Delayed and incomplete treatment may result in dural fistula development in children with Vein of Galen malformation

Abstract

The association of dural arteriovenous fistulas (DAVF) in children with Vein of Galen malformation (VGM) has recently been reported for the first time. In a larger series of cases treated with transarterial NBCA embolization, 30% had DAVF. We wanted to analyze the development of DAVF in our cohort of children with VGM and to evaluate whether their occurrence depends on different treatment timing and embolic materials.

We analyzed 43 VGM cases treated with a combined transarterial and transvenous approach between 2003 and 2016. In our early series until 2011, we used coils solely in 21 children. Since 2012, 22 children were treated with the combination of coils and Onyx.

In the early series treated with coils solely, no case presented initially with or developed DAVF over time on follow-up angiograms. In our recent series we found four cases (9%) with DAVF. In two patients (5%), DAVF were found on the initial angiogram. Both patients presented at our department at age >2 years and were not treated elsewhere before. One patient (2%) presented at our department with too proximal occlusion of arterial feeders performed at another institution before. Only one patient (2%) developed DAVF in our department after the transarterial use of Onyx. Interestingly, this child did not develop DAVF as long as we used coils solely and his DAVF was localized exactly where an Onyx cast was identified.

In conclusion, delayed and incomplete treatment may have a considerable impact on the occurrence of DAVF in VGM.

Introduction

Vein of Galen malformation (VGM) is a severe neurovascular disease of early childhood.¹ Endovascular embolization is the treatment of choice. Embolization can be achieved either by the use of n-butyl cyanoacrylate (NBCA), coils, Onyx, or by a combination.^1,2 The association of dural arteriovenous fistulas (DAVF) in children with VGM has recently been reported for the first time.³ In a larger cohort of VGM cases treated with transarterial NBCA embolization, 30% of the patients presented with DAVF. The aim of this study was to identify cases and to analyze a possible development of DAVF in our cohort of children with VGM. Further, we wanted to evaluate whether their occurrence depends on treatment timing and the use of different embolic materials.

Materials and methods

We retrospectively reviewed digital subtraction angiography (DSA) and magnetic resonance angiography (MRA) of 43 true VGM patients from our database along with their hospital charts and follow-up outcome. Vein of Galen-like malformations such as Vein of Galen aneurysmal dilatation (VGAD) and any other pial arteriovenous malformation (AVM) were excluded. All children were treated with a combined transarterial and transvenous approach between 2003 and 2016 by the senior author Professor Dr Friedhelm Brassel. Institutional review board approval was obtained from the local hospitals’ research ethics board. All available data were reviewed with respect to the patient’s age, sex, clinical history, associated underlying pathomechanism, treatment timing, and choice of embolic material as well as the location of the DAVF and the imaging appearance. In our early series until 2011, we used coils solely in 21 children. Since 2012, 22 children were treated with the combination of coils and Onyx. We evaluated all catheter angiograms performed at six months, 12 months, and two years during endovascular treatment and on routine follow-up after successful embolization. Additionally, two-dimensional time-of-flight MRA were performed as routine clinical pre-embolization and follow-up studies.

Results

We identified four patients with DAVF among a cohort of 43 children with VGM between 2003 and 2016. Two of the children were male and two were female. All 43 children received embolization in multiple treatment sessions. In our early series until 2011, we used coils solely in 21 children. Since 2012, 22 children were treated with the combination of coils and Onyx. The results are summarized in Table 1.

Table 1.

The results of the study.

Patient no.	Sex	Age at first dx of VGM	Age at first tt (mo)	Age at first ID of DAVF (mo)	Location	Feeders	Treatment	DSA follow-up (mo)
1	M	4 d (MRI)	26	26 (MRI)	SSS + VGM	Left MMA	C + O	24
2	F	2 d (MRI)	27	21 (MRI)	VGM	Left PMA + MMA	C + O	18
3	M	4 mo (MRI+DSA)	4	15 (DSA)	VGM	Right MMA	C + O	18
4	F	3 d (MRI)	4	6 (MRI+DSA)	VGM	Both MMA	C + O	24

F: female; M: male; d: days; mo: DSA: digital subtraction angiography; months; dx: diagnosis; tt: treatment; ID: identification; MMA: middle meningeal artery; MRI: magnetic resonance imaging; PMA: posterior meningeal artery; SSS: superior sagittal sinus; O: Onyx; C: coils; VGM: Vein of Galen malformation.

In our early series of 21 patients treated with coils solely, no case presented initially with or developed DAVF over time on follow-up angiograms. However, we encountered four cases (9%) of DAVF in our recent series of 22 patients while using Onyx as an addition to coiling. In two patients (5%), DAVF were already found on the initial angiogram. Both patients presented at our department at age >2 years and were not treated before at other institutions. One patient (2%) presented at our department with too proximal occlusion of arterial feeders performed at another institution before. Only one patient (2%) developed DAVF at our department after the additional use of transarterial Onyx. Interestingly, this child did not develop DAVF as long as we used coils solely and the DAVF was localized exactly where an Onyx cast was identified.

In the following, we will highlight salient features of the DAVF and its relationship to the VGM, as well as their presumable pathomechanism on the basis of illustrative cases with different etiologies.

Illustrative cases

Case 1

At the age of four days, a male neonate was diagnosed with VGM by magnetic resonance imaging (MRI). Not having been treated before, he presented at our institution at the age of 26 months. Initial MRI retrospectively showed the choroidal-type VGM and the associated DAVF. Both the VGM as well as the superior sagittal sinus (SSS) were involved. We started embolization with a combination of Onyx and coils, which was performed in eight treatment sessions up to this date. The DAVF was first remarked on DSA in the fourth session at the age of three years, demonstrating the left middle meningeal artery (MMA) supply (Figure 1). There was known cause for the SSS component. Neither Onyx ended up passing into the SSS nor was there a secondary thrombosis of the vein. The boy is currently still under treatment.

Figure 1.

A male neonate was diagnosed with VGM by MRI. Not having been treated before, he presented at our institution at the age of 26 months. We started embolization with a combination of Onyx and coils, which was performed in eight treatment sessions up to this date. The DAVF was first remarked on DSA in the fourth session at the age of three years, demonstrating the left MMA supply ((a)–(c) anteroposterior view, (d)–(f) lateral view). Arrows mark the fistula point. VGM: Vein of Galen malformation; MRI: magnetic resonance imaging; DAVF: dural arteriovenous fistula; DSA: digital subtraction angiography; MMA: middle meningeal artery.

Case 2

A female neonate was diagnosed with VGM on her second day of life. She presented at our department at the age of 21 months without any prior treatment. We performed MRI, which showed a choroidal-type VGM and a DAVF, located at the VGM and fed by the posterior meningeal artery (PMA) (Figure 2(a)). DSA confirmed the diagnosis and showed additional dural supply by the left MMA (Figure 2(b)). We intentionally treated only the VGM but not the DAVF with a combination of coils and Onyx over a period of 15 months. After four embolizations, a diagnostic DSA at the age of five showed the complete occlusion of both the VGM and the DAVF, the latter having disappeared without any specific treatment.

Figure 2.

A female neonate was diagnosed with VGM on her second day of life. She presented at our department at the age of 21 months without any prior treatment. We performed MRI, which showed a choroidal-type VGM and a DAVF, located at the VGM and fed by the posterior meningeal artery (PMA) (a) DSA confirmed the diagnosis and showed additional dural supply by the left MMA (b) Arrows mark the fistula point. VGM: Vein of Galen malformation; MRI: magnetic resonance imaging; DAVF: dural arteriovenous fistula; DSA: digital subtraction angiography; MMA: middle meningeal artery.

Case 3

Only one patient developed DAVF after being treated at our department after the first use of transarterial Onyx. This boy presented at our department at the age of four months. We first started endovascular treatment by using coils solely, following the additional use of Onyx in the course of time. At the age of 15 months, DSA showed a new DAVF fed by the right MMA (Figure 3). The localization of his DAVF was exactly where an Onyx cast was identified.

Figure 3.

This boy presented at our department at the age of four months. We first started endovascular treatment by using coils solely, following the additional use of Onyx in the course of time. At the age of 15 months, DSA showed a new DAVF fed by the right MMA ((b), (d), arrows mark fistula point). The localization of his DAVF was exactly where an Onyx cast (arrow) was identified (c). DSA: digital subtraction angiography; DAVF: dural arteriovenous fistula; MMA: middle meningeal artery.

Case 4

A 6-month-old girl with a VGM presented at our department with a history of prior treatment at another institution, where too proximal occlusion of arterial feeders were performed without having reached and closed the arteriovenous fistulous connections. The diagnosis of a choroidal-type VGM was confirmed on both MRI and DSA and the DAVF was first identified at our institution involving both MMA (Figure 4). The girl is still under treatment for both the VGM and the associated DAVF requiring multiple staged treatment sessions using various embolization techniques and with the use of different embolic materials.

Figure 4.

A 6-month-old girl with a VGM presented at our department with a history of prior treatment at another institution, where too proximal occlusion of arterial feeders were performed without having reached and closed the arteriovenous fistulous connections. The diagnosis of a choroidal-type VGM was confirmed both on MRI and DSA and the DAVF was first identified at our institution involving both MMA. Arrows mark the fistula point. VGM: Vein of Galen malformation; MRI: magnetic resonance imaging; DSA: digital subtraction angiography; DAVF: dural arteriovenous fistula; MMA: middle meningeal artery.

Discussion

This is the second large case series study aiming to identify and analyze children with VGM developing DAVF. High-flow arteriovenous shunts such as VGM require meticulous treatment to avoid complications such as hydrocephalus,⁴ heart failure, or, as described in our study, the less investigated progression into associated DAVF. DAVF in VGM can either develop as a consequence of altered flow patterns inside the vessels involved with the VGM or can be seen as a secondary complication following endovascular embolization.

Presumed pathophysiology of DAVF in VGM

DAVF are rare pathological shunts between dural arteries and venous vessels such as dural venous sinuses, meningeal veins, or cortical veins. The pathogenetic mechanisms leading to the formation of DAVF remain only partially understood. Their occurrence has been documented following a wide range of conditions, including venous sinus thrombosis, vascular malformations, inflammatory processes, trauma, intracranial tumors, congenital predisposition as well as secondary development due to iatrogenic measures. These conditions may induce a change in vascular flow patterns due to a pathological recanalization as well as endothelial growth factor-induced dural angiogenesis.^5,6 These mechanisms lead to vascular remodeling, causing turbulent flow, thrombosis, and consequently a focally increased venous sinus pressure. The change in hemodynamics may induce the development of DAVF, which explains why a decreased venous sinus pressure following embolic therapy often causes the DAVF to regress.³ Up to now, DAVF development in VGM has been documented in only one case series.³ The authors hypothesized that either sprouting or non-sprouting angiogenesis, like a possible sump-effect, might induce the formation of DAVF in VGM. This effect might be either improved or aggravated by a change in hemodynamics following endovascular embolization.

However, the results of our study lead to the assumption that treatment timing and choice of embolic material might play an additional role in the formation of DAVF in VGM.

Treatment timing

The ideal treatment timing of VGM remains unclear and certainly depends on clinical symptoms such as, for example, congestive heart failure and hydrocephalus.⁴ Some authors propose a waiting period of at least five to six months, while others prefer a rather early and aggressive treatment.^2,7,8 In our case series, two patients presenting with a DAVF on imaging without prior treatment were older infants (> 24 months). Therefore, we presume that a long waiting period before initial VGM treatment might increase the risk of spontaneous formation of DAVF. Key to the differences of the results in the DAVF group and non-DAVF group is the treatment timing in the non-DAVF group. They were generally treated earlier starting at the latest age of six months. There were no shorter gaps between embolizations in general. The non-DAVF group was rather treated in a similar way.

Choice of embolic material

In 2011, we altered our main embolization technique in treating VGM from using coils solely toward combining coils with Onyx. The technical challenges as well as the angiographic and clinical results concerning the different approaches are beyond the scope of this article. However, only one patient with VGM under treatment at our department developed a DAVF during the course of time. This child was treated by using coils solely at the beginning of the therapeutic course. Only after the additional use of Onyx as embolic material did the formation of a new DAVF become visible. The DAVF developed closely to where an Onyx cast was identified in the distal part of the VGM located at the midline. As we observed this development in only one case a simple coincidence cannot be ruled out. However, taking into consideration the higher frequency of about 30% cases in the previously published series all being treated with transarterial NBCA,³ the use of liquid embolic materials may influence the development of DAVF in VGM. A possible explanation for this phenomenon may be that transarterially delivered NBCA and Onyx are not fully controllable in high-flow arteriovenous shunts. Sometimes the casts partially fill the dilated venous sac in VGM either intentionally or by accident depending on the treatment strategy.^1,8 This may induce focal temporary partial thrombosis on the venous side of the VGM with subsequent recanalization effects leading eventually to the formation of DAVF. This phenomenon may occur depending on the time left treated incompletely. Concerning the differences of the results in the DAVF group and non-DAVF group, it is important to note that patients in the non-DAVF group were treated in a similar way—with coils and Onyx in general, depending on the individual angioarchitecture.

Management implications

Some management implications can be derived from the findings of our study. First, we suggest that a six-vessel angiogram on DSA should be performed on routine follow-up after VGM embolization to detect or rule out an associated DAVF. It has not always been our practice to perform a six-vessel angiogram on presentation in the past. However, the results of our study clearly demonstrate their importance especially in delayed treated cases. Second, the time of treatment should not be set too long, owing to a possible deterioration of the condition while left untreated or incompletely treated. Moreover and in any case, a too proximal occlusion of the arterial feeders without reaching and closing the arteriovenous fistulous connections should be absolutely avoided. Eventually, the aim of VGM treatment should be closure of all fistulous arteriovenous connections. In some cases closure of the VGM leads directly to disappearance of the associated DAVF, while in others the DAVF needs to be treated specifically. These observations concerning the treatment of the associated DAVF in VGM confirm the previously published findings of others.³ Although the use of liquid embolic materials such as NBCA or Onyx in VGM treatment seems to have an impact on DAVF formation, the influence of embolic agents on their development remains unclear and needs further investigation.

Limitations

First of all, the number of cases in our series is low as VGM is a rare disease. In addition, one should consider that some DAVFs were seen on initial angiogram, while others might have been masked by the very high flow in VGM before starting embolization. It cannot be excluded that some DAVF became visible only after successful partial treatment.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.