HSC Section 3 - Trauma, Critical Care and Sleep Medicine

pushable fibered platinum coils, detachable coils, or a combina- tion of these materials was used for hemorrhage control. Descriptive statistics were used to characterize distribu- tion of bleeding, neurological morbidity, and mortality. Survival data were also recorded for the CCA/ICA and innominate bleed- ing group. RESULTS Between 2004 and 2014, 56 patients were identified who were treated using endovascular techniques for bleeding related to extracranial head and neck vascula- ture. After excluding epistaxis (n 5 12, 21.4%), post-tonsil- lectomy/postbiopsy hemorrhage (n 5 3, 5.4%), and other benign etiologies (n 5 3, 5.4%), we identified 33 head and neck cancer patients who underwent procedures for con- trol of hemorrhage involving the extracranial head and neck vasculature (Table I). There were five patients who had evidence of clinical bleeding, but diagnostic angiogra- phy did not show any discrete evidence of vessel wall irregularity, significant tumor blush, or pseudoaneurysm, and thus no intervention was done. In these five cases, bleeding stopped with conservative management and no further episodes of hemorrhage occurred. In the 33 remaining patients who ultimately under- went intervention, the median age was 58 years (range, 45–81 years). Distribution of the primary site of cancer is detailed in Table I, with larynx and oropharyngeal sites being the most common. A total of 38 discrete events of hemorrhage were identified in these 33 patients, and all episodes were controlled successfully by the endovascular procedure. Patterns and site of hemorrhage are detailed in Table II. Twenty-three hemorrhagic events related to distal ECA circulation are also detailed in Table II. Most com- monly, the lingual branch of the ECA was involved (47.8% of ECA branch cases), whereas multiple branches were involved in 30.4% of cases and included complex blood supply from the lingual-facial trunk and branches originating from the superior thyroid and inferior thy- roid arteries. In all such cases, embolization (coil emboli- zation and/or PVA particles) was used successfully with resultant complete control of hemorrhage.

TABLE I. Characteristics of Site of Tumors and Hemorrhage.

Etiology, n 5 56 Head and neck cancer related

33 (58.9%) 12 (21.4%)

Epistaxis

Post-tonsillectomy/biopsy

3 (5.4%)

Diagnostic only

5 (8.9%) 3 (5.4%)

Benign etiology

Age, yr, median (range), n 5 33 Primary site of tumor, n 5 33 Larynx/hypopharynx

58 (45–81)

12 (36.4%)

Oropharyngeal

12 (36.4%)

Oral cavity

9 (27.3%)

debate, as there is a requirement for systemic antiplate- let therapy, a high rate of local rebleeding (30%), or con- versely, in-stent thrombosis. 5,6 In this study, we analyzed the spectrum of oropha- ryngeal hemorrhage in head and neck cancer patients treated at a single institution and elaborate our manage- ment algorithm through the discussion of selected cases.

MATERIALS AND METHODS Data Collection

After institutional review board (IRB) approval (IRB #08- 14-23C) at the University Hospitals Case Medical Center, a dedi- cated neurointerventional database was used to identify patients treated for extracranial hemorrhage. In addition, patients with a biopsy-proven diagnosis of head and neck squamous cell cancer treated between 2004 and 2014 at the Head and Neck Institute at University Hospitals Case Medical Center and who underwent surgical management of profuse oropharyngeal hemorrhage were reviewed. Patients who were treated for epistaxis, post- tonsillectomy hemorrhage, and bleeding related to benign disease (vascular malformations, endonasal surgery complications) were excluded from the study. Details regarding patient demographics, site of disease, as well as previous surgical and nonsurgical treat- ment modalities were collected in a retrospective manner. Embolization and Stenting Procedures Procedures were performed in a neurointervention suite with monitored anesthesia care using moderate sedation or gen- eral anesthesia. Endovascular large vessel occlusion was per- formed using pushable and/or detachable coils as well as self- expandable occlusion devices (Amplatzer vascular plug; AGA Medical Corp., Plymouth, MN). Flow preservation with covered stent (GORE VIABAHN Endoprosthesis; W. L. Gore & Associ- ates, Newark, DE) was used in selected cases of internal carotid artery (ICA)/common carotid artery (CCA) hemorrhage. In such cases, a postprocedure antiplatelet agent was used to prevent thrombosis. Before parent vessel occlusion of the ICA/CCA, a balloon occlusion test (BOT) was performed in stable patients to assess collateral circulation. In other cases, however, a BOT was not possible because the patient was intubated, sedated, or unstable. In such cases, angiograms of the ICA and vertebral arteries (as needed) was performed bilaterally to evaluate for adequacy of the circle of Willis, so as to estimate potential col- lateral flow via an anterior communicating artery or posterior communicating arteries. If bleeding was localized to the ECA trunk or its branches, then polyvinyl alcohol (PVA) particles,

TABLE II. Pattern of Bleeding.

Great vessel related, events n 5 15 CCA/ICA

9 (60%)

ECA

5 (33.3%)

Innominate

1 (6.6%)

ECA branches related, events n 5 23 Lingual artery

11 (47.8%)

Facial artery (isolated)

1 (4.35%) 7 (30.4%)

Multiple (lingual-facial trunk, superior thyroid, inferior thyroid)

Superior thyroid artery

1 (4.35%)

Inferior thyroid artery

2 (8.70%)

Internal maxillary artery

1 (4.35%)

CCA 5 common carotid artery; ECA 5 external carotid artery; ICA 5 internal carotid artery.

Laryngoscope 127: February 2017

Manzoor et al.: Endovascular Techniques for Management of CBS

125