Neurovascular manifestations of connective-tissue diseases: A review

Abstract

Patients with connective tissue diseases are thought to be at a higher risk for a number of cerebrovascular diseases such as intracranial aneurysms, dissections, and acute ischemic strokes. In this report, we aim to understand the prevalence and occurrences of such neurovascular manifestations in four heritable connective tissue disorders: Marfan syndrome, Ehlers-Danlos syndrome, Neurofibromatosis Type 1, and Loeys-Dietz syndrome. We discuss the fact that although there are various case studies reporting neurovascular findings in these connective tissue diseases, there is a general lack of case-control and prospective studies investigating the true prevalence of these findings in these patient populations. Furthermore, the differences observed in the manifestations and histology of such disease pathologies encourages future multi-center registries and studies in better characterizing the pathophysiology, prevalence, and ideal treatment options of neurovascular lesions in patents with connective tissue diseases.

Introduction

Patients with connective tissue diseases such as Marfan syndrome, Ehlers–Danlos syndrome (EDS), neurofibromatosis type 1 (NF1) and Loeys–Dietz syndrome (LDS) are thought to be at a higher risk for a number of cerebrovascular diseases such as intracranial aneurysms, dissections and acute ischemic strokes. The most commonly accepted causal explanation of this association is that genetic mutations involved in such connective-tissue diseases affect the collagen and proteoglycans that construct the extracellular matrix, thus resulting in the weakening of the vessel wall. Because of the purported higher prevalence of these neurovascular complications in patients with connective-tissue diseases, it is important for neurovascular specialists including neurologists, neurosurgeons, and neuroradiologists to be aware of the various cerebrovascular manifestations of these diseases. We present a comprehensive review of the neurovascular manifestations occurring in the most common inheritable connective-tissue diseases.

Description of diseases

Marfan syndrome

Marfan syndrome is the most commonly inherited connective-tissue disorder, with a reported prevalence of two or three per 10,000 individuals. An autosomal dominant mutation of the fibrillin 1 (FBN1) gene is the most frequent cause of Marfan syndrome. Sporadic de-novo mutations account for 25% of Marfan cases. Marfan syndrome has no racial or gender predisposition.¹ Aortic root disease, mitral valve prolapse, long bones and joint laxity, arachnodactyly, and ectopia lentis are some of the more well-known classical presentations of Marfan syndrome.² Currently, the revised Ghent nosology published in 2010 is used as the diagnostic criteria of Marfan syndrome, and places heavy emphasis on aortic root dilation/dissection, ectopia lentis, and genetic information.³

Once the diagnosis is made, the current management recommendations center around the use of beta blockers, restriction of vigorous physical exercise, routine and noninvasive monitoring of aortic diameter, and elective surgical repair of the aorta. There are currently no specific recommendations for screening of intracranial aneurysms and/or dissections.

The FBN1 mutation causes an alteration in the structural role of microfibrils in coordinating tissue morphogenesis, homeostasis, and response to hemodynamic stress.⁴ Histologically, fragmentation of lamellae, cystic medial necrosis (lacunar appearance of medial degeneration), fibrosis, and loss of smooth muscle cells are observed in the medial layer of the aortic root in patients with Marfan syndrome.^5–9 The wide variation in histological appearance among Marfan patients warrants further investigation of the direct cause of their vascular abnormalities.⁸

EDS

EDS describes a group of disorders with shared primary characteristics of skin hyperextensibility, joint hypermobility, and tissue fragility. This disease is seen in approximately one out of 5000 individuals, with some types being rarer than others.¹⁰ Each of the six major types of EDS has unique genetic defects and inheritance patterns.^10–12 The most commonly seen of these types are classic (EDS type I and II), hypermobility (EDS type III), and vascular (EDS type IV), with vascular EDS having the most prominent neurovascular manifestations.¹³ The diagnosis of EDS and its various types should be suspected when some combination of their shared features are present—joint hypermobility, multiple joint dislocations, translucent skin, poor wound healing, easy bruising, unusual scars, spontaneous ruptures of organs, and dissections of blood vessels. In addition, genetic testing for specific genes is available for all EDS types except for the hypermobility type. With the primary interest of our review being in the neurovascular manifestations of connective-tissue diseases, manifestations of vascular EDS (type IV) in the current literature will be emphasized.

Histological analysis of the aortic walls of patients with EDS IV show a broadened intima with fibrosis, abundant cholesterol crystals, and derangement of elastic fibers.¹⁴ Skin biopsies of EDS patients indicate collagen bundles that are thin and rare in the dermis and hypodermal septae. Under polarized light, such collagen bundles are less refringent than in normal skin.¹⁵ Type III collagen, mutated in EDS IV, is essential in providing structure and strength to connective tissue (skin, blood vessels, and internal organs) as a major component of the extracellular matrix. It is hypothesized that the poor assembly of these type III collagen fibers in EDS IV results in the neurovascular manifestations of EDS IV.

NF1

NF1, also known as von Recklinghausen disease, is the most common form of neurofibromatosis. It is an autosomal dominant mutation of the neurofibromin 1 (NF1) gene on chromosome 17 that results in the malformation of neurofibromin protein. NF1 affects one in every 2600 to 3000 individuals, with half of its presentation being familial, and the other half being de novo in nature.^16,17 It is suggested that advanced paternal age increases the likelihood of de novo NF1.¹⁸ There is a typical order of appearance of its key clinical manifestations: café-au-lait macules, axillary/inguinal freckling, Lisch nodules, and neurofibroma tumors.^19,20 Some other complications seen include osseous lesions, optic pathway glioma in children, childhood hypertension, and malignant transformations of tumors in adolescence and adulthood.^21–23 The diagnosis for NF1 is clinical, and follows the diagnostic criteria set by the United States (US) National Institutes of Health (NIH) Consensus Conference. These diagnostic criteria are highly specific and sensitive, with 97% of patients meeting the diagnostic criteria by age 8.¹⁹ Because of this, genetic testing is not required, but can be helpful in the diagnosis of individuals who do not fully meet the diagnostic criteria.

NF1 vasculopathy is a significant but under-recognized complication of systemic NF1, and its pathogenesis is hypothesized to be due to the mutant neurofibromin protein expressed in the endothelial and smooth muscle cell layers of blood vessels.^24,25 One study in rats determined the contribution of NF1 expression to NF1 vasculopathy by locating neurofibromin in blood-vessel structures.²⁶ Neurofibromin was detected in the endothelial cell layer of rat cerebral vessels, renal arteries, and aorta. Neurofibromin was also detected in the smooth muscle layer specifically of the aorta, but not of the cerebral or renal vessels. Histologic features of affected NF1 patients include concentric intimal proliferation of spindle cells, marked proliferation of the intima with fibrous thickening, aneurysmal formation with irregular smooth muscle loss, and nodular proliferation of epithelioid and spindle cells. It is hypothesized that all of the above features are contributing factors to the neurovascular malformations seen in patients with NF1.

LDS

LDS is an autosomal-dominant connective-tissue disorder that stems from a variety of genetic alterations that all result in aortic aneurysms, generalized arterial tortuosity, hypertelorism, and bifid/broad uvular or cleft palate.²⁷ The most widely understood mutations that are associated with LDS are transforming growth factor B receptor 1 (TGFBR1), transforming growth factor B receptor II (TGFBR2), decapentaplegic homolog 3 (SMAD2), and transforming growth factor B 2 ligand (TGFB2). All mutations show similar defective changes in the transforming growth factor-B (TGF-B) signaling cascade, and affected individuals show widespread arterial involvement with a proposed increase in risk of aneurysms and dissections.^28–30 The diagnosis of LDS is clinical in nature, with the finding of aortic aneurysm and characteristic facial features being the most suggestive. However, the definitive diagnosis of LDS can be accomplished only through genetic testing.³¹

In the first description of LDS by Van Laer et al., an examination of aortic tissue demonstrated the following histologic features: fragmentation of elastic fibers, loss of elastin content, and accumulation of amorphous matrix components in the aortic media.³⁰ Structural analysis showed a loss of intimate spatial association between elastin deposits and vascular smooth muscles and a marked excess of aortic wall collagen.³² Thus, although not specific, LDS aortic samples have significantly more diffuse medial degeneration than those of patients with Marfan syndrome.³³ Genetically, an interesting phenomenon of increased TGF-B signaling is observed, despite the mutations in LDS being loss of function in nature. As there is no specific understanding of the pathogenesis of vascular malformations and aneurysms in patients with LDS, it is hypothesized that similar changes to the aorta are also seen in the neurovasculature.

Intracranial aneurysms in inheritable connective-tissue diseases

In review, the current hypotheses that exist for the neurovascular manifestations of these connective-tissue diseases differ slightly from one another. In Marfan syndrome, the medial smooth muscle cell layer of the vascular wall is histopathologically affected, with suggested causes being alterations of microfibrils and/or bioavailability of TGF-beta. In EDS, fibrosis of the tunica intima is observed, with the suggested cause being the lack of collagen type III, a fundamental feature of reticular fibers. In NF1, the deposition of neurofibromin in the endothelial cell layer of cerebral vessels, concentric intimal proliferation, and irregular smooth muscle loss of the medial vascular wall layer are attributed to vessel weakening. Finally, in LDS, the fragmentation of elastic fibers and diffuse medial degeneration of vasculature is hypothesized to be the cause of neurovascular malformation. Examples of cerebral aneurysms seen in connective-tissue disease patients are provided in Figure 1.

Figure 1.

Examples of intracranial aneurysms in connective-tissue diseases. (a) Incidental 3 mm supraclinoid internal carotid artery (ICA) aneurysm in a 64-year-old male Marfan syndrome patient. The aneurysm was managed conservatively. (b) An 8 mm dolichoectatic and fusiform aneurysm of the basilar artery in a 46-year-old female with Ehlers–Danlos syndrome. This aneurysm was managed conservatively because of the lack of good treatment options for this type of aneurysm in the setting of connective-tissue disease. (c) Non-contrast head computed tomography (CT) of a 50-year-old male with neurofibromatosis type 1 (NF1) shows a small amount of subarachnoid and subdural hemorrhage adjacent to the left temporal lobe. (d) Cerebral angiography demonstrates a 3 mm ruptured superior hypophyseal aneurysm that was effectively treated with stent-assisted coiling with no complications.

Marfan syndrome

A number of studies have been performed to examine the association between intracranial aneurysms and Marfan syndrome.^34,35 Aneurysms in Marfan syndrome can be saccular, fusiform or dissecting. There is a propensity for aneurysms in Marfan syndrome to be located in the proximal intracranial carotid artery.³⁶ A series of 10 clinical reports, one pathology case, and an autopsy series of seven patients suggested that individuals with Marfan syndrome may have an increased prevalence of intracranial aneurysms. However, there are many speculations that these reports did not clearly confirm clinical manifestations of Marfan, thus leading to the questioning of a true diagnosis of Marfan syndrome in these patients.³⁷ One recently published study by Kim et al found that the prevalence of intracranial aneurysms among patients with Marfan was 14% (8/59).³³ To date, no prospective intracranial aneurysm screening study has been performed among Marfan syndrome patients to determine the true incidence of intracranial aneurysms.

EDS

The potential relationship between collagen deficiencies in EDS and cerebral aneurysms has long been speculated, but not tested in the form of a prospective screening study.^38,39 Intracranial aneurysms in EDS patients can be saccular or fusiform and are most commonly located in the cavernous sinus.³⁶ Studies examining the prevalence of aneurysm and subarachnoid hemorrhage in EDS patients have yielded varying results. In a study of 419 patients with EDS or a family history of the disease, Pepin et al. found six patients who had intracranial aneurysms.⁴⁰ In a study of 202 patients, North et al. noted that six patients had cavernous carotid fistulae (CCF), four had ruptured intracranial aneurysms and four had intracranial hemorrhage of uncertain etiology.⁴¹ However, not all patients in these studies received screening for intracranial aneurysms with imaging. The largest study to date on the prevalence of unruptured intracranial aneurysms among EDS patients with intracranial vascular imaging found a prevalence of 11%.³³

NF1

There are varying reports regarding the prevalence of intracranial aneurysms in NF1 patients. One large study found that only one of 316 NF1 patients undergoing brain magnetic resonance imaging (MRI) had an intracranial aneurysm. It is important to note only eight of these 316 patients had magnetic resonance angiography (MRA) performed. Since MRI without MRA is not sensitive for the detection of intracranial aneurysms, it is likely that the prevalence of aneurysms in this population was higher than reported.²⁴ A small autopsy study by Conway et al. found that none of the 25 patients diagnosed with NF1 had intracranial aneurysms on autopsy.⁴² On the contrary, a study conducted by Schievink et al. compared a sex- and age-matched control population with an NF1 population for the incidental detection of aneurysms with MRI brain examination. In this comparison of 39 NF1 patients to 526 control patients, the authors demonstrated a prevalence rate of 9% in NF1 patients versus 0% in the control group.⁴³ Lastly, in a study of 47 patients with intracranial neurovascular imaging, Kim et al found an aneurysm prevalence of 11%.³³ Ultimately, further studies are needed to confirm these findings.

LDS

There are only a few reports on the prevalence of intracranial aneurysms in LDS patients.^44–48 In one neuroradiologic study conducted in 25 patients with positive genetic testing for LDS, eight (32%) had intracranial aneurysms, suggesting the importance of serial noninvasive imaging monitoring in younger patients.⁴⁹ In another study of 25 patients with clinical or genetic diagnosis of LDS, Kim et al found that 28% of patients had an intracranial aneurysm.³³ In the original study describing LDS, approximately 2% of the patients died of cerebral bleeding and 10% of patients had head or neck aneurysms.

Treatment and screening recommendations

In general, screening of intracranial aneurysms in these patient populations is not advised owing to the high risk of surgical or endovascular treatment of intracranial aneurysms in these patients. Incidentally detected unruptured aneurysms are generally managed conservatively because of the extremely friable nature of blood vessels in patients with connective-tissue diseases.

There are few series and case reports that describe surgical or endovascular treatment of intracranial aneurysms in patients with connective-tissue diseases. Among patients with Marfan syndrome, connective-tissue fragility is not a major problem in the surgical treatment of intracranial lesions; however, the extreme ectasia and tortuosity of the carotid and vertebral arteries can make endovascular treatment of these lesions difficult.³⁶ Surgery in the setting of EDS is particularly challenging. In a small series of four EDS patients undergoing intracranial aneurysm surgery, Schievink et al. reported one patient who died as a direct result of the surgery and a high rate of postoperative complications including spontaneous pneumothorax and vertebral artery dissections.⁵⁰ Surgical repair of aneurysms in NF1 can be complicated by the presence of intracranial arterial occlusive disease, excessive vascular fragility and anatomic distortion from sphenoid wing dysplasia.³⁶ Some small studies and case reports suggest that despite the vessel fragility seen in LDS patients, surgical and endovascular management are safe, effective, and durable treatment modalities.^46,48,51 To our knowledge, there are no systematic studies reporting outcomes of treatment of aneurysms in connective-tissue disease.

Dissections, pseudoaneurysms, and inheritable connective-tissue disease

Cervical carotid dissections

Marfan syndrome

The annual incidence rate for spontaneous internal carotid artery dissections in the general population is 1.72 per 100,000 individuals. However, owing to its asymptomatic nature, the true incidence may be higher.^52,53 While no study has demonstrated an association between Marfan syndrome and cervical carotid dissection to date, Marfan syndrome patients are known to sometimes suffer from these dissections, both isolated to the cervical and intracranial vasculature as well as a result of extension of Stanford type A thoracic aortic dissection (Figures 2–4).^54,55 The pathophysiology of cervical and intracranial dissections is similar to that of intracranial aneurysms. Pathological studies have found that a combination of intimal proliferation, medial degeneration, and fragmentation of the internal elastic lamina is responsible for dissections in these patients.^56,57 Mucopolysaccharide deposition in the tunica media, fibromuscular dysplasia and cystic medial necrosis and focal fragmentation of the internal elastic lamina have also been suggested as potential causes.^57,58 Pathology findings of dissections in EDS, NF1, and LDS are currently not reported.

Figure 2.

A 43-year-old male with Marfan syndrome presented with infarctions in the left frontal and temporal lobes ((a) and (b)). Time of flight magnetic resonance angiography (MRA) shows a focal dissection of the left M1.

Figure 3.

A 52-year-old male with Marfan syndrome with Stanford type A dissection (a) extending into both common carotid arteries (b).

Figure 4.

A 36-year-old male with neurofibromatosis type 1 (NF1). (a) Diffusion-weighted imaging (DWI) images show an acute infarct in the right corpus callosum. ((b) and (c)) Sagittal T1-weighted images show a dissection of the distal cervical and petrous internal carotid artery with high T1 signal in the vessel wall consistent with intramural hematoma (red arrows).

EDS

While many case reports of cervical carotid and vertebral artery dissections have been reported in EDS patients, only a few studies have examined the incidence of dissection in this population.^59–64 Brandt et al. conducted a study of skin biopsies on 65 patients with proven non-traumatic spontaneous cervical artery dissection (sCAD). Although only three patients had clinical manifestations of EDS, 36 patients had ultrastructural aberrations of the connective tissue commonly found in EDS II or EDS III, suggesting that CADs are associated with ultrastructural connective tissue abnormalities such as EDS.⁶⁵ This study further supported the same association suggested in a smaller cohort of 25 cervical arterial dissection patients in which 17 (68%) patients had ultrastructural abnormalities associated with EDS II or III.⁶⁶ Ulbricht et al. conducted a study of skin biopsies on seven consecutive patients with sCADs. In five of the seven patients (71%), histologic, immunohistochemical, and ultrastructural changes similar to EDS were found, again suggesting that sCADs are associated with dermal alterations seen in EDS.⁶⁷

NF1

While no definite association between NF1 and extracranial dissections has been established, a number of case reports have been published on this topic.^68–70 It is reasonable to assume that the vasculopathy resulting from NF1 could predispose these patients to cervical dissections.

LDS

In the neuroradiologic study by Rodrigues et al. of 25 patients with a positive genetic testing for LDS, three had dissections of the carotid and vertebrobasilar arteries. However, the exact number of an intracranial or extracranial nature is not described.⁴⁹ Because LDS patients are prone to formation of aortic aneurysms and dissections, some cervical dissection cases can result from extension of a Stanford type A dissection, similar to the pattern seen in Marfan syndrome.⁷¹

Management of cervical dissections

Guidelines for management of patients with extracranial carotid and vertebral artery dissections were published by the American Heart Association in collaboration with the American Academy of Neurology and Society of Cardiovascular Computed Tomography.⁷² The guideline first recommends that a confirmatory diagnosis of cervical dissection be made using contrast-enhanced computed tomography angiography (CTA), MRA, or catheter-based angiography. Once confirmed, the first-line treatment for cervical dissections associated with ischemic stroke or transient ischemic attack (TIA) centers around medical antithrombotic treatment with anticoagulation or antiplatelet therapy for at least three to six months. However, it is of interest to note a recently published randomized control study of 250 patients with carotid and vertebral artery dissections found no differences in efficacy of antiplatelet and anticoagulation drugs at preventing stroke and death in patients with symptomatic carotid and vertebral artery dissections.⁷³ In general, invasive procedures such as carotid artery stenting are not advocated in the treatment of cervical dissections in the general population and would likely be even less warranted in the connective-tissue disease population because of the high risk of iatrogenic injury from friable vasculature. The effectiveness and safety of therapy with a-adrenergic antagonists, angiotensin inhibitor, or nondihydropyridine calcium channel antagonists to lower blood pressure are not well established.

Cervical pseudoaneurysms

Cervical pseudoaneurysms arise from a disruption within the arterial wall that allows blood to stream into the surrounding tissue to form a sac in communication with the arterial lumen, and generally result from dissections. Such dissections are primarily the result of neck trauma, iatrogenic vascular injury, or microvascular procedures in the general population. More often, it is after the dissection has healed that the pseudoaneurysm takes place between the tunica media and adventitia.

Several case reports have been published documenting the presence and treatment of cervical pseudoaneurysms in patients with inheritable connective-tissue diseases.^74–77 In the general population, pseudoaneurysms of the extracranial carotid artery are very rare, and account for less than 1% of all arterial aneurysms. These are generally not prone to rupture because of the thick layer of adventitia seen in the cervical carotid vasculature. Conservative management with antiplatelet therapy and serial imaging is generally the preferred treatment modality; however, both surgical and endovascular intervention can be used in the treatment of high-risk cervical pseudoaneurysms.⁷⁸ In general, the rupture risk of these pseudoaneurysms is very low. The main risks associated with these lesions are acute ischemic stroke and mass effect. Endovascular repair techniques reported include stenting with flow diversion or covered stents and/or coil embolization. Open surgical repair of these aneurysms through the use of interposition grafts has also been reported (Figure 5).

Figure 5.

A 38-year-old female with neurofibromatosis type 1 (NF1) presented with a large pulsatile mass in her right neck. (a) Computed tomography angiography (CTA) showed multiple pseudoaneurysms of the left internal carotid artery (ICA). (b) A cerebral angiogram was performed for balloon test occlusion prior to surgical repair of the pseudoaneurysms. The neck angiogram demonstrated the large ICA pseudoaneurysms. (c) Cerebral angiography with a balloon in the right ICA and left ICA injection shows filling of the right middle cerebral artery (MCA) and anterior cerebral artery (ACA) territories across the patient’s anterior communicating artery (Acom). (d) Surgical resection of the pseudoaneurysms with placement of a saphenous vein interposition graft was performed with no complications. Postoperative CTA shows patency of the saphenous vein graft.

Other neurovascular manifestations

NF1 and Moyamoya disease (MMD)

The occurrence of MMD in NF1 is rare, but encountered in the form of various case reports of MMD in pediatric populations.^79–89 A retrospective analysis of 197 children with NF1, 168 of whom underwent cranial MRI, found that four (2.3%) were diagnosed with MMD. It is suggested that the neurofibromin in the endothelium and smooth muscle of the blood vessels in these patients may play an important role in the development of this vascular abnormality.^24,90,91 MMD in pediatric patients with NF1 is usually unilateral and often involves anterior vascular territories.⁹² Although most patients are asymptomatic, subsequent clinical and radiologic worsening is likely to occur with events such as TIAs, infarcts, seizures with headache, and intracranial hemorrhage being the consequence.^90,93,94 For this reason, the literature urges that although MMD may be generally rare in NF1, MRA or other imaging modalities be used because of MMD and its asymptomatic and progressive nature.^91,95,96

A retrospective review of 39 patients with both NF1 and MMD, 32 of whom underwent surgical revascularization with pial synangiosis and 18 of whom had radiographic evidence of prior stroke at the time of MMD diagnosis, concluded that the clinical, radiographic, and angiographic features of MMD seen in the NF1 population are comparable to primary MMD.⁹⁵ The authors of this study recommended treatment of children with MMD and NF1 with surgical revascularization. Not only is this method shown to be safe in NF1 patients, it is also protective against further ischemic and neurological damage, with a 27-fold reduction in stroke rate.⁹⁷ In addition, the literature contains several cases of later-onset MMD in adult populations.^98–104 An example of MMD associated with NF1 is provided in Figure 6.

Figure 6.

A 23-year-old female with neurofibromatosis type 1 (NF1) presented with multiple episodes of left-sided hand symptoms thought to be related to opercular hypoperfusion. (a) Fluid-attenuated inversion recovery magnetic resonance imaging (FLAIR MRI) shows some high signal in the sulci of the right operculum and frontal lobe. (b) Right internal carotid artery (ICA) injection shows a high-grade stenosis of the supraclinoid right ICA and marked prominence of the middle cerebral artery (MCA) lenticulostriates. There is complete obliteration of the right anterior cerebral artery (ACA). (c) Left ICA injection shows mild narrowing of the supraclinoid left ICA and obliteration of the left A1. There is minimal, if any, prominence of the lenticulostriates on this side. (d) Postoperative cerebral angiogram following right superficial temporal artery to middle cerebral artery (STA-MCA) bypass shows excellent filling of the right MCA territory from the STA bypass.

A recent population-based, case controlled study addressed the risk of cerebrovascular events requiring hospitalization among patients with NF1.¹⁰⁵ Of the 21,378 patients diagnosed with NF1 between 1998 and 2009 in this study, NF1 was associated with a younger mean age (41 versus 48) of strokes, despite a lower prevalence of stroke risk factors in adults. Although pediatric and adult NF1 patients were significantly more likely to be diagnosed with any stroke than the general population (1.2:1), the odds of ischemic stroke (3.4:1) and intracerebral hemorrhage among the hemorrhagic stroke group (8.1:1) were dramatically elevated in the pediatric population. In adults, the odds of intracerebral hemorrhage among hemorrhagic strokes (1.9:1) were also elevated. Thus, this study concluded that the likelihood of any type of stroke is significantly increased in NF1 patients compared to the general population.

NF1 and vertebro-vertebral fistulas

Although vascular involvement in the setting of NF1 is commonly reported, coexistence with vertebral arteriovenous fistulae (AVF) is rare. There are several case reports indicating AVF in NF1 patients.^70,106–128 There are no systematic studies to date conducted on the prevalence and risk factors of AVF in NF1 patients. However, the association seems to be strong enough that every patient presenting with a rare vertebral-jugular fistula should be suspected to have NF1. These fistulae often form as a result of a vertebral artery pseudoaneurysm rupturing into the surrounding venous plexus. Fistulae between muscular branches of the vertebral artery and the surrounding venous plexus can also occur.

The primary method of treatment is endovascular embolization with coils or liquid embolic agents.^{112,113,117,120,121,129} A method that has shown lower success rates is the trapping of fistulas, owing to numerous potential feeding arteries that can complicate the procedure.¹⁰⁸ In some more complicated cases, direct surgery, or a combination of surgery and endovascular embolization, may be necessary.^{111,116,119,123–127}

EDS and Marfan syndrome and CCF

In a study conducted by Pepin et al., 44 out of the 419 patients with EDS IV had arterial complications of the central nervous system, with the most frequent being CCF.^40,130 There are several case studies that present CCF in EDS,^131–138 with some discussing difficulties and death associated with the treatment of EDS IV patients with extreme vessel fragility.^{130,139–149} CCF have also been reported in the setting of Marfan syndrome.

Because of the high risk associated with endovascular treatment of CCF in EDS IV patients, it is recommended that all spontaneous CCF cases be approached with the possibility of EDS IV in mind.¹⁵⁰ Given the well-known vascular fragility of these patients, even a simple catheter angiography can lead to disastrous consequences in EDS IV¹⁵¹ and therefore should be considered only if endovascular therapy is considered. As an alternative to transfemoral endovascular treatment, the ipsilateral carotid artery and internal jugular vein can be surgically exposed for direct insertion of endovascular sheaths. This allows for a reduction in risks associated with arterial dissections that often accompany transfemoral access in EDS IV patients.^{133,135,143,152} Asymptomatic CCF can be managed conservatively through manual compression of the medial canthus (Figure 7).

Figure 7.

Carotid-cavernous fistula (CCF) in a 47-year-old male with Marfan syndrome. ((a) and (b)) Time of flight magnetic resonance angiography (TOF MRA) shows high signal in the right cavernous sinus and marked enlargement of the superior ophthalmic vein, findings consistent with a CCF. (c) Right internal carotid artery (ICA) cerebral angiography shows a large direct CCF with early filling of the cavernous sinus and engorgement of the superior ophthalmic and facial veins. (d) Note the marked tortuosity and dolichoectasia of the left vertebral artery in this patient.

Conclusions

Although there are various case studies reporting neurovascular findings in connective-tissue diseases, there is a general lack of case control studies investigating the exact prevalence of these findings in these patient populations. Furthermore, little is known regarding the risks and benefits of various conservative, endovascular, and surgical management techniques in caring for patients with these lesions. Large multicenter registries may be helpful in the future in better characterizing the pathophysiology, prevalence, and ideal treatment options of neurovascular lesions in patients with connective-tissue diseases.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.