From the Western Vascular Society

The natural history of duplex-detected stenosis

after femoropopliteal endovascular therapy
suggests questionable clinical utility of routine
duplex surveillance
Trung D. Bui, MD, Joseph L. Mills, Sr, MD, Daniel M. Ihnat, MD, Angelika C. Gruessner, PhD,
Kaoru R. Goshima, MD, and John D. Hughes, MD, Tucson, Ariz
Objective: Duplex ultrasound (DU) surveillance (DUS) criteria for vein graft stenosis and thresholds for reintervention are
well established. The natural history of DU-detected stenosis and the threshold criteria for reintervention in patients
undergoing endovascular therapy (EVT) of the femoropopliteal system have yet to be determined. We report an analysis
of routine DUS after infrainguinal EVT.
Methods: Consecutive patients undergoing EVT of the superficial femoral artery (SFA) or popliteal artery were prospectively enrolled in a DUS protocol (<1 week after intervention, then at 3, 6, and 12 months thereafter). Peak systolic
velocity (PSV) and velocity ratio (Vr) were used to categorize the treated artery: normal was PSV <200 cm/s and Vr <2,
moderate stenosis was PSV 200-300 cm/s or Vr 2-3, and severe stenosis was PSV >300 cm/s or Vr >3.
Reinterventions were generally performed for persistent or recurrent symptoms, allowing us to analyze the natural history
of DU-detected lesions and to perform sensitivity and specificity analysis for DUS criteria predictive of failure.
Results: Ninety-four limbs (85 patients) underwent EVT for SFA-popliteal disease and were prospectively enrolled in a
DUS protocol. The initial scans were normal in 61 limbs (65%), and serial DU results remained normal in 38 (62%). In
17 limbs (28%), progressive stenoses were detected during surveillance. The rate of thrombosis in this subgroup was 10%.
Moderate stenoses were detected in 28 (30%) limbs at initial scans; of these, 39% resolved or stabilized, 47% progressed
to severe, and occlusions developed in 14%. Five (5%) limbs harbored severe stenoses on initial scans, and 80% of lesions
resolved or stabilized. Progression to occlusion occurred in one limb (20%). The last DUS showed 25 limbs harbored
severe stenoses; of these, 13 (52%) were in symptomatic patients and thus required reintervention regardless of DU
findings. Eleven limbs (11%) eventually occluded. Sensitivity and specificity of DUS to predict occlusion were 88% and
60%, respectively.
Conclusions: DUS does not reliably predict arterial occlusion after EVT. Stenosis after EVT appears to have a different
natural history than restenosis after vein graft bypass. EVT patients are more likely to have severe stenosis when they
present with recurrent symptoms, in contrast to vein graft patients, who commonly have occluded grafts when they
present with recurrent symptoms. The potential impact of routine DU-directed reintervention in patients after EVT is
questionable. The natural history of DU-detected stenosis after femoropopliteal endovascular therapy suggests questionable clinical utility of routine DUS. ( J Vasc Surg 2012;55:346-52.)

Duplex ultrasound (DU) surveillance (DUS) is widely

used after infrainguinal autogenous vein bypass to detect
hemodynamically significant lesions and allow timely repair
before graft occlusion.1-6 With DU-directed reintervention, the 5-year assisted-primary patency rate approaches
80%, nearly identical to the primary patency of a graft that
From the Division of Vascular and Endovascular Surgery, University of
Arizona Health Sciences Center.
Competition of interest: none.
Presented at the Twenty-fourth Annual Meeting of the Western Vascular
Society, Tucson, Ariz, September 19-22, 2009.
Reprint requests: Joseph L. Mills, Sr, MD, Professor of Surgery, Chief,
Division of Vascular and Endovascular Surgery, University of Arizona
Health Sciences Center, 1501 N Campbell Ave, P.O. Box 245072, Rm
4404, Tucson, AZ 85724 (e-mail: jmills@u.arizona.edu).
The editors and reviewers of this article have no relevant financial
relationships to disclose per the JVS policy that requires reviewers to
decline review of any manuscript for which they may have a competition of interest.
0741-5214/$36.00
Copyright 2012 by the Society for Vascular Surgery.
doi:10.1016/j.jvs.2011.08.010

346

never developed stenosis.7 Such patency results are superior

to the reported 1-year secondary patency of 20% to 35%
after treatment of thrombosed vein grafts.8
The application of endovascular therapy (EVT) to
lower extremity arterial disease has significantly expanded.
The 2007 TransAtlantic Inter-Society Consensus (TASC)
II document favors EVT of focal iliac and femoropopliteal
occlusive disease (TASC A and B lesions) over open repair.9
EVT is increasingly being used for TASC C and even TASC
D lesions. Restenosis requiring reintervention at 1 year,
however, has been reported to be as high as 30% to 60%
after percutaneous transluminal angioplasty (PTA), with
or without stenting, of the superficial femoral artery
(SFA).10-12 The natural history of DU-detected stenosis
after femoropopliteal EVT is unclear; the literature regarding the clinical utility of DUS after EVT is scarce and
conflicting. Some investigators have suggested that DUS
after EVT has no role,13,14 whereas others have reported
that residual stenosis on the initial DUS is a major risk
factor for subsequent EVT failure.15,16

JOURNAL OF VASCULAR SURGERY

Volume 55, Number 2

Bui et al 347

Fig 1. Duplex ultrasound surveillance algorithm for limbs undergoing endovascular treatment.

We routinely enroll patients into a DUS protocol after

both open bypass and EVT. Our reintervention criteria for
DU-detected lesions after infrainguinal bypass follow established parameters.8,17,18 In patients with abnormal DU
findings after EVT, however, threshold criteria for reintervention remain unclear because the natural history of DUdetected restenosis in the SFA-popliteal arterial segment is
unknown. We retrospectively examined the frequency and
natural history of DU-detected stenoses after EVT to determine its clinical utility.
METHODS
This retrospective study was approved by the Institutional Review Board.
Patient population. From September 2002 to July
2008, 94 limbs in 85 patients undergoing EVT for
isolated, chronic SFA-popliteal disease were prospectively enrolled in a DUS protocol by the Vascular Surgery Division at University Medical Center, Tucson. The
primary purpose of this study was to determine the natural
history of DU-detected lesions after SFA-popliteal EVT.
Study inclusion criteria required the performance of an
early postprocedural scan (1 week of the initial intervention) and at least two additional DU follow-up studies to monitor for progression and detect development of
de novo lesions.
Patient demographics, duration of follow-up, and standard risk factors, including tobacco smoking, diabetes mellitus, hypertension, coronary artery disease, hyperlipidemia
and renal failure were tabulated for each patient. Only
initially successful interventions were included. Excluded
were limbs in which EVT failed before the first DU study
could be performed or with inadequate surveillance (minimum of 3 postintervention studies).
Surveillance protocol and categories of stenosis.
The surveillance protocol included a focused history, peripheral pulse examination, and measurement of anklebrachial indexes and toe pressures. DU scans interrogated

the entire femoropopliteal segment and the adjacent inflow

and outflow arteries, using a 5.0- or 7.5-MHz linear-array
probe (HDI 3000; Advanced Technology Laboratories,
Bothell, Wash). Representative center-stream velocity spectra were recorded at multiple segments at a corrected
Doppler angle of 60.
Stenosis was identified by a flow disturbance with a
mosaic color pattern. Stenosis severity was graded by measuring the peak systolic velocity (PSV) and velocity ratio
(Vr), which is the ratio of the PSV at the stenotic site/PSV
of the adjacent, normal proximal arterial segment. Stenoses
were stratified into three categories by DUS findings: normal was a PSV 200 cm/s and Vr 2); moderate was
PSV 200-300 cm/s or Vr 2-3, and severe was PSV
300 cm/s or Vr 3.3,6,8
All included patients underwent initial DU scans within
the first week after intervention (Fig 1). Normal limbs were
then evaluated every 3 to 6 months during the first 12 to 18
months and annually thereafter. Limbs with moderate or
severe lesions were re-examined more frequently.
Clinical EVT failure was defined as limb occlusion or
recurrent signs or symptoms (claudication, rest pain or
nonhealing/recurrence of ischemic ulcers) requiring reintervention in the treated limbs. Limb occlusion was
defined as occlusion of the site or segment of femoropopliteal arterial intervention. Our general policy was to
reintervene only for recurrent symptoms, regardless of
DU findings. In two highly selected asymptomatic patients in whom DU-detected severe stenoses, selective
reintervention was performed at the surgeons discretion.
Statistical analysis. Data were analyzed using the
SPSS software (SPSS Inc, Chicago, Ill). All values are
presented as means standard error. Receiver operating
characteristic (ROC) curve analysis was used to determine the sensitivity and specificity of DUS to predict
limb occlusion. Predictive values of PSV and Vr were also
calculated.

JOURNAL OF VASCULAR SURGERY

February 2012

348 Bui et al

Table I. Comparison of frequency of normal scans and outcomes according to TransAtlantic Inter-Society Consensus
(TASC) classification
Limbs with normal DUS
TASCa

No.

Initial scan

Last scan

Occluded limbs

Major amputation

A and B
C and D

54
40

35 (65)
27 (68)

25 (46)
16 (40)

2 (4)
9 (23)

None
3 (7.5)

DUS, Duplex ultrasound surveillance.

a
Data are presented as number (%).

Table II. Comparison of frequency of normal scans and outcomes according to types of intervention
Limbs with normal DUS
a

Intervention

No.

Initial scan

Last scan

Occluded limbs

Major amputation

PTA
Atherectomy
Stent

44
23
27

33 (75)
11 (48)
17 (63)

16 (36)
7 (30)
15 (56)

3 (6.8)
2 (8.7)
6 (22)

1 (2.3)
1 (4.3)
1 (3.7)

DUS, Duplex ultrasound surveillance; PTA, percutaneous transluminal angioplasty.

a
Data are presented as number (%).

RESULTS
The study included 85 patients, 57 men (67%) and 28
women (32%), with a mean age of 70 3.1 years (range,
48-86 years), who underwent interventions for 94 limbs.
Limbs had a mean of four postprocedural DU studies
(range, 3-9). Mean follow-up was 19 months (range, 6-67
months). During the study period, 29 limbs (30%) underwent 38 interventions, comprising EVT in 27, open bypass
in 8 (7.9%), and major limb amputation in 3 (3.2%).
Standard risk factors were prevalent, including hypertension (81%), tobacco smoking (67%), hyperlipidemia
(55%), diabetes mellitus (47%), coronary artery disease
(39%), and end-stage renal disease (15%). The indications
for EVT using the Fontaine classification were critical limb
ischemia in 55 patients (65%), comprising tissue loss in 35
and rest pain in 20; and disabling claudication in 30 (35%).
Lesions were graded according to the TASC II classification (Table I): with 54 limbs (57%) at TASC A/B and 40
(43%) at TASC C/D.
The initial interventions in the 94 limbs were balloon
angioplasty alone in 44, angioplasty plus stenting in 27, and
atherectomy in 23. Table I and Table II compare frequency
of normal initial and last DU scans and outcomes (limb
occlusion and major limb amputation) according to TASC
classification and intervention type.
Outcomes were analyzed (Table III) by whether the
initial DU study showed no or minimal residual stenosis
(normal) or moderate or severe residual stenosis (abnormal).
Results in limbs with normal initial DU studies.
Initial DU studies were normal (PSV 200 and Vr 2) in
61 limbs (66%) in 56 patients (Fig 2), and 38 limbs (62%)
remained serially normal at a mean of 22 months (range,
6-68 months). Stenoses subsequently developed in 17

Table III. Outcomes in 91 limbs based on most recent

duplex ultrasound scan findings
Patent
Stenosis
category

Without
reintervention

With
reintervention

Occlusion

Total

Normal
Moderate
Severe

44
10
9

0
5
15a

1
9
1

45
24
25

Reinterventions because of symptomatic restenosis (n 13); reinterventions for severe asymptomatic restenosis on duplex surveillance (n 2).

limbs (28%), and 6 (10%) occluded before a secondary

intervention.
Of those 17 initially normal limbs that developed restenosis during a mean follow-up of 25 months (range, 6-50
months), 8 moderate stenoses remained stable and 9 stenotic lesions progressed to severe. Of the eight limbs with
moderate stenoses, three remained asymptomatic and required no further treatment. The other five limbs eventually became symptomatic (4 with nonhealing wounds, 1
with a new ischemic ulcer) and all underwent successful
repeat EVT (stenting in 3, atherectomy in 2). Four of these
secondary interventions remained patent and one occluded. The latter patient presented with an ischemic ulcer
after initial PTA of a TASC C SFA lesion. Repeat PTA and
stenting was performed; however, the treated segment
reoccluded 1 month, and a below knee amputation was
required because of severe concomitant tibial artery occlusive disease.
Severe stenoses were detected in nine limbs, of which
six required seven reinterventions (3 PTAs, 2 stents, 2
atherectomies) because of recurrent symptoms. Two limbs

JOURNAL OF VASCULAR SURGERY

Volume 55, Number 2

Bui et al 349

Stable
7 limbs
(25%)

Resolution
4 limbs (14%)

ABNORMAL
33/94 limbs
(35%)

Severe
5 limbs (5%)

Moderate
28 limbs (30%)

Severe
13 limbs (47%)

Occlusion
4 limbs (14%)

Fig 3. Duplex ultrasound surveillance data in 33 limbs with abnormal initial scans: results of serial surveillance for moderate
duplex-detected stenoses.
Fig 2. Duplex ultrasound surveillance data in 61 limbs with normal initial scans.

with severe stenosis remained asymptomatic and required

no further therapy. One patient with severe, asymptomatic
restenosis underwent repeat atherectomy due to surgeon
preference.
Despite a normal initial DU study, occlusion of the
treated segment occurred in six limbs (10%). The mean
time until thrombosis was 17 months (range, 6.6-48.5
months) after the index EVT procedure. DU scans had not
detected a severe stenosis before occlusion in any of these
patients. The DU scans in one patient were normal
throughout surveillance until thrombosis of the treated
segment occurred. This patient was asymptomatic and required no further treatment. In five cases, a moderate
stenosis had been identified and was being closely monitored with DUS before the occurrence of thrombosis.
These patients were symptomatic with ischemic rest pain or
nonhealing wounds, or both, and required reintervention
because of clinical indications. Three patients underwent
femorocrural bypasses, all of which remain patent. The
other two limbs underwent successful repeat EVT (atherectomy and PTA).
Results in limbs with moderate or severe stenosis at
the initial DU studies. Moderate lesions (PSV 200300 cm/s; Vr 2-3) were identified on the initial DUS in
28 limbs (29%) in 25 patients (Fig 3). The mean PSV at the
time of detection was 246 11.8 cm/s and the Vr was 2.2
0.1. The mean follow-up in patients with moderate lesions
was 11.7 1.7 months (range, 3-31 months). Intensive
DUS of these patients documented the following natural
history of moderate residual stenosis: in 11 limbs (39%), the
lesions stabilized (n 7) or completely resolved (n 4); in
13 cases (47%), the stenoses progressed from moderate to
severe; and 4 segments (14%) occluded.
In the four instances in which complete lesion resolution was observed, the initial mean PSV was 247 17.8
cm/s (range, 204-290 cm/s), and the Vr was 2.3 0.3
(range, 2.0-2.7). In the eight cases in which the stenoses
stabilized, the mean initial PSV was 229 4.7 (range,

220-250 cm/s) and the Vr was 2.4 0.5 (range, 1.6-2.7).

All limbs with resolved or stabilized lesions remain patent
without reintervention. There were no significant differences in initial PSV, Vr, or time of lesion onset for moderate
stenoses that regressed or stabilized compared with those
that progressed in severity or subsequently occluded.
During serial surveillance, 13 limbs with moderate
stenoses on initial DUS progressed to severe stenosis. The
mean initial PSV was 221 9.5 (range 184-278 cm/s),
and the mean Vr was 2.3 0.1 (range, 1.5-2.6). Six of
these 13 limbs remained asymptomatic at a mean follow-up
of 13 2.1 months and required no further treatment.
Atherectomy of the lesion in one asymptomatic limb was
done because of surgeon preference. The other six limbs
eventually became symptomatic (disabling claudication,
rest pain, and nonhealing wounds) and required reintervention (atherectomy, 4; PTA, 2). Four atherectomies for
restenosis were performed in four different limbs, only one
of which remained patent. The three failed atherectomies
presented with recurrent critical limb ischemia and eventually required open bypass. PTA was performed in the remaining two symptomatic limbs, only one of which remained patent. The other limb had a persistent nonhealing
ulcer necessitating a below-knee amputation because of the
lack of a bypass target.
Treated segments in four limbs with moderate stenoses
on initial scans occluded during surveillance. A moderate,
nonprogressive stenosis was being monitored in all four
limbs before the onset of thrombosis. The mean interval
before occlusion occurred was 6 1.8 months. One limb
remained asymptomatic and required no further treatment.
The other three limbs presented with recurrent rest pain
and nonhealing ulcers, and two were treated by femorocrural bypass and one with thrombolysis and stenting. Both
bypass grafts were patent at the last assessment, but the
patient who underwent repeat EVT required below-knee
amputation 3 months after stent thrombosis.
Only five limbs (5%) had severe residual stenoses
(PSV 300 or Vr 3) on the initial DUS (Fig 4). Mean
initial PSV was 342 14.8 cm/s (range, 301-444 cm/s)
and the mean Vr was 2.7 0.3 (range, 1.5-3.5). Four of

JOURNAL OF VASCULAR SURGERY

February 2012

350 Bui et al

DISCUSSION

Fig 4. Duplex surveillance data and outcomes in five limbs with

severe stenoses detected on initial scans.

these severe lesions did not progress; two of these treated

segments were patent at the last assessment, despite severe
stenosis, and the other two initially severe lesions completely regressed. One limb with a severe stenosis occluded
at 7 months after EVT. This patient presented with a
nonhealing wound and was treated by thrombolysis and
balloon angioplasty, with persistent patency at 16 months.
Of the two limbs with persistent severe stenoses, one was
asymptomatic and did not require treatment, and the
other eventually required reintervention for a new ischemic ulcer 3 months after EVT, with ongoing patency at
1 year (Table IV).
Natural history of severe stenosis after SFApopliteal EVT. DUS findings in 94 limbs, at the most
recent scan or at the last scan before thrombosis, were
categorized as normal in 45, moderate in 24, and severe in
25 (Table III). Severe stenoses developed in 25 limbs, of
which 13 (52%) presented with persistent or recurrent
ischemic symptoms. The symptoms would have necessitated reintervention. Nine severe stenoses (36%) were
found in asymptomatic limbs, and thus far, they have not
required reintervention at a mean follow-up of 10 1.2
months. In one instance (4%), a severe stenosis, identified
on the initial DUS, led to subsequent thrombosis 2 months
after detection.
Given that severe stenosis developed in 25 of 94 (27%)
EVT-treated limbs at some point during serial DUS, we
examined whether the presence of severe stenosis predicted
the onset of thrombosis after EVT. A total of 11 limbs
(11%) occluded, and the severity of DUS-detected stenosis
before the onset of thrombosis was normal in 1, moderate
in 9, and severe in 1. If the presence of a severe DUdetected stenosis had been used as the criterion for reintervention, only 1 of the 25 limbs (4%) would have benefited
from DUS. In fact, before thrombosis, nine limbs (82%)
exhibited only moderate stenoses. Fig 5 illustrates the
values of PSV (219 cm/s) and Vr (2.1) that best predict
limb occlusion. The calculated sensitivity and specificity of
DUS were 88% and 60%, respectively.

The natural history of vein graft stenosis has been well

studied. DUS has been widely adopted and considered to
improve graft patency and limb salvage by 10% to
15%.4,17,19-21 DUS after vein bypass is likely cost-effective.22
Despite rapid proliferation of EVT to treat peripheral arterial disease, hard data regarding the role of DUS after EVT
are sparse. A few conflicting studies have reported the
prognostic accuracy of DUS.13-16 Sacks et al14 found that
abnormal initial DU findings did not predict a higher rate
of restenosis with up to 36 months of follow-up. Tielbeek
et al13 suggested that clinical and hemodynamic assessment
alone after EVT was more useful than DUS. Their conclusions contradict studies from Spijkerboer et al,15 who observed an increased failure rate of EVT at 1 year in limbs
with abnormal DUS results the day after intervention.
Similarly, Mewissen et al16 reported that a functional residual stenosis 50% by DUS 1 week after PTA predicted
clinical failure (15% patency at 1 year). To date no reliable
data have been published documenting the natural history
of DU-detected stenosis after EVT. The present study
included only patients enrolled in a strict DUS protocol to
evaluate the natural history of restenosis and the clinical
utility of routine DUS after femoropopliteal EVT.
Limbs with TASC A/B lesions had a 65% frequency of
normal initial scans, which was similar to the 68% frequency
in limbs with TASC C/D lesions (Table I). Treated TASC
C/D lesions appeared to have a higher rate of lesion
progression. The rates of limb occlusion and major amputation were 23% and 8% for TASC C/D lesions compared
with 4% and 0% for TASC A/B lesions. Initial disease
severity likely contributes to restenosis frequency and severity and worsens outcomes.
Our data (Table II) showed that normal initial DU
scans were much more likely after PTA (75%) than after
stenting (63%) or atherectomy (48%), most likely because
PTA was performed more frequently for less severe disease.
It is also interesting to note that the rate of disease progression (the difference between the percentage of normal last
scans from normal initial scans) for lesions that were initially
stented was much less (7%) than after PTA (39%) or
atherectomy (18%). In contrast, occlusion occurred at a
higher rate after stenting (22%) than after atherectomy
(8.7%) or PTA (6.8%). Perhaps stents improve initial results
by preventing elastic recoil or intimal flap but later trigger
an adverse thrombogenic, inflammatory, or proliferative
response and resultant thrombosis.
It is worth comparing the natural history of DUSdetected lesions after EVT in this study with what is known
about vein bypass grafts. As with bypass, results of DUS
after two-thirds of interventions in our study were initially
normal23,24; however, only 62% of those initially normal
EVT-treated segments remained so during follow-up compared with 95% of initially normal vein grafts.23 After EVT,
28% of limbs with initially normal DUS results subsequently developed moderate or severe lesions, in contrast
to the reported 5% de novo stenosis rate in vein grafts with

JOURNAL OF VASCULAR SURGERY

Volume 55, Number 2

Bui et al 351

Table IV. Outcomes after occlusion after initial endovascular therapy

Limbs
1
2
3
4
5
6
7
8
9
10
11

Symptoms

Reintervention

Outcomes

Rest pain
Rest pain
New ulcer
Rest pain
Rest pain
Nonhealing ulcer
Rest pain
Nonhealing ulcer
Nonhealing ulcer
Asymptomatic
Nondisabling claudication

Femorocrural bypass
Femorocrural bypass
Femorocrural bypass
Femorocrural bypass
Femorocrural bypass
PTA
Stent
Atherectomy
PTA
None
None

Bypass graft patent

Bypass graft patent
Bypass graft patent
Bypass graft patent
Bypass graft patent
Limb reoccluded, required femorocrural bypass; graft patent
Reocclusion; required a below-knee amputation
Patent
Patent
Occlusion
Occlusion

PTA, Percutaneous transluminal angioplasty.

Fig 5. Receiver operating characteristic (ROC) curves show values of peak systolic velocity (PSV) and the velocity ratio (Vr) that
best predict limb occlusion.

normal results on initial scans.23 The 10% rate of thrombosis after EVT despite normal findings on the initial DUS is
also significantly higher than the 3% incidence reported for
vein grafts.24 The incidence of a similarly defined moderate
stenosis on the initial DUS (30%) was also higher than for
vein grafts (20%). Even though a significant proportion of
moderate lesions progressed to severe stenoses (47%) after
EVT, this incidence was somewhat lower than the 63%
progression rate we reported in vein grafts.24 Severe residual stenosis on the initial DUS was uncommon after EVT
(5% of all treated limbs), surprisingly lower than reported in
vein grafts (16%).24
In the present study, lesion stabilization or resolution
after EVT occurred in 40% and 80% of limbs with moderate
and severe lesions, respectively (Figs 3 and 4). In vein
grafts, the reported stabilization/regression rate for lesions
of similar severity is much lower, at 32% for moderate
lesions and 8% for severe lesions. The natural history of

restenosis in EVT-treated limbs thus seems markedly different from that observed in vein grafts. After EVT, the
tendency to develop restenosis is much greater, but lesions
appear more likely to stabilize or regress than those found
in vein grafts.
An interesting finding from our study was that 52% of
severe stenoses occurred in patients with ischemic symptoms who required reintervention, regardless of DUS findings. This stands in contrast to the natural history of vein
grafts, where symptomatic patients are far more likely to
present with thrombosis. In contrast to vein grafts, limbs
with severe restenosis after EVT are less frequently thrombosed and are far more often patent at the time of clinical
deterioration; such patients usually present with restenosis
rather than occlusion and can be retreated, much as they
would have been if prophylactic reintervention had been
performed. However, when thrombosis after EVT occurs,
repeat EVT is not highly effective (Table II).
Symptomatic occlusion developed in 9 of 94 limbs
(9.5%) in the study cohort: 5 (56%) required immediate
open bypass and the other 4 (44%) underwent endovascular
reintervention. Only two of four repeat endovascular interventions remain patent at 1 year. Of the latter two limbs
with recurrent EVT failure, one limb required a major
amputation and the other eventually required an open
bypass for limb salvage. Six of nine limbs (67%) that presented with symptomatic occlusion after EVT eventually
required bypass.
Another striking contrast between femoropopliteal restenosis after EVT noted in this study compared with
restenosis reported after vein graft implantation is that
severe stenosis was detected in only 1 of 11 EVT-treated
limbs before the onset of thrombosis: 82% of occlusions
occurred in limbs in which only a moderate stenosis had
been detected before thrombosis. Fig 5 shows the risk of
limb occlusion increases when the PSV and Vr values
exceed 200 cm/s and 2.1, respectively. The sensitivity of
post-EVT DUS to predict occlusion is 88%; however, the
60% specificity is low. Had these threshold DUS criteria
been used, 40% of limbs with stenoses would have undergone needless reintervention; had we applied the calculated
threshold criteria, 39 additional prophylactic reinterven-

JOURNAL OF VASCULAR SURGERY

February 2012

352 Bui et al

tions would have been required to potentially prevent 10

limb occlusions, accepting the unsubstantiated premise
that all reinterventions were successful and durable. It thus
appears that the currently accepted DUS threshold criteria
for reintervention after autogenous vein graft placement are
not applicable in patients undergoing femoropopliteal
EVT. Unless more accurate threshold reintervention criteria can be developed, the clinical utility of DUS after EVT is
questionable.

8.
9.

10.

11.

CONCLUSIONS
We recognize that our results are limited by the studys
retrospective nature and relatively low statistical power.
Nonetheless, the findings are interesting and suggest that
the natural history of restenosis after EVT differs from vein
graft stenosis. Larger prospective studies are warranted to
examine the clinical utility of DUS after femoropopliteal
angioplasty.

12.

13.

14.

AUTHOR CONTRIBUTIONS
Conception and design: JM, TB, DI, JH
Analysis and interpretation: JM, TB, AG, DI, KG
Data collection: TB, JH, KG, DI, JM
Writing the article: TB, DI, JM
Critical revision of the article: TB, DI, KG, JM
Final approval of the article: TB, DI, JH, AG, JM
Statistical analysis: AG
Obtained funding: JM
Overall responsibility: JM

15.

REFERENCES

19.

1. Cullen PJ, Lehay AL, Ryan SB, McBride KD, Moore DJ, Shanik GD.
The influence of duplex scanning on early patency rates of in situ bypass
to the tibial vessels. Ann Vasc Surg 1986;1:340-6.
2. Grigg MJ, Nicolaides AN, Wolfe JH. Femorodistal vein bypass graft
stenoses. Br J Surg 1988;75:737-40.
3. Mills JL, Harris EJ, Taylor LM, Beckett WC, Porter JM. The importance of routine surveillance of distal bypasses with duplex scanning: a
study of 379 reversed vein grafts. J Vasc Surg 1990;12:379-89.
4. Idu MM, Blankenstein JD, de Gier P, Truyen E, Buth J. Impact of a
color-flow duplex surveillance program on infrainguinal vein graft patency: a five-year experience. J Vasc Surg 1993;17:42-53.
5. Mattos MA, van Bemmelen PS, Hodgson KJ, Ramsey DE, Barkmeier
LD, Sumner DS. Does correction of stenoses identified with color
duplex scanning improve infrainguinal graft patency? J Vasc Surg 1993;
17:54-64; discussion 64-6.
6. Mills JL, Fujitani RM, Taylor SM. The characteristics and anatomic
distribution of lesions that cause reversed vein graft failure: a five-year
prospective study. J Vasc Surg 1993;17:195-206.
7. Bandyk DF, Schmitt DD, Seabrook GR, Adams MB, Towne JB.
Monitoring functional patency of in situ saphenous vein bypasses:

16.

17.

18.

20.

21.

22.

23.

24.

the impact of a surveillance protocol and elective revision. J Vasc

Surg 1989;9:286-96.
Mills JL, Sr. Infrainguinal vein graft surveillance: how and when. Semin
Vasc Surg 2001;14:169-76.
Norgren I, Hiatt WR, Dormandy JA, Nehler MR, Harris KA, Fowkes
FGR. Inter-society consensus for the management of peripheral arterial
disease (TASC II). J Vasc Surg 2007;45:S5A-S67A.
Schillinger M, Sabeti S, Loewe C, Dick P, Amighi J, Mlekusch W, et al.
Balloon angioplasty versus implantation of nitinol stents in the superficial femoral artery. N Eng J Med 2006;354:1879-88.
Amighi J, Schillinger M, Dick P, Schlager O, Sabeti S, Mlekusch W, et
al. De novo superficial femoropopliteal artery lesions: peripheral cutting
balloon angioplasty and restenosis ratesrandomized controlled trial.
Radiology 2008;247:267-72.
Schillinger M, Sabeti S, Dick P, Amighi J, Mlekusch W, Schlager O, et
al. Sustained benefit at 2 years of primary femoropopliteal stenting
compared with balloon angioplasty with optional stenting. Circulation
2007;115:2745-9.
Tielbeek AV, Rietjens E, Buth J, Vroegindeweij D, Schol FP. The value
of duplex surveillance after endovascular intervention for femoropopliteal obstructive disease. Eur J Vasc Endovasc Surg 1996;12:145-50.
Sacks D, Robinson ML, Summers TA, Marinelli DL. The value of
duplex sonography after peripheral artery angioplasty in predicting
subacute restenosis. AJR Am J Roentgenol 1994;162:179-83.
Spijkerboer AM, Nass PC, de Valois JC, van der Graaf Y, Eikelboom
BC, Mali WP. Evaluation of femoropopliteal arteries with Duplex
ultrasound after angioplasty. Can we predict results at one year? Eu J
Vasc Endovasc Surg 1996;12:418-23.
Mewissen MW, Kinney EV, Bandyk DF, Reifsnyder T, Seabrook GR,
Lipchik EO, et al. The role of duplex scanning versus angiography in
predicting outcome after balloon angioplasty in the femoropopliteal
artery. J Vasc Surg 1992;15:860-5.
Buth J, Disselhoff B, Sommeling C, Stam L. Color-flow duplex criteria
for grading stenosis in infrainguinal vein grafts. J Vasc Surg 1991;14:
716-26; discussion 726-8.
Roth S, Bandyk DF. Duplex imaging of lower extremity bypasses,
angioplasties, and stents. Semin Vasc Surg 1999;12:275-84.
Moody P, Gould DA, Harris PL, Vein graft surveillance improves
patency in femoropopliteal bypass. Eur J Vasc Surg 1990;4:117-21.
Bergamini TM, George SM Jr, Massey HT, Henke PK, Klamer TW,
Lambert GE, et al. Intensive surveillance of femoropopliteal-tibial
autogenous vein bypasses improves long-term graft patency and limb
salvage. Ann Surg 1995;221:507-16.
Lundell A, Lindblad B, Bergqvist D, Hansen F. Femoropopliteal-crural
graft patency is improved by an intensive surveillance program: a
prospective, randomized study. J Vasc Surg 1995;21:26-34.
Wixon CL, Mills JL, Westerband A, Hughes JD, Ihnat DM. An
economic appraisal of lower extremity bypass graft maintenance. J Vasc
Surg 2000;32:1-12.
Mills JL, Bandyk DF, Gahtan V, Esses GE. The origin of infrainguinal
vein graft stenosis: a prospective study based on duplex surveillance. J
Vasc Surg 1995;21:16-25.
Mills JL, Wixon CL, James DC, Devine J, Westerband A, Hughes JD.
The natural history of intermediate and critical vein graft stenosis:
recommendations for continued surveillance or repair. J Vasc Surg
2001;33:273-80.

Submitted Apr 7, 2010; accepted Aug 10, 2011.