Volume 73, No.6: 2021 Siriraj Medical Journal
https://he02.tci-thaijo.org/index.php/sirirajmedj/index
373
Original Article
SMJ
Sornsupha Limchareon, M.D.*, Trakarn Chaivanit, M.D.**, Suchanun Osatheerakul, M.D.***
*Department of Radiology, **Department of Surgery, ***Department of…………………., Faculty of Medicine Burapha University Hospital, Chonburi
20131, ailand.
Direct CT Venography in ESRD Patients: Technical
Experience and Findings
ABSTRACT
Objective: e aims of this study were to describe direct computed tomography venography (CTV) for upper limb
venous system evaluation and to report on ndings in end-stage renal disease (ESRD) patients.
Materials and Methods: Direct CTV was performed using a 64-multidetector computed tomography (MDCT)
scanner with simultaneous injection of diluted iodinated contrast (IC); 1:4 at both elbows and 2-phase scanning
namely, the direct venous, and the arterial phases. e ndings in ESRD patients evaluated between November
2013 and March 2019 were retrospectively reviewed.
Results: Forty CTV examinations (600 venous segments) were performed and the volume of IC used per patient was
38 mL. Number of lesions found in a patient ranged from 1 to 6 and the majority had 1 to 3 lesions (30/38 patients).
Stenosis and thrombosis were the two most common ndings (112/600) and were equally prevalent. e three most
common sites of steno-occlusive complications were the brachiocephalic vein (29 lesions), the internal jugular vein
(25 lesions), and the subclavian vein (16 lesions). e most common site of stenosis was the brachiocephalic vein
(18 lesions), whereas the most common site of thrombosis was the internal jugular vein (20 lesions). No venous
aneurysms or ruptures were found. IC extravasation at the site of injection occurred in one arm in one patient.
Conclusion: Direct CTV has the advantage of requiring lower IC volume while maintaining direct visualization of
the venous system similar to conventional venography.
Keywords: Computed tomography venography; contrast material; upper limb,vascular access (Siriraj Med J 2021;
73: 373-379)
Corresponding author: Trakarn Chaivanit
E-mail: trakarn.cha@gmail.com
Received 22 January 2021 Revised 15 March 2021 Accepted 15 March 2021
ORCID ID: http://orcid.org/0000-0003-4044-723X
http://dx.doi.org/10.33192/Smj.2021.49
INTRODUCTION
Venous steno-occlusive disease aer establishment
of vascular access is a common complication in end stage
renal disease (ESRD) patients. Evaluation of the venous
system is important for acquiring information on the
lesion, which then facilitates treatment planning. Various
non-invasive and invasive imaging techniques can be used
to assess the venous system. e color Doppler ultrasound
is a non-invasive technique that is easy to perform but
its use is limited in some venous segments, especially the
central veins.
1
Contrast-enhanced magnetic resonance
(MR) imaging is also non-invasive, and does not require
iodinated contrast (IC); however, it is not easily accessible
and is expensive.
2
Conventional venography (CV) is
the gold standard for evaluating the venous system;
nevertheless, it is invasive and it requires injection of a
Volume 73, No.6: 2021 Siriraj Medical Journal
https://he02.tci-thaijo.org/index.php/sirirajmedj/index
374
Limchareon et al.
considerable amount of IC. Multi-detector computed
tomography venography (MDCT-V) is fast, easy to
perform, provides high spatial resolution, and has high
accuracy.
1,3-5
Additionally, three-dimensional images
can be acquired by CTV, whereas only two-dimensional
images are obtained with CV.
e two types of CTV are direct and indirect CTV,
and the indirect CTV technique is considered to be the
standard. It is performed by injecting IC at a dose of 2
mL/kg (total 100-200 mL) in a single extremity, usually
the unaected side, and scanning during the arterial
phase (second passage) using either a test bolus or bolus
tracking.
1,4-5
However, some fresh thrombi may show
identical density as contrast-enhanced blood and, thereby,
yield false negative results.
6
Moreover, inow phenomena
can mimic an intra-luminal thrombus (Fig 1). Direct CTV
of a lower extremity was rst reported in 1994 during
which diluted IC was injected in a supercial vein and
scanning was performed before the IC entered the right
atrium (rst passage).
6
In contrast to indirect CTV, the
advantages of direct CTV are lower IC injection volume
and the fact that scanning time is independent of the
patient’s blood ow velocity.
7
ere are a few reports
of direct CTV in upper extremities and the techniques
used are variable; nevertheless, the direct CTV technique
may be considered as a replacement for indirect CTV
for evaluating the upper limb venous system.
erefore, here, we describe our technique and
experience in performing direct CTV of the upper
extremity, discuss venous complications aer vascular
access, and investigate ndings of venous complications
in ESRD patients.
MATERIALS AND METHODS
We retrospectively reviewed ndings from ESRD
patients who had undergone CTV of both upper limbs
between 1
st
November 2013 and 31
st
March 2019. We
identied 38 patients (18 males and 20 females) with an
average age of 65 years (range 29-85 y). We excluded
one patient who underwent upper extremity CTV with
only single arm injection. In the cohort, one patient
underwent three studies while another required two
studies, all at dierent times. Data from a total of 40
exams were included, and information on the patient’s
presenting symptoms, and types of vascular accesses were
collected. e presence, sites, and number of stenosis
and/or occlusion were determined. e clinical data were
accessed from medical records while CTV ndings were
accessed from the PACS database.
e study was approved by the university review
board, approval No.282/2019. Informed consent was
waived due to the retrospective nature of the study.
CTV technique
Images were obtained with patients in the supine
position with their heads directed toward the gantry.
Hands were placed in the head-holder, superior to the
head, and immobilized with tape to limit motion. A
tourniquet was not used, as recommended by Mavili
et al.
8
CTV was performed using a 64-MDCT scanner
(Toshiba Aquillion one, Tokyo, Japan) with tube potential
set at 100 kV, and current at 300 mA. e parameters
for CTV were beam collimation 0.5 mm, pitch 0.641,
slice thickness 1.0 mm, and reconstruction interval 0.8
mm.
e scanning protocol progressed in the cranio-
caudal direction from the elbow to the superior vena
cava. To explore image quality and its usefulness, the
scanning phases were initially set, as follows-non-contrast
scan and three post-contrast phases. The first post-
contrast phase was scanned aer a delay of 7 seconds
from contrast injection (rst passage). e second phase
was repeated at the same scan length immediately aer
the rst phase for arterial opacication (second passage),
and the third phase was repeated immediately aer the
second phase at the same scan length. We observed
that no additional information was acquired from the
third phase. Moreover, if the patient experienced an
arterio-venous stula development, the opacied blood
was rapidly washed out from vascular system and no
vascular enhancement was observed during the third
phase. Therefore, the third post-contrast phase was
Fig 1. “Inow phenomenon” Non-opacied blood (arrow) from
le brachiocephalic vein draining into opacied blood in superior
cava, causing partial hypodensity mimicking thrombus
Volume 73, No.6: 2021 Siriraj Medical Journal
https://he02.tci-thaijo.org/index.php/sirirajmedj/index
375
Original Article
SMJ
omitted. However, the second phase was retained because
the internal jugular veins were better opacied in the
second phase than in the rst phase. is is because,
in the absence of central vein obstruction, the internal
jugular veins are usually not opacied in the rst phase.
IC fluids, with concentrations of 320, 350 and
370 mgI/mL (Visipaque, Omnipaque, and Ultravist,
respectively) were simultaneously administered through a
20-22 gauge IV catheter into the subcutaneous supercial
veins of both forearms just below the elbows using an
automated injector. Initially, the rate of injection was 2
mL/s for each extremity but based on our observations,
it was increased to 3 mL/s. Next, IC concentrations were
optimized as follows. Initially, IC was diluted at a ratio of
1:2 in saline which led to very high IC concentration in
the veins and artifacts. erefore, we titrated IC dilution
to 1:4. e volume of diluted IC injected was 75 mL for
each extremity, and thus, a total of 38 mL IC was used
for each patient.
Image analysis and interpretation
Acquired data were transferred to a workstation
(Vitrea 5.2.512.6014), which allowed processing of
multiplanar reconstructions, including maximum intensity
projection (MIP), multiplanar reformations (MPR), and
volume rendering (VR). e images were then sent to
a PACS (Innite Healthcare, Seoul, South Korea). All
CTV examinations were arranged such that they were
performed before the patients’ hemodialysis session.
Conventional transverse images, reformatted sagittal,
and coronal reconstructions, MIP, and VR 3D images
were used for interpretation. All images were analyzed
by radiologists on duty. A total of 8 general radiologists
with variable experience, ranging from 1 to 26 years,
interpreted the images.
Statistical analyses
Data were analyzed using R soware, version 3.5.1
and SPSS statistical soware ver. 22 (SPSS, Chicago,
Illinois). Residuals were examined for assumption of
normality and heteroscedasticity. Continuous variables
are expressed as mean ± standard deviation. Categorized
variables are presented as numbers with percentages.
Association between patient’s gender and occurrence
of either stenosis or thrombosis was analyzed using
the Chi-squared test of Independence. Lastly, the eect
of a patient’s age on occurrence of both stenosis and
thrombosis was analyzed using binary logistic regression.
P values less than 0.05 were considered as signicantly
dierent.
RESULTS
Of the 40 patients in the study, two had normal
ndings; thus the incidence rate was 95%.
Table 1 illustrates patient demographics. Almost all
patients had prior vascular access either temporary or
permanent, and only one patient had no prior vascular
access. About 50% of patients had indwelling catheters
while undergoing CTV. Patients’ presenting symptoms
are listed in Table 2. Most of the patients (27/40; 67.5%)
presented with upper limb swelling.
Supercial and deep veins in both arms, and the
central veins and internal jugular veins were evaluated and
categorized into segments; as basilic vein, cephalic vein,
brachial vein, axillary vein, subclavian vein, brachiocephalic
vein, internal jugular vein, and SVC, representing a total
of 15 segments per study. is study analyzed data from
600 venous segments. e number of lesions found in
one patient ranged from 1-6, but a majority of patients
had 1-3 lesions (30/38; 78.9%) (Fig 2). Data on sites
and types of lesions are provided in Table 3. Stenosis
(Fig 3) and thrombosis (Fig 4) were the two most common
ndings, were equally prevalent, and were seen in 112 of
the 600 segments (18.7%). e three most common sites
of steno-occlusive complications were the brachiocephalic
vein (29 lesions), the internal jugular vein (25 lesions),
and the subclavian vein (16 lesions). e most common
site of stenosis was the brachiocephalic vein (18 lesions)
whereas the most common site of thrombosis was the
internal jugular vein (20 lesions). We did not encounter
any venous aneurysms or ruptures. Extravasation of IC
at the injection site occurred only once and in one arm
of one patient.
Patient gender was not signicantly related to the
occurrence of either stenosis (χ
2
(1) = 0.003, p = 1.00) or
thrombosis (χ
2
(1) = 2.308, p = 0.170). Further, patient
age did not signicantly predict the likelihood of stenosis
(Binary logistic regression, χ
2
(27) = 15.984, p = 0.9533) or
thrombosis (Binary logistic regression, χ
2
(27) = 32.763,
p = 0.205)
DISCUSSION
CTV is a highly accurate, minimally invasive
method for diagnosing steno-occlusive disease of the
upper extremity veins.
3,10
However, as false positives for
occlusion can occur upon external compression of the
patient’s body, the patient’s position on the table is critical
during the scan.
4
Additionally, indirect or direct CTV
technique, and single site or double site injections have
been used. e indirect CT technique uses 100-150 mL of
undiluted IC injected at the unaected arm at a ow rate of
Volume 73, No.6: 2021 Siriraj Medical Journal
https://he02.tci-thaijo.org/index.php/sirirajmedj/index
376
TABLE 1. Patient demographics.
No. (%)
Gender
Female 20 (52.6)
Male 18 (47.4)
Types of prior vascular access
Arteriovenousstula 17(42.5)
Arteriovenous graft 13 (32.5)
Temporary vascular access 9 (22.5)
None 1 (2.5)
Presence of indwelling catheter 21 (52.5)
while undergoing CT scan
TABLE 2. Patients’ presenting symptoms.
Presenting symptoms No. (%)
Upper limb swelling 27 (67.5)
Failed cannulation 5 (12.5)
High venous pressure* 3 (7.5)
Clot derived from catheter 2 (5)
Decreasedstulaow** 1(2.5)
Asymptomatic: follow up 2 (5)
* High venous pressure = A venous segment static pressure greater
than 0.5 in gras or stulae
9
* Decreased stula ow = An access ow < 600mL/min in gras and
< 500 mL/min in stulae
9
Fig 2. A 85 year-old female with failed le radiocephalic arterio-
venous stula, and indwelling catheter at right internal jugular vein,
presents with right upper limb and face swelling. Direct computed
tomography venography with maximum intensity projection
reconstruction, coronal view reveals stenosis at right subclavian vein
(top-arrow), and thrombosis in right internal jugular vein (bottom-
arrowhead), and thrombosis in superior vena cava (bottom-arrow)
TABLE 3. Sites and types of lesions.
Site Stenosis (No.) Thrombosis (No.) Total
Superior vena cava 3 4 7
Internal jugular vein 5 20 25
Brachiocephalic vein 18 11 29
Subclavian vein 9 7 16
Axillary vein 5 4 9
Brachial vein 3 2 5
Cephalic vein 10 5 15
Basilic vein 3 3 6
Total 56 56 112
Limchareon et al.
Volume 73, No.6: 2021 Siriraj Medical Journal
https://he02.tci-thaijo.org/index.php/sirirajmedj/index
377
Original Article
SMJ
3-5 mL/min, along with scanning in the arterial phase
using a timing bolus test or bolus tracking.
4-5
ere have
been few reports of direct CTV for upper extremities and
the techniques used are variable. A prospective study of
22 patients used a 1:2 dilution of IC that was injected
simultaneously in both arms, along with scanning using
bolus tracking.
10
e total volume of IC used per patient
in that study was 75 mL. Svensson and colleagues used
45 mL of 1:10 diluted IC, and scanning at 12 seconds;
thus, less than 5 mL of IC was used in their study.
11
We
used 1:4 diluted IC, which is similar to that used by
Kuo et al. who injected IC into both femoral veins.
12
Total volume of of IC used in our patients was 38 mL,
and we simultaneously injected IC into both arms to
avoid inow phenomenon at the superior vena cava
(SVC), which is a common problem with single arm
injection. Our protocol yielded good quality images of
bilateral upper extremity veins that were also suitable
for mapping. In cases of chronic occlusion, collateral
veins were also easily visible (Fig 5). Tanju et al. have
used direct contrast-enhanced 3D magnetic resonance
venography (MRV) with bilateral injections, and have
reported 100% sensitivity and specicity. However, there
were only 19 patients in that study.
2
Multiple contrast agents can be used for venography
other than IC such as carbon monoxide
11,13
or gadoterate
meglumine
12,14
which provide high quality images but
are not feasible for use because of their availability.
Non-contrast enhanced MRV has been proposed in
patients who want cannot endure risk from either the
contrast agent or radiation
13,15
, even though contrast-
enhanced MRV has a higher diagnostic odds ratio than
non-contrast techniques.
14,16
e incidence of steno-occlusive complications in
our cohort was 95%, which is higher than that reported
by previous studies.
15-19
is may be because almost all
our patients were symptomatic and had undergone prior
subclavian catheterization or placement of peripherally
inserted long-term catheters for hemodialysis before
stula formation. Prior catheter use is associated with
primary patency loss in patients with a stula or a gra,
and this is the strongest independent predictor of upper
extremity venous thrombosis.
18-21
Moreover, most of
our patients had an indwelling catheter, which is also
the strongest independent predictor of upper extremity
deep venous thrombosis.
20,22
Additionally, average patient
age in the current study was more than 65 years, and
available literature suggests that age more than 50 years
is a signicant risk factor for AVF thrombosis.
21,23
Fig 3. Direct computed tomography 3D volume rendered image of
a 81 year-old female shows right axillary vein stenosis (arrow),
measured 1.9 mm. in diameter, compared with 9.9 mm. of normal
diameter, about 1.0 cm. long. Much collateral veins are noted at le
upper limb.
Fig 4. A 82 year-old male presents with failed cannulation at right
internal jugular vein. Direct computed tomography venography with
multiplanar reconstruction longitudinal view (top) and axial view
(bottom) demonstrate thrombus in the superior vena cava (arrows).
Volume 73, No.6: 2021 Siriraj Medical Journal
https://he02.tci-thaijo.org/index.php/sirirajmedj/index
378
Fig 5. Direct computed tomography venography 3D Volume Rendered
image of a 52 year-old male demonstrates le brachiocephalic vein
thrombosis (arrow). Much collateral veins are noted at le hemithorax.
Limitations
e limitations of this study are a selection bias due
to the lack of a control group, because it is not possible
to perform CTV in a healthy individual. erefore, all
participants were symptomatic, which led to a high
incidence rate. Further, information on several vascular
access characteristics such as anatomic location, and
time-to complication development was not available.
Another limitation was the potential for variability in image
interpretation because of the variation in radiologists’
experience. Lastly, the study did not investigate the eects
of radiation dose which is a topic of concern. However,
as the conversion factor in this region is small, radiation
exposure is expected to be minimal
3,24
, and radiation
hazard in this cohort would be less of a concern given
the higher mean age of the patients.
25
CONCLUSION
We describe a direct CTV technique used in our
facility, which involved simultaneous injection of diluted
contrast medium in both upper extremities. is technique
not only has the benet of requiring a lower volume of
contrast medium while maintaining direct visualization of
the venous system similar to conventional venography, but
also avoids inow phenomenon at the SVC. Additionally,
mapping for venous reconstruction of both arms was
possible based on data acquired. e notable ndings
in this study are benecial for future study. Further
comparative cross-sectional study or RCT in a larger
population to demonstrate the benet of this novel
technique over the conventional method is required.
ACKNOWLEDGMENTS
e authors would like to thank Faculty of Medicine,
Burapha University for all support, and Mr.Chedhawat
Chokechaipaisarn, M.Sc., for statistical analysis, and
Enago (www.enago.com) for the English language review.
Conict of interest: All authors have to conict of interest
to disclose.
REFERENCES
1. Cansu A, Soyturk M, Ozturk MH, Kul S, Pulathan Z, Dinc H,
et al. Diagnostic value of color Doppler ultrasonography and
MDCT angiography in complications of hemodialysis stulas
and gras. Eur J Radiol 2013;82:1436-43.
2. Tanju S, Sancak T, Dusunceli E, Yagmurlu B, Erden I, Sanlidilek
U, et al. Direct contrast-enhanced 3D MR venography evaluation
of upper extremity deep venous system. Diag Interv Radiol
2006;12:74-79.
3. Li B, Li Q, Chen C, Guan Y, Liu S. Diagnostic accuracy of computed
tomography angiography and magnetic resonance angiography
in the stenosis detection of autologous hemodialysis access: a
meta-analysis. PLoS One 2013;8:e78409.
4. Karadeli E, Tarhan NC, Ulu EM, Tutar NU, Basaran O, Coskun
M, et al. Evaluation of failing hemodialysis fistulas with
multidetector CT angiography: comparison of dierent 3D
planes. Eur J Radiol 2009;69:184-92.
5. Cavagna E, D’Andrea P, Schiavon F, Tarroni G. Failing hemodialysis
arteriovenous stula and percutaneous treatment: imaging with
CT, MRI and digital subtraction angiography. Cardiovasc
Intervent Radiol 2000;23:262-5.
6. Ghaye B, Dondelinger RF. Non-traumatic thoracic emergencies:
CT venography in an integrated diagnostic strategy of acute
pulmonary embolism and venous thrombosis. Eur Radiol
2002;12:1906-21.
7. Zhou PL, Wu G, Han XW, Bi YH, Zhang WG, Wu ZY. Detection
and characterization of Budd-Chiari syndrome with inferior
vena cava obstruction: comparison of xed and exible delayed
scan time of computed venography. Eur J Radiol 2017;91:52-
56.
8. Mavili E, Ozturk M, Akcali Y, Donmez H, Yikilmaz A, Tokmak
TT, et al. Direct CT venography for evaluation of the lower
extremity venous anomalies of Klippel-Trenaunay syndrome.
AJR Am J Roentgenol 2009;192:W311-6.
9. NKF-K/DOQI Clinical Practice Guidelines for Vascular Access:
update 2006. Am J Kidney Dis 2006;48(Suppl):S176-247.
10. Kassem TW. Upper limb DVT aer hemodialysis AVF creation:
role of CT venography. Egyptian J Radiol Nuclear Med 2016;
47:897-902.
11. Svensson A, Brismer TB, Brehmer K. Computed tomography
venography of the upper extremities-using low dose bilateral
contrast media injection in a patient with suspected venous
thoracic outlet syndrome. Radiol Case Rep 2020;15:302-5.
12. Kuo YS, Chen CJ, Chen JJ, Lin CH, Ku JW, Hsu HL, et al.
Mey-urner syndrome: correlation between digital subtraction
and computed tomography venography. J Formos Med Assoc
2015;114:363-8.
Limchareon et al.
Volume 73, No.6: 2021 Siriraj Medical Journal
https://he02.tci-thaijo.org/index.php/sirirajmedj/index
379
Original Article
SMJ
13. Sullivan KL, Bonn J, Shapiro MJ, Gardiner GA. Venography
with carbon dioxide as contrast agent. Cardiovas Intervent
Radiol 1995;18:141-5.
14. Georoy O, Tassart M, Le Blanche AF, Khanlil A, Duedal Z,
Rossert J, et al. Upper extremity digital subtraction venography
with gadoterate meglumine before stula creation for hemodialysis.
Kidney Int 2001;59:1491-7.
15. Yamashita R, Isoda H, Arizono S, Ono A, Onishi N, Furuta A,
et al. Non contrast enhanced magnetic resonance venography
using magnetization-prepared rapid gradient-echo (MPRAGE)
in the preoperative evaluation of living donor candidates:
comparison with conventional computed tomography venography.
Eur J Radiol 2017;90:89-96.
16. Gao L, Xu W, Li T, Yu X, Cao S, Xu H, et al. Accuracy of magnetic
resonance venography in diagnosing cerebral venous sinus
thrombosis. romb Res 2018;167:64-73.
17. Asif A, Roy-Chaudhury P, Beathard GA. Early arteriovenous
stula failure: a logical proposal for when and how to intervene.
Clin J Am Soc Nephrol 2006;1:332-9.
18. Dumars MC, ompson WE, Bluth EI, Lindberg JS, Yoselevitz M,
Merritt CR. Management of suspected hemodialysis gra
dysfunction: usefulness of diagnostic US. Radiology 2002;222:
103-7.
19. Rooijens PP, Tordor JH, Stijnen T, Burgmans JP, Smet de AA,
Yo TI. Radiocephalic wrist arteriovenous stula for hemodialysis:
meta-analysis indicates a high primary failure rate. Eur J Vasc
Endovasc Surg 2004;28:583-9.
20. Padberg Jr FT, Calligaro KD, Sidawy AN, Sidawy S. Complications
of arteriovenous hemodialysis access: recognition and management.
J Vasc Surg 2008;48:55S-80S.
21. Ocak G, Rotmans JI, Vossen CY, Rosendaal FR, Krediet RT,
Boeschoten EW, et al. Type of arteriovenous vascular access and
association with patency and mortality. BMC Nephrol 2013;14:79.
22. Margey R, Schainfeld RM. Upper extremity deep vein thrombosis:
the o-forgotten cousin of venous thromboembolic disease.
Curr Treat Options Cardiovasc Med2011;13:146-58.
23. Jamshid R, Reza SA, Abbas G, Raha A. Incidence of arteriovenous
thrombosis and the role of anticardiolipin antibodies in
hemodialysis patients. Int Urol Nephrol 2003;35:275-82.
24. Murphy EA, Ross RA, Jones RG, Gandy SJ, Aristokleous N,
Salsano M, et al. Imaging in vascular access. Cardiovasc Eng
Technol 2017;8:255-72.
25. Preston DL, Ron E, Tokuoka S, Nishi FN, Soda M, Mabuchi K,
et al. Solid cancer incidence in atomic bomb survivors: 1958-
1998. Radiat Res 2007;168:1-64.
