Article Text

Download PDFPDF

Non-saccular vertebrobasilar aneurysms and dolichoectasia: a systematic literature review
  1. Maksim Shapiro,
  2. Tibor Becske,
  3. Howard A Riina,
  4. Eytan Raz,
  5. Daniel Zumofen,
  6. Peter K Nelson
  1. Bernard and Irene Schwartz Neurointerventional Radiology Section, NYU Langone Medical Center, New York, New York, USA
  1. Correspondence to Dr M Shapiro, Bernard and Irene Schwartz Neurointerventional Radiology Section, NYU Langone Medical Center, 660 First Avenue, 7th Floor, New York, NY 10016, USA; neuroangio{at}, maksim.shapiro{at}


Background and objective Treatment of non-saccular vertebrobasilar aneurysms remains highly challenging despite significant recent advances in endovascular techniques. Establishing the natural history of this heterogeneous disease, as best as currently available data allows, is crucial to help guide counseling and management.

Methods A review of the literature was conducted to identify publications describing the presentation and natural history of vertebrobasilar dolichoectasia and non-saccular aneurysms.

Results Nine studies of 440 patients met the analysis inclusion criteria. The majority of patients presented with ischemia, mass effect, or incidentally; hemorrhage was uncommon and overlapped with the population of vertebrobasilar dissection. Overall mortality was ∼40% after 7 years of follow-up, with 43% of these deaths resulting from non-neurologic causes. Neurologic course was dominated by ischemic stroke rather than hemorrhage. Mass effect prognosis was especially poor, with 40% mortality after ∼4 years. Incidentally discovered lesions which remain morphologically stable have a favorable long term course.

Conclusions Initial clinical presentation is a strong predictor of subsequent disease course. Although overall prognosis is poor, nearly half of all deaths resulted from non-neurologic causes, underscoring the importance of comprehensive medical management. Aneurysms characterized by expansion, established mass effect, or hemorrhage have a poor natural history, and may be considered for invasive treatment, which is increasingly endovascular in nature. Lesions presenting with ischemia or incidentally are likely best addressed with aggressive neurologic and overall medical management.

  • Aneurysm
  • Artery
  • Hemorrhage
  • Posterior fossa
  • Stroke

Statistics from


Recent advances in endovascular treatment of previously inaccessible or very high risk anterior circulation aneurysms1 ,2 generated renewed hope for the possibility of a reasonably safe and effective solution to the subpopulation of complex, non-saccular aneurysms in the posterior circulation. This initial enthusiasm was tempered by early reports of mixed treatment results,3–8 prompting reflection on the nature of this complex disease, particularly in regard to treatment risk–benefit calculus. While much remains to be determined in the arena of treatment safety and efficacy, we feel that useful information may be extracted from an updated review of the disease natural history. By all indications, the neurovascular community entertains an especially divergent set of opinions in regard to every aspect of this condition. This state of affairs attests to the heterogenous nature of vertebrobasilar aneurysmal lesions, such that practitioners, depending on their location, referral pattern, and specialty identities, may be encountering subsets of the overall mix, thus arriving at appropriate situation specific conclusions which however may not be applicable to the overall spectrum. The same considerations also apply to the currently available vertebrobasilar observational literature. A widely held view recognizes the existence of pathophysiologic differences among various non-saccular vertebrobasilar aneurysms, separable into dissecting and non-dissecting types.9–12 To what extent, however, these categories are in fact mutually exclusive or reliably demonstrable based on in vivo imaging methods, rather than post mortem investigations, is unclear. The concept of dissecting aneurysm, furthermore, broadly overlaps with the entity of vertebrobasilar dissection, leading investigation into yet another realm of heterogeneous literature, with its own uncertainties. Nevertheless, the perceived differences between vertebrobasilar aneurysmal diseases form the basis of exclusion and inclusion criteria within individual publications. The following review was undertaken with the purpose of establishing, based on currently available evidence, the incidence and prognosis of vertebrobasilar disease based on mode of initial clinical presentation rather than presumed pathophysiology, to best guide patient counseling and possible treatment recommendations.


PubMed, Medline, and Google literature search utilizing search strings ‘fusiform basilar aneurysm’, ‘vertebrobasilar dolichoectasia’, ‘vertebrobasilar aneurysm’, ‘dissecting vertebral aneurysm’, and ‘dissecting basilar aneurysm’ were carried out by a single author (MS), and the bibliography of identified sources reviewed for additional publications. From the overall group, articles concerned with observational data or non-surgical management were identified and reviewed. Publications with less than five subjects were excluded. All studies were required to contain quantitative information on the total number of subjects, initial presentation, and length of follow-up. Most publications also listed cause specific mortality. Calculations of baseline and follow-up incidence of hemorrhage, stroke, mass effect, mortality, and others were made for the overall cohort from data contained in the individual publications. Cumulative incidence of follow-up stroke, hemorrhage, overall and cause specific mortality were calculated based on follow-up times to event. Annual event incidence was calculated as a ratio of cumulative incidence and follow-up duration. The more informative Kaplan–Meyer survival curves cannot be calculated for the overall dataset. No analyses of statistical significance were felt to be applicable given the diverse nature of the patient population and study methodologies. Our medical center does not require institutional review board approval for literature review projects.


Nine studies, totaling 440 cases, were identified and are listed individually in table 1, while overall presentation and follow-up results are shown in tables 24. Over 70% of the data (315 patients) were contained within the Siena cohort of Passero et al13 and the Mayo Clinic cohort, with findings reported in three separate publications.14–16 These two populations are further distinguished by greatest length of follow-up in comparison with the remaining pool, with average follow-up of 12.9 years for the Siena group and 4.4 years for the Mayo group. In contrast with the Mayo group, which included patients defined as harboring dolichoectasia or some form of non-saccular vertebrobasilar aneurysm, the Siena cohort focused specifically on what was defined as non-aneurysmal vertebrobasilar dolichoectasia; both populations overlap and qualify for analysis.

Table 1

Representative index, comorbidity, and follow-up data from selected studies

Table 2

Presentation and comorbidity

Table 3

Follow-up data

Table 4

Follow-up mortality data

Prevalence of disease

Prevalence of disease was difficult to reliably define based on the methodology of the available studies. Estimates ranged from 0.3% to 4.4% (average 0.6%) among patients coming to neurologic attention at tertiary centers, with clinical concerns not necessarily related to vascular pathology.17

Initial presentation

Four groups were identified: incidental, ischemic stroke (distal embolic and local thrombotic), mass effect, and subarachnoid hemorrhage. Ischemia, mass effect, and incidental modes of discovery each accounted for ∼30%, with hemorrhage (subarachnoid and/or parenchymal) encountered in the remaining 7%. Stroke was further subdivided into distal embolic and local thrombotic (perforator occlusion) groups in some publications.18 Mass effect may be separated into brainstem and local peripheral nerve compression,19–21 in addition to hydrocephalus.20 Larger aneurysms were predictably more likely to present with both stroke and mass effect in comparison with smaller ones.13

Multiple groups reported a substantially higher hemorrhage incidence for both acutely dissecting aneurysms22 and vertebrobasilar dissections23 in smaller studies. For example, Pozzati computed a 50% incidence of hemorrhage in a group of 16 dissecting aneurysms.24 In Mizutani's landmark series of 80 patients, which established acute dissecting aneurysm etiology based on histopathologic findings, all 65 hemorrhagic cases were considered to be dissecting, although only four of these patients came to autopsy.9 The incidence of hemorrhage is expected to further vary substantially based on specialty, professional focus, and referral patterns.

Incidental discovery was seen in 30% of cases, and was enriched in the dolichoectatic and smaller aneurysm cases.16

Demographics and comorbidities

In contrast with saccular aneurysms, the majority of patients were men (70%). The impact of age and gender on mode of presentation and outcome was not possible to determine from the available publications, as this would require age and gender specific breakdown for both the initial presentation and outcome. There was an overall high incidence of hypertension (66%) and a moderate incidence of smoking (41%). Other cardiovascular risk factors appeared to be relatively infrequent. Several papers18 ,20 ,21 focused specifically on the medical treatment component.


The overall prognosis was 43% mortality after 7.1 years of follow-up, corresponding to a simple algebraic 6% annual mortality. Nearly half of all deaths (43% of all deaths, 2.6% annualized mortality) were due to non-neurologic causes, which varied in etiology, with no overwhelming cardiovascular driver.13 ,16 ,17 Stroke was the most common overall cause of neurologic demise.

Intracerebral hemorrhage

Within the subpopulation of patients presenting with hemorrhage, almost half died (55%) or underwent some form of invasive treatment,25 thereby excluding them from the follow-up cohort, which was thus enriched in non-hemorrhagic lesions.

The overall annualized incidence of follow-up hemorrhage, both parenchymal and subarachnoid, was 1.1%. Among patients with a non-hemorrhagic index presentation, the follow-up annual risk of first bleed was also 1.1%. Aneurysm enlargement was a strong independent predictor of hemorrhage, with a 6% annualized rupture rate in the Mayo group.15

Ischemic stroke

The annual incidence of stroke in the overall cohort was 3.6%, with corresponding annual stroke related mortality of 1.6%. For the subgroup with an index stroke presentation, the annual risk of recurrent stroke was 5.9%. Available data suggested that patients on anticoagulation may be less likely to suffer stroke recurrence compared with antiplatelet regimens.17 ,20 No publication addressed specific secondary prevention therapies and patient compliance, which are now clearly established as key drivers of improved stroke prevention strategies. The high recurrent stroke incidence in this population led multiple authors to conclude that dolichoectasia/fusiform aneurysm was either an independent risk factor for stroke or a marker of poor natural history.13 ,17 ,18 ,26 This view seems congruent with the overall modest incidence of cardiovascular risk factors within the overall cohort.

Mass effect

When brain (as opposed to cranial nerve) mass related symptoms were established, the outcome was especially poor, with 40% mortality after 3.7 years of follow-up. These values include patients with an initial mass effect presentation and those who developed it during follow-up, in which case time to death was calculated from the onset of symptoms.


Determination of growth requires serial imaging, which was available in the Mayo and Sienna cohorts only. Overall, 46% of lesions showed enlargement over 8.5 years (5.4% annual incidence). Growth was strongly associated with poor outcome, resulting in 52% mortality after 8.5 years (6.1% annual incidence). Mangrum et al15 specifically addressed this issue in the Mayo cohort, where 48% of imaged lesions demonstrated enlargement over 4 years; 68% of these patients died, corresponding to 5.7 times increased mortality relative to those whose aneurysms appeared stable. In Passero's cohort, disease progressed in 43% over ∼12 years, and was associated with significantly greater risks of ischemic stroke (63% stroke among enlarging lesions vs 19% in stable ones), hemorrhage, and death (46%). Notably, the incidence of ischemic stroke was substantially more common than hemorrhage in the enlarging aneurysm group.


The overall mortality among initially asymptomatic patients was 5.0% after 11.8 years of follow-up (0.4% annual mortality); these data were strongly driven by the Siena cohort. In the Mayo cohort of 64 initially asymptomatic patients, nine experienced transient ischemic attacks or infarction; two progressed to mass effect, of which one hemorrhaged.16 In the Passero group, among 34 initially asymptomatic patients, 30% experienced at least one adverse event (2.5% annual incidence), and 13% died after 11.7 years.


The heterogeneous patient populations and study methodologies among publications included in this review constitute the principal limitations of our study. We simply do not know to what degree these data may be generalizable to the wide spectrum of disease, which extends from dolichoectasia to complex fusiform aneurisms, or to any individual patient. In these circumstances, we attempt to extract reasonable conclusions based on what is presently available, appreciating that differences among studies may also prove enriching and insightful. Another notable limitation of our review, propagating limitations of the individual publications, is the lack of Kaplan–Meyer-type survival curves defining times to event such as death or stroke for the overall group. The annual incidence numbers listed in tables 3 and 4 represent ratios of overall event incidence and duration of follow-up. This assumes uniform probability of event occurrence over time, which is unlikely.

The issue which requires special attention is determining whether a rational basis exists for in vivo aneurysm subclassification into dissecting and non-dissecting types. Mizutani et al,9 in a cohort of 80 patients, classified vertebrobasilar disease into classic dissecting, segmental ectasia, dolichoectatic dissecting, and saccular subtypes based on histopathologic observations in 12 cases, arguing that classic dissecting aneurysms (four autopsied cases of hemorrhage) demonstrated acute elastic lamina tears without secondary intimal reaction. Based on these findings, all 65 patients with hemorrhage on presentation were considered to harbor the acute dissecting type of aneurysm, essentially equating hemorrhagic clinical presentation with dissection. Multiple issues arise, however, from attempted application of Mizutani's classification to pre-autopsy cases, as there is no established imaging correlate of described histopathologic findings. Some authors attach the dissecting label to lesions which have not ruptured,24 ,27 in contradiction with Mizutani's findings, while others describe hemorrhage in self-defined cases of dolichoectasia.28 Our conclusion is that dissecting etiology is assigned based on clinical presentation rather than pathophysiologic or imaging characteristics. In consequence, the apparent incidence of hemorrhage is fallaciously reduced in the perceived non-dissecting group to the extent that hemorrhage itself serves as a factor in the definition of dissection.

Given these uncertainties, we believe that vertebrobasilar disease is best analyzed with emphasis on clinical presentation and subsequent course, rather than presumed underlying pathology. This approach leads to the relatively straightforward conclusions that cases of subarachnoid hemorrhage and lesion growth do badly, those presenting with ischemic stroke have a variable course dominated by recurrent stroke, while incidental, stable cases are associated with a favorable neurologic prognosis. It is critical to keep in mind that nearly half of patients will die of a non-neurologic cause, often not cardiovascular in etiology,13 ,16 ,17 thus strongly emphasizing the importance of overall medical management.

All of the available data strongly support the conclusion that the risk of hemorrhage in aneurysms which do not originally present with rupture is low; even lesions demonstrating progressive growth are still more likely to produce mass effect or stroke rather than hemorrhage.9 ,13–16 ,20 The defining neurologic complications, for the entire spectrum between dolichoectasia and fusiform aneurysm, are ischemic stroke and mass effect. It is likely that overall mortality, neurologic and otherwise, can be further reduced by an aggressive medical management protocol, even if vertebrobasilar disease represents a unique risk factor apart from the usual cardiovascular predictors of mortality. This approach requires a dedicated interdisciplinary effort. To what extent medical management protocols might be customized for the unique features of vertebrobasilar disease (with some suggesting that anticoagulation might be more effective than an antiplatelet regimen),20 will likely remain unclear without further controlled studies.

The issue of mass effect is intertwined with that of aneurysm enlargement. By the time brainstem mass effect is established, prognosis is very poor. In contrast, compression of regional peripheral nerves may be successfully treated by addressing the symptomatic concern (such as gasserian thermocoagulation) with subsequent long term survival.19 The promise of reduction in mass effect following successful endoluminal (flow diversion) or other intervention remains alluring, despite uncertainly in the safety and efficacy of these heroic treatment modalities. In our opinion, mortality statistics support the option of invasive treatment, on an individual basis, in patients whose lesions demonstrate growth, before mass effect symptoms render treatment more technically challenging and neurologic recovery uncertain.


Vertebrobasilar disease represents a spectrum of conditions; neurologic outcome is best defined on the basis of the initial clinical presentation and morphologic evolution rather than the presumed underlying pathophysiology. Cases of index subarachnoid hemorrhage and progressive lesion growth/mass effect fare badly; those presenting with ischemic stroke have a variable course marked by recurrent stroke, while incidental cases carry a favorable prognosis. Nearly half of patients die from non-neurologic causes, placing a heavy premium on overall medical management. It is of the utmost importance to recognize the low (∼1% annually) risk of rupture in patients who do not initially present with subarachnoid hemorrhage. With the above observations in mind, the following broad suggestions may be considered:

  1. Asymptomatic, stable dolichoectasia or non-saccular aneurysm—benign neurologic prognosis. Emphasis on aggressive medical management. Serial imaging follow-up to monitor for disease progression (enlargement).

  2. Enlarging, still asymptomatic lesion—substantially worse long term prognosis dominated by stroke and mass effect, while the incidence of hemorrhage remains low. May possibly benefit from currently available interventional or surgical treatment prior to entering the symptomatic phase; morbidity of treatment remains to be defined.

  3. Established brainstem mass effect—especially poor prognosis, with no satisfactory medical options. Intervention, despite its uncertainties, appears to represent the most viable possibility of a meaningful solution.

  4. Focal cranial nerve mass effect—short term prognosis may still be favorable; presenting issue can be addressed by thermal ablation, pharmacologic management, etc.

  5. Acute or recurrent ischemic stroke—aggressive medical management. Antiplatelet failure may be grounds for a trial of anticoagulation.

  6. Hemorrhage—high overall morbidity and mortality—to be managed on an individual basis.



  • Contributors MS: literature search, data analysis, and primary manuscript writing. TB, HAR, ER, DZ and PKN: data analysis and manuscript editing.

  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.