Identifying risk factors for perioperative thromboembolic complications in patients treated with the Woven EndoBridge device

Abstract

Introduction

Woven EndoBridge (WEB) is a new endovascular treatment option for wide necked bifurcation aneurysms. Results from the WEB-IT trial showed a 0% risk of thromboembolic complications within 30 days post-op but the rate reported in the literature is as high as 10%. We are exploring potential risk factors associated with immediate thromboembolic complications in patients treated with the WEB device.

Methods

Retrospective study of forty-two patients with intracranial aneurysms who were treated with WEB at a single center from 2019–2021. Data was collected on patient demographics, comorbidities, aneurysm characteristics, procedural details, and hospital course. Bivariate analyses were performed to compare patients who experienced a periprocedural ischemic stroke to those who did not. Multiple logistic regression modeling was performed to identify independent risk factors for thromboembolic complications.

Results

Of the 42 patients that were treated with WEB, 6 suffered an ischemic stroke (AIS). These patients were more likely to have an underlying diagnosis of arrythmias (p value = 0.007). Furthermore, they had a median angle of 32.0° in the true neck view on diagnostic angiogram compared to 19.5° (p value = 0.046). Lastly, they had a longer procedure length of 228 min compared to 178 min (p value = 0.002). Patients with thromboembolic complications had a longer length of stay in the hospital and worse outcomes at three months follow up. On logistic regression modeling, these risk factors did not reach statistical significance.

Conclusion

Risk factors of thromboembolic complications after WEB placement include cardiac arrythmias, acute aneurysmal angle in the true neck view and a longer procedure length.

Introduction

In recent years, novel endovascular therapies have been developed for the treatment of intracranial aneurysms. ¹ These therapies include coiling and flow diverter stents, whose safety and efficacy profiles are well-established. ² An innovative device that is perhaps less well-known is the Woven EndoBridge (WEB) device. The WEB device is an example of an endosaccular flow disruptor. It is a self-expanding mesh that is placed within the aneurysmal sac and promotes thrombosis at the neck of the aneurysm, leading to complete occlusion over time. The WEB device is particularly suited for the treatment of wide necked bifurcation aneurysms (WNBAs) which have historically been a challenge to treat endovascularly. The WEB device has been used successfully to treat WNBA of the anterior and posterior circulation in Europe for the last ten years. ³ In 2019, after the results of the landmark WEB-IT study were published, the WEB device gained FDA approval for use in the United States. ⁴ Since gaining approval, the WEB device has been used to treat ruptured WNBAs and WNBAs that failed previous treatment.^5,6

A number of complications have been reported with the WEB device. Perioperative complications including thromboembolic events leading to infarction, intraoperative aneurysm rupture or other bleeding events. ⁷ Multiple studies have been published that suggest that the rate of perioperative complications after WEB treatment is low.^3,8 In the WEB-IT study, the only major adverse event was a delayed parenchymal hemorrhage three weeks status post WEB placement. ⁹ The WEB-IT study investigators did report that seven out of the 150 patients experienced a minor ischemic stroke perioperatively but in nearly all of the patients, the symptoms self-resolved with no permanent neurological sequalae. ⁹ In post-market, multicenter studies with large sample sizes, the rate of thromboembolic complications is higher than the WEB-IT study. ¹⁰ In systematic reviews that aim to evaluate the safety of the WEB device however, the rate of reported perioperative thromboembolic complications is as high as 10%.^3,8 Thromboembolic complications of WEB treatment are an interesting phenomenon. Unlike stents, the WEB device is not placed in the lumen of the parent vessel, leading to a decreased risk for thrombus formation in theory. However, further investigation into the thromboembolic complications of WEB treatment is warranted and can help address if there is a need to keep patients on an antiplatelet regimen. Thus, the main objective of our study is to identify risk factors for perioperative thromboembolic complications in patients who undergo treatment with the WEB device.

Methods

Our study is a retrospective study involving all adult patients with intracranial WNBA of the anterior or posterior circulation who were treated with the WEB device at our institution over a two-year period, from April 2019 till March 2021. A total of 42 patients were included in this study. Each patient had one aneurysm treated with WEB during the course of the neurointerventional procedure. All aneurysms were included regardless of rupture status and any prior treatments. IRB approval was obtained but informed consent was waived given the minimal risk posed by this study.

Electronic medical record review was performed and data was collected on the following variables: patient demographics, medical comorbidities, history of prior aneurysm rupture or family history of aneurysms, smoking status and alcohol use. Data was also collected on aneurysm characteristics especially the number, size and location of the aneurysms. The following aneurysmal measurements were obtained in millimeters from digital subtraction angiography (DSA): average aneurysm diameter from two views, the smallest aneurysm height and the widest aneurysm neck. The aneurysm dome height to neck width (DHNW) ratio was calculated. Angular measurements between the inlet vessel and the long axis of the aneurysm (from neck to highest point on dome) were obtained from DSA. This angle was measured in two views, the true neck view and down the barrel view, and average values were used in the analysis. All measurements were obtained by two authors who independently reviewed the imaging.

Important procedural details were collected as well including the type and dimensions of the WEB device used. In our study, most of the patients were treated with the WEB SL device as opposed to SLS and both WEB-21 and WEB-17 were used. The WEB measurements which consist of diameter and height were recorded. Additionally, the following measurements were obtained to assess how well the chosen WEB size fits into the target aneurysm: WEB diameter minus the average aneurysm width and smallest aneurysm height minus the WEB height.

Lastly, data was collected on hospital course including length of stay and ultimate disposition. Any complication that occurred from the start of the procedure till the time of discharge was documented. Thromboembolic complications were defined as an acute change in the patient's neurological status (NIHSS >4) or evidence of occlusion or infarction on radiographic imaging. Lastly, each patient's functional outcome at 90-days post-treatment as measured by the Modified Rankin scale (mRS) was documented.

Statistical analysis

Bivariate analyses were performed to compare patients who experienced a perioperative thromboembolic complication with those who did not. On bivariate analysis, a p value < 0.05 was defined as reaching statistical significance. Then, a multiple logistic regression model was constructed to identify independent predictors of perioperative thromboembolic complications after WEB treatment. Only the variables that reached a p value < 0.100 on bivariate analysis were included in the logistic regression model.

Categorical variables are presented as numbers and percentages and were compared using a Chi-squared test. Continuous variables with a normal distribution are presented as mean and standard deviation and were compared using a student's t-test. Continuous variables with a non-normal distribution are presented as median and interquartile range and were compared using a Mann-Whitney U test. All statistical analyses were performed using the Stata 16.1 statistical software.

Short description of the procedure

All patients were started on a dual antiplatelet (DAPT) regimen consisting of 81 mg Aspirin and 75 mg Clopidogrel daily one week before the procedure in case they required stenting. Thromboelastography was the Platelet Function Test (PFT) of choice to ensure therapeutic levels of the drugs with at least 60% inhibition of the AA and ADP receptors. On the day of the procedure, the patient was induced using general anesthesia. Transfemoral access was obtained using the right common femoral artery. Upon selective catheterization of the intracranial vessels, digital subtraction angiography in the anterior posterior and lateral projections was performed along with 3D rotational angiography. Aneurysm measurements were obtained and an appropriately sized WEB device was chosen as per the manufacturer's recommendations (Sequent Medical; WEB Device Selection Guide; Microvention). Intraoperative heparin was administered to achieve a clotting time that was twice the patient's baseline or at least 250 s. ACT was repeated every thirty minutes to ensure it was within the therapeutic range for anticoagulation and heparin was readministered if needed. Microcatheters were used to approach the target aneurysm and the WEB device was successfully deployed. Control angiogram was performed to confirm adequate placement. Repositioning was performed if needed. A final DSA run was performed after the WEB was detached. Microcatheters were removed and a vascular closure device was placed at the access site. The patient was taken to the recovery room in stable condition. For patients with unruptured aneurysms who were undergoing elective treatment, they were discharged home on the next day. For patients with ruptured aneurysms who presented with a subarachnoid hemorrhage (SAH), they remained in the hospital until they were stable for discharge and their disposition varied. All patients who were treated solely with the WEB device remained on baby aspirin for six weeks after the procedure unless they suffered a thromboembolic complication in which case DAPT was continued. At six months post-op, follow up diagnostic angiogram was performed to assess the occlusion status of the treated aneurysm.

Results

Out of the 42 patients who were treated with the WEB device, six of those patients experienced a thromboembolic complication (summarized in Table 1). The first patient was a 68-year-old female with an unruptured ACOM aneurysm. Intraoperative course was complicated by left M1 dissection. On post-op day 1, patient was found to have new right sided weakness and new infarcts in the left Middle Cerebral Artery (MCA) division on brain MRI. She was discharged to acute rehab one week later and returned to baseline mRS by 90-days post-op. The second patient was a 47-year-old female with an unruptured right MCA aneurysm that was treated electively with WEB. Few hours after the procedure, she developed left sided weakness and was found to have a right MCA occlusion on CTA. Thrombectomy was performed. Outpouching of the WEB device was noted and the decision was made to place an Atlas stent across the device (Figure 1). She was discharged home on post-op day 3 and returned to baseline mRS at 90 day follow up. The third patient was a 77-year-old female who presented with an unruptured right MCA aneurysm. After elective treatment with WEB, the patient awoke from anesthesia with left sided weakness. Diagnostic angiogram was performed and showed a non-occlusive thrombus around the edge of the WEB in the right MCA. Patient was discharged to a skilled nursing facility on post-op day 12. She passed away on post-op day 79 from an aortic dissection unrelated to the treatment.

Table 1.

Six cases of perioperative thromboembolic complications in our patient cohort.

Case	Age/Gender	Aneurysm Characteristics	Complication	Outcome
1	68F	Unruptured ACOM Aneurysm	Intraop left M1 Dissection. Right sided weakness on postop day 1. Had left MCA occlusion on MRA.	Dispo: Acute Rehab Returned to baseline mRS by 90 days postop
2	47F	Unruptured Right MCA Aneurysm	Left sided weakness on postop day 1. Had right MCA occlusion. Outpouching of WEB observed during thrombectomy. Atlas stent placed across WEB.	Dispo: Home Returned to baseline mRS by 90 days postop
3	77F	Unruptured Right MCA Aneurysm	Left sided weakness upon extubation. Right MCA non-occlusive thrombus & contrast extravasation on CTA	Dispo: Nursing Home Died from unrelated cause before 90-day postop
4	56F	Ruptured Right MCA Aneurysm	Intraop distal Right MCA occlusion. Given Integrilin. Asymptomatic upon extubation.	Dispo: Home with services mRS 3 at 90-day postop
5	43M	Ruptured Left MCA Aneurysm	Sticky Detachment led to WEB shape change, required adjuvant coiling. Intraop left MCA occlusion. Treated with Integrilin. Right sided weakness & aphasia upon extubation.	Dispo: Home Returned to baseline mRS by 90-day postop
6	62F	Ruptured Right Pericallosal Aneurysm	Sticky Detachment Intraop Right ACA occlusion. Thrombectomy performed	Died during hospitalization from infectious complication

ACA: Anterior Cerebral Artery, ACOM: Anterior Communicating Artery, CTA: Computed Tomography Angiography, dispo: disposition, F: female, intraop: intraoperative, M: male, MCA: Middle Cerebral Artery, MRA: magnetic resonance angiography, mRS: modified Rankin scale, postop: postoperative, WEB: Woven Endobridge Device.

Figure 1.

(a) the deployed WEB device is shown with visible outpouching as indicated by the white arrow. (b) Right MCA occlusion is seen. (c) An Atlas stent is placed across WEB device at aneurysm neck.

The fourth patient was a 56-year-old female who presented with a Hunt Hess Grade 2 mFisher 4 SAH due to a ruptured right MCA aneurysm that was treated with the WEB device. Intraoperatively, the patient developed platelet aggregation in the distal M1 segment near the aneurysm neck and was placed on an Integrilin drip. She awoke from anesthesia with no new focal deficits. She was discharged home with services on post-op day 13 and had moderate disability (mRS 3) at 90-day follow up. The fifth patient was a 43-year-old male with Hunt Hess Grade 1 mFisher 1 SAH due to ruptured left MCA aneurysm which was treated with WEB embolization. During the procedure, there was evidence of a sticky detachment. While detaching the WEB device, it inverted and led to a central area of recanalization within the aneurysm that was then treated with adjuvant coiling (Figure 2). During the final run, a small area of non-occlusive thrombus was noted in the distal left MCA and the patient received an Integrilin bolus and was placed on a low dose heparin drip. He awoke from anesthesia with RUE weakness and word finding difficulty. He was discharged home on post-op day 9 and returned to his baseline mRS by 90-days post-op. The last patient was a 62-year-old female smoker with history of multiple aneurysms who presented with Hunt Hess 4 mFisher 1 SAH due to ruptured right pericallosal aneurysm that was treated with WEB. The WEB was noted to have a sticky detachment. Control angiogram showed a right ACA occlusion which was treated with thrombectomy (Figure 3). The patient was not extubated. Hospital course was complicated by status epilepticus and hydrocephalus. The patient passed away on post-op day 12. In summary, our thromboembolic complication rate was 14% and our mortality rate was 4%.

Figure 2.

(a) WEB device is seen pre-detachment. (b) Post detachment, there is an inversion or shape change in WEB device as indicated by white arrow. (c) Adjuvant coiling of the area of recanalization was performed and coil mass is shown.

Figure 3.

(a) WEB device is shown pre-detachment. (b) Post-detachment, the proximal marker of the WEB device has moved as indicated by white arrow. (c) Right ACA occlusion is seen. (d) Restoration of blood flow has occurred post-thrombectomy.

As depicted in Table 2, there were no differences in age, gender, race or baseline functional status among patients who experienced a thromboembolic complication and those who did not. In terms of medical comorbidities, a greater percentage of patients (33%) who experienced a complication had an underlying diagnosis of cardiac arrythmias, specifically atrial fibrillation, than patients without complications (2%) (p value = 0.007). There were no statistically significant differences between both groups in regards to aneurysm location or rupture status. The average aneurysm dimensions including diameter, dome height and neck width were comparable between both groups. However, patients who experienced a thromboembolic complication had aneurysms with an angulated access; the average angle measured in the true neck view was an acute angle of 32° (p value = 0.046). Patients in both groups were treated with WEB devices of a comparable diameter and height. The differences between the dimensions of the WEB device and the size of the aneurysm were similar. Another major difference between both groups was that patients who experienced a complication had a longer average procedure length of 228 min (p value = 0.002). Lastly, the outcomes differed between both groups. Patients who experienced a thromboembolic complication had longer ICU and hospital stays and often required acute or subacute rehab services upon discharge. Additionally, patients who experienced a complication were less likely to have a low mRS score at 90-days.

Table 2.

Bivariate analyses comparing patients who experienced a thromboembolic complication versus those who did not.

Characteristic	No complication (n = 36)	Thromboembolic complication (n = 6)	P value
Age (years)	60.30 ± 9.10	58.80 ± 12.90	0.740
Male Gender	7.00 (19.4%)	1.00 (16.7%)	0.873
Race/Ethnicity			0.170
Hispanic	20.00 (55.6%)	3.00 (50.0%)
African American	7.00 (19.4%)	3.00 (50.0%)
Caucasian	9.00 (25.0%)	0.00 (0.0%)
CAD	3.00 (8.3%)	0.00 (0.0%)	0.463
CVD	4.00 (11.1%)	1.00 (16.7%)	0.697
Arrythmia	1.00 (2.8%)	2.00 (33.3%)	0.007
HTN	22.00 (61.1%)	5.00 (83.3%)	0.298
DM	6.00 (16.7%)	1.00 (16.7%)	1.000
HLD	9.00 (25.0%)	1.00 (16.7%)	0.657
Smoker status			0.567
Never	14.00 (38.9%)	3.00 (50.0%)
Current	8.00 (22.2%)	2.00 (33.3%)
Former	14.00 (38.9%)	1.00 (16.7%)
Alcohol Use	7.00 (19.4%)	1.00 (16.7%)	0.873
Family Hx of Aneurysm	4.00 (11.1%)	2.00 (33.3%)	0.150
Prior Hx of Rupture	3.00 (8.3%)	0.00 (0.0%)	0.463
Baseline mRS (<3)	35.00 (97.2%)	5.00 (83.3%)	0.139
No. of Aneurysms	1.00 (1.0–2.0)	1.50 (1.0–3.0)	0.548
No. of Aneurysms Treated	1.00 (1.0–1.0)	1.0 (1.0–1.0)	1.000
Aneurysm Diameter (mm)	5.50 ± 1.50	6.50 ± 1.50	0.136
Dome Height (mm)	5.20 ± 1.60	4.50 ± 1.00	0.358
Neck Width (mm)	4.40 ± 1.30	5.20 ± 1.90	0.283
DHNW ratio	1.20 ± 0.30	0.90 ± 0.70	0.068
Angle (true neck) (°)	19.50 (11.1–29.2)	32.00 (30.5–43.7)	0.046
Angle (down the barrel) (°)	22.60 (18.2–31.7)	28.20 (12.1–45.8)	0.914
Previously Treated Aneurysm	2.00 (5.6%)	0.00 (0.0%)	0.554
Ruptured Aneurysm	7.00 (19.4%)	3.00 (50.0%)	0.104
Location of Aneurysm			0.133
ICA	5.00 (13.9%)	0.00 (0.0%)
MCA	16.00 (44.4%)	4.00 (66.7%)
ACOM	10.00 (27.8%)	1.00 (16.7%)
Basilar	4.00 (11.1%)	0.00 (0.0%)
PCOM	1.00 (2.8%)	0.00 (0.0%)
Pericallosal	0.00 (0.0%)	1.00 (16.7%)
Procedure Length (min)	178.00 (151.0–189.0)	228.00 (197.00–285.00)	0.002
Web Type (SL)	33.00 (91.7%)	6.00 (100%)	0.463
Web Diameter (mm)	6.40 ± 1.70	7.00 ± 0.90	0.401
Web Height (mm)	3.40 ± 0.90	3.50 ± 0.50	0.951
WEB Width – Average Aneurysm Width (mm)	0.90 ± 0.80	0.50 ± 1.00	0.316
Aneurysm Smallest Height – WEB Height (mm)	1.60 ± 1.00	1.00 ± 0.80	0.137
Adjuvant Coiling	1.00 (2.8%)	1.00 (16.7%)	0.139
Adjuvant Stenting	1.00 (2.8%)	1.00 (16.7%)	0.139
ICU LOS	0.00 (0.0–2.0)	7.00 (3.0–8.0)	0.008
Hospital LOS	1.00 (1.0–3.0)	11.00 (7.0–12.0)	0.010
Disposition			0.044
Home	27.00 (75.0%)	2.00 (33.3%)
Home w/ Services	3.00 (8.3%)	1.00 (16.7%)
Acute Rehab	5.00 (13.9%)	1.00 (16.7%)
Subacute Rehab	1.00 (2.8%)	0.00 (0.0%)
Nursing Home	0.00 (0.0%)	1.00 (16.7%)
In-hospital Death	0.00 (0.0%)	1.00 (16.7%)
90-day mRS (<3)	33.00 (91.7%)	3.00 (50.0%)	0.007

Bold value means the p value reached statistical significance. ACA: Anterior Cerebral Artery, ACOM: Anterior Communicating Artery, CAD: Coronary Artery Disease, CVD: Cerebrovascular Disease, DHNW: Dome Height to Neck Width, DM: Diabetes Mellitus, DSA: Diagnostic Angiography, HTN: Hypertension, HLD: Hyperlipidemia, Hx: History, ICA: Internal Carotid Artery, LOS: Length of Stay, MCA: Middle Cerebral Artery, min: minutes, mm: millimeters, mRS: Modified Rankin Scale, PCOM: Posterior Communicating Artery, SAH: Subarachnoid Hemorrhage, WEB: Woven EndoBridge Device.

The variables that were found to be statistically significant on bivariate analyses with a p value less than 0.100 were: the presence of cardiac arrythmias, small DHNW ratio, acute angle in the true neck view and longer procedure length. They were included in a multiple logistic regression model (Table 3) and no independent predictors of thromboembolic complications after WEB treatment were identified.

Table 3.

Multiple logistic regression model for independent predictors of thromboembolic complications.

Independent predictor	Adjusted OR (95% CI)	P value
Arrythmia	4.80 (0.041–563.431)	0.519
DHNW Ratio	0.35 (0.004–31.276)	0.646
Angle (true neck) (°)	1.06 (0.994–1.132)	0.073
Procedure length (min)	1.03 (0.999–1.077)	0.053

DHNW: Dome Height to Neck Width, min: minutes.

Discussion

In our study, six (14.3%) out of the forty-two patients who were treated with the WEB device suffered a perioperative thromboembolic complication. Three of the patients were noted to have thrombus formation intraoperatively. The other three patients either awoke from anesthesia with a new focal deficit or experienced a neurological exam change within the first twenty-four hours after the procedure. One patient had an operative course complicated by known MCA dissection which may occur with any neurovascular procedure and can't be attributed to the WEB device. Of the four patients with new neurological deficits, three patients returned to their baseline functional status at 90-days post op. Patient 4 reported moderate disability at 90-days. However, it is important to note that she presented with a SAH and was asymptomatic in the immediate aftermath of the thromboembolic complication. Patient 6 died in the hospital. This patient presented with a high grade SAH and had a very complicated hospital course. Patient 3 was discharged to a skilled nursing home where she passed away before her 90-day follow up visit from an unrelated cause. Thus, our high morbidity and mortality rate can't entirely be attributed to thromboembolic complications due to WEB placement.

It is important for the authors to present additional possible explanations for high rates of thromboembolic complications associated with WEB treatment. Operator experience is a main factor to consider as there is a significant learning curve for selecting an appropriately sized WEB device and deploying it. However, at our institution, all of the neurointerventionalists who treat patients with the WEB device are experienced operators who have successfully performed more than ten procedures each. Furthermore, the WEB device has very specific indications pertaining to its on-label use. This raises the question that perhaps the off-label use of the WEB device is associated with a higher risk of complications. However, only a small minority of the aneurysms in this patient cohort, such as one PCOM and one pericallosal aneurysm, constituted an off-label use of the WEB device. These two aneurysms were treated with the WEB device because they were wide necked aneurysms whose size was suitable for WEB and case reports of successfully treating PCOM and sidewall aneurysms with WEB currently exist in the literature.^11,12 Another explanation for a high rate of thromboembolic complications is if a patient population is not on sufficient antiplatelet therapy, leading to early platelet aggregation with the WEB device. However, all of our patients were prescribed dual antiplatelet therapy one week prior to the procedure. It is possible that given the poor socioeconomic status and low health literacy of our patients, the compliance with DAPT in our patient population is suboptimal. Furthermore, a large percentage of our patient population is African American, who are more likely to be aspirin and clopidogrel nonresponders. ¹³ It is important to recognize that the conditions from the WEB-IT trial were ideal in that 3D custom models were made for operators to specifically train on each case prior to the deployment of the WEB. This rehearsal approach may in fact reduce the risk for complications. Prospective trials have stricter inclusion and exclusion criteria than post-market approval cohorts and this might lead to higher complication rates.

That being said, thromboembolic complications are a serious adverse event that can occur with on-label WEB treatment by experienced operators in patients on DAPT, and have irreversible consequences. As of date, few studies have been published in the literature for the explicit purpose of identifying risk factors for thromboembolic complications during/after WEB treatment. Identifying risk factors is the logical first step to understanding this phenomenon and taking steps to prevent it. While the results of our multiple logistic regression model yielded no clear independent predictor of thromboembolic events, the results of our bivariate analysis are illuminating as to which variables may be associated with thromboembolic complications.

Our results show that a higher percentage of patients in the thromboembolic complications group had atrial fibrillation. This result is not surprising as atrial fibrillation is a known risk factor for ischemic stroke. ¹⁴ Possible theories exist as to why patients with known risk factors are at a heightened risk for perioperative stroke. ¹⁵ These theories include changes to the patient's medication regimen (blood thinners) in the perioperative period, manipulation of the vasculature during endovascular procedures causing embolization, induction of acute arrythmias under anesthesia, etc. Another variable that may be associated with thromboembolic complications as per our study is procedure length. Increased procedure time has been linked to increased rate of complications during neurointerventional procedures. ¹⁶ This is likely because procedure length can be seen as a surrogate marker of the complexity of the procedure. In patients with tortuous vessels, anatomic variations, long standing vascular disease, and complex shaped aneurysms, performing an endovascular treatment with WEB is challenging and will take longer. Additionally, the longer the catheter remains inside the vasculature and the vasculature is actively being manipulated, the greater the likelihood that mechanical trauma will result, leading to an iatrogenic injury. Conversely, if an intraoperative complication such as thrombus formation occurs, the procedure length will increase.

For the most part, patients in both groups had comparable sized aneurysms that were treated with similar sized WEB devices. The only difference between both groups pertained to the aneurysmal angles. Patients who experienced a thromboembolic complication had, on average, an acute angle of 32.0 in the true neck view. In the literature, difficulty with deploying the WEB device to treat aneurysms with an angulated access has been described.^17,18 Furthermore, some case reports have been published that describe thromboembolic events in patients treated with WEB because of device protrusion.^19,20 Perhaps the link between angulated access and thromboembolic events is such that if the device is not deployed or positioned appropriately and there is some protrusion into the lumen of the parent vessel, then a thrombus may result. This would be similar to how malpositioned coils can herniate into the parent vessel and lead to thrombus formation. ²¹ The majority of patients in our thromboembolic complication group developed a thrombus near the site of the WEB treated aneurysm. In one patient, an outpouching of the WEB device was observed. At least two of the patients had a sticky detachment which further supports this theory. Sticky detachment of the WEB refers to a slight change in the position of the WEB as it is being detached. Right now, there is no evidence to suggest that wide necked aneurysms with an angulated access should not be treated with the WEB device. Additional studies exploring issues with WEB deployment in aneurysms with a challenging access are needed.

50% of the patients with thromboembolic complications presented with a ruptured aneurysm as compared to 19% in the no complication group. This finding did not reach statistical significance, likely due to the low number of patients. Rupture status may be an important factor for thromboembolic complication given that these patients are not routinely pretreated with dual antiplatelet therapy, and while they are heparinized intraoperatively, heparin is typically administered at a later point in the procedure. In addition to these differences compared to the unruptured cohort, which could explain why the risk for intraprocedural complications in ruptured patients is high, subarachnoid hemorrhage is a known pro-thrombotic state. ²²

As mentioned, the literature on thromboembolic complications after WEB treatment is sparse. However, a discussion of the available data is warranted. Prior to the widespread use of the WEB device, the mainstay of endovascular treatment for WNBA was stent assisted coiling. Once it became available, WEB was an attractive alternative because the complication rate was significantly lower and there was no need for antiplatelet therapy. ²³ However, despite the low rate, thromboembolic complications are still the main type of complication that occur with the WEB device. ²⁴ In a systematic review by Lv et al., the authors report that MCA bifurcation aneurysms are associated with a higher rate of thromboembolic complications than aneurysms of the posterior circulation. ⁸ However, in a meta-analysis by Zhang et al., anterior circulation and ruptured status were not found to be significant risk factors for thromboembolic events which is consistent with our results. ⁷ In a comprehensive study on risk factors for WEB-related procedural complications by Goertz et al., a low DHNW ratio was found to be a statistically significant predictor but, in our study, it only reached borderline significance. ²⁵ The general consensus in the literature appears to be that older generations of the WEB device were associated with higher rates of complications. ⁸ Studies show that WEB SL has a lower complication rate than WEB DL and web 17 has a lower complication rate than previous models.^26,27 In our study, we did not differentiate between the different models of WEB.

The major limitation of our study is the small sample size. Forty-two patients were enrolled in the study and only six patients experienced the outcome of interest. Perhaps the reason why the results of our logistic regression model were inconclusive was because the study was not sufficiently powered. Larger sized studies are needed to further investigate this topic. Additionally, our study may be subjected to bias due to its retrospective design. To adequately identify the risk factors and more specifically, to determine how to prevent thromboembolic complications, a randomized controlled trial is needed.

Conclusion

In summary, the WEB device is a safe treatment modality for WNBA. It is an improvement on existing endovascular technology. While the rate of thromboembolic complications reported in the WEB-IT study was negligible, the rate reported in the literature is around 10%. At our institution, our rate of thromboembolic complications was 14%. However, not all patients were symptomatic and half of them made a complete recovery. Three patients suffered a poor prognosis but it may have been compounded by their initial presentation with SAH. Our study identified three possible risk factors for thromboembolic complications which are as follows: presence of atrial fibrillation, acute aneurysmal angle in the true neck view and increased procedure length. These findings suggest that patients who should be monitored for risk of thromboembolic events are patients who have known risk factors for ischemic stroke at baseline. Moreover, WEB treatment carries a risk of thromboembolic complications just by virtue of the fact that it is an endovascular procedure. The one patient who suffered a dissection could have experienced this complication regardless of the device chosen to treat their aneurysm. The only characteristic inherent to WEB treated aneurysms that is a potential risk factor is an acute aneurysmal angle. It likely signifies a more challenging access and difficulty with proper deployment. While not statistically significant, the rupture status of the aneurysm also likely effects the risk of thromboembolic complications. Our study is one of hopefully many which aim to evaluate the safety profile of emerging endovascular technology. Larger prospective studies are required to further delineate the risks of treatment and guide patients and their families in decision making.