Background and purpose Flow diverters are sometimes used in the setting of acutely ruptured aneurysms. However, thromboembolic and hemorrhagic complications are feared and evidence regarding safety is limited. Therefore, in this multicenter study we evaluated complications, clinical, and angiographic outcomes of patients treated with a flow diverter for acutely ruptured aneurysms.
Methods We conducted a retrospective observational study of 44 consecutive patients who underwent flow diverter treatment within 15 days after rupture of an intracranial aneurysm at six centers. The primary end point was good clinical outcome, defined as modified Rankin Scale score (mRS) 0–2. Secondary endpoints were procedure-related complications and complete aneurysm occlusion at follow-up.
Results At follow-up (median 3.4 months) 20 patients (45%) had a good clinical outcome. In 20 patients (45%), 25 procedure-related complications occurred. These resulted in permanent neurologic deficits in 12 patients (27%). In 5 patients (11%) aneurysm re-rupture occurred. Eight patients died resulting in an all-cause mortality rate of 18%. Procedure-related complications were associated with a poor clinical outcome (mRS 3–6; OR 5.1(95% CI 1.0 to 24.9), p=0.04). Large aneurysms were prone to re-rupture with rebleed rates of 60% (3/5) vs 5% (2/39) (p=0.01) for aneurysms with a size ≥20 mm and <20 mm, respectively. Follow-up angiography in 29 patients (median 9.7 months) showed complete aneurysm occlusion in 27 (93%).
Conclusion Flow diverter treatment of ruptured intracranial aneurysms was associated with high rates of procedure-related complications including aneurysm re-ruptures. Complications were associated with poor clinical outcome. In patients with available angiographic follow-up, a high occlusion rate was observed.
- endovascular techniques
- flow diversion
- intracranial aneurysm
- subarachnoid hemorrhage
Statistics from Altmetric.com
Preventing rebleeding is the prime purpose of early intervention in cases of aneurysmal subarachnoid hemorrhage (SAH). Preceding neurosurgical therapies have failed to reliably protect patients with acute SAH from non-saccular aneurysms from early re-ruptures.1 2 Stent-assisted coiling has proved to be more effective, as shown in recent studies.3–5 However, coiling of these very delicate aneurysms carries the inherent risk of procedural aneurysm rupture with devastating consequences, since these patients are often loaded with dual antiplatelet therapy (DAPT).
Flow diverters (FDs) reconstruct the lumen of the parent artery and require no manipulation within the aneurysm sac, so the risk of intraoperative rupture is minimized. However, thromboembolic and hemorrhagic complications have been reported in the literature. The use of DAPT in the setting of acute SAH is of extra concern since these patients are often subjected to other crucial intracranial procedures. If the complication rate of FDs as treatment for acute SAH is comparable to that of stent-assisted coiling, then it would favor FD treatment for some aneurysm subtypes since FD technology is thought to have better durability with a lower recanalization rate. Several cases have been reported in the recent past, and a meta-analysis showed higher rates of long-term complete occlusion and a lower rate of retreatment for patients with ruptured blister-like aneurysms treated with FDs versus those treated with other reconstructive techniques.6–10
However, evidence is still limited and complication rates and clinical outcomes vary between published meta-analyses.11 12 The efficacy and safety of the use of FDs in the acute SAH phase is therefore still open to question.
In this study we present a multicenter experience of FDs for acutely ruptured aneurysms. The objective was to evaluate the complications, clinical and angiographic outcomes of consecutive patients treated with a FD for acutely ruptured intracranial aneurysms.
This retrospective analysis of prospectively kept data was conducted in five European centers and one in the USA. Review and approval of the study protocol was obtained from the local institutional review board (no. 2017–3511) of the coordinating center and the trial was registered in a Clinical Trials Register (DRKS-ID: DRKS000012438).
All patients with aneurysmal SAH treated with any type of FD within 15 days after the last hemorrhage were included. Indications for aneurysm treatment with a FD were left to the discretion of the participating centers. Both patients presenting with a rebleed and those with initial aneurysmal SAH were included. Fifteen days was chosen as the maximum treatment delay since delayed cerebral ischemia (DCI) mostly occurs within this timeframe and to improve comparability with existing literature.
All consecutive patients who met these inclusion criteria were included.
Clinical and radiological assessments
Medical charts of the patients with prospectively collected data were retrospectively analyzed. Data collection and reporting was performed by each center according to a standardized electronic case report form (eCRF) in an online database (Castor). The STROBE guidelines were followed for the collection and reporting of data. The following baseline parameters were collected: sex, age, treated for initial SAH or rebleed, treatment delay, World Federation of Neurosurgical Societies (WFNS) grade, modified Fisher grade, and ventriculostomy prior to FD placement. Baseline data collection also included aneurysm type (fusiform, dissecting, saccular, blood-blister like), aneurysm size (in mm), neck size (in mm), dome-to-neck ratio, location, and previous treatment of the target aneurysm. Assignment to the different aneurysm types detailed in the eCRF was left to the discretion of the participating centers. No prespecified definitions were used.
With respect to the index procedure, data were obtained on type, number, and sizes of FDs used and adjunctive treatment such as additional coiling. In addition, detailed information on antiplatelet therapy and use of platelet function tests was collected.
All neurological and hemorrhagic intra- and post-procedural complications (regardless of whether they led to neurological deficit) which occurred until last available time of follow-up were registered. Intra-procedural complications were subdivided into thromboembolic and others. Post-procedural complications were grouped as either (1) ischemic; (2) hemorrhagic (aneurysm rebleeding, intraparenchymal hemorrhage not ventricular drain-related, shunt-related hemorrhage, extra-axial hemorrhage, and other); or (3) other (e.g., hydrocephalus, meningitis, gastrointestinal bleeding). Post-procedural ischemic complications were separated into ‘disease-related’ (eg, DCI) and ‘procedure-related’ complications. If the relationship between the intervention and the complication was unclear, the complication was classified as procedure-related.
Concerning clinical follow-up, we collected data on duration of follow-up, modified Rankin Scale (mRS) score, retreatment of the target aneurysm, and occurrence of new neurologic deficits.
All angiographic data were reviewed by the local center. Centers provided data on time interval, imaging modality, position of the FD, and occlusion rate based on both the Raymond–Roy (RR) and the O’Kelly–Marotta (OKM) scales.
Study end points
The primary end point was the rate of favorable clinical outcome, defined as mRS scores of 0–2 at last available time of follow-up. Any mRS scores of 3–6 were considered unfavorable.
Secondary outcomes were complication rate, aneurysm re-rupture, and the rate of complete angiographic occlusion at last available angiographic follow-up. Complete angiographic occlusion was defined as RR class I or OKM class D.
To evaluate the statistical significance of the differences found in frequencies between the subgroups we used the χ2 test or Fisher’s exact test (if count <5). Binary logistic regression analysis was performed to assess the effect of independent variables on unfavorable clinical outcome at end of follow-up. The threshold for statistical significance was set at p<0.05. All analyses were performed using version 25 of IBM SPSS Statistics (Armonk, New York, USA).
Patient and aneurysm characteristics
Forty-four consecutive patients (44 aneurysms) treated from March 2012 to December 2017 were included. The mean age of the included patients was 55.8 years (range 19–83) and 28 were female (64%). Ten patients had been included in recent trials.13 14 Three patients (7%) were treated for a rebleed (second rebleed in all cases). The mean treatment delay was 3 days and 13 patients (30%) were treated ultra-early (within 24 hours) (table 1).
Thirteen fusiform (30%), 11 dissecting (25%), 11 saccular (25%), and 9 (20%) blister-like aneurysms were included. Twenty aneurysms (45%) were located in the posterior circulation.
Patients were treated with the following types of FD: 21 Surpass (48%), 12 FRED (27%), 4 Pipeline Embolization Device (9%), 4 SILK (9%), and 3 Derivo (7%). No Pipeline Flex embolization devices with Shield technology were reported. Placement was successful in all cases with complete neck coverage.
Additional treatment modalities were used in 10 cases. One case concerned stent placement within the FD in an attempt to improve wall apposition. In the other nine cases aneurysms were additionally coiled; in one of these treatment was staged with coiling performed at day 1 after SAH and FD placement at day 12.
Peri-procedural antiplatelet regimens were reported for all cases and varied widely. The most frequently used DAPT combination was acetylsalicylic acid with clopidogrel administered during intervention. Glycoprotein IIb/IIIA receptor antagonists were administered in 17/44 cases (39%). For a complete overview of the peri-procedural antiplatelet regimen per case see online supplementary table 1.
Platelet function tests were used for 20/44 patients (45%). The difference in the rate of either thromboembolic or procedure-related ischemic complications was not significant between patients who had or did not have a platelet function test, with rates of 25% (5/20) and 13% (3/24) (p=0.44), respectively.
The Verify Now platelet function test was used for 2/3 patients with a thromboembolic complication. In both cases the P2Y12 reaction units (PRU) level was <230. The same test was used for one patient with an ischemic stroke within 30 days after FD placement. The PRU level was also <230 in this case. Our data did not reveal any hyper-responders.
Twenty of the 44 patients (45%) had a favorable clinical outcome (mRS 0–2, table 2). Mean duration of follow-up was 7.6 months (median 3.4 (IQR 1.3–9.7) months). Rates of favorable clinical outcome were 57% (17/30) versus 21% (3/14) (p=0.05) for patients presenting with WFNS 1–3 and 4–5, respectively.
During the follow-up period eight patients (18%) died. Six patients died during the initial hospitalization, one patient died in an outside hospital shortly after discharge, and one patient died in a nursing home due to sepsis. In five of the eight deceased patients death was probably procedure-related.
Factors related to unfavorable clinical outcome
No significant difference in favorable clinical outcome was found between the aneurysm subtypes.
Binary logistic regression was used to analyze the effect of the following factors on unfavorable clinical outcome: dichotomized WFNS (favorable 1–3/unfavorable 4–5), aneurysm location (anterior/posterior), treatment ultra-early (day 0/1–15), and having had a procedure-related complication (yes/no). Presentation with unfavorable WFNS and procedure-related complications were both significantly correlated with an unfavorable clinical outcome (p=0.02, OR 7.1 (95% CI 1.3 to 38.0) and p=0.04, OR 5.1 (95% CI 1.0 to 24.9), respectively). The area under the receiver operating characteristic (ROC) curve of the model was 77 (95% CI 62 to 92) (see online supplementary table 2).
Angiographic follow-up was available for 29 of the 36 survivors (81%) with a mean follow-up time of 16.4 months (median 9.7 (IQR 6.4–24.8) months). In patients with available angiographic follow-up data the complete occlusion rate was 93% (27/29 RR class I, 27/29 OKM class D). The two other aneurysms (7%) were classified as RR class III. Digital subtraction angiography was used in 28/29 cases (97%) for follow-up imaging; the other case was assessed by subtraction CT angiography.
In four patients the reason for missing follow-up was unknown. In two cases it was discussed with the patient and/or family who decided that, due to the limited clinical condition, follow-up imaging would not have further clinical consequences. In one case follow-up had not yet been performed due to recent timing of treatment.
A total of 57 complications were registered of which 25 (44%) were procedure-related (table 3). The others were classified as disease-related (e.g., meningitis, hydrocephalus). The 25 procedure-related complications occurred in 20 different patients (45%). Of these 25 complications, five occurred intra-procedurally; three concerned a thromboembolic complication (7%). The two others were a vertebral artery dissection with stenosis and severe vasospasm which led to abortion of the procedure. See online supplementary table 3 for more details regarding the antiplatelet protocol used for patients with complications.
Twenty post-procedural complications related to the procedure were reported: 6 ischemic strokes (reported as not caused by DCI), 10 intracranial hemorrhagic strokes (5 rebleeds, 2 intraparenchymal hemorrhages not related with ventriculostomy, 2 ventricular shunt-related hemorrhages, and 1 extra-axial hemorrhage (increase subdural hematoma)), and 4 others (2 gastrointestinal bleeds and 2 retroperitoneal hematomas). The total number of procedure-related strokes was therefore 15 (in 14 different patients (32%)), leading to permanent neurological deficit in 12 patients (27%). For an illustrative case see the online supplementary illustrative case.
Three of the five aneurysms that had a rebleed had a maximum size of ≥2 cm. The rebleed rates for aneurysms >20 mm and <20 mm were 60% (3/5) and 5% (2/39) (p=0.01), respectively. In one case migration of the FD into the aneurysm was observed. Rates of re-rupture for aneurysms treated with FD solely versus those that were additionally coiled were comparable: 4/35 (11%) aneurysms treated with FD alone re-ruptured versus 1/9 (11%) aneurysms that were additionally coiled; however this last case concerned the case with FD migration (see table 4 for detailed information regarding the rebleed cases).
The difference in complication rate for patients with aneurysms in the posterior versus the anterior circulation was not statistically significant (60% (12/20) vs 33% (8/24) (p=0.08), respectively).
To the best of our knowledge, this study represents the largest series to date of patients (n=44) with aneurysmal SAH treated with flow diversion within 15 days. A high complete occlusion rate of 93% was observed. However, the rate of favorable clinical outcome was 45% with a high rate of procedure-related complications (45%) including a rebleed rate of 11%, leading to permanent neurological deficit in 12 patients (27%).
Our rates of clinical favorable outcome and complications differ considerably from the results of two recently published meta-analyses investigating FD use in ruptured aneurysms.11 12 In these studies, favorable clinical outcome rates of patients (treated within 15 and 30 days) were, respectively, 79% and 83%. This concerns a significant deviation from the 45% reported in this study.
Complication rates also differed significantly when comparing our results with those from other studies. Combined ischemic and hemorrhagic complication rates of patients treated within 15 days and 30 days were 16%12 and 15%,11 respectively.
As mentioned previously, our procedure-related complication rate was 45%. If only ischemic and hemorrhagic complications were taken into account, this rate would have been 36%, which is still double the rate of 15–16% mentioned above. However, another meta-analysis investigating complications associated with the use of FDs reported a complication rate of 31% for ruptured aneurysms.15 The thromboembolic and ischemic complications encountered in our study are similar to those found in studies in which a ruptured aneurysm was treated using stent-assisted coiling.4 16 17
Rebleed rates reported in the literature were 2% and 5%.11 12 Our overall rebleed rate was substantially higher at 11%. However, Madaelil et al reported no subgroup rate for patients treated in the acute phase. Since patients in the delayed phase had no hemorrhagic complications, we assume all rebleeds concerned patients treated in the acute phase, which would result in a rebleed rate of 10% (6/62). Aneurysms with a size ≥20 mm had a significantly higher rate of rebleed than those with a size <20 mm, with re-rupture rates of 60% (3/5) and 5.1% (2/39) (p=0.01), respectively. Madaelil et al reported similar rates of 57% and 2% (p=0.001) for these two groups, a finding which agrees with the existing literature that aneurysm size is a risk factor for rebleeding.18 Cohen et al emphasized the need to consider placing an external ventricular drain (EVD) prior to FD placement to lower the risk of hemorrhagic complications which are more likely to occur when an EVD is placed under dual antiplatelet therapy.19–21 In our study 20 patients (45%) had a ventriculostomy prior to FD placement. In 2/10 patients (20%) who had a ventriculostomy under DAPT, a hemorrhagic complication occurred. For both patients this concerned a revision (EVD was already placed prior to FD placement). Further research is required to identify those patients who benefit from having a ventriculostomy prior to FD treatment.
The noticeable differences between our results and those in the literature can be attributed to various factors. The first is publication bias. The majority (13/20) of studies included in the meta-analysis by Madaelil et al reported on ≤3 patients. It is highly likely that the ’best case scenario’ patients were selected for the included reports. In addition, data were mostly self-assessed, which is known to bias outcome as occlusion rates can differ by up to 26% when assessed by a core laboratory.22 Therefore, the rate of favorable clinical outcome reported in these meta-analyses may have been too optimistic. Furthermore, we registered both clinically silent as well as important complications during the total duration of follow-up instead of only peri-procedural. Finally, in our series we considered all reported ischemic strokes with an unclear interventional relationship as being procedure-related in order to provide a complete and transparent overview.
Several other series (not in the meta-analysis) investigating FD treatment in the acute SAH phase have been published.9 10 13 19 23–27 The pooled favorable clinical outcome (mRS 0–2) rate of these studies was 71% (range 27–92%) (57/80). However, caution is required when comparing results. Treatment delay was not always reported and was sometimes as long as 90 days.28
Although most studies report the complication rate and clinical outcome, they do not consider the contribution of complications to unfavorable clinical outcome. Procedure-related complications were significantly associated with an unfavorable clinical outcome in our series (OR 5.10 (95% CI 1.05 to 24.87), p=0.04).
To treat or not to treat?
In our opinion the use of FDs for acutely ruptured aneurysms should only be considered for selected cases where standard techniques such as coil embolization or stent-assisted coiling cannot be used. In the case of blister-like or dissecting aneurysms, FD treatment seems to improve the clinical outcome compared with other types of treatment or no treatment.2 6 29
The primary goal of early intervention is to prevent rebleeding. Although we did not find a significant difference in rebleed rates between aneurysms treated with or without additional coiling, Yang et al showed that adjunctive use of coiling achieves a higher incidence of immediate complete occlusion of blister-like aneurysms.26 However, Mokin et al found no difference in complete (blood blister-like) aneurysm occlusion regarding the use of adjunctive coiling.29 Adjunctive coiling could possibly result in a lower rebleed rate, so this should be considered especially in aneurysms with a size of ≥20 mm due to the high re-rupture rate of that subgroup. However, this was not the focus of our study and the numbers in our series were insufficient to conclude this.
Timing of treatment
If treatment with an FD is considered viable, it is crucial to perform this early after the ictus of SAH since dissecting aneurysms have a rebleed rate of up to 41% within the first 24 hours after the initial SAH, increasing to 57% after the first week.2 30 Our study showed no significant difference in unfavorable clinical outcome for patients treated ultra-early versus day 1–15, nor was there a significant difference in permanent deficit caused by procedure-related complications. This favors early treatment since most rebleeds occur in the early phase.
The limitations of our study include the retrospective design with its inherent limitations. Furthermore, we have collected a heterogeneous study sample. Aneurysm subgroups were small and we had a relatively high percentage of additionally coiled aneurysms. In addition, aneurysm classification and reporting of data were performed by the participating centers individually. To make data reporting as homogenous as possible we used a standardized eCRF. However, not for all centers' data were collected by a researcher who was not involved in the patient’s treatment. This may have led to under-reporting of procedure-related complications. Finally, we did not have insight into the rationale behind the choice for FD treatment for this heterogeneous selection of aneurysms. Altogether, the external validity of our study is limited. However, we included consecutive patients and our goal was to provide an overview of currently treated patients and their outcomes.
FD treatment of ruptured intracranial aneurysms was associated with high rates of procedural complications and aneurysm re-ruptures. In this series aneurysms with a size >20 mm had a relatively high chance of re-rupture. In these cases additional coiling should be considered. In patients with available angiographic follow-up a high occlusion rate was observed. Aneurysm size and other patient selection criteria should be the substance of future research in order to be able to select those cases that profit from this type of treatment in the acute setting.
CAT and HDB contributed equally.
Contributors MFMtB: Design of the work, data acquisition, data analysis and interpretation, drafting the manuscript. MJ: Design of the work, data acquisition, critically revising the work. JdV: Data acquisition, critically revising the work. JAG: Critically revising the work. RA: Data analysis, critically revising the work. SHM: Data acquisition, critically revising the work. RRau: Data acquisition, critically revising the work. JN: Data acquisition, critically revising the work. RRaj: Data acquisition, critically revising the work. AKW: Data acquisition, critically revising the work. ASP: Data acquisition, critically revising the work. CAT: Design of the work, data acquisition, data analysis and interpretation, critically revising the work, drafting the manuscript. HDB: Design of the work, data acquisition, data analysis and interpretation, critically revising the work, drafting the manuscript. All authors gave final approval of the version to be published.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests JdV and HDB: consultancy relationships to Stryker Neurovascular. CT: grants, personal fees and non-financial support from Microvention; personal fees and non-financial support from Stryker Neurovascular; grants, personal fees, and non-financial support from Acandis (all outside the presented work). RRau: consultant for Stryker Neurovascular and Microvention. AKW: consultant for Stryker Neurovascular, research grants from Philips Healthcare and Wyss Institute (all outside the presented work). ASP: consultant for Medtronic Neurovascular and Stryker Neurovascular; research grants from Medtronic Neurovascular and Stryker Neurovascular (all outside the presented work).
Ethics approval The study protocol (code 2017-3511) was reviewed and approved by the local institutional review board of the Radboud University Medical Center.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement Data are available upon reasonable request from the corresponding author.
Presented at Oral presentation at ESMINT, 6-8 September 2018, Nice, France. Oral presentation at the DGNR (Deutsche Gesellschaft für Neuroradiologie eV), 3-6 October 2018, Frankfurt, Germany.
Patient consent for publication Not required.
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.