Intimal Hyperplasia in Prosthetic. Arteriovenous. Fistulas for Hemodialysis. Is Associated with. Initial High Flow Velocity and Not with Mismatch in Elastic. 1.
Anastomotic Intimal Hyperplasia in Prosthetic Arteriovenous Fistulas for Hemodialysis Is Associated with Initial High Flow Velocity and Not with Mismatch in Elastic Properties 1
L. Hofstra,2
D.C.J.J.
and
Tordoir
J.H.M.
Bergmans,
K.M.L.
Leunissen,
L. Hofstra. D.C.J.J. Bergmans, P.J.E.H.M. Kitslaar, Tordoir, Department of Surgery, Cardiovascular search Institute Maastricht, University Hospital tricht, Maastricht, The Netherlands K.M.L.
Leunissen,
vascular pital
Department
Research
Maastricht,
of Nephrology,
Institute
Maastricht,
Maastricht,
A.P.G.
J.H.M. ReMaasCardio-
University
Hos-
The Netherlands
A.P.G. Hoeks, Department of Biophysics, Cardiovascular Research Institute Maastricht, University Hospital Maastricht, Maastricht, The Netherlands
M.J.A.P. vascular pital
Daemen, Research
Maastricht,
(J. Am.
Soc.
Nephrol.
The
1995;
Netherlands
6:1625-1633)
--------ABSTRACT Stenotic intimal thickening at the venous end of prosthetic arteriovenous (AV) fistulas for hemodialysis has been associated with perianastomotic mismatch in elastic properties, and low shear rates. In a prospective way, the role of these factors on the occurrence of intimal hyperplasia in prosthetic AV fistulas in hemodialysis patients was investigated. In 24 hemodialysis
patients,
the
elastic
properties
were
assessed
in
the distal graft segment and the oufflow vein postoperatively with the use of Vessel Wall Doppler Tracking (VWDT), a noninvasive ultrasound technique. In addition,
normalized
peak
systolic
velocity
(nPSV)
was
calculated from diameter (VWDT) and peak systolic velocity. The initial mismatch around the venous anastomoses and local nPSV were correlated with the occurrence of stenoses during follow-up (2 yr). The detection of a stenosis was performed with both Duplex ultrasound and angiography. In four cases, a stenosis
developed
in the
venous
eight cases, a stenosis developed flow segment; and in four cases, at both
sites. A better February
1
Received
2
Correspondence
of Maastrlcht,
21
,
initial
1995. Accepted
to Dr. L Hofstra, P0 Box 5800,
stenoses
match August
Department
anastomosis; in the venous in elastic
18,
6202 AZ, Maastrlcht,
The Netherlands.
1046.6673/0606-1625$03.00/0 Journal of the American Society of Nephrology Copyright © 1995 by the American Society of Nephrology
Journal
of the
American
Society
in
out-
developed properties
1995.
of Cardiology,
of Nephrology
around fistulas
Hospital
Kitslaar,
M.J.A.P.
Daemen,
the venous developing
anastomosis was observed in the a stenosis at this site as compared with the nonstenotic fistulas (P < 0.05). The initial local nPSV values atthe site ofthe later stenosis were higher in the fistulas developing a stenosis as compared with the nonstenotic fistulas (P < 0.05). It was concluded that the occurrence of stenoses in prosthetic AV fistulas for hemodialysis in or adjacent to the venous anastomoses is associated with a high initial flow velocity but not with a mismatch in elastic properties.
A
use
hyperplasia.
(AV)
compliance.
fistulas
of interposition
by the mostly
The
Intimal
nteniovenous development occurring
repair
of these
patients stenoses
stenotic
(2,3). The developing
unclear. Studies levels of shear reported
of
evaluating nate on the data.
AV fistulas with
often
the
complicated
intimal anastomoses
thickening. (1-4).
leads
to a mean
lesions
1 month
reasons at the
conflicting
are
of occlusive the venous
time
flow.
constructed
grafts
at
hospitalization
each
for the venous
year
for
these
high incidence anastomoses
the influence development
In animal
of are
of varying of stenoses
studies,
increased
intimal thickening was found in areas with low shear rates, indicating that low shear stresses may induce the Intimal thickening (5-12). A possible factor contnibuting to low shear rates at a site just downstream from the venous anastomosis may be the sudden increase in the capacity to stone blood volume at this site ( 1 3), beading to flow disturbances such as flow separation high shear
and rates
stasis damage
the lumen and injury, possibly concept intimal of AV
thickening
On the endothelial
was
other cells
result in hyperplasia
is supported by thickening in the fistulas. The most
observed
velocities ( 1 7, 18). A second mechanical growth may be a
around match
( 14). the
may therefore initiating intimal
The latter evaluating segments
mechanical University
P.J.E.H.M.
Key Words:
Department of Pathology. CardioInstitute Maastricht, University HosMaastricht,
Hoeks,
anastomoses, ( 12, 19-22). stretching,
in areas condition difference in
This
with
hand, very covering
vessel
animal venous severe
the
affecting elastic
wall (15,16). studies outflow intimal
highest
flow
mntimal properties
also called compliance misbeads to points of excessive possibly
resulting
in
vessel
wall injury. Furthermore, excessive mechanical stretching may Induce vascular smooth muscle cell replication, as has been demonstrated in In vitro models (23-26). However, an association between Intimal
1625
lntimal
Hyperplasia
thickening convincingly
in Arteriovenous
and compliance demonstrated
Fistulas
mismatch so far (20).
has
not
been
In this study, vessel wall elastic properties were measured around the venous anastomoses of polytetnafluonoethylene (PTFE) AV fistulas in hemodialysis patients. These measurements were performed within 2 wk after the construction of the AV fistula, with the use of Vessel Wall Doppler Tracking (VWDT) (27,28). In addition, peak systolic velocity (PSV) was measured at defined sites in the graft and in the inflow and outflow vessels with the use of Duplex ultrasound. From PSV and end-diastolic diameter, measured with VWDT. normalized peak systolic velocity (nPSV) was calculated, a rough approximation of the actual shear rate exerted on the endothellal cells covering the lumen vessel wall. Furthermore, stenotic tissue specimens were harvested in the patients who developed a stenosis during follow-up. necessitating surgical nepain. These specimens were histologically evaluated with respect to the endothelial cell coverage of the lumen, determined with the use of Immune staining. The aim of the present study was to define the influence of both mismatch in elastic properties around the anastomosis and flow velocity on the occurrence of occlusive intimal hyperplasia in or adjacent to the venous anastomoses in patients with a prosthetic hemodlalysis AV fistula. In addition, the possible influence of flow velocity on the presence of endothebial cells covering the lumen was evaluated.
MATERIALS AND METHODS Patient
Clinical
Profile
Twenty-four consecutive patients receiving a hemodialysis fistula with the use of PTFE prosthesis (WL Gore, Flagstaff. AZ) entered the study (December 1991 to June 1993). Generally. the construction ofthe fistula with the use
graftAV
of Interposition grafts was a secondary procedure. In four patients, the construction of a graft AV fistula was a primary choice.
because
of
the
absence
of
suitable
veins
In
the
forearm for the creation of a Brescia Cimino fistula (Table 1). In all patients, the arterial and venous anastomoses were constructed
at
the
level
of the
elbow.
The
brachial
artery
served as the Inflow segment; the basilic, cephalic, or cubital veins served as outflow segments. The grafts were implanted in a loop configuration in the forearm. The cause for endstage renal failure was diabetic nephropathy in two patients.
Tracking (VWDT) (27,28). ThIs system combines B-mode imaging (AlL, Ultramark IV, Bothell, WA) and off-line processing. VWDT offers the possibilIty to assess the distension and diameter accurately at defined sites In the graft, the inflow artery, and the efferent vein. In vitro studies showed that VWDT Is able to detect distenslons as low as 1 m, indicating that the expected low distension values of the PTFE grafts can be measured properly. The recording time of four s per registration allows measurement of distension (IDeltalD) and internal diastolic diameter (D) during 3 to 5 consecutive heartbeats. The values of these consecutive beats were averaged in this study. From the diameter and distension the following parameters were calculated:
Relative
Distension
(1W)
Area-Increase
Relative
distension
(RD)
(Al)
gives
the
-h-
=
=
100
(%)
irZD . D 2
strain
of the
vessel
during
the cardiac cycle. Area-increase (Al) represents the capacity of the vessel to store blood volume during the cardiac cycle. Generally. resulting coefficient,
the parameters are corrected for pulse pressure, in the distensibility coefficient and compliance respectively (29). In studies evaluating the pressure distribution in graft AV fistulas, It was shown that mean pressure and pulse pressure decrease along the graft and within the anastomoses (17, 18). The extent of the decrease varies with graft geometry. This implicates that valid pressure values can only be obtained with the use of Invasive pressure measurements at all sites Investigated. Since this would be unethical, we decided not to use pulse pressure in
the equations. Diameter
and
cm proximal
distension
In the distal graft 1 In the venous In the efferent vein, 1 cm distal from the (Figure 1 ). The positioning of the cursor
from
the
anastomoses, and venous anastomosis
line
at these
imaging
Since
of the
particular
were
venous
sites
anastomotic
we attempted
assessed
anastomoses,
was
facilitated
complex
to evaluate
with
by repeated
B-mode
the Influence
imaging.
of mismatch
in elastic properties around anastomoses on the occurrence of intimal hyperplasia, we indexed the values for A! and RD around the anastomoses, resulting in the match factor. Also
a match in the
factor size
was
calculated
differences
around
for diameter the
to obtain
anastomoses.
insight
The
match
hypertenslon/renovascular merulonephrItis in nine
disease in seven patients. gbpatients. polycystic kidney disease unknown causes In two patients. Mean
in four patients. and blood flow during dialysis
was
Measurement
Elastic
The ments
of the
diameter and were measured
TABLE 1 Patient .
to 300 mL/min.
Properties
Figure 1 Schematic
characteristics
Sex
Primary/Secondary
Graft
distension of the vessel and graft segwith the use of Vessel Wall Doppler
Age
1626
250
fistula
61 yr (39-68) 1 1 m ale/13 female 4/20
. drawing of a venous anastomosis of an arterlovenous fistula. Most of the stenoses developed at the site of the anastomoses (A) or in the venous outflow segment just distal from the anastomosis, within 1 cm (V). The lines represent the positioning of the cursor lines during a measurement of diameter and distension with VWDT.
Volume
6
.
Number
6
.
1995
Hofstra
factors
for
following Match
the
different
parameters
were
calculated
in the
way: factor
=
value
efferent
vein/value
distal
graft
A perfect match Is considered to be present when the match factor Is one. A mismatch value higher than 1 implicates an increase In the parameter within the anastomotic area, whereas a value lower than 1 implIcates a decrease of the parameter in the direction of blood flow.
Determination
of Velocity
et al
Parameters
Normalized peak systolic velocity (nPSV) was calculated from peak systolic velocity (PSV) and diameter. PSV was measured by placement of a Doppler sample midstream in the vessel at an angle of 60#{176} . At the same location , diameter was assessed with the use of VWDT. Diameter measurements with the use of VWDT have been validated in a previous study. The variation coefficIent for diameter varied between 3 and 5%. Indicating low variation in diameter assessment with the use of this system (30). PSV and diameter were assessed In the distal graft ( 1 cm from the anastomosls), within the venous anastornoses, and in the venous outflow segments (within 1 cm from the anastomoses). From PSV and diameter, nPSV was calculated:
As the chromogen. 3.3’ diamlno-benzidine was used. As a negative control, spare saphenous vein segments were used. obtained from patients undergoing peripheral bypass surgery. The degree ofendothelial cell coverage was estimated by dividing the circumference of the lumen covered by endothehal cells by the total luminal circumference in serial sections.
Analysis Data are presented as medians and 25th and 75th percentile Intervals. To compare differences between groups. the Mann-Whitney-U test was used. The local initial mechanical and flow conditions of the fistulas developing a stenosis at either the venous anastomosis or the efferent vein were compared with the conditions of the fistulas without a stenosis at these locations. Since the initial mechanical and flow conditions were assessed at well-defined sites In the anastomotic complex. good correlation of these local conditions with the site of the later stenosis was possible. To test which of the mechanical factors showed an independent correlation with the occurrence of intimal thickening. multiple regresslon analysis was performed. Mechanical factors showing a correlation with the occurrence ofintimal thickening with a P value < 0. 10 were included In a stepwise multiple regression model. A P value < 0.05 was accepted as level of statistical significance.
nPSV=PSV/Diameter
RESULTS Most patients were measured within 2 wk after the construction of the AV fistula. In some cases (N = 4) the measurements were delayed because of wound leakage or hematoma formation. These patients were measured within 4 wk after operation. All patients were measured In the supine position after an acclimatization period of 20 mm. At strict time Intervals (2. 6 and 12 wk, and every 3 months thereafter). a Duplex scan of the AV fistula was performed for the detection of possible stenoses. When a stenosis was detected with the Duplex scan, an anglogram was made to confirm the diagnosis. A stenosis was defined as a luminal narrowing resulting in substantial impairment of shunt flow, leading to difficulties In dialysis treatment. In general, this occurs when a severe stenosis ofabout 90% is present, which necessitates surgical correction of these AV fistulas. This approach differs from criteria used in studies evaluating restenosis after percutaneous transluminal coronary angioplasty (PTCA). where angiographic criteria such as 50% lumen narrowing or late lurne boss are often used. The study was approved by the medical ethical committee of the University Hospital of MaastnIcht. All patients gave written informed consent before entering the study.
Histologic
Examinations
Stenoses
Stenotic tissue was harvested for histologic evaluation in patients developing a stenosis necessitating repair with an interposition graft. These tissue specimens were fixed overnight In 10% phosphate-buffered formalin, processed according to routine histologic procedures and paraffin-embedded. Serial sections (5 .trn thick) were cut and placed on gelatin-coated glass slides. To identify endothellal cells, Ulex Europeus agglutinin ( 1 :500 dilutIon, Vector Labs, Burlingame. CA) was used. After blocking of the endogenous peroxidase activity with 0.3% H202, sections were Incubated with the lectin antibody ( 1 :250 dilutIon for 45 mm; DAKO) after predigestion with pepsin ( 1 mg/mL for, 30 mm). Subsequently. the sections were incubated with an anti-Ulex Europeus lectin peroxidase complex ( 1 : 100 dilution: DAKO).
Journal
of the
American
In the 24 patients, a total of 20 stenoses developed in 16 fistulas in the penianastomotic area at the venous site during the follow-up period (mean 12.2 ± 8.7 [SD] months); in eight patients. no fistula stenosis was found during the study period. In four fistulas, a stenosis occurred only at the venous anastomosis. In eight cases. a stenosis developed only in the effenent vein adjacent to the anastomosis (within 1 cm from the anastomosis, Figure 1 ) and In four fistulas stenoses developed at both sites. This resulted in a total of eight stenoses in the venous anastomosis and 1 2 in the effenent vein adjacent to the anastomosis. In all patients with a stenosis. surgical repair of the stenotic area was necessary to reestablish a sufficient flow for dialysis treatment. indicating that the stenosis was occlusive on suboccluslve. Stenoses occurring at sites more downstream from the efferent vein (N = 4) were not included in this study because of the difficulty to correlate the initial mechanical and flow information with the site of these stenoses.
Society
of Nephrology
at the
Venous
Anastomoses
The initial RD values in the venous anastomoses were higher (Table 2) in the fistulas developing a stenosis in the venous anastomoses as compared with the fistulas without stenosis development at this site (2.5 versus 1 .0%. P < 0.05). No differences in initial diameter match and diameter at the anastomoses were observed between the groups. The Initial match factors for A! and RD (ARM and RDM, respectively, Table 2 and Figure 2) were closer to 1 in the stenotic fistulas as compared with the fistubas without a stenosis at this site ( 1 .5 versus 5. 1 for ARM and 1.2 versus 2.6 for RDM, P < 0.05), indicatIng a better
1627
Intimal
Hyperplasia
in Arteriovenous
Fistulas
TABLE 2. Values for the different parameters measured at the venous anastomosis, two weeks after construction, for the AV fistulas with a later stenosis at this point (stenotic) and the fistulas without
a stenosis
(non-stenotic)
at this site during
fOllOW-(Jp.0’
Stenotic
(N
=
Nonstenotic
RD (%)
2.5 (1.5-3.2)
DIAM (mm) ARM RDM
4.1 (3.6-4.4) 1.5 (0.8-2.7) 1.2 (0.9-2.2)
nPSV (1/s)
(N
8)
1.0 3.5 5.1 2.6 212
308 (261-346)
=
,
value
16)
(0.4-1.6) (3.1-4.5) (2.8-8.4) (2.0-5.0) (168-276)
0.02 0.82 0.005 0.01 0.02
DISCUSSION
a RD. relative
distension; DIAM. diameter; ARM. match factor area RDM. match factor relafive distension; nPSV. normalized
increase; peak
systolic
Values vals. b
velocity.
are presented
as medians
and 25- and 75-percentile
inter-
initial match in elastic properties around the venous anastomoses in stenotic fistulas. Higher initial values for nPSV were found in the anastomoses of stenotic fistulas as compared with the fistubas without a stenosis (308 versus 2 12 1 Is, P < 0.05, Figure 3).
Stenoses
in the
Efferent
Vein
No differences in initial RD and vein diameter (Table 3) were found between the fistubas developing a stenosis in the effenent vein (within 1 cm from the anastomosis) and the fistulas without a stenosis at this site. Also no differences in diameter match around the venous anastomoses were observed between the groups. The initial match values for A! and RD (ARM and RDM. Table 3, Figure 2) around the venous anastomoses were closer to 1 In the fistulas with a later stenosis In the effenent vein as compared to the nonstenotic fistulas ( 1 .8 versus 3.7 for ARM and 1.2 versus 2.5 for RDM. P < 0.05). InitIal nPSV values in the effenent vein were higher In the fistulas with a later stenosis as compared to the fistulas without a stenosis in this area (272 versus 99 1 Is, P < 0.05, Figure 3).
Regression
Analysis
Multiple regression analysis showed that an increase in initial nPSV values was independently conrelated with the occurrence of subocclusive intimal thickening (r = -0.55, P < 0.05) at both the venous anastomosis and the effenent vein. In addition, an RDM closer to 1 around the venous anastomoses, indicating a better match, was correlated with the occurrence of a stenosis (r = 0.4 1 , P < 0.05). For match found.
in
Histologic In nine anastomosis
1628
A!
(ARM),
no
independent
correlation
was
Examinations cases, was
the repair performed
of with
the the
position graft, which allowed examination of the stenotic lesions. A total of 1 1 stenotlc specimens were derived from the anastomoses or the efferent vein adjacent to the anastomoses. In all samples the stenotic lesion consisted mainly of smooth muscle cells and extracellular matrix. The mean percentage of endothelial cells covering the lumen was 48% . In the control saphenous vein specimens (N = 15) a mean percentage of endothelial cells covering the lumen was 95%.
stenotic venous use of an inter-
This study contains two major findings. First, the occurrence of stenotic intimal hyperplasia in prosthetic AV fistulas in humans Is associated with initial high flow velocity at the site of the later stenosis,
suggesting
that
high
flow
velocity
may
initiate
the
development of stenotic intimal hypenplasia at the venous anastomoses of AV fistulas for hemodialysis. Second, a better initial match in strain around the venous anastomosis was found In fistulas developing a stenosis In this area as compared with the nonstenotic fistulas. This strongly argues against an impontant contributing effect of mismatch In elastic properties around anastomoses of prosthetic graft AV fistulas in the development of stenotic intimal hyperplasla.
Flow
Velocity
and
Intimal
Thickening
In the fistulas developing a stenosis, higher initial nPSV values were observed at the venous anastomoses and the outflow venous segments as compared to the nonstenotic fistulas. This suggests that high flow velocities may play a role in the Initiation of intimal thickening at the venous site of AV fistulas used for hemodialysis. The data from the present study confirm observations In animal studies evaluating the effect of varying levels of flow on the occurrence of intimal thickening at the venous outflow segment of AV fistubas. In these studies, more intimal thickening was observed In fistulas with the highest flow vebocities at this site ( 18,3 1 ). However, numerous investigations reported that bow shear rate promotes intimal thickening (5-12.32). Several points have to be considered when comparing the effects of shear rate on mntimal thickening. including the range of shear rate Investigated. the possible difference between the negulatlon of intimal thickening in animals and humans, and the different nonphysiological
response shear
of rate.
effect of a decreased flow was of normal arterial flow (33,34). range of shear rates investigated and low, level in between stantially tomoses 350 cm/s
arteries In some
and veins studies,
compared with This indicates lies between
to the
the effect that the normal
which is a different range compared with the our study. In normal arteries, PSV values 60 to 120 cm/s are present, leading to subbower shear rate levels. In the venous anasof an AV fistula, PSV values as high as 300 to for comparable diameters can be measured,
Volume
6
.
Number
o
-
1995
Hofstra
Stenotic
et al
Nonstenotic
6
4
RDM
*
*
2
0
Anastomosis Figure
2. Bar graph
stenosis (stenotic) a later
stenosis
showing
the initial
and nonstenotic in the
venous
values
Vein
for the match
factor
in relative
fistulas. In the fistulas with a later stenosis outflow segment, a more ideal match
distension
(RDM)
for the fistulas
at the venous anasfomoses, in relative distension was
as well observed
developing as fistulas (*, P
