Vertebral Artery Aneurysms and Cervical Arteriovenous Fistulae in Patients with Neurofibromatosis 1

Guillermo Higa; John P. Pacanowski Jr; David T. Jeck; Kaoru R. Goshima; Luis R. León Jr


Vascular. 2010;18(3):166-177. 

In This Article


Literature Search

A MEDLINE search from 1950 until 2009 (National Library of Medicine and OVID) and an extensive manual search were carried out using bibliographies from relevant published articles. The main terms for inclusion were neurofibromatosis, von Recklinghausen disease, VA aneurysms, pseudoaneurysms, AVFs, and other similar terms. As there are no prospective studies, all retrospective articles and case reports were included. These articles were organized into two tables: those reporting patients afflicted with VA aneurysms in Table 1 [1–15] and those reporting VA-dependent AVF in Table 2.[4,6,9,15–44] For space reasons, only reports following the year 2000 were included in our tables. However, for the analysis, all reports were included from 1950 and after. These cases were analyzed with regard to demographic features (age and gender), signs and symptoms at presentation, aneurysm or AVF size and location, follow-up, therapy, and outcomes.

VA Aneurysms

Sixteen articles reported at least (some reported "multiple" aneurysms, without specification of the precise number) 17 cases of VA aneurysms (affecting 16 patients) since 1981, when Pentecost and colleagues published the first report of a NF1-related VA aneurysm.[1] Three-quarters of them were male patients (13 of 17; 76.5%) with an average age of 44.5 years (range 1–74 years). These aneurysms mostly presented during adulthood, with only two reports of aneurysms affecting patients aged 11 and 18 years.[12] Most aneurysms affected the left VA (two cases unknown; 12 of 17; 70.5%) and spanned from C1 to T2. Their size was mostly unknown, but from those reported, it varied from 2 to 9 cm, with some described as "large" or "huge."

Patients with VA aneurysms presented either asymptomatic (eg, incidental finding on angiography done, for instance, for a ruptured subclavian artery aneurysm) or with a myriad of complaints, including neck or chest pain; neck mass and respiratory compromise owing to large or ruptured aneurysms; vertebral body erosion; vomiting and hemoptysis; hemodynamic instability owing to aneurysm rupture; upper extremity neurologic symptoms related to compression of the brachial plexus; and neurologic symptoms related to the VA territory. The follow-up varied from 0 to 11 years. The therapy given included observation (mostly used in old reports); open surgical therapy, including a thoracotomy and proximal and distal aneurysm ligation; surgical VA ligation; and an LSA ligation with a carotid-subclavian artery bypass. Endovascular techniques were also used, including proximal VA occlusion or aneurysm embolization by detachable balloons or microcoils and aneurysm packing with n-butyl cyanoacrylate (NBCA; B. Braun, Tuttlingen, Germany) mixed with iodized oil 1:1 (Lipiodol, André Guerbet, Aulnaysous-Bois, France). Most cases were associated with a favorable recovery, with resolution of symptoms at presentation. Mortality cases were mostly associated with untreated cases or death from a concomitant ruptured abdominal aortic aneurysm. Two patients died after therapy for the VA aneurysm was instituted.

VA Arteriovenous Fistulae (Table 2)

Our review identified 33 articles reporting 36 patients afflicted with 40 VA AVF. Their mean age was 36.4 years (range 1 month–60 years). There were 25 females and 11 males. Twenty-four cases of VA AVF were located on the left and 15 on the right (one case unknown). As in the case of patients with VA aneurysms, those with VA AVF presented as either asymptomatic or with multiple complaints, including symptoms and signs owing to a mass effect, neurologic or cardiorespiratory compromise owing to vascular rupture, or mass effect of the fistulae. Four cases were left untreated. When they were treated, these therapies included surgical ligation and removal, surgical resection, VA coagulation (bipolar coagulation of epidural varix), direct AVF puncture with glue (Eudragit, Evonik Röhm GmbH, Darmstadt, Germany) or NBCA and ethiodized oil embolization, or AVF embolization using silicone balls and coils through an open approach (VA arteriotomy). Endovascular methods included embolization by using muscle fragments, coils, polyvinyl alcohol particles, or detachable balloons. This therapy resulted in favorable outcomes and neurologic improvement in the vast majority of cases. Mortality cases were mostly observed in the early reports (before the 1980s).

Neurofibromatosis associated with vascular disorders is generally known as vascular neurofibromatosis. Vascular disease associated with neurofibromatosis is, however, rarely seen in the large cerebral arteries.[18] It is thought that neurofibromatosis-related mesodermal dysplasia causes arterial fragility and aneurysm development.[9] These aneurysms may have part of the vessel wall replaced by neurofibromatosis tissue.[9,45] Minor trauma, in the setting of arterial wall fragility, has also been implicated.[3] Extradural AVF have also been reported in the context of NF1, with or without concomitant aneurysms, probably caused by the same mechanism of origin as aneurysms. A clinical triad has been described, including symptoms of NF1, progressive radiculomyelopathy, and a bruit.[43] Several vessels have been reported, including the intercostal arteries[46] or most often the VAs.[4,6,9,15–44] In these patients, the shunt typically occurs between the V2 segment of the VA and the epidural venous plexus or involving the V3 segment, the most mobile of all VA portions. This fact and the neck mobility and arterial fragility produced by its relationship with neurofibromatosis are thought to play key roles in the development of AVF.[9] Deans and colleagues suggested that dysplastic smooth muscle or neurofibromatosis proliferation in the arterial wall could lead to aneurysm formation, leakage, and eventual rupture into adjacent veins.[20]

Our review of the literature identified only 17 cases of NF associated with aneurysms and 40 cases of association of NF1 with AVF. It has been previously noted that in most cases, the extracranial VA AVF are most often located on the left side,[23] as our review indicates. This is thought to be due to the fact that the bone defect around the asterion is predominantly seen on the left side in cases of NF1.[23]

The possibility of a vascular lesion in NF1 patients should always be considered to avoid complications or disastrous outcomes. For instance, in the case reported by Peyre and colleagues, the clinical diagnosis was that of a sixth cervical nerve neurofibroma, based on findings of severe pain following the C6 dermatome pattern in a patient with NF1, combined with a plain radiograph showing enlargement of the right C5–C6 exit foramen.[12] CT and magnetic resonance imaging (MRI) were needed to accurately diagnose a vascular lesion affecting the right VA. Horsley and colleagues treated a patient with neurofibromatosis and severe neck kyphosis.[7] MRI revealed a large left-sided anterolateral plexiform neurofibroma. This patient was treated with halo traction to correct the deformity. Unawareness of a VA aneurysm led to its rupture owing to progressively applied traction, with sudden development of respiratory distress and a neck mass with a large retropharyngeal swelling, requiring emergent tracheostomy. Westacott and colleagues and later Kähärä and colleagues treated patients with upper extremity neurologic deficits of unclear cause but possibly owing to a neck hematoma or soft tissue masses, based on preoperative imaging.[28,38] Surgical exploration revealed vascular masses with arterial leakage or massive blood loss after piercing them with needles. Surgery had to be aborted, and angiography was subsequently performed; AVF were diagnosed as the cause of their symptoms. Neck auscultation has been recommended in all patients with neurofibromatosis who present with signs of spinal cord or nerve compression. If a bruit is present, further vascular imaging should be performed.[20]

Spontaneous massive hemothorax,[11] or compressive symptoms, as in the case of our patient and others reported,[37] are potential complications of these aneurysms or AVF. Observation, surgical, and endovascular techniques for therapy have been attempted (see Table 1 and Table 2).[1–44] Withholding therapy has been done with mixed results, with death being the most frequent consequence.[6,11] However, spontaneous resolution of neurologic symptoms without intervention[28] and lack of symptom progression on follow-up[5] have been reported without any intervention. Open surgical therapy is often associated with major risks, most importantly, exsanguination.[6,10]

The endovascular therapy of complex AVF is challenging, requiring cannulation of several feeders to the AVF, which requires advanced technical skills. Often enough, as in the case of our patient, a decision has to be made as to when to quit pursuing the obliteration of all feeding vessels, once an adequate flow reduction is achieved, expecting full resolution with time. Ushikoshi and colleagues performed successful obliteration of an AVF that occurred 11 years after a ruptured VA aneurysm was treated endovascularly.[9] The fistula occurred between the reconstituted distal VA and the paravertebral venous plexus at the C1 level. This was a 2-day procedure, starting with hairy coils and alcohol embolization of the arterial side of the AVF followed by the venous side the following day. Our technique was strikingly similar to that used by Roth and colleagues, the so-called "saddling the vertebro-basilar junction."[37] In their case, the LSA was patent, but antegrade access to the left VA was unsuccessful owing to the high flow of the AVF, preferentially directing the microcatheter into the venous system. However, in our case, large feeding branches arising from the ipsilateral ECA added a great deal of technical difficulty. While treating this patient, left upper extremity ischemia was a potential threat in the face of a chronic LSA occlusion, by occluding the VA. However, given the acute and life-threatening manner of presentation, we felt that it was a risk worth taking. Roth and colleagues found on surgical exploration of a patient with a dysplastic LSA many abnormal vessels that may have represented collateral supply to the upper extremity.[37] In fact, our patient had a viable upper extremity with strong Doppler signals for the entire hospital stay.

The outcome of our patient has been favorable in the short term. Because small feeders were left untreated, there is always the potential for recurrent AVF or aneurysm formation, as has been shown in prior reports.[29,37] Murayama and colleagues suggested the use of intravascular Doppler for transvenous monitoring of blood flow velocity.[34] This technique uses a 12 MHz piezoelectric ultrasonic transducer mounted on the tip of a 0.014-inch guidewire. Most often, current techniques for AVF embolization rely on the subjective information obtained by DSA, which, as in the case reported by Murayama and colleagues,[34] can be misleading in determining remaining feeder or outflow vessels that may need to be embolized to achieve complete AVF embolization. Hence, DSA aids in determining the hemodynamic progression and end point of the procedure. Given that our patient's therapy was done on an emergent, lifesaving basis, we feel that it is prudent to follow this patient up with serial imaging studies and treat her according to our findings.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: