Full text of DOI:10.1111/j.1755-6686.2003.tb00272.x

MANAGEMENT OF VASCULAR ACCESS IN EUROPE:
PART 2 - A MULTI CENTRE STUDY OF RELATED COMPLICATIONS
Monique M. Elseviers, Jean-Pierre Van Waeleghem and Elizabeth Lindley for the Research Board of the EDTNA/ERCA.
SUMMARY
Introduction | The pilot project of the Research Board of EDTNA/ERCA handled the management of vascular accesses
(VA) in European dialysis centres. In the first part of the study, centre policies related to VA management were
investigated. This paper reports on the second part of the project, investigating VA related complications reported
during an observational prospective study.
Methods | A cohort of 1380 adult patients, randomly selected in 47 centres out of 16 European countries were
followed during one year using a computerised data collection system. Data were collected at baseline, after six and
12 months and each time a VA complication occurred.
Results | At the start of the observation period, 77% had a native AV fistula, 10% had an AV graft and 13% a catheter.
A total of 489 complications were noted. Most frequently observed were thrombosis, stenosis, infection, bleeding and
flow problems. Hospitalisation (mean duration=6.2 days) was required in 39% of complications and 29% of
complications resulted in a definitive loss of VA. Complications were more frequently observed in catheters (27%) and
AV grafts (37%) compared to AV fistulae (15%). When compared with AV fistulae, the risk for thrombosis was more
than four times higher and for bleeding more than six times higher if an AV graft was used. Catheters showed an
eightfold increased risk to develop infections and flow problems.
Conclusion | This study revealed the high complication rate in VA and strengthened the actions to promote AV
fistulae as first choice VA.
KEY WORDS
• Haemodialysis
• Vascular access
• AV fistula, graft
• Catheter
• Complication
Complications of vascular access (VA) remain a signifi-
cant clinical problem in the chronic treatment of patients with
end-stage renal disease. Recent literature suggests that access relat-
ed complications account for at least 25% of all hospital stays (1).
In addition to the financial burden, VA complications also reduce
dialysis adequacy and lead to a significant reduction in the qual-
ity of life of dialysis patients (2).
This paper reports on the second part of the research project and
will present results of an observational prospective study in indi-
vidual patients, investigating vascular access related complications.
In this first article, observed complications will be described and
their relation to the type of vascular access will be highlighted.
MATERIAL AND METHODS
In all participating European dialysis centres, a cohort of 30
chronic haemodialysis patients per centre was followed for one
year. Data concerning the type of VA in use and related param-
eters were selected at the start of the study period and after six
and 12 months. All VA related complications requiring interven-
tion were registered at the time of occurrence using an electronic
data collection system.
In 1996, the Research Board of the European Dialysis and
Transplant Nurses Association / European Renal Care Association
(EDTNA/ERCA) selected the management of VA as their pilot
research project. The first part of the project studied centre-
focused policies for the management of VA in 103 European dial-
ysis centres (3). Results showed that large differences were
observed in VA care and hygienic procedures used. It was con-
cluded that it would be extremely interesting to relate the large
differences observed in the centre policies of VA management to
individual patients outcome parameters.
| SELECTION PROCEDURE |
All centres included in the first part of the study were invited to
participate. Participating centres were asked to send a list of all
45
EDTNA|ERCA JOURNAL 2003 XXIX1

MANAGEMENT OF VASCULAR ACCESS IN EUROPE:
PART 2 - A MULTI CENTRE STUDY OF RELATED COMPLICATIONS
MedIQal® was used. After random selection of patients, the proj-
ect installation software was sent to the participating centre. For
all selected patients a standard data set was filled in at baseline,
after six months and after 12 months. The standard data set includ-
ed information concerning the type of VA, puncture techniques
used, function of the VA and medication used. Throughout the
one-year study period, VA related complications requiring inter-
vention were registered as well as details concerning related
interventions and hospitalisation.
Haemodialysis patients included
Age median (range)
n=1380
64 years (15-90)
56%
Sex (% males)
Duration of treatment for ESRD median (range)
Age of vascular access in use median (range)
Previous permanent VA used
No previous
37 months (6-370)
30 months (0-300)
59.2%
22.9%
17.9%
One
More than one
Table 1: Description of study population.
| STATISTICAL ANALYSIS |
patients matching the inclusion criteria (see below) to the coor-
dinating centre in Antwerp, Belgium. The list included patients’
initials, birth date and sex. In the coordinating centre, a cohort
of 30 patients was randomly selected from this list. If the total
number of patients in a centre was below 30, all eligible patients
were included.
Basic statistical techniques were used to describe the types of VA
used, complications observed and hygienic handling applied.
Students’ t-test was used for comparison of continuous vari-
ables and chi-square test for comparison of percentages, con-
sidering a p-value of p < 0.05 as the significance level. For com-
plications, in-depth analysis was only performed on complications
occurring frequently: thrombosis, stenosis, infection, bleeding and
flow problems. Per patient, each reported complication was only
counted once, also when it occurred more than one time dur-
ing the study period. Differences in complication rates were relat-
ed to patient characteristics, type of VA and puncture techniques
used. Complication rates were also related to centre character-
istics and hygiene policies, questioned during the first, centre based
part of the study. Risk ratios for complications were calculated using
logistic regression analysis enabling control for confounding fac-
tors.
| SELECTION CRITERIA |
All adult patients who started a chronic haemodialysis programme
at least six months prior the start of the study were included for
at random selection. Children of less than 15 years and patients
on haemodialysis for less than six months were excluded.
| DATA COLLECTION |
Centres started the study in the period 1998-1999. For data
Distribution of vascular access
Other
RESULTS
The questionnaire was distributed to 108 centres and 81 of them
sent their patients’ list to the coordinating centre enabling a
random selection of 30 patients. Finally, 47 centres out of 16
European countries completed the study. Participating centres were
mainly from United Kingdom (n=11), Belgium (n=9) and Germany
(n=6). In the remaining countries (Austria, Cyprus, Denmark,
Finland, France, Greece, Italy, Malta, the Netherlands, Norway,
Slovenia, Spain and Switzerland) less than four centres participated.
0.7%
Catheter
13.0%
Graft
9.6%
| PATIENT CHARACTERISTICS |
A total of 1380 patients were included. Median age was 64
with 56% male. Patients were on chronic haemodialysis therapy
for a median of 37 months (Table 1). During the one year study
period 15.5% of patients died, 6.4% were transplanted, 0.7%
switched to PD and 4.5% were lost to follow-up.
AV Fistula
76.7%
collection, a computer-based questionnaire developed by
Figure 1: Types of vascular access at baseline.
| DESCRIPTION OF VA IN USE AT BASELINE |
The category 'other' included 12 patients treated with a
combination of an AV fistula and a catheter.
At start of the observation period, 77% of patients had a native
AV fistula, 10% had an AV graft and 13% a catheter. AV grafts
46 EDTNA|ERCA JOURNAL 2003 XXIX1

MANAGEMENT OF VASCULAR ACCESS IN EUROPE:
PART 2 - A MULTI CENTRE STUDY OF RELATED COMPLICATIONS
were predominantly synthetic grafts (89%). Twelve patients were
treated with a combination of an AV fistula and a catheter
(Figure 1).
was 15% in AV fistulae, increasing to 27% in catheters and to
37% in grafts. Taking into account only patients with first VA
(n=792), complication rates showed even more pronounced
differences between the different types of VA with 12% in AV fis-
tulae, 23% in catheters and 47% in AV grafts. In AV fistulae and
AV grafts, thrombosis and stenosis were most frequently observed.
In catheters, about half of complications reported were infections
(Figure 3).
At baseline, 59% of patients were dialysed with their first VA and
23% with their second (Table 1). The remaining patients had
between three and ten permanent vascular accesses. Considering
only patients dialysed with their first VA (n=792), 88% had an
AV fistula and only 4% an AV graft.
In table 2, risk ratios for developing VA complications compar-
ing different types of VA were presented. Since it was found that
complication rates were highly influenced by the size of the
treating dialysis centre and a first or a subsequent VA, present-
ed risk ratios were corrected for these factors. As shown in the
last column of the table, the risk of developing a complication
almost doubled for patients with catheters and tripled for patients
with grafts compared to patients with AV fistulae. Additionally,
comparing AV grafts with AV fistulae, the risk for thrombosis was
more than four times and the risk for bleeding more than six times
higher if an AV graft was used. When catheters were compared
with AV fistulae, the risk of developing infections and flow prob-
lems both showed an eightfold increase in catheters.
Median age of patients with a first permanent VA was respec-
tively 56, 57 and 66 for those with an AV fistula, an AV graft and
a catheter. Patients had their baseline VA for a median of 30
months. AV fistulae had a median maturation time of 43 days
and one out of ten AV fistulae required a surgical intervention
before first use.
| VA COMPLICATIONS |
During the one-year observation period, 19% of patients had a
VA complication requiring intervention, varying from one to 18
patients per centre (3 to 60%). A total of 489 complications were
noted, ranging from one to 20 complications per patient (mean
1.8).
DISCUSSION
Complications most frequently observed were thrombosis (25%),
stenosis (15%), infection (14%), bleeding (13%) and flow prob-
lems (12%) (Figure 2). The category of ‘other’ complications
included pain, haematoma and high cardiac output. Balloon
dilatation was used in one third of stenosis/thrombosis. A throm-
bolyser (usually a chemical one) was used in one fourth. Bleeding,
most frequently occurring post-dialysis was mainly treated with
medication all or not combined with compression. Surgical inter-
vention was required in 39% of complications.
During this one-year observation period, 489 VA related com-
plications requiring intervention were observed in a cohort of 1380
Vascular acces related complications
Other
10.1%
Steal syndrome
Thrombosis
2.9%
25.3%
Aneurysma
3.5%
High venous
pressure
3.7%
Hospitalisation was required in 39% of complications for a peri-
od from one to 64 days (mean 6.2). Particularly for thrombosis,
stenosis and aneurisma, more than half of the patients were hos-
pitalised. The longest hospitalisation period however, was noted
for infections (mean 12 days).
Flow
problems
11.9%
Additionally, 26% of complications required the use of a temporary
catheter, 3% resulted in a change of therapy and 29% resulted
in a definitive loss of VA. Definitive loss was most frequently
observed as a result of thrombosis (46%), steal syndrome (42%)
and infection (36%).
AV Fistula
15.4%
Bleeding
13.2%
Infection
14.0%
Figure 2: Vascular access related complications reported during
the observation period of one year. The category 'other' included
pain, haematoma and high cardiac output.
| COMPLICATIONS AND TYPE OF ACCESS |
During the observation period of one year, complication rate
47
EDTNA|ERCA JOURNAL 2003 XXIX1

MANAGEMENT OF VASCULAR ACCESS IN EUROPE:
PART 2 - A MULTI CENTRE STUDY OF RELATED COMPLICATIONS
patients. Particularly thrombosis, stenosis, infection, bleeding
and flow problems were frequently noted. VA complications
required hospitalisation in 39% and resulted in a definitive loss
of VA in 29% of patients.
In this random sample of 30 patients per unit, it was demonstrated
once more that the European dialysis patient is dialysed pre-
dominantly with a native AV fistula with 88% of first permanent
VA being an AV fistula. This is in sharp contrast with US practice
where only approximately 20% of patients have a native AV fis-
tula (4,5,6). On the other hand, AV grafts were used in only 10%
of European patients. Taking into account only first permanent
VA, this percentage decreased to 4%. These frequencies are far
below the US practice where about half of patients were dialysed
with an AV graft (4,6).
The observed hospitalisation rate due to vascular access related
complications seems to be rather high. A total of 186 hospitali-
sations with a mean duration of 6.2 days was registered result-
ing in a hospitalisation rate of 0.84 days per patient year. This
means that each patient of this European cohort had a chance
of spending one day in hospital per year for a vascular access relat-
ed complication. Since we did not ask for the total number of
hospitalisation days, we cannot compare our European observation
with the observation of PJ. Held et al (1) that in the US more than
one fourth of all hospitalisations were due to vascular access relat-
ed complications.
It is clear that clinical practice in first choice VA widely differs
between Europe and the US. While native AV fistulae are pref-
erentially used in European countries, AV grafts are most com-
monly used in the US (6). The DOQI clinical practice guidelines
for vascular access (7) are appreciated worldwide as reasonable,
AV fistula
(n=1049)
Graft
(n=131)
Catheter
(n=178)
37.4%
% with complications
40%
30%
20%
10%
0%
27.5%
15.5%
AV fistula
Flow problems 6.6%
Graft
Catheter
Flow problems 1.9%
Flow problems
18.4%
Thrombosis
18.4%
Bleeding
20.8%
Bleeding
14.5%
Thrombosis
36.2%
Bleeding
8.2%
Stenosis
8.2%
Thrombosis
45.3%
Infection
5.7%
Infection
14.5%
Stenosis
26.4%
Infection
46.9%
Stenosis
28.3%
Figure 3: Frequency of complications in AV fistula, AV graft and catheter and distribution of main complications per type of vascular access.
48 EDTNA|ERCA JOURNAL 2003 XXIX1

MANAGEMENT OF VASCULAR ACCESS IN EUROPE:
PART 2 - A MULTI CENTRE STUDY OF RELATED COMPLICATIONS
appropriate and a
F
c
i
h
st
i
u
e
l
v
a
abl
G
e
ra
t
f
o
t
olsCa
f
t
o
h
c
et
u
e
s
r
edp-v
o
a
n
lueth(2)e
R
lo
R
n
(9
g
5
s
%
ta
C
n
I)d(3)ing
AV fistulae are preferred and grafts are mostly used as second
choice. It is clearly shown that preferences and experience with
a particular VA type have an important influence on outcome (5,8).
Compared to the US experience, we have the impression that in
this European study AV grafts scored lower compared to US
results and that catheters showed better results as expected.
n=1049 n=131
n=178
graft vs. fistula:
2.77 (1.75–4.40)
catheter vs. fistula:
1.86 (1.23–2.81)
graft vs. fistula:
4.10 (2.22–7.56)
graft vs. fistula:
6.68 (2.28–19.5)
catheter vs. fistula:
8.19 (3.85–17.4)
catheter vs. fistula:
8.56 (2.87–25.5)
15.5%
15.5%
5.2%
2.1%
2.1%
1.0%
37.4%
—
0.000
0.000
0.000
0.000
0.000
0.001
Total complications
–
27.5%
–
Thrombosis
Bleeding
18.3%
CONCLUSION
This study demonstrated once more that complication rates of
VA are unacceptably high. Results corroborated the opinion that
AV fistulae need to be promoted as first choice VA for chronic
haemodialysis. In view of prevention, it can be concluded that a
common action of nephrologist, surgeon, radiologist and nurse
is needed to avoid VA complications. Primary prevention will
include the correct choice of type and localisation of VA and the
use of dialysis techniques with a low risk for complications.
Within this framework, nurses have a key role and can be con-
sidered as the gatekeepers of well-functioning VA. In daily clin-
ical practice they need to handle VA correctly and to be aware
of all the signs of possible complications.
8.4%
–
Infection
–
12.9%
Flow problems
–
5.1%
(1) Only complications showing significant differences are presented.
(2) P-values are based on univariate analysis and resulted from chi-square test or Fisher exact test.
(3) Relative risk (RR) and 95% confidence intervals (95% CI) are controlled for patient/nurse ratio and
the number of previous vascular accesses.
Table 2: Complications related to type of vascular access (1)
.
maintenance of vascular access survival. Since these guidelines are
developed in the US however, it can be asked if additional
European guidelines are appropriate in view of wide differences
in practice.
ACKNOWLEDGEMENT
The project questionnaire was developed with the help of all mem-
bers of the Research Board. Technical and administrative help was
provided by Colin Aldridge, Jan Aldridge and Jago Taylor (MedIQal,
Stevenage, United Kingdom) and by Dirk De Weerdt and Eddy
Van Hout (Department of Nephrology, University of Antwerp,
Belgium). We would like to thank all the participants of this
study. Without their sustained effort of data collection during a
whole year, this study could never have been realised.
In their 2000 update, the DOQI guidelines (7), stressed once more
their preference for a native AV fistula as first choice VA. They sug-
gest the AV graft as second choice and advise avoidance of dial-
ysis with a permanent catheter. In support of this viewpoint, we
could demonstrate that the risk of complications tripled for grafts
and doubled for catheters compared to native AV fistulae. Of
course, these results were obtained in the European situation where
49
EDTNA|ERCA JOURNAL 2003 XXIX1

MANAGEMENT OF VASCULAR ACCESS IN EUROPE:
PART 2 - A MULTI CENTRE STUDY OF RELATED COMPLICATIONS
REFERENCES
1. Held, PJ. Droppleman, RF. Dickinson, DM. Pisoni, RL. Port, FK. Wolfe, RA. Mapes, DL. Inpatient and outpatient treatment of vascular
access (VA). Journal of the American Society of Nephrology 1998; 9: 172A.
2. Hakim, R. Himmelfarb, J. Haemodialysis access failure: a call to action. Kidney International 1998; 54: 1029-1040.
3. Van Waeleghem, JP. Elseviers, MM. Lindley, EJ. Management of vascular access in Europe. Part I: A study of centre based policies.
EDTNA/ERCA Journal 2000; 4: 8-34.
4. Pisoni, RL. Young, EW. Greenwood, RN. Hecking, E. Goodkin, DA. Wolfe, RA. Held, PJ. Comparison of Vascular Access Use in the US and
Europe: The DOPPS. Journal of the American Society of Nephrology 1999; 10: 215A (abstract).
5. Young, EW. Dykstra, DM. Chen, K. Held, PJ. Permanent vascular access (VA) preferences and outcomes. Journal of the American
Society of Nephrology 1999; 10: 223A (abstract).
6. United States Renal Data System 2000 Annual Report: National Institutes of Health, National Institute of Diabetes and Digestive
and Kidney Diseases. Bethesda, Md., September 2000.
7. National Kidney Foundation. K/DOQI Clinical Practice Guidelines for Vascular Access, 2000. American Journal of Kidney Diseases
2001; 37(1, suppl 1): S137-S181.
8. Burger, H. Long-term outcome of different forms of vascular access, in Conlon PJ, Nicholson ML, Schwab S (eds.): Haemodialysis
vascular access: practice and problems, chap 4. Oxford, England: Oxford University Press, 2000, pp 52-66.
STATISTICS CORNER
Invitation
• Are you involved in scientific research in your renal unit?
• Do you have questions about the best method to choose for your research project?
• Are you confronted with some specific questions about the statistical analysis of the data that you collected?
• Did you read a scientific article and do you have questions about the statistical methods presented?
We kindly invite you to send your methodological and statistical questions to the journal editor, via email preferably
106111.2270@compuserve.com. In the following issues of this journal, we will try to handle your question(s).
Your input will contribute to make this statistical corner a useful forum for all kind of research problems. And
yes you can send your questions in the same language you receive the Journal.
Monique Elseviers
50 EDTNA|ERCA JOURNAL 2003 XXIX1