Flow diversion in the treatment of carotid injury and carotid-cavernous fistula after transsphenoidal surgery

Abstract

We describe a case of iatrogenic carotid injury with secondary carotid-cavernous fistula (CCF) treated with a silk flow diverter stent placed within the injured internal carotid artery and coils placed within the cavernous sinus. Flow diverters may offer a simple and potentially safe vessel-sparing option in this rare complication of transsphenoidal surgery. The management options are discussed and the relevant literature is reviewed.

Introduction

Transsphenoidal surgery has low morbidity and mortality.¹ However, complications from internal carotid artery injury can be associated with disability or even death. The outcome depends on the ability to recognize and manage complications promptly. We report a case of iatrogenic carotid injury following transsphenoidal surgery for pituitary adenoma, presenting as a carotid-cavernous fistula (CCF), managed with FD stent placement in the injured internal carotid artery (ICA) in addition to coiling of the ophthalmic vein and cavernous sinus.

Case report

A 47-year-old female was referred to our service for endovascular treatment of a right CCF detected five days after transsphenoidal surgery performed at an outside institution for a large pituitary adenoma. She presented with right proptosis, chemosis and double vision. The ophthalmologic exam revealed a right third nerve palsy with normal visual acuity. CT demonstrated a large bony defect within the right supero-lateral wall of the sphenoid sinus and medial wall of the right orbital apex along with heterogeneous soft tissue thickening within the sphenoid sinus (Figure 1). The CT angiogram demonstrated opacification of the right CS and an enlarged right superior ophthalmic vein (SOV) (Figure 1). These findings were consistent with a surgical carotid injury, with a secondary right CCF (and reflux into the right SOV). The recent surgical tract and sizeable sphenoid sinus bony defect made us concerned about possible delayed life-threatening epistaxis. This influenced our decision to attempt immediate post-operative closure of the fistula. A cerebral angiogram was performed confirming the presence of a carotid injury, with a secondary high flow direct CCF with primary drainage into the SOV (Figure 2). A balloon test occlusion (BTO) was performed with the patient awake under local anaesthesia and was not tolerated: the patient became drowsy and hemiplegic one minute after the balloon inflation within the petrous segment of the right ICA. Therefore we sought a way to repair the carotid artery and treat the CCF. The procedure was performed under general anaesthesia after loading the patient with 650 mg of ASA and 150 mg of Clopidogrel. Coiling of the SOV and CS was performed using an arterial approach with balloon protection (Hyper Glide EV3, CA) within the ICA. The SL10 (Stryker, Kalamazoo, USA) micro-catheter was placed over the wire into the SOV through the ICA defect. Satisfactory coil packing was obtained in the SOV and anterolateral aspect of the CS, targeting the carotid wall adjacent to the presumed site of injury. A control angiogram showed satisfactory results, but contrast was still leaking from the carotid artery through the coils. Subsequently, a 4 × 20 mm Silk (Balt Extrusion, Montmorency, France) FD stent was deployed via a Vasco microcatheter (Balt Extrusion, Montmorency, France) into the supraclinoid and cavernous ICA. In-stent balloon angioplasty was performed in order to ensure good apposition of the stent to the carotid wall and hopefully facilitate repair of the carotid defect. The final angiogram demonstrated near-complete angiographic cure of the CCF, with minimal residual shunting posteriorly and inferiorly through the pterygoid venous plexus and inferior petrous sinus (Figure 2). The patient tolerated the procedure well and the occulomotor symptoms immediately improved. MRI/MRA and angiographic controls were performed at 2, 4, 8 and 12 days post procedure in order to assess treatment stability. These showed minimal transient recanalization of the shunt with sluggish flow, without cortical or ophthalmic venous reflux. A dynamic CTA performed one month post procedure showed complete occlusion of the shunt on non-invasive vascular imaging. This was confirmed with catheter angiography performed at three months, when the Clopidogrel was discontinued. A one-year follow-up angiogram showed the absence of CCF and a ‘normal’ carotid artery with no in-stent stenosis (Figure 3).

Figure 1.

Axial and coronal CTA images demonstrate right proptosis, right sphenoid sinus lateral wall bony defect, soft tissues in the sphenoid sinus and iatrogenic communication between the cavernous carotid with the superior ophthalmic vein.

Figure 2.

Right ICA angiogram. Before treatment (a) showing high flow direct CCF draining in the SOV (arrowhead). Post treatment early (b) and late (c) phase showing almost complete cure of the CCF. Minimal residual flows shunt into the pterygoid plexus.

Figure 3.

Three months follow up right ICA angiogram. AP (a) and oblique (b) demonstrating complete cure of the CCF and no in-stent thrombosis or stenosis.

Discussion

Transsphenoidal surgery is a safe and effective procedure. However, potentially life-threatening complications such as haemorrhage from carotid laceration can occasionally occur. Arterial injuries during transsphenoidal surgery have been reported in 0.14%–1% of cases, and appear to be associated with high morbidity (24%) and mortality (14%).^2–4 Carotid stenoses, occlusions, false aneurysms and CCFs have all been reported as angiographic manifestations of postsurgical vascular trauma.^5–9 Thus, after initial management of the acute bleeding with nasal packing, prompt vascular imaging is recommended to assess the carotid arteries. Definitive treatment of the injured carotid artery is necessary in the case of penetrating trauma.

Iatrogenic penetrating injuries of the carotid artery during surgical procedures are at high risk of haemorrhage, as compared to CCFs occurring after a blunt trauma, which are rarely complicated by epistaxis. The recent surgical tract and lack of bony protection from direct penetrating injuries renders the patient at risk for delayed bleeding into the sphenoid sinus and, potentially, massive fatal epistaxis.^10,11 These life-threatening events can occur hours to years after surgery, and most importantly, they can occur even when postoperative angiography shows minimal signs of a carotid injury (such as a stenosis).² Some authors have proposed that the only way to safely secure this type of traumatic lesion is to sacrifice the carotid artery.² Other options that preserve the injured carotid artery, such as coiling of the cavernous sinus, may achieve temporary results but cannot guarantee long-term success. Selective coil occlusion of post-surgical false aneurysms with preservation of the injured carotid artery can be followed by delayed fatal epistaxis (J. Raymond; personal communication). One can imagine displacement of the coils when the initial cavernous sinus haematoma retracts. This could lead to re-exposure of the carotid tear to an unprotected space lacking bony covering and hence, a risk of delayed massive epistaxis. In our case, the CCF was from a direct carotid injury. If possible, in order to prevent delayed epistaxis, we would have treated the patient by parent vessel occlusion. However, our patient failed the balloon-test occlusion and we felt that simple balloon-assisted coiling would not offer enough protection. The post-coiling angiographic control showed a significant reduction in the shunt. This theoretically suggests that the patient still was at risk prior to shunt closure; however, we feel the use of a flow diverter added additional protection. The use of dual antiplatelets to prevent potential stent-related thrombotic complications combined with the lack of previous reports in using such a strategy were additional concerns that prompted us to perform repeated control imaging to confirm treatment efficacy.

Other treatments have previously been reported. Surgical trapping with ligation of the cervical and intracranial injured ICA is possible, but appears to be associated with high incidence of complications, including stroke and death.² A number of different endovascular treatment options for iatrogenic carotid injuries are available. Parent artery occlusion (PAO) with detachable balloons or coils is traditionally considered the safest and most durable treatment for arterial injuries.^2,12,13 However, a functionally incomplete circle of Willis may not allow carotid sacrifice, especially in the acute phase. BTO is generally performed with clinical and angiographic assessment with or without hypotensive challenge and EEG recording.^14,15 However, delayed ischaemic events have been reported in 5%–20% patients who initially tolerate BTO.¹⁶

Other endovascular, vessel-preserving options include transarterial or transvenous delivery of balloons, coils, embolic liquid agents, stent assisted coiling and covered stent placement. As single or combined approaches, the success rate of endovascular closure of CCFs is in the range of 85%–99%.^4,17–23 Long-term angiographic follow-up is necessary after vessel-preserving techniques as delayed recanalization and risk of haemorrhage have been described in the literature.^24,25 Covered stent grafts, when available, can offer successful reconstruction of the injured vessel. However, covered stents are rigid and therefore have poor navigability within intracranial vessels. More recently, covered stents designed for the intracranial vasculature have been effectively used for the treatment of internal carotid artery pseudo aneurysms.²⁶

Flow diverter stents have recently been mainly used for otherwise untreatable, broad-necked, fusiform or partially thrombosed intracranial aneurysms.^27–30 They are designed to reduce flow velocity in the aneurysm sac and promote thrombosis while maintaining flow in the parent artery and branch vessels. To our knowledge, the use of flow diverter stents as a treatment option for carotid injuries and direct CCF post transsphenoidal surgery has not been described previously. Nadarajah et al.³¹ reported a case of telescopic pipeline flow diverter stents used for the treatment of a traumatic CCF with no bony defect.

In our case, the goal was to seal the communication between the artery and the CS. We first placed coils into the venous compartment in order to reduce the arterio-venous shunt before placing the FD into the ICA. The rationale was to immediately decrease the flow through the ICA defect and promote neointima formation on the FD. FD stents are often deployed in a telescopic fashion in order to minimize porosity over the target arterial defect.³¹ We postulated that a single flow diverter, combined with coil embolization of the venous side of the CCF, would suffice. Following in-stent balloon angioplasty, there was near-complete angiographic cure of the CCF.

The risk of haemorrhage may theoretically persist until complete resolution is demonstrated as long-term antiplatelet therapy is required to prevent in-stent thrombosis. Therefore, close imaging follow-up may be necessary to ensure treatment stability. Non-invasive, time-resolved CTA or MRA may be reliable modalities to follow pre- and post-treatment intracranial arterio-venous shunts.^32,33

Conclusion

Flow diverter stents may offer a simple, safe and efficient option to treat carotid injuries and direct iatrogenic CCFs in patients who cannot tolerate parent vessel occlusion.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Conflict of interest

None declared.