Introduction
Vertebrobasilar fusiform aneurysms (VBFAs), the most common spontaneous intracranial dissections, are more prevalent in younger and Asian populations [1]. The detection rate of VBFAs has increased in recent years with the advancement of non-invasive brain imaging screening techniques. The natural history of VBFAs is complex and variable, with symptoms including subarachnoid hemorrhage, ischemic stroke, and brainstem or cranial nerve compression, which can lead to high morbidity and mortality [2-4]. This makes therapeutic strategies for unruptured VBFAs of great clinical research interest.
There is no strong evidence supporting the optimal approach and timing for managing VBFAs. Surgical treatment of VBFAs is poorly reported, and even the most experienced institutions report high surgical mortality [3,5,6]. In contrast, endovascular treatment is more widely accepted and relatively safe [7]. Over the past two decades, stent-assisted coiling (SAC) has been recognized as the classic treatment approach for VBFAs. However, conventional stents are often criticized for their high recanalization rate [8]. More recently, the use of flow diverters (FDs) has been extended to unruptured VBFAs, demonstrating seemingly favorable results in some small-scale studies [9-12]. Nonetheless, despite their growing popularity, evidence supporting the comparative effectiveness of FDs in improving long-term prognosis and preventing aneurysm rupture in patients with VBFAs remains limited and inconclusive.
To fill these knowledge gaps, we conducted a large multicenter observational cohort study involving patients with unruptured VBFAs from diverse regions and medical centers. Specifically, we compared the safety and efficacy of conservative management (CM), SAC, and FD interventions in treating patients with unruptured VBFAs, focusing on long-term prognosis. We also aimed to identify patient characteristics that may influence treatment strategies and outcomes. Our findings provide valuable insights into the field of VBFA management and inform evidence-based treatment strategies, thereby improving patient outcomes and quality of life.
Methods
Study design and participants
This study analyzed data from a prospectively maintained nationwide multicenter registry, the Registry of Endovascular Treatment for Vertebrobasilar Dissecting Aneurysms in China (VBDAs China), which is a collaborative study registered at ClinicalTrials. gov (NCT 06541106). The registry was established to explore the natural history of unruptured VBFAs in Asian populations and optimal treatment strategies for patients with VBFAs. Supplementary Material 1 details data quality management information.
Patients with unruptured VBFAs who were continuously recruited from the VBDAs China cohort between January 2011 and January 2024 were reviewed. Individuals aged >18 years who underwent cranial magnetic resonance imaging (MRI) were eligible for inclusion. The exclusion criteria were as follows: (1) undergoing microsurgical treatment; (2) presence of coexisting neurological tumors or other cerebrovascular diseases; and (3) loss to follow-up. The participant selection process is illustrated in Figure 1. The Institutional Review Board of Beijing Tiantan Hospital approved this study (IRB approval number: KY2024-334-02), and all patients provided written informed consent. This study adhered to the Strengthening the Reporting of Observational Studies in Epidemiology reporting guideline [13].
VBFAs were defined when computed tomography angiography, magnetic resonance angiography, or digital subtraction angiography (DSA) showed irregularity, stenosis, or fusiform dilatation of the vertebral or basilar arteries, a pearl-and-string sign, or a string sign; when DSA showed contrast retention in the ver-tebral or basilar arteries; or when MRI showed intramural hematoma, intimal flap, or double lumen sign [14]. According to the classification proposed by Zhang et al. [14], we categorized VBFAs into four types: Type I (classic dissecting aneurysms), Type II (segmental ectasia), Type III (dolichoectatic dissecting aneurysms), and Type IV (large mural bleeding ectasia) (Figure 2). Supplementary Material 2 and Supplementary Figure 1 illustrate the typical MRI findings and specific definitions of VBFAs.
Cohort definition
Patients with unruptured VBFAs were categorized into three groups based on the treatment strategy: CM, SAC, and FD. Patients in the CM group did not undergo endovascular or surgical treatments. Routine preoperative DSA and MRI were required for procedural planning in patients undergoing interventions for aneurysms. Patients in the SAC and FD groups were treated with conventional stents and FDs, respectively. All endovascular procedures were performed by experienced neurointerventionalists with more than 15 years of experience at each participating center. Supplementary Material 3 details the endovascular treatment procedures.
Baseline characteristics
We collected demographic data (age and sex), comorbidities, clinical presentations (headache/dizziness, ischemic stroke/transient ischemic attack, cranial nerve compression, and brainstem mass effect), neurological function measured using the modified Rankin Scale (mRS) score, and VBFA imaging characteristics (location, width, branch, type, and MRI findings). The aneurysm width encompassed both the dilated aneurysm and, if present, intramural hematoma. A three-member central review board consisting of a neurointerventionalist, a radiologist, and a neurosurgeon reviewed all imaging and endpoint events. If there was a dispute regarding the assessment, the team reached a unanimous decision after discussion.
Outcomes and follow-up
The primary outcome was poor prognosis, defined as an mRS score >2 at the latest follow-up [15]. Supplementary Material 4 details the mRS score. The secondary outcomes included: (1) aneurysm rupture, defined as a clinical hemorrhagic event (newonset severe headache or worsening neurological deficits) confirmed by imaging findings such as computed tomography or MRI indicating VBFA-associated subarachnoid hemorrhage; (2) ischemic stroke related to the treated vessel; (3) compression symptoms, including aneurysm-induced cranial neuropathy or brainstem symptoms; and (4) VBFA-related death. Additionally, to evaluate stent effectiveness in the SAC and FD groups, only patients who underwent follow-up DSA were included. Supplementary Material 5 presents the definitions of complete occlusion and recanalization.
Well-trained clinical research coordinators evaluated the outcomes through visits or interviews at 3-6 months, annually (1, 2, and 3 years), and every 5 years post-treatment. Follow-up started on the VBFA diagnosis date for the CM group and on the endovascular treatment date for the SAC and FD groups, and ended when the outcome occurred or at the last follow-up date.
Controlling for confounders
The inverse probability of treatment weighting (IPTW) method, which has been widely demonstrated to be effective in several clinical studies (Supplementary Material 6) [16-18], was used to reduce imbalances in measured confounders across the three treatment groups. Using these stabilized weights, the average treatment effect can be estimated without excluding patients while preserving the original sample size. Standardized mean differences (SMDs) were calculated for each covariate to determine whether confounders differed significantly among the three treatment groups. Confounders were considered to show no between-group difference when the SMD ≤0.1 post-IPTW [19]. Despite employing the IPTW technique, unmeasured confounders could still result in bias. Consequently, we calculated E-values to explore the strength of unmeasured confounders and explain the observed association [20]. The E-value represents the minimum strength of an unmeasured confounder necessary to nullify the observed association [21].
Statistical analyses
All statistical analyses were performed using R software (v4.4.1; R Foundation for Statistical Computing, Vienna, Austria) at a two-tailed P<0.05 significance level. Continuous variables were presented as mean±standard deviation and median (interquartile range [IQR]) for normally and non-normally distributed data, respectively. Categorical variables were reported as frequencies and percentages. Risk differences (RDs) were calculated for all outcomes after IPTW. RD describes the difference in the risk of an event between an experimental and a control group [22]. Logistic regression models were used to calculate odds ratios (ORs) and 95% confidence intervals (CIs) for both primary and secondary outcomes. Time-to-event analyses were performed using the Kaplan-Meier method to estimate event-free survival. Differences between groups were assessed using the log-rank test. Stratified analyses were further conducted according to aneurysm location and type to explore heterogeneity in long-term prognosis across subgroups.
Prespecified subgroup and sensitivity analysis
Prespecified subgroup analyses were performed based on age, sex, symptoms, aneurysm location, type, branch, and width. For these comparisons, CM was used as the reference. Furthermore, interaction tests for these subgroups were conducted to assess heterogeneity across subgroups.
We performed a sensitivity analysis to investigate the stability of the primary outcomes. First, we performed an unadjusted analysis without weighting or adjustments. Second, IPTW was used to compare treatment strategies among the three groups. Third, propensity score matching was performed for the same comparisons. Fourth, we used a multivariable logistic regression model adjusting for factors included in the propensity score calculation to assess the ORs of the study outcomes.
Results
Study population and baseline characteristics
Between January 2011 and January 2024, 1,115 patients with unruptured VBFAs were recruited from 22 participating sites and enrolled in the VBDAs China study. After excluding patients who met the exclusion criteria and those lost to follow-up, 838 patients with unruptured VBFAs and long-term follow-up were included in this study (median age, 54 years [IQR 48-60]; 655 men), of whom 305 (36.4%), 241 (28.8%), and 292 (34.8%) underwent CM, SAC, and FD, respectively. The median (IQR) follow-up duration was 54 (30-70) months (52 [32-66], 54 [32- 72], and 56 [28-72] months for the CM, SAC, and FD groups, respectively). Supplementary Table 1 details the baseline comparisons between the included patients and those lost to follow-up; Supplementary Table 2 shows the breakdown of patients lost to follow-up classified into the CM, SAC, and FD treatment groups; and Supplementary Table 3 compares the baseline characteristics of included and lost patients in each group. No significant differences were found between the two groups.
VBFAs with brainstem mass effects, those located in the basilar and vertebrobasilar arteries, those with larger aneurysm widths, and those of Types II, III, and IV were more likely to receive CM. Following IPTW adjustment, most baseline characteristics met the SMD ≤0.1 threshold, indicating that baseline differences across the three treatment groups had been eliminated (Table 1 and Supplementary Figure 2). The adjusted cohort consisted predominantly of middle-aged males, and almost all patients demonstrated good neurological function before treatment. Regarding lesion characteristics, most VBFAs were Type I and located in the vertebral arteries.
Primary outcomes
Table 2 lists the frequencies of each outcome. After IPTW, the incidences of poor prognosis in the CM, SAC, and FD groups were 20.4%, 14.7%, and 11.0%, respectively. Compared with CM, FD resulted in significantly lower risks of poor prognosis (OR, 0.48 [95% CI, 0.30 to 0.77], P=0.002; RD, -9.4% [95% CI, -15.3% to -3.5%], P=0.002), while no significant difference was observed between CM and SAC. Additionally, there was no significant difference between the SAC and FD groups in terms of their impact on poor prognosis (Supplementary Table 4). The E-values for all significant outcomes implied that unmeasured confounders had to be strongly correlated (E-value range, 3.59-9.47) with the treatment modality and primary outcomes to mitigate the relationship between these variables and reduce the relative risk to 1.
Secondary outcomes
During follow-up, the three strategies showed no significant differences in ischemic stroke occurrence or VBFA-related deaths (Table 2). Notably, the FD group exhibited a significantly lower risk of aneurysm rupture (OR, 0.20 [95% CI, 0.07 to 0.60], P=0.004; RD, -5.3% [95% CI, -8.5% to -2.1%], P=0.001) and compression symptoms (OR, 0.30 [95% CI, 0.13 to 0.68], P=0.004; RD, -6.0% [95% CI, -9.8% to -2.1%], P=0.002) compared with CM, whereas no significant difference was observed between CM and SAC for either outcome. Additionally, there was no significant difference between the SAC and FD groups in terms of aneurysm rupture, ischemic stroke, compression symptoms, or VBFA-related deaths (Supplementary Table 4).
We included 306 patients who underwent follow-up DSA (136 in the SAC group and 170 in the FD group) and applied IPTW to balance the baseline characteristics between the two groups (Supplementary Table 5). We then compared stent effectiveness and found that, at a median follow-up of 26.5 months, FD was associated with a significantly higher complete occlusion rate (OR, 2.25 [95% CI, 1.38 to 3.67], P=0.001; RD, 17.5% [95% CI, 7.0% to 28.0%], P=0.001) and a significantly lower recanalization rate (OR, 0.20 [95% CI, 0.08 to 0.50], P<0.001; RD, -12.1% [95% CI, -18.8% to -5.3%], P<0.001) compared with SAC (Supplementary Table 6).
Subgroup analyses
Figure 3 summarizes the results of the prespecified subgroup analyses. In the subgroup analysis comparing CM and FD, an interaction was observed between the aneurysm location and aneurysm type subgroups (P for interaction=0.009 and <0.001, respectively). Specifically, compared to CM, FD had a better prognosis for treating VBFAs located in the vertebral artery (OR, 0.24 [95% CI, 0.11 to 0.49]; P<0.001), Type I VBFAs (OR, 0.06 [95% CI, 0.01 to 0.28]; P<0.001), and Type II VBFAs (OR, 0.11 [95% CI, 0.01 to 0.89]; P=0.039). No significant interaction effect was observed across subgroups in the subgroup analysis comparing CM and SAC. However, compared with CM, SAC was associated with a better prognosis for treating VBFAs located in the vertebral artery (OR, 0.51 [95% CI, 0.29 to 0.90]; P=0.021) and Type I VBFAs (OR, 0.34 [95% CI, 0.16 to 0.74]; P=0.006).
Time-to-event analysis stratified by aneurysm location and type
In the overall cohort, significant differences in event-free survival were observed among the CM, SAC, and FD groups (logrank P<0.001), with the FD group showing the most favorable prognosis (Figure 4A). When stratified by aneurysm location, a significant separation of survival curves was observed in patients with VBFAs located in the vertebral artery (log-rank P<0.001). In this subgroup, FD was associated with the highest event-free survival rate, followed by SAC and CM (Figure 4C). In contrast, no significant differences were observed among the treatment strategies in patients with VBFAs located in the basilar artery or vertebrobasilar junction (Figure 4B and D). Stratification by aneurysm type revealed marked heterogeneity in long-term prognosis. Significant differences in event-free survival were observed in patients with Type I (log-rank P<0.001) and Type II VBFAs (logrank P=0.031), with FD showing the most favorable prognosis (Figure 4E and F). In contrast, no significant differences were observed among the treatment strategies in patients with Type III or IV VBFAs, in whom the overall prognosis remained poor regardless of the treatment strategies employed (Figure 4G and H).
Sensitivity analyses
As shown in the sensitivity analyses (Figure 5), SAC was associated with a better prognosis than CM in the unadjusted model (OR, 0.53 [95% CI, 0.34 to 0.84]; P=0.007); however, this association lost statistical significance after adjustment for confound-ing variables. In contrast, FD consistently demonstrated a significant prognostic benefit in patients with VBFAs across all models (P<0.05).
Discussion
In this observational cohort study using data from a nationwide multicenter registry, we compared the long-term outcomes of different treatment strategies (CM, SAC, and FD) for unruptured VBFAs. Our results showed that patients undergoing FD had a better prognosis, a lower rate of aneurysm rupture, and a lower incidence of compression symptoms than patients receiving CM. Notably, FD treatment was particularly associated with improved long-term outcomes in VBFAs located in the vertebral artery as well as in Type I and Type II VBFAs. Our findings will contribute to the selection of individualized treatments for patients with unruptured VBFAs.
Vertebrobasilar dolichoectasia and vertebrobasilar non-saccular, fusiform, and dissecting aneurysms are among the terms associated with vertebrobasilar site aneurysms. Notably, these terms intersect, and the boundaries of their use in research are not well-defined [4,7,23,24]. Herein, we focused on patients with VBFAs who had undergone MRI scans and exhibited characteristics representing a broader pathological spectrum, categorizing them into four types according to Zhang’s classification criteria [14]. It is worth noting that our study did not include ruptured VBFAs; thus, Type I in this study refers specifically to type Ib in Zhang’s classification criteria.
Although the International Study of Unruptured Intracranial Aneurysms (ISUIA) and the Unruptured Cerebral Aneurysm Study of Japan (UCAS Japan) reported the natural history and treatment results of unruptured intracranial aneurysms, including VBFAs, in large cohorts, VBFAs were underrepresented in these studies [25,26]. Unruptured VBFAs accounted for only 4.9% and 3.7% of all cases in the ISUIA and UCAS Japan studies, respectively, indicating that VBFA outcomes should be interpreted cautiously. VBFAs are a specific and heterogeneous subtype of intracranial aneurysms with a poor natural history and high morbidity despite treatment [9]. Therefore, selecting between conservative and interventional treatments for VBFAs can be challenging for both neurosurgeons and patients, particularly in the absence of clearly defined clinical guidelines. It is also noteworthy that the lack of a true aneurysm neck and the localization of VBFAs in the posterior cranial fossa adjacent to the brainstem make surgical treatment extremely risky, hence the gradual transition from surgical to endovascular treatment [3,5,6].
Whether endovascular treatment improves long-term neurological status of patients with VBFAs remains unclear. Moreover, despite several previous studies reporting a satisfactory neurologic prognosis with endovascular treatment, there are limited comparisons with CM [4,9,27]. This study aimed to compare the efficacy of CM and endovascular treatment in managing unruptured VBFAs using IPTW to balance intergroup baseline characteristics. Specifically, we aimed to provide comprehensive and robust evidence for personalized recommendations in the FD era. We found that FD was superior to CM in improving patients’ long-term neurological status, whereas SAC and CM showed no significant differences. This discrepancy could be attributed to the properties of FD, which makes it well-suited for VBFA treatment. Given its reduced porosity and increased metal surface density, FD can promote endothelialization and repair of dissecting segments. FD can also promote gradual aneurysm occlusion while preserving essential blood flow to critical side branches [12].
However, the mechanisms outlined above remain largely hypothetical and should be interpreted with caution, as other potential factors may also contribute to the observed superior longterm neurological outcomes associated with FD compared with CM. First, this study was observational in design. Although IPTW and sensitivity analyses were employed to control for known confounders, residual selection bias due to unmeasured variables could not be completely excluded. Second, subtle biases in clinical decision-making (e.g., a tendency to favor FD for lesions deemed to carry a lower perioperative risk while preferentially selecting CM for patients with higher surgical risk) may also have influenced the observed outcomes. Therefore, the present findings require further validation in well-designed prospective randomized controlled trials.
Among the greatest challenges faced by neurointerventionalists during VBFA treatment is identifying specific VBFA subtypes with the best long-term prognosis following endovascular treatment. Subgroup analysis revealed that, following endovascular treatment, VBFAs located in the vertebral arteries had a better prognosis than those located in the basilar and vertebrobasilar arteries. This phenomenon could be attributed to the fact that there are more perforators along the basilar artery than the vertebral artery, an assumption consistent with the findings of Munich et al. [28], who reported that the metal surface coverage of FD on the rostral side of the anterior inferior cerebellar artery should be minimized to avoid vital perforator occlusion. Furthermore, our subgroup analyses based on aneurysm type revealed that patients with Type I and Type II VBFAs had a better prognosis after endovascular treatment, whereas those with Type III and Type IV VBFAs had a relatively poor prognosis, which is consistent with the findings of Zhang et al. [14] Not surprisingly, Type III VBFAs typically involve both vertebral and basilar arteries, extend broadly along the long axis of the vessel, and require multiple overlapping stents for treatment; hence, high metal coverage could greatly increase the risk of ischemic stroke. Type IV VBFAs have obvious intramural hematomas; patients often experience progressively worsening brainstem compression symptoms, and intramural hematomas may persistently enlarge even after complete occlusion of the aneurysm and parent artery. Theoretically, lesion resection (with or without arterial bypass surgery) is the most effective treatment. Interestingly, our findings demonstrated that FD was more effective in preventing mass effect progression than CM, implying that FD should be prioritized in treating VBFAs with brainstem compression, which is consistent with previous research findings [11]. Overall, the results of our subgroup analyses were insightful, suggesting that patients with VBFAs undergoing endovascular treatment should be carefully selected to avoid worsening their natural history.
The primary treatment goal for unruptured VBFAs is complete occlusion of the aneurysm and prevention of rupture. Our findings show that FD exerts a significant protective effect against aneurysm rupture. Following FD placement, blood flow into the aneurysm is reduced, leading to intra-aneurysmal blood stagnation and subsequent thrombus formation, which gradually results in complete occlusion. In addition, the FD serves as a scaffold that facilitates endothelialization at the aneurysm neck. By providing a structural framework for endothelial cell migration and growth, FD promotes the formation of a biologically sealed surface, thereby significantly reducing the risk of aneurysm rupture and recanalization [29]. Notably, our results also demonstrate that FD is associated with alleviation of compressive symptoms. This effect is likely due to thrombus formation within the aneurysm sac following FD placement, which leads to a substantial reduction in aneurysm size. Given its efficacy in preventing rupture and relieving compression symptoms, FD is a promising treatment option for unruptured VBFAs.
There have been concerns that stent placement for treating posterior circulation aneurysms (especially those around the brainstem) could increase the risk of ischemic stroke [30-32]. However, our findings revealed that, compared with the CM group, the ischemic stroke rates in the SAC and FD groups, although higher, were not significantly different. In our experience, the overlapping stent technique could result in higher ischemic stroke rates. Unless the dilated VBFA segment was excessively long, this technique was strictly limited in this cohort; hence, the ischemic stroke rate was relatively acceptable. Endovascular treatment has greatly altered the landscape of neurosurgical management of VBFAs, and FD seems to provide the most favorable balance between securing the aneurysm and preventing medullary infarction [33,34].
Consistent with the findings of Ji et al. [7], we observed a higher complete occlusion rate (76% vs. 58.5%) and a lower recanalization rate (3.5% vs. 15.6%) in the FD group than in the SAC group at the final imaging follow-up of unruptured VBFAs. This difference may be attributable to the distinct treatment mechanisms of the two modalities. SAC involves placement of a conventional stent within the parent artery, which allows the coils to fill the aneurysm lumen more stably and prevent blood entry. However, the stent itself does not significantly alter hemodynamics. Even when complete embolization is achieved immediately post-procedure, there remains a risk of recurrence due to coil compaction or displacement caused by blood flow impingement, as observed during DSA follow-up [35]. In contrast, FD provides a high level of metal coverage and effectively reduces post-procedural aneurysm recurrence by modifying hemodynamics, thereby significantly limiting blood flow into the aneurysm and promoting its natural occlusion. In addition, in the unadjusted model, SAC was associated with a more favorable prognosis (Figure 5); however, this difference was not statistically significant after adjusting for confounding factors, possibly because of the limited sample size or heterogeneity among the patient populations.
In summary, our study suggests that FD may be a promising treatment option for VBFAs. Detailed analysis of aneurysm characteristics (location, type, size, branches, etc.), as well as the evaluation of clinical and radiological signs such as ischemia or mass effect, will help guide treatment plans and hopefully improve the prognosis of patients with this refractory disease.
Limitation
This study has several limitations that should be considered when interpreting the results. First, as an observational study, although we applied IPTW and E-value analyses to minimize confounding by indication and assessed the potential impact of unmeasured variables, residual confounding could not be completely excluded and may have influenced our findings. Second, in clinical practice, clinicians may be more inclined to choose CM for lesions with more complex anatomical locations (e.g., involvement of the basilar artery), more challenging morphologies (e.g., Types III-IV), or a perceived higher procedural risk. Consequently, the CM group may represent a cohort with an intrinsically poorer prognosis, which could lead to overestimation of the absolute benefit of FD. Third, the generalizability of our findings is limited. As demonstrated in the subgroup analyses, evidence supporting the benefit of FD remains limited for the most challenging subtypes, such as basilar artery or vertebrobasilar aneurysms, particularly Types III-IV VBFAs. This underscores the need for highly individualized clinical decisionmaking and caution against overgeneralizing our conclusions to high-risk subgroups. Fourth, due to the technical complexity and inherent risks of surgical intervention near the brainstem, patients treated with microsurgical approaches for VBFAs were not included in this study. Overall, well-designed randomized clinical trials are required to further evaluate the benefits of endovascular treatment for unruptured VBFAs.
Conclusions
Our study provides valuable insights into treatment strategies for patients with unruptured VBFAs. Compared with CM, FD may be associated with improved long-term outcomes in patients with unruptured VBFAs, particularly in those with vertebral artery lesions and Type I-II VBFAs. However, considering the possibility of residual confounding factors, these results should be interpreted with caution.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Supplement1
Print
