CASE REPORT
Traumatic Middle Meningeal Arteriovenous Fistula Presenting with Delayed Contralateral Temporal Subcortical Hemorrhage after Surgery for Cerebral Contusion: A Case Report
2026 年 13 巻 p. 517-522
詳細
2026 年 13 巻 p. 517-522
Traumatic middle meningeal arteriovenous fistula is an uncommon and likely underrecognized dural arteriovenous shunt that develops after head trauma and may occasionally present with delayed intracranial hemorrhage. We report a case of traumatic middle meningeal arteriovenous fistula with delayed contralateral temporal subcortical hemorrhage following surgery for cerebral contusion. A man in his 50s underwent emergency hematoma evacuation and decompressive craniectomy for progressive right temporal contusion after severe head injury. During follow-up, a low-density lesion gradually enlarged in the contralateral left temporal lobe. On day 14, his consciousness deteriorated, and computed tomography revealed left temporal subcortical hemorrhage. Contrast-enhanced computed tomography failed to show any vascular abnormality, whereas magnetic resonance angiography showed abnormal dilatation of the left middle meningeal artery. Cerebral angiography demonstrated traumatic middle meningeal arteriovenous fistula supplied by the left middle meningeal artery and draining through the middle meningeal vein into the pterygoid plexus. Although definite cortical venous reflux was not demonstrated on angiography, venous-phase internal carotid angiography showed stagnation in the superficial middle cerebral vein, which improved after fistula obliteration. Cone-beam computed tomography and four-dimensional digital subtraction angiography precisely localized the fistula, and complete obliteration was achieved with transarterial coil embolization. Thus, venous congestion or venous hypertension related to traumatic middle meningeal arteriovenous fistula may contribute to delayed contralateral hemorrhage, although pre-existing cortical venous reflux could not be excluded. In severe head trauma, traumatic middle meningeal arteriovenous fistula should be considered when unexplained deterioration in consciousness or progressive low-density changes develop during follow-up. Early vascular evaluation, including magnetic resonance angiography when contrast-enhanced computed tomography is unrevealing, may facilitate prompt diagnosis and safe treatment with transarterial coil embolization.
Traumatic middle meningeal arteriovenous fistula (T-MMAVF) is an uncommon and likely underrecognized dural arteriovenous shunt that develops after head trauma. Freckmann et al.1) identified T-MMAVF in 1.8% of 446 patients who underwent angiographic evaluation after head injury. A recent systematic review showed that head trauma is the most common cause of MMAVF and that endovascular treatment represents the mainstay of management, achieving a higher obliteration rate (95.8%) and a lower complication rate than surgery.2) In contrast, T-MMAVF presenting with intracranial hemorrhage is uncommon, and the underlying pathophysiology is not fully understood. Furthermore, diagnosis may be delayed when the lesion develops contralateral to the site of impact or becomes symptomatic only days to weeks after the initial trauma.3-6) Such diagnostic delay may postpone appropriate vascular evaluation and treatment and may allow progression to hemorrhagic complications.
We report a case of T-MMAVF that presented with delayed temporal subcortical hemorrhage contralateral to the initial cerebral contusion following severe head trauma. This case highlights the importance of considering T-MMAVF when unexplained deterioration in consciousness or progressive low-density changes occur during follow-up after severe head trauma.
A man in his 50s was brought to the emergency department after a fall while intoxicated. His medical history was significant for hypertension. On arrival, his Glasgow Coma Scale (GCS) score was 9 (E2V2M5), and no focal motor deficit was observed. Head computed tomography (CT) showed a mild right temporal contusion, a mild right acute subdural hematoma, and subarachnoid hemorrhage, but no skull fracture was identified (Figure 1A and B). Contrast-enhanced CT performed at admission did not reveal any obvious vascular abnormalities (Figure 1C). Three hours after the injury, the patient developed left hemiplegia, and follow-up CT showed significant progression of the right temporal contusion (Figure 1D). Decompressive craniotomy and hematoma evacuation were performed immediately, and the patient's impaired consciousness and left hemiplegia improved (Figure 1E).
Initial and early postoperative computed tomography findings.
A: Initial non-contrast computed tomography showing a mild right temporal contusion and diffuse traumatic subarachnoid hemorrhage. B: Bone-window computed tomography showing no evidence of skull fracture. C: Contrast-enhanced computed tomography showing no apparent vascular abnormality. D: Follow-up computed tomography obtained 3 h after admission showing progression of the right temporal contusion. E: Computed tomography obtained immediately after craniotomy for hematoma evacuation and decompressive craniectomy.
Although no new neurological defects were observed, follow-up CT revealed a low-density area in the left temporal lobe that progressively enlarged from day 4 (GCS E3VTM4) to day 10 (GCS E4V1M5; Figure 2A and B). On day 14, his level of consciousness deteriorated; spontaneous eye opening, which had been present the previous day, was no longer observed, and his GCS score worsened to 6 (E1V1M4). Head CT demonstrated a left temporal subcortical hemorrhage (Figure 2C).
Serial imaging findings before diagnosis of traumatic middle meningeal arteriovenous fistula.
A: Follow-up computed tomography obtained 4 days after injury showing a low-density area in the left temporal lobe. B: Computed tomography obtained 10 days after injury showing progressive enlargement of the low-density area in the left temporal lobe (arrow). C: Computed tomography obtained 14 days after injury showing left temporal subcortical hemorrhage following deterioration in consciousness. D: Contrast-enhanced computed tomography showing no apparent vascular abnormality, similar to the admission findings. E: Magnetic resonance angiography showing abnormal dilatation of the left middle meningeal artery (arrow), visible up to the proximal segment.
Contrast-enhanced CT again failed to show any obvious vascular abnormality (Figure 2D); however, magnetic resonance angiography (MRA) suggested abnormal dilatation of the left middle meningeal artery (MMA; Figure 2E). Because these findings raised suspicion of an underlying shunt lesion, cerebral angiography was performed promptly. Left external carotid angiography demonstrated a T-MMAVF supplied by the left MMA and draining through the middle meningeal vein (MMV) into the pterygoid plexus (Figure 3A). No definite cortical venous reflux (CVR) was identified. In contrast, venous-phase left internal carotid angiography demonstrated stagnation in the left superficial middle cerebral vein (SMCV; Figure 3B1 and B2). Based on these findings, venous outflow disturbance related to the left T-MMAVF was considered the most likely contributor to the left temporal subcortical hemorrhage.
Angiographic findings and endovascular treatment of traumatic middle meningeal arteriovenous fistula.
A: Lateral view of left external carotid angiography showing traumatic middle meningeal arteriovenous fistula with connection between the left middle meningeal artery and the middle meningeal vein (arrow). B1, 2: Pre-embolization venous-phase lateral views of left internal carotid angiography showing delayed and stagnation in the left superficial middle cerebral vein (arrowheads). C: Pre-embolization cone-beam computed tomography (sagittal view) showing the fistula. D: Post-embolization lateral view of left external carotid angiography confirming complete obliteration of the fistula. E1, 2: Post-embolization venous-phase lateral views of left internal carotid angiography showing improvement of venous stagnation in the left superficial middle cerebral vein. F: Postoperative magnetic resonance angiography showing resolution of the previously dilated left middle meningeal artery.
Before treatment, the fistula was evaluated in detail using cone-beam CT (Figure 3C) and 4-dimensional digital subtraction angiography (DSA). These studies clearly localized the presumed fistula between the MMA and the MMV. Because the fistula could be accessed transarterially, transarterial coil embolization was chosen as a minimally invasive treatment. A microcatheter was then advanced across the fistulous connection into the MMV. Coils were subsequently deployed from the MMV back across the fistulous connection toward the MMA, resulting in complete obliteration of the fistula using a total of five coils (Figure 3D).
Postprocedural angiography confirmed complete obliteration of the fistula and improvement in SMCV congestion (Figure 3E1 and E2). Postoperative MRA also demonstrated resolution of the previously dilated left MMA (Figure 3F). Subsequently, endoscopic evacuation of the left temporal subcortical hematoma was performed, followed by right cranioplasty. After stabilization of his general condition, the patient was transferred to a rehabilitation hospital with a GCS score of E4V2M5.
The most important feature of this case is that, after severe head injury, a delayed intracerebral hemorrhage developed contralateral to the initial site of cerebral contusion, ultimately leading to the diagnosis of T-MMAVF. As summarized in Table 1, previously reported hemorrhagic cases share several common features, including delayed onset, discordance between the site of impact and the lesion location, the absence of skull fracture in some cases, and generally favorable obliteration outcomes following endovascular treatment.2-11) Therefore, when neurological deterioration after severe head trauma cannot be adequately explained by the initial imaging findings, delayed vascular lesions should be considered, and early vascular evaluation should be pursued.
Reported Cases of Traumatic Middle Meningeal Arteriovenous Fistula Presenting with Intracranial Hemorrhage
| Case No. | Author (Year) | Age (years)/sex | Impact side | Lesion side | Bone fracture | Hemorrhage | Interval between trauma and symptom onset | Feeder characteristics | Draining route from MMV | Treatment | Outcome |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Only definite hemorrhagic cases were included. Cases were selected when the intracranial hemorrhage was considered directly attributable to traumatic middle meningeal arteriovenous fistula. +: present; –: absent; EVT: endovascular treatment; F: female; ICH: intracerebral hemorrhage; M: male; MMA: middle meningeal artery; MMV: middle meningeal vein; NA: not available; NBCA: n-butyl cyanoacrylate; SAH: subarachnoid hemorrhage; SDH: subdural hematoma; SMCV: superficial middle cerebral vein; TAE: transarterial embolization |
|||||||||||
| 1 | Nakahara et al. (8) (1986) | 30/M | Left | Right | + | SDH | 9 years | Single MMA branch | Cortical vein | Surgery | Stable neurological status |
| 2 | Takeuchi et al. (9) (2008) | 57/M | Right | Left | + | SDH, ICH | 1 day | Single MMA branch | Cortical vein | Surgery | Survival with severe deficits |
| 3 | Sakata et al. (10) (2009) | 48/F | Right | Right | + | SAH | 15 years | Venous aneurysm | SMCV | Surgery | NA |
| 4 | Almefty et al. (7) (2016) | 60s/F | Right | Right | NA | ICH | 16 days | Pseudoaneurysm | NA | EVT (NBCA-TAE) | Stable neurological status |
| 5 | Goyal et al. (11) (2023) | 49/M | Right | Right | + | ICH | 10 days | Single MMA branch | Cortical vein | EVT (Onyx-TAE) | Good recovery at 8 months |
| 6 | Present case | 59/M | Right | Left | − | ICH | 14 days | Single MMA branch | Pterygoid plexus | EVT(Coil-TAE) | Survival with severe deficits |
In the present case, the left temporal subcortical hemorrhage was most plausibly related to focal venous congestion and/or venous hypertension caused by T-MMAVF. Although definite CVR was not demonstrated angiographically, venous-phase internal carotid angiography showed stagnation in the SMCV, which improved after fistula obliteration. This temporal change suggests that the T-MMAVF imposed a hemodynamic burden on the local venous drainage system. However, because angiography was performed after the hemorrhage development, hematoma-related mass effect may have altered the venous drainage pattern and masked pre-existing CVR. Iki et al.6) reported a case of T-MMAVF arising at a non-fracture site that caused delayed venous infarction, in which four-dimensional CT angiography demonstrated retrograde slow flow and a to-and-fro pattern, supporting the role of venous congestion in symptom development. Similarly, in the present case, impaired venous drainage likely contributed to hemorrhage formation. However, attributing this complication solely to the SMCV would be inappropriate. Rather, the shunt burden from the MMV to the pterygoid plexus likely altered the overall local venous drainage environment, resulting in functional congestion within the SMCV system.
An important lesson from this case is that a gradually enlarging low-density area had already been present in the left temporal lobe before overt hemorrhage developed. In retrospect, this finding should not have been interpreted as merely nonspecific edema. Rather, the lesion likely represented a warning sign indicating that T-MMAVF had already formed and that venous perfusion disturbance was progressively worsening. Takada et al.3) similarly emphasized that MMAVF may become symptomatic after a prolonged delay, even in the absence of temporal bone fracture, and that atypical delayed symptoms after trauma should prompt evaluation for an underlying vascular lesion. Recognition of progressive enlargement of a low-density lesion during follow-up after severe head trauma should therefore prompt consideration of newly developed posttraumatic vascular pathology in the differential diagnosis, in addition to persistent contusion-related changes or secondary ischemia. Earlier vascular investigation would likely have been warranted in this case, underscoring the clinical importance of early vascular evaluation when atypical intraparenchymal imaging findings emerge during follow-up.
An iatrogenic mechanism, including head-holder pin-related injury, was also considered because of the contralateral location. However, postoperative CT demonstrated the remoteness of the pin fixation site from the presumed fistulous site, making a direct pin-related mechanism less likely. Nevertheless, because the lesion developed after craniotomy, an iatrogenic contribution cannot be completely excluded retrospectively. Development of T-MMAVF contralateral to the site of impact or at a non-fracture site cannot be explained solely by direct vascular injury along a skull fracture line. Previous reports have proposed mechanisms such as contrecoup-related shear stress and separation between the dura and the inner table of the skull.4-6) In the present case, the initial traumatic lesions were predominantly right-sided, whereas T-MMAVF developed on the left, a pattern consistent with these proposed mechanisms. From a diagnostic perspective, contrast-enhanced CT failed to clearly identify the shunt lesion, whereas MRA demonstrated an abnormal vessel, and DSA established the diagnosis. Because MMAVF often represents a small shunt near the skull base, detection on single-phase contrast-enhanced CT may be challenging because of overlap with adjacent bony structures and venous plexuses. In contrast, time-of-flight MRA may provide an early diagnostic clue in such cases.3,12) In the present case, repeated contrast-enhanced CT remained unrevealing, whereas MRA demonstrated abnormal MMA dilatation, suggesting that MRA may provide a more useful diagnostic clue than contrast-enhanced CT in selected skull-base MMAVFs obscured by adjacent bone structures and venous plexuses.
Endovascular treatment is currently considered the most appropriate first-line strategy for T-MMAVF.2) When the fistula can be clearly identified, and a microcatheter can be safely advanced to the target, coil-only embolization may provide an effective option that balances safety and curative potential.12,13) In the present case, coil embolization was selected because the use of liquid embolic material from the proximal MMA raised concern regarding inadvertent migration through dangerous anastomoses. Precise localization of the fistula using cone-beam CT and 4-dimensional DSA enabled complete obliteration with only five coils.
This report has several limitations. First, because it describes a single case, the findings may not be generalizable. Second, the proposed relationship between T-MMAVF and hemorrhage, including the role of venous congestion, was inferred from serial imaging findings and hemodynamic changes after fistula obliteration rather than directly proven. Third, because definite CVR was not demonstrated, the precise mechanism of hemorrhage remains speculative. Considering that angiography was performed after hemorrhage development, hematoma-related mass effect may have altered the venous drainage pattern and masked pre-existing CVR. Nevertheless, the temporal progression of the imaging abnormalities and their improvement after treatment support the clinical relevance of this interpretation.
T-MMAVF may develop contralateral to the initial site of cerebral contusion after severe head trauma and may present with delayed intracranial hemorrhage. In patients who have experienced severe head trauma, T-MMAVF should be considered when deterioration in consciousness cannot be fully explained by the initial traumatic lesions or when a low-density lesion progressively enlarges during follow-up. Early vascular evaluation is warranted in such situations. Precise localization of the fistula using cone-beam CT and 4-dimensional DSA may facilitate safe and reliable treatment with transarterial coil embolization alone.
All authors have no conflicts of interest.
The Institutional Review Board of University of Miyazaki Hospital approved the study (approval no. C-0194).
Consent for publication was not required because the submission does not include any images or information that could identify the patient.
