (Table 6). In the 87 patients with an area flow of 1-100 mljmin, 10 occlusions occurred (Table 7).
mL/min.
Figure that failed from the diameter Failure grafts
(Figs
7, too, shows that the grafts could not be distinguished patent grafts on the basis of flow. (occlusion) occurred in 21 8, 9): femoropopliteal
of the
or
Ninety-two diameter flow
In the stenosis
the
criteria.
patients of 1-100
eight grafts had a focal
stenosis
from
that range
had
grafts
cmJsec;
with occlu-
failed, graft of PSV at
117 to 225 cm/sec;
Radiology
381
#{149}
.
Table
J.
:
-
;i/-1
4
Findings
in
Grafts
with
an
Average
PSV of4O cm/sec
or Less
,17’#{149}
Graft Occlusion Finding
at US
Present
Absent
Total
Positive
7
14
21
Negative
14
218
232
Total
21
232
253
Note-Numbers
tive predictive
value
are number ofgrafts. Sensitivity = 7/21 = 33.3%; specificity = 21&232 = 7121 = 33.3%; negative predictive value = 2181232 = 94%; accuracy
:!
/
fM
irivttir
I
PTG
94%; posi(7 + 218w
= =
233=89%. Figure MHz)
Table
nant
5
Findings
in
Grafts
an Average
with
PSV of Graft
at US
Finding
Total
were middle
proximal in one
were
six patients had been nation, months
in the
with normal
in one graft,
normal
6
247
21
232
253
at spectral
In
PSV exami-
analysis
(Figs 11, 12). In two patients, this was proved with angiography to have been caused by AV shunting and was associated with a normal PSV. Transient increase 13) after graft
in diastolic implantation
velocity had
4.76%, specificity = 227/232 = 227,247 = 91.9%, accuracy
=
value
that
has
the
communication
performed an average of 3 prior to diagnosis. One pa-
waveform
1/21
graft
in
range.
focal stenosis, at the prior
Total
5
graft,
distal
Absent 227
tient had had increased PSV at a focal stenosis in the same location on a prior visit. One patient had a focal increase in PSV with stenosis on his first visit. Seventeen patients showed focal elevation of the diastolic component of the
Occlusion
1
two grafts, and present in two or more sites in four grafts (Figs 10, 11). The PSVs proximal and distal to the stenosis
or Greater
cm/sec
(Fig spon-
97.8%,
=
seen
seen
at US. In one
tient with a femoropopliteal a pseudoaneurysm of the
patients
graft peroneal
distal
initially
to the had
had
26 of 85 patients (31%) and usually resolved completely over an average of 4.9 months (range, 0.5-12 months)
distal
All patients was increased 382
Radiology
#{149}
disease
beyond
the
graft.
in whom diastolic flow at US had an in situ
hindered
portion
US. Only
this group had a graft surgery with an open
inguinal with
area
that
US because
of the
one
was
not
graft
patient
infection wound
access
in
monly
fluid seen
collections initially
examined
was
were near
was
functioning
well.
(Figs
12c,
These
difficult.
the
corngraft
in
16).
fluid
collections
were rorna.
attributed to hematoma or seNo instance of fluid tracking along the graft was noted. The fluid collections caused no particular prob-
attributed
to edema
fluid,
after at the
In all other instances, both proximal and distal anastomoses were well seen, as were the interval areas of the graft (Figs 14, 15). Focal
is very faint on later, this graft DCG = distal calf graft.
immediately adjacent the compressing fluid. In nine patients, focal areas of decreased echogenicity in the soft tissues of the lower extremity were
an open
critical
that
saphe-
lems, except in one case in which the waveform of the graft was focally highly resistant just in the area of the
US anas-
wound
some
sonogram. pedis in situ
of these
taneously resolved in six patients on the next visit; follow-up was unavailable in one of these patients. In seven of nine other patients with diastolic elevation, transient increase in diastolic velocity was proved with angiography to be related to distal stenosis and AV shunting was not demonstrated (Fig 11). Three of these grafts had high focal PSV. Four grafts with abnormally low PSV and diastolic elevation had poor runoff from sites of
over
9. A longitudinal a femorodorsalis
third component of diastole the images.) Three months
became the US greater One pa-
artery that was noted at angiography but had not been examined with
it occurred
Figure strates
at angiog-
cases, the graft subsequently occluded. In another instance, scan was considered to show stenosis than the angiogram.
because tomosis. Five
months
flow
of 41 cm/sec. This remote distal ochad undergone graft occluded 4 thigh graft.
nous graft at the distal calf level, which has a PSV of 41 cm/sec but triphasic flow. (The
=
was
diminished
diastole and a PSV suggests significant disease. The patient angioplasties. The later. PTG = proximal
adjacent
=
raphy
very
(10 predomi-
to develop
the
not
flow with
sonogram shows
posi-
veins. Findings in 25 of the 36 angiograms (69.4%) in this group of patients were consistent with findings on US scans: normal flow, n = 5; low flow, n 3; stenosis, n = 12; AV shunting, n 2 (Fig 12); and graft occlusion, n = 3. Discordant findings (n = 11) occurred in two instances in which a stenosis
thigh
=
potential with
proximal
systolic
in early pattern clusive multiple
20
Note.-Nunibers are number ofgrafts. Sensitivity = live predictive value = 1/6 = 16.6%, negative predictive (1+ 227 = 228i253 = 90%.
stenoses
120
Present
Positive Negative
in the
8. A longitudinal in the proximal
which
fully
over
with
focal
spectral tions. Other
irregular patients
also time.
or disappearing
resolved Ten
increased analysis
unevent-
of 17 patients
diastolic also
had
flow fluid
findings
at US
included
rhythm
of the
heart
(Fig
17),
at
collecan
in seven
adenopathy
in six
patients, and a femoral anastomotic aneurysm in one patient. One patient had a retained venous valve that caused
increased
bulence
(Fig
PSV
18). One
and
patient
focal
tur-
had August
a 1991
Table 6 Findings
in Grafts
with
Diameter
Flow of 1-100
mL/min
Graft Occlusion at US
Finding
Present
Positive Negative
11 10
Total
21
Absent
Table
distal
anastomosis
high-grade was with days
with
stenosis
velocity.
At the
the dorsalis
pedis,
a
with a PSV of 163 cm/sec was successfully treated a sonogram obtained 6 that flow had returned to
present. This angioplasty; later showed
normal.
at US
superficial
thrombosis
seen
femoral
on the
vein
sonogram.
The
grafts
merous
tect
were
areas
local
focal either
examined
along
their
abnormalities.
in nupath
to de-
We detected
areas of high and low velocity, systolic or diastolic, or both,
and focal fluid collections. Except in five cases of open wounds overlying some areas of the graft, both proximal and distal anastomoses were well seen. Grafts that had an anastomosis with the dorsalis pedis artery could be just as easily identified as those at the tnfurcation beyond the popliteal artery. The typical normal peripheral yessel has a triphasic waveform at spectral analysis (5). The first systolic cornponent
90#{176} angle. Volume
rises
above
The 180
second Number
#{149}
the
baseline
diastolic 2
at a
corn-
value
specificity 150/160
=
150t231
=
64.9%,
=
93.8%, accuracy
=
=
mllmin
Absent
Total
77
Positive Negative
10 11
154
87 165
Total
21
231
252
= 10/21 predictive
=
47.6%, value
=
specificity 231/252
= =
1541231 = 66.6%, 91.7%, accuracy =
the next examination. However, in this group of 21 patients with a PSV of 40 cm/sec or less, seven of whom
ized
subsequently
as a high-resistance
arteries were
phasic
in diastole. has
not
triphasic.
triphasic, a dip
Biphasic
been
pattern.
were
without explained
regional
hypotension
outflow
resistance
but below
mono-
and
sion
in a graft
as being
related
These to
diminished distal
than
tnphasic
at spectral
arterial
analy-
sis, with some diastolic component above the baseline present but not elevated (Fig 14). Peripheral arteries typically
have
high-resistance
flow
with little diastole, comparable to that of the external carotid artery. Diastolic elevation, when it occurred, was not subtle and was categorized as a distinct abnormality associated with spectral
broadening
(Fig
had
occlusion
by
next visit, two died without and four others eventually
baseline
flow
of the
Most
Most
(1). A normal graft was considered to have an average PSV above 45 cm/sec but not above 115 cm/sec and was more likely biphasic or monophasic rather
DISCUSSION
predictive
Present
bed
concomitant
252
52.3%,
=
ponent descends below the baseline, followed by a final diastolic rise just above the baseline. This is character-
grafts diastolic
11/21
Graft Occlusion
femoral increased
=
with an Area Flow of 1-100
in 87 Patients
Note-Numbers are number of grafts. Sensitivity positive predictive value = 10/87 = 113%, negative (10 + 153 = 163t251 = 64.9%.
and
92 160
7
Findings
Finding
sec
81 150 231
Note-Numbers are number of grafts. Sensitivity positive predictive value = 11192 = 12.0%, negative (150 + 11 = 161)t252 = 63.9%. Figure 10. Contrast the normal spectral analysis of Figure 1 with this spectral analysis of a stenosis of the distal anastomosis in a similar graft where the PSV is 184 cm/sec, diastolic velocity is increased, and spectral broadening is prominent. G TO DP ANAS = graft to dorsalis pedis anastomosis.
Total
12).
Abnormal grafts had several different waveforms. Low blood velocity of 32 cm/sec or less was the single factor always associated with graft failure (Fig 8). Bandyk et al (2) initially suggested that low PSV of less than 40 cm/sec. and absence of diastolic flow were associated with graft occlusion, but in our study use of a PSV of only 40 cm/sec or less had a sensitivity of 33% for prediction of occlusion by
rates
(although
not
results are
Diastolic
by
indicate
the
that
the
follow-up had occlunext
visit).
low
flow
ominous.
flow
variations
cators of variance sence of diastolic sign of impending
were
mdi-
from the norm. Abflow was often a graft failure. Eleva-
tion of diastole throughout the graft in the first month after implantation suggested reactive hyperemia and was transient. Blackshear et al (4) showed that, by inflating a thigh cuff above systolic pressure for 3 minutes
and then releasing it, flow in the fernoral artery changed from a triphasic waveform to a transient monophasic one in which reverse diastolic flow disappeared
and
forward
diastolic
flow markedly increased, a response attributed to reactive hyperemia. Just as a transient stolic flow occurs
mesenteric stolic flow temporarily
increase in diain the superior
artery
after
in the after
extremity exercise.
eating,
diaincreases
Diastolic flow increases in AV shunting and also at sites of severe critical stenosis, either at focal areas along
the
graft
or as diffuse
or
increase
in diastolic flow the length of the graft in response to poor runoff (6-8). Diffuse increase in diastolic flow Radiology
383
#{149}
-
*
PAl
VEL
AVG
VEL PNGE
VTL
96
=
CM’
S
:-
.
1 M
iIl1
F1
KJG
a.
Figure 13. One week after graft placement, a longitudinal sonogram of a femorodorsalis pedis in situ saphenous graft at the calf shows normal PSV and increased diastolic velocity suggestive of AV shunting or poor runoff from distal occlusive disease, but these changes disappeared within the month, and
b.
the graft These
developed
findings
hyperemia.
;
normal
are
DCG
=
triphasic
consistent with distal calf graft.
flow. transient
w vtL,lcr
-‘,
-
-wcN 47
-=-
-;
CN”SIC
INVERTED
AFT
Figure 12. (a) Angiogram shows a taintly opacified femorodistal posterior tibial graft, which is outlined below the knee by surgical clips. The graft is shunting significant flow to the femoral venous system. (b) Sonogram of the same patient shows significant AV shunting with a normal PSV but increased diastolic velocity at the knee. Fluid is present anterior to the graft. In b and c, KIG = popliteal fossa. (c) On transverse section of the graft at the knee, the presence of a discrete fluid collection (arrowhead) is corroborated. Perimeter of the graft (A) is outlined. (d) At a site more distal than that in c, the spectral analysis of the graft in the distal calf has a more normal pattern of high resistance. DCG = distal calf graft.
throughout the graft may be transient because of reactive hyperemia in the first month after graft implantation,
but
after
1 month
such
a finding
sug-
gests vasodilatation due to arterial insufficiency from poor runoff. Focal high diastolic flow in associa-
tion
with
high
Normal
PSVs
PSVs
and
suggests focal
stenosis.
high
diastolic
AV
shunting.
velocity suggest AV shunting (two patients) or stenosis (two patients). In our study, in two cases of documented AV shunting, forward dias-
This
tolic flow was markedly but PSV was in large
increased, measure within
45
the
Estimated
range
of normal.
in milliliters
per
creased.
Prior
postocciusive
tive
and
hyperemia
increases
that than
in diastolic
with
ties greater creased
shown
than
Radiology
flow
that
findings
which showed (with a systolic graft associated cm/sec occlusive also found
in a prior
study
that monophasic waveform only) with a PSV less
flow of the than
was a predictor of a remote lesion (1,2). Bandyk et al (1) that a decrease of greater
graft shows monophasic PSV at the anastomosis
component-grafts failed.
One
that
patient
had
a normal
In one
veloci-
had suggested
distal
occlusive
lesion
in-
longer
curred graft, and This
patient
followed
than 2 years, consistently
with 25
cm/sec.
pattern
focal
readings but
over
up for
a low PSV throughout
oral graft. The stenoses ous
graft
that
bypasses
occur tend
in saphento develop
within 2 years, half of them within the first 8 months after graft placement (1). When stenoses were confirmed
on
angiograms,
the
with
time
as low
as 22
triphasic flow. was associated
range
of
PSV from
stolic
octhe
average
PSV but no diastolic flow throughout the graft. At angiography, he had a severe stenosis of the distal aortofem-
stenosis on US scans in this was 117-225 cm/sec. In two pahowever, focal graft stenoses
confirmed with angiography seen at US as areas of focal
a more
and
subsequently
cated
(1).
flow (ANAS).
with a better prognosis than those of grafts with low PSV and no diastolic
reacvelocity
400 mlJrnin
confirms
saphenous a normal
sonogram (10 pedis (DP) in situ
study tients,
mean velocity rather PSV (6). Several flow
grafts
14. A longitudinal of a femorodorsalis
Figure
MHz)
than 30 cm/sec in PSV compared with initial values postoperatively mdi-
substantial
flow
of these
TO OP
both
in-
with
estimated
diastolic
#{149}
was
postexercise
has
increase the the maximum
patients
384
minute
experience
flow
None
with high flow had occlusion. A low average PSV (32 cm/sec or lower) and absence of diastolic flow were always associated with graft failure in this study and were related to high outflow resistance; but an isolated, single reading of focal low PSV somewhere along the graft was not.
Al*IS
stenoses
flow in
grafts, patients
which (9),
Mild,
moderate,
changes
were
high diawith a normal PSV. Graft in situ saphenous vein occur in 21%-33% may be asymptomatic.
in the
or severe graft
lumen
of
stenotic can
August
be
1991
-
__;
-------
* -
-
-
ANAS
G
OP
A
Figure
15.
dorsalis (ANAS)
(DP) graft (G) anastomosis to indicate how well sonography can these small vessels. Arrowhead mdi-
define cates
Transverse
interface
pedis
scan
shows
a normal
pedis
with A
is circled.
bones graft.
=
of foot.
Dorsalis
.-
b.
a.
Figure 16. (a) Transverse sonogram shows a large fluid collection (arrowhead) anterior to the graft (A). CC = calf graft. (b) Four months after a was obtained, this fluid collection (arrowhead) has significantly decreased in size. A = graft, DCG = distal calf graft.
disease
(5,10).
that
Bandyk
the typical
sociated
et al reported
graft
with
stenosis
a decrease
sessment
was
in PSV
ascorn-
pared with that at the initial examinalion (1); Gngg et al noted that no graft with a severe stenosis had flow with a mean velocity greater than 50 cm/sec (10).
17. the irregular
Longitudinal sonogram shows rhythm of pulsus alternans on
the
analysis.
Figure
spectral
astolic MTG
=
Note
the
absence
flow. This graft subsequently middle thigh graft.
of di-
failed.
:r
cation
occurs
(12).
Of the
stolic
flow
-..
with
(1,10).
Just
a
as
standards have been suggested for reports of lower-extremity ischemia in the vascular surgery literature (11), criteria have been developed for the assessment
severe Volume
of mild,
peripheral 180
moderate,
arterial Number
#{149}
2
and
occlusive
in about
17 grafts pattern
10%
with at US,
of patients
a high
dia-
six eventu-
(14). duplex
of AV
saphenous ously
pulsed
close
in an in situ
graft
(15).
may
Kotval et al docuDoppler, an in-
shunting
vein
closed
with
that
Area
spontane-
and
diameter
flow calculations were not helpful in prediction of graft failure, but marked increases Others
suggested AV have reported
shunting. on both
frau-
matic and iatrogenic fernoral AV fistulas diagnosed with both duplex Doppler and color Doppler (16,17). While a focal high PSV and in creased diastolic velocity focally suggest graft stenosis, this is sometimes a transient pattern in the first weeks after graft placement.
middle-calf PSV of 127 cm/sec. but angiography only showed poor runoff. However, 4 months later these stenoses showed progression at US,
been described by Bandyk et al and attributed to revascularization hyperemia with associated vasodilatation and increased skin temperature (9) (Fig 13). The initial low resistant flow in in situ grafts suggests that flow is shunted until a new pathway of run-
plasty, cessful,
sonography
of all
spontaneously mented, with stance
graft
AV fistulas
Transient locity after
angiography confirmed stenoses were treated which and
them. with angio-
was only partially occlusion occurred.
all increased PSV, an indicator stenosis, leads to graft failure. patient
seen
in 73%
(13).
ally had occlusion
and The
:#{231}
Figure 18. Longitudinal sonogram shows proximal (PROK) in situ saphenous vein graft with a retained valve (arrowheads).
PSV
(35%). One patient was believed to have two graft stenoses at US because of a proximal PSV of 165 cm/sec and a
,
..s
study,
grafts, both normal and abnormal, decreased over time, and stenoses presented as areas of (a) focal high PSV, (b) normal or low PSV with increased diastolic flow, or (c) low PSV with absence of diastolic flow. A retained competent valve cusp was a source of marked focal turbulent flow and high PSV. This compli-
GRAFT*
A
In our
of the
Doppler
has
had
a proximal
sucNot
of One stenosis
with a PSV of 142 cm/sec that has not changed for over 1 year. Although a normal PSV and increased diastolic velocity suggest AV shunting, this pattern was more likely to be related to remote distal occlusive disease. This pattern occurred in in situ grafts only. AV shunting can also be detected
during
intraoperative
as-
off
has
increase in diastolic graft implantation
been
established.
One
yehas
patient
had
an initial high PSV of 274 cm/sec at US, along with high diastolic flow, but only minimal stenosis at angiography, and the PSV eventually returned to normal with no treatment, an mdication that this may have been fransient hyperemia. Pseudoaneurysm
cation
of graft
anastornotic holosystolic
and
below
sociated
raphy motic
is another
placement, aneurysms flow (ie,
the baseline with
defined aneurysm
turbulence
the
and
flow
compli-
these
typically both
have above
in systole) (18).
as-
Sonog-
proximal anastoof the femoral artery, Radiology
385
#{149}
but another aneurysm, distal to the graft, that was seen at angiography was not examined with US. The experience of surveillance with duplex Doppler of the vascular graft placed for peripheral artery disease is limited. Further experience and further refinements, such as the addition of color, promise an even more helpful noninvasive way to monitor the graft and have the potential to increase the sensitivity of predicting graft failure, allowing for intervention before occlusion develops. #{149} References 1.
Bandyk
mal human tents. 5.
6.
7.
8.
CR, Moldenhauer P. of vein graft steno-
et al. Hemodynamics sis. J Vasc Surg 1988; 8:688-695. 2.
Bandyk
DF, Cato
RF, Towne
JB.
A low
flow velocity predicts failure of femoropopliteal and femorotibial bypass grafts. Surgery 1985; 98:799-809. 3.
PolakJF,
nickJA,
4.
386
Donaldson MC, Dobkin CR, O’Leary CH. Early detection
saphenous
vein arterial
sis by color
assisted
bypass
10.
Manof
graft steno-
duplex sonography: AJR 1990; 154:857-861.
a
prospective study. Blackshear WM Jr, Phiffips DJ, Strandness DE Jr. Pulsed Doppler assessment of nor-
Radiology
#{149}
12.
Monata arterial
CL, Strandness duplex scanning.
651. Maher
F, Coen
DE. JCU
Peripheral 1987; 15:645-
11.
Nicholls Phillips
Cave-Bigley DJ, Ackroyd MA, Campbell H, Parry EW, Harris PL Complications associated with in situ vein grafts for femoropopliteal bypass. BrJ Surg 1984; 71:211221.
L,Johansen
KB, Bernstein
13.
UF, Fronek A. Post-ocdusion and postexercise flow velocity and ankle pressures in normals and marathon runners. Angiology 1976; 27:721-729.
9. DF, Seabrook
femoral artery velocity pat-
J Surg Res 1979; 27:73-83.
SC, Kohier TR, Martin RL, Neff R, DJ, Strandness DEJr. Diastolic flow as a predictor of arterial stenosis. Vasc Surg 1986; 3:498-501. Kohler TR, Nicholls SC, Zierler RE, Beach KW, Schubert PJ, Strandness DEJr. Assessment of pressure gradient by Doppler ultrasound: experimental and clinical observations. J Vasc Surg 1987; 6:460-469. Bandyk DF, Kaebnick HW, Stewart CS, Towne JB. Durability of the in situ saphenous vein arterial bypass: a comparison of primary and secondary patency. J Vasc Surg 1989; 5:256-268. Crigg MJ, Nicolaides AN, Wolfe JHN. Detection and grading of femorodistal vein graft stenoses: duplex velocity measurements compared with angiography. J Vasc Surg 1988; 8:661-666. Rutherford RB, Chairman, et al. Suggested standards for reports dealing with lower extremity ischemia. Ad Hoc Committee on Reporting Standards, Society for Vascular Surgery/North American Chapter, International Society for Cardiovascular Surgery. J Vasc Surg 1986; 4:80-94.
14.
Bandyk DF,Jorgensen traoperative assessment
RA, Towne JB. Inof in situ saphenous vein arterial grafts using pulsed Doppier spectral analysis. Arch Surg 1986; 121: 292-299. Bartlett ST. Killewich LA, Fisher REW. Duplex imaging of in situ saphenous vein bypass grafts and late failure reduction. AmJ
15.
tary 16.
17.
Surg
1988;
156:484-487.
Kotval PS, Shah PM, Barakat K. Doppler determination of in situ vein bypass tribufistula
and
its spontaneous
WD, Schwartz RA, Goldberg BB. latrogenic femoral arteriovenous fistula: diagnosis with color Doppler imaging. Radiology 1989; 170:749-752. Helvie MA, RubinJ. Evaluation of traumatic groin arteriovenous fistulas with du-
plex Doppler sonography. Med 1989; 8:21-24. 18.
closure
(letter). AJR 1989; 153:192-193. Igidbashian VN, Mitchell DC, Middleton
PolakJF,
Donaldson
MC,
J Ultrasound Whittemore
AD,
MannickJA, O’Leary DH. Pulsatile masses surrounding vascular prostheses: real-time US color flow imaging. Radiology 1989; 170:363-366.
August1991