European Journal of Radiology 31 (1999) 182 – 187

www.elsevier.nl/locate/ejrad

CO2 as a contrast medium in endoluminal treatment of high flow

vascular malformations
Elı́as Górriz *, José M. Carreira, Ricardo Reyes, Laura Gallardo, Juan M. Pulido,
Alejandro Romero, Manuel Maynar
Vascular and Inter6entional Radiology Unit, Hospital Uni6ersitario Nuestra Señora del Pino, Las Palmas de Gran Canaria Uni6ersity,
C/ Angel Guimerá, 93, 35005 Las Palmas, GC, Spain
Received 21 October 1997; received in revised form 13 May 1998; accepted 14 May 1998

Abstract

Purpose: To evaluate the diagnostic usefulness of CO2 as a radiologic contrast medium in pre and post-embolization of
hemodynamically active vascular malformations. Material and methods: Eight pre and post-embolization angiographies were
performed on five patients with hemodynamically active vascular malformations. In all cases sequentially iodinated contrast
medium and CO2 were used. Procedures were performed using Philips Integris V equipped with specific software to process images
obtained with CO2 as contrast medium and with a purpose-built gas injector pump. The images were assessed by three
interventional radiologists who evaluated the findings by consensus. They were asked whether uptake of the malformations were
better, worse or the same with iodinated contrast and CO2. The images were judged according to the following characteristics;
quality, the filling in of the contrast in afferent artery of the HAVM, caliber and number of vessels and existence of pathologic
communications in HAVM. These characteristics were assessed in all the patients of the study, both before and after embolization.
Results: The CO2 angiographic results were compared to those obtained using iodinated contrast material. Iodinated contrast
provided superior image quality in all performed studies. The filling of arterial afference of HAVM was well defined with both
contrast media. In all performed cases with CO2, the malformations showed more vessels. No inmediate nor subsequent
complications developed with CO2 injections, except in one female patient, who presented an intense sensation of cramps in the
lower limbs. Conclusion: CO2 is a useful diagnostic and assessment tool before and especially during the embolization of
hemodynamically active vascular malformations. It improves quantification and uptake of the malformation’s vascular architec-
ture, detecting collateral circulation and arteriovenous shunts. It also detects residual post-embolization disease when iodinated
contrast agent is unsuccessful. © 1999 Elsevier Science Ireland Ltd. All rights reserved.

Keywords: Vascular malformation; Angiography; Carbon dioxide; Iodinated contrast agent; Embolization

1. Introduction and severity of the malformation may vary, ranging

Vascular malformations (VM) are congenital lesions,
which occur as a result of abnormal vascular develop-
ment during the embryonic period. They appear in the
neonatal and early years. Trauma, surgery and hor-
monal changes may cause rapid growth of the lesion
[1].
Mulliken classified vascular malformations in new-
born hemangiomas and VM [2 – 4]. VM are classified as
high and low-flow VM [3]. The clinical presentation
* Corresponding author. Tel: +34 28 441092; fax: 34 28 441097.
E-mail address: vascular@ccdis.dis.ulpgc.es. (E.Górriz)
from asymptomatic macules to congestive heart insuffi-
ciency episodes.
Doppler ecography, magnetic resonance and angiog-
raphy provide diagnostic assessment [5].
In most cases, surgical treatment is impossible or
ineffective [6]. Percutaneous embolization techniques
allow to obliterate the pathologic vessels, either by
superselective catheterization of the affected afferent
arteries or by direct puncture. Different sclerosant sub-
stances are used depending on the type and morphol-
ogy of the VM [3].
Conventional angiographic methods use iodinated
liquids as a radiologic contrast agent. Using a gas for a

0720-048X/99/$ - see front matter © 1999 Elsevier Science Ireland Ltd. All rights reserved.
PII: S 0 7 2 0 - 0 4 8 X ( 9 8 ) 0 0 0 9 4 - 1
 E. Górriz et al. / European Journal of Radiology 31 (1999) 182–187 183

radiological contrast media has been a recent innova-
tion [7]. It requires multiple features due to its physical
and chemical characteristics. Specially designed equip-
ment is required. To asses the possible benefits of CO2
radiologic contrast agent a comparative study was car-
ried out betweeen both conventional contrast medium
and CO2 in patients with hemodynamically-active vas-
cular malformations (HAVM).
2. Material and methods
From January to June 1997, eight angiographic stud-
ies were performed using CO2 and iodinated contrast
agent pre and post-embolization on five patients pre-
senting with HAVM. Four women and one man be-
tween 12 and 36 years (mean age 28 years). The
malformation’s were localized in the knee (n = 2), foot
(n= 1), thigh (n= 1) and gluteal region (n =1). In all
the patients, percutaneous treatment followed episodes
of pain and functional impotence in the anatomic areas
affected by the VM.
The approachs was femoral anterograde in seven
patients and femoral retrograde contralateral in one
case. In two patients, the procedure was performed
using general anesthetic due to their young age whereas
the rest were administered local anesthetic.
In the angiographic assessment prior to treatment a 5
F pig-tail catheter was used (Cordis, Miami, Fl), a 5 F
multipurpose catheter (Cordis, Miami, Fl) and a 0.035’’
hydrophilic guidewire (Terumo, Tokyo, Japan). The
superselective studies were performed using a 3 F
Tracker catheter (Target Therapeutics, Fremont, CA)
and a 0.014’’ platinum guidewire (Advanced Cardiovas-
cular Systems, Temecula, CA).
The angiographic study showed the presence of arte-
riolo-venular HAVM, in two cases and arterio-venous
HAVM in three patients.

Fig. 1. (A) Angiography with conventional contrast material of HAVM prior to embolization. (B) Angiography of the same arterial section made
with CO2 as radiologic contrast agent. More arteriovenous communications are seen in comparation with the iodinated contrast agent study. Early
venous filling is noted. (C) Late phase of the study with CO2 in which the venous efferent of the HAVM is observed.
184 E. Górriz et al. / European Journal of Radiology 31 (1999) 182–187
Fig. 1. (Continued)
Embolization was made with either ethanol on its
own or mixed with Gelfoam particles (Spongostan
Film, Ferrosan, Söborg, Denmark) and/or Gianturco
spirals (Cook, Bjaeverskov, Denmark). To assess the
malformations CO2 was used sequentially with a non-
ionic iodinated contrast agent (Imagopaque, Nycomed,
Madrid, Spain) before and after embolization, selec-
tively injecting 8 cc of iodinated contrast manually and
60 cc of CO2 with 20 cc/sg by CO2 injector (Angiody-
namics. CO2 ject., Model CD 1000, NY).
Procedures were performed using Philips Integris V
equipped with specific software to process images ob-
tained with CO2.
The images were assessed by three interventional
radiologists who evaluated the findings by consensus.
They were asked whether the malformation had better,
worse or the same uptake with iodinated contrast or
with CO2. The characteristics seen in the images were:
quality obtained, the filling in of the contrast in afferent
artery of the HAVM, caliber and number of vessels and
existence of pathologic communications in HAVM.
These characteristics were assessed in all the patients
studied, both before and after the embolization.
CO2 angiographies were of sufficient quality to be
compared with those obtained using iodinated contrast.
The use of small caliber catheters for the superselective
angiographies and embolizations did not hinder the
administration of CO2 as contrast medium.
3. Results
The angiographic results obtained using CO2 as a
contrast medium and those using an iodinated contrast
agent were compared. Image quality was superior in all
the studies performed with iodinated contrast media.
The filling of arterial afference of the HAVM was well
detected with both contrast media. The vascular mal-
 E. Górriz et al. / European Journal of Radiology 31 (1999) 182–187 185

formation showed more vessels in all the cases per- 4. Discussion

formed with CO2 (Fig. 1A, B, C). The filling of the
malformed vessels with CO2 was crucial in the post-em-
bolization angiographic study. Residual arterio-venous
communications in the performed angiographies with
CO2 were detected which went undetected in the per-
formed angiographies using iodinated contrast agent
(Fig. 2A, B, C).
In one patient, the angiographic study with CO2 as
contrast medium showed the flow direction mainly
from the origin of the afferent artery toward the nidus
of HAVM (Fig. 1B, C). This was not detected with the
iodinated contrast image (Fig. 1A). Once embolized,
flow was reestablished in its anatomic direction (Fig.
2C).
No complications developed immediately after the
procedure nor subsequently related to the injection of
CO2, except in one female patient, who presented in-
tense cramps in the lower limbs.
Carbon dioxide was first administered intravenously
in 1950 to detect pericardial effusion [8–10]. Ten years
later it was used to evaluate the portal and hepatic
venous system [11,12].
In 1970, Hawkins pioneered the use of CO2 in vis-
ceral and peripheral arterial circulation [7]. Today an-
giographies with CO2 as a contrast medium have
extended to diagnostic procedures as well as therapeutic
purposes [13–15].
HAVM are characterized by the presence of vascular
nidus with hypertrophic afferent arteries and dilated
and tortuous efferent veins.
Traditionally, angiographic studies of HAVM have
been performed using an iodinated contrast agent, with
possible use of CO2 to note the differences in the
angiographic evaluation of HAVM [16,17]. This may be
attributed to the different physical and chemical char-

Fig. 2. (A) Conventional contrast medium post-embolization angiography in a late phase. The microcoils used are seen in the embolization. (B)
Late phase with iodinated contrast agent in which vascular malformation is not seen. (C) The same arterial section angiography taken with CO2
as a radiologic contrast agent. Residual vascular malformation is detected.
186 E. Górriz et al. / European Journal of Radiology 31 (1999) 182–187
Fig. 2. (Continued)
acteristics between the two techniques. The iodinated
contrast media, which are injected intravascularly, mix
with the blood thus increasing the radiologic density
depending on the flow and volume injected, as well as
the level of dilution. CO2, acts by displacing the blood
column, maintaining a gas/liquid interface and causing
a decrease in the radiologic density through which the
image is obtained with digital subtraction and specially
designed programs for CO2 [16].
Different types of studies of malignant tumors and in
the cases of fistulas and arteriovenous malformations,
using CO2 as contrast medium, has detected arteri-
ovenous shunts which were not evident with iodinated
contrast agents. This is mainly due to the two charac-
teristics of CO2: its buoyancy and low viscosity [14–17].
If the vascular area of study is more elevated than the
injection site, the CO2 fills the entire structure, is re-
tained and remains there for a longer period of time
[18]. The low viscosity of carbon dioxide (400 times less
that of iodinated contrast) allows it to spread through
the vascular tree and highlight arteriovenous communi-
cations and collateral vessels not visualized in studies
with iodinated contrast media [17,19].
A crucial aspect of HAVM embolization strategy is
gaining detailed knowledge of the vascular architecture,
the caliber and number of vessels which conform it as
well as the rapid rate of flow between the afferent and
efferent arteries. Thus, superselective injections of con-
trast to design the strategy and choosing the material to
use in each case are performed prior to embolization.
In all our patients both before and after embolization
more information with CO2 than with iodinated con-
trast media were obtained, showing arteriovenous com-
munications which were not visible with the iodinated
contrast media. During the procedure in angiographic
control this is especially important as it ensures greater
accuracy at the moment that the malformation has
been excluded.
One notable difference observed between performed
angiographies with iodinated contrast media and CO2 is
the detection of the flow tendency of one afferent artery
toward the ‘nidus’. This was not seen with the contrast
medium and may be useful in determining which thera-
peutic decision to take, even when a clinical situation of
arterial deprivation has not manifest itself.
 E. Górriz et al. / European Journal of Radiology 31 (1999) 182–187
As described in this paper, CO2’s many advantages as
a contrast medium make it a useful complementary tool
for investigating the hemodynamics of a malformation,
in spite of the fact that angiographic control with
iodinated contrast still is required to learn about the
state of a HAVM with conventional methods. Subse-
quently, CO2 provides the real size of the area to be
embolized. After embolization we may ascertain the
real dimensions of the excluded HAVM as well as the
residual area.
CO2 is useful in diagnostic and assessing before and
especially during the embolization of HAVM. It im-
proves the quantification and detection of the vascular
architecture of the malformation detecting collateral
circulation and arteriovenous shunts. It also detects
residual post-embolization disease when the iodinated
contrast agent is unsuccessful.
References
[1] Enjolras O. Les angiomes de l’enfant. Rev Pediatr 1982;6:168–
361.
[2] Finn HC, Glowacki J, Mulliken JB. Congenital vascular lesions:
clinical application of a new classification. J Pediatr Surg
1983;18:894 – 900.
[3] Enjorlas O, Herbreteau D, Lemerchand F, Riche MC, Laurian
C, Brette MD, Merland JJ. Hémangiomes et malformations
vasculaires superficielles: classification. J Mal Vasc 1992;17:2–19.
[4] Mulliken JB, Glowacki J. Hemangiomas and vascular malforma-
tions in infants and children: a classification based on endothelial
characteristics. Plast Reconstr Surg 1982;69:412–20.
[5] Yakes WF. Diagnosis and management of vascular anomalies.
In: Castanñeda WR, Tadavarthy SM, Qian Z, Ferral H, Maynar
M, editors. Interventional Radiology, 3rd ed. Baltimore:
Williams and Wilkins, 1997:103–38.
[6] Laurian C. Malformations vasculaires des membres. Place de la
chirurgie vasculaire et réparatrice. J Mal Vasc 1992;17:57–60.
187
[7] Hawkins IF. Carbon dioxide digital substraction arteriography.
AJR 1982;139:19 – 24.
[8] Stauffer HM, Durant TM, Oppenheimer MJ. Gas embolism:
roentgenologic considerations including the experimental use of
carbon dioxide as an intracardiac contrast material. Radiology
1956;66:686 – 92.
[9] Paul RE, Durant TM, Oppenheimer MJ, Stauffer HM. Intra-
venous carbon dioxide for intracardiac gas contrast in the roent-
gen diagnosis of pericardial effusion and thickening. AJR
1957;78:224 – 5.
[10] Scatliff JH, Kummer AJ, Janzen AH. The diagnosis of pericar-
dial effusion with intracardiac carbon dioxide. Radiology
1956;73:871 – 82.
[11] Goldenberg DB, Brogdon BG. Hepatic venography with carbon
dioxide. MD State Med J 1967;17:79 – 80.
[12] Hipona FA, Park WM. Capnosplenoportography: assesment of
portal vein patency in dogs with carbon dioxide gas. AJR
1967;99:606 – 11.
[13] Seeger JM, Self S, Harward TR, Flynn TC, Hawkins IF. Carbon
dioxide gas as an arterial contrast agent. Ann Surg
1993;217:688 – 98.
[14] Carreira JM, Reyes R, Górriz E, Pardo MD, Maynar M. CO2 as
contrast agent in lower limbs angiography. Our early experience.
Angiologı́a 1996;2:55 – 60.
[15] Ehrman KO, Taber TE, Gaylord GM, Brown PB, Hage JP.
Comparison of diagnostic accuracy with carbon dioxide versus
iodinated contrast material in the imaging of hemodialisis access
fistulas. J Vasc Interv Radiol 1994;5:771 – 5.
[16] Kerns SR, Hawkins IF. Carbon dioxide digital subtraction an-
giography: expanding applications and technical evolution. AJR
1995;164:735 – 41.
[17] Takeda T, Iko K, Nishimura G, Hasimoto S, Kyo E, Okawa S,
et al. Intraarterial digital subtraction angiography with car-
bondioxide: superior detectability of arteriovenous shunting.
Cardiovasc Interv Radiol 1988;11:101 – 8.
[18] Kerns SR, Hawkins IF, Sabatelli FW. Current status of carbon
dioxide angiography. In: Katzen BT, editor. Radiologic Clinics
of North America. Philadelphia: Saunders, 1995;33:15 – 29.
[19] Harward TRS, Smith S, Hawkins IF, Seeger JM. Follow-up
evaluation after renal artery bypass surgery with use of carbon
dioxide arteriography and color-flow duplex scanning. J Vasc
Surg 1993;18:23 – 30.