Intracranial dural arteriovenous fistulas (DAVF) is an abnormal arteriovenous connection that lies within the dura, first described in 1970 by Newton et al. [10]. Typically, DAVF involves the dura adjacent to venous sinuses. It shows markedly variable structures [5]. One such unusual configuration is that of intraosseous diploic venous drainage. Most cases of intraosseous DAVF involve the clivus and foramen magnum due to a high concentration of transosseous emissary veins [4,9]. DAVF involving diploic veins in the supratentorial cranium is extremely rare [12]. We encountered a case of DAVF involving diploic veins located in the left frontal bone, characterized by very high-flow feeders with multiple bilateral feeders. We report this case of spontaneous intraosseous DAVF involving the diploic veins, successfully treated with Onyx embolization via a transarterial route.

A 59-year-old male presented with mild right extremity hemiparalysis and aphasia that had persisted for one month. On palpation, the dilated superficial temporal arteries (STA) resembled a mass. He had no obvious history of head trauma, surgery, or venous sinus thrombosis. Brain computed tomography (CT) showed a prominent intraosseous diploic space in the left frontal bone with a small defect on the inner table of the skull (Fig. 1A). Contrast-enhanced magnetic resonance imaging demonstrated tangled vessels in the left parietal area (Fig. 1B), indicating shunt disease. Digital subtraction angiography revealed DAVF within the left parietal bone (Fig. 2A-C). It was fed by the dilated bilateral STA, left middle meningeal artery (MMA), and left occipital artery. It drained through the diploic veins into the left parasinus (parasinus is a part of a sinus that can occur due to a septum) of the superior sagittal sinus (SSS) with retrograde cortical venous reflux (CVR). No pial feeders from the internal carotid artery contributed to the shunt. This case may have presented symptoms due to venous outflow impairment caused by CVR, and we planned the surgery.

Pretreatment angiography classified DAVF as Borden type and Cognard TypeⅡa+b. The patient had symptomatic DAVF, and we performed endovascular treatment. Transarterial embolization (TAE) with Onyx-18 (Medtronic, Irvine, CA, USA) was selected due to the shunt points separated from SSS. Under general anesthesia, a 6 Fr long sheath was inserted into the right radial artery.

A 6 Fr Optimo balloon-guiding catheter (Tokai Medical Products, Kasugai, Aichi, Japan) was navigated into the left external carotid artery. Furthermore, an 8 Fr long sheath was inserted into the right femoral artery, and an 8 Fr FlowGate2 balloon-guiding catheter (Stryker, Kalamazoo, MI, USA) was navigated into the right external carotid artery. First, we planned Onyx injection from the left STA. A 3.2 Fr Tactics Plus distal access catheter (Technocrat, Kasugai, Aichi, Japan), and Marathon microcatheter (Medtronic, Irvine, CA, USA) and Chikai X010 0.010 inch/200 cm (Asahi Intecc, Seto, Aichi, Japan) were inserted into FlowGate2 and navigated to a shunt point near the diploic vein under roadmap guidance. Then, the Optimo and FlowGate2 balloons were both inflated to fully block proximal flow, and TAE of the left STA with Onyx was performed. Feeders from the left STA and middle meningeal artery were embolized using Onyx 6 vials (5.4 mL), with the “plug and push” technique [16]. Onyx embolized a portion of the draining vein beyond the shunt point but did not achieve complete occlusion. Next, another Tactics Plus and Marathon were inserted to the Optimo and navigated just proximal to the shunt point from the right STA. STA feeders were embolized with Onyx to penetrate the draining vein, using 2 vials (1.4 mL) (Fig. 3). Angiography performed at the completion of embolization confirmed resolution of the arteriovenous shunt (Fig. 4). During the injection, we inflated both balloon-guiding catheters and occluded both ECAs to prevent Onyx migration towards the SSS side [7]. Subsequently, the patient showed no neurological deficits and the mild right hemiparalysis and aphasia resolved. A 6-month follow-up angiogram confirmed persistent complete embolization of DAVF with no residual arteriovenous shunt. At that time, the patient was symptom-free with no reported neurological signs or symptoms.

DAVF draining into the diploic veins is rare, referred to as diploic arteriovenous fistula (AVF), intraosseous AVF, or intraosseous DAVF [1,2,4]. Cases draining exclusively into the diploic veins constitute an extremely rare subset [12]. Malik et al. [8] reported the first cases of DAVF with an intraosseous nidus. The nidus in those cases was located around the foramen magnum, and the patients presented with pulsatile tinnitus. Most cases of osseous DAVF involve the clivus and foramen magnum [12]. To our knowledge, 39 cases of intraosseous AVF have been reported. Of these 39, intraosseous AVF in the supratentorium comprised only 8 cases (Table 1) [1,2,6,7,12,14,17,18]. There were no case reports in which DAVF drained solely into the diploic veins, causing neurological deficits by venous stasis. Eight reports of osseous DAVF demonstrated low-flow shunts or only a few feeders. The present case, however, involved very high-flow feeders with multiple bilateral feeders and STA was dilated by the shunt and appeared mass-like. Possible causes of DAVF include head injury, cranial surgery, and venous sinus thrombosis [7]. However, the present patient had no history of them. DAVF draining into the diploic veins rarely causes CVR or neurological deficits since diploic veins generally flow into SSS and have no direct connection with cortical veins under physiological conditions. Typically, if there is a shunt that flows into the diploic vein, it should flow antegradely into SSS. However, in this case, the shunt that flowed into the diploic vein drained retrogradely, and cerebral CVR was observed.

Therefore, there may have been an obstruction that prevented the diploic veins from draining into SSS with antegrade flow. Yako et al. [18] reported a possible mechanism for DAVF draining solely into the diploic veins, as follows: The diploic veins might have connections not with SSS itself but with the venous lacuna via emissary veins. In the process of arteriovenous shunt maturation, occlusive change of the drainage site may occur in the channel between the venous lacuna and SSS. The venous lacuna might be isolated from SSS, and consequently the shunt flow might be directed via the venous lacuna to cortical veins. In the present case, we propose the following possible mechanism: The parasinus might have communicated with SSS originally. During shunt development, the parasinus may have been isolated from SSS incompletely, and flow via “diploic vein → venous lake → parasinus → cortical vein” occurred. Consequently, the shunt flow might have been directed to the cortical vein with CVR (Fig. 5).

Treatment includes transvenous embolization (TVE), surgery, and TAE, but there are no consistent conclusions. Reports of treatment with TAE have increased with the advent of Onyx and n-butyl-2-cyanoacrylate (NBCA) [3,11]. In our case, the retrograde approach to the sinus draining DAVF via the ipsilateral internal jugular vein was not technically straightforward, since stenosis was present between the pterygoid plexus and the diploic vein. We selected TAE for the following two reasons: First, TAE would not increase the risk of neurological injury, since the main feeders of DAVF were STA. Second, Onyx control with the “plug and push” technique helped avoid migration into the SSS, as a shunt point separated from the SSS. If TAE had been difficult, we planned to consider TVE. A problem regarding TAE in this case was that the shunt exhibited markedly high flow rates; STA was bilaterally dilated, with significant dilation observed on the left side. High-flow feeders could facilitate Onyx migration to SSS with blood flow. It has been reported that flow control techniques with balloon assistance are safe and effective adjunctive methods in primary endovascular Onyx embolization of high-flow DAVFs [13]. In our case, balloon-guiding catheters were placed in the bilateral external carotid arteries, and Onyx infusion was applied under flow control as the enlarged STA and the diploic veins suggested high-flow feeders.

Flow control facilitated completion of the procedure with Onyx without migration into the critical sinus. As in this case, TAE with a balloon-guiding catheter may constitute effective treatment for intraosseous DAVF with high-flow feeders.

We reported a case of intraosseous DAVF involving the diploic veins with high-flow feeders. One of the possible causative mechanisms is putative occlusive change of the venous draining system during shunt formation. In this case, using the balloon-guiding catheters, the treatment could be completed without Onyx migration to the critical sinus, even in the presence of highflow feeders. Although each case must be considered individually, TAE with flow control may be an effective surgical procedure, even in cases of DAVF with high-flow feeders such as the present patient.