Online Exclusive

Successful IVUS-Guided Reentry From Iatrogenic Coronary Arteriovenous Fistula Related to Wire Perforation Following Wiring of a Totally Occluded Vessel

Hidefumi Ohya, MD, Eisho Kyo, MD, Osamu Katoh, MD

Hidefumi Ohya, MD, Eisho Kyo, MD, Osamu Katoh, MD

ABSTRACT: We experienced a rare case in which a guidewire was advanced into a coronary vein through an arteriovenous fistula caused by wire perforation. The patient, who had chronic total occlusion (CTO) of the left circumflex coronary artery, was treated successfully with a procedure guided by intravascular ultrasound (IVUS). The IVUS-guided parallel-wire technique allowed recrossing of the guidewire into the distal true lumen of the CTO by identifying the anatomy of the occluded segment and the appropriate re-entry point. Angiography demonstrated that the fistula was completely sealed after stent deployment, and there was no extravasation.

J INVASIVE CARDIOL 2013;25(7):E139-E142

Key words: intravascular ultrasound, wire perforation, fistula


Percutaneous coronary intervention (PCI) in vessels with chronic total occlusion (CTO) represents the most technically challenging procedure for interventional cardiologists. PCI in vessels with CTO is associated with lower success rates and higher complication rates than PCI in non-CTO vessels. The most common reason for PCI failure is the inability to cross the CTO lesion with a guidewire. Therefore, aggressive wiring using tapered or stiff wires may be necessary to improve the success rates. However, the use of these wires may result in vessel perforation and life-threatening complications such as cardiac tamponade. Despite the increased risk of perforation, the actual in-hospital major complication rate was relatively low in the hands of experienced operators in Japan.1,2 Highly experienced operators have performed intravascular ultrasound (IVUS) routinely during PCI to treat CTO;3 the routine use of IVUS for this procedure may prevent complications. The usefulness of the IVUS-guided parallel-wire technique in the antegrade approach has been commonly accepted for PCI of vessels with CTO in Japan.4

The accumulated evidence suggests that most coronary fistulas are congenital and that iatrogenic coronary fistulas can be induced by cardiac trauma, ventricular biopsy, aortic valve replacement, or coronary angioplasty.5

To the best of our knowledge, this is the first case of iatrogenic coronary arteriovenous fistula between the left circumflex (LCX) artery and the posterior vein of the left ventricle due to wire perforation, which was successfully treated with IVUS-guided PCI of the totally occluded LCX. 

Case Report. An 80-year-old female with ischemic cardiomyopathy was electively admitted to our hospital for PCI of a distal LCX with CTO. She had been suffering from worsening effort angina for 3 months. Her medical history included diabetes mellitus. An electrocardiogram showed normal sinus rhythm with non-specific ST changes in leads V5 and V6. Coronary computed tomography (CT) angiography suggested triple-vessel disease with severe calcification and severely depressed left ventricular systolic function (ejection fraction, 30%).

Subsequent coronary angiography demonstrated advanced triple-vessel disease with the following characteristics: diffuse lesions and subtotal occlusion of the left anterior descending (LAD) artery, 75% stenosis of the right posterolateral branch, and subtotal occlusion of the mid-LCX with distal CTO. During previous PCI, there was successful deployment of 4 drug-eluting stents (two sirolimus-eluting coronary stents in the proximal LAD and mid-LCX, and two zotarolimus-eluting coronary stents in the mid-LAD and right posterolateral branch). Even after successful multiple stenting and optimal medical therapy, she suffered from repeated orthopnea at midnight and dyspnea on exertion. We decided to perform PCI for the distal LCX with CTO, since two-dimensional echocardiography revealed decreased regional wall motion in the distal LCX territory. Diagnostic coronary angiography demonstrated total occlusion of the distal LCX with the presence of a small side branch at the proximal end of the occlusion (Figure 1). The distal vessel was not visualized by contralateral injection, and slightly visualized via the posterolateral channel with ipsilateral injection. Angiography showed that there was no applicable collateral channel for a retrograde approach. PCI was performed using unilateral femoral access with an 8 Fr sheath. The left coronary artery was engaged with an 8 Fr XB 3.0 guide catheter after administration of unfractionated heparin (8000 units). We attempted an antegrade approach using a Rinato wire (Asahi Intec) with the support of a Corsair microcatheter (Asahi Intec) as the primary system. The Rinato wire was exchanged for a Miracle 12 g wire (Asahi Intec) after a few minutes of manipulation. However, the Miracle wire could not penetrate the proximal cap of the CTO. Then, a Conquest Pro wire (Asahi Intec) was advanced with the Corsair as the second wire (parallel-wire technique) to penetrate the proximal cap. The Conquest Pro wire passed across the proximal cap into the distal lumen with no resistance (Figure 2). The movement of the tip of the wire was likely in some part of the lumen, but it was not clear whether the wire was located in an extraluminal position or small branch by angiography. The Corsair was crossed over the wire beyond the occluded segment, and the wire exchanged for a Rinato wire to avoid wire perforation. Subsequent dilatation with a 1.3 x 10 mm Lacrosse LAXA balloon (Goodman) to a maximum of 14 atm was performed to deliver an IVUS catheter. Angiography showed restoration of antegrade flow (Figure 3A). IVUS (Revolution 45 MHz Catheter; Volcano Corporation) revealed that the guidewire had advanced into the subintima from the proximal cap, an arteriovenous fistula had formed distal to the CTO segment, and the tip of the wire was in the coronary vein (Figures 3a-3e).

After placing the IVUS catheter in the coronary vein through the fistula and identifying the optimal reentry point, the Conquest Pro wire was successfully negotiated into the true lumen (Figure 3B). Repeat IVUS confirmed that the wire was placed in the distal true lumen (Figures 4a-4e). Angiography following balloon dilatation showed recanalization of the CTO with simultaneous visualization of the parallel coronary vein and inflow into the coronary sinus (Figure 4).

A 2.5 x 18 mm everolimus-eluting coronary stent (XienceV; Abbott Vascular) was successfully implanted from the distal LCX. Subsequently, a 2.5 x 15 mm Quantum Maverick balloon (Boston Scientific) was positioned inside the stent and inflated to 20 atm. Final angiography demonstrated that the fistula was completely sealed with no extravasation (Figure 5). The patient remained stable throughout the procedure and echocardiography showed no pericardial effusion at the end of the procedure. After 6 months, follow-up angiography demonstrated no in-stent restenosis and no recurrence of fistula. One-year follow-up coronary computed tomography angiography clearly visualized the anatomical relationship between the LCX and coronary vein, and suggested the Conquest Pro wire penetrated both vessel walls into the venous lumen at the crossing point (Figure 6). 

Discussion. To the best our knowledge, this is the first case of coronary arteriovenous fistula formation between the LCX and the posterior vein of the left ventricle due to wire perforation following wiring of a totally occluded vessel, which was treated successfully with IVUS-guided PCI. There are a few previous reports of iatrogenic coronary arteriovenous fistula caused by stent implantation. Bradley et al reported a case of iatrogenic coronary arteriovenous fistula between the first obtuse marginal and great cardiac vein due to oversizing of the stents. The fistula remained in the final angiogram even after prolonged balloon inflation. However, spontaneous closure of the arteriovenous fistula was demonstrated by angiography at 7 months of follow-up.

Ziad et al reported a case of iatrogenic coronary arteriovenous fistula between the LAD and the anterior interventricular vein that occurred after stent deployment, which was successfully sealed with a covered stent.6

Although previously reported arteriovenous fistulas were induced by oversizing of the stent, the etiology of the fistula in our case was wire perforation. 

Wire perforation may be more likely to occur in CTO than in non-CTO lesions. Wire perforation may be induced by several factors, including aggressive antegrade guidewire manipulation with a tapered stiff wire, highly calcified lesions, a small distal vessel, and a poorly visualized distal vessel. Wire perforation has a potential risk of the following adverse cardiac events: intramyocardial hematoma, fistula formation into a cardiac chamber or a coronary vein, limited pericardial hematoma, and cardiac tamponade. Cardiac tamponade related to coronary perforation is one of the major reasons for discontinuation of the procedure, resulting in the failure of PCI. 

Therefore, we always confirm the wire position after it is advanced by contrast injection. However, because we could not determine the wire position from any angiographic images or by the feel of the wire during manipulation, IVUS was required to confirm the wire course by visualizing the structures around the wire. 

Fortunately, in our case of iatrogenic coronary arteriovenous fistula, we continued the procedure because the fistula did not rupture after balloon dilatation to deliver the IVUS catheter, and there was no extravasation. 

From the IVUS findings, we confirmed that the wire was in the coronary vein due to arteriovenous fistula formation. Then, we successfully redirected the wire through the optimal point of the plaque using real-time IVUS guidance, and this allowed sealing of the fistula following stenting. 

The majority of fistulas have a tendency to enlarge with time, which can induce high-output failure due to volume overload or myocardial ischemia due to coronary steal. Therefore, appropriate management for closing the fistula should be established during the procedure, if possible. Furthermore, IVUS was useful to show the anatomy when contrast injection did not convey enough information. The IVUS-guided parallel-wire technique was also effective to identify the optimal re-entry point and redirect the wire from the coronary vein into the true lumen.  

Conclusion. We describe a case where a guidewire penetrated a coronary vein during PCI of a coronary artery with CTO, which could be clearly visualized by IVUS but not angiography. Careful interpretation of real-time anatomical information provided by IVUS was the key to successful PCI in this case.


  1. Rathore S, Matsuo H, Terashima M, et al. Procedural and in-hospital outcomes after percutaneous coronary intervention for chronic total occlusions of coronary arteries 2002 to 2008: impact of novel guidewire techniques. JACC Cardiovasc Interv. 2009;2(6):489-497. 
  2. Morino Y, Kimura T, Hayashi Y, et al. In-hospital outcomes of chronic total occlusion: insights from the J-CTO Registry (Multicente CTO Registry in Japan). JACC Cardiovasc Interv. 2010;3(2):143-151.
  3. Hsu JT, Tamai H, Kyo E, Tsuji T, Watanabe S. Traditional antegrade approach versus combined antegrade and retrograde approach in the percutaneous treatment of coronary chronic total occlusions. Catheter Cardiovasc Interv. 2009;74(4):555-563.
  4. Okamura A, Iwakura K, Fujii K. ViewIT improves intravascular ultrasound-guided wiring in coronary intervention of chronic total occlusion. Catheter Cardiovasc Interv. 2010;75(7):1062-1066.
  5. Wilsmore B, Gunalingam B. Iatrogenic coronary arteriovenous fistula during percutaneous coronary intervention: unique insight into intra-procedural management. J Invasive Cardiol. 2009;22(5):460-465.
  6. Elghoul Z, Leesar MA. Iatrogenic fistula between the left anterior descending coronary artery and anterior interventricular vein following stenting. J Invasive Cardiol. 2007;19(7):E188-E191.


From the Division of Cardiology, Kusatsu Heart Center, Shiga, Japan.

Disclosure: The authors have completed and returned the ICMJE Form for Disclosure of Potential Conflicts of Interest. The authors report no conflicts of interest regarding the content herein.

Manuscript submitted November 26, 2012, provisional acceptance given December 21, 2012, final version accepted January 11, 2013.

Address for correspondence: Hidefumi Ohya, MD, Division of Cardiology, Kusatsu Heart Center, 407-1, Komaizawa-cho, Kusatsu, Shiga, 525-0014 Japan. Email: