Article Text

Download PDFPDF

Original research
Middle meningeal artery embolization reduces the post-operative recurrence rate of at-risk chronic subdural hematoma
  1. Eimad Shotar1,
  2. Louis Meyblum1,
  3. Kevin Premat1,2,
  4. Stéphanie Lenck1,
  5. Vincent Degos2,3,
  6. Teodor Grand1,
  7. Jonathan Cortese1,
  8. Arnaud Pouvelle1,
  9. Geoffroy Pouliquen1,
  10. Samuel Mouyal1,
  11. Anne-Laure Boch4,
  12. Alexandre Carpentier2,4,
  13. Nader-Antoine Sourour1,
  14. Bertrand Mathon2,4,
  15. Frédéric Clarençon1,2
  1. 1 Department of Neuroradiology, Pitié-Salpêtrière Hospital, Paris, France
  2. 2 Sorbonne Université, Paris, France
  3. 3 Neurosurgical Anesthesiology and Critical Care, Pitié-Salpêtrière Hospital, Paris, France
  4. 4 Department of Neurosurgery, Pitié-Salpêtrière Hospital, Paris, France
  1. Correspondence to Dr Eimad Shotar, Neuroradiology Department, Pitié-Salpêtrière Hospital, Paris, France; eimad.shotar{at}


Background Embolization of the middle meningeal artery (MMA) has emerged as a potential treatment of chronic subdural hematomas (CSDHs).

Objective To evaluate the impact on recurrence rate of postsurgical embolization of CSDH in patients with a higher than average risk of recurrence.

Methods A monocentric retrospective study was performed on retrospectively collected data. From March 2018 to December 2019, embolization of the MMA was proposed as an adjunct postoperative treatment after burr-hole surgery in patients operated for a recurrent CSDH or a CSDH with an independent recurrence risk factor, including antiplatelet therapy, full anticoagulation therapy, coagulation disorder, hepatopathy, or chronic alcoholism. Patients who had undergone postoperative embolization were compared with a historic group of patients operated between March 2016 and March 2018, selected based on the same inclusion criteria.

Results During the study period, 89 patients (with 74 unilateral and 15 bilateral CSDHs) were included and underwent an embolization procedure, leading to 91 out of a total of 104 MMA being embolized (88%). These were compared with 174 patients (138 unilateral and 36 bilateral CSDH) in the historic control group. One major procedure-related adverse event was registered. Four of the 89 patients (4%) required surgery for a CSDH recurrence in the embolization group, significantly less than the 24 of 174 patients (14%) in the control group (OR=0.28, 95% CI 0.07 to 0.86, p=0.02).

Conclusions Postsurgical embolization of the MMA may reduce the recurrence rate of CSDHs with a risk factor of recurrence.

  • chronic subdural hematoma
  • middle meningeal artery
  • recurrence
  • embolization

Statistics from

Request Permissions

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.


Chronic subdural hematomas (CSDHs) are among the most widespread diseases encountered in neurosurgical practice, with an annual incidence of 14 to 20 per 100 000 individuals,1 2 predicted to rise even further in the coming decades.3 It is mostly a disease of the elderly with two-thirds of cases accounted for in patients older than 65 years.1 3 Contrary to common wisdom, CSDHs are associated with significant morbidity and mortality, 4% and 11% respectively in a recent meta-analysis of randomized trials.4 Most importantly, CSDHs have been described as a sentinel health event, associated with marked reduction in life expectancy for patients in comparison with age-matched controls.5

Surgical management of symptomatic CSDHs, mostly through twist-drill or burr-hole craniostomy, is thought to be safe and effective.4 6 However, despite a demonstrated benefit of postevacuation closed-system drainage,7 recurrence remains a common preoccupation, with an estimated average rate of 11% in both randomized trials and observational studies.4 Recurrence in itself is a source of altered functional outcome and increased mortality.7 8

Middle meningeal artery (MMA) embolization has been described as a possible treatment of CSDHs.9 10 Unfortunately, evidence is lacking at many decision points in the management algorithm for the typical patient with CSDH, including the specific place of MMA embolization.2 The aim of this study is to evaluate the impact on recurrence rate of postsurgical embolization of CSDH in patients with a high risk of recurrence.

Materials and methods

Study design

This is a monocentric retrospective study performed in a tertiary care academic center. From the March 15, 2018 to the December 15, 2019, embolization of the MMA was proposed as an adjunct postsurgical treatment in patients operated for a recurrent CSDH or a CSDH with an independent recurrence risk factor. This was a concerted decision, taken before embolizations were started, by the neurosurgery and interventional neuroradiology departments after learning of the results of CSDH embolization published in the literature.9 The institutional criteria for postoperative CSDH embolization were agreed upon based on practitioners’ experience. Patients having undergone postoperative embolization were compared with a historic cohort of patients operated between the March 1, 2016 and March 14, 2018 and selected based on the same inclusion criteria. Data were collected retrospectively.


Patients were eligible if operated by single burr-hole craniostomy for a CSDH recurrence or a CSDH in the case of an independent risk factor defined as follows: (1) antiplatelet therapy11 or (2) full anticoagulation therapy11 or (3) coagulation disorder12 or (4) hepatopathy12 or (5) chronic alcoholism.12 Eligible patients were excluded if they refused the embolization procedure or were denied embolization by the attending physician because they presented in a moribund state, had a contraindication to the embolization procedure like severe renal dysfunction, or had a life expectancy below 6 months. Patients referred for embolization during the study period without the prespecified inclusion criteria were excluded.

Embolization procedures

Pre-embolization supra-aortic trunk CT-angiography was recommended to rule out technical and anatomical contraindications. The choice of the embolization technique was left to the discretion of the attending interventional neuroradiologist. However, the procedure was most often performed as follows: in a single or biplane angiosuite, under local anesthesia, through a 5F or 6F femoral or radial sheath, without intravenous heparin, a guiding catheter was navigated in the common carotid artery, ipsilateral to the CSDH. A selective angiogram was obtained to confirm that the MMA originated from the internal maxillary artery. The catheter was further advanced in the external carotid artery and a microcatheter was navigated in the MMA. Superselective angiography was performed to check for orbital branches of the MMA or anastomosis with the ophthalmic artery before embolization. Particle embolization was encouraged over liquid embolic agents because of previous reports of particle embolization safety and efficacy.9 10 The MMA embolization strategy was nevertheless left to the discretion of the attending physician and could include embolization with or without coiling of anastomosis and proximal coiling.


The primary endpoint was recurrence rate, defined as the rate of reoperation to treat a recurrent CSDH previously treated by burr-hole craniostomy.7 Immediate postoperative revision surgery (within 24 hours of index surgery) was not considered as recurrence to exclude surgery-related hemorrhage. Standard practice at our institution is to perform a CT scan only in cases of clinical deterioration and reoperate in cases of recurrent CSDH with mass effect. The secondary endpoint was rate of complications, including epilepsy and subdural empyema. Embolization procedure-related major and minor complications were defined according to previously reported generic criteria.13

Statistical analysis

Data are expressed as a percentage for binary variables and as mean±SD for continuous variables. Probability values are provided uncorrected. Probability values <0.05 were considered significant. Fisher’s exact test was used to compare frequencies, and comparison of means was performed using a Student t-test or Mann-Whitney test, depending on the data distribution.

Intention-to-treat analysis of the primary endpoint was performed. All patients who underwent an embolization procedure were included in the embolization group, regardless of whether the MMA was completely occluded or the procedure was interrupted prematurely. Only patients who were denied embolization based on the results of the supra-aortic trunk CT angiography were excluded.

Recurrence-free survival curves were drawn using the Kaplan-Meier method to compare the embolization and control groups.

In order to further evaluate the impact of embolization on CSDH recurrence rate, a multivariate stepwise logistic regression model of recurrence was constructed. Data at CSDH level from the embolization and control groups were pooled. The model was constructed with the following variables: embolization group, closed system drainage, and previous recurrence. Multivariate regression analysis was iterated with the following variables forced into the model one at a time: anticoagulation medication, antiplatelet medication, thrombocytopenia and coagulation disorders, alcohol abuse, and hepatopathy. Statistical analyses were performed using MedCalc version 19.2 (MedCalc Software, Ostend, Belgium).

Ethical statement

The institutional review board approved this study (CERAR IRB 00010254-2020-014). The need for signed patient consent was waived.


Patients’ characteristics in the embolization group

During the embolization inclusion period, 178 patients were operated for a CSDH, out of which 112 patients were referred for MMA embolization (figure 1). Nine patients were referred for indications outside the prespecified inclusion criteria. One hundred and three patients out of 178 (58%) were eligible for embolization according to the above-mentioned prespecified criteria. Six patients were deemed unfit to undergo embolization. Six patients refused the procedure. Of the remaining 91 patients, 81 (89%) underwent CT angiography and two were excluded because of extensive supra-aortic trunks atheroma. Subsequently, 89 underwent embolization procedures of 104 MMAs (74 unilateral and 15 bilateral CSDH).

Figure 1

Flow chart. CSDH, chronic subdural hematoma; E°, embolization.

Embolization procedures

Of the 89 embolization procedures, 86 (97%) were performed under local anesthesia and 3 (3%) under conscious sedation. The procedures were performed via a femoral and radial access in 75 (84%) and 14 (16%) cases as a first-line option, respectively. A 5F or 6F introducer sheath was used in 86 cases (97%) and a larger introducer sheath was used in three cases (3%). The embolization procedure was interrupted prematurely in eight patients (9%): in four cases because of agitation and in four cases because of a highly tortuous vasculature, with a procedural risk deemed excessive. An angiographic anastomosis between the MMA and the ophthalmic artery or an orbital branch of the MMA was identified in 22 of the 104 cases (21%), leading to three MMAs not being embolized, including two cases where the MMA originated from the ophthalmic artery. In seven cases, an anastomosis was occluded by coiling prior to MMA embolization and in the remaining 12 cases, embolization was performed without prior coiling of the orbital branch, including three cases where only proximal coiling was performed. Ninety-one out of a total of 104 MMAs were embolized (88%). Embolization was performed with calibrated triacryl gelatin microspheres (300 to 500 µm in diameter) in 81 instances, with associated proximal MMA coiling in 27 of the 81 cases, using an n-butyl cyanoacrylate liquid embolic agent in five cases, and by proximal MMA coiling alone in the remaining five cases. Five minor complications (6%) were registered: one patient experienced partial seizure during the procedure, one reported reversible headache, two described transient diplopia, and one asymptomatic iatrogenic meningo-meningeal fistula was treated by MMA coiling during the procedure. One major procedure-related complication (1%) occurred during the study period: postprocedural femoral artery occlusion at the puncture site, requiring bypass surgery. This patient died 43 days after the embolization procedure of multiple organ failure.

Comparison with the historic control group

In the historic control group, 174 out of 310 patients (56%) were included after surgery, 138 unilateral and 36 bilateral CSDH, totaling 210 individual cases of CSDH. The demographics of the embolization and control groups were comparable (table 1). Patients in the embolization group were more frequently referred to embolization because of CSDH recurrence surgery than expected from the historic control group. Burr-hole surgery alone, without closed-system drainage, was more frequent in the historic group.

Table 1

Patients’ characteristics

Four out of the 89 patients (4%) in the embolization group required surgery for a CSDH recurrence. All four patients had undergone embolization of the MMA with particles 300 to 500 µm in diameter. For two of these four patients, additional proximal MMA coiling was performed. In contrast, 24 of the 174 patients (14%) in the historic control group reached that outcome (odd ratio (OR)=0.28, 95% CI 0.07 to 0.86, p=0.02). Three cases of early surgical revisions (within 24 hours of the index surgery) were not counted as recurrence events in the historic control group, in line with the definition of the primary outcome. In the embolization group, three patients experienced clinical recurrence of an operated CSDH that did not lead to reoperation, therefore not meeting the primary endpoint definition. One of these three recurrences occurred in a patient for whom the embolization procedure was aborted. The rate of non-operated recurrences in the historic control group is unknown.

Six (7%) of the 89 patients in the embolization group died during the 3-months follow-up period, comparable to the 12 deaths for the 174 patients (7%) in the historic control group. In the embolization group, one death followed the above-mentioned embolization procedure-related complication. Another death was due to a compressive CSDH contralateral to the operated and embolized CSDH and one was due to status epilepticus in relation to a CSDH recurrence which was not operated (the above-mentioned non-operated recurrence after an aborted embolization procedure), neither of which met the primary endpoint definition. The other three deaths were unrelated (cardiac arrest, brain neoplasm, and death of undetermined extraneurological origin). Figure 2 shows recurrence-free survival to be significantly higher in the embolization group (HR for recurrence 0.39, 95% CI 0.18 to 0.85, p=0.02).

Figure 2

Kaplan-Meier curves of recurrence-free survival of operated patients with chronic subdural hematomas.

When individual CSDHs were considered rather than patients, 4 recurrences in 104 CSDH (4%) were registered in the embolization group, significantly less than the 26 recurrences in 210 CSDH (12%) in the control group (OR=0.28, 95% CI 0.07 to 0.85] p=0.01). Three cases of early surgical revisions were not counted as recurrence in the control group.

Given the above-mentioned imbalance between the embolization and historic control groups for previous recurrence and percentage of surgery performed with closed-system drainage, a sensitivity analysis was performed to rule out the potential confounding effect of these factors. When patients eligible because of recurrence surgery where excluded, 3 recurrence in 67 patients (4%) were noted in the embolization group, significantly less than the 22 recurrences in 154 patients (14%) in the control group (three cases of early surgical revisions were not counted as recurrence in the control group) (OR=0.28, 95% CI 0.05 to 0.99, p=0.04). When CSDHs operated without closed-system drainage were excluded, 4 recurrences in 97 CSDH (4%) were noted in the embolization group as compared with 19 recurrences in 159 CSDHs (12%) in the control group (two cases of early surgical revisions were not counted as recurrence in the control group) (OR=0.32, 95% CI 0.08 to 0.99, p=0.04). Multivariate analysis found that the only factor independently associated with recurrence of an individual CSDH in the study population was the embolization group with a protective effect (OR=0.3, 95% CI 0.1 to 0.87, p=0.03).


Main results

This non-randomized study shows the benefit for recurrence-rate reduction of postsurgical embolization of CSDH in patients with a high risk of recurrence. One procedure-related serious adverse event was noted in the embolization group (1%).

MMA embolization in the treatment of CSDHs

MMA embolization has emerged in recent years as a potential treatment of CSDHs.14 The rationale for this treatment is provided by refined understanding of the physiopathology of the disease. CSDHs are most often related to minor head trauma, even though as many as 40% of patients deny such an event.15 16 This initial insult is thought to cause a cleavage within the dural border-cell layer, resulting in either an acute subdural hematoma or a subdural hygroma, both of which can evolve into a CSDH. Whatever the initial insult and ensuing subdural lesion, CSDHs subsequently demonstrate an autonomous evolution, mediated by multiple drivers promoting expansion. These include inflammation, membrane formation, angiogenesis, and hyperfibrinolysis. Morphologically, the microvasculature alterations of the outer layer of CSDHs include increased vessel density, capillary diameter, and the occurrence of large intercellular gaps between the endothelial cells.17 18 These so-called sinusoid neovessels of the CSDH outer membrane are known to be anatomically connected to meningeal arteries from the adjacent dura—namely, the MMA.19 Several radiological findings corroborate these histological observations, including MMA selective angiograms in patients with CSDH demonstrating an abnormal dural capillary blush,20 21 MMA increased caliber ipsilateral to CSDHs as compared with contralateral MMAs and control vessels22 and CSDH membrane enhancement on angiography-DynaCT imaging during MMA embolizations of CSDHs.23

As mentioned earlier, CSDH postsurgical recurrence is a significant source of morbidity and mortality in the natural history of the disease.7 8 Recurrence rate in this study is higher than the average rate of 11% described by Almenawer et al 4 in their meta-analysis, in line with the selected nature of the study population. Indeed, higher recurrence rates have been described in the literature, up to 33%.24 A possible explanation for such variations is the absence of definitive data for risk factors of recurrence. A great many number of risk factors have been proposed, including poor postoperative brain re-expansion, alcoholism, liver dysfunction, coagulopathies, larger residual hematomas following surgery, presence of septations and loculations preoperatively, antiplatelet and anticoagulant medication, advanced age, male sex, and diabetes mellitus among others, with significant discrepancies between studies. More importantly, description of recurrence rates and risk factors is hindered by the absence of a common definition of this complication in the literature, as highlighted by several investigators.2 4

Embolization of the MMA has emerged as a possible treatment of CSDH.9 10 Embolization has been proposed both as a possible curative treatment of CSDH or CSDH recurrence9 12 20 25 or as an adjunct to surgery.9 23 26 Regardless of the embolic agent employed, particles, coils, n-butyl cyanoacrylate or ethylene vinyl alcohol copolymer liquid embolic agents, most publications have concluded that MMA embolization is safe and effective in the treatment of CSDH.9 12 14 27 Ban et al, in particular, compared 72 consecutive patients who underwent MMA embolization, as sole therapy (27 patients) or prior to surgical evacuation (45 patients), with 469 patients who underwent conventional treatment in a historic control group. They found treatment failure rate to be significantly lower in the embolization group (1%) than in the control group (27.5%).9 Despite the apparent simplicity of the MMA embolization procedure, however, the possibility of technical failure and major complications must not be overlooked, as demonstrated herein, especially given the age of the target patients, in average more than 5 years older than the population treated by mechanical thrombectomy for acute ischemic stroke in published trials.28 The recent surge in interest in radial access for endovascular neurointerventional procedures29 may prove to be timely with regard to the emergence of MMA as a possible treatment of CSDH, given its lower complication rate demonstrated in the cardiology literature.30

Study limitations

The major limitations of this study are its retrospective, monocentric, and non-randomized design, with inherent risks of bias. The definition of recurrence rate was chosen in order to limit the risk of bias, as compared with clinical recurrence, for instance. Moreover, patients in the control group who were lost to follow-up were included and considered not to have reached the study endpoint, reducing the risk of falsely concluding in favor of the efficacy of the tested intervention. Another potential source of bias derives from the absence of preoperative imaging data for a significant proportion of the historic control group. It was therefore not possible to ensure that the embolization and control groups were comparable for potentially confounding imaging factors like presence of septations for instance. Finally, non-randomization led to a significant differences in potentially confounding factors between the two study populations, probably due to referral bias in the embolization group for the rate of previous recurrence. Sensitivity and multivariate analysis demonstrated that the effect of embolization on recurrence rate was independent and persisted both in direction and amplitude after elimination of potential confounding factors.


Postsurgical embolization of the MMA may reduce the recurrence rate of CSDHs with a risk factor of recurrence. This finding needs to be confirmed by a randomized controlled trial.



  • Contributors Study design: ES, LM, BM, FC; Study oversight: FC; Data collection: ES, LM, TG, JC, AP, GP, SM; Statistical analysis: ES, LM; Manuscript preparation: ES; Critical review of the manuscript: all coauthors.

  • 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 FC reports conflict of interest (unrelated) with Medtronic, Guerbet, Balt Extrusion (payment for readings), Codman Neurovascular (core laboratory). N-AS is consultant for Medtronic, Balt Extrusion, Microvention. The other authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper. The manuscript is not supported by industry. ES is principal investigator of a randomized controlled trial related to chronic subdural hematomas embolization financed by a PHRC-IR public grant (non-related to the present study).

  • Patient consent for publication Not required.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data availability statement Data are available upon reasonable request from the corresponding author.