A prospective study of the clinical outcome

of femoral pseudoaneurysms and

arteriovenous fistulas induced by
arterial puncture
K. Craig Kent, M D , Colin R. McArdle, M D , Bernadette Kennedy, R D M S ,
D o n a l d S. Baim, M D , Elaine Anninos, R N , and John J. Skillman, M D , Boston, Mass.

Purpose: Although spontaneous thrombosis of femoral false aneurysms (FAs) and

arteriovenous fistulas (AVFs) has been reported, the frequency of this occurrence is
unknown. This prospective study was designed to establish the natural history of FA and
AVF and to evaluate factors that might predict eventual thrombosis of these lesions.
Methods: Twenty-two patients with either femoral FAs (n = 16) or AVFs (n = 6) induced
by percutaneous arterial punctures were evaluated prospectively. After an initial duplex
scan, all patients were monitored with serial scans, either in hospital or weekly as
outpatients, depending on the stability of the process. Operative repair was performed for
the following indications: (1) a greater than 100% increase in size ofa FA by duplex scan,
(2) the development of symptoms, or (3) continued patency of the lesion after 2 months
of observation.
Results: Nine of 16 FAs and four of six AVFs closed spontaneously; FAs greater than 6 cm s
(1.8 cm in diameter) required repair more often (p = 0.065). However, size was not an
absolute predictor of the need for repair because two small aneurysms (1.6 and 0.7 cm a)
remained patent, although both patients were discharged safely from the hospital, and two
large aneurysms (13.2 and
10.7 cm s) thrombosed spontaneously. Three of seven patients whose aneurysms required
repair received anticoagulation continuously from the time of catheterization until repair
became necessary. None of the patients whose FAs closed spontaneously were receiving
anticoagulants at the time of thrombosis (p = 0.02). Neither length of the FA neck,
velocity in the FA cavity, size of original arterial puncture, nor velocity in the AVF
correlated with thrombosis.
Conclusions: We conclude that (1) all FAs do not thrombose spontaneously and at least one
third require surgical repair, (2) patients receiving continuous anticoagulation should
undergo aneurysm repair, (3) discharge of patients with FAs less than 6 cm s is safe (the
majority of these FAs will eventually thrombose spontaneously), and (4) many AVFs close
spontaneously and repair is not required unless symptoms or signs of progressive
enlargement develop. (J VASC SURG 1993;17:125-33.)

The development of complex percutaneous vas-
cular procedures, including angioplasty, atherec-
tomy, stenting, and valvuloplasty has resulted in a rise
in the incidence of arterial injury after femoral
From the Departmentsof Surgery(VascularDivision) (Drs. Kent
and Skillmanand Ms. Aaminos),Radiology(Dr. McArdleand
Ms. Kennedy), and Medicine (CardiovascularDivision) (Dr.
Balm), Beth Israel Hospital and Harvard Medical School,
Boston.
Presented at the Forty-sixthAnnual Meeting of the Societyfor
Vascular Surgery, Chicago,Ill., June 8-9, 1992.
Reprint requests: K. Craig Kent, MD, Beth Israel Hospital, 330
Brookline Ave., Boston,MA 02215.
24/6/41707
0741-5214/93/$1.00 + .10.
catheterization. ~The larger sheaths that are necessary
for these more complex interventions, and the more
frequent use of anticoagulation, are two factors that
contribute to the increased frequency of complica-
tions resulting from arterial puncture. 2,3
Although uncontrollable hemorrhage and limb-
threatening ischemia mandate urgent surgical inter-
vention, controversy exists regarding the proper
management of femoral false aneurysms (FAs) and
arteriovenous fistulas (AVFs). Thought to be unsta-
ble lesions, FAs were traditionally repaired immedi-
ately after their discovery.4 Reports describing spon-
taneous thrombosis of FAs have appeared recently, s#
125

126 Kent et al.
Kresowik et al. 5 monitored seven pseudoaneurysms
prospectively and found that within 2 weeks of their
discovery, all seven thrombosed spontaneously. In
the same study three AVFs were observed for 4 weeks
and all remained patent.
Although spontaneous thrombosis has been ob-
served, FAs may also rapidly expand, rupture, pro-
duce a femoral neuralgia, or lead to femoral venous
thrombosis. Considering the risk associated with
each of these events, we were concerned that a
nonoperative approach for all FAs would be associ-
ated with increased morbidity. Also, in contrast to
the findings of Kresowik et al.,s we and others6 have
frequently observed spontaneous closure of femoral
AVFs.
Thus a prospective study of the chnical outcome
of FAs and AVFs was undertaken. Our study was
designed to determine how often FAs and AVFs will
thrombose spontaneously. We searched also for
variables that might predict eventual thrombosis of
FAs or AVFs. The abihty to predict the fate of a FA
or AVF would allow patients soon after the discovery
of these lesions either to be safely discharged from the
hospital or to undergo repair of their femoral artery
injury.
MATERIAL A N D M E T H O D S
From May 1, 1991, to January 1, 1992, all
patients undergoing femoral artery catheterization at
the Beth Israel Hospital were eligible for study.
Patients were routinely examined by the medical or
surgical staff after catheterization for the presence of
a new femoral bruit or a widened pulse. A vascular
surgery consultation and duplex scan were obtained
in all patients suspected of having a FA or AVF.
After verification of the presence of either of these
two lesions, informed consent was obtained and the
patient was entered into the study.
Patients were excluded from the study if they had
progressive hemorrhage from the femoral artery after
catheterization (n = 2). Femoral artery repair was
performed urgently in these patients. Several patients
were found to have a pulsatile groin mass that
expanded rapidly over a period of hours (n -- 6).
Although these lesions could be considered to be
FAs, we included in our prospective study only those
lesions that appeared to be stable for at least 24 hours.
All duplex scans were performed by one ultra-
sonographer (B.K.) and reviewed by one radiologist
(C.M.). An Acuson 128 XP with color-flow Doppler
(Acuson, Inc., Mountain View, Calif.) was used for
the duplex scans. The following information was
derived from the initial duplex scan: FAs-three-
Journal of
VASCULAR
SURGERY
dimensional size of FA cavity, velocity of flow within
the FA, and length of FA neck; and AVF-velocity
of flow in the AVF.
Patients remained in the hospital for at least 2
days after initial discovery of a FA; patients with
AVFs were discharged without delay. Daily in-
hospital inspections were made by a vascular surgeon
(J.J.S. or K.C.K.), and all patients with FAs under-
went repeat duplex scanning at 48 hours. Patients
with stable FAs were then discharged and followed at
weekly intervals by a vascular surgeon. Ultrasound
examinations were performed at 1-week intervals.
False aneurysms were repaired for the following
reasons: (1) development of symptoms, (2) expan-
sion greater than 100%, or (3) failure to thrombose
after 2 months of observation. Arteriovenous fistulas
were repaired if they did not close after 2 months of
observation.
Femoral artery repair, when required, was per-
formed under local anesthesia with anesthesia
standby.
RESULTS
During the 8-month period of the study, 1838
femoral catheterizations were performed at this
institution (1454 cardiac and 384 other). Twenty-
four patients were discovered to have either a FA
(n = 18) or AVF (n = 6) (incidence 1.3%). Two
patients with FAs were not studied prospectively
because appropriate arrangements for follow-up
could not be made. Both of these FAs were repaired.
Sixteen FAs and six AVFs were monitored prospec-
tively. The average age of all patients was 71 years.
Nine patients were female and 13 were male.
Femoral artery puncture was performed in these
patients for a variety of reasons, including diagnostic
cardiac catheterization (n = 8), percutaneous trans-
luminal coronary angioplasty (PTCA) ( n - 7),
PTCA with stent (n = 4), aortic vaivuloplasty
(n = 1), coronary atherectomy (n = 1), and place-
ment of a dialysis line (n = 1). Sheath sizes were
standard for each procedure: diagnostic catheteriza-
tion, 7F; PTCA, 8F; PTCA with stent, 8F; and
valvuloplasty, 12F.
Nine FAs thrombosed spontaneously, and seven
required repair. The average interval from catheter-
ization to spontaneous closure of FAs was 22 days,
with a range of 3 to 34 days. All seven FAs that
required repair expanded during the period of
in-hospital observation. Reasons for repair included
expansion greater than 100% (n = 4), severe femoral
pain (n = 2), and femoral neuralgia (n = 1). It was
necessary for the five patients with large FAs who
Volume 17
Number 1
January 1993 Pseudoaneurysms and fistulas induced by arterial puncture 127

Table I. Data in 16 patients with FAs

Initial size Outcome
Largest After
diameter Volume catheterization Increasein Length of Anticoagulation
Procedure (cm) (cms) ClosedRepaired (days) size (%) Symptoms neck (mm) heparin/Coumadin
PTCA 3.3/3.3 13.2/i4.8" x 34 0 Heparin, 24 hr
Diag cath 3.3 10.7 x 9 17 Heparin, procedure
only
PTCA 2.2 5.2 x 26 7 Heparin, 72 hr
PTCA 1.7 4.1 x 33 12 Heparin, 24 hr
Diag cath 2.1 4.0 x 15 0 Heparin, procedure
only
Diag carla 2.3 2.7 x 3 2i Heparin, procedure
only
Diag cath 1.9 2.5 x 21 3 Heparin, procedure
only
Dialysis line 2.6 2.1 X 27 0 None
PTCA 1.4 1.2 X 31 2 Heparin, 2.4 hr
PTCA/stent 5.9 41.3 x 6 40 Severe groin pain 0 Heparin/Cotunadin
through repair
PTCA/stent 3.4 18.0 x 8 232 5 Heparin/Coumadin
through repair
Aortic valve 3.5 10.1 x 4 289 20 Heparin, 24 hr
Diag cath 2.8 10.0 x 4 106 2 Heparin through
repair
PTCA 2.3 7.4 x 7 30 Severe groin pain 0 Heparin, 24 hr
PTCA 1.6 1.6 x 13 79 Femoral neuralgia 2 Heparin, 24 hr
Diag cath 1.0 0.7 x 51 203 17 Heparin, procedure
only
Diag cath, Diagnostic catheterization; aortic valve, aortic valvailoplasty.
*Bilobed.

eventually required repair to remain in the hospital
for a mean of 6 + 2 days. The patients with smaller
aneurysms were discharged from the hospital. One
patient returned at 7 days with femoral neuralgia
prompting repair, and a second patient underwent
elective repair 51 days after catheterization.
Table I shows the profile of each patient with a
FA. Aneurysms less than 6 cm 3 (1.8 cm in diameter)
had a greater tendency to thrombose spontaneously
compared with aneurysms greater than 6 cm 3
(p = 0.065, Fisher's exact test). However, as the p
value reflects, size was not an absolute predictor of
aneurysm closure. Two of seven large aneurysms
thrombosed and two of nine FAs less than 6 cm 3
remained patent.
The connection between the native femoral artery
and the FA was identified by ultrasonography in all
cases. In those aneurysms that closed, the average
length of this neck was 8 mm, compared with 9 mm
in aneurysms that required repair (p > 0.4, t test).
The velocity of flow in the FA cavity was determined
in 13 of the 16 patients. The average velocity in the
aneurysms that closed was 90 cm/sec versus 123
cm/sec in patients whose aneurysms remained patent
(p > 0.1, t test). Thus neither the length of the neck
nor the velocity of flow could be used to predict the
propensity for closure of an aneurysm. Also, neither
the type of procedure nor the sheath size was useful
in predicting thrombosis of FAs or AVFs (p = 0.9).
All patients but one received heparin at the time
of femoral artery puncture. In 10 patients, heparin
was continued for 24 hours after the procedure.
Anticoagulation (heparin, Coumadin, or both) was
continued for greater than 24 hours in only four
patients. None of the nine patients whose aneurysms
closed were receiving anticoagulants at the time that
closure occurred. Three of seven patients requiring
repair were receiving anticoagulants continuously
from the time of catheterization until expansion or
the development of symptoms necessitated operative
repair (p = 0.02, Fisher's exact test).
Table II shows the data of six patients with AVFs.
A fistulous tract could not be identified by ultra-
sonography in two patients, and the diagnosis was
made by auscultation of a systofic/diastolic bruit in
the groin and a duplex scan demonstrating turbu-
lence in the femoral vein. Four AVFs closed sponta-
neously and two required repair. One of the two
repairs was performed at 4 weeks in conjunction with
a coronary bypass. The second AVF was repaired 4

128 Kent et al.
Table II. Data in six patients with AVFs
Outcome After
catheterization
Procedure Closed Repaired (days)
Diag cath x 51
Diag cath x 37
PTCA/stent x 29
PTCA/stent x 32
Coron ather x 127
PTCA x* 30
Diag cath, Diagnostic catheterization; coron ather, coronary atherectomy.
*Repaired at 4 weeks at time of scheduled coronary artery bypass.
months after the cardiac catheterization. The average
interval between catheterization and spontaneous
closure of the AVF was 37 days.
During the course of this prospective study, eight
additional patients underwent urgent operation for
femoral artery hemorrhage. Two additional patients
were found to have FAs related to catheterizations
that had been performed 6 weeks and 3 years earlier.
Both of these lesions were repaired.
DISCUSSION
False aneurysms. A study of the natural history
of femoral pseudoaneurysms requires a reproducible
and accurate method for evaluating FA size. Previous
studies s have used the greatest diameter of an
aneurysm as an estimate of size. However, FAs are
often oblong, irregular, or even multilobed (Fig. 1).
In this study all three dimensions were measured
precisely, and the size was reported as the product of
these dimensions. In calculating the volume of an
aneurysm, it is important to separate the cavity
containing flow from the surrounding hematoma.
This is accomplished most accurately with a duplex
scan. 8"1°The use of color enhances further the ability
of the duplex scan to define the region of flow. n
Fifty-six percent of the FAs that we studied
thrombosed spontaneously (Fig. 2). Of those FAs
that required repair, progressive enlargement was
noted in all and symptoms developed in three
patients (Fig. 3). Our criteria for repair were defined
prospectively. We assumed that most vascular sur-
geons would consider the development of symptoms
or expansion of more than 100% as appropriate
indications for repair. There was no morbidity
associated with the observation of patients with FAs
when these criteria were used.
Patients were more likely to require repair ofa FA
if they were receiving anticoagulant drugs continu-
ously. Three of seven patients with aneurysms that
Journal of
VASCULAR
SURGERY
Velocity in
fistulous tract Anticoagulation
(cm/sec) heparin/Coumadin
Neck not visible Heparin, procedure only
150 Heparin, procedure only
120 Heparin/chronic Coumadin
Neck not visible Heparin/chronic Coumadin
150 Heparin, 24 hr
30-50 Heparin, 24 hr
required repair were treated with anticoagulants from
the time of catheterization until repair. None of the
nine patients whose aneurysms thrombosed sponta-
neously were receiving anticoagulants at the time the
aneurysm closed.
Size of the aneurysm on the initial duplex scan
was not a statistically significant predictor of closure.
A size of 6 cm s allowed the best statistical division
between large and small aneurysms (p -- 0.065).
Two of seven FAs greater than 6 cm 3 thrombosed
spontaneously, and two of nine FAs less than 6 cm 3
required repair. All patients with FAs less than 6 cm s
were discharged safely and monitored as outpatients.
We felt that discharge of patients with larger FAs
could be potentially hazardous.
Although many of the FAs in this study were
related to complex cardiologic procedures requiring
large sheaths, the type of procedure or sheath size
could not be used to predict whether an FA would
close. It might be expected that aneurysms with
longer necks would have a greater propensity to
occlude. However, there was no correlation between
the length of the neck and spontaneous closure.
Velocity of flow in the aneurysm cavity was also not
a useful predictor of closure. Because velocity varies
throughout the cavity, it was difficult to know which
value to choose. The swirling nature of blood flow
within the cavity made an accurate estimation of the
Doppler angle of the interrogating beam difficult.
Based on these findings, we recommend that
aneurysms be repaired in patients who are receiving
continuous anticoagulation. Patients with aneurysms
that are less than 6 cm 3 can be discharged safely from
the hospital. The majority of these FAs will throm-
bose spontaneously (seven of nine did so), and
outpatient monitoring of these patients appears to be
safe. For the remaining patients with FAs greater
than 6 cm 3 who are not receiving continuous
anticoagulation, a period of in-hospital observation
Volume 17
Number 1
January 1993
Fig. 1. Femoral false aneurysms are usually irregular in shape and determination of true size
requires measurement of three dimensions. A, Transverse image demonstrates aneurysm that is
4.8 cm in width and 2.4 cm in depth. B, Longitudinal projection shows aneurysm to be 2.7 cm
in length.
allowed 50% (two of four) to thrombose spontane-
ously.
Arteriovenous fistula. Although iatrogenic fem-
oral AVFs pose no immediate risk for the patient,
occasionally high-output congestive heart failure,
lower extremity edema, or even arterial insufficiency
may develop if the fistula remains patent. 12'is The
long-term fate of these lesions is unknown, and the
proportion of patients who will eventually have
complications has not been determined. Most pa-
tients with AVFs have underlying coronary artery
disease. The insidious onset of congestive heart
failure could be potentially dangerous. It seems
prudent to be certain that all AVFs eventually dose.
Four of six AVFs in this study closed spontane-
ously. One of the remaining two was repaired 4
weeks after its recognition in conjunction with a
coronary artery bypass. The bruit in this patient had
been gradually increasing in intensity during the
4-week period. It seemed unlikely that spontaneous
closure would occur. The second AVF was repaired
4 months after the initial catheterization.
Velocity of flow through the aneurysm neck
could be measured only in four of the six AVFs. 14
Velocities ranged from 30 to 150 cm/sec, but
unfornmately velocity did not predict closure or
continued patency of the AVF. Anticoagulation was
also not a useful predictor of whether these lesions
would close.
Because two thirds of these lesions developed
thrombosis spontaneously, we recommend that
Pseudoaneurysms and fistulas induced by arterial puncture 129
AVFs not be repaired immediately. Repair can be
performed later if the AVF becomes symptomatic or
shows signs of progressive enlargement.
Ultrasound-guided compression repair. It has
been proposed that all femoral FAs and AVFs should
be treated with ultrasound-guided compression re-
pair. is Ultrasonography is used to identify the neck of
the pseudoaneurysm or AF. A downward force is
applied with the transducer until flow in the FA or
through the AVF is eliminated. FeUmeth et al}s were
able to treat successfully 27 of 33 patients in whom
ultrasound-guided compression repair was at-
tempted. Drawbacks of this technique include dis-
comfort associated with compression and the poten-
tial for femoral artery thrombosis. The utility of this
procedure is uncertain in patients receiving antico-
agulation) 6 It would seem that the spontaneous
resolution of an FA or AVF would be preferable to
any type of intervention, whether this be ultrasound-
guided compression or surgery. Spontaneous closure
of FAs occurred in 59% of our patients.
However, the excellent results reported by Fell-
meth et al}s are encouraging. Our most recent
protocol includes an initial attempt to close pseudo-
aneurysms by ultrasound-guided compression in
patients who are receiving continuous anticoagula-
tion or in patients whose FAs are large.
Because two thirds of AVFs thrombose sponta-
neously and the majority do not become symptom-
atic, we do not see a role for this method in the
treatment of AVFs.
 Journal of
VASCULAR
130 K e n t et al. SURGERY

Fig. 2. A, A 2.7 cm 3 FA was noted immediately after diagnostic catheterization. B, Duplex

scan 3 days later shows no flow in previous aneurysm cavity, indicating that closure has occurred.

Fig. 3. A, After femoral artery puncture for coronary angioplasty, duplex scan shows false
aneurysm 1.7 cm in length and 1.0 cm in depth. B, Six days later, FA has increased in size and
now measures 3.0 cm in length and is 1.3 cm deep.

H e m o r r h a g i c complications. During the course
of this prospective study, eight additional patients
underwent femoral artery exploration for hemor-
rhage. The usual indication for operation was an
expanding hematoma, which in several patients was
associated with hemodynamic instability. Although
these patients were not included in the analysis of
FAs, they serve to demonstrate the risks associated
with continued patency of a laceration after FA
puncture.
Summary. We make the following conclusions
and recommendations: (1) All false aneurysms do not
thrombose spontaneously; at least one third will
require surgical repair. (2) Patients receiving contin-
uous anticoagulation should undergo repair of the
aneurysms. (3) Patients with aneurysms less than 6
cm 3 can be discharged safely from the hospital if close
follow-up is maintained. The majority of these FAs
will thrombose spontaneously. (4) Many AVFs close
spontaneously. Repair is not required unless symp-
toms or signs of progressive enlargement develop.
All FA and AVF repairs are performed under local
anesthesia. Morbidity and mortality rates of the
procedure are low, particularly in patients who are
hemodynamically stable at the time of operation. 3 O f
the seven patients who underwent repair during the
course of this study, there were no deaths or
morbidity, although three of seven patients required
transfusion. We believe that the potential risks
associated with symptomatic or expanding FAs far
outweigh the small risk that is associated with repair.
Spontaneous thrombosis of an FA or AVF is
Volume 17
Number 1
January 1993
preferable to any interventional procedure.
U l t r a s o u n d - g u i d e d compression repair m a y be useful
in the treatment o f those FAs that are unlikely to
t h r o m b o s e spontaneously, and we await further
studies to verify the efficacy o f this technique.
REFERENCES
1. Oweida SW, Roubin GS, Smith RB, Salam AA. Postcathe-
terization vascular complications associated with percutane-
otis transluminal coronary angioplasty. J VAsc Suv,G 1990;
12:310-5.
2. Wyman RM, Safian R_D,Portway V, Skillman JJ, McKay RG,
Balm DS. Current complications of diagnostic and therapeu-
tic cardiac catheterization. J Am Coil Cardiol 1988;12:
1400-6.
3. Skillman JJ, Kim D, Baim DS. Vascular complications of
percutaneous femoral cardiac interventions. Arch Surg 1988;
123:1207-12.
4. Mills JL, Wiedeman JE, Robison IG, Hallett JW. Minimizing
mortality and morbidity from iatrogenic arterial injuries: the
need for early recognition and prompt repair, l V^sc St3RG
1986;4:22-7.
5. Kresowik TE, Khoury MD, Miller BV, et al. A prospective
study of the incidence and natural history of femoral vascular
complications after percutaneous transluminal coronary an-
gioplasty. ~ VASCSURG 1991;13:328-36.
6. McCann R_L, Schwartz LB, Pieper KS. Vascular complica-
tions of cardiac catheterization. J VASCSURG 1991;14:375-
81.
7. Kotval RS, Khoury A~ Sha PM, Babu SC. Doppler sono-
graphic demonstration of the progressive spontaneous throm-
bosis ofpseudoaneurysms. J Ulttasound Med 1990;9:185-90.
DISCUSSION
Dr. Malcolm O. Perry (Lubbock, Texas). Why did
you decide on 6 cm 3, and how did you determine those
dimensions? I must have missed that in the early part.
Dr. K. Craig Kent. For each FA, width, length, and
depth were determined. The volume was calculated by
taking the product of these three dimensions.
Dr. Perry. Because there is laminated clot and you
could not visualize all that with the duplex scan, was it a
combination of tactile work as well?
Dr. Kent. Color-flow duplex scanning is able to
differentiate between a central area of flow and outer
laminated clot. When determining the volumes of these
FAs, we measured only the portion of the cavity where flow
was present.
We observed that smaller FAs were more likely to
thrombose spontaneously and larger FAs were more likely
to remain patent. The value that provided the best statistical
separation between large and small was 6 cm 3 (p = 0.065).
Although not statistically significant, this was practically
useful; we were able to send all of the patients with FAs less
Pseudoaneurysms and fistulas induced by arterial puncture 131
8. Sacks D, Robinson ML, Perlmutter GS. Femoral arterial
injury following catheterization: duplex evaluation. J Ultra-
sound Med 1989;8:241-6.
9. Coughlin BF, Paushter DM. Peripheral pseudoaneurysms:
evaluation with duplex US. Radiology 1988;168:339-42.
10. Helvie MA, Rubin JM, Silver TM, Kresowik TF. The
distinction between femoral artery pseudoaneurysms and
other causes of groin masses: value of duplex doppler
sonography. AJR 1988;150:1177-80.
11. Sheikh KH, Adams DB, McCann R, Lyerly HK, Sabiston
DC, Kisslo J. Utility of Doppler color flow imaging for
identification of femoral arterial complications of cardiac
catheterization. Am Heart J 1989;117:623-8.
12. Glaser RL, McKellar D, Scher KS. Arteriovenous fistulas after
cardiac catheterization. Arch Surg 1989;124:1313-5.
13. Kron J, Sutherland D, Rosch J, Morton MJ, McAnulty JH.
Arteriovenous fistula: a rare complication of arterial punc-
ture for cardiac catheterization. Am J Cardiol
1985;55:1445-6.
14. Igidbashian VN, Mitchell DG, Middleton WD, Schwartz RA,
Goldberg BB. Iatrogenic femoral arteriovenous fistula: diag-
nosis with color Doppler imaging. Radiology 1989; 170: 749-
52.
15. Fellmeth BD, Roberts AC, Bookstein JJ, et al. Postangio-
graphic femoral artery injuries: nonsurgical repair with
US-guided compression. Radiology 1991;178:671-5.
16. Dorfman GS, Cronan JJ. Postcatheterization femoral artery
injuries: is there a role for nonsurgical treatment? Radiology
1991;178:629-30.
Submitted June 12, 1992; accepted Aug. 7, 1992.
than 6 cm 3 home, and most of the FAs (seven of nine)
thrombosed spontaneously.
Dr. Perry. So we are talking about 1.5 cm in diameter?
Dr. Kent. Yes, the cube root is approximately 1.8 cm.
Dr. Robert B. Smith (Atlanta, Ga.). Your 1.7% local
vascular complication rate, after a variety of transfemoral
procedures, compares favorably with reports from other
centers and emphasizes that pseudoaneurysms and arterial
venous fistulas are infrequent complications of very com-
mon procedures. When they do occur, the vascular surgeon
must recommend a course of action.
All would agree that large or symptomatic false
aneurysms should be repaired immediately, but what about
those 2 cm in diameter or less? Can they be observed safely
on an outpatient basis provided the patient lives locally, is
reliable to return for frequent office visits, and has sufficient
financial resources to comply in such a program of care?
Obviously, a conservative treatment plana is not appropri-
ate for regional referral centers where patients live at
considerable distance from the hospital and the local

132 Kent et al.
physician may be intolerant o f being asked to follow
complications created at the tertiary facility. Our own
experience, and that o f Dr. Kent's group, suggests that
anticoagulation therapy may further reduce the likelihood
of spontaneous thrombosis of false aneurysms.
To these considerations should be added the acknowl-
edgement that repair of a pseudoaneurysm under local
anesthesia is safe and effective and the patient is usually
discharged from the hospital the following day. The
relatively small number of cases that meet the criteria for
nonoperative care, the definitiveness of surgical treatment,
and the inability to guarantee successful outcome with
conservative management argue against a wair-and-see
approach for most false aneurysms, in our opinion. We
tend to repair them when they are diagnosed.
Although we seem more aggressive than you in our
approach to false aneurysms, we are more conservative in
our management of iatrogenic arteriovenous fistulas. Our
experience suggests that many small arteriovenous com-
munications resolve over time and few are responsible for
serious problems. Accordingly, we simply observe most
femoral arteriovenous fistulas not associated with false
aneurysms, delaying repair for up to 6 months. At that
time, elective surgical repair is recommended for persistent
fismlas, not out of concern about imminent consequences
of the fistula but from apprehension about eventual loss of
control of the patient with the possibility of a remote
complication resulting in criticism of our cardiology
associates. Clearly, an extended multicenter study is needed
to define better treatment alternatives in such cases.
Would it be cost-effective or clinically useful to search
for iatrogenic vascular complications by routine duplex
scanning of all patients undergoing transfemoral proce-
dures before their release from care?
Dr. Kent. I do not believe it would be efficacious to
perform routine duplex scanning of all patients who
undergo femoral artery catheterization. In a center where
attending and house staff are ammed to these complica-
tions and reliably examine patients after catheterization, it
would be remote that a "missed" lesion would lead to any
eventual harm.
In the past we were very aggressive about repairing all
FAs. At the meeting of the International Society for
Cardiovascular Surgery, North American Chapter, 2 years
ago, Dr. Kresowik reported that all of the FAs in his series
thrombosed spontaneously. Unfortunately, about 2 weeks
later an FA that we were monitoring ruptured. In our
experience some FAs thrombose spontaneously and some
do not. We hoped, with this study, to establish criteria that
would help determine which ones we should monitor and
which ones should be repaired. Six cubic centimeters turns
out to be a very practical value. We were able to discharge
patients with aneurysms smaller than 6 cm 3, and most of
these anenrysms thrombosed. As you mentioned, we also
found that aneurysms in patients who received anticoagu-
lant medication were less likely to thrombose.
Dr. Geoffrey S. Cox (Cleveland, Ohio). You have
Journal of
VASCULAR
SURGERY
partly answered my question, which was going to be what
your experience with ultrasound-guided compression re-
pair has been. Instead, I will answer in part your question
and give you our experience.
During the last 12 months at the Cleveland Clinic, we
have seen 79 patients with postcatheterization pseudoan-
eurysms, including 22 who were given anticoagulants and
an additional who were about to undergo open-heart
surgery. Although, as you have clearly demonstrated, many
pseudoaneurysms thrombose spontaneously, some ~ and
do develop symptomatic enlargement. We now prefer to
treat all these lesions with ultrasound-guided compression
repair. Of 74 pseudoaneurysms treated in this way, we have
successfully occluded 93%, including 98% of patients who
are not given anticoagulants and 82% of those receiving
heparin or Coumadin. There have been no complications,
and all patients treated successfully were able to be
discharged the following day if their other conditions
permitted.
We believe that this method of treatment is safe and
reliable and reduces the need for the close follow-up that
you have recommended, especially in patients taking
anticoagulants and for those traveling long distances.
Furthermore, with a conservative approach, the mean time
to occlusion is 32 minutes overall. For those patients with
aneurysms present for greater than 30 days, this rises to 75
minutes. In our experience the size of the aneurysms did
not influence our ability to occlude the aneurysms.
Therefore we prefer to see them acutely and thrombose
them, rather than have to follow up on them.
Dr. Richard L. McCann (Durham, N.C.). I also
support the concept of compression therapy. We were
quite interested in this problem until about 11/2 years ago
when we instituted a similar program with compression
therapy, and in our last 40 cases we have failed to occlude
the false aneurysm in only two cases.
We use a slightly different system. We do use ultra-
sound color-flow guidance but also the C clamp and find
that is more comfortable and more reproducible in the
patients. On the other hand, we have found that this
therapy has not been helpful in resolving arteriovenous
fistulas, but, like Dr. Smith, we find that these cause very
few problems in the long term. Our only question is, what
are the long-term implications and indications for repair of
arteriovenous fistulas? We think that very few of these do
become hemodynamically significant and we have seen
several of them that have been present for 5 and 6 years,
without increasing in size.
Dr. Timothy F. Kresowik (Iowa City, Iowa). I am the
author of the report on this subject presented at these
meetings 2 years ago. I do not think that there is a large
discrepancy between our results and yours. The differences
that do exist can probably be explained by differences in our
patient populations. Our study was a series of unselected
patients who underwent transfemoral coronary angioplasty
and routine color duplex scanning irrespective of any
clinical suspicion of puncture-related complications. Your
Volume 17
Number 1
January 1993
patients were specifically referred for evaluation because a
puncture complication was suspected. I would assume that
a common cause of this suspicion was the presence of a
large hematoma. We have observed that the presence of a
large surrounding hematoma may indicate a less stable
pseudoaneurysm. Was this in fact the case with your
patients? Because three of your seven patients who required
repair were operated on for symptoms, I wonder whether
the symptoms were caused by a large hematoma rather than
the pseudoaneurysm itself. You also mentioned that three
o f your seven patients who required repair were receiving
continuous anticoagulation. Was there a relationship
between continuous anticoagulation and the pseudoaneu-
rysm expansion that you described? In your pseudoaneu-
rysms that did expand did this enlargement occur in one
specific dimension or was it generalized?
With respect to arteriovenous fistulas, we had observed
a relationship between a puncture site distal to the femoral
bifurcation and arteriovenous fistula formation. Specifically
a common cause appeared to be a puncture into the
superficial artery and the circumflex femoral vein, which
passes between the superficial and deep femoral arteries.
Did this site of fistula formation occur in your experience?
Dr. Robert J. Hye (San Diego, Calif.). I just wanted
to echo the comments o f the previous discussants. We have
been using ultrasound-guided compression at the Univer-
sity of California, San Diego, for the past 3 years and have
treated well over 70 patients, with a success rate of greater
than 90%. You alluded to cost and discomfort of the
procedure as being potential problems.Although there is a
cost associated with the compression procedure itself, we
believe it is more cost-effective than performing serial
ultrasound examinations or seeing the patient repeatedly in
the clinic to evaluate the pseudoaneurysm. The discomfort
of the procedure is occasionally a problem but can generally
be managed with small doses of intravenous sedatives or
narcotics.
We also believe that it is important to treat these lesions
early to optimize the oppommity for successful ultrasound-
guided compression. Our experience has been that the
more long-standing pseudoaneurysms are much more
difficult to obliterate with this method.
Finally, we have had one incidence of femoral arterial
thrombosis with embolization to the popliteal artery. This
occurred in a patient with a 6-week-old pseudoaneurysm in
whom the required compression resulted in cessation of
flow in the femoral artery. This resulted in thrombosis and
embolization. This was managed successfully; however,
because of the consequence o f this experience we do believe
Pseudoaneurysms and fistulas induced by arterial puncture 133
it is important to monitor femoral artery blood flow and be
certain that it is not interrupted during the compressive
maneuver.
Dr. Kent. I do not want anyone to think that we are
against ultrasound-guided compression repair. We are
presently using this technique at our institution. I still
believe that spontaneous resolution o f a FA is preferable to
any interventional procedure, particularly if this procedure
uses resources and has the potential to produce complica-
tions. We have identified a group of patients (those patients
with FAs less than 6 cm s) in whom spontaneous throm-
bosis is very likely to occur. I do not believe ultrasound-
guided compression repair should be used as the initial
treatment in these patients. Our present protocol includes
an initial attempt to compress FAs that are greater than 6
cm 3, as well as FAs in patients who are given anticoagulant
medication.
We have attempted to compress four very large
aneurysms. We were unsuccessful in three out of the four.
All of these patients had very large aneurysms, and they all
had large hematomas. In two of the patients we were not
able to achieve adequate compression because of patient
discomfort; and in the third patient, every time we
occluded the neck of the aneurysm, we occluded the
femoral artery. I am certain there is a learning curve. I hope
that we and other institutions can duplicate your results,
particularly in the patients that have these very large,
painful aneurysms.
Dr. McCann asked me about the long-term fate of an
AVF. Unforttmately, we do not know in how many late
complications will eventually develop in patients with
AVF. One of the late complications that can develop is
congestive heart failure. Because these patients usually have
myocardial disease, the insidious onset of congestive heart
failure could be harmful. If an AVF does not close after
several months of observation, repair of the fistula may be
indicated, particularly if long-term follow-up cannot be
assured.
As Dr. Kresowik outlined, the FAs in his study were
derived from a group of patients who were screened by use
of duplex scanning. The FAs in our study were discovered
in patients who had symptoms. These are two different
populations, and I am sure this, at least in part, explains
why his rate of spontaneous thrombosis was greater than
ours. We did not specifically examine the relationship
between hematoma size and the potential for FA throm-
bosis. FAs expanded in every direction. We have noted that
both FAS and AVFs are more often associated with
puncture below the femoral bifurcation.