Intra‐arterial Selective Bevacizumab Administration in the Middle Meningeal Artery for Chronic Subdural Hematoma: An Early Experience in 12 Hemispheres

Abstract

Background

Chronic subdural hematoma (cSDH) has a rising incidence associated with an increasing burden of disability and mortality worldwide. Vascular endothelial growth factor plays an integral role in the inflammation and formation of subdural membranes responsible for the origin and propagation of cSDH. We report an early experience of intra‐arterial bevacizumab, a vascular endothelial growth factor receptor antagonist, to the middle meningeal artery of 12 hemispheres in 8 patients with cSDH.

Methods

Eight patients with either unilateral or bilateral cSDH received intra‐arterial infusion of 2 mg/kg bevacizumab into the middle meningeal artery of each treated hemisphere. The primary outcome was hematoma recurrence or reaccumulation requiring surgical drainage or middle meningeal artery embolization within 3 months posttreatment.

Results

Of 12 hemispheres treated, no treatment‐related complications were reported. Median duration of follow‐up was 5 months (interquartile range 3–7.5). By 3 months posttreatment, no patients experienced hematoma recurrence or reaccumulation. One patient required concurrent evacuation at the time of bevacizumab administration. There were no major strokes or mortality within 3 months. Four hemispheres (33.3%) demonstrated complete radiographic hematoma resolution by 3 months. All hemispheres achieved 50% reduction in hematoma size by 3 months.

Conclusion

For all hemispheres treated, there was no hematoma recurrence or progression requiring surgical drainage or middle meningeal artery embolization within 3 months except 1 who required concurrent evacuation 24 hours after treatment. Our initial experience supports bevacizumab as a novel, potentially viable agent for cSDH treatment in select patients. Future studies in larger cohorts are necessary to confirm efficacy and safety and appropriate dosing.

Nonstandard Abbreviations and Acronyms

cSDH

chronic subdural hematoma

IA

intra‐arterial

MMA

middle meningeal artery

MMAE

middle meningeal artery embolization

Clinical Perspective

• Chronic subdural hematoma that is symptomatic or associated with mass effect is traditionally treated with open surgical procedures or endovascularly with occlusive embolic agents or coils.

• Bevacizumab, a vascular endothelial growth factor inhibitor, may tip the balance toward resorption of blood products, and it may provide targeted treatment to subdural membranes.

• Our experience demonstrates initial, positive results with intra‐arterial injection of bevacizumab into the middle meningeal artery.

There has been increased interest in the pathophysiology and treatment of chronic subdural hematoma (cSDH). The prevalence of symptomatic cSDH treatment has risen with the aging population and more widespread use of antithrombotics. ¹ , ² Surgical evacuation, the current standard of care, remains with high rates of recurrence, reported between 8% and 37%. ³ The persistence of a vascular endothelial growth factor (VEGF)‐mediated inflammatory cascade plays an integral role in the formation and propagation of subdural membranes involved in the development and recurrence of cSDH. ⁴ , ⁵ , ⁶ , ⁷ , ⁸ VEGF, angiopoietin‐2, and hypoxia‐inducible factor 1α constitute the 3 key angiogenic factors involved in cSDH inflammation. ⁷ , ⁹ Traditionally involved in vascular repair, the pathological upregulation and downregulation of VEGF‐mediated intracellular pathways cause the continuous creation of fragile blood vessels, leading to extravasation and recurrent microbleeds. VEGF regulates endothelial cell survival through the phosphatidylinositol 3‐kinase/Akt/endothelial nitric oxide synthase pathway. ¹⁰ Along with interleukin‐6, it activates the Ras/mitogen‐activated protein kinase/extracellular signal‐regulated kinase pathway involved in endothelial cell proliferation upstream of the transforming growth factor β/activin receptor‐like kinase 1 pathway, which is essential for the formation and remodeling of new vessels. ¹¹ VEGF is produced by macrophages, plasma cells, and endothelial cells of the fragile microcapillaries of the outer membrane and is found in higher concentrations in subdural fluid compared to the serum. ¹²

Middle meningeal artery embolization (MMAE) with liquid embolic agents, particles, or coils is a new treatment approach for cSDH, and evidence has increasingly supported its safety and effectiveness. ¹³ , ¹⁴ , ¹⁵ , ¹⁶ The aim is to “devascularize” the dura, stop the continuous leakage of blood, and tip the balance toward resorption of the hematoma. ¹⁷ As a minimally invasive approach, MMAE is an appealing option, either as primary or as an adjunct to surgical evacuation, for treating patients with smaller hematoma sizes, mild symptoms, or those who are poor open surgical candidates. ¹⁸ , ¹⁹ , ²⁰ , ²¹ Several recent randomized controlled trials, such as EMBOLISE (Embolization of the Middle Meningeal Artery With ONYX Liquid Embolic System for Subacute and Chronic Subdural Hematoma; NCT04402632), ²² MAGIC‐MT (Managing Non‐acute Subdural Hematoma Using Liquid Materials: A Chinese Randomized Trial of MMA Treatment; NCT04700345), ²³ and STEM (SQUID Trial for the Embolization of the Middle Meningeal Artery for Treatment of Chronic Subdural Hematoma; NCT04410146), ²⁴ have demonstrated favorable results supporting MMAE. ¹⁵ , ²⁵ However, the optimal agent and embolization technique are yet to be identified, and the procedure can prove to be complicated, with the risk of unintended migration of embolic material from the middle meningeal artery (MMA) to eloquent neural structures through anastomoses, longer procedural times, and the need for distal catheterization. ²⁶ , ²⁷

Intra‐arterial (IA) delivery of a nonobstructive pharmacological agent has the potential to mitigate the current limitations of MMAE. Bevacizumab is an anti‐VEGF monoclonal antibody that inhibits angiogenesis and has widespread intravenous and intravitreal therapeutic uses. ¹⁷ , ²⁸ We have previously presented our case report describing positive experience with a single patient with cSDH treated with IA bevacizumab to the MMA. ²⁹ We herein present our follow‐up experience with IA bevacizumab to the MMA for unilateral or bilateral cSDH treatment in 8 patients, for a total of 12 hemispheres. This article serves as an exploratory study evaluating a novel treatment for cSDH using targeted pharmacological therapy, potentially serving as an option for select patients.

Methods

Deidentified data may be made available upon reasonable request to the corresponding author. Observational data collection was approved by the institutional review board with waiver of informed consent. The results were reported in accordance with Strengthening the Reporting of Observational Studies in Epidemiology guidelines. ³⁰

Patient Selection and Treatment

This prospective case series included consecutive adult patients treated with IA bevacizumab presenting with either unilateral or bilateral cSDH, who were referred for endovascular treatment with MMAE between December 2020 and November 2023 with 3 months or longer duration of clinical follow‐up available. Patients aged 18 to 90 years with a premorbid modified Rankin Scale score of 0–3 were considered. Treatment candidacy was determined by the treating physician and referring neurosurgeon based on imaging findings and clinical presentation. Patients with mild to moderate symptoms (ie, headache, gait instability, cognitive complaints, seizures, and sensory symptoms) who did not initially require surgical drainage and who were found to have mixed density hematoma subtype per the Nomura classification on baseline computed tomography were eligible. ³¹ Mixed density hematomas were found to have the highest levels of VEGF concentration. ³² , ³³ After identifying eligible patients, both treatments with MMAE and selective bevacizumab infusion were discussed with the patients and their families. Informed consent was obtained for those who agreed to treatment with bevacizumab.

Using standard angiographic technique, 2 mg/kg of bevacizumab was infused via microcatheter intra‐arterially to the proximal MMA of the hemisphere corresponding to the site of cSDH. Patients with bilateral cSDH received 2 mg/kg IA bevacizumab to each hemisphere. Bevacizumab has widespread use intravitreally in a variety of ocular diseases without known adverse events. ³⁴ Distal catheterization was deemed unnecessary and was not performed where meningo‐orbital anastomoses were observed. The dose of 2 mg/kg represents one fifth of the one used for the IA treatment of recurrent glioblastoma, chosen conservatively to minimize theoretical concerns with wound healing. ³⁵ Bevacizumab solution (AVASTIN, Genentech USA Inc. [Roche Group], South Lake, TX; 100 mg/mL single use vial) was injected over 5–10 minutes depending on the size of the vessel and the position of the microcatheter to ensure adequate distal migration without reflux into the internal maxillary artery. Patients were followed with clinical evaluation and serial imaging with noncontrast computed tomography of the head at 2 weeks, 6 weeks, 3 months, and then as needed until resolution.

Demographics and Outcomes

Demographic information (ie, age, sex, race, past medical history including hypertension, coronary artery disease, hyperlipidemia, tobacco use, atrial fibrillation, prior evacuation, prior antithrombotic use, and presenting symptomatology) was collected from documentation at initial clinical evaluation. Baseline computed tomography imaging was reviewed to determine side(s) of cSDH, subdural membranes, and hematoma size. Procedural characteristics collected included type of access, type of anesthesia, procedural time, intraprocedural complications, and observation of dangerous anastomoses. Follow‐up computed tomography images were reviewed for changes in hematoma size, measured as the greatest thickness on axial view. Hematoma recurrence or need for rescue drainage in the follow‐up period was noted.

The primary outcome was hematoma recurrence or reaccumulation requiring surgical drainage or MMAE within 3 months posttreatment. The safety outcome was any major stroke or any death at 3 months from treatment. The secondary outcome was percent of hemispheres with complete hematoma resolution at 3 months compared to baseline.

Statistical Analysis

Categorical variables were summarized as frequencies with percentages, and continuous variables were analyzed as means with SDs or medians with interquartile range. All statistical analyses were performed using JMP Pro v17.0.0 (SAS Institute Inc., Carey, NC).

Results

Eight patients with cSDH were treated with bevacizumab – 4 unilateral, 4 bilateral – for a total of 12 hemispheres (Table 1). The average age was 73 years (interquartile range 69–79) (Table 1). The most common comorbidity was hypertension (6/8; 75.0%), and 3 (37.4%) patients were receiving antithrombotic therapy on presentation. Seven of the 8 patients treated had symptomatic cSDH. Three patients presented with gait instability and dizziness, 3 patients presented with headache and cognitive symptoms, and 1 patient presented with new onset seizures.

Table 1.

Baseline Demographics

Variable	Patients (n = 8)
Age, y, median (IQR)	73 (69–79)
Male sex, no. (%)	8 (100%)
Race, no. (%)
White	7 (87.5%)
Black	1 (12.5%)
Past medical history, no. (%)
Hypertension	6 (75.0%)
Coronary artery disease	2 (25.0%)
Hyperlipidemia	6 (75.0%)
Atrial fibrillation	1 (12.5%)
Tobacco use	3 (37.5%)
Prior antithrombotic use, no. (%)	3 (37.5%)
Prior hematoma evacuation, no. (%)	1 (12.5%)

IQR indicates interquartile range.

Four patients with unilateral cSDH received 2 mg/kg IA bevacizumab, and 4 patients with bilateral cSDH received a total of 4 mg/kg IA bevacizumab (2mg/kg per hemisphere). The median procedure duration per hemisphere treated was 22 minutes (interquartile range 21–26.3 minutes) (Table 2). There were 2 patients (25%) with dangerous anastomoses. Both patients had a prominent lacrimal branch of the MMA anastomosing with the ophthalmic artery (Figure 1). All procedures were completed with monitored anesthesia care (Table 2).

Table 2.

Procedural Details

Variable	Patients (n = 8)
Access type, no. (%)
Radial	4 (50.0%)
Femoral	4 (50.0%)
Sheath size, no. (%)
5‐French	6 (75.0%)
6‐French	1 (12.5%)
7‐French	1 (12.5%)
Procedural time (minute) per hemisphere, median (IQR)	22 (21–26.3)
MAC anesthesia, no. (%)	8 (100%)
Complication rate, no. (%)	0 (0%)
Presence of dangerous anastomoses, no. (%)	2 (25.0%)

IQR indicates interquartile; and MAC, monitored anesthesia care.

Figure 1.

Bevacizumab treatment in 2 patients with dangerous anastomoses. A and C, AP and (B, D) lateral projections of left MMA injections in 2 patients demonstrating a prominent lacrimal branch of the MMA anastomosing with the opthalmic artery, better demonstrated on lateral projections. Both patients were treated with intra‐arterial bevacizumab without complications. AP indicates anterior posterior; and MMA, middle meningeal artery.

Outcomes

Median time to last follow‐up was 5 months (interquartile range 3–7.5 months) (Table 3). One patient with fluctuating symptoms was taken for surgery 24 hours after treatment, which was considered as a concurrent surgical intervention. Otherwise, there was no incidence of hematoma recurrence or reaccumulation requiring surgical drainage or MMAE within 3 months posttreatment. There were no immediate procedural complications, and no patient had either a major stroke or death within 3 months after treatment. The secondary outcome – defined as complete radiographic resolution at 3 months compared to baseline – was achieved in 33.3% (4/12) of treated hemispheres. All treated hemispheres achieved 50% hematoma reduction at 3 months compared to baseline (Table 3). Six out of 8 patients (75%) were discharged from outpatient care 6 months after treatment with no planned repeat follow‐up imaging. Examples of posttreatment response to bevacizumab are shown in Figures 2, 3, 4.

Table 3.

Hematoma Characteristics and Response to Treatment

Variable	Hemispheres (n = 12)
Laterality, no. (%)
Left	6 (50%)
Right	6 (50%)
SDH thickness at baseline (mm), median (IQR)	14 (12–18.5)
SDH recurrence by 3 mo, no. (%)	0 (0%)
Stroke or death by 1 mo, no. (%)	0 (0%)
Complete SDH resolution by 3 mo, no. (%)	4 (33.3%)
50% hematoma reduction by 3 mo, no. (%)	12 (100%)
Time to last follow‐up (mo), median (IQR)	5 (3–7.5)
Treatment to discharge from care, no. (%)
3 mo	5 (41.7%)
6 mo	4 (33.3%)
>1 y	3 (25%)

IQR indicates interquartile range; and SDH, subdural hematoma.

Figure 2.

Axial noncontrast CT of the head. A patient with unilateral cSDH at baseline and at 6 months after bevacizumab demonstrated complete resolution. cSDH indicates chronic subdural hematoma; and CT, computed tomography.

Figure 3.

Axial noncontrast CT of the head. A patient with bilateral cSDH at baseline and 2 months after bevacizumab showed interval reduction in hematoma thickness bilaterally. cSDH indicates chronic subdural hematoma; and CT, computed tomography.

Figure 4.

Axial noncontrast CT of the head. A patient with bilateral cSDH at baseline and 2 months after bevacizumab showed near‐complete resolution with minimal residual hematoma bilaterally. cSDH indicates chronic subdural hematoma; and CT, computed tomography.

Interval Follow‐Up

One patient with a single hemisphere treated had a prior craniotomy for cSDH evacuation 38 days prior with asymptomatic cSDH recurrence. This patient had severe hepatic coagulopathy rendering femoral access high risk. Radial access (7‐French) and tortuous arch anatomy did not provide sufficient support for the distal catheterization of the MMA required for MMAE with liquid embolic agents, and hence the treatment with bevacizumab was proposed as a potentially safer alternative to treat the recurrence given that it can be administered proximally. This patient demonstrated complete resolution of the hematoma at 6‐month follow‐up. However, this patient passed away from complications of advanced liver cirrhosis at 9 months after treatment. No other mortalities were noted in the extended follow‐up period for all other patients.

One patient, who presented with progressive headache and bilateral cSDH with mass effect, developed fluctuating diplopia and bilateral cranial nerve VI palsies 1 day following treatment, prompting bilateral burr hole evacuation of the hematomas (Figure 5). This was believed to be secondary to elevated intracranial pressure and unrelated to bevacizumab treatment. There were no postsurgical complications related to impaired wound healing, dehiscence, or hemorrhage. This patient also required re‐hospitalization within 30 days of treatment due to encephalopathy and gait instability. Repeat imaging showed no recurrence of bilateral cSDH. The patient was discharged home without complications on levetiracetam due to concern for possible seizures, and made a full recovery back to baseline mental status.

Figure 5.

Images 8 days after IA bevacizumab injection to the bilateral MMAs and 6 days after bilateral burr hole evacuations for cSDH showing uncomplicated wound healing. cSDH indicates chronic subdural hematoma; IA, intra‐arterial; and MMA, middle meningeal artery.

Discussion

We present our initial, expanded experience with 8 patients with cSDHtreated with 2 mg/kg IA bevacizumab per hemisphere administered via selective MMA injection. There were no symptomatic recurrences or hematoma progression requiring surgical drainage by 3 months. No major stroke or death was reported within 3 months. One third of hemispheres treated achieved complete resolution within 3 months. Thus, our initial experience supports bevacizumab as a potentially promising agent for cSDH treatment in select patients.

Understanding the proinflammatory cascade underlying cSDH illustrates new potential therapeutic targets and mechanisms. VEGF plays a key role in angiogenesis of new and immature capillaries, leading to continuous leakage in cSDH. ³² In particular, hematomas with mixed‐density appearance have high rates of exudation and have been shown to have increased VEGF concentrations in the outer membrane and subdural fluid. ³² Consequently, inhibiting VEGF with bevacizumab in the subdural outer membrane is likely therapeutic by breaking the proinflammatory cycle responsible for neovascularization and halting persistent exudation.

Currently available pharmacologic treatments targeting similar proinflammatory processes have had mixed clinical results. Corticosteroids alter transcription of inflammatory proteins, such as chemokines and cytokines, at the gene expression level. ⁶ However, as both standalone or adjunct therapy to surgical evacuation in cSDH, intravenous corticosteroids were associated with higher rates of adverse events, including the need for recurrent surgery, without clear improved functional outcomes. ³⁶ , ³⁷ Hydroxymethylglutaryl coenzyme A reductase inhibitors (ie, statins) have been shown to affect biochemical signaling at several points within the inflammatory and coagulation pathways, and they have been shown to be associated with greater improvement in cSDH hematoma size, particularly in nonsurgically treated patients. ³⁸ , ³⁹ Although statins are useful agents for conservative management, they only have moderate clinical efficacy when used as a standalone therapy. ³⁹ Additionally, tranexamic acid inhibits fibrinolysis, which drives extravasated hematoma liquefaction. Although there are several ongoing clinical trials, current data, demonstrating favorable hematoma reduction and comparable recurrence rates in standalone and adjunctive tranexamic acid therapy, were derived from studies that excluded patients receiving anticoagulation or antiplatelet agents, which encompasses a large proportion of patients with cSDH. ⁴⁰

Endovascular treatment options for cSDH, including coiling and MMAE with particle or liquid embolic agents, have demonstrated favorable safety and efficacy. ³ However, MMAE may cause higher risks and be procedurally difficult in specific patients. The process of administering certain embolic agents may lead to severe pain, which may increase the likelihood of patient movement during catheterization and necessitate use of general anesthesia, posing additional risks in an aging population of patients with cSDH. Technical nuances, such as degree of distal penetration versus proximal occlusion and type of embolic agent, are to be considered for adequate and effective embolization. ¹³ Moreover, the injection of obstructive agents in the MMA can carry significant disabling ischemic risk due to unintended migration of embolic material through anastomoses. Recognition and avoidance of variant anastomoses between the MMA and the eye or the neural tissue require careful evaluation of the ophthalmic artery and the MMA with selective injections. ²⁷

In contrast to MMAE, bevacizumab can be injected even in the presence of dangerous anastomoses, as ischemic risk due to embolic agent migration is essentially eliminated. In our study, 2/8 patients (25%) treated with bevacizumab had dangerous anastomoses that, if attempted to treat with MMAE, would have had higher risk of new deficits and otherwise required either careful distal selection or aborting of the procedure (Figure 1). Moreover, bevacizumab delivery reduces the necessity for distal catheterization and the need for larger vascular access for support. Bevacizumab was also administered in all patients with minimal, conscious sedation, as this treatment avoids the pain associated with embolic agent injection. Lastly, it is still not clear if there are any long‐term potential consequences related to the complete sacrifice of the MMA. This technique preserves MMA patency for potential future interventions.

It is important to note that impaired wound healing is a potential, rare side effect associated with VEGF inhibitors. ²⁹ This risk is particularly relevant when bevacizumab therapy may need to be subsequently followed with surgical hematoma evacuation in cSDH. In our case series, 1 patient underwent bilateral burr hole evacuation 1 day after IA bevacizumab, and no complications related to wound healing were observed (Figure 5). However, further evaluation in larger studies are required to fully characterize the risk of impaired wound healing with IA bevacizumab in patients with cSDH.

Additional adverse effects, including leukopenia, diarrhea, hypertension, neutropenia, gastrointestinal perforation, wound dehiscence, and severe hemorrhage, have been associated with systemic bevacizumab administration for oncologic indications. These adverse events have been described in patients with cancer with recurrent doses of 5–10 mg/kg intravenously. ⁴¹ However, in our series, we conservatively administered one fifth of the dose locally in order to minimize the potential for systemic side effects; no such effects were observed in our patients. The lower dose in this series was derived from the concept that bevacizumab administered into the MMA was thought to be mostly retained at the target site, similar to how contrast injected into the MMA accumulates within the subdural membranes and space, and thus, a lower total local quantity was thought to potentially be adequate to yield a similar therapeutic effect compared to the quantity needed for systemic administrations. Future studies incorporating a dose‐response assessment may help further characterize optimal dosing, including the potential benefit of repeat dosing, for maintaining adequate therapeutic benefit while minimizing administered quantity.

Limitations

Although the findings from these 8 patients show promise, there are inherent limitations to the study. The analysis is limited by biases inherent to observational studies, including confounding and selection bias. Moreover, the nature of uncontrolled observational data precludes inference of causality. The sample size is small, patients are not selected from representative population samples, and there are limited follow‐up data. These factors limit the generalizability of this study. Moreover, there is variability in the management of patients, including preoperative conservative approaches, hematoma size, symptoms, and follow‐up duration. The study also lacked a control group and independent clinical assessments at follow‐up. Further studies in a larger patient cohort with extended follow‐up are necessary to advance our understanding of this treatment. Additionally, it is possible that not all patients are suitable for successful treatment with bevacizumab, and future studies may help elucidate specific clinical, angiographic, or radiographic findings predictive of treatment success.

Conclusion

We present our follow‐up experience of 8 patients with cSDH, for a total of 12 hemispheres, treated with IA bevacizumab to the MMA without hematoma recurrence or reaccumulation within 3 months. This novel, targeted, and selective infusion therapy presents a possible adjunctive or alternative to MMA embolization in select patients. Additional prospective studies are warranted to confirm efficacy and safety.

Sources of Funding

None.

Disclosures

Ajith J. Thomas is a consultant for Qure.ai. Tudor G. Jovin is a consultant for Anaconda, Route 92 Medical,Viz.AI, FreeOx, Blockade Medical, and Methinks. He serves in an advisory capacity for Cerenovus and Contego Medical. He has equity in Corindus. He receives research support from Medtronic and Stryker. All other authors have no disclosures to declare. Ajith Thomas is an Guest Associate Editor for S:VIN and was not involved in the handling or final disposition of this article. Disclosures provided by Ajith Thomas in compliance with American Heart Association’s annual Journal Editor Disclosure Questionnaire are available at https://www.ahajournals.org/editor‐coi‐disclosures. Tudor Jovin serves on the Editorial Board of S:VIN. Editorial Board Members are not involved in the handling or final disposition of submissions.