Nephrol Dial Transplant (2004) 19: 1231–1236
DOI: 10.1093/ndt/gfh073
Advance Access publication 19 February 2004
Original Article
The natural history of autogenous radio-cephalic wrist arteriovenous
fistulas of haemodialysis patients: a prospective observational study
Carlo Basile, Giovanni Ruggieri, Luigi Vernaglione, Alessio Montanaro and Rosa Giordano
Division of Nephrology, Hospital of Martina Franca, Italy
Abstract
Background. Clinical practice guidelines have supported vascular access surveillance programmes on
the premise that the natural history of the vascular
access will be altered by radiological or surgical
interventions after vascular access dysfunction is
detected. The primary objective of this study was to
assess the actual risk of thrombosis of autogenous
radio-cephalic (RC) wrist arteriovenous fistulas (AVFs)
without any pre-emptive interventions.
Methods. We enrolled 52 randomly selected adult
Caucasian prevalent haemodialysis (HD) patients, all
with autogenous RC wrist AVFs, into this prospective,
observational study aimed to follow the natural history
of their AVFs for 4 years. The protocol prescribed
avoiding any surgical or interventional radiological
procedures until access failure (AVF thrombosis or a
vascular access not assuring a single-pool Kt/V 1.2).
The subjects underwent yearly assessments of vascular
access blood flow rate by means of a saline ultrasound
dilution method.
Results. All failures of vascular access were due to
AVF thrombosis; none were attributed to an inadequacy of the dialysis dose. AVF thrombosis occurred
in nine cases; a rate of 0.043 AVF thrombosis per
patient-year at risk. A receiver operating characteristic
curve, evaluating the diagnostic accuracy of baseline
vascular access blood flow rate values in predicting
AVF failure, showed an under-the-curve area of
0.82±0.05 SD (P ¼ 0.01). The value of vascular
access blood flow rate, identified as a predictor of
AVF failure, was <700 ml/min with an 88.9%
sensitivity and 68.6% specificity. When subdividing
the population of AVFs into two groups according to
the baseline vascular access blood flow rates, two out
of the nine thromboses occurred among the AVFs that
had baseline blood flow rates >700 ml/min (n ¼ 31),
whereas seven occurred among the AVFs that had
Correspondence and offprint requests to: Carlo Basile, MD, Via
Battisti 192, 74100 Taranto, Italy. Email: nefromartina@topvideo.
net
baseline blood flow rates <700 ml/min (n ¼ 21). The
4 year cumulative actuarial survival was 74.36 and
20.80%, respectively (log-rank test, P ¼ 0.04). The 24
AVFs that remained patent at the end of the 4 years
maintained a median blood flow rate 900 ml/min at
all time points studied. Worth noting is that, five of
them (20.8%) remained patent throughout the study
with a blood flow rate consistently 500 ml/min.
Conclusions. This study shows a very low rate of AVF
thrombosis per patient-year at risk and a high actuarial
survival of autogenous RC wrist AVFs, particularly of
those having a blood flow rate >700 ml/min. Thus, a
vascular access blood flow rate <700 ml/min appears
to be a reliable cut-off point at which to start a closer
monitoring of this parameter—which may lead to
further investigations and possibly interventions relevant to the function of the AVFs.
Keywords: autogenous radio-cephalic wrist arteriovenous fistula; haemodialysis; ultrasound dilution;
vascular access blood flow
Introduction
Maintaining a functional haemodialysis (HD) is one
of the important challenges nephrologists face today.
Vascular access failure is associated with significant
morbidity. There is also a significant cost associated
with it—8000 US dollars per patient per year at risk in
the United States, 15% of Medicare’s expenditure for
end-stage renal disease [1]. This, however, represents
only the tip of the iceberg. Vascular access complications account for 16–25% of hospital admissions, and
the type of vascular access in use is correlated with
overall and cause-specific mortality [2].
The high cost of access failure calls for a closer look
at strategies to prevent vascular access thrombosis.
NKF K/DOQI clinical practice guidelines recommend
regular monitoring of vascular accesses [be they
arteriovenous grafts (AVGs) or arteriovenous fistulas
Nephrol Dial Transplant Vol. 19 No. 5 ß ERA–EDTA 2004; all rights reserved
1232
C. Basile et al.
(AVFs)], by one of several methods, such as measurements of the vascular access blood flow rate [3]. A
technique employing saline ultrasound dilution (USM)
(Transonics Hemodialysis Monitor HD01; Transonics
Systems Incorporated, Ithaca, NY, USA) has been
developed [4], and published data support its efficacy
for the regular monitoring of both AVGs and AVFs
[3,4]. Such monitoring has received support [3] based
on the premise that the natural history of the vascular
access will be altered by radiological or surgical interventions after vascular access dysfunction is detected.
The main point of our 4 year prospective observational
study was to follow the natural history of mature AVFs
without altering them by radiological or surgical
interventions after detection of vascular access dysfunction. Thus, the aim of the present study was to
assess the actual risk of thrombosis of AVFs while
desisting from any pre-emptive interventions.
Table 1. Demographic, comorbidity data and vascular access
characteristics in the group of 52 patients at the baseline time point
Subjects and methods
for 4 years without any surgical or radiological interventions
until vascular access failure (AVF thrombosis or inadequacy
of dialysis dose, i.e. single-pool Kt/V <1.2). Thus, the primary objective was to assess the actual risk of thrombosis of
AVFs in the absence of any pre-emptive interventions. All
patients were dialysed three times per week with high- and
low-flux membranes and standard bicarbonate dialysate
using 15-gauge fistula needles. Patients were anticoagulated
with systemic heparin.
Patients
Our policy regarding vascular access includes two points: (i)
the forearm, native-vessel AVF is considered the vascular
access of first choice, and whenever possible must be
attempted in every patient referred for surgery; (ii) all
autogenous radio-cephalic (RC) wrist AVFs are constructed
by the experienced nephrologists of our unit (A.M. and
R.G.). Out of the 123 HD patients treated in our unit by 31
December 1998, 115 had native-vessel AVFs (93.5%) (112
autogenous RC wrist, three humerobasilic), five had AVGs
(4.1%) and three had central venous catheters (2.4%). All
AVFs are created in our unit after standardized vascular
mapping, including both physical examination and colour
Doppler sonography of the venous as well as of the arterial
beds of the upper extremity.
We invited 112 adult Caucasian patients prevalent as of 31
December 1998 to participate in this study. All of them had
mature autogenous RC wrist AVFs. For the purpose of the
study, an AVF was defined as mature if it allowed a dialyser
blood flow of at least 300 ml/min. Informed consent was
obtained from 105 patients. We randomly selected 52 of
them to undergo the present prospective observational study.
Comorbidity scoring was performed by a single clinician
(C.B.) familiar with the patient’s case histories, utilizing
medical and nursing records and investigation data. A very
simple comorbidity index was derived based on the presence
(scored 1) or the absence (scored 0) of the following
comorbidity domains: 1, diabetes mellitus; 2, heart disease
(previous myocardial infarction, angina pectoris, atrial fibrillation, pacemaker, left ventricular dysfunction); 3, systemic or
other significant pathologies—particular attention was paid
to the presence of peripheral vascular diseases as detected by
colour Doppler sonography or arteriography, or both. The
highest the cumulative score could be was 3 [5] (Table 1).
Age (years)
Dialysis vintage (months)
Females (%)
Comorbidity index
Comorbid conditions (%)
Atherosclerotic vascular disease
Diabetes mellitus
Arrythmia
Arterial hypertension
Congestive heart failure
RC wrist AVFs
Blood flow rate (ml/min)
Time interval between AVF placement
and start of the study (months)
Time interval between AVF placement
and first venipuncture (months)
Prior number of AVF failures/patient
66.6±11.5
53.1±56.6
42.3
1.3±0.5
49.8
11.5
2.6
76.3
6.7
1056.9±776.9
64.5±53.4
4.4±1.2
0.4±0.6
The data are expressed as mean±SD.
Studies
At the baseline, the 52 patients were cannulated with two
needles in the standard way; the distance between the tips,
pointing in opposite directions, ranged from 13 to 21 cm, and
the arterial needle faced upstream. Vascular access blood
flow rate was evaluated by means of the ultrasound dilution
Transonics Hemodialysis Monitor HD01 as previously
described [4]. For the measurement of vascular access
blood flow rate, the dialyser blood lines were reversed in a
sterile manner and a temporary recirculation channel was
created. The blood pump rate was set to 300 ml/min. The
baseline vascular access blood flow rate was determined by
averaging two separate measurements taken 5–10 min apart
during the first 30 min of dialysis.
Subsequently, the patients underwent follow-up studies
consisting of the periodic assessment —every 12 months for
4 years—of vascular access blood flow rate rate by means
of USM conducted in the same manner as for the baseline
measurements. Arterial blood pressure was measured immediately after vascular access blood flow rate measurement,
and is expressed as mean arterial pressure (MAP).
All tests throughout the study were performed by the same
operator (G.R.).
Blood samples were collected for the determination of
BUN, which was done using routine automated methods
(Cobas Mira S, Roche, Italy). Single-pool Kt/V was estimated
monthly.
Study protocol
The main point of our 4 year prospective observational
study, which started on 1 January 1999, was to follow the
natural history of the 52 mature autogenous RC wrist AVFs
Statistics
The survival of AVFs was evaluated by means of Kaplan–
Meier analysis. The log-rank test was used for life-table
The natural history of arteriovenous fistulas
1233
Table 2. Events occurring during the 4 year prospective observational study in the 52 patients with autogenous RC wrist AVFs
Vascular access failures
Rate of AVF thrombosis
Per patient-year at risk
Transplants
Deaths
1 years
2 years
3 years
4 years
1
0.019
4
0.075
2
0.030
2
0.061
2
2
2
2
1
6
0
4
comparisons. To identify threshold values of vascular access
blood flow rate that predict AVF failure, a receiver operating
characteristic (ROC) curve was constructed [6] by plotting
sensitivity vs the false-positive rate (FPR) at different cut-off
levels of vascular access blood flow rate. Sensitivity was
defined as the percentage of failed AVFs with positive test
results; specificity was defined as the percentage of patent
AVFs with negative test results, and FPR (equal to 1 –
specificity) as the percentage of patent AVFs that tested
positive. Friedman’s test was used for assessing the trends
of repeated measures of MAP and of vascular access blood
flow rate during the follow-up period. Data are expressed as
means±SD and median/range. All the statistical inferences
were derived using the SPSS 10.1 (SPSS Inc., Chicago, IL,
USA) software package. A P-value <0.05 was considered
significant.
Results
All vascular access failures resulted from AVF thrombosis, no non-thrombosed vascular access failed to
allow an adequate dialysis dose, i.e. a single-pool Kt/V
1.2. AVF thrombosis occurred in nine patients; 19
cases were censored out of the study (14 deaths and five
transplants). The rate of AVF thrombosis per patientyear at risk was 0.043 (Table 2). A thrombosed AVF
was abandoned and a new vascular access was created
either on the same or on the opposite arm.
Figure 1 shows the ROC curve for the diagnostic
accuracy of baseline vascular access blood flow rate
values in predicting AVF failure. The area under the
curve was 0.82 ± 0.05 (P ¼ 0.01) and the value of
Fig. 1. Receiver operating characteristic curve analysis of vascular
access blood flow rate measurements as a predictor of AVF failure.
The area under the curve (AUC) was statistically significantly >0.5
(P ¼ 0.01) and the cut-off value for predicting vascular access
failure was <700 ml/min.
Global
9
0.043
5
14
vascular access blood flow rate identified as a predictor
of AVF failure was <700 ml/min with an 88.9%
sensitivity and 68.6% specificity. When subdividing
the population of AVFs into two groups according to
baseline vascular access blood flow rates, two out of the
nine thromboses occurred among the AVFs that had
baseline vascular access blood flow rates >700 ml/min
(n ¼ 31), whereas seven occurred among those that had
baseline vascular access blood flow rates <700 ml/min
(n ¼ 21). The 4 year cumulative actuarial survival was
74.36 and 20.80%, respectively (log-rank test, P ¼ 0.04)
(Table 3 and Figure 2). The 24 AVFs that remained
patent at the end of the 4 years of follow-up kept a
median vascular access blood flow rate 900 ml/min at
all time points studied, and did not show any
differences in MAP values (Figure 3). Worth noting
is that five of them (20.8%) remained patent throughout the study with blood flow rates consistently
500 ml/min.
Discussion
Two main results derive from this study. The first, a
very low occurrence rate of vascular access failures/
AVF/year and a high actuarial survival of mature
autogenous RC wrist AVFs is a reality we confirm.
Two points need to be clarified about this outcome.
(i) We had a low prevalence of diabetic patients in our
population (11.5%). As is known, diabetes is a risk
factor for vascular access thrombosis. (ii) We studied a
population of prevalent AVFs. Thus, it is possible that
we selected a series of AVFs with different thrombotic
and survival rates than would be present in an incident
population of AVFs. All of the above considerations
notwithstanding, a number of studies have shown that
mature AVFs have superior performance, when compared with AVGs, in terms of patency rates and the
incidence of infection [7,8]. Thus, the first message of
the present study is that autogenous RC wrist AVFs
perhaps should be the initial type of permanent
vascular access in the majority of HD patients. Two
recent papers support our thesis: Rodriguez et al. [9]
showed that AVFs have longer survival than AVGs and
that, among AVFs, RC ones have the highest long-term
function rate (45% at 10 years and 38% at 12 years).
Furthermore, Konner et al. [10] published on a series
of 748 consecutive primary AVFs made at a single
center. The thrombosis rates were 0.03 and 0.07 per
patient-year at risk for non-diabetics and diabetics,
respectively. Despite these data, the number of
1234
C. Basile et al.
Table 3. Events occurring during the 4 year prospective observational study in the 52 patients with autogenous RC wrist AVFs: the latter
were subdivided into two groups, one including those AVFs with a baseline vascular access blood flow rate (VA BFR) <700 ml/min, the
other including those AVFs with a baseline VA BFR >700 ml/min
VA BFR <700 ml/min (n ¼ 21)
VA BFR >700 ml/min (n ¼ 31)
VA failures
Cumulative survival (%)
Transplants
Deaths
7
2
20.80
74.36
2
3
7
7
Fig. 2. When subdividing the population of 52 autogenous RC wrist AVFs into two groups according to the baseline vascular access blood
flow rate (VA BFR), two out of the nine thromboses occurred in the group of AVFs which had a baseline VA BFR >700 ml/min (n ¼ 31),
whereas seven occurred in the group of AVFs which had a baseline VA BFR <700 ml/min (n ¼ 21). The 4 year cumulative actuarial
survival was 74.36 and 20.80%, respectively (P ¼ 0.04).
synthetic grafts placed in the United States has been
increasing relative to AVFs [8,11]. The high rate of
maturation failure of the RC AVFs (13–70%) [12,13],
late referral to surgeons, timing of dialysis, the
economics related to the type of access, and the
inadequacy of veins for distal AVFs have all been
implicated in the preference given to AVGs [7,8,11].
Our second result, that a vascular access blood flow
rate <700 ml/min appears to be a reliable cut-off point
at which to implement a policy of closer monitoring of
this parameter, may lead to further investigations
and possibly interventions relevant to the function of
AVFs. For both AVFs and AVGs, it is recommended
that vascular access blood flow rate measurements be
performed monthly [3]. The guidelines further suggest
that a vascular access with a blood flow rate 600 ml/
min or 1000 ml/min that has decreased by more than
25% over 4 months should be referred for a fistulogram
[3]. Recommendations [3] have supported such monitoring on the premise that the natural history of the
vascular access will be altered by radiological or
surgical interventions after vascular access dysfunction
is detected. However, there is no clear evidence that
such a strategy of prospective monitoring of vascular
access blood flow rate followed by radiological or
surgical intervention significantly improves patency
[14–20]. Most studies that have evaluated surveillance,
whether done by vascular access blood flow rate or
venous pressure, suggest that prophylactic repairs of
vascular access reduce thrombosis rates and increase
cumulative graft survival; however, those studies have
used historic control groups or sequential groups, or
have been retrospective [14–16]. Actually, some very
recent randomized controlled studies in AVGs
have not confirmed that the prophylactic repair of
vascular access reduces thrombosis rates and increases
The natural history of arteriovenous fistulas
1235
Fig. 3. Comparison of vascular access blood flow rate (VA BFR) and of the MAP in the 24 patients with mature autogenous RC wrist
AVFs which remained patent during the 4 year follow-up. No statistically significant difference was found when using Friedman’s test
(P ¼ 0.09 for VA BFR; P ¼ 0.63 for MAP). The median VA BFR was 900 ml/min at all time points studied.
cumulative graft survival [17–19], whereas a preliminary report on a randomized controlled study of AVFs
shows that the pre-emptive correction of subclinical
stenoses improves AVF longevity [20]. Our study did
not clarify this issue, because its design does not include
a second arm (regular monitoring of vascular access
blood flow rate coupled with early intervention), which
might have given the answer that, even at the very low
rate of AVF thrombosis per patient-year at risk shown
in this study, a policy of regular monitoring of vascular
access blood flow rate coupled with early intervention
could be worthwhile. Nonetheless, we think that our
data (the very low rate of AVF thrombosis per patientyear at risk associated with the long-term patency of
five AVFs with a consistent blood flow rate 4500 ml/
min on the one hand, and the different outcomes of
AVFs associated with different baseline blood flow
rates on the other hand) allow us to suggest that: (i)
performing monthly blood flow rate measurements in
all AVFs as recommended by the NKF K/DOQI
guideline may be excessive; (ii) a vascular access blood
flow rate <700 ml/min appears to be a reliable cut-off
point at which to implement a policy of closer
monitoring of this parameter.
In conclusion, our present prospective observational
study shows a very low rate of AVF thrombosis per
patient-year at risk and a high survival of autogenous
RC wrist AVFs, particularly of those with vascular
access blood flow rates >700 ml/min. Thus, a vascular
access <700 ml/min would appear to be a reliable
cut-off point at which to start closer monitoring of this
parameter, which can culminate in further investigations and perhaps in eventually initiating interventions
relevant to the function of AVFs.
Conflict of interest statement. None declared.
References
1. Excerpts from the United States Renal Data System 1997
Annual Data Report. Am J Kidney Dis 1997; 30: S1–S195
2. Dhingra RK, Young EW, Hulbert-Shearon TE, Leavey SF,
Port FK. Type of vascular access and mortality in U.S.
hemodialysis patients. Kidney Int 2001; 60: 1443–1451
3. National Kidney Foundation. K-DOQI clinical practice guidelines for vascular access: update 2000. Am J Kidney Dis 2001;
37: S137–S181
4. Krivitski NM. Theory and validation of access flow measurement by dilution technique during hemodialysis. Kidney Int
1995; 48: 244–250
5. Tattersall J, Greenwood R, Farrington K. Urea kinetics and
when to commence dialysis. Am J Nephrol 1995; 15: 283–289
6. Tessitore N, Bedogna V, Gammaro L et al. Diagnostic
accuracy of ultrasound dilution access blood flow measurement
in detecting stenosis and predicting thrombosis in native forearm
arteriovenous fistulae for hemodialysis. Am J Kidney Dis 2003; 42:
331–341
7. Hakim R, Himmelfarb J. Hemodialysis access failure. A call to
action. Kidney Int 1998; 54: 1029–1040
8. Chertow GM. Grafts vs. fistulas for hemodialysis patients.
Equal access for all? J Am Med Assoc 1996; 276: 1343–1344
1236
9. Rodriguez JA, Armadans L, Ferrer E et al. The function of
permanent vascular access. Nephrol Dial Transplant 2000; 15:
402–408
10. Konner K, Hulbert-Shearon TE, Roys EC, Port FK. Tailoring
the initial vascular access for dialysis patients. Kidney Int 2002;
62: 329–338
11. Peters VJ. Increasing the prevalence of the arteriovenous fistula
in hemodialysis patients: an ongoing network-wide quality
improvement project on vascular access. Dial Transplant 1999;
28: 207–209
12. Tordoir JHM, Rooyens P, Dammers R, van der Sande FM,
de Haan M, Yo TI. Prospective evaluation of failure modes
in autogenous radiocephalic wrist access for haemodialysis.
Nephrol Dial Transplant 2003; 18: 378–383
13. Hakaim AG, Nalbandian M, Scott T. Superior maturation and
patency of primary brachiocephalic and transposed basilic vein
arteriovenous fistulae in patients with diabetes. J Vasc Surg
1998; 27: 154–157
14. Schawb J, Oliver MJ, Suhochi P, McCann R. Hemodialysis
arteriovenous access: detection of stenosis and response to
treatment by vascular access blood flow. Kidney Int 2001; 59:
358–362
C. Basile et al.
15. McCarley P, Wingard RL, Shir Y, Pettus W, Hakim RM,
Ikizler TA. Vascular access blood flow monitoring
reduces access morbidity and costs. Kidney Int 2001; 60:
1164–1172
16. Sands JJ, Miranda CL. Prolongation of hemodialysis access
survival with elective revision. Clin Nephrol 1995; 44: 329–333
17. Ram SJ, Work J, Caldito GC, Eason JM, Pervez A,
Paulson WD. A randomised controlled trial of blood flow and
stenosis surveillance of hemodialysis grafts. Kidney Int 2003; 64:
272–280
18. Moist LM, Churchill DN, House AA, et al. Regular
monitoring of access flow compared with monitoring of
venous pressure fails to improve graft survival. J Am Soc
Nephrol 2003; 14: 2645–2653
19. Dember LM, Holmberg EF, Kaufman JS. Randomized
controlled trial of prophylactic repair of hemodialysis
arteriovenous (AV) graft stenosis. J Am Soc Nephrol 2002;
13: 600A
20. Tessitore N, Lipari G, Mansueto G et al. Prospective
randomized trial of effect of preemptive correction of stenosis
on arteriovenous fistula (AVF) survival. J Am Soc Nephrol
2002; 13: 34A
Received for publication: 18.5.03
Accepted in revised form: 12.11.03