Endovascular management of carotid blowout syndrome in patients with head and neck cancers

Summary

Endovascular treatments for carotid blowout syndrome (CBS) have been advocated by interventional neuroradiologists. We therefore retrospectively evaluated the efficacy of endovascular treatments of CBS in 16 patients with head and neck cancers (HNC). The clinical, treatment and outcome data were evaluated in 16 HNC patients with CBS, all of whom underwent permanent embolization or covered stent graft of the affected carotid artery. All patients received multimodal treatments, including radiotherapy (mean total dose, 78.5 Gy). CBS was caused by tumor carotid invasion in 8 patients, pharyngocutaneous fistula in 7, and laryngeal chondroradionecrosis in 1, with the external and common carotid arteries being the most common rupture sites. CBS was occluded by embolization or revascularized by covered stent placement. Immediate hemostasis was achieved in all patients; however, 7 patients had recurrent CBS, all of whom were retreated effectively by endovascular management. Three patients had strokes and four had extrusion of intervention materials from the infected wounds. Most patients died of tumor progression, with a mean survival time of five months from initial CBS; only two patients survived. Endovascular therapy, by both permanent occlusion and stent grafts, is effective in hemostasis of CBS but its long-term efficacy may not be high in these HNC patients.

Introduction

Carotid blowout, defined as the rupture of the extracranial carotid arteries or their major branches, is one of the most devastating complications associated with therapy for head and neck cancers. The clinical spectrum of signs and symptoms related to rupture of the carotid artery is known as carotid blowout syndrome (CBS), which can manifest as either an episode of acute hemorrhage or as the potential threat of bleeding.1, 2, 3 CBS tends to occur in patients with conditions such as recurrent or persistent head and neck tumors, pharyngocutaneous fistula, and radiation necrosis.⁴ The mortality and neurovascular morbidity rates associated with this complication are as high as 40% and 60%, respectively.1, 5, 6, 7, 8

Surgical management of CBS is usually difficult because the surgical exploration is performed in previously irradiated or infected fields. In addition, emergency surgical ligation of the common or internal carotid artery without testing of the collateral cerebral circulation can result in major neurologic morbidity and mortality.6, 7, 8 Such poor outcomes have been substantially improved by endovascular management of acute bleeding in the head and neck, using methods such as occlusion of the offending carotid arteries with detachable balloons, coils, or microparticles.1, 5, 9 In 15–20% of patients, however, these procedures can induce delayed cerebral ischemic complications,¹⁰ resulting from an incomplete circle of Willis, thromboembolism arising from an acutely occluded carotid artery, and/or delayed collateral failure.¹¹ Alternatively, CBS can be managed by reconstructing the damaged artery with covered stents.10, 12, 13 These recent developments in the endovascular treatment of CBS have significantly reduced morbidity rates to 0–8% and mortality rates to 0%.¹⁰

Although endovascular management has resulted in high rates of arrest of acute carotid hemorrhage and low morbidity rates, the long-term sequelae of endovascular treatment has not been determined in head and neck cancer (HNC) patients.12, 13, 14 Endovascular treatment of CBS, including permanent occlusion and stent grafts of the affected carotid artery, has been advocated mainly by interventional neuroradiologists. We therefore evaluated the efficacy of endovascular treatments of CBS in HNC patients, focusing on arresting bleeding, complications, and survival.

We retrospectively reviewed the records of all HNC patients with CBS who had been referred for interventional radiologic treatment in our single institution. Patients managed by embolization or stent-assisted endovascular treatment of the carotid artery were identified from our neurointerventional database and HNC registry. Patients with a present or past history of HNC and who underwent neurointervention to control acute or impending bleeding of the carotid artery were included, whereas patients with no history of HNC treatment and those lost to follow-up were excluded.

The clinical severity of CBS for each patient was classified into three separate entities1, 2, 3: Grade 1, “threatened” carotid blowout, characterized by physical and radiologic findings suggestive of inevitable hemorrhage from an exposed segment of the carotid artery, neoplastic invasion of the carotid artery or nonhemorrhagic pseudoaneurysm; Grade 2, “impending” carotid blowout or sentinel hemorrhage, characterized by an episode of transcervical or transoral hemorrhage, typically from a pseudoaneurysm, which usually resolves spontaneously or with surgical packing; and Grade 3, acute carotid blowout, characterized by profuse, poorly controlled hemorrhage that cannot be stopped by packing or pressure. Patients who experienced subsequent episodes of recurrent CBS were also identified.²

All interventional procedures were performed under local anesthesia. Digital subtraction angiography using a percutaneous, transfemoral arterial route was used to evaluate the common, internal, and external carotid and intracranial vascular anatomy. If bleeding was localized to the external carotid artery or branches, they were occluded with detachable balloons, coils, gelfoam, glue, or polyvinyl alcohol particles. If bleeding was from the internal or common carotid artery, collateral cerebral circulation was evaluated by angiography. Stent-assisted endovascular treatment was performed in patients with an incomplete circle of Willis, a thromboembolism from the acutely occluded artery, or a positive carotid occlusion test during or prior to intervention (Fig. 1). The covered stent (Taewong Medical, Seoul, Korea), composed of a self-expanding nitinol wire covered with polytetrafluoroethylene (PTFE), was passed over the exchange guidewire and positioned at the level of the bleeding site including a pseudoaneurysm and deployed across the corresponding segment of the pseudoaneurysm.¹⁵ The common or internal carotid artery was completely occluded in patients with uncontrolled massive carotid hemorrhage and good collateral cerebral circulation on angiograms, usually at the second session of hemorrhage (Fig. 2).

Complications from procedures, additional treatments, and outcome and survival data were collected from the patients’ medical records. The survival time of each patient was calculated from the time of initial CBS to death. Our institutional review board waived informed consent and approved the design of this retrospective study.