| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Radiology, Tenon University Hospital, 4, rue de la Chine,
F-75970 Paris Cedex 20, France.
2 Department of Radiology, Charles Foix University Hospital, 7, ave. de la
République, F-94205 Ivry-sur-Seine Cedex, France.
3 Department of Nephrology B, AP-HP, Tenon University Hospital, F-75970 Paris
Cedex 20, France.
Received January 2, 2002;
accepted after revision April 1, 2002.
Address correspondence to A.-F. Le Blanche.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Over a 20-month period, 23 patients (15 women, eight men) with an age range of 42-87 years (mean, 63 years) having end-stage renal insufficiency and with recent hemodialysis fistula surgical placement underwent gadoterate-enhanced digital subtraction angiography with a digital 1024 x 1024 matrix. Opacification was performed on the forearm, arm, and chest with the patient in the supine position using an injection (retrograde, n = 14; anterograde, n = 8; arterial, n = 1) of gadoterate meglumine into the perianastomotic fistula segment at a rate of 3 mL/sec for a total volume ranging from 24 to 32 mL. Percutaneous transluminal angioplasty was performed in three patients and required an additional 8 mL per procedure. Examinations were compared using a 3-step confidence scale and a two-radiologist agreement (Cohen's kappa statistic) for diagnostic and opacification quality. Tolerability was evaluated on the basis of serum creatinine levels and the development of complications.
RESULTS. No impairment of renal function was found in the 15
patients who were not treated with hemodialysis. Serum creatinine level change
varied from -11.9% to 11.6%. All studies were of diagnostic quality. The
presence of stenosis (n = 14) or thrombosis (n = 3) in
arteriovenous fistulas was shown with good interobserver agreement (
=
0.71-0.80) in relation to opacification quality ( = 0.59-0.84). No
pain, neurologic complications, or allergiclike reactions occurred. Three
percutaneous transluminal angioplasty procedures (brachiocephalic, n
= 2; radiocephalic, n = 1) were successfully performed.
CONCLUSION. Gadoterate-enhanced digital subtraction angiography is an effective and safe method to assess causes of malfunction of hemodialysis fistulas. It can also be used to plan and perform percutaneous transluminal angioplasty.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Hemodialysis is required earlier in the more advanced cases of renal insufficiency. Although conventional or digital subtraction fistulography is considered a reference standard for assessment of fistula status [4], it is nevertheless performed with potentially nephrotoxic iodinated contrast agents. In patients with severe renal insufficiency, fistulography enhanced with iodinated contrast agents may be deleterious and responsible for permanent deterioration of renal function. This examination can precipitate the patient into hemodialysis and can also increase the cost of treatment. Although the incidence of nephropathy induced by iodinated contrast agents ranges from 2% to 7% in the general population undergoing a contrast-enhanced procedure [5], the risk of nephrotoxicity with iodinated contrast agents in patients with preexisting azotemia is close to 12% [5,6,7] and increases to 33% in patients with associated diabetes mellitus [8]. Moreover, the incidence of nephropathy induced by iodinated contrast agents would be even higher in patients with known end-stage renal disease if iodinated contrast agents were not contraindicated in this setting. A nonnephrotoxic contrast agent would, therefore be useful to protect the remaining renal function of patients with end-stage renal disease.
Gadolinium-based contrast agents are effective and well tolerated and are less toxic to the nephrons than iodinated contrast agents [9, 10]. In patients having renal insufficiency, IV administration of gadopentetate dimeglumine at doses of up to 0.4 mmol/kg for MR imaging does not induce nephrotoxicity [9, 10]. MR angiography has been accepted by nephrologists as the reference technique to evaluate renal artery stenosis in patients with renal insufficiency [9]. Gadolinium chelates have also been used in patients with a history of allergic-like reactions to iodinated contrast agents as an alternative radiographic contrast agent, initially for CT [11, 12] and then for digital subtraction angiography [13]. Gadolinium chelates have also been proposed as a contrast agent for angiography [14] and for various types of interventional procedures in patients with renal insufficiency [15,16,17], but until now, little experience with hemodialysis fistulas has been reported [18]. The aim of our study was to evaluate the feasibility, safety, and reliability of a gadolinium-based complex as a contrast agent for opacification of arteriovenous fistulas in diagnostic procedures or before percutaneous transluminal angioplasty.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Imaging Techniques
The upper limb examined was immobilized in the supine position with 45°
of abduction. Fistulography was performed on a high-resolution digital
subtraction system (Multistar T.O.P.; Siemens, Erlangen, Germany) after
retrograde (n = 14) and anterograde (n = 8) puncture of the
fistula or puncture of the inflow artery of the fistula (n = 1) with
an 18-gauge angiocatheter. The examinations were performed with three
sequential subtracted acquisitions, including the forearm in the case of a
radiocephalic fistula, the arm, and the superior vena cava in the thorax.
Digitalized images were obtained with the following technical parameters: 96
kVp, 6-25 mAs, a sequence of three images per second, a 1024 x 1024
matrix, and a 40-cm field of view. Each step of this technique was performed
without any movement of the table. The first digital images of one series were
acquired just before the beginning of the injection and constituted the
"mask" that could be subtracted from opacified images.
The same automatic injector (Mark V Plus; Medrad, Philadelphia, PA) as that used for vascular examinations enhanced with iodinated contrast agents was used. Undiluted gadoterate meglumine (Dotarem; Laboratoire Guerbet, Aulnay-sous-Bois, France), a gadolinium complex with extracellular biodistribution similar to that of gadopentetate dimeglumine, was injected at a volume of 8 mL for the forearm and the arm and 16 mL for the thorax, at an injection rate of 3 mL/sec. The gadolinium complex was injected using a low-resistance extension tubing set (length, 150 cm; lumen diameter, 3.0 mm; external diameter, 4.1 mm; Codan, Lensahn, Germany) corresponding to a volume of 8 mL. One extension tubing set was filled with 8 mL for opacification of the forearm and arm. The tubing was loaded with a gadolinium complex directly with a syringe via a three-way stopcock, and the injector was loaded with saline. The injector pushed the column of contrast agent with 20 mL of saline to flush the tubing set. The gadolinium-based contrast agent was pushed by the outflow of the fistula. A tourniquet was placed to obtain reflux in the anastomosis and retrograde opacification of the inflow artery. The tourniquet was cautiously released during imaging to allow reflux and antegrade flow through the fistula and to perform the imaging of the entire segment simultaneously. A total volume of 8 + 16 = 24 mL was used for brachiocephalic or brachiobasilic fistulas and 8 + 8 + 16 = 32 mL for radiocephalic fistulas. An additional 8 mL was injected for checking fistula diameter when percutaneous transluminal angioplasty was performed.
Image Analysis
Subtracted and nonsubtracted images from the same patients were interpreted
independently by two radiologists. The radiologists were instructed to grade
the quality of opacification of the fistula, the cephalic and basilic veins of
the forearm and arm, and their capacity to show fistula abnormalities. The
radiologists graded the veins from 0 (insufficient visualization) to 2 (good
visualization) for opacification quality and for diagnostic confidence of
normal status, stenosis, or thrombotic occlusion. This diagnosis was based on
the analysis of the diameter of the vein and its uniformity. The subclavian
veins, innominate veins, and superior vena cava were graded for opacification
quality according to the same scale, from 0 to 2. The vascular radiologic
images of the same patient were presented randomly so that the images were not
examined in sequence except by chance. The patients' names were obscured from
the radiologists. The interobserver correlation was evaluated using Cohen's
kappa correlation coefficient
[20]. The correlation was
calculated according to each vascular segment, for diagnosis and quality of
opacification. Cohen's kappa statistic is a statistical measure designed to
assess agreement between two or more observations for categoric or nominal
data. It determines the proportion of decisions in which observers agree while
accounting for the possibility of agreements attributable solely to chance.
Perfect agreement results in a kappa value of 1.0; a kappa value of 0
indicates the level of agreement expected on the basis of chance alone. Less
agreement than that expected by chance results in a negative kappa value.
Kappa values of 0.20 or less indicate slight agreement; 0.21-0.40, fair;
0.41-0.60, moderate; 0.61-0.80, substantial; and 0.81-1.00, almost perfect
agreement between observers.
Analysis of Tolerability
Serum creatinine levels before (1-3 days) and after (1-5 days) the
examination were measured because they appear to be the most effective and
least expensive laboratory parameter to evaluate renal function
[7]. Patients with
contraindications to iodinated contrast agents were 15 patients not treated
with hemodialysis (end-stage renal insufficiency, n = 10; renal
transplant recipients, n = 5), patients treated with dialysis with
residual diuresis (n = 4), or patients with recurrence of an
allergiclike syndrome to iodinated contrast agents despite usual preventive
antihistamine chemotherapy (n = 4).
The definition of renal failure was the one most commonly used in the literature [5,6,7]: more than 25% increase of serum creatinine levels 48-72 hr after the examination. A variation of the [(creatinineafter — creatininebefore) / creatininebefore] x 100 ratio expressed as a percentage was considered to be more appropriate for our group of patients with high creatinine levels than the rise in creatinine levels alone. No hyperhydration or fasting was required for the examination. The development of neurologic disorders, pain, or discomfort during gadolinium chelate injection was noted by the radiologist performing the examination.
Follow-Up
The course of immature fistulas was monitored at 2 and 6 months after
fistulography.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The vein of the forearm ( = 0.71) and the vein of the arm ( =
0.78) were visible in all cases (100%), whereas the anastomotic area (
= 0.80) was visible in all fistulas opacified by retrograde injection (14
[61%] of 23 fistulograms). A good interobserver correlation was also observed
for grading of opacification quality with kappa values ranging from 0.62 to
0.84. Opacification quality also allowed visualization of the subclavian
(100%, = 0.66), brachiocephalic (100%, = 0.72), and innominate
veins and the superior vena cava (82%, = 0.59; moderate agreement).
The contrast agent in one of the innominate veins was washed out by the
contralateral flow. Decreased contrast agent was visualized in the innominate
veins and superior vena cava because of X-ray attenuation through the tissues
combined with respiratory motion. However, visualization of the thoracic veins
kept sufficent accuracy for diagnosis in all patients (Tables
1 and
2). The findings of the
examination were considered to be normal in six (26%) of 23 patients. Twelve
(52%) of 23 venous stenoses were visualized (cephalic vein, n = 10;
basilic vein, n = 1; innominate vein, n =1). One (4%) of 23
stenoses of the inflow artery and one stenosis of the anastomosis (4%) were
shown. Fistula mal-function was related to venous occlusion (basilic vein,
n = 1; cephalic vein, n = 1; innominate vein, n =
1) in three (13%) of 23 patients. Overall, gadolinium-enhanced digital
subtraction fistulography contributed to the choice of the treatment on the
basis of percutaneous transluminal angioplasty or surgical recreation of the
fistula.
|
|
Tolerability
No adverse event related to an allergiclike reaction and no hot sensations
were reported for any of the examinations. No neurologic disorders nor
extravasation occurred. The examination was considered painless, whereas
injection of CO2 was painful for the patients (n = 2) who
had undergone both types of angiography.
All patients presented renal insufficiency pending hemodialysis, and no impairment of renal function was observed in any patients with end-stage renal insufficiency. The serum creatinine levels remained unchanged before (range, 135-1093 µmol/L; mean, 502 ± 284 µmol/L) and after (range, 133-1103 µmol/L; mean, 485 ± 288 µmol/L) the examination. Individual variations ranged from -60 µmol/L (-11.9%) to 70 µmol/L (11.6%), always remaining below the reference 25% threshold for the definition of renal failure.
Follow-Up and Treatment Modalities
Percutaneous transluminal angioplasty was performed in three patients (Fig.
1A,1B,1C),
and new arteriovenous fistulas were created in two patients. In another two
patients with normal findings on fistulography, superficialization of the
fistula proved to be effective. One immature fistula at the time of evaluation
(2 months) subsequently presented with normal development. Percutaneous
transluminal angioplasty procedures (n = 3) were technically
successful in the immediate postprocedural period. Surgery consisted of
creating a brachiobasilic fistula in two patients. The 6-month follow-up
showed that the 21 other fistulas matured, remaining patent with effective
flow.
|
|
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The exact time interval between creation of the arteriovenous fistula and hemodialysis is not easy to predict, and some patients may therefore require opacification of their arteriovenous fistulas before initiation of hemodialysis. Fistulography enhanced with iodinated contrast agents is associated with a high risk of permanent deterioration of renal function in these patients, leading to premature dialysis and an increased cost of treatment. Hemodialysis can also be precluded by a malfunctioning arteriovenous fistula, which then requires an additional invasive catheter access procedure. When the anastomotic area has been evaluated by both CO2 and gadolinium, the quality of opacification of the arteriovenous fistula is better with gadolinium than with CO2 [18].
It is difficult for the reviewer to appreciate the grading because one cannot determine the basis for the grades in the absence of a gold standard (digital subtraction angiography enhanced with iodinated contrast agents). Nevertheless, interobserver agreement calculation does not require a gold standard. Interobserver agreement then is provided by reviewers whose expertise is used daily in similar vascular procedures. Although observers do not have a gold standard to grade against, good interobserver correlation likely means that abnormalities have many chances to be correctly found or excluded with a high confidence level.
Gadolinium and iodine have similar properties in terms of radiopacity: gadolinium has an atomic number of 64 and a k-absorption edge of 50 keV, whereas iodine has an atomic number of 53 and a k-absorption edge of 33 keV. Gadolinium-based contrast agents can absorb sufficient energy to be opaque during digital subtraction angiography. Gadoterate meglumine (376.9 mg/mL) results in 3.5-fold lower attenuation rate than that induced by iodixanol (320 mg/mL), between 73 and 96 kVp and 100 mA [21]. Gadolinium-based contrast agents provide weaker vascular enhancement during digital subtraction angiography than iodinated contrast agents because of the semimolar concentration of gadolinium in commercially available preparations of MR contrast agents, fivefold less than that of iodinated contrast solutions. In practice, digital subtraction must be used to obtain informative images. A fistulography protocol using successive series allowing subtraction during all sequences improves visualization of the vascular segments of the arm and forearm.
In contrast with the findings of a previous study indicating that opacification of the thoracic veins was of poorer quality because of a dilution effect and respiratory motions [22], sufficient opacification was obtained in our study with a maximal amount of 40 mL of gadoterate meglumine, because venous flow was improved by the arterial push of the arteriovenous fistula. As the contrast agent passes from a smaller to a larger lumen and is accompanied by unopacified inflow from jugular and then brachiocephalic veins contralateral to opacified fistula, the most central veins appear to be patent, but the superior vena cava may not completely opacify. This result can be a limitation of the technique.
In our center, 30-60 mL of gadoterate meglumine is commonly injected for MR angiography of the aorta, with dose adjustments depending on the patient's weight. Additional injections might be planned if inadequate images are obtained. Regarding alternative contrast agents to iodine, opacification quality is commonly considered to be less with CO2 than with iodinated contrast agents, which induce overestimation of stenoses by CO2 [23].
Diagnostic Performance and Follow-Up
Good interobserver correlation was found for opacification quality and
diagnosis of the cause of fistula malfunction. Opacification quality was good
for all segments, inflow artery segments or fistula segments of the forearm
and arm. This quality determines the reliability of the therapeutic choice
between percutaneous transluminal angioplasty and repeated surgery
[24]. Despite the good
interobserver correlation, clinical evidence and follow-up data are considered
to be the most reliable method to evaluate a diagnostic technique
[25], and all patients treated
with percutaneous transluminal angioplasty, surgical repair, or re-creation of
their arteriovenous fistula based on gadolinium fistulography data obtained a
good clinical result in our study.
Tolerability
The safety of gadolinium in MR imaging has been clearly shown, with a 2.4%
incidence of side effects, most of which (94%) are minor
[26]. In our study, no hot
sensations or allergiclike reactions were noted, even in the four patients
with a history of allergiclike reactions to iodinated contrast agents;
gadoterate meglumine was therefore routinely administered without preliminary
fasting. No deterioration of renal function occurred with total administered
doses (40 mL) of less than 0.3 mmol/kg (0.6 mL/kg). This finding confirms the
high level of safety reported in the literature. Absence of nephrotoxicity has
been observed for doses of up to 0.4 mmol/kg
[9]. Gadolinium chelates have
been assessed to more precisely evaluate areas of suspected stenosis than
CO2; in a recent series in the literature, the range of volume of
gadolinium administered was 24-60 mL (mean, 37 mL) with no signs of
nephrotoxicity [18].
CO2 fistulography could be recommended in the presence of a
contraindication to iodine, but CO2 classically requires specific
experience with a significant learning curve, causes pain during venous
injections [27], overestimates
stenosis [28], and can lead to
serious complications
[29,30,31].
Regarding the literature reporting CO2 hazards and renal safety of
gadolinium, no proposal of a randomized comparison study of CO2
versus gadolinium-based agents versus iodinated contrast media was considered
acceptable by our institutional ethics committee.
Hemodynamic Safety
A hazardous aspect of CO2 imaging is the possibility of
inadvertent delivery of massive volumes of air due to contamination by room
air: CO2 and air are invisible gases, inducing undetected
contamination that may result in a potentially fatal air embolism. An increase
of pulmonary artery pressure and a decrease of systolic pressure were shown
experimentally when a large dose (6.4 mL/kg) was delivered
[32]. Dogs promptly died after
IV injection of 2000 mL of CO2, apparently because of gas trapping
in the right side of the heart; excessive volumes can flood the right side of
the heart, leading to pump failure. No side effect of this kind has been
reported with gadolinium complexes.
Neurologic Safety
The neurotoxicity of CO2 is not anecdotal for intraarterial
injections, and supradiaphragmatic injection is clearly contraindicated to
avoid intracranial delivery of CO2. The compressibility of the gas
may result in an unpredictable and explosive delivery
[33] and a possible flow in
the opposite direction toward the cerebral circulation in the case of
arteriovenous fistula [23] or
axillary-femoral graft [29]:
one patient was reported to raise his head, experience dizziness, and become
unconscious after several diagnostic injections. In a recent study of 32
patients, five received an injection of CO2 to evaluate their
arteriovenous fistulas with deliberate countercurrent reflux into the
anastomotic area and the artery followed by severe neurologic complications:
two patients experienced seizures and another one experienced loss of
consciousness with a 30-sec respiratory arrest
[23]. CO2 injection
is contraindicated for reflux into the anastomotic area of an arteriovenous
fistula [29,
30], but reflux can also occur
accidentally in the presence of stenosis or even occlusion. This finding
constitutes the main indication for fistulography. Because CO2 may
accidentally reflux into the native artery, gadolinium has been reported to be
useful for assessment of the arterial anastomosis or in patients with an
occluded hemodialysis graft
[18]. Intracardiac septal
defects are another possible cause of exposure of carotid vessels to
CO2 in cases of IV injections.
Although CO2 is substantially less expensive than gadolinium complexes, CO2 requires an expensive injector dedicated to gas injection, whereas gadolinium is injected with a standard injector for iodinated contrast agents. MR angiographic sequences now available on recent MR imaging systems cannot visualize the hemodynamics of arteriovenous fistulas and thus overestimate stenoses [34]. Although MR angiography has been shown to be effective for visualization of the arterial system, it is considerably less effective for exploring the peripheral venous system.
Although gadolinium chelates are five to 10 times more expensive than nonionic iodinated contrast agents, this cost should be compared with the expenses related to renal insufficiency and premature dialysis. Approximately 20% of patients with a serum creatinine level greater than 2.0 mg/dL develop acute renal insufficiency [35] after iodinated contrast agent injection. The cost of dialysis sessions three times weekly is considerably higher than that of gadolinium fistulography. However, gadolinium digital subtraction angiography may represent the ideal opacification method for patients with end-stage renal insufficiency because of the complete absence of nephrotoxicity with gadolinium doses of less than 0.4 mmol/kg [9, 10], although a recent study showed the ability of antioxidant acetylcysteine to prevent nephrotoxicity induced by iodinated contrast agents in patients with only low-stage renal insufficiency [36].
Gadolinium fistulography is feasible and appears to be a safe and effective technique to assess various causes of hemodialysis fistula malfunction such as stenosis or thrombosis. Surgery may be planned and percutaneous transluminal angioplasty may be performed after thorough evaluation. The absence of nephrotoxicity makes gadolinium chelates suitable for patients with renal insufficiency not yet requiring hemodialysis. A maximal volume of 40 mL is sufficient to safely evaluate the anastomotic area and the entire venous segment of the arteriovenous fistula and to perform balloon dilation. The higher cost of gadolinium chelates should preclude their use in patients who are already being treated with hemodialysis for the reasons outlined. A more detailed analysis could be initiated to compare the tolerability and efficacy of the gadolinium complexes versus iodinated contrast agents associated with acetylcysteine in end-stage renal disease.
Acknowledgments
We thank Sandrine Boucheteil and Fabienne Cheminant for secretarial
expertise and assistance in editing and Anthony Saul for manuscript
revision.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  P. M. Colletti Nephrogenic Systemic Fibrosis and Gadolinium: A Perfect Storm Am. J. Roentgenol., October 1, 2008; 191(4): 1150 - 1153. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
