Vascular remodeling after the creation of an arteriovenous fistula as angioaccess for hemodialysis: the predictive value of a patient-specific computer simulation model.

Hemodialysis dependent patients, suffering from end-stage renal disease, need a
well-functioning vascular access to connect them with the dialyzer. Usually,
the vascular access is surgically created by making a connection between an
artery and a vein (arteriovenous fistula=AVF). Due to the low resistance leak
created and the high arterial-venous pressure difference, vessel adaptation
(remodeling) would occur with an enormous increase in blood flow and proximal
vessel dilatation. As a results, proper cannulation and the connection with the
dialyzer becomes possible. The use of an autogenous wrist AVF (between the
radial artery and cephalic vein) is preferred due to its better long-term
patency and its smaller complication rate compared to more proximal AVF's and
grafts. However, a significant amount of wrist AVF's fail directly after the
surgical intervention due to thrombosis and/or insufficient vessel remodeling.
Unless the fact that all those patients are selected based on preoperative
diagnostics (duplex, physical examination and CE-MRA), 30% of all newly created
wrist AVF's fail within six weeks after surgery and are useless for
hemodialysis. For the planning of the type of vascular access for each
individual patient, it is very important to preoperatively predict the
postoperative flow increase and vessel remodeling. When insufficient vessel
remodeling is likely, based on preoperative diagnostics, another type of
vascular access can be chosen.
The hypothesis is that the amount of failing AVF's can be reduced, if the blood
flow increase through the AVF can be better predicted with the available
preoperative diagnostics, since vessel adaptation is related to blood flow
increase. A patient-specific computer simulation model, based on preoperative
MRA, duplex data and bloodexaminations for biomarkers can possibly give insight
in the postoperative blood flow increase and failure incidence for different
fistula configurations. The big advantage of the computer model is that the
combination of the different factors, influencing AVF failure, like vessel
diameters, vessel compliances and accessory veins, can be varied and examined.

The objective of this study is to examine if a patient-specific computer
simulation model is able to predict the blood flow increase after AVF creation.
The interaction between the different factors that influence the blood flow
increase will be investigated. In addition, the ability of the model to predict
the peripheral perfusion is examined, which ultimately should be able to
predict hypoperfusion (ischemia). Second, a non-contrast enhanced MRA sequence
will be developed to overcome the problem of contrast administeration in
patients with end-stage renal disease. Finally, a biomarker ucMGP will be
correlated to vascular compliance, which is an important factor in calibrating
the model.

A computer model that is able to preoperatively predict the postoperative blood
flow increase through the AVF, is developed and tested. The choice for the type
of vascular access in the individual patient in this study, is based on the
results of preoperative diagnostics. The preoperative computer simulations are
in this study not used to determine the clinical policy, but only to compare
the postoperative measured pressure and flow with the simulated pressure and
flow.
The preoperative models are adapted to patient-specific conditions by using
data that are obtained from available preoperative modalitities: duplex,
CE-MRA, bloodexamination, blood pressure measurements and cardiac output/flow
examinations. The measurements protocols are developed in earlier research, but
should be adapted for our application. Therefore, a pilot study with 5
volunteers is performed before the main project starts. This pilot study should
result in a measurement protocol that is specific for our application. This
protocol is used in the main project to obtain the patient-specific data for
the model.
The results of the preoperative model are compared with preoperative pressure
and flow measurements to validate the model. Thereafter the postoperative blood
flow and pressure are simulated. The simulated blood flow is correlated with
the measured blood flow (duplex and MRA) to determine the accuracy of the
postoperative model predictions. In addition, the simulated and measured
postoperative blood pressure are compared. Follow-up after surgery will take
place on a regular basis. MRA examination will take place preoperative, 1 week
post-operative, 6 weeks post-operative, 6 months post-operative and 18 months
post-operative. Bloodwithdrawel will take place previous to the OR, out of the
venous canula which is placed in order to perform surgery. No additional
venapunctions will be performed.
For both the patients with a failed AVF and the patients with a
well-functioning AVF, is examined if the model was able to accurately predict
the postoperative blood flow increase.

All participants in this study will receive all pre- and postoperative MRA
examinations according to the follow-up protocol, a duplex examination and
pressure measurements (radial artery and finger). The duration of all
examinations in total is approximately 500 minutes (90 minutes per MRI and 10
minutes per duplex). All measurements are performed when the patient is already
in the hospital for hemodialysis, for another examination not related to this
study or when they are in the hospital one day before surgery.
For the CE-MRA examination a Gadolinium containing contrast agent is injected
by intravenous infusion. The total incidence of adverse events related to
gadolinium use for CE-MRA is less than 5%. The incidence of any single event is
even lower than 1%. The most common events are nausea, headache and emesis.
When used intravenously, no nephrotoxicity has been detected and the rates of
adverse events are extremely low. However, recently, concerns have arisen
regarding the use of gadolinium because some gadolinium containing contrast
agents are associated with Nephrogenic Systemic Fibrosis (NSF) in patients with
severe renal impairment. NSF is a rare, debilitating and sometimes fatal
disease. Therefore, the contrast agent gadobutrol (Gadovist(c), Bayer Schering
Pharma AG, Germany) with a very low dose of gadolinium is used in this study.
Although, this contrast agent contains a low dose of gadolinium, this contrast
agent is uptil now not associated with the development of NSF and is at the
moment specifically used in the clinic to make CE-MRA scans in patients with a
renal disorder. If this contrast agent will be associated with the development
of NSF in future, the use of gadobutrol is stopped immediately. Before the
CE-MRA scan, the patient will be informed thoroughly by the surgeon.