Keywords falx artery - vasospasm - nimodipine - subarachnoid hemorrhage
Introduction
The falx cerebri is a multilayered interhemispheric invagination of the dura. Anchored
at the crista galli anteriorly and internal occipital protuberance posteriorly, it
assumes a sickle-like morphology as it traverses the supratentorial midline fissure.[1 ] Unopposed layers of the falcine dura enclose blood-filled sagittal sinuses. The
superior sagittal sinus (SSS) is enclosed in its superior margin of the falx cerebri,
while the inferior sagittal sinus rides along the free margin just rostral to the
corpus callosum.[2 ] The most common pathology afflicting the falx is a falcine meningioma (not invading
the SSS) that constitutes approximately 9% of all intracranial meningiomas.[3 ]
Until the latter half of the last century, the general impression of the falcine dura
was that of an avascular membranous structure, similar to fascia. It is now acknowledged
that the dura is indeed a vascularized sheet and that the anterior segment of the
falx is perfused by the anterior falcine artery (AFA), a branch of the anterior ethmoidal
artery (AEA), a branch of the ophthalmic artery (OA).[4 ] Moyamoya disease, a progressive neurovascular pathology of the distal internal carotid
arteries (ICA), is characterized by hypertrophic AEA and AFA that subserve dural–pial
anastomoses in the ischemic anterior parasagittal parenchyma.[5 ] It is then not unreasonable to hypothesize that in other ischemic states, as with
post-subarachnoid hemorrhage vasospasm (SAH-V), the AFA may reactively increase in
dimensions to perfuse the at-risk ischemic parenchyma. In present study, we tried
to evaluate the angiographic attributes of the AFA and its reactionary role in post-SAH-V
and the effect of chemical angioplasty on AFA.
Materials and Methods
This was a retrospective observational study performed at a tertiary referral center.
Cases were defined as patients with clinical and angiographic evidence of vasospasm,
whose age at evaluation was between 18 and 85 years, for whom chemical angioplasty
was achieved with intra-arterial nimodipine infusion (Nimodipine chemical angioplasty
[NCA]). One hundred cases whose digital subtraction angiography (DSA) was performed
between July 2019 and March 2020 were retrieved from the Picture Archiving and Communication
System (PACS). Informed consent was taken from all participants. All angiographies
were done on either Artis Zee (Siemens Healthineers AG, Erlangen, Germany) or Allura
(Koninklijke Philips N.V., Amsterdam, the Netherlands) biplane DSA systems at six
frames per second as a standard. The institutional intra-arterial NCA protocol involves
infusion of nimodipine (3 mg maximum dose diluted in 35 mL normal saline) over a period
of 20 minutes) via a 5-Fr vertebral catheter placed in the distal cervical ICA under
strict blood pressure monitoring and intermittent fluoroscopy. A subjective/visual
grading of angiographic vasospasm into (1) mild, (2) moderate, and (3) severe categories
was performed for the angiograms obtained before and after NCA. The imaging and angiographic
attributes that were tabulated included the following: presence/absence of AFA; length
of AFA from the anterior cranial fossa (ACF) base computed from freehand tracing with
the available tools in the institutional Radiology Information System (Centricity
RIS-I, GE Healthcare) in the corresponding DSA frame in which the AFA was visualized
to its longest ([Fig. 1A, B ]), change in length of AFA after nimodipine, aneurysm location (anterior cerebral
artery vs. other), treatment methods (surgical clipping and endovascular coiling),
and involved territory in vasospasm (anterior cerebral artery [ACA] vs. others). Angiograms
of an equivalent number of “disease controls” defined by patients without subarachnoid
hemorrhage (in whom elective angiography was done for indications other than acute
subarachnoid hemorrhage and in which the vascular abnormality was restricted to the
posterior fossa and wherein the pathological vessels involved were not anterior cerebral
artery) were evaluated on the same lines.
Fig. 1 (A–C ) A 56-year-old male patient following anterior communicating artery aneurysm clipping.
The lateral angiogram shows clip in situ (white arrow in A ) with mild vasospasm in the A2–A3 segments of the anterior cerebral artery. The anterior
falcine artery (AFA) is demonstrated (encircled in B ). The AFA length was measured by joining multiple small segments along its course
starting from the anterior cranial fossa base (gray dashed line in C ).
Statistical Analysis
Data were entered in Excel Spreadsheet 2010 (Microsoft, Redmond, Washington, United
States). Mean and standard deviation were calculated for the variables. Mixed effect
analysis was adopted for pre- and postintervention comparisons among different groups.
For comparison of paired categorical data, the chi-squared test was used, while the
Mann–Whitney U test was used for comparison of paired nonparametric data. A p -value of ≤0.05 was considered significant. The binary logistic regression model was
used to model the probability of vasospasm with respect to the length of the AFA.
Results
One hundred cases the fulfilled inclusion criteria (M:F= 48:52; age: 26–75 years;
mean age: 53.07 years). Eighty-nine patients were treated by surgical clipping, while
the remaining were managed with endovascular coiling. Preclipping/coiling DSA was
available for 82 cases, among which 43 (52.4%) had a demonstrable AFA. The AFA was
present in 59/100 cases in pre-NCA DSA (with a mean length of 3.45 ± 3.3 cm from the
ACF base) and in 57/100 cases in post-NCA DSA. Forty-seven (47%) cases of the angiographic
controls had AFA visualized on DSA (2.19 ± 2.54 cm from the ACF base). This intergroup
difference was statistically significant (p = 0.004). Although there was a trend for a longer length of AFA in the cases with
aneurysms located in the ACA territory (3.68 vs. 3.33 cm, with p = 0.7237), and when the vasospasm primarily involved the ACA and its branches (3.50
vs. 3.03 cm, respectively, with p = 0.6308), and cases with a higher modified Fisher grade of SAH (longer by 0.99 cm
in grade 4 SAH; p = 0.276), the differences were not statistically significant.
Following NCA, the mean reduction in AFA length was 0.49 cm in “mild,” 0.78 cm in
“moderate,” and 0.81 cm in cases with “severe” vasospasm ([Table 1 ], [Figs. 2 ]
[3 ]
[4 ]). Mixed effect analysis results for the pre- and postintervention comparison for
the various degrees of vasospasm were the following: irrespective of the severity
of the vasospasm, the reduction in the length of the AFA on post-nimodipine DSA was
significant (p < 0.001). The intergroup differences in the absolute change in the length of the
AFA post-NCA was significant (“mild” vs. “severe”: 3.653 cm; p = 0; “mild” vs. “moderate”: 0.629 cm in moderate; p = 0.354). There was no significant interaction effect in the “moderate versus mild”
spasm and “severe versus mild” spasm comparison. There was a trend toward more reduction
in the AFA length post-NCA in a higher-grade SAH (0.458 cm in grade 4 SAH, [p = 0.088] vs. 0.338 cm [p = 0.144], 0.09 cm [p = 0.84], and 0.192 cm [p = 0.493] in grades 1, 2, and 3, respectively; [Table 2 ]).
Table 1
Mean lengths of anterior falcine artery (AFA) according to vasospasm severity (length
in cm)
Time point
Vasospasm severity
Mean length of AFA
Standard deviation
Pre-nimodipine
Mild
2.74
3.17
Post-nimodipine
Mild
2.25
2.75
Pre-nimodipine
Moderate
3.37
3.15
Post-nimodipine DSA
Moderate
2.59
2.55
Pre-nimodipine
Severe
6.39
2.72
Post-nimodipine DSA
Severe
5.58
2.64
Abbreviation: DSA, digital subtraction angiography.
Fig. 2 Line graph of mixed effect analysis of the length of the anterior falcine artery
before and after administration of intra-arterial nimodipine in mild, moderate, and
severe vasospasm groups.
Fig. 3 A 51-year-old male patient. Following anterior communicating artery aneurysm clipping,
(A ) the pre-nimodipine lateral angiogram shows clip in situ (white arrow ) with mild vasospasm in the A2–A3 segments of the anterior cerebral artery. (B ) Decrease in the length and prominence of the anterior falcine artery (black arrows ) above the anterior cranial fossa base is demonstrated after administration of intra-arterial
nimodipine. Note the improvement of brain perfusion after intra-arterial nimodipine
infusion.
Fig. 4 A 35-year-old female patient following anterior communicating artery aneurysm clipping.
(A ) The pre-nimodipine anteroposterior angiogram shows severe spasm of the supraclinoid internal carotid artery, A1 segment of the anterior cerebral
artery, and M1–M2 segments of the middle cerebral artery. (B ) The pre-nimodipine lateral angiogram shows clip in situ (white arrow ) with conspicuously visualized the anterior falcine artery (AFA; black arrow ). (C ) Marked decrease in the length of the AFA and its poor visualization (black arrow ) are demonstrated after administration of intra-arterial nimodipine.
Table 2
Results of mixed effect analysis in pre- and post-nimodipine data with different degrees
of vasospasms (length in cm)
Variable (length of AFA from ACF base)
Estimate
p
Main effect: change in AFA length with nimodipine infusion irrespective of vasospasm
severity
−0.488
< 0.001
Main effect: change in AFA length in moderate spasm compared with mild spasm (irrespective
of nimodipine infusion)
0.629
0.354
Main effect: change in AFA length in severe spasm compared with mild spasm (irrespective
of nimodipine infusion)
3.653
< 0.001
Interaction effect: moderate vs. mild spasm (with nimodipine infusion, pre vs. post)
−0.295
0.171
Interaction effect: severe vs. mild spasm (with nimodipine infusion, pre vs. post)
−0.326
0.224
Abbreviations: ACF, anterior cranial fossa; AFA, anterior falcine artery.
Binary logistic regression analysis revealed that for every centimeter increase in
the length of the AFA, there was an 18.9% increase in the probability of angiographic
vasospasm.
Discussion
A summary of the key results from our study is presented thus: the AFA was visualized
on ICA angiography in higher than 53% of all subjects who underwent angiography. The
mean length of the AFA from the ACF base was significantly higher in patients with
SAH-V compared with the angiographic controls. The AFA significantly reduced in length
following intra-arterial nimodipine/vasodilator infusion in patients with severe angiographic
vasospasm, compared with those with milder degrees of vasospasm. The AFA is in essence
the continuation of the AEA supplying the midline falx dura around the SSS as far
posterior as the coronal suture.[6 ] The AEA, the parent artery of the AFA, has been studied in cadaveric dissections
by White et al.[7 ] The definition of the origin of the anterior falx artery in their work is as follows:
the AEA originates from the OA, approaches the medial wall of the orbit, and via the
anterior ethmoidal foramen (AEF) courses through the ethmoid air cells. At the cribriform
plate, it turns superiorly forming the anterior falx artery to travel between the
two layers of dura. In this work, they demonstrate three sites for hemostatic control
of the AEA, namely, (1) AEF in the medial orbital wall (lamina papyracea), (2) anterior
ethmoid canal (in the location of the lateral ethmoid wall), and (3) at the cribriform
plate (extradural). The third site denotes the origin of the AFA for hemostatic control
of lesions of the ACF base via a single-flap craniotomy in the fronto-orbital location.
In another prior work, Müller had highlighted that the AFA is a continuation of the
AEA and that the course may be further posterior than what was once believed. A continuation
into the parietal dura of the SSS later uniting with the branches (frontal) of the
middle meningeal artery (MMA) at the coronal suture was found. The surgical anatomy
of the AEA leading onto the AFA has also been reported by Moon et al with similar
observations.[8 ] Cross-connections with the companion AFA (paired) exist.[9 ] Venae comitantes are absent in these paired vessels. The AFA is also involved in
anastomosis with the posterior ethmoidal artery.[10 ] In OA occlusion, the AFA, by virtue of its connections with the MMA, can restitute
orbital collateral circulation.[9 ] Collateralization with the branches of the lacrimal artery has also been described.
Displacement of the AFA by intracranial masses is rare.[4 ] The vessel is usually opacified only from one side. Given that the course of the
AFA is not leptomeningeal, it is unlikely to be compromised by post-SAH-V.
During 1979, Müller described the AFA and its anatomic relationships to the dural
veins, the SSS, and the arachnoid granulations. Histology of the artery highlighted
its modification to be compliant to longitudinal stretching. The intricacy of its
relationships hinted a functional importance greater than just nourishment of the
dura.[6 ] Thick walled arteries were noted flanking the SSS whose pulsations would help with
moving the venous blood. It is reasonable to believe that the description alludes
to the AFA.[11 ]
The angiographic visibility of the anterior falx artery has been variably reported,
ranging from 8.7 to 21% in apparently normal clinical conditions.[4 ]
[12 ] Across all the groups evaluated in the current study (including the angiographic
controls), the percentage of angiograms with AFA visualization was consistently higher.
It ought to be noted that the percentages cited above are sourced from angiographic
literature from ethnically different populations and from a time when the angiographic
resolution and the sophistication for digital subtraction was lesser. The rates of
AFA visibility that we report are to be interpreted in light of these determinants.
While it is now considered that the AFA does not invariably regress in its entirety
after the period of development, definite data regarding its identification (percentages)
in surgical/dissection specimen are lacking.
We refer to the literature on the dynamically responsive nature of the caliber of
the AFA. Quite understandably, a greater proportion of patients (up to 35%) whose
pathologies are locoregionally related to the anterior falx show a prominent falcine
artery.[12 ] Hypertrophy of the AFA has hitherto been reported in pathological contexts of dural
arteriovenous fistulas of the anterior cranial fossa base,[13 ] arteriovenous malformations,[14 ] subdural empyema in the interhemispheric region,[15 ] primary neoplasms (meningiomas and glioblastomas),[16 ] metastasis,[17 ] Paget's disease,[4 ] and dural–pial collateralization.[5 ] In their report of a case of anterior subdural (interhemispheric) empyema, Mitsuoka
et al interpreted that exaggerated angiographic dimension of the AFA was a reactive/reversible
phenomenon to pachymeningeal inflammation/irritation.[15 ] Such observations testify the sensitive nature of the AFA.
Dural–pial collateralization between the AFA and the mediofrontal/parasagittal branches
of the callosomarginal trunk are well described by Hawkins.[18 ] Proof of their existence and compensatory nature is drawn from moyamoya disease,
wherein this dural–pial collateralization hypertrophies to sustain the at-risk ischemic
parenchyma. In fact, OA collaterals via the AFA are among the most frequent in moyamoya
disease with their development and dimensions in balance with the perfusion needs
of the at-risk anterior cerebral artery territories.[5 ] While it seems incongruent to compare collateralization in moyamoya disease (an
indolent slow-progressing disease) to the acute course of post-SAH-V, it is worth
noting that in stroke collaterals do develop in as early approximately 3 to 5 days.[17 ] Post-SAH-V is unusual before 3 days of the ictus.[19 ] Cerebral vasospasm is maximally symptomatic and angiographically most evident in
the ACA territory.[20 ] That the anterior communicating artery is among the most common sites for ruptured
intracranial aneurysms is known.[21 ] Summarizing by taking into account the time course of the SAH-V evolution, at-risk
parenchyma being the distal ACA territory, the availability of dural–pial collateralization
from the AFA, and the evidence for reactive/sympathetic nature of the AFA in locoregional
pathology (alluded to earlier), we state that the observed statistically significant
higher percentage of AFA visualization in our study testifies the AFA as a key contributor
to sustain the vasospastic territory. We substantiate the a priori hypothesis of our
study by emphasizing that AFA dural–pial collaterals hypertrophy due to their reciprocal
balance with the constricted pial vasculature is in cases with SAH-V especially if
the latter involves ACA territory. The correlation between the degree of vasospasm
with the length of the AFA that we noted also is in line with these observations.
Our purpose of using the “length” of the AFA above the anterior cranial fossa floor
as the objective parameter as against the “diameter” was to ensure better reproducibility
and eliminate the possible chances of error in calibration/measurement of the extremely
small girth of the AFA.
The reciprocal balance between the AFA and the pial vasculature is discussed earlier.
Intra-arterial vasodilators, such as nimodipine, decrease resistance and increase
blood flow through the circle of Willis. With such changes, it is expected that postinfusion
of vasodilators, dural–pial collateralization, although transient, becomes less vigorous.
The postinfusion decrease in the length of the AFA that we noted to be significant
in the group with severe vasospasm aligns with the prediction. It thus reaffirms the
socialist role of the AFA for deprived brain parenchyma. Although not statistically
significant, we found a trend of longer lengths of the AFA if aneurysm and subsequent
vasospasm involve the ACA territory, higher grades of SAH, and clipping as compared
with coiling. In binary logistic regression analysis also, we found a positive correlation
between the probability of the presence of vasospasm with the length of the AFA and
patient age. To the best of our knowledge, no report of the role of AFA in SAH-V and
its dynamic changes in angiography exists in the literature.
We acknowledge shortcomings of this retrospective study. The control population that
was included were diseased controls (not matched for age) who had no clinical/angiographic
evidence of vasospasm and in whom in the vascular abnormality was remote from the
AFA/ACA/anterior parasagittal territory. “Healthy controls” in angiography, given
the radiation/ethical concerns, are not feasible. Vasospasm was judged based on visual
assessment and may be prone to observation errors. The numbers of participants in
the case and control groups were not matched. The case group was heterogeneous with
regard to the location of aneurysm and treatment offered. Further influence of gender,
ethnicity, etc., on the skull shape can affect the AFA measurements. Other potential
dural–cortical collaterals, which can potentially supply the at-risk cerebral parenchyma,
were not evaluated in our study. Also, subgroup analysis of the clipped versus coiled
group and the effect of location of the aneurysm and the associated hematoma on AFA
was not performed.
Conclusion
The AFA is visualized on ICA angiography at a greater frequency than has been earlier
reported in the literature. As an angiographic equivalent to the presence of vasospasm,
the mean length of the AFA from the ACF base was significantly higher in patients
with SAH-V compared with angiographic controls. The AFA significantly reduced in length
following intra-arterial nimodipine infusion in patients with severe angiographic
vasospasm, compared with those with milder degrees of vasospasm, further substantiating
this claim of its role in autoregulation in vasospasm. A trend toward a longer length
of the AFA is seen in grade 4 SAH, ACA territory aneurysms, and vasospasm involving
the ACA territory. Further studies are needed for exploring and validating the compensatory
role of the anterior falx artery in cerebral vasospasm and to probe its significance
as an angiographic sign that antedates the onset of angiographic vasospasm.
