Journal of the American Society of Nephrology 523

Hemodialysis Vascular Access Morbidity1Harold I. Feldman,2 Sidney Kobrin, and Alan Wasserstein

In a whom radial arteriovenous fistula is a viable

HI, Feldman, S. Kobrin, A. Wasserstein, Department ofMedicine, Renal Electrolyte and Hypertension Divison,University of Pennsylvania School of Medicine, Phila-delphia, PA

HI, Feldman, Department of Biostatistics and Epidemi-ology, and Center for Clinical Epidemiology and Bio-statistics, Universily of Pennsylvania School of Medi-cine, Philadelphia, PA

(J, Am, Soc. Nephrol. 1996; 7:523-535)

ABSTRACTComplications associated with hemodialysis vascularaccess represent one ofthe most important sources ofmorbidity among ESRD patients in the United Statestoday. In this study, new data on the magnitude andgrowth of vascular access-related hospitalization inthe United States is presented, demonstrating that thecosts of this morbidity will soon exceed $1 billion peryr. This study also reviews published literature on themorbidity associated specifically with native arterio-venous fistulae, polytetrafluoroethylene bridge grafts,and permanent central venous catheters. Next, newinformation on the changing palterns of vascularaccess type in the United States is presented, demon-strating the continuing evolution of medical practiceaway from the use of arteriovenous fistulae in favor ofmore reliance on synthetic bridge grafts. Based onthese data, a discussion is provided of the tradeoffsamong the most commonly available modalities ofvascular access today. Although radial arteriovenousfistulae continue to represent the optimal access mo-dality, the appropriate roles for brachial arterio-venous fistulae, synthetic bridge grafts, and centralvenous catheters are less certain because of made-quate data on the long-term function of the first andthe high rates of complications associated with thelatter two. To reduce vascular access-related morbid-ity, strategies must be developed not only to preventand detect appropriately early synthetic vascularaccess dysfunction, but to better identify the patients

1 ReceIved June 2, 1995. Accepted January 3, 1996.

2 Correspondence to Dr. H.!. Feldman, Center for Clinical Epidemiology and

Biostatistics, University of Pennsylvania School of Medicine, 720 Blockley Hall, 423ServIce Drive, Philadelphia, PA 19104.6021.

104&667310704.0523$03.00/0Journal of the American Society of NephrologyCopyright C 1996 by the American Society of Nephrology

clinical option.

Key Words: Arteriovenous fistula, polytetrafluorene bridgegraft, permanent central venous catheter, hospitalization,

ESRD

C hronlc hemodialysis for patients with ESRD first

became technically feasible in 1960 with theintroduction of the Quinton-Scribner shunt (1). Al-

though this external arteriovenous (AV) shunt pro-vided, for the first time, direct access to the circulation

so that hemodialysis could be performed in the out-

patient setting, patient care was complicated by fre-

quent thrombosis and infection (1-3). Autologous ax-

teriovenous fistulae (A’IF’), introduced by Brescia and

Cimino in 1966, overcame many of the problems of

external dialysis shunts (4). Subsequent to this ad-

vance in dialysis vascular access surgery, AVF be-came the primary mode of achieving vascular access

for chronic hemodialysis. Once an AVF has maturedand been used for dialysis, the subsequent or second-

ary failure rate is low, with most patients enjoying

long-term fistula function lasting for many years (3-

5).Since the early 1970s, the rapid growth of the U.S.

ESRD Program has been accompanied by a diminu-

tion in the use ofAVF in favor ofalternative types of AV

bridge grafts and, more recently, permanent indwell-ing central venous catheters. This evolution has been

attributed to the growing primary failure rate of AVF,

often attributed to the increasing age and comorbidity

of the dialysis population, as well as the more rapidusabifity of alternative vascular accesses as compared

with AVF (6-1 1). Supporting this formulation is the

extensive change in the the demographics of the U.S.ESRD population over the past two decades. Although

only 6.6% ofESRD patients had diabetes in 1972 andfewer than 20% of ESRD patients were older than 65

yr of age, 36% of the ESRD population in 199 1 had

diabetes and nearly 1 in 2 (45%) was 65 yr of age or

older (12). It has been estimated that fewer than 25%

of incident ESRD patients successfully undergo theconstruction of a native AV fistula, an estimate re-

cently borne out by special studies conducted by the

U.S. Renal Data System (10,13,14).

Early alternative surgical techniques utilizing autol-

ogous saphenous vein grafts and implanted bovine

carotid heterografts to form bridge AV anastomoses

had unacceptably high rates of thrombosis and pseu-

doaneurysm (2,3,6, 15-17). In an attempt to achieve

better long-term success for dialysis vascular ac-

cesses, the same synthetic materials used for artery to

Hemodialysis Vascular Access Morbidily

524 Volume 7 ‘ Number 4 ‘ 1996

TABLE 1 . Hospitalization for vascular access morbidity, 1984 to 19860

Diagnosis/Hospital Stay 1984 1985 1986

Total number of ESRD hospital stays for any diagnosis 150,775 158,634 170,572Frequency of hospital stays with ICD-9 996. 1b 6,079 5,755 5,953Frequency of hospital stays with ICD-9 996.#{243}c 5,83 1 6,837 7,889Frequency of hospital stays with ICD-9 996#{149}7d 1 1 791 13,452 15,899Total number of stays with 996.1, 996.6, or 996.7 listed 23,701 26,044 29,741

as any diagnosisTotal number of stays with 996. 1 , 996.6, or 996.7 listed 9, 170 21 01 6 24,025

as first diagnosisAccess-related stays as percentage of Total 15.7 16.4 17.4Mean length of access-related stays 7.6 7.3 7.4

a Adapted from data presented in Reference 19: Feldman HI, Held PJ. Hutchinson 1. Stoiber E. Hartigan, MF, Berlin JA: Hemodialysis vascular accessmorbidity in the U.S. Kidney mt 1993:43:1091-1096.b Mechanical complication of vascular device, implant. and graft.C Infection and inflammatory reaction as a result of internal prosthetic device, implant. and graft.d Other complications of internal prosthetic device, implant, and graft.

artery vascular bypass procedures were implanted

into dialysis patients as bridging AV anastomoses.

Dacron#{174} (E.I. duPont de Nemours and Co. , Wilming-ton, DE) was used initially chosen for these access

devices but was rapidly replaced by polytetrafluoro-

ethylene (PTFE) (GoreTex#{174}, W.L. Gore and Assoc.,

Flagstaff, AZ: Impra#{174}, Impra, Inc. , Tempe, AZ) after

early experience suggested better short-term patency

rates with this newer synthetic material (10, 1 8). As

will be discussed below, the substitution of synthetic

access grafts for AVF has been accompanied by enor-

mous growth in the rate of secondary vascular access

failure, along with its attendant costs.

Below. we provide new data from Medicare’s ESRD

Program and summarize published research describ-

ing the morbidity associated with hemodialysis vascu-

lar accesses in current use today in the United States,

including AVF, PTFE bridge grafts, and indwelling

central venous catheters. We also provide recent data

from the U.S. Renal Data System on the continuing

transition from AVF to alternative synthetic access

devices. Finally, we present a discussion of the

tradeoffs involved in choosing a type of vascular ac-

cess in the clinical setting, and the data needed to

establish appropriate guidelines for making this

choice.

VASCULAR ACCESS COMPLICATIONS IN THEUNITED STATES

We have previously reported on the magnitude of

vascular access morbidity in the United States by

using hospitalization claims data from Medicare’s

End-Stage Renal Disease database, the ESRD Pro-

gram’s Medical Management Information System

(ESRD PMMIS) (19). Vascular access complications

were tabulated by counting the number of hospital

stays for one of three ICD-9 diagnosis codes recorded

as the first diagnosis in the hospitalization record of

PMMIS.*

Prevalent Medicare-entitled U.S. ESRD patients in

1984, 1985, and 1986 were studied (see Table 1).

Complications of vascular accesses alone accounted

for one of the most important causes of hospitalization

in the ESRD population. Between 1984 and 1986, the

number of hospitalizations for revision of dialysis

vascular accesses rose by 25.0% and represented

14. 1% of all ESRD hospitalizations that year.

Data recently published by Lazarus et at. (20) from

the network of dialysis units managed by National

Medical Care (NMC) also examined vascular access-

related morbidity as reflected in hospitalization forvascular access complication-related diagnoses by us-

ing a similar methodology. Between 1986 and 1990,

Lazarus et aL studied over 25,000 NMC patients who

were monitored for vascular access-related hospital-

ization. The authors reported that the proportion of all

hospital days related to vascular access dysfunctiongrew from about 6% in 1988 to nearly 1 1% by the end

of 1990, suggesting that vascular access-related hos-

pitallzation is growing faster than hospitalization for

other types of morbidity in the ESRD population.

Recently we have extended our analyses character-

izing the burden of vascular access morbidity in the

United States between 1 986 and 1 99 1 on the basis ofdata from Medicare (see Figure 1). During this time

period, the number of vascular access-related stays

grew to over 70,000. Although this rise in access-related hospitalization emanates, in part, from the

growth in the U.S. ESRD population, Figure 1 alsodemonstrates that vascular access-related hospital-

ization as a percentage ofall-cause hospitalization has

also risen from approximately 1 7% in 1 986 to greater

* The three ICD-9 codes searched for included: 996.1, Mechanical complicationof vascular device, Implant, and graft 996.6, InfectIon and inflammatory reac-tion due to Internal prosthetic device, implant, and graft: and 996.7, Othercomplications of internal prosthetic device, Implant, and graft.

0.5

0.45

0.4

0.35

80000

70000

60000

50000

40000

30000

20000

10000

0

Figure 1 . Vascular access morbidity-related hospitalization In the United States and vascular access morbidity-relatedhospitalization among all U.S. dialysis patients as a proportion of all-cause hospitalization.

0.3

0.25

Cl,

E5,

.00a.(#5C’,4,UU

,(

CC’

U(I,CC’

>

0

CC,

0

CC’

.1:�

0.U,0

I

(C’

0I-C0

t00.0

0 a.

0.2

0.15

0.1

1986 1987 1988 1989 1990 1991

0.05

Feldman et al

Journal of the American Society of Nephrology 525

U,>.

U,‘I,U,0C.,U

U,

0

CUN

CU

0.U,0

I

than 20% in 199 1 . These estimates of hospitalization

because of vascular access complications are most

likely underestimates of access complications in theESRD population, as they do not consider access-

related morbidity diagnosed and treated in the outpa-

tient setting, outpatient procedures performed in a

hospital, or hospitalization paid for by insurers other

than Medicare. Although the disproportionate in-crease in morbidity may be a result of increasing

severity of illness among U.S. ESRD patients, the

increasing demand for accesses that can accommo-

date higher blood-flow rates as well as the use of new

diagnostic technologies such as color-flow Doppler

ultrasonography may also have increased the detec-

tion of heretofore occult vascular access problems.

Unfortunately, no single data source exists that pro-

vldes information on outpatient access-related mor-

bidity or morbidity paid for by insurers other than

Medicare. Medicare does maintain data ifies that

record bills for outpatient procedures such as fistu-lography and color-flow Doppler ultrasonography

studies, but to our knowledge, these data have notbeen exploited to estimate outpatient vascular access

morbidity.

Costs of Vascular Access Dysfunction in theUnited States

Approximately one-half of Medicare’s ESRD budget(over $6 billion in 199 1) is spent on the procedural

costs of delivering dialysis to ESRD patients (21,22).

The remainder of ESRD expenses arise primarily from

the costs ofmorbidity and associated hospitalizations,of which greater than 14% have been estimated to

result from complications of vascular access (19).

Assuming a similar rate of growth in vascular access-

related hospitalization as observed between 1986 and

199 1 , we project that Medicare will pay for more than

90,000 hospItalizations for vascular access-related

problems in 1995. If vascular access repair proce-

dures are increasingly performed in the outpatient

setting, the number ofhospitaJlzations may not rise as

quickly as it has In the past. Nonetheless, based on a

conservative estimated cost of $7500 per hospital

stay, the projected costs to Medicare for these hospi-talizations alone will rise above $675 million annually,representing nearly 10% of its ESRD budget. Thisanalysis represents a gross underestimate of the total

cost for vascular access-related morbidity in the

United States for at least two important reasons. First,

the hospitalization data cited do not include hospital-

ization paid for by insurers other than Medicare.

Second, the costs of diagnostic screening procedures(e.g., fistulography, Doppler ultrasonography, etc.)and outpatient therapies (e.g., thrombolysis, anglo-

plasty, etc.) are also not included. Thus, the annual

cost ofvascuiar access-related morbidity in the United

States will rapidly approach and exceed $ 1 billion per

yr.

Hemodlalysis Vascular Access Morbidity

526 Volume 7 . Number 4 ‘ 1996

ARTERIOVENOUS FISTULAE

Arteriovenous Fistula FunctionNumerous investigators have studied the likelihood

of successful maturation ofAVF into functional hemo-

dialysis vascular accesses (5,8, 17,23-32). Primary

failure of AVF is defined as insufficient dilation and

arterialization needed for function as a dialysis vascu-

lar access. The observed risk of primary failure of

these accesses has ranged from 9% (24) to 30% (1 7) in

surgical experience accumulated since the late 1960s.

Once AVF mature, numerous investigators have doc-

umented their excellent long-term function. Winsett

and Wolma reported the complication rates for 273AVF after successful maturation into functional vas-

cular accesses ( 1 7). After 2 yr of dialysis, 90% of these

accesses continued to function, with 80% still in use

after 3 yr. They observed a complication rate of five

events per 1000 patient-months of follow-up for AVF,which was considerably and statistically smaller than

the 37 events per 1000 patient-months observed in

recipients of PTFE synthetic dialysis grafts. Similar

fistula survival rates were also cited by Reilly and

colleagues in their report of 148 AVF created in En-gland between 1976 and 1981 (25) and more recently

by Kherlakian and colleagues (8). Bonalumi and co-

workers have reported the longest follow-up of AVF,

observing that 52% ofend-to-end radial artery fistulae

continued to be useful after 6.5 yr of hemodlalysis (5).

More recently, native AVF constructed using the

brachial artery have been evaluated as an alternative

access procedure when a radiocephalic fistula was not

able to be constructed (33-35). Although uncontrolled

studies have documented primary patency rates as

high as 86% at 2 yr of follow-up (33), the long-term

function of these accesses has not yet been evaluated.

Risk Factors for AVF DysfunctionDespite the excellent long-term function of fistulae

after successful maturation, little consensus has beenachieved regarding the predictors of the primary fail-

ure ofAVF. A number ofpotential risk factors, includ-

ing gender, vessel size, surgical technique, surgical

skill, and primary diagnosis have been postulated as

having an Important impact on the likelihood of suc-

cessful AVF maturation. Several studies in the early

1970s reported that women had a higher risk of

primary failure than men (24,26). Numerous otherreports (5,27-30) have looked for but failed to identify

specific risk factors of primary AVF failure. Many of

these studies were small in size and carried out lim-

ited analyses to control for potential confounding

factors.Reilly and coworkers presented one of the best

documented and one of the few prospective risk-factor

studies of dialysis fistula function published to date

(25). They studied 157 consecutive patients between

1 976 and 1 98 1 . Intraoperative data, including size

and type of anastomosis, size of vessels, suture mate-

rial, and adequacy of flow were recorded. Univariate

and multivariate analysis demonstrated that vein size

was an important predictor of long-term access com-

plications. The power of this study, however, was

limited because of its sample size, and no specific

analysis of the predictors of inadequate fistula matu-ration was performed.

Few recent data are available regarding the risk

factors for inadequate maturation ofAVF. Pourchez etaL reported that 4 1 of 47 French dialysis patients

older than 75 yr of age underwent successful con-struction of an AVF (36). Kobrin et al. evaluated 48

AVF constructed between 1989 and 199 1 at the Uni-versity of Pennsylvania (37). Overall, 44% ofAVF had

successfully matured at 3 months, whereas primaryfailure occurred in 56%. Although age, gender, loca-

tion of the AVF, and the presence of diabetes mellitus

did not affect AVF maturation, the small size of the

study limited statistical power to detect clinically rel-

evant differences in access function. Few investigators

have analyzed the impact of the use of recombinant

erythropoietin (EPO) on AVF function (38,39) but the

few data available seem to suggest that AVF thrombo-sis is not a very common complication. Neither Be-

sarab et aL (39) nor Tang et aL (40) found thrombosis

of AVF to be more common with EPO use.

Thus, extensive published experience with AVF’

demonstrate their long-term utility and low rate of

complications among patients in whom primary fis-

tula function is achieved. However, despite a large risk

of primary fistula failure, little is known about the

specific surgical, comorbidity, or demographic factorspredictive of primary fistula function. Although it is

probably true that an older and sicker ESRD popula-

tion may less commonly undergo successful AVF con-

struction, the subgroup of patients in whom fistulaeshould be created is poorly defined. Several recent

reports (23,33-37) suggest that despite the older age

and worse comorbidity of U.S. ESRD patients as com-

pared with two decades ago, a substantial number of

patients remain viable candidates for placement ofAVF. This is especially true if upper-arm fistulas such

as brachial artery-basilic vein fistulas are considered.

A review of alternatives to AVF readily demonstrates

that unnecessary substitution of synthetic grafts for

autologous AVF greatly enhances both the morbidity

of and costs for the U.S. ESRD population.

POLVTETRAFLUOROETHYLENE DIALYSIS ACCESS

Patency and Complications ofPolyfetrafluoroethylene Dialysis Access

PTFE was introduced as a material for vascular

bypass grafts in 1976 (10). Since that time, thismaterial has become the mainstay for dialysis vascu-

lax accesses when an autologous AVF is either be-

lieved to be technically impossible or has failed to

mature, or commonly when there is insufficient time

for an AVF to mature before the need for hemodialysis

because of the late presentation of a patient’s symp-

51).

Feldman et al

Journal of the American Society of Nephrology 527

tomatic advanced renal failure. PTFE dialysis grafts

are currently the most common type of dialysis ac-

cess, accounting for as many as 83% of access place-

ments among prevalent ESRD patients (38). A great

majority of hemodialysis patients undergo only PTFE

graft placement despite the high probability that a

significant proportion of them would have been able to

undergo successful AVF construction. However, few

data exist documenting the proportion of accesses

that are constructed of PTFE but that could have been

successfully created as a native AVF. Nonetheless, the

potential excess use of PTFE vascular accesses and

the protean complications possible after their place-

ment, including stenosis, occlusion, infection, aneu-

rysm and pseudoaneurysm formation, cardiovascular

instability, and nerve injury, are all potentially re-

sponsible for a large component of the high morbidity

in the U.S. ESRD population.

Numerous investigators have reported 1 -and 3-yr

patency rates for PTFE grafts. Between 63 and 90% of

grafts still function after 1 yr. whereas only 42 to 60%

remain patent after 3 yr (9, 13-15, 17,41-47). Most of

these reports defined patency as persistent graft func-

tion, regardless of treatment for access complications

such as thrombectomy or percutaneous angioplasty.

Complication-free graft survival has been reported by

a number of investigators (17,41,45). Sabanayagam

and colleagues (45) reported the complication-free

survival ofPTFE grafts to be 77. 1% at 1 yr; Palder and

colleagues reported complication-free PTFE graft sur-

vival to be 51% at 2 yr (41): and Winsett and Wolma

reported complication-free PTFE graft survival to be

39% at 3 yr (1 7). Outflow obstruction and thrombosis

accounted for the vast majority of the complications

observed. Unfortunately, none of these studies re-

ported the influence of patient comorbidity on access

dysfunction. By prospectively studying a subgroup of

52 healthier patients who lived longer than 3 yr on

dialysis, Schuman and colleagues grossly controlled

for comorbidity and observed one access complication

for every 5 1 5 patient-days of follow-up (14).

In recent years, increasing attention has been given

to the detection and repair of stenosis of PTFE grafts.

Color-flow Doppler ultrasonography and contrast fis-

tulography are widely utilized. Repair by transluminal

angioplasty or surgical vein-patch angioplasty may

promote delivery of adequate dialysis and prevent

graft thrombosis. Whether repair before thrombosis

extends the useful life of PTFE grafts has yet to be

proven. A large, but not precisely determined, cost is

incurred by these diagnostic tests and interventional

procedures, which are performed for the most part on

an outpatient basis. A detailed discussion of the diag-

nosis and therapy of stenosis of PTFE grafts is beyond

the scope of this paper, but the reader is referred

elsewhere for recent discussions of these issues (48-

Risk Factors for Complications from PTFEDialysis Access

Thromboses of VFFE access grafts have been

thought to result from a number of causes, including

venous outflow obstruction and inadequate arterial

inflow. Outflow obstruction has usually been attrib-

uted to neointimal hyperplasia (52-54) or central ye-nous stenosis/thrombosis from prior central venous

access (55). Inadequate arterial inflow has resulted

from a small or diseased arterial source as well as

from recurrent bouts ofintradialytic hypotension (56-

59). Postdialysis dehydration with increased blood

viscosity; low serum albumin; inadequate anticoagu-

lation; external compression either from perigraft he-

matomas, overly vigorous pressure at needle sites,

occlusive bandages, or inopportune arm positioning

during sleep; the presence of anticardiolipin antibod-

ies, diabetes mellitus, and high hematocrit with EPO

therapy have all been claimed, with varying justifica-

tion, to contribute to PTFE graft thromboses

(39,40,52,57,59-64).

Despite these numerous hypothetical risk factors,

few studies have examined specific risk factors in a

controlled fashion. Aman et aL reported on 9 1 PT’FE

grafts followed for 24 months and observed a signifi-

cantly increased graft patency rate with the use of

double-lumen dialysis needles (54). They were unable

to find a relationship between graft survival or com-

plication-free graft survival and sex, race, degree of

hypertension, heparin dosing, or intradialytic hypo-

tension, although the statistical power of their study

was limited. Munda and colleagues followed 67 PTFE

grafts for 2 yr and found that a forearm loop configu-ration had greater than twice the rate of secondary

patency of forearm straight grafts (57).Although controversial, several reports have sug-

gested an association between diabetes mellitus andPTFE graft thrombosis (19,65,66). Windus and col-

leagues observed a rate of PTFE graft thrombosis at 1

yr of 72% among patients with diabetes, compared

with a rate of 49% among patients without diabetes

(65). Data from Medicare and the U.S. Renal Data

System (19,66) have been consistent, demonstrating

higher rates of hospitalization for vascular access

problems among the subset of ESRD patients with

diabetes.

Several investigators have evaluated the impact of

therapy with recombinant EPO on synthetic vascular

access function and reported conificting results

(39,40,62). Besarab and colleagues compared the rateof thrombotic events in 164 patients receiving EPO for

at least 2 months with that among 142 patients

treated earlier, before the availability of EPO. Theywere unable to detect any differences in the rate of

thrombosis either for synthetic grafts or AVF (39).

Consistent with these findings have been a number ofother studies (67-70), each failing to note an increase

in access thrombosis with EPO.In contrast to these findings, Dy and colleagues

Hemodialysis Vascular Access Morbidity

528 Volume 7 ‘ Number 4 . 1996

performed a nonconcurrent cohort study of 46 dialy-

sis patients before and after treatment with EPO (62).They reported a greater incidence of graft thrombosis

in the EPO period. Similar findings were reported by

the Canadian Erythropoletin Study, a randomized

comparison between treatment with placebo (N = 40)

and treatment with EPO targeting either a hemoglobin

concentration of 9.5 to 1 1 .0 g/dL (N = 40) or 1 1 .5 to13.0 g/dL (N = 38). A 2.5% rate of access thrombosis

at 6 months was observed in the placebo group with a

mean hemoglobin of 7.4 g/dL. In contrast, the two

groups treated with EPO achieved mean hemoglobin

concentrations of 10.2 g/dL and 1 1.7 g/dL with

6-month thrombosis rates of 10% and 18%, respec-

tively (7 1). The low rate of thrombosis in the placebo

group was considerably lower than that published by

Eschbach and colleagues (67), making this study’s

findings difficult to interpret. Although several inves-

tigations reporting substantial thrombosis rates inpatients with PTFE grafts on EPO have recenfly been

reviewed (68), none of these studies have includedcontrol groups off EPO.

Thus, several investigators have independently con-

firmed the impression derived from the hospitalizationdata in Medicare’s ESRD database: the rate of compli-

cations, especially stenosis and thrombosis, arising

from synthetic dialysis vascular accesses is very large.

The mean duration of the complication-free accesssurvival appears to be between 9 and 16 months, with

a great majority of patients experiencing a complica-

tion by the end ofthe 3rd yr after graft placement. This

extraordinarily high complication rate, with its atten-dant suffering and high utilization of U.S. Dialysis

Program resources, makes the search for better strat-egies to improve outcomes among recipients of syn-

thetic vascular accesses urgent. Formulation of suchstrategies will require, in part, better delineation of the

risk factors for failure of PTFE access grafts.

INDWELLING CENTRAL VENOUS DIALYSISCATHETERS

Intended originally as a temporary measure, thesilicone double-lumen catheter with Dacron#{174}cuff hasinevitably come into long-term use in patients inwhom peripheral vascular access (fistula or graft) Is

unsuitable. The Dacron#{174} cuff decreases infection andmakes long-term use possible. When catheters are

intended for use for less than 1 month, a polyure-

thane, non-cuffed central venous access is normallyused. However, if catheter use is expected to exceed 1month, the cuffed SILASTIC (Dow Corning, Midland,

MI) catheter is preferred (73). The proportion of hemo-

dialysis patients in the United States in whom cuffed

SILASTIC� catheters provide long-term vascular ac-

cess is unknown. Based on data from the 1991 U.S.Renal Data System (USRDS) Special Study of CaseMix, It can be estimated that 4% ofU.S. ESRD patientsbegin dialysis with a permanent catheter. These data

underestimate the actual prevalence of catheter use,

as they do not account for usage after failure of more

traditional vascular accesses, AVF, or PTFE grafts. In

a survey of 1 7 dialysis units in the Chicago area, theproportion ofpatients using permanent double-lumencatheters ranged between 3.3 and 45%; of 1372 pa-

tients, 210 (15%) received or already had such cathe-

ters (74). Increasing long-term survival of the end-

stage renal population has enhanced the incidence of

depletion of potential sites for AV vascular access, and

the use of central venous catheters for permanentaccess is likely to be increasing.

A jugular vein has become the preferred site forpermanent central venous catheter insertion. Percu-

taneous jugular venous insertion is safer than in thesubclavian location. Subclavian catheters have

caused subclavian thrombosis and stenosis (75),

which can preclude creation of AV vascular access in

the ipsilateral ann, and there may be less mechanical

malfunction in catheters placed in the jugular rather

than subclavian location (76), partIcularly the right

jugular veins (77). Blood-flow rates of 200 to 300mL/min are routinely achieved (78-80); flows up to

400 mL/min can be achieved in some patients, but

sometimes not consistently. Recirculation in variousstudies has ranged from 5.5 to 8.6% (79-82). Unlike

AV fistulas and grafts, dialysis through central venous

catheters is not affected by cardiopulmonary recircu-lation.

Actuarial catheter survival at 12 months has ranged

from 30 to 65% (76,79,80,83). Catheter failure has

been the result primarily of malfunction (thrombosis

or poor blood flow) or to infection, in roughly equalproportions. Catheter malfunction can be reduced by

proper placement of the catheter tip in the right

atrium and by the use of chronic anticoagulation,beginning with aspirin and using warfarmn if necessary(74). Urokinase and streptokinase and mechanical

means have been used to open thrombosed catheters

and those with poor flow (74,78-80,84-86). Infection

with septicemia occurs with a frequency between 0.25and 1.0 infections per patient-year (76-79). Although

Gram-positive infections have been successfullytreated with antibiotics alone, many infections, espe-cially those resulting from Gram negative organisms,have required catheter removal.

Use of permanent central venous catheters hasimportant morbidity. Insertion can be complicated by

arterial puncture, pneumothorax, hemothorax, ar-rhythmia, and perforation or laceration of the brachial

plexus, trachea, superior vena cava, or myocardium.Jugular venous cannulation Is generally safer than

subclavian cannulation and the incidence of these

complications is low in experienced hands (73). Late

complications include catheter malfunction because

of thrombosis or malposition, infection, and subcla-

vian or superior vena cava thrombosis or stricture(73,87,88). The latter complications, which occur with

internal jugular as well as subclavian catheters, mayprevent subsequent use of the ipsilateral arm or of

both arms for AV access. Superior vena cava throm-

Feldman et al

Journal of the American Society of Nephrology 529

bosis may be more common than has been reported

(88), but the actual rates of occurrence of these mor-bidities are unknown.

Permanent dual-lumen central venous cathetershave some advantages over AV access, including easeof insertion, removal, and replacement, immediacy of

use, absence of hemodynamic stress or steal syn-

drome, absence of cardiopulmonary recirculation,

and avoidance of venipuncture. Indications for suchcatheters include loss of other vascular access sites,

prolonged period of maturation of other accesses

(usually AVF), cardiovascular instability or steal syn-

drome, and short expected duration of dialysis (e.g.,

malignancy, impending renal transplantation). If

rates of poor flow and thrombosis and infection are as

low as in some series, permanent catheters may even

be considered an acceptable alternative to PTFE grafts

in the general population of patients in whom creation

of an AVF is not feasible (80).

PATFERNS OF VASCULAR ACCESS UTILIZATION INTHE UNITED STATES

The only national data permitting study of specific

types of vascular access were collected as part of the

special studies of Case Mix conducted by the USRDS,

first for patients in 1986 and 1987, and subsequentlyfor patients in 1 99 1 . Although these data are older,

they provide us with an opportunity to examine the

evolution of the patterns of vascular access utilization

C4,C.,

4,a.

in the United States that have not been previously

published.

By using data from the USRDS 1 986 to 1 987 Special

Study of Case Mix, we are able to estimate the distri-bution of vascular access type in the United States.

This study examined the initial medical records for a

nearly random sample of 4964 U.S. dialysis patients

initiating dialysis in 1986 and 1987. Among them,

1 5 1 8 were identified as beginning hemodialysis with

an AVF, 1498 with a VTFE graft, and 77 with a central

venous catheter. The remainder of the patients either

initiated peritoneal dialysis or could not be catego-

rized into one specific vascular access group. As is

demonstrated in Figures 2 and 3, older age and female

gender were associated with lower utilization ofAVF in

1986 and 1987. A similar analysis failed to demon-

strate an association between AVF utilization and

either race or the presence of diabetes mellitus.

The 199 1 USRDS Special Study of Case Mix dem-onstrated that the pattern of AVF utilization had

substantially changed over the span of 4 to 5 yr. This

study, which analyzed the initial vascular access

among 1673 randomly selected hemodialysis pa-

tients, demonstrated that AVF utilization among dia-

betics had fallen substantially from over 40% to about

25% (see Figure 4). Similarly, but to a lesser extent,

AVF utilization had also fallen among ESRD patients

without diabetes (see Figure 4). Although these

changes may be a reflection of further increases in the

morbidity of the ESRD population, leading to less

.< 19 yrs

#{149}20-44yrs

�45-64yrs

D65-74yrs

D75 + yrs

AV Fistula PTFE Graft

Figure 2. Vascular access type by age: data from USRDS Special Study of Case Mix: 1986 to 1987.

Hemodialysis Vascular Access Morbidity

530 Volume 7 . Number 4 ‘ 1996

C4,

C.,I-

4,a.

AV Fistula PTFE Graft

OMale

#{149}Female

Figure 3. Vascular access type by gender: data from USRDS Special Study of Case Mix: 1986 to 1987.

successful efforts to place AVF, these trends may also

reflect changes in patterns ofpractice that favor place-

ment of synthetic vascular accesses available for usealmost Immediately after insertion. Finally, the esti-

mates of AVF utilization defined here among incident

ESRD patients probably overstate the prevalence of

AVF use among all hemodialysis patients. Windus hasrecently estimated that approximately 1 1% of Amen-

can hemodialysis patients in 1990 used AVF for an-

gioaccess (38).

CLINICAL SELECTION OF VASCULAR ACCESSMODALITY: WHAT ARE THE TRADEOFFS?

Numerous options currently exist for achieving vas-

cular access for ESRD patients today. Selectionamong these options requires evaluation of the poten-

tial toxicitles and benefits of each of the access

choices, so as to understand the tradeoffs inherent in

choosing a particular access method (see Figure 5).

Data currently available strongly suggest that whentechnically feasible, radial AVF represent the optimalsolution for vascular access. Identification of patients

in whom a radial AVF would mature, however, isdifficult with currently available information. At-tempting to place radial AVF before the need fordialytic therapy would minimize the morbidity andcosts associated with use of central venous catheters

while awaiting AVF maturation. Even when a radialAVF falls to mature adequately, a VFFE graft or a

brachial AVF can usually be placed in the ipsilateral

ann. Success of these accesses may even be enhanced

by venous dilatation that may have arisen as a result

of the initial radial AVF. Selection of the patients in

whom a radial AVF should be attempted, however,

often presents a more difficult decision when patients

require immediate or imminent dialysis. Under thesecircumstances, both the high probabifity of unsuc-

cessful maturation especially among older and sicker

patients, and the morbidity associated with central

venous catheters necessary for dialysis while waiting

for AVF maturation, lead many clinicians directly toan alternative access choice. Usually their choice is aVrFE graft, which permits dialysis very soon after

placement. Better delineation of both the profile ofpatients likely to experience successful maturation of

a radial AVF as well as quantification of the morbidity

associated with temporary central venous catheters

necessary for expanded utilization of AVF are neededto determine the appropriate clinical thresholds for

AVF use.

Currently, three distinct alternatives exist to radialartery AVF; brachial AVF, VFFE bridge grafts, and

permanent indwelling catheters. Although availabledata demonstrate that radial AVF are superior toVT’FE grafts and catheters, the appropriate role of

brachial artery AVF Is still not entirely defined. Al-though these accesses are likely to enjoy some of the

benefits of radial AVF (e.g., only one anastomosis isrequired and many infections are treatable withoutaccess removal), their long-term patency is uncertain.

Graft

Figure 4. Evolution of vascular access use in the United States: diabetic and nondlabetic hemodlalysis patients <65 yr of age.Data from USRDS Special Studies of Case Mix: 1986 to 1987, and 1990.

Feldman et al

Journal of the American Society of Nephrology 531

C,U _________L. � � �

C, IO1986-7�a. � I

_0

Diabetic - Diabetic - Nondiabetic Nondiabetic

AV Fistula PTFE Graft - AV Fistula - PTFE

Although several small studies suggest that patency of

brachial artery AVF at least rivals VT’FE graft patency,more research is needed. Whether problems such as

steal syndrome and heart failure secondary to the

high rates of blood flow through these accesses will

limit their use is still unknown. Despite these un-

knowns, we currently believe that the preponderance

of evidence supports attempting to place a brachial

AVF before a PTFE graft in patients in whom a radial

AVF is either not possible or has failed.In the tradeoff between use of a PTFE graft and a

permanent central venous catheter for angloaccess,several considerations must be weighed. Permanent

dual-lumen central venous catheters have some ad-vantages over FTFE grafts, including ease of insertion,

removal, and replacement, immediacy ofuse, absenceof hemodynamic stress or steal syndrome, absence of

cardiopulmonary recirculation, and avoidance of ye-

nipuncture. However, the permanent venous catheter

may not support blood flow at rates adequate toensure optimum dialysis; this problem, together with

loss of time because of the need for declotting withurokinase, could lead to underdialysis. Episodic poorflow often makes use of these catheters frustrating for

technical staff. However, despite high rates of poor

flow, thrombosis, and infection in some series, perma-

nent catheters have even been considered an alterna-

tive to PTFE grafts in the general population of ESRDpatients in whom creation of a native AVF Is not

feasible (80). Like central venous catheters, PTFE

grafts similarly are complicated by thromboses and

high recirculation from stenoses, leading to possible

underdialysis and necessitating anatomical correc-

tion (e.g., angioplasty, surgical revision). Both PTFEgrafts and catheters may require anticoagulation,adding a risk of hemorrhage, including subdural he-

matoma. Although infection of a central venous cath-eter may be treated with antibiotics without catheter

loss and catheter replacement is relatively easy, infec-

tion of a PTFE graft almost always requires graftremoval, which can entail significant morbidity.

Despite the problems associated with PTFE grafts,

one final distinction between them and central venouscatheters weighs most heavily in our opinion that

catheters not be used as a primary permanent accesswhen either an AVF or VFFE graft can be placed.Central venous stricture or thrombosis subsequent toplacement of a central venous catheter may eliminate

all potential access sites in one or both arms. Giventhe often catastrophic impact of this loss of upper-

extremity access sites, the PTFE graft, despite itscomplications, is preferable to the central vein cathe-ten used as a permanent access.

SUMMARY

Vascular morbidity In the United States is substan-

tial and is growing faster than Is attributable to the

Principal Vascular Access Choice

Radial

AVF

I Pros �J Cons

NFirst Alternatives to Radial

AVF

Longer Survival � High rate ofprimary failure

Fewer Complications � Delayed availability of access

�J Complications of temporaryuse of central venous

catheters

Secondary Alternative to RadialAVF

Brachial Artery Arteriovenous Fistulae Permanent Central Venous Catheter

Pros Cons

Feasibly placed in

virtually all patients

Available immediately

No hemodynamic stressor steal

Avoids venipuncture

Hemodialysis Vascular Access Morbidity

532 Volume 7 ‘ Number 4 ‘ 1996

Vascular Access Selection: What Are the Tradeoffs?

Pros ConsFeasibly placed in many Unknown long-term

patients not function

candidates for radial High rate offlow may

AVFs lead to

May have nsorbidity cardiovascularprofile similar to instability

radial AVFs

PTFE AV Bridge Graft

Pros Cons

Feasibly constructed in High rate of occlusive

most patients failure

Available for use Infective failure

typically within 2 typically demandsweeks removal

Shortened secondary

patency vs. AVFs

May cause central

venous stenosis,limiting subsequent

access procedures

High rate of infectiveand occlusive failure

Figure 5. Vascular access selection: what are the tradeoffs?

growth of the ESRD population or growth in other

morbidity among ESRD patients. Morbidity from ac-

cess dysfunction continues to grow faster than sug-gested by available data on hospitalization for vascu-

lar access dysfunction, In part because of the shift tooutpatient therapies such as thrombolysis and anglo-

plasty. The growth in vascular access morbidity may

be a result ofthe rising age ofthe ESRD population, itsworsening severity of illness, increasing demands on

vascular accesses to tolerate greater blood-flow rates,

and increasing diagnostic activities focused on detect-

ing occult access dysfunction.

In addition to the growing age and comorbidity of

patients independent of surgical technique, an evolv-

ing pattern of surgical practice away from the creation

of autologous AVF to the construction of syntheticPTFE dialysis vascular access grafts is also responsi-

ble for the increase in vascular access-related morbid-ity. The substitution of PTFE grafts for AVF contrib-utes to the extensive access-related morbiditycurrently confronting dialysis patients and providers.

Although these synthetic accesses typically functionimmediately after their construction, their rate of later

secondary failure is much higher than that for radialAVF. A great challenge before the dialysis community

is the identification of patients in whom an AVF is a

viable vascular option so that shorter-lived synthetic(VrFE) vascular access grafts can be reserved forpatients in whom an AVF is unlikely to mature suc-

cessfully. To accomplish this, the identification of riskfactors for inadequate maturation of AVF into long-

lasting hemodialysis vascular accesses must be a high

priority. In addition, it is essential that prospective

recipients of AVF be identified early and that everyeffort be made to avoid venipuncture in the arm

intended for the vascular access.

PTFE AV bridge grafts are accompanied by a high

rate of complications primarily from stenosis and

occlusion, and to a lesser extent, infection. Risk fac-

tons for early PTFE access complications, too, are

incompletely defined. Such data are needed to preventnot only the morbidity of vascular access dysfunction

per se, but potential undendialysis from access dys-

function that represents a most important, but oftenhidden and insidious, cost ofvascular access morbid-

ity.

Permanent central venous catheters may not onlyrepresent an access of last resort when other accesssites have been exhausted, but may offer an altenna-tive to PTFE grafts. The population in whom perma-

Feldman et al

Journal of the American Society of Nephrology 533

nent central venous catheters may be a superior

alternative to PTFE grafts, even when sites for grafts

have not been exhausted, needs to be better defmed incontrolled studies.

Clinical strategies to optimize the diagnosis andtherapy ofvascular access dysfunction also need to be

identified. These strategies must define the appropri-

ate threshold of abnormal morphology and functionthat justifies radiologic and/or surgical intervention.Clearly, early intervention may preserve the second-

ary patency of synthetic vascular access grafts. How-

ever, aggressive diagnostic activities plausibly may

lead to a higher rate of radiologic and surgical inter-ventions for detectable but not yet clinically important

stenoses. These interventions (e.g. , angioplasty or

surgical repair) themselves may reduce the feasibifity

of later interventions and may damage the vascularaccess directly, thereby shortening its long-term func-

tion.Finally, identification of the fundamental mecha-

nisms of vascular access dysfunction should also bethe focus of investigation so that specific therapies

and preventive strategies can be developed and imple-

mented.

ACKNOWLEDGMENTSThe authors gratefully acknowledge the assistance of the U.S. RenalData System for providing data from their Special Studies ofCase Mix.

This work was supported, in part. by a Cooperative Agreement withthe Health Care Financing Administration (HCFA).

REFERENCES1 . Quinton W, Dullard D, Scribner BH: Cannulation of blood

vessels for prolonged hemodlalysis. Trans Am Soc ArtifIntern Organs 19606:104-113.

2. Rosenthal JJ: Vascular access for hemodialysis. MajorProbl Clin Surg 198 1;4:455-468.

3. Wilson SE, Stabile BE, Williams RA, Owens ML: Currentstatus of vascular access techniques. Surg Clin NorthAm 1982;62:531-551.

4. Brescia MJ, Cimino JE, Appel K, Hurwich BJ: Chronichemodialysis using venipuncture and a surgically cre-ated arteriovenous fistula. N Engi J Med 1966;275:1089-1092.

5. Bonalumi U, Civalleri D, Rovida S, Adami GF, GianettaE, Griffanti-Bartoli F: Nine years’ experience with end-to-end artenlovenous flstula at the anatomical snufiboxfor maintenance haemodlalysis. BrJ Surg 1982;69:486-488.

6. Brems J, Castaneda M, Garvin PJ: A five-year experiencewith the Bovine Heterograft for vascular access. ArchSurg 1986:121:941-944.

7. Giacchino JL, Geis WP, Buckingham JM, Vertune LL,Bansal VK: Vascular access: Long-term results, newtechniques. Arch Sung 1979; 114:403-409.

8. Kherlakian GM, Roedersheimer LR, Arbaugh JJ, New-mark LI, King LR: Comparison of autogenous fistulaversus expanded polytetrafluoroethylene graft fistula forangioaccess in hemodialysis. Am J Surg 1986; 152:238-243.

9. Munda R, First MR, Alexander JW, Linnemann CC,FidlerJP, Kittur D: Polytetrafluoroethylene graft survivalin hemodialysis. JAMA 1983;249:219-222.

10. Rapaport A, Noon GP, McCollum CH: Polytetrafluoro-ethylene (VI’FE) grafts for haemodlalysis in chronic renalfailure: Assessment of durability and function at threeyears. Aust NZJ Surg 198 1:51:562-567.

1 1 . Veith FJ, Wilson SE, Hobson RW, Rosenthal JJ, TellisVA, Dagher FJ: Vascular access complications and new

methods. Trans Am Soc Artif Intern Organs 1982;28:647-657.

12. Office of Statistics and Data Management. End-stagerenal disease patient profile tables, 1987. Baltimore:Office of Statistics and Data Management, Health CareFinancing Administration; 1987.

13. Bell DD, Rosental JJ: Arteniovenous graft life in chronichemodlalysis. Arch Surg 1988; 123:1169-1172.

14. Schuman ES, Gross GF, Hayes JF, Standage BA: Long-term patency of polytetrafluoroethylene graft flstulas.AmJ Surg 1988:155:644-646.

15. Haimov M, Burrows L, Sclianzer H, et at.: Experiencewith arterial substitutes in the construction of vascularaccess for hemodialysis. J Cardiovasc Surg 1980:21:149-154.

16. Mohaideen AR, Mendivil J, Avram MM, Malnzer RA:Artenlo-venous access utilizing modified bovine arterialgrafts for hemodlalysis. Sung Clin North Am 1977:62:531-550.

17. Winsett OE, Wolma FJ: Complications of vascular ac-cess for hemodialysis. South Med J 1985:78:513-517.

18. Kester RC: Arteniovenous grafts for vascular access inhemodialysis. Br J Surg 1979;66:23-28.

19. Feldman HI, Held PJ, Hutchinson T, Stoiber E, HartiganMF, Berlin JA: Hemodialysis vascular access morbidityin the U.S. Kidney Int 1993:43:1091-1096.

20. Lazarus JM, Huang WI!, Lew NL, Lowrie EG. Contnibu-lion of vascular access-related disease to morbidity ofhemodlalysis. Vascular Access for Hemodialysis III. Chi-cago: WL Gore & Assoc. , Inc. , and Precept Press, Inc.:1993:3- 13.

2 1 . Held PJ, Pauly MV, Bovbjerg R, Diamond L: Large andsmall transplant programs: Implications for cost andoutcomes. IPublication No. 3293-22] WashIngton, DC:The Urban Institute; 1986.

22. US Renal Data System. USRDS 1994 Annual Data Re-port. Bethesda, MD: The National Institutes of Health,National Institute of Diabetes and Digestive and KidneyDiseases: 1994:23-42.

23. Churchill DN, Taylor DW, Cook RJ, LaPlante P, et at.:Canadian hemodialysis morbidity study. Am J KidneyDis 1992:19:214-234.

24. Kinnaert P. Vereerstraeten P, Toussaint C, Van Geer-truyden J: Nine years’ experience with internal artenio-venous fistulas for haemodialysls: A study of some fac-tors influencing the results. BrJ Surg 1977:64:242-246.

25. Reilly DT, Wood RFM, Bell PRF: Prospective study ofdialysis fistulas: Problem patients and their treatment.BrJ Surg 1982:69:549-553.

26. HIggins MR. Grace M, Bettcher KB, Silverberg DS,DossetorJB: Blood access In hemodialysis. Clint Nephrol1976:6:473-477.

27. Aim A, Lundberg M: Clinical experience with therapeuticarteriovenous flstulae. Scand J Urol Nephrol 1977; 1 1:53-57.

28. Rohr MS. Browder W, Frentz GD, McDonald JC: Arterio-venous flstulas for long-term dialysis. Arch Surg 1978;113:153-155.

29. Dunlop MG, Mackinlay JY, Jenkins AM: Vascular ac-cess: Experience with the brachiocephalic fistula. Ann RColl Sung Engl 1986:68:203-206.

30. ZIbari GB, Rohr MS. Landreneau MD, Bridges RM, etaL: Complications from permanent hemodialysis vascu-lar access. Surgery (St. Louis) 1988; 104:681-686.

3 1 . Bourquelot P, Cussenot 0, Corbi P, et at.: Microsurgicalcreation and follow-up of arteriovenous flstulae forchronic haemodialysis in children. Pediatr Nephrol1990:4:156-159.

32. Wehrli H, Chenevard R, Zaruba K: Chlrurgische Erfahr-ungen mit arterio-venosen Hamodialyse-Shunts (1970-1988). Helv Chir Acta 1989:56:621-627.

33. Coburn MC, Carney WI Jr: Comparison of basilic veinand polytetrafluorethylene for brachial arteriovenousfistula. J Vasc Surg 1994:20:896-902..

34. RIvers SP, Scher LA, Sheehan E, Lynn R, Veith FJ:Basific vein transposition: An underused autologousalternative to prosthetic dialysis angioaccess. J VaseSurg 18:391-396.

Hemodialysis Vascular Access Morbidity

534 Volume 7 . Number 4 . 1996

35. Bender MH, Bruyninckx CM, Genlag PG: The brachioce-phalic elbow fistula: A useful alternative angioaccess forpermanent hemodialysis. J Vase Surg 1994:20:808-813.

36. Pourchez T, Moriniere F, St Priest A, et at: Outcome ofvascular access for hemodialyses in the elderly (>75years) lAbstracti. Proc Eur Dial Transplant Assoc 1990:27: 160A.

37. Kobrln 5, Feldman H, Morley C, Deno M: Outcome ofarteniovenous fistula (AVF) placements (Abstract).ASAIO 38th Annual Meeting: 1992:84A.

38. Windus DW: Permanent vascular access: A nephrolo-gist’s view. Am J Kidney Dis 1993:21:457-471.

39. Besarab A, Medina F, Musial E, et aL: Recombinanthuman erythropoietin does not increase clotting in vas-cular accesses. ASAIO (Am Soc Artif Intern Organs)Trans 1990;36:M749-M753.

40. Tang TY, Vrahnos D, Valaitis D, Lau AR: Vascular accessthrombosis during recombinant human erythropoietintherapy. ASAIO J 1992:38:M528-M531.

4 1 . Palder SB, Kirkman RL, Whittenmore AD, Hakim RM,Lazarus JM, Tilney NL: Vascular access for hemodialy-sis. Ann Surg 1985:202:235-239.

42. Tauches LA: Immediate, safe hemodialysis with arterio-venous fistulas created with a new tunneler. Ann Sung1985:150:212-215.

43. Raju 5: PTFE grafts for hemodlalysis access. Ann Surg1987:206:73, 206-266.

44. Mandel SR. Martin PL, Blumoff RL, Mattern WD: Vas-cular access In a university transplant and dialysisprogram. Arch Surg 1977:112:1375-1380.

45. Sabanayagam P. Schwartz AR, Soricelli RR, Chinitz JL,Lyons P: Experience with one hundred reinforced ex-panded PTFE grafts for angioaccess in hemodialysis.Trans Am Soc Artif Intern Organs 1980:26:582-583.

46. Kraev Z, Kirjakov Z, Antonov S. Stantchev I, ZlatarskaS, Sofia: Polytherafluoroethylene graft’s complications inpatients on hemodialysis. Przegl Lek 1985:42:442-443.

47. Tellis VA, Kohlberg WI, Bhat DJ, Dniscoll B, Veith FJ:Expanded polytetrafluoroethylene graft fistula forchronic hemodialysis. Ann Surg 1979: 189: 10 1-105.

48. Beathard GE: Percutaneous transvenous angioplasty inthe treatment of vascular access stenosis. Kidney mt1992:42:1390-1397.

49. Schwab SJ, Raymond JR, Saeed M, Newman GE, Den-nis PA, Bollinger RR: Prevention of hemodialysis fistulathrombosis. Early detection of venous stenoses. Kidneymt 1989:36:707-711.

50. Windus DW, Audrain J, Vanderson R, Jendnisak MD,Picus D, Delmez JA: Optimization of high-efficiencyhemodialysis by detection and correction of fistula dys-function. Kidney Int 1990:38:337-441.

5 1 . Sands J, Young 5, Miranda C: The effect of Doppler flowscreening studies and elective revisions on dialysis ac-cess failure. ASAIO (Am Soc Artif Intern Organs) Trans1992:38:M524-M527.

52. Bell DD, Rosental JJ: Arteriovenous graft life in chronichemodlalysis. Arch Surg 1988:123:1169-1172.

53. Haimov M, Burrows L, Schanzer H, Neff M, Baez A,Kwun K, Slkfkin R: Experience with arterial substitutesin the construction of vascular access for hemodialysis.J Cardiovasc Surg 1980:21:149-154.

54. Aman LC, Levin NW, Smith DW: Hemodialysis accesssite morbidity. Proc Clin Dial Transplant Forum 1980:10:277-284.

55. Schwab SJ, Quarles D, Middelton JP, Cohan RH, SaeedM, Dennis VW: Hemodialysis-associated subclavian veinstenosis. Kidney Int 1988:33:1156-1159.

56. Kherlakian GM, Roedersheimer LR, Arbaugh JJ, New-mark KJ, King LR: Comparison of autogenous fistulaversus expanded polytetrafluoroethylene graft fistula forangloaccess in hemodialysis. Am J Sung 1986:152:238-243.

57. Munda R, First MR Alexander JW, Linnemann CC,Fidler JP, Kittur D: Polytetrafluoroethylene graft sur-vival in hemodialysis. JAMA 1983:249:219-222.

58. Winsett OE, Wolma FJ: Complications of vascular ac-cess for hemodialysis. South Med J 1985:78:513-517.

59. Kraev Z, Kirjakov Z, Antonov 5, Stantchev I, Zlatarska5, Sofia: Polytherafluoroethylene graft’s complications inpatients on hemodialysis. Przegl Lek 1985:42:442-443.

60. Rosenthal JJ: Vascular access for hemodialysis. MajorProbl Clin Surg 198 1:4:455-468.

6 1 . Tellis VA, Kohlberg WI, Bhat DJ, Driscoll B, Veith FJ:Expanded polytetrafluoroethylene graft fistula forchronic hemodialysis. Ann Sung 1979: 189: 10 1-105.

62. Dy GR, Bloom EJ, Ijelu GK, Merritts GW, Kramer MS,Raja RM: Effect of recombinant human erythropoietin onvascular access. ASAJO (Am Soc Artif Intern Organs)Transactions 199 1 ;37:M274-M275.

63. Garcia-Martin F, DeArriba G, Carrascosa T, et at.: Anti-cardiolipin antibodies and lupus anticoagulant in end-stage renal disease. Nephrol Dial Transplant 1991:6:543-547.

64. Page B, ZingraffJ, Houhou 5, et at.: Clotting of vascularaccess in dialysis patients associated to anticardiolipinantibodies lAbstracti. J Am Soc Nephrol 1992;3:385A.

65. Windus DW, Jendrisak MD, Dalmez JA: Prosthetic fis-tula survival and complications in hemodialysis pa-tients: Effects of diabetes and age. Am J Kidney Dis1992:20:448-452.

66. US Renal Data System. USRDS 1991 Annual Data Re-port. Bethesda, MD: The National Institutes of Health,National Institute of Diabetes and Digestive and KidneyDiseases; 1991.

67. Eschbach JW, Abdulhadi MH, Browne JK, et at.: Re-combinant human erythropoietin in anemic patientswith end-stage renal disease. Ann Intern Med 1989: 1 1:992-1000.

68. Sundal E, Businger J, Kappeler A: Treatment of trans-fusion dependent anemia of chronic renal failure withrecombinant human erythropoietin. Nephrol Dial Trans-plant 1991:6:955-965.

69. Bennett WM: A multicenter clinical trial of epoetin betafor anemia of end-stage renal disease. J Am Soc Nephrol1991; 1:990-998.

70. Kawalski EA, Trenkle JA, Natke E, et aL: The effect oferythropoietin on the incidence of chronic vascularthrombosis (Abstracti. J Am Soc Nephrol 1991:2:380.

7 1 . Canadian Erythropoietin Study Group. Association be-tween recombinant human erythrepoietin and qualityof life and exercise capacity of patients receiving hemo-dialysis. BMJ 1990;300:573-579.

72. Muirhead N: Erythropoietin as a cause of access throm-bosis. Semin Dial 1993:6:184-188.

73. Bander SJ, Schwab SJ: Central venous angioaccess forhemodialysis and its complications. Semin Dial 1992:5:12 1-128.

74. Uldail R, Besley ME, Thomas A, Salter S, Nuezca LA,Vas M: Maintaining the patency of double-lumen silasticjugular catheters for haemodialysis. mt j Artif Organs1993; 16:37-40.

75. Fant GF, Dennis V, Quarles LD: Late vascular complica-tions of the subclavlan dialysis catheter. Am J KidneyDis 1986:7:225-228.

76. Swartz RD, Messana JM, Boyer CJ, Lunde NM, WeitzelWF, Hartman TL: Successful use of cuffed central ye-nous hemodialysis catheters inserted percutaneously. JAm Soc Nephrol 1994:4: 17 19-1725.

77. Dc Meester J, Vanholder R, De Reese J, Ringoir 5:Factors and complications affecting catheter and tech-nique survival with permanent single-lumen dialysiscatheters. Nephrol Dial Transplant 1994:9:678-683.

78. Uldall R, DeBruyne M, Besley M, McMillan J, Simons M,Francoeur R: A new vascular access catheter for hemo-dialysis. Mn J Kidney Dis 1993:21:270-277.

79. Shusterman NH, Kloss K, Mullen JL: Successful use ofdouble-lumen, silicone rubber catheters for permanenthemodialysis access. Kidney mt 1989:35:887-890.

80. Moss AH, Vasilakis C, Holley JL, Foulks CJ, Pillal K,McDowell DE: Use ofa silicone dual-lumen catheter witha Dacron cuff as a long-term vascular access for hemo-dialysis patients. Am J Kidney Dis 1990: 16:21 1-2 15.

8 1 . Blake PG, Huraib 5, Wu G, Uldall PR: The use of duallumen jugular venous catheters as definitive long term

Feldman et al

Journal of the American Society of Nephrology 535

access for haemodialysis. IntJ ArtifOrgans 1990:13:26-31.

82. Schanzer H, Kaplan 5, Bosch J, Glabman 5, Burrows L:Double-lumen, silicone rubber, indwelling venous cath-eters. A new modality for angioaccess. Arch Surg 1986:121:229-232.

83. Shaffer D, Madras PN, Williams ME, D’Elia JA, KaldanyA, Monaco AP: Use ofDacron cuffed silicone catheters aslong-term hemodialysis access. ASAIO J 1992:38:55-58.

84. Dunea G, Domenico L, Gunnerson P, Winston-Willis F:A survey of permanent double lumen catheters in hemo-dialysis patients. ASAIO (Am Soc Artif Intern Organs)Trans 1991:37:M276-M277.

85. Schwab SJ, Buller GL, McCann RL, Bollinger RR,Stickel DL: Prospective evaluation of a Dacron cuffed

hemodialysis catheter for prolonged use. Am J KidneyDis 1988:11:166-169.

86. Shrivastava D, Lundin AP, Dosunmu B, Rae TKS, BeyerMM, Friedman EA: Salvage of clottedjugular vein hemo-dialysis catheters. Nephron 1994:68:77-79.

87. Caruana RI, Raja RM, Zeit RM, Goldstein SJ, KramerMS: Thrombotic complications of indwelling centralcatheters used for chronic hemodialysis. Am J KidneyDis 1987:9:497-501.

88. Grote J, Luift V. Nikkuta PVD, Lieth H, Bahlmann J,Daniel WG: Transesophageal echocardiographic assess-ment of superior vena cava thrombosis in patients withlong-term central venous hemodialysis catheters. ClinNephrol 1994:42:183-188.