Technical Results and Effects of OperatorExperience on Uterine Artery Embolization forFibroids: The Ontario Uterine FibroidEmbolization TrialGaylene Pron, PhD, John Bennett, MD, Andrew Common, MD, Kenneth Sniderman, MD, Murray Asch, MD,Stuart Bell, MD, Roman Kozak, MD, Leslie Vanderburgh, MD, Greg Garvin, MD, Martin Simons, MD,Cuong Tran, MD, and John Kachura, MD, for the Ontario UFE Collaborative GroupPURPOSE: Todocument thetechnical resultsandspectrumofpracticeofuterinearteryembolization(UAE) forfibroids in the health care setting in Canada. The effects of interventional radiologists (IRs) experience with UAE onprocedure and fluoroscopy time were also investigated.MATERIALS AND METHODS: The study involved a multicenter prospective single-arm clinical treatment trial andincluded the practices of 11 IRs at eight university-affiliated teaching and community hospitals. Vascular access withpercutaneous femoral artery approach was followed by transcatheter delivery of polyvinyl alcohol (PVA) particles intouterine arteries with fluoroscopic guidance. Technical success, complications, procedural time, fluoroscopy time, andeffects of operator experience were outcomes analyzed.RESULTS: Between November 1998 and November 2000, 570 embolization procedures were performed in 555 patients.UAE was bilaterally successful in 97% (95% CI: 95%98%). Variant anatomy was the most common reason for failure toembolize bilaterally. The procedural complication rate was 5.3% (95% CI: 3.6%7.4%). Of the 30 events, three involvedmajor complications (one seizure and two allergic reactions) that resulted in additional care or extended hospital stay.Procedure time and fluoroscopy time averaged 61 minutes (95% CI; 5863 minutes) and 18.9 minutes (95% CI; 1819.8) andvaried significantly among IRs (P < .001; P < .001). The average 27% reduction in procedure time (20 minutes; P < .001)and 24% reduction in fluoroscopy time (5.1 minutes; P < .001) with increasing UAE experience were significant.CONCLUSIONS: A high level of technical success with few complications was obtained with a variety of operatorsin diverse practice settings. Increased experience in UAE significantly reduced procedure and fluoroscopy time.J Vasc Interv Radiol 2003; 14:545554Abbreviations: IR interventional radiologist, PVA polyvinyl alcohol, UAE uterine artery embolizationUTERINEarteryembolization(UAE)hasincreasinglybeenreportedasanalternativetreatmenttohysterectomyforsymptomaticfibroids. Theresultsof UAE have been reported as clinicalseries inseveral countries includingtheUnitedStates(15), France(68),andEngland(9,10). However, thesestudies involvedtheexperienceof asingleinstitutionorindividual inter-ventional radiologist (IR) and havetended to be retrospective and involvesmall patient groups.Thisstudyisthefirst multicenterprospective clinical study of emboliza-tion therapy to involve the practices ofa large collaborative group of IRs. Theoverall objectives of this study were toevaluate technical success, tolerability,safety, effectiveness, and durability ofthis therapy for symptomatic fibroids.The purpose of this paper is to docu-ment the technical results and theFromthe Department of Public Health Sciences(G.P.), University of Toronto, Toronto; Departmentof Radiology(J.B., R.K., G.G.), St JosephsHealthCare, London; Department of Medical Imaging(A.C), St Michaels Hospital, Toronto; Department ofMedical Imaging (K.S, J.K), Toronto General Hospi-tal, Toronto; Department of Medical Imaging (M.A),MountSinai Hospital, Toronto; DepartmentofDi-agnostic Imaging (S.B.), Rouge Valley Health Systemandthe ScarboroughHospital, Scarborough; De-partment Radiology(L.V.), WilliamOsler HealthCenter, Brampton; Department Medical Imaging(M.S.), TorontoWesternHospital, Toronto; Depart-ment Medical Imaging(C.T.), McMaster UniversityMedical Center, Hamilton, Ontario, Canada. ReceivedNovember18, 2002; acceptedJanuary15, 2003. TheOntario UFE Trial was funded in part by Boston Sci-entific Corporation. Address correspondence to G.P.,Department Public Health Sciences, Faculty of Medi-cineUniversityof Toronto, 100CollegeSt, Rm513Banting Building, Toronto, Ontario, Canada M5G 1L5;E-mail: g.pron@utoronto.caNone of the authors have identified a potential con-flict of interest. SIR, 2003DOI: 10.1097/01.RVI.0000071099.76348.DFClinical Studies545spectrumof practicewithrespect tofibroid UAE therapy in the health caresetting in Canada. The details of prac-tice spectrum that were evaluated in-clude: the hospital setting; equipment;interventional team; andbackgroundand experience of the IR. The technicalaspects that were studied include:technical success; procedural and flu-oroscopictime; materials; proceduralcomplications; and reasons for failure.The effects of IR experience with UAEon procedure and fluoroscopy timewas also investigated because the ther-apywas beingnewlyintroducedatthese hospitals. This report is based onthe procedural details of fibroid UAEperformed on 555 women in eight uni-versity-affiliated teaching and com-munity hospitals.MATERIALS AND METHODSStudy Design and ParticipatingCentersTheOntarioUFETrialwasamul-ticenter clinical study that involvedthe prospective follow-up of a consec-utive series of women undergoingUAE as an alternative to hysterectomyfor symptomatic uterinefibroids be-tween November 1998 and November2000. Treatment wasprovidedby11IRs at eight hospitals; seven hospitalswere university-affiliated teachinghospitalsandonewasacommunityhospital. Approvals for the study wereobtained fromInstitutional ReviewBoards at all institutions.Patient EligibilityThestudygroupincludedwomenwho were referred to IRs for treatmentofsymptomaticultrasound(US)-doc-umented uterine fibroid(s). Symptomsincluded, but werenot restrictedto:menorrhagia, pelvicpain, orbulkre-latedsymptoms(suchasurinaryur-gency or frequency). Patient age or fi-broid size, number, and location werenot criteria for exclusion. Althoughwomen who wanted to have childrenwere not excluded fromthe study,they were further informed of the un-certaineffectsof UAEonconceptionor carrying to full term. Women withactive pelvic infection, undiagnosedpelvic mass, endometrial carcinoma,pregnancy, or renal insufficiency (cre-atinine levels 150 mol/L) were in-eligible for the study.Study Patient CharacteristicsPatients were predominantly white(66%) and the mean age was 43 years(range, 1859 years). More than half ofthe patients (56%) were marriedorwith regular partners and 50%ofpatients had one or more children.Thirty-one percent of patients wishedto retain their fertility. Fifty-seven per-cent of patients reported their generalhealth to be very good or better.Heavy menstrual bleeding was re-portedby80%of patients. Most pa-tients (70%) had multiple fibroids anda minority of patients had undergonesurgical treatment for their fibroidsmyomectomy(14%) andendometrialablation (3%). Pre-UAE uterine and fi-broidaveragevolumeswere680cm3(95%CI: 626734cm3) and293cm3(95%CI: 259327cm3). Meandomi-nantfibroidlengthwas8cm(range,124 cm). The complete details on thiscohort have been reported (11).Interventional TeamsEmbolization therapy was pro-vided by 11 IRs in the Departments ofMedical Imaging or Radiology. Allwere fellowship-trained in vascularand interventional radiology and hadat least 3 years of experience in periph-eral angiography and embolizationtechniques (Table 1). Only two IRswere 100% dedicated to interventionalpractice. Others hadcombinedprac-tices that included diagnostic imaging.At some teaching hospitals (numbers 3and 4) fellows occasionally assistedwith UAE. The standard personnel forUAEproceduresconsistedofoneIR,at least one nurse, andat least oneradiology technologist.Preprocedural InvestigationsAll patients underwent agyneco-logical examination (performed by gy-necologists) before UAEtorule outothercausesfortheirsymptoms. Theprocedure, risks, indications, andal-ternatives wereexplainedtothepa-tient indetail bytheIR, afterwhichinformedconsent wasobtainedfromTable 1Background and Experience of Interventional Radiologists in the Ontario UFE TrialInterventionalRadiologist HospitalFellowshipTraining(months) Residency1 1 12 Montreal, Quebec, Canada2 2 24 Toronto, Ontario, Canada3 3 6 Toronto, Ontario, Canada4 4 24 United Kingdom5 5 12 London, Ontario, Canada6 5 12 London, Ontario, Canada7 6 12 London, Ontario, Canada8 5 3 London, Ontario, Canada9 7 12 Vancouver, British Columbia, Canada10 8 12 Montreal, Quebec, Canada11 2 12 Toronto, Ontario, Canada546 Ontario UFE Trial: Technical ResultsMay 2003 JVIRallpatientsforthetreatmentandthestudy. Preadmissionbloodworkandinstructionwereperformedapproxi-mately1 weekbefore treatment. Atsome sites, patients viewed a video toorient them with the use of a patient-controlledanalgesic pumpfor man-agement of postprocedural pain. In-vestigations included prothrombintime and international normalized ra-tio, partial thromboplastin time, com-plete blood count with platelet count,and serum creatinine. Transabdominaland/or transvaginal US examinationswere performedbefore embolizationto document baseline characteristics ofthe uterus and fibroid(s). Uterine andfibroid (largest) volume measure-mentswereestimatedwiththeequa-tion for an ellipsoid shape (d1 d2 d3 0.5233).Admission and PreparationOnthemorningoftheprocedure,patients were generallyadmittedasinpatients, prepared, andbrought toInterventional Suites in the Depart-ment of Medical Imaging. Antibioticprophylaxis was routine at four hospi-tals(numbers2, 3, 4, and6) andre-served for patients at higher infectionrisk at four hospitals (numbers 1, 5, 7,and 8). Routine antibiotic coverage in-volved 1 g of cefazolin (Ancef; Smith-KlineBeechamPharm, Oakville, On-tario) administered intravenously 1hour before the procedure. Patientswhowereallergictopenicillinwereadministered500mgvancomycinin-travenously(Vancocin; Eli LillyCan-ada, Scarborough, Ontario, Canada).Foley catheters were placed by nursesbefore treatment.Sedation and AnalgesiaPreprocedural analgesics included30 mg intramuscularly or intravenousToradol (Ketorolac Tromethamine;Hoffman-LaRoche, Mississauga, On-tario, Canada), 30 mg ibuprofen intra-muscularly (Motrin; McNeil Con-sumer Healthcare, Guelph, Ontario,Canada), 600mgibuprofenorallyor50mgindomethacin(Indocid; MerckFrosst Canada, Kirkland, Quebec,Canada) suppository.Patients were administered con-scious sedation with use of benzodiaz-epines such as midazolam(Versed;Hoffman-LaRoche) ordiazepam(Va-lium, Roche, Mississauga, Ontario,Canada). Narcotics suchas fentanylcitrate (Sublimaze; Abbott Laborato-ries, St. Laurent, Quebec, Canada),meperidine hydrochloride (Demerol;Abbott Laboratories), or morphinesulfate were used as necessary for an-algesia. Epidural anesthesiawas ad-ministered to 13 patients at one center(number 8). In that protocol, epiduralcatheters were inserted in the medicalimaging suites by anesthesiology andcontinuous epidural infusions wereadministered with bupivicaine hydro-chloride 0.125% (Sanofi-SynthelabCanada, Markham, Ontario, Canada)and fentanyl citrate 5g/mL titratedper pain usually at 8 mL/h. Epiduralcatheters were removed several hoursafter the procedure.Angiography and Arterial AccessDigital subtraction angiographywas performedwithavarietyof C-armequippedunits capableof roadmapping and pulsed fluoroscopy (ex-ceptatinstitutionnumber3). Patientvital signs, includingoxygensatura-tion, heart rate, andbloodpressure,were monitored by pulse oximetry,electrocardiography (except at institu-tion number 3), and a blood pressurecuff. Duringthe procedure, patientswere administered an intravenous so-lution of 5% dextrose solution or0.45% saline running at 100120mL/h.Withlocal anesthetic andaseptictechnique, hemostatic sheaths wereplacedinthecommonfemoralarterythroughwhicha 4- or 5-Fselectivesingle-holecatheterwasusedforse-lectivecatheterizationof theanteriordivisionoftheinternaliliacandsub-sequently the left and right uterine ar-teries. Two IRs (numbers 3 and 11) didnot routinely use a hemostatic sheath.Thecatheterwasthenadvancedintothehorizontal portionof theuterineartery and satisfactory position forsafe embolization was confirmed withangiography. When required, pelvicflush arteriography with nonionic con-trastmaterial wasperformedtomapthe origin, size, and course of the uter-ine, ovarian(if abdominal aortogra-phy was performed), and other pelvicarteries. Microcatheters were not rou-tinely used but were reserved for caseswith difficult uterine artery access. Inmost cases, both uterine arteries wereTable 1(Continued)Fellowship Training (Location)Years PerformingPeripheral AngiographyEmbolization(Post-fellowship)Time asDedicatedInterventionalist(%)AdmittingPrivilegesMontreal, Quebec, Canada 15 90 NoBoston, MA; Cornell, New York, NY 26 100 YesToronto, Ontario, Canada 11 60 NoToronto, Ontario, Canada 7 50 NoDuke, Durham, NC; London, Ontario, Canada 11 75 NoLondon, Ontario; Montreal, Quebec, Canada 13 50 NoToronto, Ontario, Canada; Detroit, MI 3 50 YesDetroit, MI 5 20 YesToronto, Ontario, Canada 15 80 YesMontreal, Quebec, Canada 3 60 YesBoston, MA 3 100 NoPron et al 547Volume 14 Number 5catheterized fromthe right femoralapproach. The ipsilateral uterine ar-tery was catheterized with a variety oftechniques, includingdirect catheter-izationandformationof aWaltmanloop (12).EmbolizationThe primary embolic agent waspolyvinyl alcohol (PVA) particles 355500minsizeandpackagedas1ccper vial (Contour; Target Therapeu-tics, Boston Scientific Corporation,Mississauga, Ontario, CanadaorIva-lon; Cook, Stouffville, Ontario, Cana-da). One IR (number 4) routinely useda larger particle size (500710 m) forembolization. Embolic material washandmixedinasyringewithsalineand nonionic contrast material in a ra-tiodependentoncontrastagentden-sity. The slurry was then carefully in-jected to prevent reflux out of theuterine artery. Embolization pro-ceeded until a standing column of con-trast in the uterine artery was ob-served or contrast refluxed toward theuterine artery origin or into the inter-nal iliac artery.FourIRs(numbers1, 5, 6, and8)routinely used supplemental metalcoils. Protocol for IR number 1 was toinject pledgets of absorbable gelatinsponge (Gelfoam; Pharmacia &Up-john, Mississauga, Ontario, Canada)whenendpoints withPVAparticleshad been established, followed byplacement of asinglemetal coil perside (Tornado; Target Therapeutics,Boston Scientific). Protocol for theotherthreeIRs(numbers5, 6, and8)wastoplacetwometalcoilsperside(Gianturco; Cook, Stouffville, Ontario,Canada) with gelatin sponge pledgetssandwiched between. After emboliza-tion, the sheath and catheter were re-moved and hemostasis at the puncturesite was obtained by directcompression.After UAE, morphinesulfatewasadministeredthrougha patient-con-trolled analgesic pump. The anesthesi-ology pain service programs in thehospitals were responsible for the pa-tient-controlledanalgesic pump. Pa-tients allergic to morphine wereswitched to meperidine-based pumps.Nausea was managed by metoclopra-mide (Reglan; Wyeth-Ayerst, St. Lau-rent, Quebec, Canada), dimenhydri-nate (Gravol; Carter Horner,Mississauga, Ontario, Canada),and/or ondansetron (Zofran; GlaxoWellcome, Mississauga, Ontario, Can-ada). Patients recovered in the postan-esthesia recovery room for 1 to 2 hoursafterUAEandwerethentransferredto the gynecology ward and admittedovernight.At three hospitals (numbers 6, 7,and8),patientswereadmittedtothesole supervision of the IR. Patientswho underwent epidural anesthesia atone hospital (number 8) were dis-chargedthe same day, usually 45hours after the epidural was stopped,if they were ambulatory, toleratingoral medications, andtherewerenocomplications. Foley catheters wereremoved8hoursaftertheprocedureand, on the following day, intravenouslines were removed when the patientwas tolerating clear fluids. Patientsweredischargedthenext morningifambulatory and tolerating oral intake.At discharge, patients were given pre-scriptions for ibuprofen, codeine, oroxycodonewithacetaminophen(Per-cocet; DuPont Pharma, Mississauga,Ontario, Canada) and stool softeners.Follow-up was performedby re-searchers through a central universitycoordinating site. Telephone inter-viewswereperformedbytrainedin-terviewerswithpreparedscriptsat2weeks, 3 months, 6 months, andplannedannuallyfor 5years. PelvicUS examinations were also scheduledafter embolization therapy at intervalsof3months, 6months, andannuallyfor 5 years.Data Management and AnalysisData forms ontechnical informa-tion were completed at the end of eachprocedure by the IR and forwarded toresearch staff at the central coordinat-ingsite. Informationcollectedontheprocedureincluded: operator; proce-duredate; procedurestartandfinishtime; fluoroscopy time; material used,including contrast agent; catheters;embolicmaterial(s); medications; andprocedural complications. Procedureduration was defined as the start timefrom arterial puncture to the end timeat puncture closure and catheter with-drawal. Fluoroscopytime was auto-matically generatedby angiographyunits and was recorded by thetechnologist.Technical success was definedasbilateral embolization, verified by flu-oroscopy as stasis in the uterine arterytrunk and nonfilling of the uterine ar-tery on postembolization angiogra-phy. The effect of experience on tech-nical performance was investigated bycomparison of the outcomes in theearly experience, defined as the first 20consecutive procedures, with later ex-perience, definedasthenext20con-secutive procedures. A cut-off point of20 procedures was based on study re-sultsreportedbyAndrewsetal(13),who reportedthat fluoroscopytimestabilizedafter 20procedures. Com-plication events were classified basedon criteria established by the Stan-dards of PracticeCommitteefor theSocietyof Cardiovascular andInter-ventional Radiology (SCVIR) (14). Mi-nor complications included eventsthat involvednominal therapyof noconsequencesuchashematomasthatresolvespontaneously.Majorcompli-cations included events that involvedminortherapywithashort hospital-ization, major therapy with an un-planned increase in care and pro-longedhospitalization(48hours), orpermanent adverse sequelae or death.Descriptive measures such asmeans, medians, and95%CIs wereused to detail variation in proceduraltechnical details such as complicationrates, procedure time, fluoroscopytime, and contrast volume. Complica-tion rates included events that oc-curred during the first and second at-tempted procedures. Exact CIs forrates were based on the binomial dis-tribution with use of StatXact software(Cytel Software, Cambridge, MA). Dif-ferencesinproceduretimeandfluo-roscopy time between two groupswereevaluatedbyStudentttestandbetween more than two groups byone-way analysis of variance. Statisti-calsignificancewasdeterminedbyPvalues of 0.05 or less and all tests weretwo sided.RESULTSTechnical SuccessIn total, 570 UAE procedures wereperformedon555womenby11IRs(Table2). Themajoritywereelectiveprocedures. Five procedures (0.9%)wereperformedbecauseofurgentoremergent situations, including fourprocedures for fibroid related vaginal548 Ontario UFE Trial: Technical ResultsMay 2003 JVIRhemorrhageandoneprocedureper-formed after an aborted hysterectomy(becauseofextensiveadhesionsfromprevious myomectomies). The totalnumberofUAEsperformedbyindi-vidual IRs ranged from 5 to 203; witheight IRs performingmore than25.UAE was bilaterally successful in 97%of patients(538of 555patients; 95%CI: 95%98%]. Bilateral success wasachieved in 523 patients (94%) on thefirstattemptand15patients(3%)onthe second attempt.Failureofbilateralembolizationatthe first attempt occurredin32 pa-tients (5.8%) and resulted in three dif-ferentoutcomes(Table3): successfulsecond uterine artery embolization onthe second attempt in 15 patients(group1), unilateral embolizationin14patients(group2), andbilaterallyunsuccessful embolization in three pa-tients (group 3).Themostcommonreasonforfail-ure to bilaterally embolize at the firstattemptwasforanatomical variation(n18). Uterinearterieswereeithertoo small or too tortuous to catheterizeor thevessel originangles weretootortuous or steep for access. However,most of thesepatients (13of 18pa-tients) were successfully treated with asecond procedure, usually with a con-tralateral approach. In one case, previ-ous treatment with lupron depot (Leu-prolite Acetate for Depot Suspension;Abbott Laboratories) resultedin re-duced arterial diameter that preventedaccess and necessitated procedure ter-mination. The patient was successfullyembolized at the second attempt aftera 2-month wait to allow for the uterinearteries to normalize. Allergic reac-tions occurredintwocases, onein-volved classic signs of anaphylaxis. Inboth cases, patients returned at a laterdate and were successfully embolizedafterpremedicationanduseof alter-nate contrast agents.Of the14patients withunilateralembolizations (group 2), the majorcause was absence or nonidentifica-tionof theuterineartery(fiveright,five left). In three of these patients (oneright, twoleft)alargeovarianarterywas noted to be the dominant sourceof uterine vascularity. In the other fourpatients, perforations in the right uter-inearterypreventedembolizationintwo patients and vessel tortuosity pre-vented catheterization in two patients.Twootherpatientsinthisgrouphadpreviously undergone myomectomyand the uterine artery may have beendamaged or ligated during surgery.Procedural ComplicationsThe overall procedural complica-tion rate was 5.3% (95% CI: 3.6%7.4%).Complication events (n 30) that oc-curredduringembolizationare out-linedinTable4. Themost commoncomplications were angiography re-lated and included minor reactions tocontrastmaterial(n10)andhema-tomas (n 4). Perforations occurred infour patients; three involved the uter-ine artery and one involved the obtu-Table 2Technical Results of Embolization TherapyInterventionalRadiologist HospitalBilateral EmbolizationUnilateralEmbolizationBilateralFailureTotalCasesTotalProceduresAt firstprocedureAt secondprocedure*1 1 190 6 5 2 203 2092 2 64 1 65 663 3 49 1 1 51 524 4 43 3 1 47 505 5 40 40 406 5 38 1 2 41 427 6 38 2 40 408 5 31 3 2 1 37 409 7 13 13 1310 8 12 1 13 1311 2 5 5 5Total 523 15 14 3 555 570* Cases undergoing successful bilateral emoblization at the second procedure had an earlier first procedure.Table 3Reasons for Unsuccessful Bilateral Embolization at First ProcedureGroup 1 Group 2 Group 3TotalBilateralEmbolization atSecond ProcedureUnilateralEmbolizationBilateralFailure1. Unable to catheterize; smalltortuous uterine artery, steepangles13 2 3 182. Uterine artery absent or unableto locate 10 103. Allergic response 2 24. Perforation uterine artery 2 2Total 15 14 3 32Pron et al 549Volume 14 Number 5rator artery (successfully treated withtwo platinumcoils). Perforations inthe uterine artery prevented completeembolizationintwo patients. Bloodpressure decreased in five patientsand increased in one patient. One pa-tient experienced numbness in her legafter epidural anesthesia and was keptovernight for observation.Of the30immediatecomplicationevents, three (0.5%; 95%CI: 0.1%1.5%) weremajor complications thatresultedinextracareoranextendedhospital stay. One woman experiencedmultiple seizures approximately 10hours after UAE and was furtherhospitalizedfor3daysfortreatmentand observation. Her embolizationhadbeenuneventful, therehadbeenno anaphylactoidsigns or reactionsduring or just after the procedure.She did not have a history of seizuresand her medical history was unevent-ful. Her seizures were controlled withphenytoin (Dilantin; Parke-Davis,Scarborough, Ontario, Canada) andshe had no further neurologic se-quelae at 3- and 6-month follow-up.Allergic reactions wereserious intwocases. Thefirstcaseinvolvedananaphylactic response. After accessingtherightinternaliliacartery, 1.5to2mLof iodixanol (Visipaque270; Ny-comedAmersham, Oakville, Ontario,Canada) was injected. The patient be-came flushed and developed itchinessinthe throat andpalms. Her pulseincreasedto110andbloodpressuredecreasedto93/54mmHg. Shewasgiven oxygen by mask, a 500 mL bolusof normal salineand0.4mL1/1000epinephrine subcutaneously. Symp-tomsresolvedwithin15minutes. At20 minutes, the patient developed con-fluent hives centered on the pelvis thatbecamemorescatteredwithincreas-ing distance from the pelvis and otherthan pruritis was asymptomatic.Hivesslowlyfadedafteradministra-tionof50mgintravenousdiphenhy-dramine HCI (Benadryl; Warner-Lam-bert, Scarborough, Ontario, Canada)and recovery was uneventful.Asecond embolization was suc-cessfully performed 3 weeks later. Thepatient was given prophylaxis withthreeoraldosesofprednisone50mgandonedose50mgintravenousdi-phenhydramineHCI beforethepro-cedure. Alternate contrast agentsincluding 20 mL Gadopentetate dime-glumine (Magnevist, Princeton, NJ)and 40 mL CO2 were used. In the re-covery room, 30 minutes after the pro-cedure, thepatientdevelopednauseaandmildlydecreasedbloodpressureof 115/68that respondedto0.6mgintravenous atropine and 50 mg intra-venous dimenhydrinate. Recoverywas otherwise uneventful.The secondcase involvedanun-usual allergic type reactionthat oc-curred early after a small volume testinjectionof iohexol (Omnipaque300;Nycomed Amersham, Oakville, On-tario, Canada). The patient becameflushedsawstars, felt light headed,had chest heaviness and had difficultybreathing. Pulse increased to 124 andbloodpressurewas 180/85. Thepa-tient was turned on her side and oxy-genanda 250 mLbolus of normalsaline was administered. Epinephrinewasnotadministeredbecausesymp-toms rapidly resolved. The patient didnot have typical signs of contrast reac-tionurticaria, lip or tongue swelling,or bronchospasm. Two weeks later theprocedure was repeated with io-dixanol contrast agent. Prophylaxisincludedthreedosesof 50mgpred-nisone orally and 50 mg diphenhydra-mineHCLintravenously. Otherthanone hive, the procedure was unevent-ful and was successfully completed.Procedure TimeProcedure time for the group aver-aged 61 minutes (95% CI: 5863 min-utes) and the median time was 55 min-utes. Theaverageproceduretimeforbilaterally successful cases at the firstattempt was 59 minutes (95% CI:5761minutes) andforunsuccessfulcases was 88 minutes (95% CI: 75100minutes). Thelongertimeforunsuc-cessful cases(mean, 29minutes; 95%CI: 1839 minutes) was statisticallysignificant (P .001). Procedure timealso varied significantly (P .001)among IRs from 45 minutes to 99 min-utes (Fig 1).Overall, the average time for proce-dures performedinearlyexperiencewas75minutes(95%CI: 7080min-utes) andinlaterexperiencewas55Table 4UAE Procedural ComplicationsN (%)Minor contrast reaction 10 1.8Anaphylactoid response* 2 0.4Hematoma 4 0.7Vasovagal attacks, bradycardia 5 0.9Pseudo aneurysm 1 0.2Hypertension 1 0.2Perforation uterine artery 3 0.5Perforation obturator branch 1 0.2Arthritic flare 1 0.2*Seizure 1 0.2Leg numbness 1 0.2Total 30 5.3* Complications were classified asmajor because of resultant extra careand extended length of hospital stay.Figure 1. UAEprocedure time byinterventionalradiologist.550 Ontario UFE Trial: Technical ResultsMay 2003 JVIRminutes (95% CI: 5259 minutes). Theaverage difference in procedure time,20minutes (95%CI: 1326minutes)representinga25%reductioninpro-cedure time, was significant (P .001).Of the eight IRs who performed 40 ormore procedures, average time de-creasedforsix, rangingfrom12%(8minutes) to42%(44minutes) reduc-tion. Procedure time was significantlyreduced for two IRs (numbers 2 and 6)(Table 5).Fluoroscopy TimeFluoroscopy time for the group av-eraged 18.9 minutes (95%CI: 18.019.8minutes). The average fluoroscopytime for cases that were bilaterally suc-cessful at thefirst attempt was 18.1minutes(95%CI: 17.318.91minutes)and for cases that were not successfulatthefirstattemptwas31.5minutes(95% CI: 24.338.6 minutes). Thelonger time for unsuccessful cases, onaverage 13.4 minutes (95% CI: 5.920.8minutes), wassignificant (P.001).Fluoroscopytimealsovariedsignifi-cantly among IRs from15 minutes(95% CI: 1318 minutes) to 27 minutes(95%CI: 2123 minutes) (P.001)(Fig 2).Overall, the average fluoroscopytime for procedures performed inearly experience was 21.3 minutes(95%CI: 19.423.3 minutes) andinlater experience was 16.2 minutes(95% CI: 14.817.5 minutes). Theshorter fluoroscopy time with in-creasedexperience, 5.1minutes(95%CI: 2.77.6 minutes) or 24% reductionin time, was significant (P .001). Ofthe eight IRs whoperformed40 ormore procedures, average fluoroscopytimedecreasedwithsixIRs rangingfrom a 4% (2.7 minutes) to a 42% (8.1minutes) reduction. Fluoroscopy timewas significantly reduced for two IRs(numbers 2 and 6) (Table 6).Procedural MaterialsThe majority of embolizations wereperformed with a variety of 4- and 5-Fcathetersin42%and53%ofthepro-cedures (N570), respectively. Aright femoral approachwas usedin535 procedures (94%). Microcatheters,usuallythe Fas Tracker 325(BostonScientific, Target Therapeutics), wereused in 28 procedures (5%). Vasodila-tors, such as intraarterial nitroglycerinor sublingual nifedipine, were used toTable 5Effects of Interventional Radiologists Experience on UAE Procedure TimeInterventionalRadiologistEarly Experience* Time(Minutes)Late Experience Time(Minutes)Change in Time(Minutes)P Mean 95% CI Mean 95% CI Mean Change (%)1 66 5873 58 5067 8 12 ns2 100 87112 59 5167 41 41 .0013 53 4363 43 3550 10 19 ns4 64 5375 62 5271 2 3 ns5 50 4259 40 3346 10 20 .056 104 81128 60 5368 44 42 .0017 67 5083 70 6178 3 4 ns8 76 5597 55 4568 21 28 ns9 88 789910 99 8611211 58 4571Group 75 7080 55 5259 20 25 .001* First twenty consecutive cases. Next twenty consecutive cases. Student t-test.ns not significant.Figure 2. UAEfluoroscopy timeby interventionalradiologist.Pron et al 551Volume 14 Number 5controlspasmin23procedures(4%),primarily by two IRs intheir earlycases. All IRsusednonioniccontrastagents. Theaveragevolumeinjectedperprocedurewas147mL(95%CI:141154 mL). Various contrast agents,includingOmnipaque300(NycomedAmersham), Visipaque 270 (NycomedAmersham), Isovue 300 (Iopamidol,BraccoDiagnostics) andOmnipaque240 (Nycomed Amersham), were usedin 60%, 18%, 16%, and 6% of the pro-cedures, respectively. PVAparticlessized355500 mwereusedin501procedures (88%). The average totalnumber of vials used per case was 3.6(median, 3.3; range, 110). PVA parti-cles sized 500710 m were used in 50procedures by one IR (number 4). Theaverage total number of vials used percase was 7.4 (median, 6.5; range, 226).Supplemental metal coils were used in308 procedures (54%).DISCUSSIONThe objectives of this report were todocument the spectrumof practiceand technical results for UAE therapyfor symptomatic fibroids in a collabo-rative multicenter setting (OntarioUFE Trial). Examination of the proce-dural variation in a new and evolvingtherapy is a useful method to identifythe parameters to be used to define thelimits of practice. It is alsoausefulmethod to identify aspects of the tech-niquethatarecontentiousandbasedmore on personal preference than ob-jective evidence. This multicenter col-laborative groupincludeda diversespectrumof interventional practice.Recruitment to the different centersdidvaryandwasmostlyattributableto the varied time of application sub-missions and approval processes fromtheir Ethics ReviewBoards. Despitethis, most IRs hadperformed40 ormore procedures during the study.Thehighdegreeof technical suc-cess, 97% reported in this trial agreeswith results reported in American(14) and European (7,8,10) studiesand is also higher than the suggested95%thresholdofsuccessforembolictherapies established by the SCVIRguidelines(15). Themainreasonforfailuretobilaterallyembolizeat thefirst attempt (ie, difficultiesincathe-terizing small or tortuous arteries) hasalso been reported by others: Siskin etal (2) (2%; 1 of 49); Hutchins et al (5)(0.7%; 2 of 305); Ravina et al (6) (5%; 4of 80); Braude et al (16) (2.5%; 6 of 200)and McLucas et al (17) (1.8%; 3 of 167).Although difficulty in accessing arter-ieswasthemostcommonreasonforfailureinthisseries, caseswereusu-ally successfully embolized at the sec-ond attempt. One case in this trial thatinvolved previous treatment with Lu-pron was an avoidable failure. Use ofhormonesbeforeUAEhasalsobeenreportedbyothers(1,6,18)asacausefor embolization failure. Because of itseffects on the uterine arteries, it is rec-ommended that medical therapyshould be discontinued several weeksbefore embolization therapy.In some cases, variant anatomywith arteries that are congenitally ab-sent or traumatically damaged made itimpossible to complete bilateral embo-lization. Inthisstudy, singleuterinearteries wereidentifiedin10of 555cases (1.8%). Two of these had previ-ous myomectomies. Absent uterine ar-teries have also been reported byHutchins et al (5) (1.6%; 5 of 305) andby Kirsch et al (19) (1.6%; 12 of 746) inareviewof experienceof threeIRs.However, damagetouterinearteriescan also occur during embolization insome cases, resulting in incompleteembolizations. In this study, the rate of0.5%(3of555cases)waslowerthanthat reported by others (5,8).Vasospasmcanalsobeacauseoftechnical failure. Spasmcanresultinrestricted blood flow, limiting the flowof embolic material to distal regions ofsmaller vessels and potentially result-ing in incomplete or less effectivetreatment. Spasmwasnot systemati-cally reported in this study but Pelageatal(18)notedvasospasmin26%ofcases (97 of 197) that prevented embo-lizationin3%ofcases(6).Brunereauet al (8) also reported vasospasm as asignificant problem in 10% of cases (6of58). Themajority(5of6)ofthosecases were treated with the use of mi-crocatheters. Occurrence of vasospasmappearstobeasvariableastheap-proaches to prevent or treat it. Optionsinclude waiting until the spasmre-solves, catheterizingpast the spasmwithahydrophilicwire,useofami-crocatheter, or treatment with vasodi-lators. Some IRs early in this trial usedTable 6Effects of Interventional Radiologists Experience on UAE Fluoroscopy TimeInterventionalRadiologistEarly Experience* Time(Minutes)Late Experience Time(Minutes)Change in Time(Minutes)P Mean 95% CI Mean 95% CI Mean Change (%)1 15.6 8.922.2 16.8 13.719.8 1.2 8 ns2 24.9 19.530.4 14.1 11.616.7 10.3 42 .0013 12.9 9.017.0 12.4 9.615.3 0.5 4 ns4 18.9 7.430.4 18.9 13.624.2 0.0 0 ns5 17.7 14.121.4 13.0 8.917.2 4.7 26 ns6 28.9 22.235.5 20.8 17.324.3 8.1 28 .037 15.6 12.518.7 12.9 10.215.7 2.7 17 ns8 28.2 18.837.6 21.1 14.627.7 7.1 25 ns9 21.3 16.825.910 27.3 20.534.111 16.0 5.426.6Group 21.3 19.123.3 16.2 14.817.6 5.1 26 .001552 Ontario UFE Trial: Technical ResultsMay 2003 JVIRvasodilators but, because of subse-quent incidents of vasovagal attacksand hypotension, the practice was dis-continued. Microcatheterswereusedinfrequently (5%) in this series, inkeeping with the low rates reported inEuropean(Pelageet al (7), [6%; 5of80], Brunerau et al (8), [11%; 6 of 53])trials. While this variabilitymaybeduetoindividual preference, institu-tional cost constraints may also play arole.Nontarget embolizationis anim-portant risk consideration for thoseproviding or undergoing UAE. In thistrial, theovarianarterywasnotedtobethemajoruterinevascularitysup-plyinseveral cases. Vascularsupplytoother regions of thebladder, rec-tum, or vagina in association was alsonotedinsomeofthesecases. Unilat-eral embolizationwas performedinthese cases because of a concern withdisruptingovariansupply. Althoughthe rich network of communicatingvessels(20)inthepelvisofferssomeprotection, judgment, experienceanddetailed knowledge of pelvic anatomyis required to ensure that inadvertentembolization of nontarget organs doesnotoccur. Theserisksshouldbedis-cussedwiththepatientbeforeembo-lization as part of the informedconsent.Procedural complications duringUAEwere uncommonandinmostcases minor. Onlythreewereclassi-fiedasmajorand, althougheachre-quiredadditional careandincreasedhospital stay, none involved any sub-sequentclinical sequelae. Anaphylac-tic reactions to iodinated contrast ma-terialsoccurinfrequently(21). Inthistrial, two procedures were abandonedbecause of anaphylactic reactions.However, both cases were successfullymanagedatalaterdatebypremedi-cating the patients and using alternatecontrast agents. Gadoliniumuse tosupplement CO2angiographic tech-niques, includingUAE, hasbeenre-ported (22,23). Previous contrast mate-rial reactions therefore, may not be anabsolute contraindication for UAE,giventhesuccessful management ofpatients.Although the training standards forcompetencyinUAEhas beenestab-lished at 25 procedures for fibroids bythe SCVIR(15), the authors believethat it wasreasonabletoexploretheeffects of experience based on 20 pro-cedures. Thedecisionwasalsobasedon study results by Andrews et al (13),whoreportedthatfluoroscopictimesduring fibroid embolization had stabi-lized after 20 procedures. A cut pointof 20 procedures also allowed thecomparison of experiences of a greaternumber of IRs. Bothprocedure andfluoroscopic time are reflections oftechnical experience. Procedures timesreportedinthistrialwerelowerthatthose reported by others, even thoughembolizationwas performedwithasingle access site. In a review of tech-nical outcomes after UAE, medianprocedure times were higher for bothEuropean (87 minutes) and American(74 minutes) studies than for this Ca-nadian trial (55 minutes) (24). ThelongerproceduretimeintheAmeri-can studies can be explained to someextent by the greater use of microcath-eters, 55%versus 5%inthis study.However, the14%microcatheterusein the European studies was closer tousage in this study but their proceduretime was even greater than the Amer-ican studies. Other factors, such as op-eratorexperienceorcasedifferences,may also explain some of thesedifferences.Experience was foundto have amarkedaffect onprocedure time inthistrial. Withexperience, proceduretimes decreased significantly by 20minutes. WithsomeIRs, thelearningcurve was evensteeper withproce-dure times decreasing by 44 minutes,which represented a 42% reduction. Itwould seem that IRs in this study, de-spitehavingseveral yearsof experi-ence in angiographic and embolictechniquesbeforeUAE, continuedtoincrease their knowledge of femalepelvic vascular anatomy.Fluoroscopy time during UAE pro-cedures is an important considerationbecausetheovariesareinthefluoro-scopicfieldduringtheprocedure. Inthis study, the average fluoroscopytime (18.9 minutes) was also lowerthan those reported by others: An-drewset al (13) (30.6minutesfor35cases) and Nikolic et al (25) (21.9 min-utes for 20 cases). Significant decreasein fluoroscopy time was also noted inthisstudywithincreasedexperience.The decrease in this trial (21.316.2minutes) closely paralleled the de-crease with experience reported byAndrews et al (13).WithUAE, as withanynewandevolvingtherapy, therearetechnicalaspects that are opinion or preferencebased. The use of different embolicagents, types, andsizescontinuestobe a matter of debate. There is no com-parative evidence to suggest that out-comesofembolizationtherapydifferwith embolic particle size. In general,the short-term results of studies (2,68,10)thatusedsmallerembolicparti-cles (355500 m) are similar interms of fibroid reduction and symp-tomrelief tothe studies (1,3,4) thatused larger embolic particles (500710m). There is some thought thatsmaller particle size may be associatedwithmore ischemic pain, but therehave been no studies that confirm thisimpression.The use of secondary embolicagents, such as gelfoamand metalcoils, is also a source of debate. SomeIRs used metal coils routinely becauseof the ease in determining the techni-cal endpoint. Although the angio-graphic endpoint usedinthis studywas complete stasis in the uterine ar-tery trunk, it is still not generallyknownhowmuchorforhowlongaperiod vascular occlusion is needed toproduce adequate fibroid ischemiaandcell death. Useofcoilsincreasesthe cost of the procedure and preventsreembolization at a later date if it be-comesnecessary. Therefore, theiruseis probably not supported unless theyprove to be more effective or produceresults that are more durable. Furtherfollow-up of this trial cohort isplannedto determine if the use ofmetal coils or embolic particle size in-fluencesthedegreeof fibroidreduc-tion or symptom improvement.Acknowledgments: Thetrial representsthe collaborative efforts of members of theOntario UFE Collaborative Group and con-sists of the following individuals and insti-tutions: Dr. Gaylene Pron, DepartmentPublicHealthSciences, Universityof To-ronto, Toronto, Ontario. Dr. Andrew Com-mon, Dr. JaneWall; Department MedicalImaging, St. MichaelsHospital, Toronto,Ontario. Dr. AnthonyCecutti, Dr. JorzyCupryn, Dr. AdolmoMartoglio, Dr. PaulMcCleary, Dr. EvaMocarski, Dr. DonnaSteele, Dr. Karen Tessler, Department Ob-stetrics and Gynecology, St. Michaels Hos-pital, Toronto, Ontario. Dr. JohnBennett,Dr. RomanKozak, Dr. GregGarvin, De-partment Radiology, St. Josephs HealthCare, London, Ontario. Dr. GeorgeVilos,Department Obstetrics and Gynecology, St.Pron et al 553Volume 14 Number 5Josephs Health Care, London, Ontario. Dr.Stuart Bell, Department DiagnosticImag-ing, Rouge Valley Health Systems and theScarborough Hospital, Scarborough, On-tario. Dr. Jennifer Blake, Dr. Deborah But-ler, Dr. Rose Kung, Dr. Michael Shier, De-partment Obstetrics and Gynecology,Sunnybrook and Womens College HealthSciences Center, Toronto, Ontario. Dr. Mar-shaCohen, CenterforResearchinWom-ens Health; Sunnybrook and WomensCollegeHealthSciences Center, Toronto,Ontario. Dr. Murray Asch, DepartmentMedical Imaging, Mount Sinai Hospital,Toronto,Ontario.Dr.PeterHawrylyshyn,Dr. NickLeyland, Dr. Michael Sved, De-partment Obstetrics and Gynecology,Mount Sinai Hospital, Toronto, Ontario.Dr. TerryColgan, PathologyandLabora-toryMedicine, MountSinai Hospital, To-ronto, Ontario. Dr. Kenneth Sniderman,Dr. John Kachura, Department Medical Im-aging, Toronto General Hospital, Toronto,Ontario. Dr. MartinSimons, DepartmentMedical Imaging, Toronto Western Hospi-tal, Toronto, Ontario. Dr. LeslieVander-burgh, Department Radiology, WilliamOsler HealthCenter, Brampton, Ontario.Dr. Cuong Tran, Department Medical Im-aging,McMasterUniversityMedicalCen-ter, Hamilton, Ontario.References1. Goodwin SC, McLucas B, Lee M, et al.Uterine artery embolization for thetreatment of uterine leiomyomata mid-term results. J Vasc Interv Radiol 1999;10:11591165.2. SiskinGP, StainkenBF, Dowling K,Meo P, Ahn J, Dolen EG. Outpatientuterinearteryembolizationforsymp-tomaticuterinefibroids:experiencein49 patients. J Vasc Interv Radiol 2000;11:305311.3. SpiesJB, AscherSA, RothAR, KimJ,LevyEB, Gomez-JorgeJ. Uterinear-tery embolization for leiomyomata.Obstet Gynecol 2001; 98:2934.4. Worthington-Kirsch RL, Popky GL,HutchinsFL. Uterinearteryemboli-zation for the management of leiomyo-mas: quality-of-life assessment andclinical response. Radiology 1998; 208:625629.5. Hutchins FL, Worthington-Kirsch R,Berkowitz RP. Selective uterine ar-tery embolization as primary treatmentfor symptomatic leiomyomata uteri.J AmAssoc Gynecol Laparosc 1999;6:279284.6. Ravina JH, Bouret JM, Ciraru-VigneronN, et al. Particulate arterial emboliza-tion: a new treatment for uterineleiomyomata-related hemorrhage. LaPress Medicale 1998; 27:299303.7. Pelage JP, LeDref O, Soyer P, et al.Fibroid-related menorrhagia: treat-ment with superselective embolizationof the uterine arteries and midterm fol-low-up. Radiology 2000; 215:428431.8. Brunereau L, Herbreteau D, Gallas S, etal. Uterine artery embolization in theprimary treatment of uterine leiomyo-mas; technical features and prospectivefollow up with clinical and sono-graphic examinations in 58 patients.AJRAmJRoentgenol2000; 175:12671272.9. Bradley E, Reidy J, Forman R, Jarosz J,Braude B. Transcatheter uterine ar-tery embolization to treat large uterinefibroids. Br J Obstet Gynecol 1998; 105:235240.10. WalkerW,GreenA,SuttonC. Bilat-eral uterine artery embolization formyomata: results, complications andfailures. Min Invas Ther Allied Technol1999; 8:449454.11. Pron G, Cohen M, Soucie J, Garvin G,VanderburghL, Bell S. TheOntariouterine fibroid embolization trial. Part1. Baseline patient characteristics, fi-broid burden and impact on life. FertilSteril 2003; 79:112119.12. Waltman AC, Courey WR, Athanasou-lis CA, BaumS. Technique for leftgastric arterycatheterization. Radiol-ogy 1973; 109:732734.13. Andrews RT, Brown PH. Uterine ar-terial embolization: factors influencingpatient radiationexposure. Radiology2000; 217:713722.14. DroozAT, LewisCA, AllenTE, etal.SCVIRStandardsofPracticeCommit-tee. Quality improvement guidelinesfor percutaneous transcatheter emboli-zation. J Vasc Interv Radiol 1997;8:889895.15. Spies J, Niedzwiecki G, Goodwin S, etal. Training standards for physiciansperforming uterine artery embolizationforleiomyomata. JVascIntervRadiol2001; 12:1921.16. BraudeP, ReidyJ, Nott V, TaylorA,Forman R. Embolization of uterineleiomyomata: current concepts in man-agement. HumReprodUpdate2000;6:60360817. McLucas B, Adler L, Perrella R. Uter-ine fibroidembolization: nonsurgicaltreatment for symptomatic fibroids.J Am Coll Surg 2001; 192:95105.18. Pelage JP, Soyer P, LeDref O, et al.Uterine arteries: bilateral catheteriza-tionwitha single femoral approachand a single 5-F catheter-technical note.Radiology 1999; 210:573575.19. Worthington-Kirsch R, Walker WJ,Adler L, Hutchins FL. Anatomic vari-ation in the uterine arteries: a cause offailureof uterinearteryembolisationfor the management of symptomaticmyomata. Min Invas Ther Allied Tech-nol 1999; 8:397402.20. Pelage JP, LeDref O, Soyer P, et al.Arterial anatomy of the female genitaltract: variations and relevance to trans-catheter embolization of the uterus. AJRAm J Roentgenol 1999; 172:989994.21. Katayama H, Yamaguchi K, Kozuka T,Takashima T, Seez P, Matsuura K.Adverse reactions to ionic andnon-ioniccontrast media. Radiology1990;175:621628.22. Spinosa DJ, Hagspiel KD, Angle JF,Matsumoto AH, Hartwell GD. Gado-linium-based contrast agents in an-giographyandinterventional radiolo-gy: uses and techniques. J Vasc IntervRadiol 2000; 11:985990.23. EnglanderMJ, SiskinGP, DowlingK,Quarfordt S. Uterinefibroidemboli-zation without the use of iodinatedcontrast material. J Vasc Interv Radiol2002; 13:427429.24. SubramanianS, SpiesJB. Uterinear-tery embolization for leiomyomata: re-source use and cost estimation. J VascInterv Radiol 2001; 12:571574.25. NikolicB,SpiesJB,LundstenMJ,Ab-bara S. Patient radiation dose associ-atedwithuterinearteryembolization.Radiology 2000; 214:121125.554 Ontario UFE Trial: Technical ResultsMay 2003 JVIR