Uterine fibroid embolization (UFE) is a widely practiced procedure for the treatment of symptoms due to uterine leiomyomata. In recent years, studies have contributed a tremendous amount of knowledge about the impact of selectively embolizing both uterine arteries, demonstrating UFE to be an extremely safe and effective therapy for the treatment of symptomatic uterine fibroids.
is a Clinical Instructor and
is an Associate Professor of Radiology in the Division of
Interventional Radiology, University of Pennsylvania Medical
Center, Philadelphia, PA.
Uterine fibroid embolization (UFE) is a widely practiced
procedure for the treatment of symptoms due to uterine leiomyomata.
Originally performed as uterine artery embolization (UAE) for the
treatment of acute uterine bleeding, the majority of procedures are
now performed for bleeding or bulk symptoms due to fibroids, hence
the term UFE. Since the original description of the technique by
we have gained enormous knowledge about the impact of selectively
embolizing both uterine arteries, demonstrating UFE to be an
extremely safe and effective therapy for the treatment of
symptomatic uterine fibroids.
Prior to the discovery of UAE for the treatment of fibroids,
embolotherapy for benign disease was reserved for bleeding
complications associated with gynecologic procedures, obstetrical
emergencies, and arteriovenous malformations. The first reports of
percutaneous transcatheter embolization of postpartum hemorrhage
and postoperative gynecological bleeding was in 1979.
The first pelvic embolization specifically for the treatment of
symptomatic uterine fibroids was performed in France initially as a
treatment preceding myomectomy to reduce blood loss.
At that time, Ravina et al
discovered that many of the patients canceled their surgery due to
almost complete disappearance of their symptoms. Goodwin et al
published the first U.S. results in 1997.
In October 2000, the Society of Interventional Radiology (SIR)
performed a survey of its members performing UFE.
Information from 330 institutions revealed that approximately
10,500 procedures had been performed worldwide, including 8644 in
the United States, at both academic and nonacademic sites in 47
states and the District of Columbia in addition to 17 foreign
countries. The number of women currently treated worldwide probably
has doubled since the last SIR survey.
Uterine leiomyomata or fibroids are the most common tumor found
in women. They are benign clonal tumors arising from the
smooth-muscle cells of the uterus. Approximately 25% of
premenopausal women suffer from fibroids,
while the overall prevalence of these tumors could be as high as
The incidence of fibroids in African-American women is three times
that of Caucasian women.
Fibroids may occur at any age, but are most common in women over
the age of 40 years. After menopause, fibroids usually regress in
size due to the lack of hormonal stimulation, which may result in
More than 50% of the patients with fibroids are asymptomatic.
When present, there are usually three types of symptoms associated
with fibroids: bleeding, pelvic pressure (pain, urine frequency),
and infertility. The location of the leiomyoma is classified as
subserosal, submucosal, or intramural and is probably more
important than size in causing the symptoms (Figure 1).
However it may be difficult to determine its position within the
wall of the uterus as the tumor increases in size. Prolonged and
intensive cyclical menstrual bleeding (menorrhagia) is the most
frequent clinical presentation of submucosal fibroids. Bulk
symptoms are usually caused by serosal fibroids that compress the
nearby pelvic structures, resulting in the associated symptoms:
bladder (urinary frequency), sciatic nerve (pain), or bowel
(constipation). Uterine fibroids may impact fertility,
usually due to a submucosal fibroid distorting the endometrial
In infertile women, uterine cavity distortion by fibroids was found
in 56% of hysterectomy specimens.
The patient's clinical history is very important in evaluating
patients for UFE. The typical patient is a premenopausal woman
suffering from prolonged regular bleeding during menses without
bleeding between periods (menorrhagia). Irregular or intermenstrual
gynecological bleeding (menometrorrhagia) should raise the
suspicion of another pathology. The differential diagnosis of
vaginal bleeding includes fibroids, adenomyosis, endometrial
sarcoma, polyps, and other endometrial disorders. An endometrial
biopsy is generally recommended when the complaint is bleeding, but
in cases of menometrorrhagia, it is essential to exclude
endometrial carcinoma. The initial diagnosis is confirmed by
magnetic resonance imaging (MRI) or ultrasound. MRI is advocated by
many as the most reliable and reproducible modality. It allows for
the differentiation of the fibroids from other pathological
conditions, such as adenomyosis (Figure 2), potentially predicts
outcome based on vascularity,
and delineates precisely the fibroid's location within the
One concern in evaluating a patient for UFE is the risk of
embolizing a leiomyosarcoma. In premenopausal women, the incidence
of the leiomyosarcoma is very low with a rate of 0.13% to 0.29%.
A biopsy is not helpful to exclude this malignancy and it is almost
indistinguishable from a leiomyoma on ultrasound or MRI.
Unfortunately, a high growth rate of the uterine mass doesn't
necessarily predict malignancy
; however, symptoms such as fatigue and weight loss may raise
suspicion. The risk of inadvertently embolizing a sarcoma is very
small with only two cases reported in the literature.
Both of these cases presented as therapeutic failures with a
uterine mass that continued to grow after the embolization, making
follow-up imaging important.
Uterine and ovarian artery anatomy
The main portion of the uterine artery can be divided into three
The first part runs close to the lateral pelvic sidewall, the
second part runs in the parametrium, and the third part, called the
marginal artery, runs along the lateral margin of the uterus toward
the uterine horn. The main branches from the uterine artery in the
order of their branching are: the ureteric branch supplying the
distal portion of the ureter (usually not visible); the
cervico-vaginal branch supplying the vagina and cervix; the uterine
intramural branches supplying the myometrium; and the terminal
fundal branches supplying the fundus. In 9% of cases, the
cervico-vaginal branch comes off of the internal iliac. Finally,
the medial tubal and ovarian branches arise from the terminal
uterine artery. In approximately 10% of cases, the blood to the
uterine fundus is supplied by the ovarian arteries. Rarely, the
ovarian artery may supply blood to most of or all of the
Generally the ovarian arteries arise from the ventral surface of
the aorta in approximately 83% of cases, usually as a single origin
but they may have multiple origins.
In 17% of cases, the ovarian arteries may arise from the renal
arteries or other pelvic vessels. The blood supply to the ovaries
comprises flow from the uterine artery from branches that create a
tubal and ovarian arterial arcade and anastomose with the lateral
tubal and ovarian branches from the ovarian artery. In 40% of
cases, the ovarian artery solely supplies the ovary while there is
a shared supply with the uterine artery in 30% of cases. In 10% of
patients, the uterine artery is the main supply to the ovaries. The
uterine artery supplies the fallopian tube in 60% of patients while
there is shared supply in another 56% of cases. In only 4% of cases
does the ovarian artery solely supply the fallopian tube.
Knowledge of these anastomoses is important since they provide for
the collateral blood flow that may result in the failure of
percutaneous embolization or ovarian nontarget embolization.
Several different techniques utilizing a variety of catheters to
access the uterine arteries have been described.
Due to extensive intrauterine collateral flow, bilateral uterine
artery embolization is necessary to achieve fibroid infarction.
Catheterization is typically performed by the right common femoral
artery approach with the use of a 4F or 5F Cobra catheter (Boston
Scientific Corp., Watertown, MA) or Levin catheter (Cook Inc,
Bloomington, IN) (Figure 3). Some practitioners use a coaxial
microcatheter to access the uterine artery to reduce the potential
for spasm. Spasm may result in nontarget embolization to the
ovaries or in premature termination of the embolization.
Ipsilateral catheterization is commonly performed using the
Waltman loop technique. As originally described by Waltman
in 1973, the catheter is looped back over itself by selecting a
suitable visceral vessel and pushing the catheter up into the
aorta. This technique works well with braided catheters; however,
based on our experience, newer hydrophilically coated catheters are
more susceptible to kinking while forming the Waltman loop. We have
described a modification of a Cope suture technique to form the
Waltman loop to access the contra- and ipsilateral sides.
Others use the "Roberts" catheter (Cook Inc., Bloomington, IN),
which is specifically designed for UFE with a long reversed
secondary curve that tapers from 5F to 3.8F. Spies et al
advocate using a bilateral femoral approach to reduce the radiation
dose to the ovaries by simultaneously injecting the embolic agent
into both uterine arteries.
The main goal of the procedure is to embolize the perifibroid
plexus where the median size of these arteries is approximately 500
For that reason, most investigators use particles >350 µm.
Several European groups advocate the use of the smaller 150 to 300
Most reports describe the use of polyvinyl alcohol (PVA) as the
embolic agent. The advantages of PVA are that it has been used
safely as an embolic agent for more than 20 years and is relatively
inexpensive. The main drawback of these particles is their
propensity to aggregate in the syringe or catheter, causing
catheter blockage or potentially proximal occlusions.
Recently, Embospheres (Biosphere Medical, Rockland, MA)
particles have gained popularity. Their main advantage over PVA is
a hydrophilic feature that prevents the aggregation and facilitates
delivery of the particles.
Some authors advocate embolization with Gelfoam (Pharmacia &
Upjohn Co., Kalamazoo, MI) since it is a temporary embolization
Because it allows reperfusion to the uterus after initial
embolization, it has the theoretical benefit of potentially
limiting some of the complications of UFE and may help maintain
fertility in patients. Stancato-Pasik et al
described a series of 12 patients who underwent embolization with
Gelfoam and resumed normal menstruation; 3 of them eventually
carried a pregnancy to term and had a normal delivery.
The end point of the embolization generally depends on the
embolization material and personal experience of the operator
(Figure 4). As a rule, the Embospheres propagate deeper than PVA
because of the elastic features of the particle. For this reason,
most researchers recommend a much less aggressive approach to
embolization than the use of PVA. Moderate flow is maintained to
the uterine vessels while achieving a "pruned tree" appearance to
the fibroids. This is opposed to the "stagnant" flow that is
usually achieved with PVA.
Since UFE is done under fluoroscopic guidance, radiation
exposure is a concern, especially in women who are of reproductive
age. Several published studies demonstrated that the overall dose,
22.34 cGy, is comparable to the other diagnostic studies and far
below the doses that are used for treatment of the pelvic
malignancies (up to 3500 cGy).
With proper collimation and the use of pulsed fluroscopy, the
estimated absorbed ovarian dose could be decreased to 9.5 cGy
(range 2.21 to 23.21 cGy).
Periprocedural pain control is of the utmost importance since it
represents the major morbidity of the procedure. Pain generally
starts early after the embolization and reaches the highest
severity 24 to 48 hours after the embolization. Most pain protocols
use a combination of opioids, such as a oxycodone derivative, and a
nonsteroidal anti-inflammatory (NSAID), such as ibuprofen.
Successful pain control potentially allows this procedure to be
performed on an outpatient basis. Early studies attempting to
perform UFE as an outpatient procedure reported that 15% of
patients returned to the hospital for pain control.
In a recent article by Siskin et al,
the authors reported 95.9% successful discharge after 8 hours of
postprocedure observation. One should not use intra-arterial
lidocaine in an attempt to reduce pain since it causes a large
amount of spasm.
Postembolization syndrome with severe pain, fever, and an elevation
in the white blood count occurs in as many as 34% of patients.
Two main measurable outcomes are reported in all published UFE
studies: clinical (improvement of symptoms) and radiological
(decrease of the fibroid and uterus sizes). Even though the change
in size of the fibroids and uterus is a more objective measurement,
the goal of the therapy is to improve clinical symptoms. Patients
who have minimal change in the size of their fibroids and uterus
will often have dramatic improvement in their symptoms.
The technical success rate of the procedure is very high: 84% to
Difficulties in accessing the uterine artery or arterial spasm are
the main causes of failure. Besides making it difficult to
catheterize, uterine arterial spasm may also cause early proximal
thrombosis of the uterine artery that may result in immediate
recanalization causing procedural failures.
The fibroids and uterus decrease in size following the
embolization (Figure 5). Overall shrinkage of the dominant fibroid
is 40% to 70% and of the uterus is 40% to 60% (Table). It has been
shown that the uterus continues to shrink over time.
Fibroids vary considerably in their histologic composition. Some
of them are extremely vascular with a significant cellular
component and others are partially degenerated. Jha et al
demonstrated that hypervascular fibroids by MRI criteria decrease
in size more than those that are avascular, which suggests that the
MRI vascularity pattern may help predict fibroid shrinkage after
The overall clinical response of UFE is close to 90% (Table).
Menorrhagia is the most common indication for UFE and has the best
response rate. In all published series, 81% to 100% of patients
For bulk-related symptoms (bloating, pelvic pain, urinary
frequency), the success rate is 61% to 100%. Some of the
inconsistencies in these results may be due to the different ways
that the outcomes were measured.
Lack of response to embolization may be due to perfusion to the
fibroids by ovarian collaterals (Figure 6). As discussed, the
ovarian artery may supply the uterus and fibroids in a small
percentage of patients. For that reason, we complete each procedure
with flush aortography with the tip of the catheter placed at the
level of renal arteries. In cases of an enlarged ovarian artery,
selective catheterization may be performed to demonstrate arterial
supply to the fibroids. Some investigators advocate ovarian
embolization if definite fibroid supply is demonstrated or if the
patient's clinical symptoms do not improve.
Ovarian embolization may increase the risk of ovarian failure.
The effect of the UFE on fertility is still under investigation.
Ravina et al
identified 12 pregnancies in 9 women in their series of 184 women
who underwent UFE. Among the 12 pregnancies, there were 4 full-term
and 3 premature deliveries, and 5 early miscarriages.
Unfortunately, the authors didn't describe how many women tried to
conceive. Due to the risk of ovarian failure, possible
hysterectomy, and devascularization of the uterus, we generally do
not recommend UFE for patients who desire pregnancy when surgical
and medical options are available. If these other options have
failed or are not possible, such as in a patient with numerous
fibroids in whom a myomectomy cannot be performed, UFE may be
Adenomyosis, which is ectopic growth of endometrial glands and
stroma into endometrium, is less common and sometimes coexists with
fibroids. Some investigators originally considered it responsible
for UFE failure.
However, in a recent study in which UAE was performed for 15
patients with adenomyosis, Siskin et al
found that 92.3% of patients reported significant improvement of
symptoms and quality of life at 1-year follow-up. This suggests
that adenomyosis may be treated by UAE; although further
information is needed.
Complications of UFE
The overall complication rate of UFE is very low. Currently
there are 4 known deaths related to UFE: 2 have been reported in
and 2 unreported among >15,000 cases.
Two are from pulmonary embolus and 2 from infection. This compares
favorably to the mortality rate for hysterectomy performed for
benign disease, which is 1 in 1600 operations.
The mechanism of pulmonary embolism may be related to a derangement
of the clotting cascade due to tissue necrosis. Some clinicians
advocate leg pressure stockings to eliminate venous stasis, while
others advocate the use of heparin to prevent deep venous
Complications related to the catheterization are extremely low,
with < 1% reported throughout most of the studies (Table). Most
of these complications are self-limited and do not require
additional therapy. Hematoma is the most common complication but
the incidence is difficult to estimate. Other rare reported
complications include allergy to iodine contrast,
pseudoaneurysm formation, and dissection.
The more commonly reported complications are related to fibroid
expulsion, uterine infection, and nontarget embolization. Fibroid
expulsion is a common event after UFE, with up to a 5% rate in two
The intrauterine necrotic fibroid tissue may become infected and
cause significant morbidity. We recommend MRI for any women with
suspected intracavitary sloughing of a fibroid, since it
demonstrates whether the fibroid is detached from the wall and
lying in the endometrial cavity near the cervix. Patients may expel
the necrotic tissue without any help but often will require
gynecologic assistance. Although patients will have a dramatic
volume reduction with expulsion, suggesting that the process may
accentuate their response to embolization, there remains a risk of
hysterectomy if the debris cannot be removed or if the patient
Infection is another complication that is related directly to
ischemia of the fibroids or uterus. Endometritis and pyometra are
responsible for most postembolization hysterectomies, and this was
the reason for 1 of the reported deaths.
Nontarget embolization to the ovary is possible due to the
ovarian collaterals from the uterine artery. Pelage et al
identified the collaterals between the uterine artery and ovarian
artery in 11% of the patients that underwent pelvic angiogram for a
variety of reasons. In 5% of patients, the ovarian artery can be
absent and the ovaries supplied solely from the uterine artery. It
has been suggested that the embolization material passes into the
ovarian artery through these tubal collaterals. This is supported
by Ryu et al,
who demonstrated that 50% of the patients who had UFE had a
decrease in ovarian Doppler flow.
Many patients may experience amenorrhea for 1 or 2 months after
UFE. Most will have their periods resume. However, a few will go
onto permanent amenorrhea; in most studies the incidence is <5%
The etiology is probably nontarget embolization; however, other
etiologies are possible, such as radiation. A higher rate of
ovarian dysfunction is suspected in patients older than 45 years.
Chrisman et al
demonstrated that 15% of their 66 patients did not resume regular
menses and all of them were older than 45 years of age. In the
subgroup of patients who were >=45 years, 9 of 23 (43%) became
menopausal. Spies et al
conducted a study in which they measured follicle stimulating
hormone (FSH) levels, an indicator of menopause, before and up to 6
months following UFE. They could not identify a statistically
significant difference of the basal FSH change in the whole group,
but when a subgroup analysis was performed for the patients aged 45
and older, 15% of patients had an increase in FSH into the
perimenopausal range. These findings are not surprising, since the
average age of menopause is 51 years. One could speculate that the
ovaries in an older patient may be less tolerant of
While sexual dysfunction may be common after hysterectomy,
there is only 1 report of a patient losing pelvic and clitoral
orgasm after a successful UFE.
The patient's orgasm eventually returned to baseline after a
certain period of time. Trying to avoid embolization of the
cervical-vaginal branch of the uterine artery may prevent this rare
Comparison with surgical methods
There is only one reported prospective Phase II clinical trial
that compares UFE using Embospheres with hysterectomy.
The abstract was presented at the recent SIR annual meeting. The
report demonstrated that UFE with Embospheres gave a similar
improvement in the symptoms when compared with hysterectomy, but
had the advantage of a faster recovery.
In a small nonrandomized retrospective study comparing
myomectomy with UFE,
12% of the myomectomy patients required blood transfusion and 25%
overall suffered from various complications while the UFE group had
no complications or blood transfusion. Hwang et al
demonstrated differences in complication rates of 19.4% versus 4.1%
for the myomectomy and UFE groups, respectively. In addition,
recovery was much faster in the UFE group. Future, prospective
studies are needed to compare these two treatment modalities. The
results from the prospective Boston Scientific Corp.sponsored PVA
UFE versus myomectomy study should be available soon.
In the future, a better understanding of the impact of UFE will
be obtained from the large Fibroid Registry for Outcomes Data
(FIBROID), sponsored by the Cardiovascular and Interventional
Radiology Research and Education Foundation, the research affiliate
of SIR. Its main goal is to establish the effectiveness and safety
of UFE in a prospective longitudinal fashion.
Uterine artery embolization is a minimally invasive alternative
for treatment of symptomatic fibroids. The success rate is
excellent, with a low complication rate and a recovery time that is
much shorter than conventional surgery. Although the experience of
single-center small series makes it difficult to give accurate
rates for many of these reported complications, the incidence
appears to be low. In addition, further research must clarify the
impact of UFE on fertility. We expect the FIBROID registry
initiative to answer many of these questions and further support
the advantages of using UFE for treating women with symptomatic