Influence of procedural factors associated with injection of liquid embolics on outcomes after middle meningeal artery embolization for chronic subdural hematoma

Abstract

Background

Middle meningeal artery embolization (MMAe) is an effective treatment for chronic subdural hematoma. There are various procedural aspects related to liquid embolic (LE) injection that could influence radiographic resolution rates but require further study.

Methods

A retrospective review of consecutive MMAe procedures using LEs performed at a single institution was performed. Technical factors investigated included type of LE, microcatheter position during embolization, number of branches embolized, and depth of penetration. Outcomes included rescue treatment within 90 days, complete hematoma and midline shift (MLS) resolution, and modified Rankin Scale (mRS) 0–2 at 90 days.

Results

A total of 122 MMA embolizations performed on 95 patients (mean age 70.3 years, 66% male) were included. Hematoma recurrence requiring rescue treatment occurred for 12 hematomas (9.8%) among 10 patients. None of the technical factors were associated with rescue treatment, early MLS resolution, final hematoma resolution, final MLS resolution, or 90-day mRS 0–2. Contralateral penetration of embolisate was associated with increased odds of complete hematoma resolution (OR 8.4, 95% CI 1.79–50.38; p = 0.011) at early follow-up (median 2 months). The mean (± SD) hematoma reduction at early follow-up for contralateral penetration was 76.8% (37.2) compared to 56.5% (36) for only ipsilateral penetration (p = 0.047).

Conclusion

Contralateral penetration of LE may be associated with faster complete hematoma resolution but not the final radiographic result. These results require validation in larger cohorts.

Introduction

Chronic subdural hematoma (cSDH) is one of the most common neurosurgical conditions and disproportionately affects the elderly. Hematoma evacuation through a craniotomy or burr holes has historically been the standard of care, but middle meningeal artery embolization (MMAe) has gained popularity as a standalone or adjunct treatment. Multiple randomized controlled trials (RCTs) support the safety and efficacy of MMAe for treatment of cSDH.^1–3 These studies showed reduced hematoma recurrence rates with MMAe compared to surgery alone.^1,3 Numerous observational studies have confirmed the safety of MMAe. ⁴

MMAe is a relatively new procedure and there is no consensus on how it should be performed. Several embolic agents are available including coils, particles, and liquid embolics (LEs), but the latter is most used. There are many technical factors related to the injection of LEs that could influence outcomes. Although most neurointerventionalists aim for distal embolization to penetrate the fragile neovasculature of hematoma membranes, there is limited data to quantify how distal penetration affects hematoma resolution and clinical outcomes. Contralateral penetration, where LE travels across midline to contralateral distal MMA branches, sometimes occurs but its effects have not been studied. The optimal microcatheter position for embolization is unknown. It is also unclear whether embolization of both frontal and parietal MMA divisions affords better outcomes compared to a single division. Finally, multiple LEs are available including Onyx, n-BCA, and Squid without consensus on the best agent.

We sought to analyze various technical factors related to MMAe with LEs to determine how these influence rescue treatment rates as well as radiographic and clinical outcomes. This data could inform optimal techniques for MMAe since early hematoma or midline shift (MLS) resolution may be desired in some patients to rapidly relieve symptomatic mass effect. We hypothesized that more distal penetration of LE ipsilaterally and contralateral penetration would result in enhanced membrane devascularization leading to improved radiographic outcomes.

Methods

Study design

After obtaining institutional review board approval, consecutive MMAe procedures performed at a single institution from 1 July 2020 to 31 July 2024 were reviewed. Inclusion criteria included: (1) Adults ≥ 18 years old, (2) MMAe performed with a LE agent with or without adjunctive coiling, and (3) pre-embolization non-enhanced computed tomography (NECT) available for review. Exclusion criteria included: (1) prior ipsilateral MMAe, (2) predominantly acute SDH (>50%), and (3) SDH due to underlying vascular lesion or arachnoid cyst. Patients were included regardless of whether hematoma evacuation was performed.

The choice of LE agent was based on physician discretion. In general, the most distal microcatheter position that could be safely achieved for embolization was pursued. Injection of the embolisate was carried out under blank roadmap until there was no further forward penetration or an undesirable amount of reflux was seen. The most distal delivery of LE along the vertex was targeted including contralateral penetration to maximally devascularize the hematoma membranes. The number of MMA branches embolized was based on physician discretion. n-BCA was typically diluted to 20% with ethiodol. All procedures were performed under general anesthesia.

Treatment variables

Data was collected from the electronic medical record (EMR) regarding patient demographics, comorbidities, presenting laboratory values, and presenting symptoms. Radiographic data was collected by a single attending neurosurgeon without knowledge of outcomes. SDH thickness was measured on coronal NECT perpendicularly from the inner table of the skull to the edge of the hematoma. The slice showing the maximal thickness was selected for measurement. MLS was measured on axial NECT from the midline to the Foramen of Monroe. SDHs were characterized as chronic (primarily hypodense), subacute (primarily isodense), or a mixture of both. Hematomas were additionally categorized as homogeneous, laminar, separated, or trabecular based on Nakaguchi et al.'s classification. ⁵

Procedural data was collected from a combination of the operative notes and angiographic images. Variables included type of LE, size of the MMA measured just distal to the foramen spinosum on lateral DSA, identity of MMA branch(es) embolized (anterior and/or posterior division), location of microcatheter during injection of LE, whether embolisate reached the midline or contralateral hemisphere, procedure duration, fluoroscopy time, and type/timing of surgical hematoma evacuation. Contralateral penetration was recorded dichotomously based on the presence of any LE contralateral to the midline on a straight AP radiograph. Microcatheter position for embolization was classified as proximal, intermediate, or distal in relation to the branching pattern of the given MMA division visualized on baseline external carotid artery injection (proximal: microcatheter distal to the primary trunk's bifurcation but proximal to the first major branching point; intermediate: microcatheter distal to the first major branching point of the MMA division but proximal to the second; distal: microcatheter distal to the second major branching point). This classification scheme is illustrated in Supplemental Figure 1. When more than one MMA division was embolized, the most distal microcatheter position was considered in the analysis. Final depth of penetration was also evaluated based on distance from the most distal LE to the midline on a straight AP radiograph.

Outcomes

The primary outcome was hematoma recurrence or growth requiring rescue treatment of any modality within 90 days of MMAe. The decision to perform rescue treatment was based on physician discretion. Secondary outcomes included complete SDH and MLS resolution. Two timepoints for radiographic follow-up were selected: the first and last NECTs obtained following MMAe after the early perioperative period (first two weeks) and within one year of the procedure. The modified Rankin Scale (mRS) was derived from clinic notes as close as possible to 90 days following MMAe. The follow-up mRS was dichotomized to 0–2 or 3–6. Procedural complications were recorded from the operative reports and EMR. Major access complications were defined as hematoma requiring transfusion, limb ischemia, or any vascular abnormality requiring surgical repair, while all others were considered minor.

Statistical analysis

Binary logistic regression analyses were performed to determine the associations between the following technical factors and each outcome: type of LE, whether both anterior and posterior divisions were embolized, most distal microcatheter position for embolization, distance from midline of most distal LE penetration, and whether contralateral penetration occurred. The models were further adjusted based on whether hematoma evacuation was performed and for preoperative hematoma thickness for hematoma resolution outcomes and preoperative MLS for MLS resolution outcomes. The model for 90-day mRS included technical factors, age, mRS at the time of presentation, preoperative hematoma thickness, and whether hematoma evacuation was performed. Covariates were selected a priori based on clinical relevance. All analyses were performed with R version 4.1.3.

Results

Patient characteristics

After removing two patients without preoperative imaging, one patient who was made hospice shortly after MMAe, and six patients who underwent MMAe with particles or standalone coiling, a total of 122 MMA embolizations performed on 95 patients were included. The mean (± SD) age was 70.2 years (12.9), and most patients were male (66%). The median mRS (IQR) at presentation was 3 (2–4). There were 43 (45%) patients on antiplatelet or anticoagulation medication on presentation. Most patients had a history of trauma (63%) and were symptomatic from their cSDH (82%). Additional patient data is provided in Table 1.

Table 1.

Demographics, baseline clinical characteristics, and baseline radiographic characteristics.

Variable	N (%)
Mean age in years (SD)	70.2 (12.9)
Sex: Male Female	63 (66.3) 32 (33.7)
Race: White African American Hispanic Other	45 (47.4) 46 (48.4) 2 (2.1) 2 (2.1)
Comorbidities: Hypertension Diabetes Chronic kidney disease Congestive heart failure Alcohol abuse	76 (80) 27 (28.4) 8 (8.4) 9 (9.5) 11 (11.6)
Presenting mRS: 0 1 2 3 4 5	19 (20.2) 18 (19.1) 31 (33.0) 10 (10.6) 15 (16.0) 1 (1.1)
Antiplatelet medication	29 (30.5)
Anticoagulation	22 (23.2)
Median presenting INR (IQR)	1.1 (1–1.14)
Median presenting platelet count (IQR)	209 (178–249)
History of trauma	60 (63.2)
Preoperative symptoms: Headache Neurologic deficit Gait ataxia Altered mental status Seizure	45 (47.4) 27 (28.4) 23 (24.2) 40 (42.1) 4 (4.2)
Median preoperative hematoma thickness in mm (IQR)	15 (11–19)
Median preoperative MLS in mm (IQR)	3.9 (0–7)
Preoperative SDH characteristic: Mixed Chronic Subacute	82 (67.2) 25 (20.5) 15 (12.3)
Preoperative hematoma type: Homogeneous Separated Laminar Trabecular	40 (32.8) 20 (16.4) 26 (21.3) 36 (29.5)
Bilateral hematomas present	31 (32.6)
Laterality: Left Right	67 (54.9) 55 (45.1)

MLS: midline shift, mRS: Modified Rankin Scale, SDH: subdural hematoma.

The median (IQR) preoperative hematoma thickness and MLS were 15 mm (11–19) and 3.9 mm (0–7), respectively. Most hematomas were mixed density (67.2%), followed by chronic (20.5%) and subacute (12.3%). The most common hematoma type was homogeneous (32.8%), followed by trabecular (29.5%), laminar (21.3%), and separated (16.4%).

Treatment characteristics

Onyx 18 was the most common LE used (81%). Adjunctive coil embolization was performed in 12 patients (12.3%). Hematoma evacuation was performed in 64 patients (67.4%) and timing/type of surgery is described in Table 2. Both anterior and posterior MMA branches were embolized in 31.9% of embolizations. Microcatheter position was proximal in 28.7%, intermediate in 45.9%, and distal in 25.4%. As shown in Supplemental Figure 2, this classification scheme resulted in three discrete microcatheter positions in relation to the foramen spinosum. The median (IQR) procedure duration was 120 min (84.5–150). Procedural complications occurred in five patients (5.3%). These included two major access site complications, one minor access site complication, one ischemic stroke, and one iatrogenic MMA injury resulting in fistula formation.

Table 2.

Treatment characteristics.

Variable	N (%)
Access: Femoral Radial	66 (69.5) 29 (30.5)
Median diameter of MMA trunk in mm (IQR)	1.2 (1.0–1.4)
Liquid embolic: Onyx n-BCA Squid	99 (81.1) 12 (9.8) 11 (9.0)
Location of embolization (liquid embolic): Anterior branch Posterior branch Both branches	55 (45.1) 28 (23.0) 39 (31.9)
Location of coil placement: Trunk Anterior branch Posterior branch No coils used	4 (3.3) 7 (5.7) 4 (3.3) 107 (87.7)
Number of coils used: 0 1 2 3	107 (87.7) 4 (3.3) 5 (4.1) 6 (4.9)
Midline penetration of liquid embolic	54 (44.3)
Contralateral penetration of liquid embolic	16 (13.1)
Median fluoroscopy time in minutes (IQR)	48 (33.5–70.1)
Median procedure duration in minutes (IQR)	120 (84.5–150)
Surgery performed*	79 (64.8)
Timing of surgery compared to embolization:^ Before After	55 (69.6) 24 (30.4)
Type of surgery performed:^ Burr hole Craniotomy Subdural evacuating port system	35 (44.3) 14 (17.7) 30 (38.0)

MMA: middle meningeal artery.

*Not including for recurrent hematomas.

^Denominator is 79.

Hematoma recurrence

Hematoma recurrence requiring rescue treatment occurred for 12 hematomas (9.8%) among 10 patients. The median (IQR) time to retreatment was 11 days (7.25–18.75). Detailed clinical and radiographic data for these patients are shown in Supplemental Table 1. None of the technical factors were associated with rescue treatment (Supplemental Table 2).

Early radiographic outcomes

Early radiographic outcomes were evaluated at a median (IQR) of 2 months (1–3) after MMAe. The rates of complete hematoma resolution and MLS resolution at first radiographic follow-up were 32.7% and 69.3%, respectively. Improvement in hematoma thickness and MLS over time are shown in Figures 1 and 2, respectively. As shown in Table 3, contralateral penetration of LE was associated with increased odds of complete hematoma resolution (OR 8.4, 95% CI 1.79–50.38; p = 0.011), but none of the other technical factors were. The mean (± SD) hematoma reduction at early follow-up for contralateral penetration was 76.8% (37.2) compared to 56.5% (36) for only ipsilateral penetration (p = 0.047). As a post-hoc sensitivity analysis a linear regression evaluating % reduction in hematoma thickness that included the same variables as the logistic regression was performed and contralateral penetration was also associated with % thickness reduction (β = 34.32, 95% CI 11.16–57.48; p = 0.004). An additional sensitivity analysis was performed to evaluate technical factors associated with complete hematoma resolution at first follow-up while stratifying patients based on whether up-front hematoma evacuation was performed (Supplemental Table 3). Contralateral penetration approached statistical significance in the surgery group (OR 10.3, 95% CI 1.02–251; p = 0.072) and was significantly associated with complete hematoma resolution in the non-surgical group (OR 18.6, 95% CI 1.49–586.4; p = 0.042). None of the technical factors were associated with complete MLS resolution.

Figure 1.

Distributions and changes in hematoma thickness over time. Thick points in (b) indicate mean values and small points represent individual patients.

Figure 2.

Distributions and changes in MLS over time. Thick points in (b) indicate mean values and small points represent individual patients.

MLS: midline shift.

Table 3.

Logistic regression analysis of technical factors associated with radiographic outcomes at first follow-up.

	Hematoma resolution			MLS resolution
	OR	95% CI	p	OR	95% CI	p
Preoperative hematoma thickness	1.01	0.96–1.05	0.740	−	−	−
Preoperative MLS	−	−	−	0.98	0.82–1.17	0.819
Liquid embolic type: Onyx n-BCA Squid	Ref 0.11 3.43	− 0.01–0.82 0.73–19.0	− 0.070 0.126	Ref 0.21 4.34	− 0.008–2.46 NA	− 0.250 0.994
Embolization of both anterior and posterior MMA branches	0.61	0.22–1.62	0.336	0.32	0.09–1.06	0.065
Microcatheter position: Proximal Intermediate Distal	Ref 0.63 0.40	− 0.20–1.93 0.09–1.67	− 0.415 0.220	Ref 0.74 0.43	− 0.17–2.99 0.07–2.33	− 0.681 0.327
Distance from midline of most distal embolysate penetration	0.99	0.97–1.02	0.570	0.99	1.03	0.666
Contralateral penetration of embolysate	8.40	1.79–50.38	0.011	1.72	NA	0.994
Surgery performed	0.89	0.35–2.33	0.811	1.03	0.24–4.07	0.967

Bolded values indicate statistical significance.

MLS: midline shift, MMA: middle meningeal artery.

Final radiographic outcomes

Final radiographic outcomes were evaluated at a median (IQR) of 4.5 months (2–7.6) after MMAe. The rates of complete hematoma resolution and MLS resolution at final follow-up were 62.7% and 92.3%, respectively. None of the technical factors were associated with complete hematoma or MLS resolution (Table 4).

Table 4.

Logistic regression analysis of technical factors associated with radiographic outcomes at final follow-up.

	Hematoma resolution			MLS resolution
	OR	95% CI	p	OR	95% CI	p
Preoperative hematoma thickness	1.01	0.53–8.09	0.525	−	−	−
Preoperative MLS	−	−	−	1.19	0.92–1.61	0.211
Liquid embolic type: Onyx n-BCA Squid	Ref 0.63 5.53	− 0.15–2.60 0.89–108.12	− 0.510 0.124	Ref 3.75 2.83	− NA NA	− 0.997 0.996
Embolization of both anterior and posterior MMA branches	0.72	0.30–1.73	0.457	0.49	0.05–4.54	0.506
Microcatheter position: Proximal Intermediate Distal	Ref 0.68 0.64	− 0.23–1.96 0.17–2.38	− 0.480 0.510	Ref 1.10 1.27	− 0.11–9.60 0.08–37.34	− 0.932 0.868
Distance from midline of most distal embolysate penetration	0.99	0.96–1.01	0.253	0.99	0.94–1.05	0.783
Contralateral penetration of embolysate	1.46	0.35–6.81	0.612	9.53	NA	0.997
Surgery performed	1.21	0.50–2.90	0.664	0.45	2.10–3.56	0.498

MLS: midline shift, MMA: middle meningeal artery.

Clinical outcomes

The median (IQR) clinical follow-up duration was 3 months (1.5–4). The median (IQR) length of stay was 4 days (1–9) and 68.9% of patients were discharged home. There were 4 (4.2%) mortalities. The final mRS was 0–2 in 64.9% of patients. Older age (OR 0.93, 95% CI 0.88–0.97; p = 0.003) and higher baseline mRS (OR 0.44, 95% CI 0.28–0.66; p < 0.001) were associated with lower odds of 90-day mRS 0–2 (Table 5). Use of n-BCA was also associated with lower odds of 90-day mRS 0–2 compared to Onyx (OR 0.16, 95% CI 0.03–0.73; p = 0.020). This finding was further investigated by comparing baseline characteristics of patients embolized with n-BCA to those embolized with Onyx (Supplemental Table 4). Patients embolized with n-BCA were significantly less likely to have undergone up-front surgery (n-BCA: 50%, Onyx: 69.7%; p = 0.049). Furthermore, all n-BCA patients were symptomatic compared to 79% of Onyx patients, although this did not reach statistical significance (p = 0.118).

Table 5.

Logistic regression analysis of technical factors associated with mRS stabilization or improvement. .

	OR	95% CI	p
Age	0 . 93	0.88–0.97	0.003
Preoperative hematoma thickness	0.99	0.95–1.04	0.961
Liquid embolic type: Onyx n-BCA Squid	Ref 0.16 0.48	− 0.03–0.73 0.08–2.99	− 0.020 0.427
Embolization of both anterior and posterior MMA branches	1.11	0.40–3.11	0.843
Microcatheter position: Proximal Intermediate Distal	Ref 1.74 1.19	− 0.50–6.47 0.24–6.09	− 0.391 0.833
Distance from midline of most distal embolysate penetration	1.01	0.98–1.04	0.533
Contralateral penetration of embolysate	1.95	0.42–10.32	0.406
Baseline mRS	0.44	0.28–0.66	<0.001
Surgery performed	1.65	0.60–4.63	0.333

Bolded values indicate statistical significance.

MMA: middle meningeal artery, mRS: Modified Rankin Scale.

Factors associated with distal or contralateral penetration

LE on average reached closer to the midline with a distal microcatheter position (mean 4.9 mm) compared to intermediate (13.7 mm) or proximal (32 mm; p < 0.001 (Figure 3)). Contralateral penetration was most often achieved from a distal microcatheter position (62.5%) compared to intermediate (37.5%) or proximal (0%; p < 0.001). Among each type of LE, n-BCA reached contralateral more often than Onyx or Squid (33.3% vs 10.1% and 18.2%, respectively; p = 0.048), but it was not associated with distance to midline (p = 0.715). Embolization of both MMA divisions was not associated with contralateral penetration (p = 0.170) but was associated with closer penetration of LE to midline (p = 0.014).

Figure 3.

Depth of penetration based on distance from the most distal liquid embolic to the midline for different microcatheter positions (a), liquid embolic types (b), and number of middle meningeal artery divisions embolized (c).

MMA: middle meningeal artery.

Discussion

There is substantial variability in the way MMAe can be performed, including the type of embolisate used, number and identity of branches embolized, microcatheter positioning, and degree of distal penetration achieved. There is no consensus on the best way to perform the procedure. Traditionally, the goal of MMAe with LE has been to obtain as distal penetration with as much embolisate as safely possible to penetrate the microvasculature of the cSDH membrane that leads to hematoma recurrence. There is little data quantifying how this approach affects radiographic or clinical outcomes and the literature evaluating the effect of distal penetration is limited by small sample sizes, inconsistent definitions of distal penetration, and lack of adjustment for confounders. Additionally, there is data suggesting proximal embolization with coils is effective at reducing hematoma recurrence. ⁶ Therefore, we sought to determine how technical factors related to MMAe affect outcomes. We found that contralateral penetration of LE was associated with quicker hematoma resolution, but not rescue treatment rates or final radiographic outcomes. This suggests that although distal penetration is desirable, it may not be necessary to achieve cSDH resolution.

The rate of rescue treatment for hematoma recurrence or growth was 9.8%, which is similar to the literature on MMAe. In the three RCTs evaluating MMAe, the recurrence rates were between 4.1% and 10%.^1–3 A cohort of 78 patients embolized with particles had a surgical rescue rate of 8.5%. ⁷ Meta-analyses of individual institutional cohorts have reported even lower rates of rescue treatment, ranging between 3% and 4.4%.^8–10 Our rescue treatment rate is still much lower than that for the surgical arms of studies comparing MMAe and surgery, which range from 16.4% to 27.7%.^9,10 Several factors can affect the rate of rescue treatment including different thresholds for retreatment and varying rates of hematoma evacuation at the time of MMAe. None of the procedural variables were associated with rescue treatment, which could have been due to the low incidence of this outcome. One of the few studies comparing depth of penetration with surgical rescue also did not find a significant association. ¹¹ The lack of association between embolic agent and rescue treatment is consistent with a recent meta-analysis. ¹²

Hematoma resolution at first follow-up was associated with contralateral penetration of LE but not distance of embolisate from midline. This suggests deeper penetration contributes to more rapid hematoma resolution but in a non-linear fashion. Ma et al. also found shorter time to hematoma resolution with more aggressive penetration but measured depth of penetration qualitatively and did not evaluate contralateral penetration. ¹¹ Catapano et al. noted that depth of penetration was associated with rapid hematoma resolution but defined distal penetration as within 2.5 cm of the vertex. ¹³ Other studies have not evaluated contralateral penetration's effect on outcomes. Since contralateral MMAe branches have been shown to produce neovascularization of previously embolized ipsilateral MMA territories,^14,15 upfront reduction of this could enhance hematoma resolution. The contralateral contribution may be particularly strong in patients undergoing concomitant surgical evacuation in which distal MMA branches are sacrificed. ¹⁶ One study showed faster hematoma resolution when n-BCA reached the falx, which could have occurred by occluding the falcine MMA branches that anastomose with contralateral distal MMA branches. ¹⁷ Since early hematoma resolution is strongly influenced by upfront surgical evacuation and baseline hematoma thickness, we adjusted for these in the regression models. Most other studies evaluating technical aspects of MMAe have not adjusted for these factors.^7,13,17–21

Our results suggest that contralateral penetration of LE should be targeted in patients with large hematomas and mass effect when early resolution is desirable. As our results show, contralateral penetration is more likely to be achieved through distal microcatheter positioning or use of n-BCA. Use of a dual lumen balloon microcatheter is an alternative to achieve distal penetration. If there is no reflux or penetration of dangerous collaterals, a longer injection may be pursued to deliver LE contralaterally. If contralateral penetration is not achieved from one division, the other division may be embolized to try again.

We did not find an association between number of branches embolized and hematoma or MLS resolution. There have been mixed results reported regarding the influence of embolizing both the anterior and posterior MMA branches. In one of the first studies to evaluate technical factors of MMAe, Catapano et al. found that embolization of both divisions was associated with complete cSDH resolution but did not adjust for confounders. ¹⁸ In a larger study by Hung et al., there was no difference in hematoma resolution based on number of MMA branches embolized. ¹⁹ Khorasanizadeh et al. found that that the number and identity of MMA branches embolized were not associated with surgical rescue or hematoma resolution. ⁷ This is likely due to the extensive anastomoses between MMA branches, which allows embolization from a single division to reach other territories. Flow control or “wedged” catheter positions are helpful to achieve this. There is also variability in the configuration of MMA divisions, with one division sometimes being “dominant” rather than equally sized anterior and posterior branches.

The complete hematoma resolution rate at final follow-up was 62.7%. This is similar to the literature, with reported rates between 33% and 55% for LEs.^19–21 None of the procedural variables were associated with complete hematoma or MLS resolution at final follow-up. This suggests that contralateral penetration of embolisate may increase the rate of hematoma resolution but not the final outcome. Other studies have also found that deeper LE penetration reduces time to cSDH resolution but not whether complete resolution is achieved. ²⁰

Our results corroborate other studies reporting low rates of complications with MMAe. Our rate of 5.3% is similar to a recent meta-analysis in which the pooled complication rate was 4.9%. ⁴ None of our complications were related to injection of LEs, which can lead to blindness or facial weakness through inadvertent penetration of the central retinal artery or facial arcade, respectively. Although distal microcatheter position was not associated with outcomes, this has the benefit of providing a greater safety margin for reflux.

As expected, older age and higher baseline mRS were associated with worse clinical outcomes. It is unclear why n-BCA was also associated with worse clinical outcomes. A recent large propensity-score matched comparison showed no difference in outcomes with n-BCA compared to Onyx. ²² Given there were very few patients who were embolized with n-BCA in our cohort and that they had similar radiographic outcomes compared to other LEs this result could have been confounded by variables that we did not control for. For example, there was less up-front hematoma evacuation and more symptomatic patients in the n-BCA group. Since clinical follow-up occurred relatively early this could have favored the Onyx group.

While distal penetration of LEs and particles has historically been the goal of MMAe there is some evidence that proximal embolization, such as with coils, can also lead to low rates of cSDH recurrence. ²³ This approach has the benefit of being easily performed under conscious sedation and not risking penetration of branches to the orbit or facial nerve. In some situations MMA tortuosity can preclude distal microcatheter position, which makes coiling desirable. Only a few case series have been published regarding this approach. Studies comparing standalone coiling to distal penetration with LEs are lacking. Although we did not have a coiling group for comparison, our findings support the role of distal embolization with the goal of contralateral penetration when early hematoma resolution is desired.

Limitations of this study include its retrospective and monocentric nature. There were no standardized protocols for how MMAe was performed, which could have introduced multiple sources of bias. The indications, type of LE, and microcatheter positioning were at the discretion of the treating physician. The decision to perform up-front surgery or rescue treatment were also not standardized, and all outcomes were self-reported by a single reviewer. Furthermore, follow-up times were not standardized. The study is strengthened by its sample size and inclusion of multiple time points for radiographic follow-up. It is also the first study to evaluate the effect of contralateral LE penetration.

Conclusion

MMAe is a safe and effective treatment for cSDH that is associated with high rates of complete hematoma and MLS resolution as well as low rates of retreatment and complications. Contralateral penetration with LEs may be associated with more rapid cSDH resolution but not the final radiographic result. Other technical factors such as type of LE, microcatheter positioning, and number of branches embolized are not associated with hematoma resolution. Our results should be interpreted with caution given there were multiple potential sources of bias related to non-standardization of treatments. Further studies are needed to confirm these findings in larger, multicenter cohorts.