Journal oftheAmerican Society ofNephrology 523Hemodialysis Vascular AccessMorbidity1 [604210]

Journal oftheAmerican Society ofNephrology 523Hemodialysis Vascular AccessMorbidity1
Harold I.Feldman,2 Sidney Kobrin, andAlanWasserstein
Inawhomradialarteriovenous fistulaisaviable
HI,Feldman, S.Kobrin,A.Wasserstein, Department of
Medicine, RenalElectrolyte andHypertension Divison,
University ofPennsylvania School ofMedicine, Phila-
delphia, PA
HI,Feldman, Department ofBiostatistics andEpidemi-
ology,andCenter forClinical Epidemiology andBio-
statistics, Universily ofPennsylvania School ofMedi-
cine,Philadelphia, PA
(J,Am,Soc.Nephrol. 1996;7:523-535)
ABSTRACT
Complications associated withhemodialysis vascular
accessrepresent oneofthemostimportant sourcesof
morbidity among ESRDpatients intheUnitedStates
today.Inthisstudy,newdataonthemagnitude and
growth ofvascular access-related hospitalization in
theUnitedStatesispresented, demonstrating thatthe
costsofthismorbidity willsoonexceed $1billionper
yr.Thisstudyalsoreviewspublished literature onthe
morbidity associated specifically withnativearterio-
venousfistulae, polytetrafluoroethylene bridgegrafts,
andpermanent centralvenouscatheters. Next,new
information onthechanging palterns ofvascular
accesstypeintheUnitedStatesispresented, demon-
strating thecontinuing evolution ofmedical practice
awayfromtheuseofarteriovenous fistulaeinfavorof
morereliance onsynthetic bridgegrafts.Basedon
thesedata,adiscussion isprovided ofthetradeoffs
among themostcommonly available modalities of
vascular accesstoday.Although radialarteriovenous
fistulae continue torepresent theoptimal accessmo-
dality, theappropriate rolesforbrachial arterio-
venousfistulae, synthetic bridgegrafts,andcentral
venouscatheters arelesscertainbecause ofmade-
quatedataonthelong-term function ofthefirstand
thehighratesofcomplications associated withthe
lattertwo.Toreduce vascular access-related morbid-
ity,strategies mustbedeveloped notonlytoprevent
anddetectappropriately earlysynthetic vascular
accessdysfunction, buttobetteridentifythepatients
1ReceIved June2,1995.Accepted January 3,1996.
2Correspondence toDr.H.!.Feldman, CenterforClinical Epidemiology and
Biostatistics, University ofPennsylvania SchoolofMedicine, 720Blockley Hall,423
ServIce Drive,Philadelphia, PA19104.6021.
104&667310704.0523$03.00/0
Journal oftheAmerican Society ofNephrology
Copyright C1996bytheAmerican Society ofNephrologyclinicaloption.
KeyWords: Arteriovenous fistula,polytetrafluorene bridge
graft,permanent central venous catheter, hospitalization,
ESRD
Chronlc hemodialysis forpatients withESRD first
became technically feasible in1960withthe
introduction oftheQuinton-Scribner shunt (1).Al-
though thisexternal arteriovenous (AV)shuntpro-
vided, forthefirsttime,directaccess tothecirculation
sothathemodialysis could beperformed intheout-
patient setting, patient carewascomplicated byfre-
quentthrombosis andinfection (1-3).Autologous ax-
teriovenous fistulae (A’IF’), introduced byBrescia and
Cimino in1966, overcame many oftheproblems of
external dialysis shunts (4).Subsequent tothisad-
vance indialysis vascular access surgery, AVFbe-
cametheprimary mode ofachieving vascular access
forchronic hemodialysis. OnceanAVFhasmatured
andbeenusedfordialysis, thesubsequent orsecond-
aryfailure rateislow,withmostpatients enjoying
long-term fistula function lasting formany years (3-
5).
Since theearly1970s, therapidgrowth oftheU.S.
ESRD Program hasbeenaccompanied byadiminu-
tionintheuseofAVF infavorofalternative typesofAV
bridge grafts and,morerecently, permanent indwell-
ingcentral venous catheters. Thisevolution hasbeen
attributed tothegrowing primary failure rateofAVF,
oftenattributed totheincreasing ageandcomorbidity
ofthedialysis population, aswellasthemorerapid
usabifity ofalternative vascular accesses ascompared
withAVF(6-11).Supporting thisformulation isthe
extensive change inthethedemographics oftheU.S.
ESRD population overthepasttwodecades. Although
only6.6%ofESRD patients haddiabetes in1972and
fewerthan20%ofESRD patients wereolderthan65
yrofage,36%oftheESRD population in1991had
diabetes andnearly 1in2(45%) was65yrofageor
older(12).Ithasbeenestimated thatfewerthan25%
ofincident ESRD patients successfully undergo the
construction ofanative AVfistula, anestimate re-
cently borne outbyspecial studies conducted bythe
U.S.Renal DataSystem (10,13,14).
Earlyalternative surgical techniques utilizing autol-
ogous saphenous veingrafts andimplanted bovine
carotid heterografts toformbridge AVanastomoses
hadunacceptably highratesofthrombosis andpseu-
doaneurysm (2,3,6, 15-17). Inanattempt toachieve
better long-term success fordialysis vascular ac-
cesses, thesamesynthetic materials usedforartery to

Hemodialysis Vascular Access Morbidily
524 Volume 7‘Number 4‘1996TABLE1.Hospitalization forvascular accessmorbidity, 1984to19860
Diagnosis/Hospital Stay 1984 1985 1986
Totalnumber ofESRDhospital staysforanydiagnosis 150,775 158,634 170,572
Frequency ofhospital stayswithICD-9996.1b 6,079 5,755 5,953
Frequency ofhospital stayswithICD-9996.#{243}c 5,831 6,837 7,889
Frequency ofhospital stayswithICD-9 996#{149}7d 11791 13,452 15,899
Totalnumber ofstayswith996.1,996.6,or996.7listed 23,701 26,044 29,741
asanydiagnosis
Totalnumber ofstayswith996.1,996.6,or996.7listed 9,170 21016 24,025
asfirstdiagnosis
Access-related staysaspercentage ofTotal 15.7 16.4 17.4
Meanlengthofaccess-related stays 7.6 7.3 7.4
aAdapted fromdatapresented inReference 19:Feldman HI,HeldPJ.Hutchinson 1.StoiberE.Hartigan, MF,BerlinJA:Hemodialysis vascular access
morbidity intheU.S.Kidneymt1993:43:1091-1096.
bMechanical complication ofvascular device,implant.andgraft.
CInfection andinflammatory reaction asaresultofinternal prosthetic device, implant. andgraft.
dOthercomplications ofinternalprosthetic device,implant,andgraft.
artery vascular bypass procedures wereimplanted
intodialysis patients asbridging AVanastomoses.
Dacron#{174} (E.I.duPont deNemours andCo. ,Wilming-
ton,DE)wasusedinitially chosen forthese access
devices butwasrapidly replaced bypolytetrafluoro-
ethylene (PTFE) (GoreTex#{174}, W.L.GoreandAssoc.,
Flagstaff, AZ:Impra#{174}, Impra, Inc. ,Tempe, AZ)after
earlyexperience suggested better short-term patency
rateswiththisnewer synthetic material (10,18).As
willbediscussed below, thesubstitution ofsynthetic
access grafts forAVFhasbeenaccompanied byenor-
mous growth intherateofsecondary vascular access
failure, alongwithitsattendant costs.
Below. weprovide newdatafromMedicare’s ESRD
Program andsummarize published research describ-
ingthemorbidity associated withhemodialysis vascu-
laraccesses incurrent usetodayintheUnited States,
including AVF, PTFE bridge grafts, andindwelling
central venous catheters. Wealsoprovide recent data
fromtheU.S.Renal DataSystem onthecontinuing
transition fromAVFtoalternative synthetic access
devices. Finally, wepresent adiscussion ofthe
tradeoffs involved inchoosing atypeofvascular ac-
cessintheclinical setting, andthedataneeded to
establish appropriate guidelines formaking this
choice.
VASCULAR ACCESS COMPLICATIONS INTHE
UNITEDSTATES
Wehavepreviously reported onthemagnitude of
vascular access morbidity intheUnited States by
using hospitalization claims datafromMedicare’s
End-Stage Renal Disease database, theESRD Pro-
gram’s Medical Management Information System
(ESRD PMMIS) (19).Vascular access complications
weretabulated bycounting thenumber ofhospital
staysforoneofthreeICD-9 diagnosis codes recordedasthefirstdiagnosis inthehospitalization record of
PMMIS.*
Prevalent Medicare-entitled U.S.ESRD patients in
1984, 1985, and1986werestudied (seeTable 1).
Complications ofvascular accesses alone accounted
foroneofthemostimportant causes ofhospitalization
intheESRD population. Between 1984and1986,the
number ofhospitalizations forrevision ofdialysis
vascular accesses roseby25.0% andrepresented
14.1%ofallESRD hospitalizations thatyear.
Datarecently published byLazarus etat.(20)from
thenetwork ofdialysis units managed byNational
Medical Care(NMC) alsoexamined vascular access-
related morbidity asreflected inhospitalization for
vascular access complication-related diagnoses byus-
ingasimilar methodology. Between 1986and1990,
Lazarus etaLstudied over25,000 NMCpatients who
weremonitored forvascular access-related hospital-
ization. Theauthors reported thattheproportion ofall
hospital daysrelated tovascular access dysfunction
grewfromabout 6%in1988tonearly 11%bytheend
of1990,suggesting thatvascular access-related hos-
pitallzation isgrowing faster thanhospitalization for
othertypesofmorbidity intheESRD population.
Recently wehaveextended ouranalyses character-
izingtheburden ofvascular access morbidity inthe
United States between 1986and1991onthebasisof
datafromMedicare (seeFigure 1).During thistime
period, thenumber ofvascular access-related stays
grewtoover70,000. Although thisriseinaccess-
related hospitalization emanates, inpart,fromthe
growth intheU.S.ESRD population, Figure 1also
demonstrates thatvascular access-related hospital-
ization asapercentage ofall-cause hospitalization has
alsorisenfromapproximately 17%in1986togreater
*ThethreeICD-9codessearched forincluded: 996.1,Mechanical complication
ofvascular device, Implant, andgraft996.6,InfectIon andinflammatory reac-
tionduetoInternal prosthetic device, implant, andgraft:and996.7,Other
complications ofinternal prosthetic device, Implant, andgraft.

0.5
0.45
0.4
0.3580000
70000
60000
50000
40000
30000
20000
10000
0
Figure1.Vascular access morbidity-related hospitalization IntheUnitedStatesandvascular access morbidity-related
hospitalization among allU.S.dialysis patients asaproportion ofall-cause hospitalization.0.3
0.25Cl,
E5,
.0
0
a.
(#5C’,
4,
U
U,(
CC’
U(I,
CC’>
0
CC,
0
CC’
.1:

0.
U,
0
I
(C’
0
I-
C
0
t
0
0.
0
0a.0.2
0.15
0.1
1986 1987 1988 1989 1990 19910.05Feldman etal
Journal oftheAmerican Society ofNephrology 525U,
>.
U,
‘I,
U,0
C.,U
U,
0
CU
N
CU
0.
U,
0
I
than20%in1991.These estimates ofhospitalization
because ofvascular access complications aremost
likely underestimates ofaccess complications inthe
ESRD population, astheydonotconsider access-
related morbidity diagnosed andtreated intheoutpa-
tientsetting, outpatient procedures performed ina
hospital, orhospitalization paidforbyinsurers other
than Medicare. Although thedisproportionate in-
crease inmorbidity maybearesult ofincreasing
severity ofillness among U.S.ESRD patients, the
increasing demand foraccesses thatcanaccommo-
datehigher blood-flow ratesaswellastheuseofnew
diagnostic technologies such ascolor-flow Doppler
ultrasonography mayalsohaveincreased thedetec-
tionofheretofore occult vascular access problems.
Unfortunately, nosingle datasource exists thatpro-
vldesinformation onoutpatient access-related mor-
bidity ormorbidity paidforbyinsurers other than
Medicare. Medicare doesmaintain dataifiesthat
record billsforoutpatient procedures suchasfistu-
lography andcolor-flow Doppler ultrasonography
studies, buttoourknowledge, these datahavenot
beenexploited toestimate outpatient vascular access
morbidity.
CostsofVascular AccessDysfunction inthe
UnitedStates
Approximately one-half ofMedicare’s ESRD budget
(over$6billion in1991)isspent ontheproceduralcostsofdelivering dialysis toESRD patients (21,22).
Theremainder ofESRD expenses ariseprimarily from
thecostsofmorbidity andassociated hospitalizations,
ofwhich greater than14%havebeenestimated to
result fromcomplications ofvascular access (19).
Assuming asimilar rateofgrowth invascular access-
related hospitalization asobserved between 1986and
1991,weproject thatMedicare willpayformorethan
90,000 hospItalizations forvascular access-related
problems in1995. Ifvascular access repair proce-
duresareincreasingly performed intheoutpatient
setting, thenumber ofhospitaJlzations maynotriseas
quickly asithasInthepast.Nonetheless, based ona
conservative estimated costof$7500 perhospital
stay,theprojected coststoMedicare forthesehospi-
talizations alonewillriseabove$675million annually,
representing nearly 10%ofitsESRDbudget. This
analysis represents agrossunderestimate ofthetotal
costforvascular access-related morbidity inthe
UnitedStates foratleasttwoimportant reasons. First,
thehospitalization dataciteddonotinclude hospital-
ization paidforbyinsurers other thanMedicare.
Second, thecostsofdiagnostic screening procedures
(e.g.,fistulography, Doppler ultrasonography, etc.)
andoutpatient therapies (e.g.,thrombolysis, anglo-
plasty,etc.)arealsonotincluded. Thus, theannual
costofvascuiar access-related morbidity intheUnited
States willrapidly approach andexceed$1billion per
yr.

Hemodlalysis Vascular Access Morbidity
526 Volume 7.Number 4‘1996ARTERIOVENOUS FISTULAE
Arteriovenous FistulaFunction
Numerous investigators havestudied thelikelihood
ofsuccessful maturation ofAVF intofunctional hemo-
dialysis vascular accesses (5,8,17,23-32). Primary
failure ofAVFisdefined asinsufficient dilation and
arterialization needed forfunction asadialysis vascu-
laraccess. Theobserved riskofprimary failure of
theseaccesses hasranged from9%(24)to30%(17)in
surgical experience accumulated sincethelate1960s.
OnceAVFmature, numerous investigators havedoc-
umented theirexcellent long-term function. Winsett
andWolma reported thecomplication ratesfor273
AVFaftersuccessful maturation intofunctional vas-
cularaccesses (17).After2yrofdialysis, 90%ofthese
accesses continued tofunction, with80%stillinuse
after3yr.Theyobserved acomplication rateoffive
events per1000patient-months offollow-up forAVF,
which wasconsiderably andstatistically smaller than
the37events per1000patient-months observed in
recipients ofPTFE synthetic dialysis grafts. Similar
fistula survival rateswerealsocitedbyReilly and
colleagues intheirreport of148AVFcreated inEn-
gland between 1976and1981(25)andmorerecently
byKherlakian andcolleagues (8).Bonalumi andco-
workers havereported thelongest follow-up ofAVF,
observing that52%ofend-to-end radial artery fistulae
continued tobeuseful after6.5yrofhemodlalysis (5).
More recently, native AVFconstructed using the
brachial artery havebeenevaluated asanalternative
access procedure whenaradiocephalic fistula wasnot
abletobeconstructed (33-35). Although uncontrolled
studies havedocumented primary patency ratesas
highas86%at2yroffollow-up (33),thelong-term
function oftheseaccesses hasnotyetbeenevaluated.
RiskFactorsforAVFDysfunction
Despite theexcellent long-term function offistulae
aftersuccessful maturation, littleconsensus hasbeen
achieved regarding thepredictors oftheprimary fail-
ureofAVF. Anumber ofpotential riskfactors, includ-
inggender, vessel size,surgical technique, surgical
skill,andprimary diagnosis havebeenpostulated as
having anImportant impact onthelikelihood ofsuc-
cessful AVFmaturation. Several studies intheearly
1970s reported thatwomen hadahigher riskof
primary failure thanmen(24,26). Numerous other
reports (5,27-30) havelooked forbutfailedtoidentify
specific riskfactors ofprimary AVFfailure. Many of
thesestudies weresmall insizeandcarried outlim-
itedanalyses tocontrol forpotential confounding
factors.
Reilly andcoworkers presented oneofthebest
documented andoneofthefewprospective risk-factor
studies ofdialysis fistula function published todate
(25).Theystudied 157consecutive patients between
1976and1981.Intraoperative data, including size
andtypeofanastomosis, sizeofvessels, suture mate-
rial,andadequacy offlowwererecorded. Univariateandmultivariate analysis demonstrated thatveinsize
wasanimportant predictor oflong-term access com-
plications. Thepower ofthisstudy, however, was
limited because ofitssample size,andnospecific
analysis ofthepredictors ofinadequate fistula matu-
ration wasperformed.
Fewrecent dataareavailable regarding therisk
factors forinadequate maturation ofAVF. Pourchez et
aLreported that41of47French dialysis patients
olderthan75yrofageunderwent successful con-
struction ofanAVF(36).Kobrin etal.evaluated 48
AVFconstructed between 1989and1991attheUni-
versity ofPennsylvania (37).Overall, 44%ofAVF had
successfully matured at3months, whereas primary
failure occurred in56%.Although age,gender, loca-
tionoftheAVF,andthepresence ofdiabetes mellitus
didnotaffect AVFmaturation, thesmall sizeofthe
study limited statistical power todetect clinically rel-
evantdifferences inaccess function. Fewinvestigators
haveanalyzed theimpact oftheuseofrecombinant
erythropoietin (EPO) onAVFfunction (38,39) butthe
fewdataavailable seemtosuggest thatAVFthrombo-
sisisnotaverycommon complication. Neither Be-
sarabetaL(39)norTangetaL(40)found thrombosis
ofAVFtobemorecommon withEPOuse.
Thus, extensive published experience withAVF’
demonstrate theirlong-term utility andlowrateof
complications among patients inwhom primary fis-
tulafunction isachieved. However, despite alargerisk
ofprimary fistula failure, littleisknown about the
specific surgical, comorbidity, ordemographic factors
predictive ofprimary fistula function. Although itis
probably truethatanolderandsicker ESRD popula-
tionmaylesscommonly undergo successful AVFcon-
struction, thesubgroup ofpatients inwhom fistulae
should becreated ispoorly defined. Several recent
reports (23,33-37) suggest thatdespite theolderage
andworse comorbidity ofU.S.ESRD patients ascom-
pared withtwodecades ago,asubstantial number of
patients remain viable candidates forplacement of
AVF.Thisisespecially trueifupper-arm fistulas such
asbrachial artery-basilic veinfistulas areconsidered.
Areview ofalternatives toAVFreadily demonstrates
thatunnecessary substitution ofsynthetic grafts for
autologous AVFgreatly enhances boththemorbidity
ofandcostsfortheU.S.ESRD population.
POLVTETRAFLUOROETHYLENE DIALYSIS ACCESS
Patency andComplications of
Polyfetrafluoroethylene Dialysis Access
PTFE wasintroduced asamaterial forvascular
bypass grafts in1976(10).Since thattime, this
material hasbecome themainstay fordialysis vascu-
laxaccesses when anautologous AVFiseither be-
lieved tobetechnically impossible orhasfailed to
mature, orcommonly when thereisinsufficient time
foranAVFtomature before theneedforhemodialysis
because ofthelatepresentation ofapatient’s symp-

51).Feldman etal
Journal oftheAmerican Society ofNephrology 527tomatic advanced renalfailure. PTFE dialysis grafts
arecurrently themostcommon typeofdialysis ac-
cess,accounting forasmany as83%ofaccess place-
ments among prevalent ESRD patients (38).Agreat
majority ofhemodialysis patients undergo onlyPTFE
graftplacement despite thehighprobability thata
significant proportion ofthemwould havebeenableto
undergo successful AVFconstruction. However, few
dataexistdocumenting theproportion ofaccesses
thatareconstructed ofPTFEbutthatcouldhavebeen
successfully created asanative AVF.Nonetheless, the
potential excess useofPTFE vascular accesses and
theprotean complications possible aftertheirplace-
ment, including stenosis, occlusion, infection, aneu-
rysmandpseudoaneurysm formation, cardiovascular
instability, andnerve injury, areallpotentially re-
sponsible foralargecomponent ofthehighmorbidity
intheU.S.ESRD population.
Numerous investigators havereported 1-and3-yr
patency ratesforPTFEgrafts. Between 63and90%of
grafts stillfunction after1yr.whereas only42to60%
remain patent after3yr(9,13-15, 17,41-47). Mostof
thesereports defined patency aspersistent graftfunc-
tion,regardless oftreatment foraccess complications
suchasthrombectomy orpercutaneous angioplasty.
Complication-free graftsurvival hasbeenreported by
anumber ofinvestigators (17,41,45). Sabanayagam
andcolleagues (45)reported thecomplication-free
survival ofPTFE grafts tobe77.1%at1yr;Palder and
colleagues reported complication-free PTFEgraftsur-
vivaltobe51%at2yr(41):andWinsett andWolma
reported complication-free PTFE graftsurvival tobe
39%at3yr(17).Outflow obstruction andthrombosis
accounted forthevastmajority ofthecomplications
observed. Unfortunately, noneofthese studies re-
ported theinfluence ofpatient comorbidity onaccess
dysfunction. Byprospectively studying asubgroup of
52healthier patients wholivedlonger than3yron
dialysis, Schuman andcolleagues grossly controlled
forcomorbidity andobserved oneaccess complication
forevery515patient-days offollow-up (14).
Inrecent years, increasing attention hasbeengiven
tothedetection andrepair ofstenosis ofPTFEgrafts.
Color-flow Doppler ultrasonography andcontrast fis-
tulography arewidely utilized. Repair bytransluminal
angioplasty orsurgical vein-patch angioplasty may
promote delivery ofadequate dialysis andprevent
graftthrombosis. Whether repair before thrombosis
extends theuseful lifeofPTFE grafts hasyettobe
proven. Alarge, butnotprecisely determined, costis
incurred bythesediagnostic testsandinterventional
procedures, which areperformed forthemostparton
anoutpatient basis. Adetailed discussion ofthediag-
nosisandtherapy ofstenosis ofPTFEgrafts isbeyond
thescope ofthispaper, butthereader isreferred
elsewhere forrecent discussions oftheseissues (48-RiskFactorsforComplications fromPTFE
DialysisAccess
Thromboses ofVFFE access grafts have been
thought toresult fromanumber ofcauses, including
venous outflow obstruction andinadequate arterial
inflow. Outflow obstruction hasusually beenattrib-
utedtoneointimal hyperplasia (52-54) orcentral ye-
nousstenosis/thrombosis frompriorcentral venous
access (55).Inadequate arterial inflow hasresulted
fromasmall ordiseased arterial source aswellas
fromrecurrent bouts ofintradialytic hypotension (56-
59).Postdialysis dehydration withincreased blood
viscosity; lowserum albumin; inadequate anticoagu-
lation; external compression either fromperigraft he-
matomas, overly vigorous pressure atneedle sites,
occlusive bandages, orinopportune armpositioning
during sleep; thepresence ofanticardiolipin antibod-
ies,diabetes mellitus, andhighhematocrit withEPO
therapy haveallbeenclaimed, withvarying justifica-
tion, tocontribute toPTFE graft thromboses
(39,40,52,57,59-64).
Despite these numerous hypothetical riskfactors,
fewstudies haveexamined specific riskfactors ina
controlled fashion. Aman etaLreported on91PT’FE
grafts followed for24months andobserved asignifi-
cantly increased graftpatency ratewiththeuseof
double-lumen dialysis needles (54).Theywereunable
tofindarelationship between graftsurvival orcom-
plication-free graftsurvival andsex,race,degree of
hypertension, heparin dosing, orintradialytic hypo-
tension, although thestatistical power oftheirstudy
waslimited. Munda andcolleagues followed 67PTFE
grafts for2yrandfound thataforearm loopconfigu-
ration hadgreater thantwicetherateofsecondary
patency offorearm straight grafts (57).
Although controversial, several reports havesug-
gested anassociation between diabetes mellitus and
PTFE graftthrombosis (19,65,66). Windus andcol-
leagues observed arateofPTFEgraftthrombosis at1
yrof72%among patients withdiabetes, compared
witharateof49%among patients without diabetes
(65).DatafromMedicare andtheU.S.Renal Data
System (19,66) havebeenconsistent, demonstrating
higher rates ofhospitalization forvascular access
problems among thesubset ofESRD patients with
diabetes.
Several investigators haveevaluated theimpact of
therapy withrecombinant EPOonsynthetic vascular
access function andreported conificting results
(39,40,62). Besarab andcolleagues compared therate
ofthrombotic events in164patients receiving EPOfor
atleast2months withthatamong 142patients
treated earlier, before theavailability ofEPO.They
wereunable todetect anydifferences intherateof
thrombosis either forsynthetic grafts orAVF(39).
Consistent withthesefindings havebeenanumber of
otherstudies (67-70), eachfailing tonoteanincrease
inaccess thrombosis withEPO.
Incontrast tothese findings, Dyandcolleagues

Hemodialysis Vascular Access Morbidity
528 Volume 7‘Number 4.1996performed anonconcurrent cohort study of46dialy-
sispatients before andaftertreatment withEPO(62).
Theyreported agreater incidence ofgraftthrombosis
intheEPOperiod. Similar findings werereported by
theCanadian Erythropoletin Study, arandomized
comparison between treatment withplacebo (N=40)
andtreatment withEPOtargeting either ahemoglobin
concentration of9.5to11.0g/dL (N=40)or11.5to
13.0g/dL(N=38).A2.5%rateofaccess thrombosis
at6months wasobserved intheplacebo group witha
mean hemoglobin of7.4g/dL. Incontrast, thetwo
groups treated withEPOachieved mean hemoglobin
concentrations of10.2g/dL and11.7g/dL with
6-month thrombosis ratesof10%and18%,respec-
tively(71).Thelowrateofthrombosis intheplacebo
group wasconsiderably lowerthanthatpublished by
Eschbach andcolleagues (67),making thisstudy’s
findings difficult tointerpret. Although several inves-
tigations reporting substantial thrombosis rates in
patients withPTFEgrafts onEPOhaverecenfly been
reviewed (68),noneofthese studies haveincluded
control groups offEPO.
Thus, several investigators haveindependently con-
firmedtheimpression derived fromthehospitalization
datainMedicare’s ESRD database: therateofcompli-
cations, especially stenosis andthrombosis, arising
fromsynthetic dialysis vascular accesses isverylarge.
Themeanduration ofthecomplication-free access
survival appears tobebetween 9and16months, with
agreatmajority ofpatients experiencing acomplica-
tionbytheendofthe3rdyraftergraftplacement. This
extraordinarily highcomplication rate,withitsatten-
dantsuffering andhighutilization ofU.S.Dialysis
Program resources, makesthesearchforbetterstrat-
egiestoimprove outcomes among recipients ofsyn-
thetic vascular accesses urgent. Formulation ofsuch
strategies willrequire, inpart,better delineation ofthe
riskfactors forfailure ofPTFEaccess grafts.
INDWELLING CENTRAL VENOUS DIALYSIS
CATHETERS
Intended originally asatemporary measure, the
silicone double-lumen catheter withDacron#{174} cuffhas
inevitably come intolong-term useinpatients in
whom peripheral vascular access (fistula orgraft) Is
unsuitable. TheDacron#{174} cuffdecreases infection and
makes long-term usepossible. When catheters are
intended foruseforlessthan1month, apolyure-
thane,non-cuffed central venous access isnormally
used. However, ifcatheter useisexpected toexceed 1
month, thecuffed SILASTIC (DowCorning, Midland,
MI)catheter ispreferred (73).Theproportion ofhemo-
dialysis patients intheUnited States inwhom cuffed
SILASTIC
catheters provide long-term vascular ac-
cessisunknown. Based ondatafromthe1991U.S.
RenalDataSystem (USRDS) Special StudyofCase
Mix,Itcanbeestimated that4%ofU.S. ESRD patients
begin dialysis withapermanent catheter. These data
underestimate theactualprevalence ofcatheter use,astheydonotaccount forusage afterfailure ofmore
traditional vascular accesses, AVF,orPTFEgrafts. In
asurveyof17dialysis unitsintheChicago area,the
proportion ofpatients usingpermanent double-lumen
catheters ranged between 3.3and45%;of1372pa-
tients, 210(15%) received oralready hadsuchcathe-
ters(74).Increasing long-term survival oftheend-
stagerenalpopulation hasenhanced theincidence of
depletion ofpotential sitesforAVvascular access, and
theuseofcentral venous catheters forpermanent
access islikelytobeincreasing.
Ajugular veinhasbecome thepreferred sitefor
permanent central venous catheter insertion. Percu-
taneous jugular venous insertion issaferthaninthe
subclavian location. Subclavian catheters have
caused subclavian thrombosis andstenosis (75),
which canpreclude creation ofAVvascular access in
theipsilateral ann,andtheremaybelessmechanical
malfunction incatheters placed inthejugular rather
thansubclavian location (76),partIcularly theright
jugular veins (77).Blood-flow rates of200to300
mL/min areroutinely achieved (78-80); flowsupto
400mL/min canbeachieved insome patients, but
sometimes notconsistently. Recirculation invarious
studies hasranged from5.5to8.6%(79-82). Unlike
AVfistulas andgrafts, dialysis through central venous
catheters isnotaffected bycardiopulmonary recircu-
lation.
Actuarial catheter survival at12months hasranged
from30to65%(76,79,80,83). Catheter failure has
beentheresult primarily ofmalfunction (thrombosis
orpoorblood flow)ortoinfection, inroughly equal
proportions. Catheter malfunction canbereduced by
proper placement ofthecatheter tipintheright
atrium andbytheuseofchronic anticoagulation,
beginning withaspirin andusingwarfarmn ifnecessary
(74).Urokinase andstreptokinase andmechanical
means havebeenusedtoopenthrombosed catheters
andthosewithpoorflow(74,78-80,84-86). Infection
withsepticemia occurs withafrequency between 0.25
and1.0infections perpatient-year (76-79). Although
Gram-positive infections havebeensuccessfully
treated withantibiotics alone,manyinfections, espe-
ciallythoseresulting fromGram negative organisms,
haverequired catheter removal.
Useofpermanent central venous catheters has
important morbidity. Insertion canbecomplicated by
arterial puncture, pneumothorax, hemothorax, ar-
rhythmia, andperforation orlaceration ofthebrachial
plexus, trachea, superior venacava,ormyocardium.
Jugular venous cannulation Isgenerally saferthan
subclavian cannulation andtheincidence ofthese
complications islowinexperienced hands (73).Late
complications include catheter malfunction because
ofthrombosis ormalposition, infection, andsubcla-
vianorsuperior venacavathrombosis orstricture
(73,87,88). Thelattercomplications, which occurwith
internal jugular aswellassubclavian catheters, may
prevent subsequent useoftheipsilateral armorof
botharmsforAVaccess. Superior venacavathrom-

Feldman etal
Journal oftheAmerican Society ofNephrology 529bosismaybemorecommon thanhasbeenreported
(88),buttheactual ratesofoccurrence ofthesemor-
bidities areunknown.
Permanent dual-lumen central venous catheters
havesomeadvantages overAVaccess, including ease
ofinsertion, removal, andreplacement, immediacy of
use,absence ofhemodynamic stress orstealsyn-
drome, absence ofcardiopulmonary recirculation,
andavoidance ofvenipuncture. Indications forsuch
catheters include lossofothervascular access sites,
prolonged period ofmaturation ofother accesses
(usually AVF), cardiovascular instability orstealsyn-
drome, andshort expected duration ofdialysis (e.g.,
malignancy, impending renal transplantation). If
ratesofpoorflowandthrombosis andinfection areas
lowasinsomeseries, permanent catheters mayeven
beconsidered anacceptable alternative toPTFEgrafts
inthegeneral population ofpatients inwhom creation
ofanAVFisnotfeasible (80).
PATFERNS OFVASCULAR ACCESS UTILIZATION IN
THEUNITEDSTATES
Theonlynational datapermitting study ofspecific
typesofvascular access werecollected aspartofthe
special studies ofCaseMixconducted bytheUSRDS,
firstforpatients in1986and1987,andsubsequently
forpatients in1991.Although these dataareolder,
theyprovide uswithanopportunity toexamine the
evolution ofthepatterns ofvascular access utilization
C
4,
C.,
4,a.intheUnited States thathavenotbeenpreviously
published.
ByusingdatafromtheUSRDS 1986to1987Special
Study ofCaseMix,weareabletoestimate thedistri-
bution ofvascular access typeintheUnited States.
Thisstudy examined theinitial medical records fora
nearly random sample of4964U.S.dialysis patients
initiating dialysis in1986and1987. Among them,
1518wereidentified asbeginning hemodialysis with
anAVF,1498withaVTFEgraft,and77withacentral
venous catheter. Theremainder ofthepatients either
initiated peritoneal dialysis orcould notbecatego-
rizedintoonespecific vascular access group. Asis
demonstrated inFigures 2and3,olderageandfemale
gender wereassociated withlowerutilization ofAVF in
1986and1987. Asimilar analysis failed todemon-
strate anassociation between AVFutilization and
either raceorthepresence ofdiabetes mellitus.
The1991USRDS Special Study ofCaseMixdem-
onstrated thatthepattern ofAVFutilization had
substantially changed overthespanof4to5yr.This
study, which analyzed theinitial vascular access
among 1673 randomly selected hemodialysis pa-
tients, demonstrated thatAVFutilization among dia-
betics hadfallensubstantially fromover40%toabout
25%(seeFigure 4).Similarly, buttoalesser extent,
AVFutilization hadalsofallen among ESRD patients
without diabetes (seeFigure 4).Although these
changes maybeareflection offurther increases inthe
morbidity oftheESRD population, leading toless
.<19yrs
#{149}20-44yrs

45-64yrs
D65-74yrs
D75+yrs
AVFistula PTFEGraft
Figure2.Vascular accesstypebyage:datafromUSRDSSpecialStudyofCaseMix:1986to1987.

Hemodialysis Vascular AccessMorbidity
530 Volume 7.Number 4‘1996C
4,
C.,
I-
4,a.
AVFistula PTFEGraftOMale
#{149}Female
Figure3.Vascular accesstypebygender: datafromUSRDSSpecial StudyofCaseMix:1986to1987.
successful efforts toplaceAVF,thesetrends mayalso
reflect changes inpatterns ofpractice thatfavorplace-
mentofsynthetic vascular accesses available foruse
almost Immediately afterinsertion. Finally, theesti-
mates ofAVFutilization defined hereamong incident
ESRD patients probably overstate theprevalence of
AVFuseamong allhemodialysis patients. Windus has
recently estimated thatapproximately 11%ofAmen-
canhemodialysis patients in1990usedAVFforan-
gioaccess (38).
CLINICAL SELECTION OFVASCULAR ACCESS
MODALITY: WHATARETHETRADEOFFS?
Numerous options currently existforachieving vas-
cular access forESRD patients today. Selection
among theseoptions requires evaluation ofthepoten-
tialtoxicitles andbenefits ofeach oftheaccess
choices, soastounderstand thetradeoffs inherent in
choosing aparticular access method (seeFigure 5).
Datacurrently available strongly suggest thatwhen
technically feasible, radial AVFrepresent theoptimal
solution forvascular access. Identification ofpatients
inwhom aradial AVFwould mature, however, is
difficult withcurrently available information. At-
tempting toplace radial AVFbefore theneedfor
dialytic therapy would minimize themorbidity and
costsassociated withuseofcentral venous catheters
while awaiting AVFmaturation. Evenwhenaradial
AVFfallstomature adequately, aVFFE graftora
brachial AVFcanusually beplaced intheipsilateralann.Success oftheseaccesses mayevenbeenhanced
byvenous dilatation thatmayhavearisen asaresult
oftheinitial radial AVF.Selection ofthepatients in
whom aradial AVFshould beattempted, however,
oftenpresents amoredifficult decision whenpatients
require immediate orimminent dialysis. Under these
circumstances, boththehighprobabifity ofunsuc-
cessful maturation especially among olderandsicker
patients, andthemorbidity associated withcentral
venous catheters necessary fordialysis whilewaiting
forAVFmaturation, leadmanyclinicians directly to
analternative access choice. Usually theirchoice isa
VrFE graft, which permits dialysis verysoonafter
placement. Better delineation ofboththeprofile of
patients likelytoexperience successful maturation of
aradial AVFaswellasquantification ofthemorbidity
associated withtemporary central venous catheters
necessary forexpanded utilization ofAVFareneeded
todetermine theappropriate clinical thresholds for
AVFuse.
Currently, threedistinct alternatives existtoradial
artery AVF;brachial AVF,VFFE bridge grafts, and
permanent indwelling catheters. Although available
datademonstrate thatradial AVFaresuperior to
VT’FE grafts andcatheters, theappropriate roleof
brachial artery AVFIsstillnotentirely defined. Al-
though theseaccesses arelikelytoenjoysomeofthe
benefits ofradial AVF(e.g.,onlyoneanastomosis is
required andmanyinfections aretreatable without
access removal), theirlong-term patency isuncertain.

Graft
Figure4.Evolution ofvascular access useintheUnitedStates:diabeticandnondlabetic hemodlalysis patients<65yrofage.
DatafromUSRDSSpecial Studies ofCaseMix:1986to1987,and1990.Feldman etal
Journal oftheAmerican Society ofNephrology 531C,
U _________
L.



C, IO1986-7

a.
I
_0
Diabetic – Diabetic – Nondiabetic Nondiabetic
AVFistula PTFEGraft -AVFistula -PTFE
Although several smallstudies suggest thatpatency of
brachial artery AVFatleastrivalsVT’FEgraftpatency,
moreresearch isneeded. Whether problems suchas
stealsyndrome andheart failure secondary tothe
highratesofblood flowthrough these accesses will
limittheiruseisstillunknown. Despite these un-
knowns, wecurrently believe thatthepreponderance
ofevidence supports attempting toplace abrachial
AVFbefore aPTFEgraftinpatients inwhom aradial
AVFiseither notpossible orhasfailed.
Inthetradeoff between useofaPTFEgraftanda
permanent central venous catheter forangloaccess,
several considerations mustbeweighed. Permanent
dual-lumen central venous catheters havesomead-
vantages overFTFEgrafts, including easeofinsertion,
removal, andreplacement, immediacy ofuse, absence
ofhemodynamic stress orstealsyndrome, absence of
cardiopulmonary recirculation, andavoidance ofye-
nipuncture. However, thepermanent venous catheter
maynotsupport blood flowatrates adequate to
ensure optimum dialysis; thisproblem, together with
lossoftimebecause oftheneedfordeclotting with
urokinase, could leadtounderdialysis. Episodic poor
flowoftenmakes useofthesecatheters frustrating for
technical staff.However, despite highratesofpoor
flow,thrombosis, andinfection insomeseries, perma-
nentcatheters haveevenbeenconsidered analterna-
tivetoPTFEgrafts inthegeneral population ofESRD
patients inwhom creation ofanative AVFIsnotfeasible (80).Likecentral venous catheters, PTFE
grafts similarly arecomplicated bythromboses and
highrecirculation fromstenoses, leading topossible
underdialysis andnecessitating anatomical correc-
tion(e.g.,angioplasty, surgical revision). BothPTFE
grafts andcatheters mayrequire anticoagulation,
adding ariskofhemorrhage, including subdural he-
matoma. Although infection ofacentral venous cath-
etermaybetreated withantibiotics without catheter
lossandcatheter replacement isrelatively easy,infec-
tionofaPTFEgraftalmost always requires graft
removal, whichcanentailsignificant morbidity.
Despite theproblems associated withPTFE grafts,
onefinaldistinction between themandcentral venous
catheters weighs mostheavily inouropinion that
catheters notbeusedasaprimary permanent access
wheneither anAVForVFFEgraftcanbeplaced.
Central venous stricture orthrombosis subsequent to
placement ofacentral venous catheter mayeliminate
allpotential access sitesinoneorbotharms. Given
theoftencatastrophic impact ofthislossofupper-
extremity access sites, thePTFE graft, despite its
complications, ispreferable tothecentral veincathe-
tenusedasapermanent access.
SUMMARY
Vascular morbidity IntheUnited States issubstan-
tialandisgrowing faster thanIsattributable tothe

Principal Vascular AccessChoice
Radial
AVF
I Pros
J ConsN
FirstAlternatives toRadial
AVF
Longer Survival
Highrateofprimary failure
FewerComplications
Delayed availability ofaccess

J Complications oftemporary
useofcentral venous
cathetersSecondary Alternative toRadial
AVF
Brachial ArteryArteriovenous Fistulae Permanent Central Venous Catheter
Pros Cons
Feasibly placed in
virtually allpatients
Available immediately
Nohemodynamic stress
orsteal
Avoids venipunctureHemodialysis Vascular Access Morbidity
532 Volume 7‘Number 4‘1996Vascular AccessSelection: WhatAretheTradeoffs?
Pros Cons
Feasibly placed inmany Unknown long-term
patients not function
candidates forradial Highrateofflow may
AVFs leadto
Mayhavensorbidity cardiovascular
profilesimilar to instability
radialAVFsPTFEAVBridgeGraft
Pros Cons
Feasibly constructed inHighrateofocclusive
mostpatients failure
Available foruse Infective failure
typically within2 typically demands
weeks removal
Shortened secondary
patency vs.AVFsMaycausecentral
venous stenosis,
limiting subsequent
accessprocedures
Highrateofinfective
andocclusive failure
Figure5.Vascular accessselection: whatarethetradeoffs?
growth oftheESRD population orgrowth inother
morbidity among ESRD patients. Morbidity fromac-
cessdysfunction continues togrowfaster thansug-
gested byavailable dataonhospitalization forvascu-
laraccess dysfunction, Inpartbecause oftheshiftto
outpatient therapies suchasthrombolysis andanglo-
plasty. Thegrowth invascular access morbidity may
bearesult oftherisingageoftheESRD population, its
worsening severity ofillness, increasing demands on
vascular accesses totolerate greater blood-flow rates,
andincreasing diagnostic activities focused ondetect-
ingoccult access dysfunction.
Inaddition tothegrowing ageandcomorbidity of
patients independent ofsurgical technique, anevolv-
ingpattern ofsurgical practice awayfromthecreation
ofautologous AVFtotheconstruction ofsynthetic
PTFEdialysis vascular access grafts isalsoresponsi-
blefortheincrease invascular access-related morbid-
ity.Thesubstitution ofPTFE grafts forAVFcontrib-
utes totheextensive access-related morbidity
currently confronting dialysis patients andproviders.
Although these synthetic accesses typically function
immediately aftertheirconstruction, theirrateoflater
secondary failure ismuch higher thanthatforradial
AVF.Agreatchallenge before thedialysis communityistheidentification ofpatients inwhom anAVFisa
viablevascular option sothatshorter-lived synthetic
(VrFE) vascular access grafts canbereserved for
patients inwhom anAVFisunlikely tomature suc-
cessfully. Toaccomplish this,theidentification ofrisk
factors forinadequate maturation ofAVFintolong-
lasting hemodialysis vascular accesses mustbeahigh
priority. Inaddition, itisessential thatprospective
recipients ofAVFbeidentified earlyandthatevery
effort bemade toavoid venipuncture inthearm
intended forthevascular access.
PTFEAVbridge grafts areaccompanied byahigh
rateofcomplications primarily fromstenosis and
occlusion, andtoalesser extent, infection. Riskfac-
tonsforearly PTFE access complications, too,are
incompletely defined. Suchdataareneeded toprevent
notonlythemorbidity ofvascular access dysfunction
perse,butpotential undendialysis fromaccess dys-
function thatrepresents amostimportant, butoften
hidden andinsidious, costofvascular access morbid-
ity.
Permanent central venous catheters maynotonly
represent anaccess oflastresortwhenotheraccess
siteshavebeenexhausted, butmayofferanaltenna-
tivetoPTFE grafts. Thepopulation inwhom perma-

Feldman etal
Journal oftheAmerican Society ofNephrology 533nentcentral venous catheters maybeasuperior
alternative toPTFEgrafts, evenwhen sitesforgrafts
havenotbeenexhausted, needs tobebetter defmed in
controlled studies.
Clinical strategies tooptimize thediagnosis and
therapy ofvascular access dysfunction alsoneedtobe
identified. These strategies mustdefine theappropri-
atethreshold ofabnormal morphology andfunction
thatjustifies radiologic and/or surgical intervention.
Clearly, earlyintervention maypreserve thesecond-
arypatency ofsynthetic vascular access grafts. How-
ever,aggressive diagnostic activities plausibly may
leadtoahigher rateofradiologic andsurgical inter-
ventions fordetectable butnotyetclinically important
stenoses. These interventions (e.g. ,angioplasty or
surgical repair) themselves mayreduce thefeasibifity
oflaterinterventions andmaydamage thevascular
access directly, thereby shortening itslong-term func-
tion.
Finally, identification ofthefundamental mecha-
nisms ofvascular access dysfunction should alsobe
thefocus ofinvestigation sothatspecific therapies
andpreventive strategies canbedeveloped andimple-
mented.
ACKNOWLEDGMENTS
Theauthors gratefully acknowledge theassistance oftheU.S.Renal
Data System forproviding datafromtheirSpecial Studies ofCase Mix.
Thisworkwassupported, inpart.byaCooperative Agreement with
theHealthCareFinancing Administration (HCFA).
REFERENCES
1.Quinton W,Dullard D,Scribner BH:Cannulation ofblood
vessels forprolonged hemodlalysis. TransAmSocArtif
Intern Organs 19606:104-113.
2.Rosenthal JJ:Vascular access forhemodialysis. Major
ProblClinSurg1981;4:455-468.
3.Wilson SE,Stabile BE,Williams RA,Owens ML:Current
status ofvascular access techniques. SurgClinNorth
Am1982;62:531-551.
4.Brescia MJ,Cimino JE,AppelK,Hurwich BJ:Chronic
hemodialysis usingvenipuncture andasurgically cre-
atedarteriovenous fistula. NEngiJMed1966;275:
1089-1092.
5.Bonalumi U,Civalleri D,Rovida S,Adami GF,Gianetta
E,Griffanti-Bartoli F:Nineyears’experience withend-
to-end artenlovenous flstula attheanatomical snufibox
formaintenance haemodlalysis. BrJSurg1982;69:486-
488.
6.Brems J,Castaneda M,Garvin PJ:Afive-year experience
withtheBovine Heterograft forvascular access. Arch
Surg1986:121:941-944.
7.Giacchino JL,GeisWP,Buckingham JM,Vertune LL,
Bansal VK:Vascular access: Long-term results, new
techniques. ArchSung1979;114:403-409.
8.Kherlakian GM,Roedersheimer LR,Arbaugh JJ,New-
markLI,KingLR:Comparison ofautogenous fistula
versus expanded polytetrafluoroethylene graftfistula for
angioaccess inhemodialysis. AmJSurg1986;152:238-
243.
9.Munda R,FirstMR,Alexander JW,Linnemann CC,
FidlerJP, KitturD:Polytetrafluoroethylene graftsurvival
inhemodialysis. JAMA 1983;249:219-222.
10.Rapaport A,NoonGP,McCollum CH:Polytetrafluoro-
ethylene (VI’FE) graftsforhaemodlalysis inchronic renal
failure: Assessment ofdurability andfunction atthree
years.AustNZJSurg1981:51:562-567.
11.VeithFJ,Wilson SE,Hobson RW,Rosenthal JJ,Tellis
VA,Dagher FJ:Vascular access complications andnewmethods. TransAmSocArtifIntern Organs 1982;28:
647-657.
12.OfficeofStatistics andDataManagement. End-stage
renaldisease patient profile tables, 1987.Baltimore:
OfficeofStatistics andDataManagement, Health Care
Financing Administration; 1987.
13.BellDD,Rosental JJ:Arteniovenous graftlifeinchronic
hemodlalysis. ArchSurg 1988; 123:1169-1172.
14.Schuman ES,GrossGF,HayesJF,Standage BA:Long-
termpatency ofpolytetrafluoroethylene graftflstulas.
AmJSurg1988:155:644-646.
15.Haimov M,Burrows L,Sclianzer H,etat.:Experience
witharterial substitutes intheconstruction ofvascular
access forhemodialysis. JCardiovasc Surg1980:21:
149-154.
16.Mohaideen AR,Mendivil J,Avram MM,Malnzer RA:
Artenlo-venous access utilizing modified bovine arterial
graftsforhemodlalysis. SungClinNorthAm1977:62:
531-550.
17.Winsett OE,Wolma FJ:Complications ofvascular ac-
cessforhemodialysis. SouthMedJ1985:78:513-517.
18.Kester RC:Arteniovenous graftsforvascular access in
hemodialysis. BrJSurg1979;66:23-28.
19.Feldman HI,HeldPJ,Hutchinson T,Stoiber E,Hartigan
MF,Berlin JA:Hemodialysis vascular access morbidity
intheU.S.Kidney Int1993:43:1091-1096.
20.Lazarus JM,Huang WI!,LewNL,Lowrie EG.Contnibu-
lionofvascular access-related disease tomorbidity of
hemodlalysis. Vascular Access forHemodialysis III.Chi-
cago:WLGore&Assoc. ,Inc. ,andPrecept Press, Inc.:
1993:3- 13.
21.HeldPJ,PaulyMV,Bovbjerg R,Diamond L:Largeand
smalltransplant programs: Implications forcostand
outcomes. IPublication No.3293-22] WashIngton, DC:
TheUrban Institute; 1986.
22.USRenalDataSystem. USRDS 1994Annual DataRe-
port.Bethesda, MD:TheNational Institutes ofHealth,
National Institute ofDiabetes andDigestive andKidney
Diseases: 1994:23-42.
23.Churchill DN,Taylor DW,CookRJ,LaPlante P,etat.:
Canadian hemodialysis morbidity study. AmJKidney
Dis1992:19:214-234.
24.Kinnaert P.Vereerstraeten P,Toussaint C,VanGeer-
truyden J:Nineyears’experience withinternal artenio-
venous fistulas forhaemodialysls: Astudyofsomefac-
torsinfluencing theresults. BrJSurg1977:64:242-246.
25.ReillyDT,WoodRFM,BellPRF:Prospective studyof
dialysis fistulas: Problem patients andtheirtreatment.
BrJSurg1982:69:549-553.
26.HIggins MR.Grace M,Bettcher KB,Silverberg DS,
DossetorJB: Bloodaccess Inhemodialysis. ClintNephrol
1976:6:473-477.
27.AimA,Lundberg M:Clinical experience withtherapeutic
arteriovenous flstulae. Scand JUrolNephrol 1977;11:
53-57.
28.RohrMS.Browder W,Frentz GD,McDonald JC:Arterio-
venous flstulas forlong-term dialysis. ArchSurg1978;
113:153-155.
29.Dunlop MG,Mackinlay JY,Jenkins AM:Vascular ac-
cess:Experience withthebrachiocephalic fistula. AnnR
CollSungEngl1986:68:203-206.
30.ZIbari GB,RohrMS.Landreneau MD,Bridges RM,et
aL:Complications frompermanent hemodialysis vascu-
laraccess. Surgery (St.Louis) 1988;104:681-686.
31.Bourquelot P,Cussenot 0,CorbiP,etat.:Microsurgical
creation andfollow-up ofarteriovenous flstulae for
chronic haemodialysis inchildren. Pediatr Nephrol
1990:4:156-159.
32.Wehrli H,Chenevard R,Zaruba K:Chlrurgische Erfahr-
ungenmitarterio-venosen Hamodialyse-Shunts (1970-
1988).HelvChirActa1989:56:621-627.
33.Coburn MC,Carney WIJr:Comparison ofbasilic vein
andpolytetrafluorethylene forbrachial arteriovenous
fistula. JVascSurg1994:20:896-902..
34.RIvers SP,Scher LA,Sheehan E,LynnR,VeithFJ:
Basific veintransposition: Anunderused autologous
alternative toprosthetic dialysis angioaccess. JVase
Surg18:391-396.

Hemodialysis Vascular Access Morbidity
534 Volume 7.Number 4.199635.Bender MH,Bruyninckx CM,Genlag PG:Thebrachioce-
phalicelbowfistula: Ausefulalternative angioaccess for
permanent hemodialysis. JVase Surg 1994:20:808-
813.
36.Pourchez T,Moriniere F,StPriestA,etat:Outcome of
vascular access forhemodialyses intheelderly (>75
years)lAbstracti. ProcEurDialTransplant Assoc1990:
27:160A.
37.Kobrln 5,Feldman H,Morley C,DenoM:Outcome of
arteniovenous fistula (AVF) placements (Abstract).
ASAIO 38thAnnual Meeting: 1992:84A.
38.Windus DW:Permanent vascular access: Anephrolo-
gist’sview.AmJKidney Dis1993:21:457-471.
39.Besarab A,Medina F,Musial E,etaL:Recombinant
human erythropoietin doesnotincrease clotting invas-
cularaccesses. ASAIO (AmSocArtifIntern Organs)
Trans1990;36:M749-M753.
40.TangTY,Vrahnos D,Valaitis D,LauAR:Vascular access
thrombosis during recombinant human erythropoietin
therapy. ASAIO J1992:38:M528-M531.
41.Palder SB,Kirkman RL,Whittenmore AD,Hakim RM,
Lazarus JM,Tilney NL:Vascular access forhemodialy-
sis.AnnSurg1985:202:235-239.
42.Tauches LA:Immediate, safehemodialysis witharterio-
venous fistulas created withanewtunneler. AnnSung
1985:150:212-215.
43.Raju5:PTFEgraftsforhemodlalysis access. AnnSurg
1987:206:73,206-266.
44.Mandel SR.Martin PL,Blumoff RL,Mattern WD:Vas-
cularaccess Inauniversity transplant anddialysis
program.ArchSurg1977:112:1375-1380.
45.Sabanayagam P.Schwartz AR,Soricelli RR,Chinitz JL,
Lyons P:Experience withonehundred reinforced ex-
panded PTFEgraftsforangioaccess inhemodialysis.
TransAmSocArtifInternOrgans 1980:26:582-583.
46.KraevZ,Kirjakov Z,Antonov S.Stantchev I,Zlatarska
S,Sofia:Polytherafluoroethylene graft’scomplications in
patients onhemodialysis. PrzeglLek1985:42:442-443.
47.TellisVA,Kohlberg WI,BhatDJ,Dniscoll B,VeithFJ:
Expanded polytetrafluoroethylene graftfistula for
chronic hemodialysis. AnnSurg1979:189:101-105.
48.Beathard GE:Percutaneous transvenous angioplasty in
thetreatment ofvascular access stenosis. Kidneymt
1992:42:1390-1397.
49.Schwab SJ,Raymond JR,SaeedM,Newman GE,Den-
nisPA,Bollinger RR:Prevention ofhemodialysis fistula
thrombosis. Earlydetection ofvenous stenoses. Kidney
mt1989:36:707-711.
50.Windus DW,Audrain J,Vanderson R,Jendnisak MD,
Picus D,Delmez JA:Optimization ofhigh-efficiency
hemodialysis bydetection andcorrection offistula dys-
function. Kidney Int1990:38:337-441.
51.SandsJ,Young 5,Miranda C:TheeffectofDoppler flow
screening studies andelective revisions ondialysis ac-
cessfailure. ASAIO (AmSocArtifIntern Organs) Trans
1992:38:M524-M527.
52.BellDD,Rosental JJ:Arteriovenous graftlifeinchronic
hemodlalysis. ArchSurg1988:123:1169-1172.
53.Haimov M,Burrows L,Schanzer H,NeffM,BaezA,
KwunK,Slkfkin R:Experience witharterial substitutes
intheconstruction ofvascular access forhemodialysis.
JCardiovascSurg1980:21:149-154.
54.Aman LC,LevinNW,Smith DW:Hemodialysis access
sitemorbidity. ProcClinDialTransplant Forum 1980:
10:277-284.
55.Schwab SJ,Quarles D,Middelton JP,Cohan RH,Saeed
M,Dennis VW:Hemodialysis-associated subclavian vein
stenosis. Kidney Int1988:33:1156-1159.
56.Kherlakian GM,Roedersheimer LR,Arbaugh JJ,New-
markKJ,KingLR:Comparison ofautogenous fistula
versus expanded polytetrafluoroethylene graftfistulafor
angloaccess inhemodialysis. AmJSung1986:152:238-
243.
57.Munda R,FirstMRAlexander JW,Linnemann CC,
Fidler JP,Kittur D:Polytetrafluoroethylene graftsur-
vivalinhemodialysis. JAMA 1983:249:219-222.
58.Winsett OE,Wolma FJ:Complications ofvascular ac-
cessforhemodialysis. SouthMedJ1985:78:513-517.59.KraevZ,Kirjakov Z,Antonov 5,Stantchev I,Zlatarska
5,Sofia:Polytherafluoroethylene graft’scomplications in
patients onhemodialysis. PrzeglLek1985:42:442-443.
60.Rosenthal JJ:Vascular access forhemodialysis. Major
ProblClinSurg1981:4:455-468.
61.TellisVA,Kohlberg WI,BhatDJ,Driscoll B,VeithFJ:
Expanded polytetrafluoroethylene graftfistula for
chronic hemodialysis. AnnSung1979:189:101-105.
62.DyGR,Bloom EJ,IjeluGK,Merritts GW,Kramer MS,
RajaRM:Effectofrecombinant human erythropoietin on
vascular access. ASAJO (AmSocArtifIntern Organs)
Transactions 1991;37:M274-M275.
63.Garcia-Martin F,DeArriba G,Carrascosa T,etat.:Anti-
cardiolipin antibodies andlupusanticoagulant inend-
stagerenaldisease. Nephrol DialTransplant 1991:6:
543-547.
64.PageB,ZingraffJ, Houhou 5,etat.:Clotting ofvascular
access indialysis patients associated toanticardiolipin
antibodies lAbstracti. JAmSocNephrol 1992;3:385A.
65.Windus DW,Jendrisak MD,Dalmez JA:Prosthetic fis-
tulasurvival andcomplications inhemodialysis pa-
tients: Effects ofdiabetes andage.AmJKidney Dis
1992:20:448-452.
66.USRenalDataSystem. USRDS 1991Annual DataRe-
port.Bethesda, MD:TheNational Institutes ofHealth,
National Institute ofDiabetes andDigestive andKidney
Diseases; 1991.
67.Eschbach JW,Abdulhadi MH,Browne JK,etat.:Re-
combinant human erythropoietin inanemic patients
withend-stage renaldisease. AnnIntern Med1989:11:
992-1000.
68.Sundal E,Businger J,Kappeler A:Treatment oftrans-
fusion dependent anemia ofchronic renalfailure with
recombinant human erythropoietin. Nephrol DialTrans-
plant1991:6:955-965.
69.Bennett WM:Amulticenter clinical trialofepoetin beta
foranemia ofend-stage renaldisease. JAmSocNephrol
1991;1:990-998.
70.Kawalski EA,Trenkle JA,NatkeE,etaL:Theeffectof
erythropoietin ontheincidence ofchronic vascular
thrombosis (Abstracti. JAmSocNephrol 1991:2:380.
71.Canadian Erythropoietin StudyGroup. Association be-
tween recombinant human erythrepoietin andquality
oflifeandexercise capacity ofpatients receiving hemo-
dialysis. BMJ1990;300:573-579.
72.Muirhead N:Erythropoietin asacauseofaccess throm-
bosis.SeminDial1993:6:184-188.
73.Bander SJ,Schwab SJ:Central venous angioaccess for
hemodialysis anditscomplications. Semin Dial1992:5:
121-128.
74.UldailR,Besley ME,Thomas A,Salter S,Nuezca LA,
VasM:Maintaining thepatency ofdouble-lumen silastic
jugular catheters forhaemodialysis. mtjArtifOrgans
1993;16:37-40.
75.FantGF,Dennis V,Quarles LD:Latevascular complica-
tionsofthesubclavlan dialysis catheter. AmJKidney
Dis1986:7:225-228.
76.Swartz RD,Messana JM,BoyerCJ,Lunde NM,Weitzel
WF,Hartman TL:Successful useofcuffed central ye-
noushemodialysis catheters inserted percutaneously. J
AmSocNephrol1994:4:1719-1725.
77.DcMeester J,Vanholder R,DeReese J,Ringoir 5:
Factors andcomplications affecting catheter andtech-
nique survival withpermanent single-lumen dialysis
catheters. Nephrol DialTransplant 1994:9:678-683.
78.UldallR,DeBruyne M,Besley M,McMillan J,Simons M,
Francoeur R:Anewvascular access catheter forhemo-
dialysis. MnJKidney Dis1993:21:270-277.
79.Shusterman NH,KlossK,Mullen JL:Successful useof
double-lumen, silicone rubber catheters forpermanent
hemodialysis access. Kidneymt1989:35:887-890.
80.MossAH,Vasilakis C,Holley JL,Foulks CJ,PillalK,
McDowell DE:Useofasilicone dual-lumen catheter with
aDacron cuffasalong-term vascular access forhemo-
dialysis patients. AmJKidney Dis1990:16:211-215.
81.Blake PG,Huraib 5,WuG,UldallPR:Theuseofdual
lumenjugular venous catheters asdefinitive longterm

Feldman etal
Journal oftheAmerican Society ofNephrology 535access forhaemodialysis. IntJArtifOrgans 1990:13:26-
31.
82.Schanzer H,Kaplan 5,BoschJ,Glabman 5,Burrows L:
Double-lumen, silicone rubber, indwelling venous cath-
eters.Anewmodality forangioaccess. ArchSurg1986:
121:229-232.
83.Shaffer D,Madras PN,Williams ME,D’EliaJA,Kaldany
A,Monaco AP:UseofDacron cuffedsilicone catheters as
long-term hemodialysis access. ASAIO J1992:38:55-58.
84.Dunea G,Domenico L,Gunnerson P,Winston-Willis F:
Asurvey ofpermanent double lumen catheters inhemo-
dialysis patients. ASAIO (AmSocArtifIntern Organs)
Trans1991:37:M276-M277.
85.SchwabSJ,BullerGL,McCannRL,BollingerRR,
Stickel DL:Prospective evaluation ofaDacron cuffedhemodialysis catheter forprolonged use.AmJKidney
Dis1988:11:166-169.
86.Shrivastava D,Lundin AP,Dosunmu B,RaeTKS,Beyer
MM,Friedman EA:Salvage ofclottedjugular veinhemo-
dialysis catheters. Nephron 1994:68:77-79.
87.Caruana RI,RajaRM,ZeitRM,Goldstein SJ,Kramer
MS:Thrombotic complications ofindwelling central
catheters usedforchronic hemodialysis. AmJKidney
Dis1987:9:497-501.
88.GroteJ,LuiftV.Nikkuta PVD,LiethH,Bahlmann J,
Daniel WG:Transesophageal echocardiographic assess-
mentofsuperior venacavathrombosis inpatients with
long-term central venous hemodialysis catheters. Clin
Nephrol 1994:42:183-188.