We present a case that documents a minimally invasive endovascular repair of a very rare iatrogenic AVF between the internal iliac artery and the internal iliac vein.
Etiology and Demographics
Despite the low incidence rates of iliac AVF, the etiology of this vascular pathology (Table
1) may have several cardinal causes, which may be categorized as spontaneous and traumatic. Brewster et al. [
1] investigated aortocaval and iliac AVF over a 30-year observation period. Their report of 20 abdominal AVFs revealed that the primary etiology was aneurysm erosion (70%), followed by iatrogenic causes (20%) and gunshot wounds (10%). Lumbar disk surgery was reported to be the most prominent iatrogenic cause of iliac AVF [
2,
3]. However, Wang and coworkers [
2] also pointed out that besides aneurysm formation (mycotic, syphilitic), neoplastic erosion affecting surrounding vasculatures and connective tissue pathologies (e.g., Marfan syndrome) may also be clinically significant causes of iliac AVF. Trauma may not always be related to penetrating injuries (gunshot and stab wounds); it may also occur after seat-belt trauma [
4] and laparoscopic appendectomy [
5].
Table 1
Summary table for iliac arteriovenous fistulas
Definition | Iliac arteriovenous fistulas are abnormal connections between the iliac arteries and the corresponding iliac veins (common, internal, external) |
Etiology | Strongly associated with trauma (shot and stab wounds) or iatrogenic interventions (lumbar disk surgery), they may also be related to aneurysm formation (mycotic, syphilitic), neoplastic erosion affecting surrounding vasculatures, or connective tissue pathologies (e.g., Marfan syndrome) |
Incidence | Occur in less than 1% of all common iliac artery aneurysm cases and in (most likely) < 1% of all lumbar disk surgeries; strongly related to lower quadrant “piercing” traumatic injuries with other concomitant symptoms (signs of congestive heart failure, dyspnea, back pain, lower extremity edema, abdominal bruit) |
Gender ratio | Not gender specific |
Age predilection | Not related to age |
Risk factors | Preceding lumbar disk surgery and, in particular, repeat lower back surgeries and traumatic injuries to the lower left and right quadrants |
Treatment | Vascular reconstruction (open surgery) and percutaneous interventions (covered self-expanding stents) |
Prognosis | Early detection: expected mortality < 10% and alleviation of symptoms; in the case of preceding ruptured aneurysm: mortality > 60% |
Findings on imaging | Contrast-enhanced MR: easily discernible CT and US: clearly communicating flow across the fistula |
Clinical and Imaging Findings
Depending on the etiology, the size, and the anatomic location of the iliac AVF, symptoms may be quite nonspecific, consisting of back pain [
1], dyspnea commonly observed as pulmonary edema [
2], or unilateral leg edema [
1,
5,
6]. Due to the dramatically compromised hemodynamic environment, specific cardiovascular abnormalities may aid the decision regarding additional imaging diagnostics. The presence of a pulsatile abdominal mass with concomitant vascular bruit and thrill [
6] has frequently been reported. Furthermore, intermittent claudication [
2,
7], congestive heart failure [
8,
9], pulmonary hypertension [
3], T-wave abnormalities and renal failure [
3,
10], and left ventricular hypertrophy [
11] are described in the literature. Imaging findings based on exploratory ultrasound (US) are able to detect iliac AVFs with sufficient arterial leak flow rates into the venous system.
The cross-sectional area of the AVF in this case report was estimated at 2.2 cm
2 (
Aoval = ¼π ×
daxial ×
dlongitudinal = ¼π × 1.9 cm × 1.5 cm). Unfortunately, there is a paucity of data on actual cross-sectional areas of AVFs. However, Cronin et al. [
12] reported a diameter of the left IIA of 24 mm, which corresponds well to the pathoanatomy in our case. They used a thoracic endograft that was longer (10 vs. 6 cm) and had a larger diameter (28 mm vs. 9/14 mm). We were able to use a shorter stent graft, since the iliac trunk was quite short in our patient. Ramacciotti and coworkers [
13] conducted flow experiments in a canine AVF model and found that the flow through the fistula increases by a factor of 5 if the diameter of the AVF is 50% larger than the arterial diameter. This scenario from preclinical work appears to be similar to our case, without having measured actual flow rates.
Differential Diagnoses
All imaging modalities (Table
2) will show the focal widening caused by either a pseudoaneurysm or an aneurysm. These can be clearly differentiated from an AVF, which communicates between an artery and a vein. With duplex ultrasound, a pseudoaneurysm can have a characteristic yin–yang sign, which indicates the presence of bidirectional flow due to swirling of the blood within the lumen of the lesion. This swirling flow can also be seen in conventional angiograms. In contrast-enhanced CT angiography and MR angiography, this swirling flow can be detected from differences in the degree and shape of opacification of the contrast media within the lesion and/or in the vessels adjacent to the lesion. On duplex ultrasound, a true aneurysm of fusiform shape may show turbulent, disorganized flow, whereas a saccular aneurysm may show the yin–yang sign of swirling flow.
Table 2
Differential diagnosis table for iliac arteriovenous fistulas
Iliac arteriovenous fistula | Detectable arteriovenous flow patterns Arterialized venous flow is visible; pulsatile flow | Dilated inferior vena cava and/or iliac arteries originating from the aneurysmal fistulous tract (detectable even for a small-sized AVF) | Findings similar to CT, especially in gadolinium-enhanced MR: dilated inferior vena cava and/or iliac arteries originating from the aneurysmal fistulous tract | Findings similar to CT, even with small leak flows of << 0.5 l/min |
Pseudoaneurysm | No arteriovenous communication, but large fluid accumulation that can have a characteristic yin–yang sign, which indicates the presence of bidirectional flow due to swirling of the blood within the lumen of the lesion | Contrast-enhanced CT scan with axial images reveal large contrast-filled spaces around the iliac artery | Jet of bleeding from the pseudoaneurysm; outpouching of the vessel wall; focal widening edema; gas (without signs of infection) | Irregular lumen Focal widening No communication with the vein; the pseudoaneurysm may have irregular jagged margins; the angles between the pseudoaneurysm and the vessel lumen may be more acute; a neck connecting the pseudoaneurysm to the vessel wall may be seen |
Aneurysm | Opacification of the extended vessel and a fusiform shape may show turbulent, disorganized flow, whereas a saccular aneurysm may show the yin-yang sign of swirling flow | Outpouching of the lumen of the vessel, usually no venous opacification | Focal widening; time-resolved flow; surrounding tissue does not have bleeding; gas w/o infection | Regular focal widening; no communication with the vein; may have smoother margins; the angles between the aneurysm and the vessel lumen are more obtuse; usually there is no neck connecting the aneurysm to the vessel wall |
For the sake of discussion (Table
2), it can be stated that in CT or MRI, it may be difficult to differentiate a pseudoaneurysm from a true aneurysm. In a true aneurysm, there are usually no edema, blood products, and/or air in the tissues surrounding the aneurysm, unless the aneurysm has ruptured or is infected. If the pseudoaneurysm is chronic, the edema, blood products, and/or air in the tissues surrounding the pseudoaneurysm have usually dissolved. Consequently, exploratory ultrasound imaging is able to detect an iliac AVF with a sufficient arterial leak flow into the venous system [
6], which may be in the range of 3 l/min [
5]. Gadolinium-enhanced magnetic resonance (MR), computed tomography (CT), and angiography [
7] are all suitable to detect even small-sized iliac AVFs with certainty. As concluded by Iijima et al. [
6], any combination of imaging diagnostics are useful for detecting iliac AVF and deciding on the treatment modality.
Treatment and Prognosis
Surgical repair in these scenarios is associated with high morbidity and mortality due to the complex anatomy and the unfavorable access site [
6]. Taking into consideration post-laparotomy abdominal adhesions, surgical access to the pelvis would be very demanding, and is associated with a high procedural risk. Large incisions and significant blood loss would most likely be inevitable. Furthermore, vessel wall adhesion between the IIA and IIV could have led to vessel wall rupture or wall dissection during surgical repair, which is associated with fatal bleeding.
There is also a significant risk while obtaining vascular access. Perforation of the intestine or bladder could have occurred, which inevitably would have resulted in peritonitis. Such a situation would have made it impossible to use a surgical repair with vascular patches. The endovascular approach, however, requires only a puncture on the left side and a small incision on the right side exposing the right CIA. Since we were not routinely using closure systems, we used a slightly longer incision to create a vessel loop so that bleeding could be better controlled. We are also convinced that a decreased introducer diameter may have facilitated the puncture and may have even avoided an incision. The endovascular approach, however, became a valuable alternative when dealing with patients with an iliac AVF [
11].
The in-hospital outcome of this endovascular approach with a short recovery time is noteworthy. The patient could walk 2 h after the procedure and was discharged after 2 days. A similar immediate improvement using surgical repair was also reported by Machado-Atías [
3]. Thus, despite the potential higher cost of stenting in comparison to open surgery, the very short hospitalization made the procedure cost effective.
It is also of paramount importance to note the fact that the patient is in her procreative years. Surgical access to the pelvis might have led to complications in future pregnancies.
The alternative implantation of the covered stent in the corresponding vein in conjunction with coil implantation at the arterial lesion site was also reported in the literature [
14]. However, we believe that it cannot be recommended due to difficult-to-detect microbleeding. The higher arterial pressure may exert enough forces on the stent and the vessel wall to provoke leakage at the lesion site.
Due to the low incidence rates of iliac AVF (less than 1% [
6]) and the associated pathoanatomy, the manufacture of standard PTFE-covered stents is most likely not attractive for device manufacturers, despite their efficacy in this niche indication. In our case, we implanted a custom-made device whose dimensions were based on precise CT measurements obtained prior to our intervention. Another interesting solution, described by Cronin and coworkers, is the use of preexisting thoracic stents [
12]. Nevertheless, despite the advantages of endovascular techniques, Kuehnle and coworkers [
5] cautioned about the use of alloplastic material in patients who are not fully grown: even in very young patients, the open surgical repair option may still be associated with a higher benefit/risk ratio than endovascular repair with covered stents. Another potential risk factor for iliac AVF formation is the occurrence of thrombotic occlusions in the common iliac vein, as reported by Weyrich and Beck [
15]. This rare complication may also necessitate an open surgical approach.
An important step, and the most technically demanding one, turned out to be the cannulation of the distal IIA. Despite the contralateral approach, the small size of the distal portion of the IIA and the significant blood flow through the fistula made it very cumbersome to cross the lesion with the guidewire and over-the-wire catheter. The other challenge for the operator was to cross the aortic bifurcation with the introducer. Significant device stiffness, the young age of the patient, and the fixation of the aorta due to the adhesions made the vessel much less movable. This increased the risk of potential dissections and/or wall perforations. The puncture of the left CFA, through which we introduced the pigtail catheter, greatly facilitated angiographic imaging. Furthermore, if the stent graft had migrated, the outlet of the external iliac artery would have been blocked. However, with the second guidewire in place, it would have enabled us to implant another “kissing” stent. Finally, both guidewires crossed each other, marking the beginning of the IIA.
It would have been preferable to use MRI imaging modalities to reduce radiation exposure to the patient. However, in our clinical setting, mandatory pregnancy tests permitted angiographic CT images, which were the design inputs for the custom-made device.