Continuing Education Activity

This activity reviews the minimally invasive medical therapy known as uterine artery embolization (UAE). It discusses which patients may or may not benefit from this therapy. It reviews how an interprofessional team works together for these patients.

Objectives:

• Know indications, contraindications, and possible alternatives to UAE.
• Explain patient preparation and some of the procedural technique for UAE.
• Identify known complications of UAE.
• Recognize clinical benefits that can be achieved by UAE with the help of an interprofessional team evaluating, treating, and educating the patient on the procedure.

Introduction

Embolization of uterine leiomyomata (fibroids) has been performed since 1995. Because the embolization procedure is performed either solely or primarily through the uterine artery, the procedure is often referred to as uterine artery embolization (UAE).  Some authors prefer to refer to the procedure as uterine fibroid embolization or UFE, but UAE has indications other than for treatment of fibroids, so the term UAE is preferred here.

The procedure's technical success is defined as delivery of the embolic particles to their intendend distention.  Technical success of bilateral UAE by experienced physicians is about 95%. 

The goal of therapy is to relieve symptoms, and not all women whose procedures are deemed to be a technical success remain symptom free, as discussed later.  Symptom improvement from fibroid tumor therapy is correlated with tumor death.  A retrospective review of 84 fibroids treated over one year showed an average decrease in the size of 60%.  Submucosally located fibroids showed the greatest shrinkage, while subserosally located fibroids showed the least.  Ten percent of the fibroids increased in size.[1][2][3]

Anatomy and Physiology

The uterine artery is a branch of the internal iliac artery (IIA).  The IIA has two divisions, each with multiple branches:

Anterior Division

• Obturator artery
• Superior vesical artery
• Vaginal (females)/inferior vesical (males) artery
• Middle rectal artery
• Internal pudendal artery
• Inferior gluteal artery
• Uterine artery (UA)

Variations in this organization can occur.  For example, the vaginal artery may arise from the uterine artery itself.

Posterior Division

• Superior gluteal artery
• Lateral sacral artery
• Iliolumbar artery 

The uterine artery is typically the first branch of the IIA anterior division, whereas the dominant (largest) branch of the IIA anterior division is usually the inferior gluteal artery.  Five distinct IIA branching patterns have been described.  The most common is as described above.  The second most common is a trifurcation of the IIA into the UA, the IIA anterior division, and the IIA posterior division.

The UA has three segments.  Its first segment descends along the pelvic sidewall to the broad ligament.  The second segment courses transversely toward the abdominal midline.  As it reaches the uterus, it again turns, and the third segment ascends the uterine sidewall.  Thus, the segments are termed the descending, transverse, and ascending segments.  The descending segment has no side branches.  A branch termed the cervicovaginal branch usually arises from the mid to distal transverse segment, but it can arise from the ascending segment.  From the ascending segment, numerous branches course over the surface of the uterus and in turn give rise to numerous arterioles that perforate the myometrium.  Arterioles supplying individual fibroids are usually too small to be seen or to be catheterized.

Communication (anastomosis) between the UA and the ovarian artery (OA) can be demonstrated in almost half of female cadavers, but during angiography this anastomosis is visible about 10% of the time.  A classification for OA-UA anastomoses has been proposed. 

• Flow from the OA into the descending or transverse UA is the most common (type 1).
• An OA branch can bypass the main UA to an ascending UA branch supplying a fibroid (type 2).
• In about 1 out of 20 women, an OA is absent, and the ovary itself is supplied by the UA (type 3).

Conversely, the OAs (instead of the IIA) rarely can supply the uterus directly with no UA derived from the IIA (which is not alotted a type number but is similar to type 1).  There are other rarer variants as well. The OAs  typically arise from the aorta a few centimeters below the origins of the renal arteries.  Their courses are often too tortuous to insert a standard catheter for more than a few centimeters.  The importance of the OAs in the management of patients underoing embolization is discussed below.

Indications

The success of UAE depends on appropriate patient selection, which requires an investigation of multiple pertinent positive and negative factors. [4][5]  Patient symptom control and post-procedure satisfaction are highest in "ideal" candidates, which are women with no contraindications to UAE and with all of the following characteristics:

• Heavy regular menstrual bleeding or dysmenorrhea associated with intramural fibroids
• Premenopausal
• No desire for future pregnancy

Fibroids characteristically cause heavy bleeding cycles, not irregular bleeding.  Increased frequency of bleeding (fewer than 21-day cycles) and increased duration of bleeding (greater than 10 days) are more likely to be related to endometrial pathologies such as polyps or cancer.

Instead of or in addition to bleeding, fibroids can cause "bulk-related" symptoms:

• The sensation of pressure in the lower abdomen, legs, and back
• Dysuria (i.e., nocturia, polyuria, incontinence)

Prospective cohort studies and randomized controlled trials (RCT) of women who have undergone UAE have found a long-term improvement in bulk-related symptoms, [5][6] but bulk-related symptoms in general are not as successfully treated as bleeding.

There are other indications for UAE.

Adenomyosis

A retrospective review reported that 76% of 511 women having undergone UAE for the treatment of adenomyosis reported significant improvement, [7] and other studies with smaller patient groups have also claimed high rates of improvement.  However, there has not yet been a study that has included a control group to account for a placebo effect or to enable determination of a number needed to treat (NNT).  Patients with adenomyosis can also complain of worsened pain after UAE.

Postpartum Hemorrhage (PPH)

Whether UAE is viewed as the first or second line treatment (i.e., after incomplete/ineffective surgical ligation) for PPH tends to be institution and/or gynecologist dependent.  Factors predicting the need for UAE as the first line before surgery include abnormal placental attachment and abnormal clotting function (e.g., as determined by serum fibrinogen/INR).  Performing UAE after a failed uterine artery ligation is more difficult. 

Persistent bleeding after surgical treatment can indicate the source of bleeding to be an extrauterine arterial source that may be better identified with arteriography.  In one case series, arterial embolization was successful in 10 of 11 cases after failed surgical ligation therapy.

Prophylaxis of Intra-Operative or Post-Operative Bleeding

In some institutions, gynecologists request UAE as a prophylactic measure in an attempt to reduce operative hemorrhage, such as for:

• resection of benign or malignant uterine tumors
• before fetal delivery in the setting of placenta accreta.

Failed Medical Therapy or Poor Candidate for Other Therapies for Fibroids

UAE also becomes a relevant consideration if a patient without contraindications to the procedure declines or has already attempted other therapies to treat bleeding or pelvic bulk symptoms (e.g., hysterectomy, myomectomy, endometrial ablation, gonadotropin-releasing hormone agonists, or focused ultrasound).

An in-depth discussion of medical treatment for fibroids and abnormal uterine bleeding is beyond the scope of this article, but a brief overview is included here.  A Cochrane database meta-analysis of randomized trials concluded that nonsteroidal anti-inflammatory drugs are more effective than placebo at decreasing menstrual pain and heavy menstrual bleeding but are less effective than tranexamic acid or the levonorgestrel-releasing IUD than placebo.

Gonadotropin-releasing hormone analogues (GnRHa) can shrink fibroids and limit menorrhagia but usually initially worsen symptoms.  Because they cause osteopenia with long term use, GnRHa can only be used for 2 to 6 months and when there is a clear therapeutic goal such as scheduled surgery or when menopause is incipient (Friedman 1991). 

Gonadotropin-releasing hormone antagonists (GRHa) avoid the symptom flare that is associated with GnRHa but also cause osteopenia with long term use.  Rebound fibroid growth occurs after cessation of both GnRHa and GRHa.

Randomized trials have shown that the progesterone-receptor antagonists (PRa) mifepristone and ulipristal acetate decrease fibroid symptoms and reduce fibroid volume without osteopenia.  A woman undergoing therapy with a PRa cannot become pregnant, which is a goal for some women simultaneously wanting therapy for fibroids.  

Contraindications

Different organizations have different lists of "absolute" and "relative" contraindications to UAE.  The following list roughly orders contraindications from strongest to weakest.[4]

• Viable active pregnancy
• Active endometritis
• Malignancy of the uterus/cervix without concurrent surgical treatment planned
• Postmenopausal patient with bleeding of undiagnosed etiology
• Fibroid that is already infarcted (based on MRI)
• Fibroid that is smaller than 1 cm
• Fibroid with pedunculated morphology (such as stalk width less than 50% of the maximum width; some people use different percentage cutoffs)
• Fibroid located in the cervix
• Concurrent use of a GnRH agonist
• Prior pelvic radiation therapy
• Immunocompromised state
• Fibroid growth resulting in overall uterus size equivalent or greater than the expected size at 24 weeks gestation (uterus craniad aspect reaching the umbilicus)
• Severe contrast allergy
• Severe renal insufficiency not receiving dialysis therapy
• Uncorrectable coagulopathy

Of some controversy:

• Patient desire for future pregnancy
• Adenomyosis

For both of these topics, the 2014 SIR guidelines state that in the absence of further evidence, neither should be seen as a contraindication but only as a point of discussion.  The 2009 SIR guidelines listed patient desire for future pregnancy as a relative contraindication and were ambivalent on adenomyosis, only stating that "extensive endometriosis or adenomyosis" could result in UAE failure.  Evidence regarding these issues is discussed elsewhere in this article.[8][9][10]

Gonadotropin-releasing hormones that may have been prescribed as medical therapy should be stopped at least three months before UAE because they constrict the uterine arteries and may preclude their catheterization. 

If the stalk of a pedunculated fibroid is much less than the overall fibroid diameter, then laparoscopic resection can avoid the chance of the fibroid sloughing and forming a loose necrotic tissue mass in the abdomen that can become infected and form an abscess.  A fibroid that is very large and has an extensive blood supply is unlikely to be successfully embolized, whereas a fibroid that is very small is unlikely to be the source of symptoms.  Fibroids of the cervix are often able to be removed surgically via a vaginal approach.

Equipment

In general, tumors are embolized with particles, which are more likely than coils or glue to reach arterioles and capillaries and induce tissue necrosis.  Several types of particles are available, including tris-acryl gelatin (TAG) spheres, non-spherical polyvinyl alcohol (PVA), and spherical PVA.

Regarding the size of particles used, no clinical difference was found between the use of 500 micron or smaller PVA compared to 500 to 700 micron PVA particles in a RCT.  However, it is generally believed that smaller particles result in better infarction rates.  This can be desireable or undesireable depending on the situation, because particles can reflux within an artery and cause "inadvertent" embolization of other nearby tissues (such as the ovary).  Therefore, large (700 to 900 micron) particles may be chosen if there is a visualized ovarian artery anastomosis, whereas small (300 to 500 micron) particles may be chosen when the microcatheter has been successfully placed deep within the uterine artery without a visualized uterine artery-ovarian artery anastomosis to attempt better fibroid infarction.  A RCT comparing TAG microspheres to nonspherical PVA particles found no difference in short-term clinical outcomes.  A follow-up RCT by the same physicians found a statistically significant increase in fibroid infarction rates with TAG compared to PVA and a trend toward improved short-term clinical outcome with TAG.  A third RCT comparing PVA and TAG had eight treatment failures in 27 patients with PVA and one treatment failure in 26 patients TAG, but the physicians used smaller particle sizes for TAG than for PVA, and the study was performed under the conflict of interest that the TAG manufacturers paid for the study.  There has not been a RCT comparing clinical symptoms after using TAG and PVA of the same size particles. [11][12][13][14][15]

Preparation

The possible causes of the symptoms should be investigated in order to exclude pathologic or anatomic contraindications to UAE (as discussed above).

Cancer Screening 

The patient should be current in her gynecologic screening for cervical cancer.  American College of Obstetrics and Gynecology (ACOG) recommendations include HPV testing at age 30 thereafter every 3 years if a woman is HPV-positive and every 5 years if HPV-negative.  Papanicolaou smears are not recommended after age 65 unless there is a special reason to continue.  Women with any endometrial abnormalities on imaging, particularly women over age 45, should undergo endometrial biopsy before UAE, because UAE is not an approved therapy for the treatment of endometrial cancer. 

Serum Parameters to Evaluate

• Electrolytes
• Estimated glomerular filtration rate (eGFR)
• Complete blood count (CBC)
• Platelet count greater than 50,000 per microliter
• International normalized ratio (INR) less than or equal to 1.5
• Beta-human chorionic gonadotropin for pregnancy exclusion

Imaging

MRI provides a significant amount of additional information compared with ultrasound (size and position of fibroids, adenomyosis status) and should be performed prior to UAE.

Informed Consent

The 2009 SIR guidelines stress the importance of informed consent before UAE:

"Any and all treatment options should be presented with the understanding that, for patients interested in gynecologic alternatives, additional consultation will be required. Choice of embolic agents, routes of vascular access, and the anticipated use of an arterial closure device should be discussed. Descriptions of the expected process of admission to the hospital (and by whom), the immediate post-uterine artery embolization recovery period, convalescence and medications at home, and the time expected away from work should also be made.

Written materials given at the time of consultation can enhance this process. Contact information should be given to patients so they can access a responsible physician on a 24-hour availability basis if they perceive that a problem has arisen." [15]

The 2014 SIR guidelines state:

"Interventional radiologists should inform patients about...the possibility of a missed diagnosis of cancer and a delay in definitive treatment." [16]

These recommendations are set with the intention to preserve patients' rights of autonomy.

Pre-Procedure Care

Blood-borne pathogens can colonize UAE-induced necrotic tissue.  The 2018 SIR guidelines recommend that antibiotics, such as 1 g cefazolin or 500 mg levofloxacin, be given IV to reduce the chance of abscess development. [17]

Some physicians prescribe pain medication (e.g., ketorolac 30 mg IV) prior to the procedure. 

If the patient has risk factors for a difficult-to-intubate airway (e.g., screens positive on a sleep apnea questionnaire), then precautions should be taken (e.g., having the patient monitored by capnography at all times) to prevent arrest from respiratory depression from sedation.  Some states have laws requiring capnography, and the 2013 SIR standards of practice recommend it. [18]

Post-Procedure Care

There are a number of clinical factors that may need to be addressed after the procedure. 

Pain

Options for management include NSAIDs and/or opioids.

One regimen includes:

• Ketorolac 30 mg given before the procedure and after, then 30 mg every 6 hours intravenously (IV) as an inpatient
• Ibuprofen scheduled during post-procedure observation and opioid of choice available PRN for several days after discharge

Puncture site pain or leg pain is common, which can be from injury of the medial femoral cutaneous nerve.  It usually responds to NSAIDs and time.  For severe pain thought to be related to nerve injury, gabapentin can be prescribed. 

If a patient has persistent (longer than 1 week) abdominopelvic pain, then imaging workup for non-target embolization should be considered.

Chronic pain syndromes (such as inflammatory bowel disease) may worsen after UAE.

Pulmonary Emboli Prevention

Sequential compression devices and prophylactic enoxaparin should be considered.  Patients have developed and died from pulmonary emboli after UAE.

Bleeding

Vaginal bleeding in the first week after therapy could indicate a developing endometrial infection.  A patient with endometritis should have flu-like symptoms and fever.  Bleeding after the first week with cramping could indicate that there is a sloughing or sloughed fibroid. 

Nausea

Nausea can be treated with ondansetron 4 mg IV after the procedure continued every 6 hours until discharge. 

Hypertension

Blood pressure often increases in the short term after embolization.  Potential sequelae of hypertension include pulmonary edema, myocardial infarction, and stroke. 

Post Discharge Communication

Communication between the patient and the physician or physician representative should continue during the first week.  Physicians may wish to schedule a return clinic visit at 1 to 3 months.

Technique or Treatment

The technique for UAE is described in brief below:

• Access the common femoral artery and insert a 5 or 6 French sheath.
• Obtain an arteriogram to demonstrate the arterial anatomy and find the UA origins.
• Place a catheter with a selective tip shape into the IIA anterior division.
• Advance a smaller catheter (a microcatheter) coaxially through the selective tip catheter and then into the UA horizontal segment beyond any cervicovaginal branch(es).
• Obtain another arteriogram through the microcatheter to identify any collateral branch(es) to the ovaries/bladder/vagina that might affect specific choices made in particle delivery.
• Perform embolization.
• Obtain final images showing the state of intra-arterial contrast flow to the uterus.
• Repeat the procedure on the contralateral side.
• Obtain hemostasis at the site(s) of the arterial puncture(s).

Variables are many and include:

• Bilateral femoral artery punctures (to decrease procedure time) vs. single femoral artery puncture (to decrease the likelihood of artery access site complication)
• Initial or final overview arteriogram from the aorta (to identify an OA contribution/variant anatomy) versus only selective common iliac arteriograms (to reduce contrast load)
• Delivery of medications during the procedure (such as nitroglycerin to prevent/treat arterial spasm, lidocaine intra-arterially or ketorolac intravenously to reduce post-procedure pain)
• Embolization endpoint (contrast stasis in the UA for 5 to 10 heartbeats, versus visualization of contrast reflux toward the UA origin, versus subjectively achieving a UA "pruned" appearance)

Complications

The overall complication rate differs by published report and can be as high as 40% to as low as 5%.[19][20][21]

Major Complications

• Death (less than 1/1000)
• Need for surgery due to a complication of the procedure, such as arterial perforation (2 to 3 in 100)
• Abscess/other serious intrauterine infection (1 in 100)
• Pulmonary embolism
• Subsequent pregnancy-related complications, such as spontaneous uterine rupture at the mid-posterior wall during subsequent pregnancy or placenta accreta
• Premature ovarian failure
• Buttock/leg ischemia

Death

Deaths have occurred from UAE via uterine infection/sepsis and pulmonary embolism.

Infection-Spectrum Findings

Postembolization syndrome is the most common adverse event and includes pain, low-grade fever, malaise, nausea, and appetite loss.  Symptoms and signs of bacterial endometritis include pain, fever, vaginal discharge, and/or leukocytosis.  Initial therapy includes intravenous antibiotics and medications to reduce pain and inflammation.  Noninfectious endometritis can manifest as watery vaginal discharge, fever, and/or leukocytosis occurring days to weeks after the procedure.  Imaging findings do not reliably differentiate infected fibroids from non-infected fibroids immediately after UAE.

Amenorrhea Spectrum Findings

Amenorrhea of any kind occurs in less than 10% of cases when considering patients treated for all UAE indications and is much more likely in patients older than 45 to 50 years old.  It is usually limited to a few cycles and is not typically associated with increased FSH levels or menopausal symptoms.  Premature ovarian failure is defined as the presence of persistent amenorrhea, increased FSH levels, and symptoms of menopause (night sweats, mood swings, irritability, and/or vaginal dryness).

Expulsion of Fibroids

Expulsion is defined as a detachment of fibroid tissue from the uterine wall and its subsequent transvaginal passage, most commonly occurring with submucosal fibroids that have thin points of attachment.  This process may present with uterine contractions, abdominal pain, fever, nausea, vomiting, and vaginal bleeding or discharge.  Surgical intervention may be necessary to relieve persistent discomfort and remove the tissue.

Angiography Complications

These include hematoma, dissection, pseudoaneurysm, and contrast-induced nephropathy.

Complications Other

• Patients may have persistent pain.  Contrast-enhanced CT or MRI may be used in these circumstances to assess for findings of non-target embolization or for fibroid sloughing into the pelvis.
• The radiation doses for UAE are an order of magnitude less than those required to produce radiation burns.
• There have been reports of petechial rash on the torso and limbs (1% to 7%) after UAE.

Clinical Significance

For women with ideal procedure selection characteristics, a high symptom control rate, procedure satisfaction rate, and self reported quality of life rate can be achieved for up to 10 years after treatment, as discussed below.

The SIR guidelines indicate that about 90% of women should experience a reduction in bulk symptoms and that more than 90% should expect the elimination of abnormal uterine bleeding.

Causes of UAE failure include but are not limited to:

• Poor patient selection
• Large fibroids
• Inability to catheterize the uterine arteries
• Incomplete artery embolization
• Arterial spasm leading to poor flow of embolic arterial 
• Clumping of embolic material leading to false embolization endpoint
• Uterus anatomic variants
• Presence of important collateral arteries (e.g., ovarian arteries) feeding the uterus
• Coexisting adenomyosis or leiomyosarcoma
• Recanalization of embolized arteries

Trials and Guidelines for Treatment of Fibroids

There have been seven RCTs including 793 women comparing UAE versus surgery for fibroids. Three trials compared UAE against abdominal hysterectomy, two against myomectomy, and two against either type of surgery. In a meta-analysis of these patients, Gupta et al. (2014) found no significant difference between UAE and any type of surgery concerning patient satisfaction or major adverse events.  UAE had a higher rate of minor complications and need for additional surgical interventions.  Gupta calculated that within 2 years of initial therapy, 7% of women require further surgery after hysterectomy or myomectomy, whereas 15% to 32% require further surgery after UAE.  UAE patients were found to have a lower likelihood of requiring a blood transfusion, shorter procedural time, shorter length of hospital stay, and faster resumption of usual activities.[22]

The first randomized controlled trial (RCT), Embolization versus Hysterectomy for Symptomatic Uterine Fibroids (EMMY), enrolled patients between 2002 and 2004. This multicenter trial has the longest follow-up to date (ten years).  Enrolled patients had both fibroids and menorrhagia and were all eligible for hysterectomy.  One hundred seventy-seven patients were randomized to UAE (n = 88) or hysterectomy (n = 89).  Ten years after treatment, 33% of UAE patients had undergone secondary hysterectomy for recurrent symptoms (8% of initial hysterectomy patients also required a second procedure), but there were no significant differences between hysterectomy and UAE with respect to successful eradication of pain, bulk-specific complaints, complications, or overall satisfaction.  The study authors concluded that hysterectomy is better at achieving a certain cure for bleeding problems but that patients with the appropriate combination of indications and lack of contraindications should be allowed to choose UAE instead of a hysterectomy.[23]

The Randomized Controlled Trial of Embolization versus Surgical Treatment for Fibroids (REST) is the largest study to date. It was a multicenter RCT comparing all surgical interventions to UAE and achieved 80% power with 1-year follow-up.  It found no statistically significant difference between UAE and surgery in quality of life or major complications.  UAE had a higher minor complication rate (34% versus 20%) but also had a statistically significant reduced length of hospitalization, 24-hour pain score, and cost.[24]

In one study, patients having intramural fibroids larger than 4 cm were randomized to UAE (n = 58) or to myomectomy (open or laparoscopic according to surgeon preference; n = 63) and followed for two years. This is the primary study used as the rationale for making myomectomy the first-line invasive therapy in women who still wish to conceive, as 50% of UAE versus 78% of myomectomy patients were able to conceive.  UAE in this study had an increased calculated relative risk compared to myomectomy for (1) inability to conceive and (2) miscarriage (95% CI of [1] 1.11 to 4.44 and of [2] 1.25 to 6.22).  However, Gupta (2014), included Mara's patients with others having undergone myomectomy in RCTs and deemed the evidence that myomectomy provides a fertility benefit over UAE to be "low-quality evidence" that "should be regarded with extreme caution."  Another systematic review on myomectomy concluded that there is currently insufficient evidence that myomectomy improves fertility compared to UAE whether by the laparoscopic or open approach.[22]

A Cochrane systematic review of 15 prospective cohort studies and RCTs found that loss of ovarian function after UAE occurred primarily in women older than 45 years of age.

Despite the conclusions of Gupta (2014), guidelines by gynecological societies in France and the United States recommend myomectomy over UAE for symptomatic intramural and subserosal fibroids in women who wish to have a subsequent pregnancy.  In the United Kingdo, the Royal College of Obstetricians and Gynecologists does not have guidelines on the subject, but the British National Institute for Health and Care Excellence guidelines do not favor UAE or surgery over the other even for patients considering future pregnancy.

Trials and Guidelines not for the Treatment of Fibroids

Outside of the setting of treatment of fibroids, there are no RCTs comparing UAE to other treatments and almost no society or national guidelines regarding the use of UAE over other therapies.

For the control of PPH, emergent UAE has about a 90% to 95% success rate (with failures usually requiring hysterectomy), and the morbidity of UAE is less than with emergent laparotomy.

Enhancing Healthcare Team Outcomes

Healthcare Team

Uterine artery embolization is usually performed by an interventional radiologist, but the gynecologist and obstetrician should also be able to help educate the patient about pros and cons of this technique compared to surgical and/or medical treatments that the gynecologist or other primary care physician can offer.  The patient should be informed that infertility may result and that subsequent pregnancy may be complicated (such as by premature birth). [25][26](Level V and see above section, Level I).  The nurse taking care of the patient post-procedure may have to administer physician-prescribed analgesics and other medications.  This nurse also observes for signs of complications, primarily those resulting from arterial puncture.  

Outcomes Summary

Most randomized (Level I, see above section) and non-randomized series report high odds of clinical success and subsequent pregnancy success for UAE, although UAE is not necessarily always superior to surgical techniques.  UAE is best suited for women who prefer a minimally invasive procedure to surgery or other methods of treatment and are willing to accept the risks of UAE. [27][28]