Neurosurgical Review

, Volume 31, Issue 2, pp 131–140

Dissecting aneurysms of the vertebrobasilar system. A comprehensive review on natural history and treatment options


  • Jorge Arturo Santos-Franco
    • Department of Neurological Endovascular TherapyNational Institute of Neurology and Neurosurgery
    • Department of Neurological Endovascular TherapyNational Institute of Neurology and Neurosurgery
    • Comprehensive Stroke CenterHospital Ángeles del Pedregal
    • Instituto Nacional de Neurología y Neurocirugía
  • Angel Lee
    • Comprehensive Stroke CenterHospital Ángeles del Pedregal
    • Department of NeurosurgeryInstituto Nacional de Ciencias Médicas y Nutrición Salvador Zubirán

DOI: 10.1007/s10143-008-0124-x

Cite this article as:
Santos-Franco, J.A., Zenteno, M. & Lee, A. Neurosurg Rev (2008) 31: 131. doi:10.1007/s10143-008-0124-x


Vertebral artery dissection has been recognized as an uncommon cause of ischemic stroke. However, it is less well known as a cause of subarachnoid hemorrhage. Even if dissecting aneurysms of the vertebral artery are rare, their importance arise from their high morbidity and mortality with rebleeding occurring more often than in cases of saccular aneurysms. Dissecting aneurysms of the vertebrobasilar system are a complex entity which requires a rapid and effective treatment to prevent rerupture. The sole stenting technique stands as a promising approach, allowing to occlude the aneurysm while preserving the vessel patency and reconstructing the diseased segment.


Intracranial aneurysmsVertebral artery aneurysmDissecting aneurysmSole stentingEndovascular treatmentSubarachnoid hemorrhage

Epidemiology and anatomical pathology

The annual incidence of the vertebral artery (VA) dissection is about 1/100,000 and 1.5/100,000 in America and in France, respectively [23, 56, 57, 74, 76]. The age of onset is slightly higher than the one of carotid dissection, as the latter dissection occurs in the second or third decade of life, whereas VA dissection appears in the fourth [57, 59, 65].

VA dissections may be classified into intracranial and extracranial, the latter classification being more common [50, 52, 65]. The most frequently involved segments are the first (V1) and the third (V3) as they are highly mobile, while the second (V2) is fixed within the foramina transversaria and the fourth segment (V4) is less mobile, as it is fixed when piercing the dura. The dissecting aneurysms of the vertebral artery (DAVA) are sometimes termed as “pseudoaneurysms” and represent around 28% of the posterior circulation aneurysms and 3.3% of all intracranial aneurysms [12, 37, 42, 68]. Considering all the patients who underwent a surgical or an endovascular treatment of an intracranial aneurysm at our institution in the last 2 years, the dissecting aneurysms of the VA account for 4% of the total and 10% of the endovascularly managed aneurysms.

Subarachnoid hemorrhage (SAH) caused by DAVA explains less than 10% of all causes of nontraumatic SAH [2, 56], with a reported mortality ranging from 19% to 83% [52]. Intracranial dissections cause more generally a SAH, while extracranial dissections produce ischemia [8, 22, 40, 49, 77]. The intradural VA is more susceptible to rupture than the extradural portion because of its thicker internal elastica lamina, lack of external elastica lamina, thinner adventitia, and fewer elastic fibers in the media [23, 52]. In our series, 73% of DAVA is revealed by SAH, whilst 27% of them gave bulbar or cerebellar ischemia. Although they are rare, some cases of SAH by a DAVA of the extracranial portion have been reported [23, 41, 78]. In our series, we have encountered one such case, with a high peculiar pattern, as the aneurysm arose at the V1 segment, having no particular explanation for this. The patient had a complete radiologic workup including spinal angiography, which discarded any other cause of SAH [83].

DAVA arises in the space between the internal elastic lamina and the media: a defect in the internal elastic lamina has been regarded as the primary triggering factor, as this layer normally determines the strength of the vessel’s wall [2, 18, 44, 56, 58, 75]. Sasaki et al. have established an interesting classification of vertebrobasilar system dissection into two histological groups which have clinical implications [56]. Group 1 includes dissections limited to V4 and group 2 extends farther into the basilar artery (BA). In group 1, the intramural hematoma arises within the media or in the space between the media and the adventitia, frequently giving birth to a DAVA [19, 56, 76]; the rupture of the three layers of this lesions often cause a massive SAH. In the lesions of group 2, the hematoma is usually subintimal, occluding the vessel lumen and causing, therefore, more often brainstem ischemia rather than SAH. However, attention should be drawn on the fact that in some cases, lesions of group 2 can bleed. After an initial ischemic presentation, a SAH can occur several weeks later or even months later [56], making of this condition a formal indication for a well-timed management. It could be risky to consider clinical management alone the treatment of choice of group 2 lesions, as risk factors for bleeding cannot be anticipated on the basis of angiographic aspect alone.

Clinical features and natural history

These aneurysms frequently rupture causing a massive SAH with devastating neurological consequences [1, 2, 30, 37, 43, 52, 56, 76]. The rebleeding rate is particularly high, arising to 30% [2, 52, 58]. In a retrospective study including 42 patients, Mizutani et al. recorded a rebleeding rate of 40.5% at 24 h and 57.1% within the first week of the initial bleeding [43]. Yamada et al. reported on the conservative management of 24 patients with a DAVA: 58% of them rebled (14 patients) and 11 patients died [73]. Takagi et al. reported a rebleeding rate of 37% in 62 patients, most of them within the first 24 h of the opening episode [64]. The wall of these aneurysms is extremely thin and friable and, as shown in surgical procedures and autopsy series [2, 18, 19, 39, 43, 76] and among other factors, this finding could explain this high rerupture frequency. The healing process is long, as shown in angiographic and in autopsy studies, and starts by a neointimal hyperplasia, which begins 1 week after the rupture and lasts for a month. The duration of this healing process might also explain the high rebleeding rate [2, 19]. Some aneurysms are angiographic findings without clinical evidence of rupture, and, in some cases, they are incidental findings [3, 12, 30, 61]. Even if some cases of spontaneous involution have been reported, they often tend to expand [1, 2, 4, 8, 58], sometimes manifesting with compressive symptoms and/or posterior fossa ischemia. These lesions have a dynamic evolution with high morbidity and mortality rate and can either rebleed or expand; they must, therefore, be treated as soon as possible [24, 30]. On the contrary, nonaneurysmal VA dissections tend to be managed conservatively [8, 14, 38, 45, 58, 65] even if our team disagrees with this passive attitude toward VA dissections, as we shall discuss in a future communication.

Imaging findings

The gold standard in diagnostic imaging is still represented by digital subtraction angiography (DSA). Classical findings are: pearl and string, double lumen, rosette, or a simple fusiform dilatation, in addition to a delayed clearance of the dilatation or a false lumen [1, 15, 24, 30, 61, 62]. The string and pearl sign is nothing else but a dilatation (pseudosaccular or fusiform) proximal or distal to a narrowed segment (Figs. 1a and b, 2a–c and 3a–e), while the double lumen is a double channel image, due to the visualization of both the true lumen of the vessel and the false lumen which causes the dissection (Figs. 4c and 5b). The rosette pattern is similar to string and pearl but with a multilobulated aneurysm (Fig. 6). In our series, 91% of the cases displayed a string and pearl pattern, while 18% and 27% also presented with a fusiform pattern and double lumen sign, respectively [83]. We had only one case of a rosette pattern aneurysm.
Fig. 1

AP (a) and lateral (b) DSA views of the right VA. The DAVA is clearly seen (hollow arrow in a and b), as well as the pre- and post-aneurysmal stenosis (thin black arrows in a and b). A severe vasospasm of the BA and a fenestration can also be appreciated (arrowhead in a)
Fig. 2

DSA views of the left VA. Diagnostic 2D runs (a and b) and pretreatment (c) and post-treatment 3D runs (d). A fusiform dissecting aneurysm (thick black arrow in ac) with pre- and post-lesional stenosis is visualized (thin arrows in b and c). The 1-year control 3D angiogram shows both the healing of the aneurysm and the correction of the stenosis (thin arrows) with integrity of the distal flow. A balloon-expandable stent was used
Fig. 3

2D-DSA selective diagnostic run of the right VA (ac), MIP reconstruction angio-CT (d) and volumetric reconstruction of the VB system (e). One-year post-treatment follow-up angiograms: 2D-DSA (f) and 3D-DSA (g) runs. A string and pearl image shows a pre-aneurysmal stenosis (thin arrowsae) and a dissecting aneurysm of the BA (hollow arrowsae). The 1-year control 2-D and 3-D angiograms show both the healing of the aneurysm and the correction of the stenosis with integrity of the distal flow. The fully deployed stent (thin arrows in g) was a balloon-expandable stent. Notice how the arteries emerging from the diseased and stented segment are preserved (arrowheads in g)
Fig. 4

Pre-treatment imaging of a right DAVA: 2D-DSA (ac), angio-CT images (d: native, e: MIP reconstruction images). Post-treatment control 2D-DSA at 1 year (f). Several features can be noticed: fusiform pattern DAVA (hollow arrowsae), predilatation stenosis (ac), double lumen image (black arrow in c). Contrast-enhanced axial CT demonstrates the focal dilatation of the right VA. A complete occlusion (and no recurrence) of the aneurysm is shown in the stented segment by 1-year follow-up 2D-DSA (thin arrows in f). Notice how one artery emerging from the diseased and stented segment is preserved (arrowhead in f)
Fig. 5

2D-DSA pretreatment imaging of a right V1 segment DAVA (a), pretreatment angio-CT MIP reconstruction (b) compared to post-stenting imaging (c) with the same technique. A dissecting aneurysm is clearly seen in V1 segment (hollow arrow in a) with a double lumen image (thin arrow in a and b). After the stent placement (thin arrows in c), both the double lumen and the aneurysm have disappeared
Fig. 6

Illustrative case of a rosette pattern. 3D-DSA of the right VA. A multilobulated aneurysm (hollow arrow) is followed by a post-lesional stenosis (thin arrow)

We might summarize imaging findings of DSA in VA dissection like this: (1) dilatation (fusiform or rosette), (2) stenosis and occlusion (tapered or abrupt occlusion, prelesional stenosis and post-lesional dilatation, string sign), (3) coexistence of dilatation and stenosis (string and pearl sign), (4) double lumen images (intimal flap). This imaging modality has another advantage, as it can visualize either spontaneous progression or resolution of the dissection (Fig. 7a compared to b).
Fig. 7

Illustrative images of the progression of a dissection of the VA from one day to another (no aneurysm in this case). 3D-DSA of the VA (a: day 0, b: day 1) with immediate post-stenting control 3D-DSA (c). The stenosis is greater at day 1 (thick arrow in b) than the previous day (thick arrow in a). Dual reconstruction of the DSA shows the stent (thin arrows in c) with healing of the stenosis

The computed tomography angiography (CTA) is also a useful tool in the diagnosis of this disease, with some images close to those displayed in a DSA. We might summarize imaging findings of CT and magnetic resonance angiographic (MRA) techniques in VA dissection like this: (1) a focal dilatation (native images of CTA, maximum intensity projection (MIP) of CTA and MRA; Figs. 3d and e, 4d and e and 8a–e), (2) an intimal flap (MIP of MRA or CTA; Fig. 5b), (3) an increase in the outer diameter of the vessel due to the intramural hematoma (yielding an eccentric increase in intensity signal in T1- and T2-weighted images; Fig. 8a and b), (4) a narrowing or an occlusion of the vessel lumen (better seen in CTA and MRA MIP images; Figs. 3d and e, and 8d and e). We should however point out that T2-weighted images might not be reliable enough as a misleading can occur and get confused with cerebrospinal fluid signal, even if it remains an excellent tool to determine if there is a decrease of flow or flow void.
Fig. 8

Multimodality imaging of a right DAVA. The intramural hematoma is visualized as a hyperintense signal in T1-weighted images (red thick arrow in a) and in T2-weighted images (red thick arrow in b). Angio-CT images with native images (c), MIP reconstruction (d) and volumetric reconstruction (e). An increase in the size of the VA is clearly demonstrated (red thick arrows in a and b, hollow arrow in c). The reconstruction of the angio-CT show the aneurysm (hollow arrows in d and e) as well as a severe postdilation stenosis (thin arrows in d and e)

Finally, all imaging findings may have a predictive value. For instance, when considering imaging aspects, Takagi et al. found in a multivariate analysis the following DSA criteria as independent risk factors for rerupture: predilatation stenosis, lateralized dilatation, and a pre or in-posteroinferior cerebellar arteries (PICA) location [64].

Definitive treatment

The management of DAVA is a technical challenge due to their histopathological features and their localization [16, 17]. Among the surgical strategies employed, we could cite the trapping of the aneurysm, the proximal ligation of the VA, or the wrapping of the aneurysm. The two latter techniques have not shown their efficacy in comparison with the trapping of the lesion in preventing rebleeding of the aneurysm [2, 6, 19, 23, 43] as this trapping prevents any retrograde flow into the lesion. A high-flow bypass has been preconized to be a good solution when the occlusion of the parent vessel is mandatory and collateral supply is poor [2, 8, 27, 51], especially because treatment of pre-PICA lesions may induce a cerebellar infarction if the VA is occluded in cases of inadequate collateral supply [81, 83]. With its evolving, endovascular techniques are more preferred in some cases of DAVA because surgical trauma is avoided; complete workup of collateral supply is possible, leading to its preservation; probability of ischemia due to the surgical therapeutic approaches decreases; and it is less time-consuming [23, 30, 37, 43].

Among the different causes of morbidity and mortality or reasons for surgical failure, we could mention an unnoticed damage to tiny perforator branches during dissection or clipping of the aneurysm and the nonperfusion of these branches by an insufficient retrograde flow, leading to nonfatal and fatal ischemic lesions of the brainstem [19, 23, 43]. Concerning this very last point, the anticoagulation and/or the antiplatelet regime needed in endovascular procedures prevents, somehow, the thrombosis of these perforating vessels. On the contrary, the use of antithrombotic medications is impossible in open surgery [43].

Since their origin, the interventional radiological techniques have greatly developed [5, 32, 47, 48, 53], and two different categories of therapeutic approaches in DAVA can be mentioned: the so called deconstructive techniques and reconstructive techniques [1, 16, 30, 37, 43, 61]. The first consist substantially in a parent vessel occlusion, and the second involve the treatment of the lesion with preservation of the parent vessel.

Endovascular deconstructive procedures have been inspired by the surgical experience in vessel occlusion and involve the use of detachable balloons or Guglielmi detachable coils (GDC) [2, 3, 911, 23, 24, 27, 33, 34, 47, 51, 55, 72, 74]. The occlusion of a vessel requires a previous balloon test occlusion and triggers the problem of lesions whose treatment might involve the origin of the PICA or of the anterior spinal artery with insufficient collateral supply. Even if some of the deconstructive techniques have been widely used and fair good long-term results have been described, severe and frequent complications have also been reported, including ischemic lesions of the brainstem or retrograde recanalization of the aneurysm [27, 28, 30, 34, 46, 75]. Reconstructive techniques have recently and progressively been developed. In spite of this, many authors still report their experience to parent vessel sacrifice, yield excellent results (prevention of further rupture and ischemic events), and consider the latter approach as a first option if allowed by collateral supply [2, 16, 24, 39, 47, 52, 60, 63].

Reconstructive techniques foster the thrombosis of the aneurysm while preserving, or even improving the flow in the parent vessel [1, 31, 37]. From a conceptual point of view, reconstructive techniques are more solid than deconstructive ones as the future behavior of the noninvolved VA or of the collateral supply is not predictable [30]. Following the past experience in endovascular occlusion of intracranial aneurysms, the best option to occlude an aneurysm is the use of GDC [79]. However, the wide neck or the fusiform pattern of some of the DAVA [19, 56, 76] holds back the use of this technique; therefore, the use of intracranial stents has gained acceptance and has tried to shape a better neck and prevent coil herniation causing occlusion of the parent vessel, thus, reconstructing the dissected vessel [25, 26, 36, 68, 71]. Notwithstanding these considerations, some authors still regard the stent-assisted coiling as an alternative option to parent vessel sacrifice, or even as a third option when the collateral supply from the contralateral VA is insufficient, esteeming the occlusion of the parent vessel plus a surgical bypass to be the gold standard of management in such cases [2, 24, 27, 39, 51].

These last years, we have worked on a notion that might not be brand new but which had neither been thoroughly conceptualized nor clinically applied: the exclusive use of a stent in endovascular management of aneurysm, or sole stenting technique [8083], as we have named this technique.

In early experience, computerized and in vitro experimental models of aneurysms which underwent a stent placement seem to promote some of the following effects: (1) redirection and deviation of flow into the parent vessel away from the aneurysm, (2) blood stasis, and (3) sluggish intraneurysmal vortex motion [20, 21, 6870]. In a later phase in animal models, the stents have confirmed their ability to shift the flow pattern within the sac form a helical pattern to a noncoherent (turbulent) pattern, a phenomenon which may induce thrombosis [13, 66, 78] and later may lead to a permanent occlusion of the aneurysm, which might be explained by the above-described hemodynamical changes [66, 69, 70]. Finally, once the intra-aneurysmal thrombosis has occurred, the sprouting of the neointima over the luminal surface of the stent leads to a remodelling of the stented arterial segment. Currently, the use of this sole stenting technique in human patients has been described [80, 82, 83]. The findings of these previous experiences in sole stenting may let us think that the stent alone can provide some protection against further rupture, explained by the hemodynamic changes within the aneurysm [1, 4, 7, 13, 16, 29, 31, 35, 37, 42, 54, 67, 78]. In a previous article, we published our initial experience on sole stenting at our institution in V4 aneurysms and included the experience of other groups, emphasizing on the efficacy and safety of the technique [80]. Our series is a prospective study on posterior circulation aneurysms treated by this technique which is about to be published [83], including 11 dissecting aneurysms, (Figs. 2d, 3f & g, 4f and 5c) representing a 55% of the total. In this subgroup, 73% presented as a SAH (eight patients) and 27% with a brain stem or cerebellar infarction (three patients). The reason for scheduling these patients for this new technique is twofold: (1) the coiling of this thin-walled lesion could cause rupture as there are weak and fragile aneurysms; 2) the preservation of the parent vessel is a requisite whenever possible. All our patients have complied with established protocols by our institutional Committee of Ethics and are in accordance with the Declaration of Helsinki.

Considering the cases with a previous SAH, 75% of them (six cases) were treated 2 weeks after the initial episode and the rest were treated within the first 15 days (one of them within the first 3 days). Hydrocephalus was present in four cases (36%) and half of them required shunting before stenting. All the 11 cases were treated with a balloon-expandable stent, and no complication occurred.

Of course, the imaging goal is the occlusion of the aneurysm with preservation of the parent vessel. The DSA demonstrated an immediate postprocedural complete occlusion in four cases (36.4%) and sluggish intra-aneurysmal vortex motion in five cases (45.5%). Subsequent imaging follow-up at 3, 6, and 12 months showed progression of the occlusion in a further percentage of cases—64% (n = 7), 73% (n = 8), and 91% (n = 10). The parent vessel was kept patent in all cases at all times. The clinical purpose is prevention of rebleeding, which never happened during the first year of follow-up. Only one case (9%) of spontaneous recanalization occurred, which was further successfully coiled. Another series of 2006, published by Ahn [1], includes nine DAVA treated by sole stenting. Complete occlusion is obtained in four cases and near-complete occlusion in five. No procedural complication happened and no further bleeding of these lesions occurred during the follow-up period (6 to 12 months). In a recently published article, Peluso et al. reported on the endovascular treatment of symptomatic DAVA (14 lesions) [47]. In most of cases, the treatment consisted of coil occlusion of the arterial diseased segment. In one patient, stent placement over the aneurysm as the only therapy was done; follow-up angiograms after 6 months showed complete occlusion of the aneurysm with a patent right VA and PICA.

These initial results uphold the safety and effectiveness of the technique. Anyhow, it may be overhasty to assert that the sole stenting technique is a panacea for the management of DAVA or of posterior circulation or even of intracranial aneurysms.

Larger, prospective, or randomized series are warranted to establish accurate criteria for the selection of appropriate cases for surgical management (clipping of the aneurysm and trapping of the diseased segment with or without bypass) or endovascular treatment (coiling, endovascular trapping, stenting-assisted coiling, or sole stenting). Watchfulness should be kept on DAVA starting with an ischemic event, and in account of the unpredictable risk of rupture, we deem that the conservative management is a risky attitude.


DAVA represents a significant group within all the aneurysms of the posterior circulation. Their rupture is a major source of morbidity and mortality in all the reported series, and their rerupture is far more frequent than in saccular aneurysms. As a matter of consequence, they must be managed as soon as possible. The endovascular approach is a safe and effective procedure even if no consensus has been reached concerning the most suitable technique for their endovascular occlusion. Overall, reconstructive techniques are more adequate as they have both targets in mind: to treat definitively the vascular anomaly and to preserve the parent vessel. To attain this goal, the intracranial stents are to become a useful tool.

Copyright information

© Springer-Verlag 2008