Keywords
chronic subdural hematoma - computed tomography - embolization - meningeal artery
- risk factors
Introduction
Chronic subdural hematoma (SDH) occurs as a result of minor trauma that causes extravasation
of blood into the subdural space, followed by inflammation and angiogenesis.[1] Following a variable latency period during which the hematoma slowly increases in
size, the patients present with symptoms such as altered mental status, memory disturbances,
headache, seizures, and focal neurological deficits due to compression of the adjoining
brain parenchyma by the hematoma.[2] Although surgical evacuation is the primary treatment for SDH, it is associated
with a high recurrence rate of 11.2 to 27.7%.[3]
[4] Various factors contributing to recurrence are greater thickness of SDH, greater
midline shift, hyperdense content, laminar and separated morphology, antiplatelet
or anticoagulant use, bilaterality, and pneumocephalus.[5]
[6]
[7]
[8]
[9]
Middle meningeal artery (MMA) embolization reduces the incidence of chronic SDH recurrence
after surgical evacuation by reducing the vascularity of the hypervascular membrane
surrounding the SDH cavity.[10] Although the factors that influence the recurrence of SDH following surgical evacuation
have been well studied, the factors affecting its recurrence after MMA embolization
are yet to be studied in detail. SDH recurrence can occur due to vascular supply from
the contralateral MMA.[11] Dural blood supply from arteries other than the ipsilateral MMA may be a cause of
persistent hypervascularity in the dural membrane following MMA embolization.
Various angiographic findings seen during MMA embolization include cotton wool-like
staining, contrast pooling around distal vessels, opacification of the membrane, and
contrast pooling in the SDH cavity.[12] Extravasation of contrast into the SDH cavity indicates leakiness of the membrane.
It needs to be studied whether these angiographic features can predict the clinical
outcome after MMA embolization.
The objective of this study was to evaluate the factors that influence the risk of
recurrence following MMA embolization in patients with chronic SDH.
Materials and Methods
A prospective observational study was conducted between September 2022 and October
2024 on patients with chronic SDH who underwent MMA embolization with or without surgical
evacuation. The exclusion criteria were acute kidney injury, severe chronic kidney
disease (estimated glomerular filtration rate < 30 mL/min/1.73 m2), history of severe contrast reaction to iodinated contrast agents, and follow-up
duration < 90 days after the procedure. The study was initiated after obtaining approval
from the Institute Ethics Committee, and written informed consent was obtained from
all participants.
Preprocedure Imaging
All computed tomography (CT) scans were done on a SOMATOM Definition Flash dual-source
256-slice CT machine (Siemens Healthcare, Erlangen, Germany). Interpretation of the
CT images was performed by a single observer with 10 years of experience in diagnostic
radiology, before the MMA embolization. The imaging parameters that showed a significant
association with SDH recurrence were reassessed by another independent observer with
7 years of experience in diagnostic radiology who was blinded to the treatment outcome,
and the interrater agreement was ascertained. The parameters assessed on baseline
CT were the maximum thickness of SDH, whether unilateral or bilateral, the degree
of midline shift, and the morphology of SDH. The maximum thickness of SDH was measured
on the axial section of CT above the level of the temporal bones and lateral ventricles.[13] The morphology of SDH was described as homogenous, laminated, separated, or trabecular
according to the classification proposed by Nakaguchi et al.[14] In patients who underwent surgical evacuation, a noncontrast CT scan was repeated,
and the thickness of the residual SDH cavity, degree of midline shift, and the presence
of hyperdense contents within the cavity were noted. A multiphasic contrast-enhanced
CT, with intravenous administration of 1.2 mL/kg of iohexol (Omnipaque 350, GE Healthcare,
Shanghai, China), was performed before MMA embolization to ascertain the enhancement
of the membrane covering the SDH cavity. Membrane enhancement was classified as involving
only the outer membrane, and both outer and inner membranes ([Fig. 1]).[15] The morphology of enhancement was classified as thin, thick (≥ 2 mm in thickness),
nodular, and septated ([Fig. 2]). Presence of the spandrel sign, defined as a triangular thick enhancement at the
junction of the outer and the inner membranes of the SDH cavity, was ascertained ([Fig. 2B]).[16]
Fig. 1 Inner and outer membrane enhancement on preprocedure computed tomography (CT). (A, B) Noncontrast (A) and contrast-enhanced venous phase CT images showing enhancing outer membrane (arrow
in B). (C, D) Noncontrast (C) and contrast-enhanced venous phase CT images showing enhancing inner (white arrow
in D) and outer (black arrow in D) membranes.
Fig. 2 Patterns of membrane enhancement on preprocedure computed tomography (CT). (A–D) Contrast-enhanced venous phase CT images showing thin enhancement of the inner membrane
(arrow in A), thick enhancement of the outer membrane (white arrow in B), spandrel sign (black arrow in B), nodular enhancement (arrow in C), and septated enhancement (arrow in D).
Middle Meningeal Artery Embolization
All cases were performed on an Artis Zee biplane digital subtraction angiography machine
(Siemens Healthcare), with local anesthesia used for most cases and general anesthesia
for uncooperative and intubated patients. The procedure was performed by one of the
three interventional radiologists with at least 8 years' experience in neurovascular
interventions. A 5-F 70-cm long sheath (Flexor Raabe, Cook Medical, Indiana, United
States) was placed in the common carotid artery through the common femoral artery
access. The external carotid artery was catheterized using a 4-F Berenstein catheter
(Cordis, Florida, United States). After priming the external carotid artery with 0.5 mg
nimodipine diluted in 25 mL normal saline to prevent vasospasm, the MMA was super
selectively catheterized using a Merit Maestro 2.4 F microcatheter (Merit Medical,
Utah, United States). The parameters observed on the MMA angiogram were the diameter
of MMA just before it entered the foramen spinosum, branching pattern, anastomosis
with the internal carotid artery, membrane blush, and collateral supply across the
midline. Note that 0.5 mg of diluted nimodipine was injected into the MMA before the
advancement of the microcatheter into the divisions to prevent vasospasm. Embolization
was performed using 100 to 300 µm tris-acryl gelatin microspheres (Embospheres, Merit
Medical, Utah, United States) until stasis within the MMA. MMA on the unaffected side
was also embolized to reduce the risk of recurrence due to collateralization across
the midline. A cone-beam CT was performed immediately after embolization to look for
hyperdensity in the subdural space, suggesting contrast pooling.
Outcome Measures
The primary outcome was clinical recurrence of SDH within 90 days after the embolization,
defined as new onset or worsening of existing neurological symptoms, associated with
a residual or recurrent SDH with hyperdense contents on CT. Periprocedural adverse
events occurring within 48 hours of the procedure were recorded as per the adverse
event severity scale of the Society of Interventional Radiology.[17]
Statistical Analysis
The data were compiled and analyzed for their potential effects on the postprocedural
outcome. Stata software, version 18 (StataCorp, Texas, United States), was used for
statistical analysis. Fisher's exact test or chi-square test was used for categorical
variables. Numerical variables were expressed as mean ± standard deviation (SD) or
median with interquartile range (IQR) and analyzed using Student's t-test or Wilcoxon rank sum test. For variables that showed a significant association
with recurrence, relative risks were ascertained. The kappa coefficient was used to
determine the interrater agreement of variables that had a significant association
with recurrence. All statistical tests were two-sided, and a p-value less of than 0.05 was considered significant.
Results
Demographics of the Study Population
MMA embolization was performed on 77 patients. After excluding 3 patients who were
lost to follow-up, 74 patients were included in the final analysis. The mean age was
67.3 ± 11 (SD) years, and 89.2% of the patients (66/74) were males.
Clinical and Imaging Features
The mean duration of symptoms was 10.2 ± 14.2 (SD) days. Sixty-six patients (89.2%)
had a history of antecedent head injury. Ninety-six SDHs were seen in the 74 patients.
The mean thickness of SDH was 17.5 ± 7 (SD) mm, and bilateral SDH was seen in 22 patients
(29.7%). Sixty-five patients (87.8%) had a midline shift in baseline CT with a mean
shift of 8 ± 4.9 (SD) mm. A contrast-enhanced CT was performed after the surgical
evacuation, on the day before MMA embolization in all patients. Membrane enhancement
was best visualized in the venous phase of the scan. While only outer membrane enhancement
was seen in 27/96 (28.1%) hematomas, both inner and outer membrane enhancement were
seen in 36/96 (37.5%) hematomas. Surgical evacuation was performed before MMA embolization
in 65 patients (87.8%) with 84 hematomas. The mean thickness of the residual SDH cavity
on postoperative CT was 9.2 ± 4.7 (SD) mm. Hyperdense contents were seen on the postoperative
CT in 53/84 (63.1%) hematomas, and residual midline shift was seen in 37/65 (56.9%)
patients with a mean shift of 3 ± 3.3 (SD) mm.
MMA Embolization
MMA embolization was performed as an adjunct to surgery in 65 patients (87.8%) and
as the primary treatment in 9 patients (12.2%). The findings observed in 147 MMAs
during the angiography in 74 patients are depicted in [Table 1]. In one patient, MMA on the affected side could not be visualized on angiography
as the internal maxillary artery was found to be occluded. Of the 147 MMAs, 95 were
on the affected side and 52 were on the unaffected side. The MMA was significantly
larger on the affected side (mean diameter ± SD, 1.15 ± 0.03 mm) than on the unaffected
side (mean diameter ± SD, 1.03 ± 0.03 mm) (p = 0.003). Ophthalmic artery arising from the MMA and MMA arising from the ophthalmic
artery were the most common variants. Cone-beam CT was performed in 68 patients with
82 SDHs immediately after MMA embolization.
Table 1
Angiographic findings observed during middle meningeal artery embolization
|
Variable
|
Number of patients (%)/Number of hematomas (%)/Number of arteries (%)/Mean ± SD
|
|
Origin
|
|
Internal maxillary artery
|
140/147 (95.2%)
|
|
OA
|
2/147 (1.4%)
|
|
Direct origin from ECA
|
1/147 (0.7%)
|
|
Variant anatomy
|
|
OA arising from MMA
|
3/147 (2%)
|
|
Anterior division arising from OA
|
2/147 (1.4%)
|
|
Absent posterior division
|
1/147 (0.7%)
|
|
ECA-ICA anastomosis
|
|
MMA to OA via lacrimal artery
|
1/147 (0.7%)
|
|
Dural supply from accessory meningeal artery
|
0
|
|
MMA supply across the midline
|
27/74 (36.5%)
|
|
Membrane blush during angiogram
|
54/95 (56.8%)
|
|
Contrast pooling on post-embolization cone-beam CT
|
74/82 (90.2%)
|
Abbreviations: CT, computed tomography; ECA, external carotid artery; ICA, internal
carotid artery; MMA, middle meningeal artery; OA, ophthalmic artery; SD, standard
deviation.
A total of 143 MMAs were embolized in 74 patients—93 MMAs on the affected side and
50 on the unaffected side. Only one of the MMAs could be embolized in three patients
with bilateral SDH due to an occluded internal maxillary artery, MMA arising from
the ophthalmic artery, and ophthalmic artery arising from MMA, with failure to advance
the microcatheter distal to the ophthalmic artery origin. MMA on the unaffected side
could not be embolized in two patients due to its origin from the ophthalmic artery,
and a small caliber of MMA in one patient each. In two cases in which the ophthalmic
artery arose from the MMA on the unaffected side, there was collateral supply from
the MMA across the midline into the dura of the affected side. Embolization was therefore
performed by advancing the microcatheter distal to the origin of the ophthalmic artery.
Similarly, in one case in which there was anastomosis of the MMA with the ophthalmic
artery, embolization was performed after advancing the microcatheter distal to the
anastomosis. No procedure-related complications were observed in these cases.
Factors Affecting SDH Recurrence
The patients were followed up for a mean duration of 195.2 ± 137.6 (SD) days after
MMA embolization, and eight recurrent SDHs were observed in six patients (6/74, 8.1%).
The surgical rescue rate was 5.4% (4/74), and two patients with recurrence were managed
conservatively. No recurrences were observed in patients in whom MMA embolization
was performed as a standalone treatment. No procedure-related death or severe adverse
events were observed. Minor adverse events were headache (26/74, 35.1%) and jaw pain
(2/74, 2.7%), both lasting less than 24 hours.
The association of SDH recurrence following MMA embolization with the various clinical
features, imaging findings, angiographic findings, and embolization parameters is
depicted in [Table 2]. Membrane enhancement on the preprocedure CT was significantly less common in hematomas
that recurred than those that did not (2/8, 25% vs. 61/88, 69.3%; p = 0.018). However, there was no significant association between the location or morphology
of enhancement and recurrence ([Table 2]). None of the other parameters had a significant association with recurrence. The
absence of membrane enhancement was a significant risk factor for clinical recurrence
(relative risk, 5.7 [95% confidence interval [CI], 1.2–26.8], p = 0.01) ([Fig. 3]). Out of the 33 hematomas with no membrane enhancement, 6 recurred (18.2%). Membrane
enhancement had an excellent interrater agreement between the two observers (kappa
coefficient, 0.95; p-value < 0.001). In patients with recurrence, there was 100% agreement between the
two observers.
Fig. 3 Recurrence of chronic subdural hematoma (SDH) after middle meningeal artery (MMA)
embolization. (A) Coronal reformatted noncontrast computed tomography (CT) image of an elderly patient,
on antiplatelet therapy for coronary artery disease, who presented with slurring of
speech and giddiness, showing bilateral chronic SDHs. Bilateral burr-hole evacuation
was done. (B) Coronal reformatted contrast-enhanced CT image showed no enhancing membrane around
the SDH. (C, D) The right (C) and the left MMA (D) were superselectively catheterized and embolized using 100 to 300 µm tris-acryl
gelatin microspheres. (E, F) Lateral projections of post-embolization angiograms of right (E) and left (F) MMAs showing stasis within. (G) The patient presented with paraparesis 40 days later, and noncontrast CT showed
bilateral recurrent SDH.
Table 2
Association of clinical, imaging, angiographic, and embolization parameters with chronic
subdural hematoma recurrence after middle meningeal artery embolization
|
Variable
|
Number of patients (%)/Number of hematomas (%)/Mean ± SD/Median (IQR)
|
p-Value
|
|
No recurrence
|
Recurrence
|
|
Male sex
|
60/68 (88.2%)
|
6/6 (100%)
|
1.00
|
|
Age
|
67.2 ± 10.9
|
68.7 ± 13.4
|
0.76
|
|
Duration of symptoms before treatment (d)
|
7 (2.5–10)
|
6 (3–12)
|
0.84
|
|
Diabetes mellitus
|
34/68 (50%)
|
3/6 (50%)
|
1.00
|
|
Hypertension
|
34/68 (50%)
|
4/6 (66.7%)
|
0.68
|
|
History of trauma
|
44/68 (64.7%)
|
4/6 (66.7%)
|
1.00
|
|
Antiplatelet or anticoagulant use
|
26/68 (38.2%)
|
2/6 (33.3%)
|
1.00
|
|
Glasgow coma scale at presentation
|
15 (14.5–15)
|
15 (14–15)
|
0.60
|
|
Focal neurological deficit at presentation
|
26/68 (38.2%)
|
5/6 (83.3%)
|
0.08
|
|
Bilateral SDH
|
19/68 (27.9%)
|
3/6 (50%)
|
0.35
|
|
SDH thickness on baseline CT (mm)
|
17.4 ± 6.8
|
18.6 ± 9.2
|
0.64
|
|
Presence of midline shift on baseline CT
|
60/68 (88.2%)
|
5/6 (83.3%)
|
0.55
|
|
Morphology
|
|
Homogenous
|
19/88 (21.6%)
|
3/8 (37.5%)
|
0.51
|
|
Laminated
|
8/88 (9.1%)
|
0
|
|
Separated
|
33/88 (37.5%)
|
4/8 (50%)
|
|
Trabecular
|
28/88 (31.8%)
|
1/8 (12.5%)
|
|
SDH thickness on postoperative CT (mm)
|
9.1 ± 4.8
|
9.8 ± 4.0
|
0.71
|
|
Presence of midline shift on postoperative CT
|
33/59 (55.9%)
|
4/6 (66.7%)
|
0.69
|
|
Presence of hyperdense contents on postoperative CT
|
50/76 (65.8%)
|
3/8 (37.5%)
|
0.14
|
|
Membrane enhancement on pre-embolization CT
|
61/88 (69.3%)
|
2/8 (25%)
|
0.018
|
|
Location of membrane enhancement
|
|
Only outer membrane
|
25/61 (41%)
|
2/2 (100%)
|
0.18
|
|
Both outer and inner membranes
|
36/61 (59%)
|
0/2
|
|
Morphology of membrane enhancement
|
|
Thin
|
44/61 (72.1%)
|
1/2 (50%)
|
0.49
|
|
Thick
|
9/61 (14.8%)
|
1/2 (50%)
|
|
Nodular
|
1/61 (1.6%)
|
0/2
|
|
Septated
|
7/61 (11.5%)
|
0/2
|
|
Spandrel sign in hematomas with membrane enhancement
|
7/61 (11.5%)
|
0/2
|
1.00
|
|
Interval between last surgery and embolization (d)
|
4.7 ± 9.4
|
3 ± 1.8
|
0.66
|
|
Diameter of MMA on the affected side (mm)
|
1.16 ± 0.26
|
1.1 ± 0.15
|
0.53
|
|
Collateral supply from contralateral MMA
|
26/68 (38.2%)
|
1/6 (16.7%)
|
0.41
|
|
Membrane blush on angiogram
|
50/87 (57.5%)
|
4/8 (50%)
|
0.72
|
|
Bilateral MMA embolization
|
63/68 (92.7%)
|
6/6 (100%)
|
1.00
|
|
Microcatheter position during embolization of MMA on the affected side
|
|
Main trunk of MMA
|
52/85 (61.2%)
|
7/8 (87.5%)
|
0.25
|
|
Anterior and posterior divisions
|
33/85 (38.8%)
|
1/8 (12.5%)
|
|
Contrast pooling on post-embolization cone-beam CT
|
69/77 (89.6%)
|
5/5 (100%)
|
1.00
|
|
Postprocedure headache
|
24/68 (35.3%)
|
2/6 (33.3%)
|
1.00
|
Abbreviations: CT, computed tomography; IQR, interquartile range; MMA, middle meningeal
artery; SD, standard deviation; SDH, subdural hematoma.
Discussion
MMA embolization is an effective treatment for chronic SDH, both as an adjunct to
surgery and as a standalone treatment.[3]
[18]
[19]
[20]
[21]
[22] In our study, we observed low clinical recurrence and surgical rescue rates of 8.1
and 5.4%, respectively, after MMA embolization. In the limited number of patients
in whom MMA embolization was performed as a standalone procedure, no recurrences were
observed. The low recurrence rate observed in our study is similar to that observed
in prior studies.[3]
[18]
[19]
[20]
[21]
[22] However, the recently published EMPROTECT trial failed to show a significant reduction
in SDH recurrence after MMA embolization using tris-acryl gelatin microspheres compared
with standard treatment (14.8% vs. 21%; odds ratio, 0.64 [95% CI, 0.36–1.14]; p = 0.13). The trial included only those patients who were at high risk of SDH recurrence
after surgical evacuation, and the embolization was done using large 300 to 500 µm
particles.[23] The use of smaller particles (100–300 µm) and the inclusion of all patients with
chronic SDH, regardless of the risk of recurrence, may have contributed to the lower
recurrence rate in our study.
We investigated the factors influencing the risk of SDH recurrence following MMA embolization.
The absence of an enhancing membrane on the pre-embolization contrast-enhanced CT
was identified as a risk factor for recurrence. Membrane formation in SDH occurs as
a result of the proliferation of dural border cells, which differentiate into connective
tissue.[24] The outer membrane of SDH contains pathological sinusoidal capillaries, which contribute
to recurrent bleeding into the SDH cavity.[25] Membrane enhancement around the SDH cavity is associated with early recurrence after
surgery, and such patients will benefit from MMA embolization as it reduces the hypervascularity
of the membrane.[15] In contrast to our study results, a study by Weinberg et al did not find any difference
in the recurrence rate in membranous and nonmembranous SDH. Notably, in their study,
postoperative neurological improvement after MMA embolization was better in patients
with membranous SDH.[26] However, MMA embolization was done as a standalone treatment in most of the patients
in their study, indicating that the characteristics of their study population differ
from ours. As MMA embolization was performed as a standalone procedure in only 12.2%
of the patients in our study, a subgroup analysis on the impact of membrane enhancement
on the outcome could not be performed in this subset. The recurrence rate for hematomas
without membrane enhancement was high (18.2%), similar to that reported after surgical
evacuation without MMA embolization (11.2–27.7%).[3]
[4] Therefore, a routine pre-embolization contrast-enhanced CT would help identify patients
with no membrane enhancement and high risk of recurrence after MMA embolization. The
high rate of recurrence in these patients suggests that other factors may contribute
to the recurrence. Increased levels of proinflammatory cytokines are observed in the
SDH, indicating that inflammation and increased vascular permeability contribute to
hematoma expansion.[27] Hyperfibrinolysis within the SDH cavity also plays a role in hematoma expansion
by preventing stable clot formation and promoting continued bleeding.[24] Surgical techniques like membranectomy may help reduce the rate of recurrence in
patients with nonenhancing membranes.[28]
Apart from the absence of membrane, none of the other variables had an association
with SDH recurrence following MMA embolization in our study. There was no association
of location or morphology of enhancement with recurrence. Notably, although the lack
of membrane enhancement had an association with recurrence, the membrane blush observed
on the MMA angiogram did not have a significant association with recurrence. This
is likely because the membrane blush was influenced by the volume and rate of contrast
injection into the microcatheter, which was not standardized in our study. A study
by Salem et al found that the use of anticoagulant agents, midline shift, MMA diameter < 1.5 mm,
and superselective embolization of the MMA without targeting the main trunk were significant
predictors of treatment failure following the embolization.[29] Supportive evidence from a study by Fuentes et al also indicates that anticoagulant
use is a predictor for recurrence following MMA embolization.[30] Evaluation of the factors affecting the resolution of SDH following embolization
has led to the finding that mixed density and separated types, postoperative SDH thickness
and midline shift, and antiplatelet or anticoagulant use are associated with delayed
hematoma resolution.[31]
We observed that the MMA on the same side as that of the SDH was significantly larger
than the MMA on the unaffected side (mean diameter, 1.17 ± 0.02 vs. 1.03 ± 0.03 mm,
p < 0.001). Similar results were observed in a study by Pouvelle et al, who found that
the MMA on the same side as that of the SDH was significantly larger than the contralateral
MMA (median [IQR], 1.6 [1.4–1.8] vs. 1.4 [1.25–1.6]; p < 0.001).[32] Enlargement of the MMA has been reported to occur when SDH develops following trauma.[33] However, our study did not find a significant association between MMA diameter and
SDH recurrence.
Apart from the normal origin of MMA from the internal maxillary artery, anatomical
variants are rarely observed. Fantoni et al observed MMA originating from the ophthalmic
artery in 13.8% of patients undergoing MMA embolization.[34] However, in our study, the MMA had a variant origin from the ophthalmic artery in
only 1% of cases. MMA embolization is risky in such cases due to the potential risk
of injury to the ophthalmic artery during superselective catheterization and nontarget
embolization into the central retinal artery.
The MMA can have anastomoses with the branches of the internal carotid artery. We
observed an anastomosis between the MMA and the ophthalmic artery through the recurrent
meningeal branch of the lacrimal artery in one of the cases. Further, the ophthalmic
artery originated from the MMA instead of the internal carotid artery in 2.5% of the
cases. There is a potential risk of developing blindness following nontarget embolization
into the ophthalmic artery in such cases. However, embolization can safely be performed
by navigating the microcatheter distal to the dangerous anastomosis.[35] Selective catheterization of the anterior and posterior branches of MMA above the
level of the anterior clinoid process is recommended to avoid nontarget embolization
into the dangerous anastomoses.[13] No complications were observed in our patients in whom the embolization was performed
by advancing the microcatheter distal to the origin of the ophthalmic artery.
Our study has various limitations. The sample size was small, and therefore, the number
of patients with SDH recurrence after MMA embolization was small. The small number
would have affected the assessment of factors influencing the recurrence. The assessment
of membrane blush on the MMA angiogram was subjective and thus prone to errors. Follow-up
CT was not performed for all patients, because of which asymptomatic recurrences could
have been missed.
Conclusion
The absence of membrane enhancement on baseline CT is a risk factor for SDH recurrence
after MMA embolization. The recurrence rate was high in hematomas without membrane
enhancement, indicating that MMA embolization may not be useful in reducing the recurrence
rate in this subset of patients.
